ICESA 2014 Internatıional Civil Engineering & Architecture Symposium for Academicians

ICESA 2014 Internatıional Civil Engineering & Architecture Symposium for Academicians

TRANSPORTATION ENGINEERING ANALYSIS of SIMPLE DISTANCE (EDSP-BW TYPE) GUARDRAILS with SAP 2000 PROGRAM

1

INVESTIGATION ON ADVANTAGE DISADVANTAGE OF INSTALLATION OF COMPOSITE & PIERS IN HEIGHT

6

THE STUDY OF FOAM BITUMEN MIXTUES IN ADVERSE WEATHER

14

EFFECT OF THE SOLID PYROLYSIS PRODUCT OBTAINED FROM WASTE HARD PLASTIC POLYPROPYLENE ON BITUMINOUS BINDERS

20

SOME MAJOR SIGNALIZED INTERSECTION EXAMINATION IN KONYA BY USING SIDRA INTERSECTION 5 1 SOFTWARE

27

STATEMENT OF TRAFFIC PROBLEMS AND PROPOSALS FOR A CONGESTED AREA WITHIN BAGHDAD CITY

33

A STUDY ON THE RHEOLOGICAL PROPERTIES OF RUBBER MODIFIED ASPHALT MIXTURES

41

MODELING OF TRAVELLING TIME FOR PLANNING VERTICAL PROFILE OF HIGH-SPEED RAILWAYS

55

ANALYZING THE EFFECTS OF NORTHERN MARMARA HIGHWAY AND 3rd BOSPHORUS BRIDGE PROJECT AND 3rd AIRPORT PROJECT ON ISTANBUL

61

MODELLING OF THE EFFECTS OF HYDRATED LIME ADDITIVES ON HOT MIX PAVEMENTS USING A FUZZY LOGIC APPROACH

67

STATISTICAL ANALYSIS OF VEHICLE DELAY MEASUREMENTS CONSIDERING DIFFERENT TIME DURATIONS

75

ASSESSMENT OF QUALITY INDEX FOR SUBWAY WITH CONCRETE SLAB

81

ASSESSMENT OF ANALYTICAL TECHNIQUES OF FLEXIBLE PAVEMENTS

90

A MODEL APPLICATION TOWARDS SOLUTION OF CAR PARKING PROBLEM IN TURKIYE: IZMIR ALSANCAK MULTI-STOREY FULLY AUTOMATED CAR PARK

97

EXAMINING THE USE OF CEBECİ DOLOMITE LIMESTONE, GÖLCÜK SANDSTONE AND KARATEPE BASALT IN STONE MASTIC ASPHALT MIXTURES

104

STUDY ON GROUND TREATMENT FOR SUBWAY SHALLOW TUNNEL CROSSING AN URBAN RIVER BY SHIELD TUNNELING METHOD 110 ASSESSMENT BALLAST LAYER THICKNESS FOR DIFFERENT TRACKS

115

RISK ACCIDENT ON URBAN ROADS RELATED WITH OVERTAKING MANEUVER

121

UTILIZATION OF RFID DATA TO EVALUATE CAMPUS COMMUTE CHARACTERISTICS

127

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

ANALYSIS of SIMPLE DISTANCE (EDSP-BW TYPE) GUARDRAILS with SAP 2000 PROGRAM . Oruç1, A. Z. Dilbero lu2, B. Y lmaz3, M. Bostanc 1

lu4

Assoc. Prof. Dr., Karadeniz Technical University, Faculty of Engineering, Department of Civil Engineering, Trabzon 2 M.Sc. Civil Engineering, Arsin Municipality, Trabzon 3 Res. Assist., Bayburt University, Faculty of Engineering, Department of Civil Engineering, Bayburt, [email protected] 4 Assist. Prof. Dr., Cumhuriyet University, Faculty of Engineering, Department of Civil Engineering, Sivas

ABSTRACT Reducing the loss of life that occurred on highways is the primary objective of relevant institutions and researchers who deals with traffic accidents. Loss of life due to single vehiche which goes of the road is 30% of the total loses in accidents. Guardrail is a protective railing which prevents major accidents. In this study , in order to reduce the losses that may occur in the case of vehicles get out of the road , 12 meters EDSP -BW type guardrail at an angle of 90 º, 1.3 ton and 2 meters were distributed installation. For this load case, deflections of the guardrails are calculated with SAP 2000 Program for 4, 3, 2, 1.33 and 1meters of the strut distance in 0.75 fixed strut height and for 0.75, 0.85, and 0.95 meters of the strut heights in 2 m fixed strut distance, seperately. Keywords: Accident, Deflection, Guardrail, Sap 2000, Traffic, Transportation INTRODUCTION Substructures of the country, main roads, are planned according to the surroundings. Without considering this harmony only main road platform constructing is not only a sub-structural design mistake but generates a big hazard and security concern for the vehicles leaving the main road as well [1]. Traffic accident numbers and severity could be decreased considerably by designing main road constructions according to the traffic security and besides taking precautions on the main roads. Therefore it is getting more important to increase the road security by improving design of road side elements and existing road geometrical specifications [2]. 30% of casualties in all the traffic accidents in our country is due to single vehicle accidents [3]. Table 01. Year 2012 mean single vehicle accident statistics Accident Type The ratio of all accidents (%) Single vehicle– Fixed object collision 10 Single vehicle – Tilting 7 Single vehicle – Get out of the road 13 Total 30 As shown in Table-1 the top most loss of life single car accidents are due to vehicles running of the road. Since main roads are crossing different locations, the objects have shown differences that cars are colliding them when they running of the road. Most of the objects that cause accidents are barriers (guardrails), trees, shoulders, and service polls [4]. Guardrails are protecting vehicles from hazards from road sides and protecting lane invasions due to both vehicles on each side of the refuges, taking back to the road when a car collide and absorbing the effect of collision protecting mechanisms [5]. Since guardrails it selves are also artificial tackles that mounted on road platforms, they must be used where colliding to the obstacle is more dangerous. Below conditions must be evaluated before mounting guardrails [6]. No obstacles could be found in clean road side distance Obstacle cannot be broken or abandon Road side slopes in dangerous rates Vehicle volume dense on the road Traffic accidents high at accident point Guardrails used in Turkey are 7 types: Simple guardrails Simple distance (wedge) guardrails Double sided guardrails Double sided distanced (wedge) guardrails Heavy duty guardrails Steel cable guardrails 1

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Cement guardrails [7]. Between those guardrail types, the most used guardrail is simple distance guardrails. Details of simple distance guardrails details are shown in Figure1. This distance is kept by using 480 mm wedge. Also there is band behind the guardrail that has role of a second crossbeam. Due to this crossbeam, the system becomes more flexible and enduring. The height of the system from the cover is 24,4 kg [8].

Figure 01. Detail of simple distance (wedge) guardrail [9] Crash Tests of Guardrails The crash test is generally to crash a vehicle to the guardrail which its endurance is determined with a specific speed and angle. The vehicles less than 1500 kg are crashed with 50-110 km/s speed and 8-20 degrees angle and the vehicles over 10 tones are crashed with 65-80 km/s and 8-15 degrees angle. The vehicle test of 1500 kg is essential for low and normal obstruction guardrails and the vehicle test of 10 tones, is required for guardrails with high and very high obstruction these tests are used for measuring the endurance of the guardrails [10]. Examining the Crashing Behavior of Guardrails in Virtual Environment Today, with the development of the technology many commercial software like ANSYS, NASTRAN, LS-DYNA, ABAQUS, PATRAN are used in the analysis process of the complex problem. This package software provides easiness in solving the non-linear and dynamic analysis aroused from complex designs actively and in a short time. Due to the commercial software, the correct result of the crashing simulations is obtained without using vehicle and human prototype in virtual environment and by making many analyses. This led economic saving from money and time. The actual crashing tests are actualized by having the highway vehicles of approximately 900 kg crash to the guardrail with a specific angle and with a sped of approximately 100 km/s. During the test results, it is observed that the vehicle stays in its road if the type of the vehicle, crashing speed and angle, lateral deformation of the guardrail and severity level of the crashing is in conformity with EN1317-2 standard [11]. Together with the developing technology, different materials are used in guardrail manufacturing and studies are made on manufactured guardrails. The prototypes of guardrails with fiber supported polymer (FRP) composite materials are developed and these guardrails are subjected to the static and dynamic stress in experimental environment by using computer aided simulation and the analysis software LS-DYNA. As the result, it is observed that it is poor in elastic and plastic deformation and has a lower capacity than the steel guardrails but its energy absorption is better [12]. In the study made by Coon vd.(2005), energy blocker caps that are installed at the end of the guardrails and their crashing behaviors are examined. By considering the deformations that occur after the crashing, a method is suggested for reforming the crashing. For reforming the crashing, rules of preserving momentum and enegery are used. It is emphasized in the study that determining the conditions of actual accidents and estimating the behavior provides decreasing the number and severity of the accidents beside manufacturing more effective systems [13]. Atahan (2008), has researched the effect of height of New Jersey type concrete guardrails to the stability of the vehicle. In this study made by using heavy vehicles of 10 and 30 tones, the optimum guardrail height is determined as 1050 mm by using LS-DYNA simulations. It is observed that the concrete guardrails which are constructed as lower than this height, are insufficient for keeping the heavy vehicles of 30 tones [14]. In an another study, by using W-beamed guardrail system modeling and computer aided analysis software MADYMO (Mathematical Dynamic Model), the crashing of the motorcycle to the guardrail with 45 and 90 degrees angle and with various speeds is simulated. In this software, ASCII text file is formed for both guardrail and motorcycle and it is simulated. As the result of the study, it is emphasized that it is required to distribute the energy aroused from the crashing by using alternative materials in guardrails and to prevent the crashing of those who fall from the motorcycle to the sharp surfaces by changing the mast of the guardrails with smooth surfaces [5]. In this study, it is aimed to examine the EDSP-BW which can be applied for decreasing the losses in case of getting out of the road in terms of deflections aroused in different strut and height.

2

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

MATERIALS AND METHODS In this study, EDSP- BW type (simple distance on bridge) guardrail systems of 12 meters are formed in SAP 2000. In the formed systems three W-rails of 4 meters are connected to each other and these rails are connected to paneled struts with chocks of 480 meters. The purpose of the usage of the paneled struts, is to be fixed to the concrete with anchors and their ability to work like fixed support. The panel struts are fixed to the concrete with 4 units of anchors. This guardrail system is generally used in cases which scintillation cannot be made to concrete pavements, on bridges or viaducts. The stay beams that are connected from the rear sides of wedges, are the parts of 4 meters like W-rails and has the duty as a second beam. The EDSP-BW type guardrails are shown in Figure 2.

Figure 02. EDSP-BW type (simple distance on bridge) guardrails The stress is loaded with 2 meters and 1,3 ton of distribution to EDSP-BW type guardrails of 12 meters that is modeled in SAP2000 program with 90º angle. The situation that the distributed load is placed to guardrails with gaps of 4 meters, is shown in Figure-3.

Figure 03. Situation that the 2 meters distributed load is placed to guardrails with strut distance of 4 meters The deflections made by the system by changing the strut distance and heights, are compared. RESULTS With the aid of SAP2000 program, EDSP type guardrails are designed as to have strut distance with 4, 3, 2, 1.33, and 1 meters and according to the change in strut distance and heights, the deflections occurred in these systems, are determined separately. Comparison of The Deflections of Guardrails in Changing of Strut Distance For the guardrails with the strut height of 0,75, the loading of 1,3 ton is made separately as 4 m, 3 m, 2 meters, 1,33 meters and 1 meter strut distance. Figure 4 shows how the deflections change as a result of the loadings in case the strut distances change.

3

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Deflections 0,025 0,02104

0,020

0,01421

0,015

0,01042 0,01006

0,010

0,00662

0,005 0,000 4,00 m

3,00 m

2,00 m Strut Distances

1,33 m

1,00 m

Figure 04. With the change of the erection gaps, changes in deflections made by the system in erection height of 0,75 meter [15]. Today, in EDSP type guardrails, the systems with strut distance bigger than 2.0 meters, are not used. If the EDSP-BW type guardrails are used on bridges or viaducts, considering the strut distance of 3,00 and 4,00 for these systems, can be risky. When the strut is made in guardrails with distance of 2,00 meters, the system causes deflections of 0,01042 meter. When we decrease the strut distance to 1,33 meters, the maximum deflection in system, decreases to 0,01006 meters. However, there is not much difference between the strut distance system of 2,00 meters and strut distance system of 1,33 meters. When the strut distance are decreased to 1,00 meters from 1,33, a decrease of %3,45 is observed. But when the strut distance is decreased to 1,00 meters from 1,33, it is observed that the decrease in system deflection is % 34.1 . In other words, to decrease the strut distance from 2,00 meters to 1,33 meters, shall not decrease the system deflection much. If the location has the possibility to cause dangerous results in case vehicles leave the road platform, strut distance with 1 meters may be used. If there is not such a case, it is not required to be used. Cause to the poor deflection of strut system of 1 meters, physical injury shall be increased in case of the crashing of the vehicles to this system. This system can be applied on suspension bridges, very high viaducts or at the edges of steep slopes. But for applying this system, the surface must be concrete pavement or concrete beam. For sure these evaluations can be made for vehicles of 1,3 tons. If the heavy vehicles are dense in traffic, it is not right to use this system. In that case, heavy service guardrails shall be used in same anchored system. In Case of The Change of Strut Heights, Comparison of The Deflections Caused by Guardrails The loading of 1,3 tons, are made separately in strut heights of 0,75 meters, 0,85 meters and 0,95 meters for the strut distances guardrails. If the strut heights change, the deflection amounts according to the loading are shown in Figure 5. Deflections 0,015

0,01042

0,01066

0,01071

0,010 0,005 0,000 0.75 m

0.85 m Strut Heights

0.95 m

Figure 05. With the change of the strut height, changes in deflections made by the system in strut distance of 2 meter [15]. In our country and foreign countries, the height of the EDSP type guardrails is 0,75 meters. 0,85 meters and 0,95 meters EDSP type guardrails systems aren’t used. The EDSP type guardrails is designed as 0,75 meters, 0,85 meters and 0,95 meters with the help of SAP 2000 program and the deflections are determined. The changes in the deflections with the change of the strut heights, can be poorly realized. The deflection caused by the changing of strut height, is 0,01042. When the height of the system is increased to 0,85 meters, the deflection of the 4

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

system is 0,01066. When we increase the height of the system from 0,75 meters to 0,85 meter, an increase of % 0,46 is seen in deflection. With the change in the height of the system, the change in the deflection of the system is poor. If the system has not been fixed to the concrete with anchors, an important change would occur when the strut heights are changed. But in this system, when the strut heights are changed, there are not much changes in the deflection made by the system. DISCUSSION As the result of the study, when EDSP type guardrails are modeled with 1,00 meters strut distances, it is observed that the deflection they caused, have decreased. It is observed that the locations where flexible guardrails and dense vehicle traffic can cause problems, the guardrails with strut distance of 1,00 meters can be used. The guardrails with strut distance of 1,00 meters never used in our country and in foreign countries. This system can be used in our country by testing with actual crashing experiments. But first the test for actual crashing shall be made and a specification shall be formed for these guardrails according to the result of these tests. The system has been fixed to the concrete with anchors, so there has not been such change in the deflection of the system with the change of the height of strut distances. At the result of the studies made by changing the strut heights of the EDSP type guardrails, major changes can be observed in the deflections occurred in the system. REFERENCES [1] Atahan, A.O and Ross, H.E., (2002). Computer Simulation of Recycled Content Guardrail Post, Journal of Transportation Engineering, ASCE. [2] Mutlu, O., (2010). Metrobüs Güzergah nda Kullan lan Halatl Otokorkuluklar n ncelenmesi ve Alterantif Sistemlerin Güvenlik Dayan n Belirlenmesi, Yüksek Lisans Tezi, Bahçe ehir Üniversitesi, Fen Bilimler Enstitüsü, stanbul. [3] Karayollar Genel Müdürlü ü, (2013). Trafik Kaza Özeti 2012, Ankara. [4] Devlet statistik Enstitüsü, 1996-2000 Y llar Aras ndaki Kaza statistikleri, Ankara. [5] Ibitoye, A.B., Hamouda A.M.S.,Wong, S.V. and Radin, R.S., (2004). Simulation of Motorcyclist’s Kinematics During Impact With W-Beam Guardrail, Advances in Engineering Software, 37, 56-61. [6] AASHTO, (1996). A Standardized Guide to Highway Barrier Hardware, Washington D.C. [7] Karayollar Genel Müdürlü ü, (2006). Karayollar Teknik artnamesi, Ankara. [8] European Comittee for Standardization, 1998. Road Restraint Systems-Part 2, Performans Classes Impact Test Acceptance Criteria and Test Methods for Safety Barriers European Standard EN-1317-2. [9] http://www.alkagroup.com.tr/EN/Highway_Safety_System.php, 5Ekim 2010. [10] Plaxico, C.A., Ray, M.H. and H ranmayee, K., (2000). Comparison of the Impact Performance of the G4(1W) and G4(2W) Guardrial Systems Under NCHRP Report 350 Test 3-11 Conditions, Transportation Research Record, 000525,Washington D.C. [11] Atahan, A.O. and Cansiz, O.F., (2005). Improvements to G4(RW) Strong-Post Round-Wood W-Beam Guardrail, Journal of Transportation Engineering, 131, 63-73. [12] Bligh, R.P., Menges, W.L. and Alberson, D.C., (2001). Testing and Evaluation of Recycled Materials in Roadside Safety Devices, Texas Transportation Institute, Research Report: 1458-3. [13] Coon, B.A., Reid, J.D. and Rohde, J.R., (2005). Dynamic Impact Testing of Guardrail Posts Embedded in Soil, Transportation Research Report No: TRP-03-77-98, Midwest Roadside Safety Facility, Nebraska. [14] Atahan, A.O., (2008). Effect of Permanent Jersey-Shaped Concrete Barrier Height on Heavy Vehicle PostImpact Stability, Int. J. Heavy Vehicle Systems, 16, 1/2. [15] Dilbero lu, A.Z., (2011).Karayollar nda Otokorkuluklar n Kazalar n Önlenmesindeki Önemi, Yüksek Lisans Tezi, KTU, Fen Bilimler Enstitüsü, Trabzon.

5

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

INVESTIGATION ON ADVANTAGE DISADVANTAGE OF INSTALLATION OF COMPOSITE & PIERS IN HEIGHT S. Nobakht1, S.-M.-R. Fakhrefatemi2, S. Rezaei3, M. Khordehbinan4 [email protected] 1

Faculty member, Payame Noor University, Tehran, Iran. M.Sc. of civil engineering, Payame Noor University, Tehran, Iran. 3 Faculty member, Pooyesh Institute of Higher Education, Qom, Iran. 4 Department of Civil Engineering, Payame Noor University, Tehran, Iran. 2

ABSTRACT Authentic management of bridge from initial states of design to maintenance levels plays an important role in qualification of vital infrastructures. Hence, fewest attentions in this regards, can result in saving national capitals. Previously, most of the concentrations were on the design of bridge's deck and its material. Besides, it was prevalent to assume a single concrete pier for bridges just for facility in design. In this research, four different models of substructure were analyzed and put in comparison. In other words, Life Cycle Cost Analysis was calculated for these four models. The three first models are included Single concrete pier, Moment Concrete frame and braced Steel frame. The main aim of this study is investigation on fourth model -Composite column in height- during the three levels of design, installation and maintenance. Based on the results, composition of steel Pier with concrete ones would provide an optimal range of costs and time demand. Therefore, with using composite substructures, designers are able to manage time and cost more effectiveness. This research is novel from the view point that it is the first time that composite pier of bridge in height are employed for analysis. Keywords: Pier of Bridge, Composite pier in height, Cost of bridge's construction, Costs of bridge's maintenance, Life Cycle Cost Analysis. INTRODUCTION One of the most effective performances by a road construction engineer is replacing old systems with the novel methods so that productivity is increased, industry is improved and costs are reduced. As it is clear, road construction accounts for a considerable share of national financial resources and thus, cost-effective methods provision allows allocating more cost to other national development sectors. By investigating the material and resources in great construction industry, structure systems of the bridge pier can be classified into steel and concrete types. Formerly, when use of steel was not as common as today, engineers used cast iron for construction. The first cast iron bridge was constructed in 1779 in England [1]. It was named Coalbrookdale which passes Severn River with a span length of 30.5 m. Although vehicles do not pass it anymore; it is still standing strong after more than 200 years. The idea for using Cast Iron Bridge was due to its light weight and high strength compared to its dimensions which allowed designing spans with longer length. At the second half of the nineteenth century, with the development of steam machines and equipment for melting, steel could replace pre-made iron (cast iron) because of its higher resistance characteristics and various steel bridges were built during 1855 to 1859 in order to develop the rail industry across Europe and America [1]. Since then, considering different conditions, concrete and steel are widely used always in bridge construction as two famous options of Steel and Concrete with their special characteristics. Today designers utilize recent advancements in molding industry and are able to create and install any unique type of the shapes. It is widely observed in both steel industry and the concrete placement. Following these developments, all sources and references for designing concrete and steel bridges [1-7] and also bridges built during last 150 years were studied and thus structure systems of the bridge pier were summarized in Table 1.

6

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Pier Type Multiple-Pier of reinforced concrete

Table 01. Types of structure systems used in the bridge pier Used Usage Case Material Reinforced concrete

Single-pier of reinforced concrete

Reinforced concrete

Integrated caps

Concrete or steel

Wall-shaped pier

Reinforced concrete

Stands

Precast concrete or steel piles

It is widely used in different conditions This type of pier is particularly used for low-width bridges where there is not adequate space for placing two or more Pier. Also long pier are considered in which using one Pier with larger dimensions is more preferred. Hammer tip shaped pier are regarded as a solution in the locations where water current may lead to obstruction of the space between Pier of a multiple-Pier pier It is sometimes tended to structure pier caps as integrated form with the superstructure construct. It leads to improved structure efficiency and sometimes it is considered in terms of aesthetics Wall-shaped pier are usually used in such locations where use of multi-Pier pier is problematic due to probability of accumulating garbage between Pier resulting from water current, or effective design for collision force is needed Stands are often used in wet lands or in the water, where using individual pile and pier is not cost-effective

One of the usual situations for use of steel pier is construction of temporary bridges such as temporary access bridges in construction sites. In such cases, contractors usually prefer using steel substructure due to light weight, ease of use, installation Steel profiles Steel speed and reuse potential. Also, many steel bridges with integrated tilted steel pier or steel delta-shaped pier have been built so that complicated design problems in challenging sites are overcome. Reviewing common pier types in references and books indicate that Composite Pier of bridge in Height has not been used. After ensuring this fact, studies for identifying this system's advantages and disadvantages in bridge substructure was initiated. Definition of Composite Piers in Height Composite pier in height is a pier that a ratio of its height is composed with reinforced concrete, and steel sections are placed in its upper part. Ratio of 0.1 to 0.6 for replacement of the section with concrete was modeled in this study. For instance, sixth tenth of the Pier's height is made of reinforced concrete in a combined Pier with ratio of 0.6, and the remaining is made of steel. METHODOLOGY Modeling Method Case study was integrated with real condition of a bridge. The name of the bridge is Forth Bridge located in the East of Scotland and in the western region of Edinburgh [9-8]. Four basic models for bridge pier include: (1) concrete singlepier, (2) Concrete moment frame, (3) Steel braced frame, and (4) Composite Pier in Height which is investigated with 6 different replacement ratios (from 0.1 to 0.6). In addition, high early strength concrete type and normal concrete type were used in concrete sections [10]. Diagram of Fig. 4 indicates this classification. In order to obtain proper dimensions of section of the piers, all four piers were modeled in SAP2000 Software version 14, 2, 2. Spectral analysis was conducted on four models based on Iran Regulation 463 and design was carried out based on optimal sections [11-17]. Tables 2 and 3 give assumptions relating to the design parameters and final sections,respectively. Concrete singlepillar Concrete

types of bridge pier

Steel

Concrete moment frame Ordinary concrete

Composite

Ordinary concrete High early strength concrete ordinary concrete High early strength concrete

High early strength concrete

Figure 04. Types of pier in terms of the material used in this research 7

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Table 02. Spectral design parameters Parameter Definition Regulations 463 Soil Type IV Earthquake direction In line with bridge width Dead load 650 tons Live load 113 tons Height 48 m

Type Pier dimensions Beam dimensions Brace Dimensions

Table 03. Characteristics of pier sections Concrete singleConcrete moment Steel braced frame pier frame 1500 × 2300 mm 800×800 mm Web 500, Flange 500 mm Pier head: Web 850, Flange 500 × 700 mm 700 mm 2 UNP: Web 200, Flange 100 mm

Cost Calculation Following modeling and specifying the dimensions of sections, calculations can be done in three stages: 1. construction cost, 2. Installation costs, and 3. Maintenance costs. Three stages are described in the following: Construction Cost Construction costs was estimated based on the basic cost list for Iran's buildings and toll in 2013. Installation Cost Following identifying methods of installation on the water, required equipment were categorized as Table 4. This table includes initial estimation of costs and it should be updated in any new region and conditions. Table 04. Installation services costs on the water in 2013 Service Type Daily Cost (Dollar) Crane vessel (barge) 377.4 Ship carrying Truck Mixer and Concrete Pump 377.4 Armature team 56.6 concrete placement team in special circumstances 188.7 supervision team in special circumstances 79.2 surveyor team in special circumstances 52.8 Health and Safety Executives (HSE) 52.8 Installation cost calculation method was modeled according to the following equation: 1 1 1 1 Cost = F×H [ ( Crane)+( Transport)+( Concrete)+( Bar reinforcement)+( Personnel)] F F F F Where, (F) is time factor and (H) is height of the bridge’s pier. F has determined based on Iranian Concrete Code (ABA). Thus, F coefficient for normal concrete type and high early strength concrete type is and , respectively. Followings are times for installation of the piers in this work where the height is 48 m: 5 t = F×48 = 2 ×48=120 days, in case using normal concrete type; 3

t = F×48 = 2 ×48=72 days, in case using high early strength concrete; t = 3 days, in case using steel braced frame with any height. Maintenance Cost Following investigating available facilities and identifying common methods for corrosion protection, three methods were provided for 50-year life time of this bridge and applied in calculations: 1. Epoxy color is used every 20 years for concrete elements. It should be noted that pozzolan of Silica fume has been employed in the construction of concrete pier in order to reduce water absorption [18]. 2. Two methods were considered for steel elements: type I method is repeated every 5 years, and after sandblasting, the 50-micron-thick oil color is used, while in type II method, which is conducted once every 10 years, after sandblasting, epoxy paint with 100 micron total thickness is applied. In order to calculate maintenance costs, calculation instruction provided by America’s Federal Road Department and Engineering Economics formulas were used. Annual inflation rate was considered as 20 percent, which seems logical 8

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Composite frame in height

Composite pier in height

currently due to economic instability in Iran. Five-year term deposit is used to cover the costs of maintenance [19]. RESULTS Stages for selecting the best structure system are as follows: Step 1: results for construction costs calculation Table 5 indicates construction costs for steel braced frame is higher than concrete single-pier by 1.8 times and higher than concrete moment frame by 2.7 times regardless of the costs for pier building. This trend becomes descending by replacing steel with concrete section, and it is directed toward a cheaper construct. Weight of the pier made of concrete single-pier is 11 times higher than Steel braced frame and it is about 2 times higher than concrete moment frame. Replacing concrete with steel section reduces the structure weight. Significant difference between weights of steel structure and concrete structure impose considerable costs for pier building sector. However, after summing pier costs, steel braced frame is still more expensive than other models. Table 05. Construction cost of bridge pier with different sections (Dollar) Weight Construction cost Pier cost Final cost Section (ton) (Dollar) (Dollar) (Dollar) Steel braced frame 37.4 56,071 37,068 93,139 0.1 height: concrete single-pier 75.1 53,564 38,695 92,259 0.2 height: concrete single-pier 112.7 51,058 40,322 91,380 0.3 height: concrete single-pier 150.4 48,551 41,949 90,500 0.4 height: concrete single-pier 188.0 46,044 43,576 89,620 0.5 height: concrete single-pier 225.7 43,538 45,203 88,741 0.6 height: concrete single-pier 263.4 41,031 46,830 87,861 concrete single-pier 414.0 31,004 53,338 84,342 0.1 height: concrete moment 52.9 52,526 37,806 90,332 frame 0.2 height: concrete moment 68.3 48,982 38,544 87,526 frame 0.3 height: concrete moment 83.8 45,437 39,282 84,719 frame 0.4 height: concrete moment 99.3 41,892 40,020 81,912 frame 0.5 height: concrete moment 114.8 38,348 40,758 79,106 frame 0.6 height: concrete moment 130.2 34,803 41,496 76,299 frame concrete moment frame 192.1 20,624 44,448 65,073 Step 2: time and installation cost calculation results Fig 5 indicates that 120 days are needed for construction of a pier with 48 m height and using normal concrete. If high early strength concrete is used, this time is reduced to 72 days which is a little more than its half. If steel structure is installed, 117 days are saved in time. Any time composite pier in height is employed, time saving would be considerable. For example, with construction of a Pier that 0.6 of its height is made of normal concrete, 45 days (= 120 – 75) is saved in time.

Time for Installation of 48 meters Pier 140

120

Time (Day)

120 100 80 60 40 20

3

3

10,2 15

17,4

27

39 24,6

51 31,8

75

63 39

72

46,2

0 Braced Steel 0.1 height = 0.2 height = 0.3 height = 0.4 height = 0.5 height = 0.6 height = Full Concrete Frame Concrete Concrete Concrete Concrete Concrete Concrete Pier

Section type Early Concrete Strength Normal Concrete

Figure 05. Bar diagram for the number of days needed for installing pier 9

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

The Diagram in Fig 6 indicates that installing steel structure is cheaper than other concrete constructs on water due to higher speed and reduced time-dependent costs. Additionally, using high early strength concrete in concrete structure allows cost reduction by 8.9 thousands of dollars.

Installation Cost 50.000

Cost ($)

40.000 30.000 20.000 10.000 0 Braced Steel 0.1 height = 0.2 height = 0.3 height = 0.4 height = 0.5 height = 0.6 height = Frame Concrete Concrete Concrete Concrete Concrete Concrete Using Normal Concrete

Section type

Full Concrete Pier

Using Early Concrete Strength

Figure 06. Installation costs diagram (Dollar) Step 3: time and maintenance cost calculation results Table 6 indicates summary of maintenance costs. This table shows that using type II maintenance method or application of epoxy color every 10 years is cheaper than type I method in steel constructs. In addition, this table shows that maintenance costs for concrete constructs is much less than steel constructs and its reason is higher durability and no need of concrete structure to sandblast in wide area. Table 06. Summary of maintenance costs Maintenance costs (Dollar) Pier height: 48 m

Concrete singlepier + type I steel frame

Concrete singlepier + type II steel frame

Concrete moment frame + type I steel frame

Concrete moment frame + type II steel frame

Steel braced frame

15,471

13,358

15,471

13,358

0.1 height: concrete section

14,173

12,271

14,205

12,303

0.2 height: concrete section

12,874

11,184

12,940

11,249

0.3 height: concrete section

11,576

10,097

11,674

10,195

0.4 height: concrete section

10,278

9,010

10,409

9,141

0.5 height: concrete section

8,980

7,923

9,144

8,087

0.6 height: concrete section

7,682

6,836

7,878

7,033

Full-concrete pier

2,489

2,489

2,816

2,816

DISCUSSION Following calculating costs related to three mentioned steps in Results section, now all costs are summed and provided in the form of 6 models. Final total cost for each model includes sum of construction, installation and maintenance costs. Given the diagrams resulting from the final total cost calculations in Fig 7 and Fig 8, it can be stated that concrete structure in concrete single-pier form (in high early strength concrete and normal concrete manner) is more expensive than steel braced frame with type II maintenance method. Moment frame with normal concrete has similar maintenance costs compared to steel braced frame with type I maintenance method. It is logically acceptable if time factor is not important for the contractor. Using diagrams in Fig 8, it is possible to compare time factor. These diagrams indicate that in the first place, using moment frame with high early strength concrete has the top place in terms of economic matter. Moment frame of composite pier in height and steel braced frame with type II maintenance method are in the second and third places, respectively.

10

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Curves of Final Cost 140.000 Cost ($)

130.000 120.000 110.000 100.000 90.000

Braced Steel Frame

0.1 height = Concrete

0.2 height = Concrete

0.3 height = Concrete

0.4 height = Concrete

0.5 height = Concrete

0.6 height = Full Concrete Concrete Pier

Section type Single Normal Concrete Pier

Single Early Concrete Strength Pier

Moment Frame Normal Concrete

Moment Frame Early Concrete Strength

Braced Steel Frame I

Braced Steel Frame II

Figure 07. Final costs diagram (Dollar)

Curves of Final Cost - Time Cost ($)

140.000 130.000 120.000 110.000 100.000 90.000 0

10

20

30

40

50

60

70

80

90

100

110

120

130

Time (day) Single Normal Concrete Pier

Single Early Concrete Strength Pier

Moment Frame Normal Concrete

Moment Frame Early Concrete Strength

Braced Steel Frame I

Braced Steel Frame II

Figure 08. Cost- time diagram By selecting the plan of concrete moment frame with high early strength concrete it can be concluded that daily time purchase cost is: (cost of steel braced frame cost of moment frame with high early strength ) = 100.5 $ 72 3 This equation suggests that if the project manager is going to shorten the project time for, say, 28 days, he should use Composite Pier that 0.6 of the height of them are made of concrete moment frame with high early strength concrete and the remaining are made of the steel braced frame. Cost of this time saving approximately would be: 100.5×28= 2,814$ Finally, diagram in Fig 9 allows calculation of purchase cost in different times with interpolation. Vertical axis shows difference of costs of composite pier in height in comparison with steel braced frame.

11

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Time purchase Curve

8.000 7.000 6.000

Cost ($)

5.000 4.000 3.000 2.000 1.000 0 0

5

10 15 20 25 30 35 40 45 50 55 60 65 70 75 80

Time (day)

Composite Pier…

Figure 09. Time purchase curve in Composite Pier in $ (concrete moment frame with high early strength concrete + steel braced frame) CONCLUSION One type of material was used earlier in designing bridge pier so that mostly steel or concrete were used in bridge pier. By investigating Composite Pier in Hight, following findings were obtained in the current work: 1. Regarding construction costs, construction cost for steel braced frame is higher than concrete single-pier by 1.8 times and higher than moment frames by 2.7 times regardless of the costs for pier foundation. Weight of the pier made of concrete single-pier is 11 times higher than Steel braced frame and it is about 2 times higher than concrete moment frame. It leads to increase costs of pier foundation. Replacing concrete with steel section (using composite pier in height) leads to reduced structure weight. 2. In structure installation part, it should be mentioned that costs for installing concrete structure on the water are much higher than steel structure due to spending much more time. Installing steel pier, concrete pier with high early strength concrete and normal concrete would last 3, 72, and 120 days, respectively. Thus, wherever concrete pier is used, using cement or known high early strength additives are strongly recommended. In addition, using high early strength concrete allows reduction of the costs in full-concrete structure by 8.9 thousands of Dollars. 3. In the final stage which is related to selection of the best model, it can be stated that moment frame with high early strength concrete is in the first place in terms of economic matters. Type II steel braced frame is in the first place in terms of time saving. However, the important point is that always there are not optimal financial or temporal conditions. The plan is top plan which can have both conditions together. Thus, using composite pier in heights allows reduction of installation time compared to concrete moment frame and also has lower construction cost compared to steel braced frame. Hence, composite pier in height may have superior place compared to concrete moment frame or steel braced frame depending design conditions. REFERENCES [1] Chatterjee, Sukhen, 2003, "The Design of Modern Steel Bridges", Second edition. [2] Virginia Department of Transportation Bridge Manuals, Volume V, Part 2, Design Aids Typical Details. [3] FHWA, "Steel Bridge Design Handbook: Substructure Design" -IF-12-052-Vol. 16. [4] The Scottish Office bridge management procedures, 2005, Handbook of Bridge management. [5] Department of Transportation, England. [6] Association of Sea Ways, England. [7] National Bridge Management of Italy. [8] Keay, John; Keay, Julie, 2000, Collins Encyclopedia of Scotland. London: HarperCollins. p. 409. [9] The Gazetteer for Scotland, 2006, "Overview of Forth Bridge". [10] Fakhrfatemi, S.A. (2013), Reviewing performance of combination of metal structure on concrete Pier in elevated road bridges, MSc Thesis, Payam Noor University (PNU), North Tehran, Tehran. [11] M.J.N. Priestly, Displacement-Based Seismic Design of Structures. [12] SAP2000 Basic Analysis Reference Manual: Linear and nonlinear Static and Dynamic Analysis and Design of Three-Dimensional Structures. [13] Khairuddin, A. and Hemmati, A. (2003), Determining the optimal location of transitional floor in combined elevated buildings, Fourth International Conference on Seismology and Earthquake Engineering, Tehran. 12

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

[14] Tahoony, S.H. (2004), Designing Bridge (reinforced concrete, steel and pre-stressed concrete bridges). [15] Iranian steel bridge design guidelines, publication No. 395, strategic and planning department of presidency. [16] Reinforced concrete bridge design and calculation regulations in Iran, Journal No. 389, strategic and planning department of presidency. [17] Road and rail bridge design regulations against earthquake, Journal No. 463, strategic and planning department of presidency. [18] Fakhretaha, S., (2013), "Properties of Self Consolidating Concrete containing ternary blends of Silica fume and natural pozzolan", Construction Material Institute. [19] Federal Highway Administration, 2005, Life cycle Cost Analysis of Bridges.

13

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

THE STUDY OF FOAM BITUMEN MIXTUES IN ADVERSE WEATHER R. Bagheri1, S. Rezaei2, M.-M. S.-Hashemi3, M. Khordehbinan4 [email protected] 1

M.Sc. of civil engineering, Payame Noor University, Tehran, Iran. Faculty member, Pooyesh Institute of Higher Education, Qom, Iran. 3 Instructor, Engineering faculty, Islamic Azad University of Germi, Ardabil, Iran 4 Department of Civil Engineering, Payame Noor University, Tehran, Iran. 2

ABSTRACT In situ cold recycling with foamed asphalt technology is an appropriate method for reconstruction of deteriorated roads. However, the main problem is lack of comprehensive guideline for this method including the essential points for its application in cold weather condition and its strength drop in moisture especially in cold climates. In this method, designing is usually done based on Marshall Strength, indirect tensile strength and unlimited compressive strength. Its moisture sensitivity is measured through TSR (Tensile Strength Ratio) and for reducing strength drop, filler is used, however, this design method is not responsive to cold weather. In this study, it is tried to investigate the status of foamed asphalt mixtures for cold mountainous weather using micro-fillers. The effect of micro-fillers with different percentages on foamed asphalt mixture has been investigated concerning adverse weather condition of Iran which has not been investigated so far. The results of studies show that the type and percentage of micro-fillers can affect the results to a great extent. Keywords: Foamed asphalt mixture, Micro-fillers, Indirect tensile strength, Marshall strength, Adverse weather. INTRODUCTION Foamed asphalt technology has been recently considered by researchers due to high strength and being economic. This technology has been evaluated as appropriate for tropical and temperate climates and has been used in most projects of tropical and temperate climates such as Iran, South Africa and Australia, etc. However, for cold climates no instruction has been given due to the presence of high porosity and lack of coarse aggregates with asphalt which have high strength loss in freezing- thawing condition in this technology. One of the moisture sensitivity reduction parameters in foamed asphalt mixtures is filler. Castedo and et al (1983) found that additives like lime reduce the moisture sensitivity of mixture [1]. However, in recent years, with advances in technology, micro-fillers have been produced which can be used for filling very fine spaces of the mixture and help better distribution of asphalt in mixture. Thus, in this project, micro-fillers (which have not been evaluated in foamed asphalt mixture) with cement, and America Standard (AASHTO T 283), intended for hot asphalt mixture (the porosity of hot asphalt mixture is less than foamed asphalt mixture) and not recommended for foamed asphalt mixture, have been used to evaluate the specification of the mixture in adverse weather to have necessary precision in selection of optimum design. In addition to showing the effect of micro-fillers on freezing- thawing, this study provides the possibility of avoiding early destruction of foamed asphalt by consideration of adverse weather for cold climates. Particles smaller than 0.075 mm are called filler and it is usually classified intro natural and active filler categories. Natural fillers are those which are naturally available in gradation and just act as filler, however, active filler are fillers like cement, lime, fly ash etc. with hydration reaction which are usually added to materials for two reasons: a. in case of lack of natural filler, b. for improving some required specifications in the mixture like increase of strength. Asi (2001) obtained significant improvement in the mixture specification by adding 2% cement to Sebkha soil [2]. For preparing recycling mixtures with foamed asphalt, a sufficient amount of filler should be used. If filler is not sufficient, compressive and indirect tensile strengths reduce [3]. Iran Specifications and Performance of Cold Reclaimed Asphalt Pavement (Iran Publication 339) limited the maximum consuming cement to 2% of sample weight and recommends using 1 to 2 percent cement as desired percentages [4]. Halles and et al. (2009) carried out some experiments and showed that cement as filler has better performance than cement, lime and fly ash. Furthermore, fly ash acts as filler the same as mineral (natural) filler [5]. Iwa skia and et al. (2013) showed that in 2.5% asphalt, the moisture sensitivity of mixture significantly reduces by consuming 2% cement [6]. Use of 3% limes in foamed asphalt mixture leads to increase of porosity of mixture since by increase of consuming lime, the fine aggregates sticks to other materials before acting as asphalt carrier so the amount of fine aggregates reduces and porosity between coarse aggregates increases [7]. 14

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

It is possible to produce cold mixture or foamed asphalt mixture from moisture aggregates [8]. Foamed asphalt can be used for stabilization of various rock materials and materials obtained from available Reclaimed Asphalt Pavement (RAP) and is limiting factor of materials’ gradation which if required, it can be modified with appropriate materials. According to Iran regulations, it is possible to add 25% new materials to reclaimed materials to obtain desired gradation [4]. For stabilization with foamed asphalt to be successful, continuous gradation with proper fine aggregates percentage (at least 5% filler passed from sieve number 200) is required [4]. The proposed range by South Africa Asphalt Academy has been regulated based on minimizing the empty space between aggregates which is shown in figure (1). Producing foamed asphalt through this gradation leads to obtaining a mixture with optimum specifications [8].

Figure 01. Proposed range of Asphalt Academy of South Africa [8] Xu and et al (2012) concluded that if foamed asphalt mixture is prepared with proper moisture and then compacted, it could cover the stone granular with size up to 1.18 mm with asphalt [9]. Investigations of tests As previously mentioned, in foamed asphalt mixture, uniaxial compressive strength, indirect tensile strength and Marshall strength are used. Imagine the foamed asphalt mixture prepared through above method is used in two cities with different weather conditions (e.g. Ardebil and Ahwaz in Iran) with proper asphalt for each city. Would the performance of this plan be the same in these cities while Ardebil is located in mountainous climate with annual temperature of -30 to +30 and Ahwaz with temperature of 50°C in summers, which rarely experience the temperature below 10 °C? The answer to this question is certainly nay and this issue would be more outstanding when it is argued that in foamed asphalt mixture design, the temperature effect is not considered as dry and saturated with uniaxial compressive strength, Marshal Strength and indirect tensile strength and for both weather conditions, one percent asphalt is determined. Nowadays, the researchers don’t just use TSR criterion in various climates and weathers for sensitivity of foamed asphalt mixture. In addition to the mentioned tests, Iwa skia and et al. (2011) investigated the effects of cold for investigation of foamed asphalt mixture and its comparison with asphalt emulation mixture for Poland weather which is considered cold climate based on AASHTO T 283. Their results indicate that according to Marshall, ITS and TSR tests (according to Iran publication 339, the minimum proper TSR for moisture weather is 0.7), the minimum value of Guideline is fulfilled in 2% asphalt [10]. (1) TSR= ITSD/ ITSs 0.7 Where, ITSD is indirect tensile strength of saturated sample, and, ITSS is the indirect tensile strength of dry sample. However, in refer to AASHTO T 283, (it worth noting that AASHTO T 283 has been developed for hot asphalt mixture which has less porosity than foamed asphalt mixture), according to the figure, the strength reduces in processing in cold rather than processing in saturation and with a bit negligence in 2.5% asphalt, the least values of guideline are fulfilled [7]. Therefore 0.5% deficiency of asphalt leads to reduction of strength in cold climates. It is possible to obtain the ratio of indirect tensile strength in freezing state to dry state (TSRc) through equation (2). According to Judycki and et al. (1997) and Tunnicliff and et al. (1981), the mixtures prepared by traditional technology are more resistive against freezing when TSRC is bigger than 0.7 [11 and 12]. (2) TSRc= ITSC/ ITSA 0.7 15

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Where, ITSC is indirect tensile strength of processed sample in freezing-thawing process, and, ITSA is the indirect tensile strength of processed dry sample. MATERIALS AND METHODS The selection of consumi ng materials Since, the objective of this project was to investigate the performance of foamed asphalt and micro-fillers in cold climates, the aggregated were prepared from RAP of Vahdat Street of Ardebil. The status quo was investigated without adding new materials and with separating aggregates bigger than 19 mm. for investigation of fillers’ performance in freezing and thawing condition, three kinds of filler including cement, micro-silica and micro-lime were used according to AASHTO T 283. The gradation of RAP materials is according to figure (2).

Figure 02. Gradation of consumed materials graph The selection of asphalt and the specifications of foamed asphalt 85/100 asphalt (used in Iran for cold climate) has been used with softness point of 49 degree Celsius. For evaluation of foamed asphalt specifications including half-life (HL) and expansion rate (ER), first hot asphalt (usually in 160, 170 and 180 degree) was transformed to foam by various percentages of water (with 1, 2, 3 and 4 weight percent) and the above parameters were measured and recorded in the special expansion case. Then, the graph representing parameters’ variation with variation of water percentage was drawn and the crossing point of two graphs was used for optimization of both properties as proper percent of water for production of foamed asphalt and controlled with the least recommended. In this project, after the above tests, the specifications of foamed asphalt used are as table (1).

Water percent (%) 2.7

Table 01. The specification of foamed asphalt in WLB10 Asphalt temperature Water pressure ER HL(s) (°C) (atmosphere) 170 12 18 8

Air pressure (atmosphere) 7

The preparation of sample and performing tests In respect to 2 different percent of foamed asphalt (2% and 3% asphalt), the cylindrical samples were prepared with one gradation type through Marshall method according to ASTM D 1559 instruction. Since the foamed asphalt mixtures are usually considered for heavy traffic, 75 strokes were considered in each side for compression. Moreover three kinds of filler including cement, micro-silica and micro-lime were individually evaluated with different combinations according to table (2).

Gradation

Constant

Table 02. Kinds of samples’ combination Asphalt percentage Combinations Kind of combination A Cement 1.5% Micro-silica 0. 33% B 2 C Micro-lime 1.5% D 0.1% micro-silica+ 1.4% cement 3 E 0.75% micro-lime+ 0.75% cement 16

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

According to [4 and 13], about 80% of optimum moisture of mixture obtained from proctor test was considered for greasing the mixture and separating fine aggregates for better combination with asphalt; transfer and dispersion of asphalt drops was considered in mixage as 6.2%. 24 hours after sample preparation, they were removed from mold and processed in two states: a) dry processing where materials were placed in 40° Celsius for 72 hours, b) freezing- thawing state where materials were kept in +60°C for 72 hours. Then it was saturated with desiccator (in vacuum), then freezing process was done in -16° for 16 hours and after than keeping in 60°c bath for 24 hours and then in +25°C bath for 2 hours were prepared for test [AASHTO T 283]. After processing, the samples were tested in terms of dry and freezingthawing Marshall Strength and indirect tensile strength. After determining the most desired design, the considered design was tested in terms of indirect tensile tests with saturation processing to determine whether the intended design has obtained the least recommended values for foamed asphalt mixtures or not? RESULTS

Marhsal strength (Kgf)

Marshall Strength Marshal Strength is an expression of mixture strength. As can be seen in figure (3), Marshal Strength of samples which micro-filler and cement were used for their construction, has significantly increases. Micro-fillers can fill the porosities due to having fine aggregates; furthermore, it can be combined with cement due to cement Pozzolan property which makes rigid object. By increase of used asphalt from 2 to 3 percent, Marshall Strength is reduced which is due to insulation feature of asphalt which prevents hydration operation and its combination with micro-fillers to some extent. 2500 2000 1500 1000 500 0

C D E 2% Asphalt 3% Asphalt Samples Figure 03. The results of Marshall Strength of foamed asphalted samples A

B

Indirect tensile strength in dry and freezing- thawing state Indirect tensile test is an appropriate method for evaluation of tensile strength and the fatigue life of asphalt mixtures. Furthermore, this test can be used for evaluation of moisture sensitivity of asphalt mixtures [14]. The tensile strength tests were performed in dry and freezing- thawing conditions. The results indicate that according to figures (4 and 6), the greatest indirect tensile strength in dry state is related to design C and in freezing- thawing condition, it is related to design D. According to TSRC results (figures 5 and 7), design D has less sensitivity to freezing- thawing in 2% asphalt and design E in 3% asphalt. Moreover, figures 4 and 6 show that when the asphalt changes from 2 to 3%, it leads to increase of dry and saturated indirect tensile strength. In dry condition, due to high cohesiveness of asphalt, tensile strength in 3% asphalt is more than 2%; and in freezing-thawing condition, higher strength of asphalt to freezing- thawing due to cohesiveness of active filler, the tensile strength increases. The largest value of TSR in 2 and 3% asphalt is for design D and E, respectively which indicates, the presence of micro-fillers and their combination with cement, lead to significant increase of the mixture strength against freezing- thawing. However, TSR value is less than the minimum value for hot asphalt mixture (minimum 0.7). Since in foamed asphalted mixture, the porosity is so much, furthermore, coarse aggregates are not fully covered by asphalt so the strength significantly reduces. However, as it can be observed, microfillers have less strength than design A due to combining with cement and having filler properties (according to figure 4, 5, 6 and 7). This method can be used for better understanding and selection of proper design for mountain areas so by comparison, the best design can be used for cold climates.

17

ICESA 2014

ITS (kPa)

Internat ional Civil Engineering & Architecture Symposium for Academicians

600 500 400 300 200 100 0 A

B

C

D

samples by 2% asphalt

E

ITS dry ITS cold

Figure 04. Indirect tensile strength in 2% asphalt 0,5

TSR

0,4 0,3 0,2 0,1 0 A

B C D samples by 2% asphalt

E

Figure 05. Ratio of tensile strength of freezing- thawing to dry state in 2% asphalt

ITS (kPa)

600 400 200 0 A

B C D E ITS dry samples by 3% asphalt ITS cold Figure 06. Indirect tensile strength in 3% asphalt

TSR

0,6 0,4 0,2 0 A

B C D E samples by 3% asphalt Figure 07. Ratio of tensile strength of freezing-thawing to dry state in 3% asphalt Complementary tests After performing the tests related to dry and freezing- thawing conditions, designs D and E were selected (since they have higher TSRC ratio). Then, their specification in saturation state was investigated in 3% asphalt and compared with recommended values of Asphalt Academy Regulation 2002, and at the end, the best design was selected (Design A has been brought for comparison of results). To this end, samples were prepared according to previous conditions, however, to determine the indirect tensile strength of saturation, this time, after dry processing they were placed in 25° Celsius water for 24 hours, and then saturation indirect strength test was performed. For evaluation of mixture flexibility, mixture Q (ratio of Marshall Strength to fluidity) was used. As can be seen in table (3), all designs are proper since they have fulfilled the recommended values of guideline. However, since design D has the highest Marshall Strength, it can 18

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

be selected as the best design for heavier traffic condition. Moreover, design E can be selected as the best design for average traffic condition with adverse weather due to less sensitivity to freezing-thawing (TRS= 0.84). Table 03. The results of complementary test Samples ITSdry (kpa) ITSsat (kpa) TSR Q Criterion [8]: minimum Criterion [8]: minimum Criterion [8]: minimum Criterion [8]: 200 to 300 200 0.65 600 A 584 440 0.75 570 D 476 350 0.73 500 E 440 370 0.84 280 DISCUSSION For cold climates, it is better to use AASHTO T 283 for evaluation of mixtures’ quality and after determining appropriate design for cold areas, it is possible to control it in saturation state to control the terms of regulation in terms of moisture sensitivity. The results indicate that by obtaining TSRc value as 0.3 and more, it is possible to obtain the minimum TSR value for saturation. The combination of micro-fillers with cement leads to increase of mixture strength rather than using each of them individually. The combination of micro-fillers with cement leads to increase of mixture durability. Furthermore, micro-lime preserves the flexibility of mixture in addition to increasing its durability. By reduction of consuming asphalt from 3 to 2%, it is observed that the tensile strength of freezing-thawing has decreased which indicates asphalt resists against freezing-thawing more rather than Pozzolans. This is while in saturation state, it is usually expected that due to decrease of asphalt consumption there is more opportunity for Pozzolan’s reaction and the strength of mixture in saturation state increases, however, in freezing- thawing state, the mixture experiences strength reduction. For future studies, it is recommended to perform the tests with various percentages of asphalt and pozzolans to be able to obtain the proper least TSRc. Furthermore, it is recommended to evaluate lime combination which is good for reducing moisture sensitivity. Furthermore, it is proposed to control various percentages of chips in freezing- thawing condition which can be beneficial in determining the minimum coefficient. REFRENCES 1-Castedo, L.H. and Wood, L.E, (1983), “Stabilization with foamed asphalt of aggregates commonly used in lowvolume roads”, Transportation Research Board, Issue (898), P. (297-302), Washington D.C.. 2-Asi, I. (2001), ”Stabilization of Sebkha Soil Using Foamed Asphalt.” J. Mater. Civ. Eng., 13(5), 325–331. 3-Gonzalez, A., Cubrinovski, M., Pidwerbesky, B., and Alabaster, D. (2011), ”Strength and Deformational Characteristics of Foamed Bitumen Mixes under Suboptimal Conditions”, J. Transp. Eng., 137(1), 1–10. 4- Iran Management and Planning Organization (2006), Specifications and Performance of Cold Reclaimed Asphalt Pavement, Publication 339. 5- Halles, F.A. and Thenoux, G.Z., (2009), "Degree of Influence of Active Fillers on Properties of Recycled Mixes with Foamed Asphalt", Transportation Research Board, Issue (2095), P. (127–135), Washington D.C.. 6- Iwa skia M. and Chomicz-Kowalskab, A., (2013), "Laboratory Study on Mechanical Parameters of Foamed Bitumen Mixtures in the Cold Recycling Technology. Procedia Engineering, Vol. (57), P. (433 – 442). 7-Kendall, M., Baker, B., Evans, P. and Ramanujam, J. (2001), “Foamed Bitumen Stabilisation-The Queensland Experience” 20th ARRB Conference held in Melbourne, Australia, 19-21 March 2001, 1-4. 8- Asphalt Academy, (2002), "Interim technical guidelines (TG2): The design and use of foamed bitumen treated materials", CSIR, Pretoria, South Africa. 9- Xu, J. Z., Hao, P.W., Ma, Y.F. and Liu, N., (2012), "Study on the optimization design of mixing moisture content in foamed asphalt mix". Materials and Structures, Vol. (45), Issue (7), P. (1071–1085). 10- Iwa ski, M. and Chomicz-Kowalska, A., (2011), “The effects of using foamed bitumen and bitumen emulsion in the cold recycling technology”, 8th International Conference, Environmental Engineering, May 19-20, pp. 1089-1096, Vilnius, Lithuania. 11- Judycki J., Jasku a J., (1997), "Badania betonu asfaltowego na oddzia ywanie wody i mrozu (Investigations into water and frost resistance of asphalt concrete)", Drogownictwo Nr 12, 1997.s. 374-378. 12- Tunnicliff, D. G. and Root, R. E., (1981), "Anti-stripping in Asphalt Concrete State-of-Art", The Association of Asphalt Technology, vol. 51. 13-Technical Guideline (TG2), “The design and use of foamed bitumen treated materials”, (2009), Published by Asphalt Academy, Pretoria, South Africa. 14- Kim, Y. and Lee, H. (2006), ”Development of Mix Design Procedure for Cold In-Place Recycling with Foamed Asphalt”, J. Mater. Civ. Eng., 18(1), 116–124.

19

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

EFFECT OF THE SOLID PYROLYSIS PRODUCT OBTAINED FROM WASTE HARD PLASTIC POLYPROPYLENE ON BITUMINOUS BINDERS O. N. Çelik1, N. Atasa un1, M.A. Lorasokkay2 [email protected], [email protected], [email protected]

1

Department of Civil Engineering, Selcuk University, Konya, Turkey 2 Department of Construction, Selcuk University, Konya, Turkey

ABSTRACT The performances of both the binder and the mixture could be enhanced by modifying them. In this study, bitumen was modified by using solid pyrolysis product of waste hard plastic polypropylene (WHPP) and the effects of this additive on the rheological properties of bitumen was examined. 50/70 penetration bitumen was modified with this new additive at different concentrations ranging from 2% to 6% in weight. The rheological properties of these modified bitumens were determined by using some traditional and some superpave binder test methods. In this way, it was investigated that the possibility of producing a new additive from waste plastic polypropylene liquefied by pyrolysis method for bitumen modification. Thus, it was aimed to investigate whether this new additive from waste hard plastic polypropylene usable for bituminous binders. Keywords: bitumen, modification, road materials, wastes. INTRODUCTION The strength of bitumen and bituminous mixtures could increase by incorporating some additives in both bitumen and bituminous mixtures. Different additives were used by researchers to improve the performance of both bitumen and bituminous mixtures. In this study, bituminous binder was modified by using the solid pyrolytic product obtained from waste hard plastic polypropylene. Pyrolysis is the thermal degradation in which liquid, solid and gas products are obtained through heating the raw material to high temperatures in oxygen free environment. In pyrolysis, the polymeric materials are heated to high temperatures, so their macromolecular structures are broken down into smaller molecules and a wide range of hydrocarbons are formed. These pyrolytic products can be divided into a gas fraction, a liquid fraction consisting of paraffin, olefins, naphthenes and aromatics, and solid residues. Pyrolysis appears to be a technique which is able to convert petroleum based plastic wastes into gasoline-range hydrocarbons [1]. MATERIALS AND METHODS In this study, 50/70 penetration degree base bitumen was used. The solid pyrolysis product from waste hard plastic polypropylene (WHPP) was used as an additive for bitumen modification. This additive was mixed with the base bitumen to produce modified bituminous binders containing 2%, 4% and 6% by weight of the base bitumen. Waste hard plastic polypropylene used in this study is shown in Figure 1. Pyrolysis Method In the pyrolysis process (heating in an oxygen free atmosphere), the organic components of the material are decomposed generating liquid and gaseous products, which can be useful as fuels and/or sources of chemicals. The inorganic materials (fillers, metals) remain practically unaltered and free of the binding organic matter; therefore, metals could be separated and the remaining solid may be reused (additive, filler, pigment) or as a last resort, it would be a minimum waste to be land filled. Pyrolysis is an especially appropriate recycling technique for waste streams containing different plastics and other materials, for which mechanical recycling is not feasible [2, 3]. Pyrolysis of the waste hard plastic polypropylene was carried out on laboratory scale pyrolysis system which is shown in Figure 2.

20

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Figure 01. Waste hard plastic polypropylene and pyrolytic product of WHPP

Figure 02. Laboratory scale pyrolysis system Penetration and Softening Point Tests Penetration and softening point tests are used to determine the consistency of modified and unmodified bituminous binders according to the ASTM D5, 2006 and ASTM D 36, 2006, respectively. In addition, the temperature susceptibility of bituminous binders are determined by using penetration and softening point test results with PI values. Rotational Viscometer (RV) Test The RV is used to determine the workability of bituminous binders at high temperatures. High temperature binder viscosity is measured to ensure that the asphalt is sufficiently fluid when pumping and mixing [6]. The specification stipulates that the binder must have a maximum viscosity of 3 Pa.s at a test temperature of 135°C [7]. The RV can also be used to establish the viscosity – temperature relationship for binders. This relationship can be used as a guideline to determine mixing and compacting temperatures [7]. Dynamic Shear Rheometer (DSR) Test The DSR is used to characterize the viscous and elastic behavior of asphalt binders. It does this by measuring the complex shear modulus (G*) and phase angle ( ) of asphalt binders. G* is a measure of the total resistance of a material to deforming when repeatedly sheared. is an indicator of the relative amounts of recoverable and non-recoverable deformation [6]. Permanent deformation is controlled by limiting G*/sin at the test temperatures to values greater than 1.0 kPa (before aging) and 2.2 kPa (after aging). Fatigue cracking is controlled by limiting G*.sin of pressure aged material to values less than 5000 kPa at the test temperature [6]. Binder Aging Methods In this study, bituminous binders were aged by using rolling thin film oven (RTFO) and pressure aging vessel (PAV) tests. During mix production and construction, is simulated by aging the binder in RTFO test. RTFO test exposes films of binder to heat and air approximates the exposure of asphalt to these elements during hot mixing and handling. Binder aging occurs a long term period in a pavement. This is simulated by use of a PAV. This test exposes binder samples to heat and pressure in order to simulate, in matter of hours, years of in service aging in a pavement [6]. RESULTS In this study, bituminous binder was modified by mixing the solid pyrolysis product obtained from WHPP containing 2%, 4% and 6% in bitumen. The pure bitumen and modified bituminous binders were subjected to the penetration [4] and softening point [5] tests and PI values were determined by using these two test results. 21

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Also, pure bitumen and modified bituminous binders’ high temperature properties were tried to determine by using Superpave binder tests including dynamic shear rheometer (DSR) test [8], rotational thin film oven test for short term aging (RTFOT) [9], pressing aging test for long term aging (PAV) [10] and rotational viscometer test (RV) [11]. Pyrolysis Test Results Pyrolysis of hard plastic polypropylene wastes were carried out at 5-50 oC/min., at 550 oC and nitrogen flow on the samples. The loaded reactor was placed into the oven and then the air in the reactor was purge out with nitrogen flow. The reactor was externally heated from room temperature to the desired operating temperature. Then, the reactor was removed from the oven and then it was pull out to cool to room temperature. Later on, the solid pyrolysis product obtained was used for bitumen modification.

Penetration Value (1/10 mm)

Penetration and Softening Point Test Results Conventional binder tests including the penetration test and softening point test were performed according to the standards. The graphical display of the penetration and softening point test results are shown in Figure 4 and Figure 5, respectively.

Pyrolysis Product

55,00

54,07

54,00

53,11

53,00 52,00 51,12

51,00

50,97

50,00 0

2

4

6

Content of additive in binder (%) Figure 04. Penetration values of bituminous binders

Softening Point °C

According to the test results, it can be seen that the additive content decreased the penetration value of the bitumen. This result indicates that the bitumen get harder consistency.

Pyrolysis Product

62 60 58 56 54 52 50 48 46 44

62,2 53,7 46,9

44,0 0

1

2

3

4

5

6

Content of additive in binder (%) Figure 05. Softening points of bituminous binder

According to the softening point test results, it was seen that the additive used in this study increased the softening point of bituminous binder. In addition, as the additive content increased, the softening point of bituminous binder increased. Consequently, this result indicates that the bituminous binder get harder consistency. Penetration Index Test Results The penetration index values which indicate the susceptibility of bituminous binders against the temperature were calculated with equation below by using the penetration and softening point test results. The penetration index test results conducted on bituminous binders used in this study are shown in Table 1. 22

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

PI

1952 500 * log( Pen) 20 * SP 50 * log( Pen) SP 120

Table 01. PI values Additive Content Penetration Index (PI) 0%

-2.66

2%

-1.96

4%

-0.18

6%

1.49

In this formula, (Pen) shows the penetration value of bituminous binder at 25 °C and SP shows the softening point of bituminous binders. When the PI results are examined, it can be seen that the additive used in this study increased the PI values of bituminous binder. Consequently, it was seen that as the additive content increased, the temperature susceptibility of bituminous binders decreased and it was seen that this additive made positive effect on bituminous binders. Rotational Viscometer (RV) Test Results The viscosity values of base bitumen and bituminous binders modified by using solid pyrolysis product of waste hard plastic polypropylene (WHPP) with pyrolysis method at 135 °C were determined with Brookfield Viscometer. The workability levels of base bitumen and modified bituminous binders were determined by using viscosity values obtained. The graphical display of RV test results are shown in Figure 6.

Figure 06. Viscosity values of base bitumen and modified bituminous binders at 135°C When the test results were examined, it was determined that the viscosity values of all bituminous binders were not exceed the 3000 cp which is the specification criterions and it was seen that both base bitumen and modified bituminous binders were suitable for workability. Also, according to the test results given in Figure 6, when comparing the viscosity values of base bitumen with modified bituminous binders at 135 °C, it was observed that the viscosity value of 2% additive modified bitumen and 6 % additive modified bitumen were approximately 7.6% and 15.8% lower than base bitumen, respectively. On the other hand, the viscosity of 4% additive modified bitumen was approximately 2% higher than of base bitumen. According to this result, it was seen that the 4% additive modified bituminous binder had the lowest temperature susceptibility and it had harder consistency as compared to the other bituminous binders used in this study. DSR Test Results In this study, the rheologic properties of base bitumen and bituminous binders modified by using solid pyrolysis product of WHPP were tried to be determined by using DSR test. In order to determine the effect of solid pyrolysis product of WHPP on rutting resistance of both unaged bituminous binders and aged bituminous binders by using RTFOT method, G*/ sin values of all bituminous binders were tried to be determined. Changes in rutting parameters (G*/ sin ) of both unaged bituminous binders and aged bituminous binders by using RTFOT method to temperature is shown in Figure 7 and Figure 8. The test results obtained were evaluated by comparing with the specification criterions. According to the test results, for unaged and aged bituminous binders, from 23

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

the Figure 7 and Figure 8, we can see that as the temperature increases there is a decreasing in rutting parameter G*/sin .

Figure 07. Changes in rutting parameters of unaged bituminous binders to temperature Also, for unaged bituminous binders, as shown in Figure 7, when G*/sin was equal to the specification criterions, 1.0 kPa, the temperature of base bitumen and 2 % additive modified bituminous binder were 64°C. On the other hand, when the rutting parameter G*/sin was equal to the 1.0 kPa, the temperature of 4% additive modified bituminous binder was 70°C and the temperature of 6% additive modified bituminous binder was 76°C.

Figure 08. Changes in rutting parameters of aged bituminous binders to temperature For aged bituminous binders, as shown in Figure 8, while G*/sin was equal to the specification criterion 2.2 kPa at 64°C for base bitumen and 2% additive modified bituminous binder, this parameter was equal to the 2.2 kPa at 70°C and 82°C for 4% and 6% additive modified bituminous binders, respectively. According to the test results, for unaged bituminous binders, it was seen that this additive had a positive effect on rutting parameter G*/sin at high temperatures. Also, from the test results, when comparing the G*/sin values of base bitumen with modified bituminous binders, it was observed that the G*/sin value of 2% additive modified bituminous binder and 4% additive modified bituminous binder were approximately 9% and 52% more effective than the G*/sin value of base bitumen, respectively. And, for the G*/sin value of 6% additive modified bituminous binder was approximately 2.64 times more effective than of base bitumen. It was given in Figure 9 that the effect of additive content on rutting parameters (G*/sin ) at 70°C for unaged bituminous binders and aged bituminous binders by using RTFOT method.

24

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Figure 09. The effect of additive content on rutting parameters (G*/sin ) at 70°C According to the test results from Figure 9, we can see that as the additive content increases there is an increasing in rutting parameter (G*/sin ) for both unaged and aged bituminous binders at 70°C. Also, for unaged bituminous binders from Figure 8, when comparing the G*/sin values of base bitumen with modified bituminous binders at 70°C, it was observed that the G*/sin value of 2% additive modified bituminous binder and 4% additive modified bituminous binder were approximately 16% and 68% more effective than the G*/sin value of base bitumen, respectively. And, for the G*/sin value of 6% additive modified bituminous binder was approximately 3.2 times more effective than the G*/sin value of base bitumen. On the other hand, for aged bituminous binder by using RTFOT method, when comparing the rutting parameter (G*/sin ) values of base bitumen and modified bituminous binders at 70°C, it was seen that the G*/sin value of 2% additive modified bituminous binder and 4% additive modified bituminous binder were approximately 33% and 92% more effective than the rutting parameter (G*/sin ) value of base bitumen, respectively. Also, the fatigue cracking resistance of long term aged bituminous binders were determined by using G*.sin values. Changes in G*.sin parameters of long term aged bituminous binders to temperature is shown in Figure 10.

Figure 10. The effect of additive content on fatigue parameter (G*.sin ) When Figure 10 was examined, it was seen that the additive obtained from WHPP by pyrolysis method increased the G*.sin parameter of long term aged bituminous binder. In addition, according to the test results, when comparing the G*.sin parameter of long term aged base bitumen with long term aged modified bituminous binders, it was observed that the G*.sin value of 2%,4% and 6% additive modified bituminous binders were approximately 41%, 32% and 47% higher than the G*.sin parameter of long term aged base bitumen, respectively.

25

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

DISCUSSION In this study, it was investigated that the possibility of producing a new additive from pyrolysis product of waste hard plastic polypropylene for bitumen modification. Conventional test methods including penetration test and softening point test, and superpave binder test methods including DSR, RV, RTFOT and PAV tests were performed to determine the rheologic properties of bituminous binders. According to the penetration test and softening point test results, it was determined that the additive content decreased the penetration value of base bitumen and increased the softening point of base bitumen. These results showed that the bituminous binders modified by using solid pyrolysis product of WHPP get harder consistency. Also, the PI test results demonstrated that the additive used in study reduced the temperature susceptibility of bituminous binders and it made positive effect on bitumen. RV test result shows that the viscosities at 135°C are less than 3000cp which is the specification criterions for all bituminous binders. And it can be concluded that all bituminous binders are suitable for workability. According to the DSR test results, we can see that as the additive content increases, there is an increasing in rutting parameter (G*/sin ) for both unaged and aged bituminous binders by using RTFOT method. And so, it can be concluded that the solid pyrolysis product of WHPP used as an additive in this study improve the rutting resistance of bitumen. Also, it was seen that the additive used in this study increased the fatigue parameter (G*.sin ) of long term aged bituminous binders. It can be concluded that this additive increases the fatigue cracking resistance of base bitumen. According to the results of this study, it is thought that the solid pyrolysis product of WHPP can be used as an additive for bitumen modification. The additive used in this study is waste material. So, it is thought that recycling of wastes in this way will make good contribution to the environment and to the economy. Acknowledgemets: This study was prepared by using a part of PhD thesis of Neslihan Atasa un. And the authors would like to thank the Turkish Scientific and Technological Research Foundation (TUBITAK) (Project Number: 112M116) and the Scientific Research Projects Coordination Department of Selcuk University (SU BAP) (Project Number: 11101031) for their financial support. REFERENCES [1] Demirba A., 2004, Energy Edu. Sci. Technol. 13, 1–12. [2] Al-Salem, S.M., Lettieri, P., Baeyens, J., 2009, Recycling and recovery routes of plastic solid waste (PSW): a review, Waste Manag. 29, 2625–2643. [3] Carvalho, M.T., Ferreira, C., Portela, A., Santos, J.T., 2009, Application of fluidization to separate packaging waste plastics, Waste Manag. 29, 1138–1143. [4] ASTM D5, 2006, Standard test method for penetration of bituminous materials. American Society for Testing and Material. [5] ASTM D 36, 2006, “Standard test method for softening point of bitumen (ring-and ball apparatus)”, American Society for Testing and Materials. [6] FHWA-SA-94-069, 1994, “Bacground of Superpave Asphalt Binder Test Methods”, National Asphalt Training Center Demonstration Projects 101. [7] SHRP-A-410, 1994, “Superior performing asphalt pavements (Superpave) The product of the SHRP asphalt research program”, National Research Council Washington, DC. [8] AASHTO T 315, 2008, Standard Method of Test for Determining the Rheological Properties of Asphalt Binder Using a Dynamic Shear Rheometer (DSR), American Association of State and Highway Transportation Officials. [9] TS EN 12607-1, 2008, Bitumen and bituminous binders - Determination of the resistance to hardening under the influence of heat and air - Part 1: RTFOT method. [10] TS EN 14769, 2012, Bitumen and bituminous binders - Accelerated long-term ageing conditioning by a Pressure Ageing Vessel (PAV). [11] ASTM D 4402, 2013, Standard Test Method for Viscosity Determination of Asphalt at Elevated Temperatures Using a Rotational Viscometer, ASTM International.

26

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

SOME MAJOR SIGNALIZED INTERSECTION EXAMINATION IN KONYA BY USING SIDRA INTERSECTION 5.1 SOFTWARE O. N. ÇEL K1, M. M. AKMAZ2 [email protected], [email protected]

1

2

Prof. Dr., Department of Civil Engineering, Selcuk University, Konya, Turkey, Res. Assist., Department of Civil Engineering, Selcuk University, Konya, Turkey,

ABSTRACT Signalized intersections are widely used in today's cities where the traffic flows from various directions and the pedestrians have right to pass one by one. In cities, vehicles are increasing with the increase in population and economic developments. This highlighted the need of traffic flow arrangement using by signals. If it is ignored, delays will inevitable. Therefore, the increase in delays will cause labor losses, increase in fuel consumptions due to waiting vehicles and other negative effects on vehicle drivers. In this study, some of the intersections in Konya were investigated such as Kule and Nalçac –Sille intersections that are important in terms of urban traffic in Konya. New cycle times are proposed with the aim to minimize delays and increase the capacities and level of services in these intersections. The intersections were examined using Sidra Intersection 5.1 software based on the Australian methods that the traffic counting results were obtained by shooting cameras at intersections to enter intersection data into the computer program. Then the information of signalization such as signal plans, phase times, green times and cycle times and the information about the geometric layout of intersections were provided. The intersection analyses were performed according to the current cycle times and optimum cycle times proposed by the method. The analytical results obtained were compared and some solution suggestions were given. When the results were analyzed, the decrease in the delays and the increase in the capacities at the intersections were generally observed as a result of the proposed cycle times. Keywords: Delay, Capacity, Sidra Intersection, Signalized Intersection INTRODUCTION Signals are used together with other traffic signs to control traffic flows, to warn and to give necessary information to vehicle drivers and pedestrians [5]. Signals are traffic control tools that ensure regular and safe flows on roads and at intersections. For the first time, hand–controlled traffic signals in the form of semaphores were used in London in 1868. The first signalization system with red and green lights was established in 1914 in Cleveland, USA. Yellow lights were started to use in Detroit in 1920 [3]. In the late 1920s, electrically operated signals became the primary intersection traffic control device. Signals varied but became standardized as major manufacturers entered the traffic signal business and transportation projects gained priority after World War I. After the 1920s, major technological advances made signal controls more flexible so that multiple cycle length units could vary timing plans by time of day [6]. Urban transportation, as a whole, covers many traffic compositions including subway train, tram, bus, minibus, car, bike and pedestrians. Signalization has an important position in integrated urban transportation solutions in terms of regulating and managing traffic flows. Therefore, signalization has to be considered as a part of the whole transportation system instead of considering it individually. When we look at the transportation system of Konya city, we see that investments to speed up vehicle traffic ignoring other traffic groups in the system are not very successful because grade separated intersections constructed at different locations in the city prevent the signalization system from working with high performance. A successful signalized intersection incorporates three conditions [4]: Intersections have to be designed in the form of islands which are in conformity with traffic flows. True signal orders (phase diagrams) have to be set up. True cycle times have to be calculated in accordance with flow volumes. In addition to these conditions, to get better results and to increase success, a device which is sensitive to changes in flows has to be set up to synchronize cycle times with day–long flow changes. The most important point is to set up appropriate cycle order and to apply the best cycle time which will minimize the medium of delays as per vehicles [4]. This study aims to find out optimum cycle times to minimize delays and to increase capacities at Kule and Nalçac –Sille signalized intersections. These are the most intensive intersections of Konya city. They are located at the important points of North–South connecting roads. Again, these intersections are very close to each other so proposed solutions 27

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

are important for interactions between different traffic flows. For this reason, these intersections have been selected in the study. Data about traffic, signalization and geometric layout in the intersections has been collected. Using the software of Sidra Intersection 5.1, current traffic problems have been revealed. MATERIALS AND METHODS The Sidra Intersection 5.1 software is an advanced micro–analytical tool for evaluation of alternative intersection designs in terms of capacity, level of service and a wide range of performance measures including delay, queue length and stops for vehicles and pedestrians, as well as fuel consumption, pollutant emissions and operating cost. It has been a valuable technology transfer tool based on extensive research carried out in Australia, USA and elsewhere. It has been developed continuously in response to feedback from practising traffic engineers and planners. It is for use as an aid for design and evaluation of signalised intersections, signalised pedestrian crossings, single point interchanges, roundabouts, roundabout metering, two–way stop sign control, all–way stop sign control, and give–way / yield sign– control. Sidra Intersection traffic models can be calibrated for local conditions. It provides various facilities for this purpose. In the USA, Sidra Intersection is recognised by the US Highway Capacity Manual, TRB Roundabout Guide (NCHRP Report 672) and various local roundabout guides. In Australia and New Zealand, it is endorsed by AUSTROADS and various local guidelines [2]. The research report ARR 123 is related to the intersection analysis methodology used in the Sidra Intersection software. Since the publication of this report, many related aspects of the traffic model have been further developed in later versions of Sidra Intersection [1]. The methods given in report follow the basic framework established in earlier publications which have influenced the Australian and U.K. signal design practices (Miller 1968b; Webster and Cobbe 1966). This report presents techniques for the analysis of capacity and timing requirements of traffic at signalised intersections. The present report introduces several changes to the traditional techniques, a basic change being from ‘phase–related’ methods to ‘movement–related’ methods. An important aspect of this change is the use of ‘movement lost time’ concept instead of ‘phase lost time’ concept, which leads to a definition of the intersection lost time as ‘the sum of critical movement lost times’ rather than ‘the sum of phase lost times’ [1]. Signal phasing is the basic control mechanism by which the operational efficiency and safety of a signalised intersection is determined. It is therefore important to understand clearly how traffic movements and signal phases relate to each other. Each separate queue leading to the intersection and characterised by its direction, lane usage and right of way provision is called a movement. Signal phase is a state of the signals during which one or more movements receive right of way. Signal phases will be defined in such a way that when there is a change of right of way, that is when a movement is stopped and another started, there is a phase change [1]. One complete sequence of signal phases is called a signal cycle. The time from the end of the green period on one phase to the beginning of the green period on the next phase is called the intergreen time (I). It consists of yellow and all–red periods. During the all–red period, both the terminating and the starting phases/movements are shown red signal simultaneously. The sum of all phase intergreen and green times is the cycle time: C = (I + G). The basic movement characteristics are saturation flow, effective green time and lost time. They are illustrated in Fig. 1 in relation to a corresponding signal phasing arrangement. The basic model used is essentially a traditional one. However, some definitions are new. The model assumes that when the signal changes to green, the flow across the stop line increases rapidly to a rate called the saturation flow (s), which remains constant until either the queue is exhausted or the green period ends. The departure rate is lower during the first few seconds while vehicles accelerate to normal running speed. Similarly, the departure rate is lower during the period after the end of green because some vehicles stop and others do not. In this way, the saturation flow is the maximum departure rate which can be achieved when there is a queue [1]. As indicated by the dotted line in Fig. 1, the basic model replaces the actual departure flow curve by a rectangle of equal area, height of which is equal to the saturation flow (s) and whose width is the effective green time (g). The time between the start of displayed green and the start of effective green periods (ee’) is considered to be a start loss. Similarly, the time between the end of displayed green and the end of effective green periods (ff’) is considered to be an end gain. Therefore, the effective green time: (g = G + ff ee ). The start and end times of the effective green period for a movement are best defined with reference to phase change times. For this purpose, a start lag (a) is defined as the sum of the movement intergreen time and start loss, and an end lag (b) is defined simply as the end gain (a = I + ee and b = ff ). The movement lost time (l) is defined as the difference between the start and end lag times: l = a b. Therefore, the movement lost time: (l = I + ee ff ).

28

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Figure 01. Basic model and definitions [1] The capacity of a movement at traffic signals depends on the maximum sustainable rate at which vehicles can depart, the saturation flow (s) and the proportion of the cycle time (c) which is effectively green for that movement, and is given by the formula: Q = S . The proportion of effective green time to cycle time is called the green time ratio for the movement: u =

. Another useful movement parameter is the ratio of arrival flow (q) to saturation flow. This is

called the flow ratio: y = x=

=

=

. The movement degree of saturation is the ratio of arrival flow to capacity. It is given by:

. The degree of saturation is the ratio which relates these two parameters. To provide adequate

movement capacity, “Q>q” or “x<1” is should be [1]. Data obtained from intersections is required to examine intersections with the help of Sidra Intersection 5.1 software. This data can be grouped as intersection layout, traffic volumes and signal phasing. Data about intersection layout: Any turn bans, one way approaches or exits, all lanes (exclusive or shared) with clear indication of lane disciplines, slip lanes and continuous (uninterrupted traffic) lanes, upstream and downstream short lanes (turn bays, approach parking, and loss of a lane at the downstream side), lane widths and median widths, pedestrian crossings, grade information, any data related to adjacent parking, buses stopping, tram, etc., direction of north, intersection control. Data about traffic volumes: Volume counts for vehicles, heavy vehicles data, pedestrian volume, peak flow periods, peak flow factor and flow scale. Data about signal phasing: Timing data (yellow and all–red times, start loss and end gain, minimum and maximum green time, etc.), phase descriptions and phase sequences showing movements which have right of way in each phase, signal phasing diagrams indicate differences between normal vehicle movements and pedestrian movements clearly, basic saturation flows [2]. Data about the intersection layout and signal phasing of selected intersections has been collected from the Greater Konya Municipality, Department of Reconstruction and Urbanism, Branch of Transportation and Traffic Signalization. Traffic counts have been conducted to find out traffic volumes in the intersections. Traffic at intersections has been recorded on video cameras. Camera has been settled at a high point to see all traffic flows in the intersection. Traffic flows have been taped 15 hours between 07:00–22:00 to determine maximum traffic volume. Traffic flows have been taped for every intersection separately in the third Friday of that month considering that traffic density might be higher due to market mobility and trade in this time. Counting forms have been filled hourly for every traffic flow (which turn right and left and go straight) and for every vehicle separately in the intersections. Typing hourly traffic volumes for every flow in tables, maximum hour traffic at the examined intersections has been found. To determine peak hour factor (PHF), which has been entered in the computer software, maximum hour traffic rates have been recounted with 15–minute periods and then typed in the 29

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

related forms and tables. In this way, maximum traffic volume rates which are used to find PHF have been found. The intersections of Kule and Nalçac –Sille have been recorded to tapes on Fridays at the 21st of October 2011 and at the 18th of November 2011. RESULTS Data obtained from intersections has been entered in Sidra Intersection 5.1. As a result of the software analysis, some numerical values for each intersection directions such as degree of saturation, capacity, level of service, control delay per vehicle have been gained. Firstly selected intersections have been analysed by using current phase plans and times (cycle times, green times, etc.) to understand the present situation of intersections. Then the software has been calculated optimum cycle times by using the same or different phase plans. In this way, new cycle times have been found to decrease delays and to increase capacities. This calculation entails entering maximum / minimum cycle times and a reference interval in the software to find out optimum cycle times. Analysis results obtained from current and proposed (optimum) cycle times have been presented in Table 1 and Table 2 for selected intersections.

Figure 02. Kule Intersection handles and the point of camera

Figure 03. Kule Intersection

Figure 4. Nalçac –Sille Intersection

30

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Flow Direction 1–2 1–3 1–4 2–1 2–3 2–4 3–1 3–2 3–4 4–1 4–2 4–3

Table 01. Analysis results for current and proposed cycle times at Kule Intersection CURRENT CYCLE T ME (110 Sec.) PROPOSED CYCLE T ME (125 Sec.) Delay Degree of Capacity Level of Delay Degree of Capacity Level of (sec.) Saturation (veh./hour) Service (sec.) Saturation (veh./hour) Service 534.5 2.02 204.0 F 356.4 1.62 255.4 F 519.9 2.02 218.5 F 342.5 1.62 273.6 F 522.1 2.02 236.9 F 344.3 1.62 296.6 F 7.8 0.24 1431.7 A 7.7 0.25 1396.8 A 332.6 1.58 268.0 F 360.5 1.62 260.7 F 156.8 1.21 1362.9 F 336.8 1.59 1037.3 F 489.4 1.96 156.4 F 337.7 1.61 190.6 F 492.8 1.96 241.0 F 340.4 1.61 293.6 F 507.1 1.96 226.2 F 354.4 1.61 275.6 F 141.3 1.14 254.6 F 137.1 1.11 261.3 F 58.0 0.94 1244.0 E 147.2 1.17 1000.9 F 7.8 0.17 1345.4 A 7.8 0.17 1351.9 A

Figure 05. Nalçac –Sille Intersection handles and the point of camera

Flow Direction 1–3 1–4 2–1 2–4 3–1 3–2 4–2 4–3

Table 02. Analysis results for curr. and prop. cycle times at Nalçac –Sille Intersection CURRENT CYCLE ME (75 Sec.) PROPOSED CYCLE T ME (70 Sec.) Delay Degree of Capacity Level of Delay Degree of Capacity Level of (sec.) Saturation (veh./hour) Service (sec.) Saturation (veh./hour) Service 21.0 0.64 1448.8 C 18.9 0.62 1496.8 B 99.2 1.08 508.9 F 84.1 1.05 525.8 F 37.3 0.89 589.0 D 26.0 0.79 665.3 C 157.3 1.26 1855.9 F 78.5 1.08 2163.9 E 108.8 1.14 1063.7 F 91.8 1.10 1099.0 F 118.4 1.14 308.2 F 101.3 1.10 318.4 F 28.9 0.88 1836.9 C 14.5 0.75 2141.8 B 21.3 0.29 566.0 C 17.1 0.26 644.4 B

31

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

DISCUSSION Table 1 shows that level of service values do not differ for the current and the proposed cycle times but it also shows that there is an improvement in terms of delay and capacity values with the proposed cycle time. In the solution according to the current cycle time; the delay and degree of saturation values of 1 (1–2, 1–3 and 1–4) and 3 (3–1, 3–2 and 3–4) intersection handles are higher than the values of other intersection handles (2 and 4). Similarly, the capacity values are lower than the others. In the solution according to the proposed cycle time; it is observed that the delay and degree of saturation values of 1 and 3 intersection handles decreased about 35% and 20% respectively and the capacity values increased about 25%. Also, it is observed that changes in the values of delay, degree of saturation and capacity are not significant for 2 and 4 intersection handles (except 2–4 and 4–2 directions). In this way, all values for intersection handles were balanced distributed with the proposed cycle time by the software. Therefore, it is observed that there is an improvement in Kule Intersection. Table 2 shows that level of service values of the some intersection handles and capacity values of the all intersection handles increase, and delay and degree of saturation values of the all intersection handles decrease with the proposed cycle time by the software. In the solution according to the current cycle time, it is observed that the 2–4 direction has the biggest delay value. In the solution according to the proposed cycle time, it is observed that the delay and degree of saturation values of this direction (2–4) decreased about 50% and 15% respectively and the capacity values increased about 15%. Also, it is observed that the delay and degree of saturation values of the all intersection handles decreased about 25% and 10% respectively and the capacity values increased about 10% in Nalçac –Sille Intersection. In conclusion, some improvements have been seen at all intersections, the delay and degree of saturation values decreased and the capacity values increased with the proposed cycle times. However, it is found that Kule Intersection work at F level of service and many traffic flow directions are insufficient in terms of the level of service in Nalçac – Sille Intersection. This finding reveals that new solutions to manage traffic flows have to be considered in the intersections. That’s why, the integration of traffic compositions in urban transportation has to be considered; and public transportation systems has to be developed to carry more people with less vehicle.

ACKNOWLEDGEMENTS This paper was prepared by using a part of MS Thesis of M. Mevlüt Akmaz.

RESOURCES 1Akçelik, R. 1998. Traffic signals: capacity and timing analysis. Research Report ARR No. 123 Seventh reprint. Australian Road Research Board Transport Research Ltd. Australian. 1–35. 2Anonymous. 2011. Sidra intersection user guide. Akçelik & Associates Pty. Ltd. Australian. Part 1 (1–8); Part 3 (17–18). 3Ayfer, M. Ö. 1977. Trafik sinyalizasyonu. No: 226. T.C. Bay nd rl k Bakanl Karayollar Genel Müdürlü ü Matbaas . Ankara. 7–82. 4Gedizlio lu, E. 2004. Kentlerimizde trafik yönetimi. Türkiye mühendislik haberleri. 434–2004/6. 20–21. 5Kutlu, K. 1993. Trafik tekni i. Üçüncü Bask . .T.Ü. aat Fakültesi Matbaas . stanbul. 261–298. 6Yauch, P. J. 1997. Traffic signalization–a history. Institute of Transportation Engineers 67th annual Meeting. Washington D.C.

32

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

STATEMENT OF TRAFFIC PROBLEMS AND PROPOSALS FOR A CONGESTED AREA WITHIN BAGHDAD CITY

A. AL-MASHHADANI1, Y. DEM R1

[email protected]

1

Gaziantep University, Civil Engineering Department, Gaziantep.

ABSTRACT This paper studies the circumstance of a congested signalized network selected inside Baghdad city. It is includes the description of the study area and the state of it and the reasons of congestion, in addition to viewing the data collection of the network. The adopted program is PTV-VISSIM 6. It is used in simulation the network and obtaining results of flow and measure of performance parameters. Finely, a brief discussion of produced results is explained in the end of paper. Key words: Baghdad, Traffic, Network, Intersection, Simulation, Calibration INTRODUCTION Iraq has been undergoing a rapid industrialization and urbanization funded largely by oil revenues. As a result, demand for transport has been increasing rapidly throughout Iraq in recent years [1]. To imagine that, we can take a general review of vehicles in Baghdad. In 1986, less than 200000 private passenger cars are registered in Baghdad [2]. Before April 2003 the number of private passenger cars is about 300000, since after April 2003 the number of vehicles in Baghdad increased rapidly and they became more than 1000000 (private & taxi) until April 2013 [3]. The vehicles increased without considering increase the capacity of roads and intersections, which gave rise to a congested system of traffic. A little number of overpasses was executed at some of intersections in recent years, but traffic problems still in Baghdad city. Therefore we selected a congested network inside Baghdad city to study it. This paper explains the description of the study area and the data collection which includes determination of peak hour and the volume at peak hour. Also it shows the adopted software program used in simulation and obtaining performance results. MATERIALS AND METHODS Description of the study area The study area represents a network of seven signalized intersections within Baghdad which is the capital of Iraq. It contains commercial, educational and trading activities. A large number of buses and heavy goods vehicles in addition to small cars contribute in raising congestion severity, especially at a.m. and p.m. peak hours, since it connects the west of Baghdad with its center, as shown in figure 1. The previous study of this network in 2003 evaluated the flow data and suggested some proposals to improve the performance and decrease the congestion of the network, but no proposal of those was executed (because of the bad situation that faced Iraq after 2003).

33

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Figure 01. Location of the study area in Baghdad city Now, after ten years, the congestion increased and became more complex because of increasing in number of vehicles rapidly during this period, in addition to another reasons related to the security situation inside Baghdad which lead to make many of check points in the roads. In another hand there is no signal works good and continuously because the electric is available for a little time during the day (therefore, traffic policemen always control the flow at intersections manually when there is no electric). Figure 2 describes the network links, depending on the method of single node representation. It is also explain the directions of flow for each nodes and the distance of connection between them.

Figure 02. Description of the network links 34

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Traffic problems in the study area There are many factors cause traffic congestion problems along the roads and intersections in the study area such as heavy vehicles, poor geometry condition, turnings, the intersections themselves and other reasons which make the problem more complicated, for instance on- street parking and the buses. Data collection of the network After fixing the location and counting the links of the network and calculating their geometric parameters, now it is important to determine the peak hour period and so on the volume at peak hour. Determination of peak hour period: In order to illustrate the best time to collect data of peak volumes, volumes must be calculated for twelve hour continually for each intersection within selected network by taking the average volumes of working days a week (from Sunday to Thursday). On the other hand, personal discussions were made with traffic policemen and some users of the road. From all of these considerations, it was concluded that the peak hour of the network is at 7:30 – 8:30 A.M. Figure 3 shows two graphs of volume with time during twelve hours a day at 15 min. division for tow nodes, selected randomly, from the network. Counting traffic volume at peak hour The following table 1 shows the volume, direction and name of links at peak hour period. The operation of collecting the average traffic volume was manually for all the links of intersections at five working days in week (from Sunday to Thursday) during A.M peak hour (v/h). It is started in January 2014 and ended in April 2014. Volume counting was divided into 15- min periods.

Figure 03. Graphs of volume with time during 12 hours, with division every 15 min. 35

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

(a) Links name and direction Node no. Link no. Link volume

1 101

102

103

104

105

106

107

108

109

110

111

112

*2050

845

*2625

1165

940

1910

520

990

230

855

820

730

(*) This volume is passing the intersection by an overpass. Node no.

2

Link no.

201

202

203

204

205

206

207

208

209

210

211

212

Link volume

1775

350

1760

1790

1620

620

1135

1050

475

740

560

540

Node no.

3

Link no.

301

302

303

304

305

306

307

308

309

310

311

312

Link volume

1960

—

1610

2455

—

—

—

1055

1040

—

—

525

Node no.

4

Link no.

401

402

403

404

405

406

407

408

409

410

411

412

Link volume

—

—

—

1790

1675

2730

1370

—

—

760

—

670

Node no.

5

Link no.

501

502

503

504

505

506

507

508

509

510

511

512

Link volume

—

750

—

—

2710

—

1030

1125

1750

—

1340

—

Node no.

6

36

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Link no. Link volume

601 *1050

602

603

**

*970

340

604 **

1760

605 #

1930

606

607 #

1320

---

608

609

610

611

612

---

355

1020

650

430

(*) This volume is passing the intersection by a tunnel. (**) This volume will turn left around the square. (#) This volume will be distributed to through and left directions around the square. Node no. Link no. Link volume

7 701 #

1925

702 ---

703 #

1815

704

705

706

707

708

709

710

711

712

---

1245

1270

440

455

635

880

620

465

(#) This volume will be distributed to through and left directions around the square. (b) Volume of links v/h Table 01. Volume and direction of links at peak hour period for all nodes of network The adopted software program In order to study the case of traffic in the network, we need a suitable software program to represent the data and making simulation for it inside the program. There are many programs specialized in studying traffic networks. They are developed for analyzing and simulating microscopic and macroscopic networks, signalized and unsignalized. For example, CORSIM, TRAF-NETSIM, TRANSYT-7F, AIMSUN, and PTV-VISSIM which is the adopted one in this study. It is one of the PTV GROUP programs. We used VISSIM 6 which is a modern version and easy to use for representing the network by drawing all links and nodes and connect links with each others, in addition to setup the parameters and input vehicles and signal controllers... etc. RESULTS Simulation and Calibration After representing the network in adopted program and setting up all required data and parameters for simulation, it is necessary to make calibration between observed (actual) volumes and simulated volumes, in order to depend on this representation of the network and check the proposals of improving the performance of network. Simulated flow must represent the indeed flow closely as it is possible. Calibration is the most important step in studying network problems and trying to solve them. We can do it by using Theil’s U formula which contains both actual and simulated data as shown below in equation (1) [4].

(1) Where: n is the number of observations, yi is the ground truth value for observation i, and xi is the simulated or detector reported value for observation i. The results of U value will be between or equal 0 to 1. The best values are the nearest of 0 and the bad values are those nearest of 1. There is no exactly trustworthy value of U, but often less than or equal 0.2 values give good forecasted results. The following classification of U, as shown in table-2, is dependent by some specialists in simulation and forecasting data by interpreting the value (1 - Theil’s U) [5].

37

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Table 02. Interpreting Theil’s U value 1 - Theil’s U

Classification (interpretation)

From 1.00 to 0.80

Strong high forecasting power

From 0.80 to 0.60

Moderately high forecasting power

From 0,60 to 0.40

Moderate forecasting power

From 0.40 to 0.20

Weak forecasting power

From 0.20 to 0.00

Very weak forecasting power

Determination of dependent limits of values related with the accuracy required in the work. In this work high accuracy is required, so that we will adopt the first class shown in table- (strong high) especially with main links that affects the flow inside the network. In other hand, we can adopt the second class (moderately high) for the links that have less effectiveness on the network flow. For example the links resort to out of the network. After making simulation to the network by using adopted program (PTV-VISSIM) several times and checking the simulated volumes with the observed by Theil’s U, the results was obtained of current case. Table 03. shows U results for two congested intersections of the network. Node no.

1 101

102

103

104

105

106

107

108

109

110

111

112

7:30-7:45

461

190

590

262

211

429

117

222

51

192

184

164

7:45-8:00

553

228

708

314

253

515

140

267

62

230

221

197

8:00-8:15

594

245

761

337

272

553

150

287

66

248

237

211

8:15-8:30

441

181

565

251

202

410

111

213

49

184

178

158

Link no. Observed Vol.

Simulated Vol. 7:30-7:45 7:458:00 8:00-8:15 8:15-8:30

355 358 362 363

152 155 154 156

651 641 651 648

266 263 265 266

173 175 344 176 343 176 345 345

118 271 115 117 274 117 274 275

44 43 218 44 45 219 218 220

192 194 195 192

185 184 184 185

Theil’s U

0.188

0.170

0.063

0.080

0.160

0.176

0.085

0.077

0.148

0.063

0.065

0.062

1 - Theil’s U

0.812

0.830

0.937

0.920

0.840

0.824

0.915

0.923

0.852

0.937

0.935

0.938

(a) U results of node 1

38

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Node no.

7

Link no.

701

Observed Vol. 7:30-7:45

462

7:45-8:00

702

703

705

706

707

708

709

710

711

712

436

299

305

105

109

153

211

149

112

529

499

342

349

121

125

174

242

170

128

8:00-8:15

510

481

330

336

117

120

168

233

164

123

8:15-8:30

423

274

279

97

101

140

194

136

102

Simulated Vol. 7:30-7:45

332

398

258

262

103

104

106

234

128

105

7:45-8:00

335

399

261

265

102

106

109

234

129

104

8:00-8:15

334

401

259

265

101

104

111

235

127

107

8:15-8:30

336

261

263

101

105

113

236

122

106

Theil’s U

0.186

0.078

0.101

0.102

0.059

0.058

0.189

0.053

0.107

0.067

1 - Theil’s U

0.814

0.922

0.899

0.898

0.941

0.942

0.811

0.947

0.893

0.933

---

704

399

---

---

402

---

(b) U results of node 7 Table 03. Calibration results between observed and simulated volumes of nodes Evaluation of existing traffic network results The method used to evaluate the performance of the existing traffic network system is by simulating the patterns of this network. The simulation runs performed using PTV-VISSIM 6 program. The tables 4, 5 and 6 below show the results that evaluated for selected links, vehicle travel time and network performance. The nodes and zones are shown in figure 2 above. Table 04. Link segment results for important links selected in existing network Link no.

11

20

30

32

33

38

56

58

4-3

3-4

5-4

5-4

5-6

6-5

6-7

7-6

4

4

2

3

3

3

3

3

1392.9

1401.5

203.4

177.2

1016.8

779.0

2215.4

2313.7

277.1

276.2

114.6

245.1

29.4

49.5

241.5

119.2

Av. speed (km/h)

11.97

9.90

21.84

10.28

46.92

47.95

8.95

18.47

Av. loss time (s)

53.7

81.9

59.3

81.1

10.1

7.9

81.9

60.3

Between nodes No. of lanes Link length (m) Av. density (v/km)

Table 05. Vehicle travel time results for more congested paths selected in existing network Paths

Path 1

Path 2

Path 3

Path 4

Path 5

Path 6

1-4

1-5

8-5

8-4

10 - 7

7 - 10

Distance (m)

1280.5

2957.8

2337.4

3758.8

2850.2

2870.0

Av.tr. time (s)

409.7

1006.4

526.1

1121.6

525.6

1016.0

11.25

10.58

16.0

12.06

19.52

10.17

From zone to zone

Av. speed (km/h)

39

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

performance

Table 06. Evaluation of network performance results Total Average Average Total Average Delay Delay delay delay speed stops stops tot. stops av. stops (s) (s) (km/h) (s) (s)

Sim. Run 1

850144

130.51

13.03

117223

17.75

441470

68.05

Sim. Run 2

857531

130.73

13.01

122549

18.40

439271

67.29

Sim. Run 3

867236

132.27

12.73

124631

18.79

444055

67.95

Sim. Run 4

879623

132.96

12.69

126912

18.93

452045

68.60

Sim. Run 5

881247

133.08

12.68

125801

18.73

450075

68.29

Average

867156.2

131.91

12.83

123423.2

18.52

445383.2

68.04

performance Average

(a) Network performance results for five runs of simulation Total distance Total travel time Vehicles active Vehicles arrived (km) (s) 4034.4

1155470

3928

2593

(b) Other Average results of network performance from same simulation of (a) DISCUSSION This study deal with a very congested network having many security check points. Counting of volumes was collected manually by eye observing because it was not possible to use cameras or other instruments. The currency results of network obtained by VISSIM 6 give an idea for condition of the network, so that we can submit suitable proposals to improve the performance of the network. It is clear from results in tables 4, 5 and 6 above that most of the links have very high density and have level of service class F which is the worst operating condition of network flow, depending on HCM [6]. Also the average speed is very low which increase the travel time that gives rise in delay. There is two main reasons cause congestion in this network, the high volume and the security check points. In addition to other reasons, such as lack of optimum signal heads, geometric of intersections…etc. The solutions must focus on improve capacity of intersections by setting optimum signal heads and change geometric design for some of them, furthered more, it is very important to solve the delay due to check points, which cause big queuing leads to high density, delay and bad case of network flow. We recommend for future studies to use suitable approved instruments in collecting volumes. Also it is necessary to use update release of adopted software programs. REFERENCES [1] R. L. Dapre and J. P. Munro lafon, The Baghdad Transportation Study, Traffic Engineering Control, Vol. 28, No. 12, September 1987. [2] Mayoralty of Baghdad, By Japan International cooperation Agency, "The Baghdad City, Urban Transport Improvement Study, Progress Report (I)," December 1986. [3] Traffic Police of Baghdad, Department of Vehicles Data, "Tables of The Vehicles Numbers," April 2013. [4] Hourdakis, J., P. G. Michalopoulos and J. Kottommannil, "Practical Procedure for Calibrating Microscopic Traffic Simulation Models," Transportation Research Record: Journal of the Transportation Research Board of the National Academies, Washington, D.C., no. 1852, pp. 130-139, 2003. [5] D. C. E. "Statistical Forecasting Models," 27 may 2008. [Online]. [Accessed 6 march 2014]. [6] Transportation Research Board, National Academy of Science, 2000 Highway Capacity Manual, Washington , 2000.

40

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

A STUDY ON THE RHEOLOGICAL PROPERTIES OF RUBBER MODIFIED ASPHALT MIXTURES

M. Karacasu1, V. Okur1*, A. Er2

1

Eskisehir Osmgangazi University, Departmant of Civil Engineering, Meselik, Eskisehir, Turkey 2 Akdeniz University, Department of Civil Engineering, Antalya, Turkey *Corresponding author:[email protected], Phone: 90 222 239 3750, Fax: 90 222 239 36 13

ABSTRACT Using waste rubber in asphalt mixes has become a common practice in road construction. This paper present the results of a study on the rheological characteristics of rubber-modified asphalt (RMA) concrete under static and dynamic loading conditions. A number of static and dynamic creep tests were conducted on RMA mix specimens with different rubber sizes and contents, and a series of resonant column tests were conducted to evaluate the shear modulus and damping values. To simulate the stress-strain response of traffic-induced loading, the measurements were taken for different confining pressures and strain levels. The results of the study indicated that rubber modification increases stiffness and damping ratio, making it a very attractive material for use in road construction. However the grain size of the rubber is very important. Although RMA may cost up to 100 % more than regular asphalt, the advantages it brings, such as an increased service life of the road and proper waste utilization contributing to a more sustainable infrastructure, may justify the added cost. Keywords: Damping, Dynamic Creep, Rubber-Modified Asphalt (RMA), Static Creep, Stiffness INTRODUCTION Asphalt concrete is the leading paving material for roads and runways. Understanding the characteristics of the asphalt being used in a project is important to ensure long-term performance and stability.Various environmental conditions and traffic-induced loadings must be taken into account in the design stage. Sometimes the design does not result in acceptable, safe and sound usability due to early deterioration. Several factors contribute to deterioration, such as the quality of materials and construction, traffic loading on the road, road geometry, and environmental conditions. In general, most of the deterioration is due to rutting or bottom up fatigue and thermal cracking. Enhancing the pavement life is possible with some modifications if possible factors causing deteriorations are taken into consideration during the design stage. Plastomeric polymeric materials, such as polyethylene (PE) and polypropylene (PP), have evoked considerable interest among engineers and manufacturers for use in road paving modification because of their viscoelastic properties and good adhesion to mineral aggregates (Al-Hadidy and Tan 2011, Pasetto and Baldo 2010, Yoon et al 2006, Metcalf et. Al. 2000,). The main purpose of using polymers in asphalt concrete is to increase binder stiffness at high service temperatures and reduce stiffness at low service temperatures (Chen and Qian, 2003, Mull et al. 2002). Polymers used for the modification of asphalt concrete can be divided into three main categories: thermoplastic elastomers, plastomers, and reactive polymers. Thermoplastic elastomers are apparently capable of high elastic response characteristics and therefore resist permanent deformation by stretching and recovering their initial shape on the modified binder layer, whereas plastomers and reactive polymers modify asphalt by forming a tough, rigid, three-dimensional network to increase stiffness and decrease deformations (Zhang et al. 2009, Airey, 2004). Because of its suitability in these conditions, one of the leading polymer modifiers for bitumen among the larger group of copolymers is styrenebutadiene-styrene (SBS) block. SBS is a synthetic hard rubber copolymer that is used in applications where durability is important and is often substituted in part for natural rubber based on the comparative raw materials' costs (Alonso et al. 2010, Omran et al. 2009, Lu and Isacsson 2001, Fawcett and Nally 2003, Airey 2003, Navarro et. al., 2002, Stastna, 2004). SBS is a cost-effective material that is used to stretch dwindling natural rubber resources, especially in tire manufacturing. The disposal of used automobile tires, however, has caused many environmental and economic problems. The annual global production of used tires estimated 17 million tonnes in which China, European Union Countries, USA, Japan and India lead the way to produce the largest amounts of tyre (Leung and Wang 1998, Sienkiewicz et al. 2012). A small percentage of these tires are recapped or reused as lower-quality rubber, but around 80% of these tires are accumulated in dumps, posing health hazards and adversely impacting the environment (Mastral et al 2002).

41

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians Due to its elastic nature, crumb rubber can be used in road construction to improve deformation resistance. Rubbermodified asphalt (RMA) is a bituminous mix consisting of blended aggregates, recycled crumb rubber and bitumen. Crumb rubber is a granulated tire rubber from which the fabric and steel belts have been removed. The processing method such as ambient grinding, buffing, rasping, and cryogenic grinding, significantly affects the reaction of the rubber with asphalt and the resultant properties of the asphalt-rubber binder. It is found out that rubber structure is the most important factor affecting elastic properties. Crumb rubber has a granular texture and ranges in size from a very fine powder (<0.1 mm) to coarse particles (>5 mm). Due to its low specific gravity, crumb rubber has been successfully used as an alternative lightweight aggregate in asphalt concrete (TNRCC 1999, Pierce and Blackwell 2003). RMA has been known to improve the rheological properties at low and high temperatures and provides a lifespan that is up to three times longer than conventional asphalt (Borah and Wang, 2012). There are also other reasons why rubber is useful in both highways and railways, such as decreasing thermal instability, increasing resistance to low-temperature cracking, reduction of noise levels, and reduction of vibrations generated by heavy axle loads (Roschen, 2000, Azizian et al, 2003, Zhong et al. 2002,). Modelling of RMA behaviour in compliance with strain-dependent deformation characteristics is important in the design process. Despite several researches in the field a review of the literature did not provide any insight into the efffect of tire shape and size. The main objective of the this research was to study the performance of asphalt concrete with or without waste rubber considering the shape and size of waste rubber, in order to evaluate their efficiency. The study was carried out in the laboratory, to demonstrate, through, that crump rubber from scrap tires can be used as a high quality modifier agent in asphalt concrete. MATERIALS AND METHODS The focus of the test program was to examine the performance of asphalt concrete when mixed crumb rubber as aggregate. The study concentrated on the admixture mechanism rather than bitumen additive of crump rubber. Same type rubber in different sizes and shapes were added to the mix without any modification to the bitumen content. The digestion mechanism of the rubber sizes is not evaluated in this paper, but should be considered in the future research. The appropriate aggregate gradation for hot-mix bitumen was designed according to the technical specifications of the General Directorate of Turkish Highways, (GDTH) 2006. The aggregates have a mean grain size (D50) between 0.303.0 mm and a coefficient of uniformity (Cu) between 2.0-3.0. The margins of the GDTH and the prepared grading curves are given in Fig. 1, and the physical characteristics of the aggregates are given in Table 1.

100

Sample lower limits

Passsing, (%)

80

upper limits

60

40

20

0 100.000

10.000

1.000 Sieve size, (mm)

0.100

0.010

Figure 01. Aggregate grading curves for asphalt mixtures compared with the current GDTH Table 01. Physical characteristics of the aggregates used in the tests Properties Test Values Standards o 3 Specific gravity of coarse aggregate, 25 C, gr/cm 2.62 ASTM C127-07 Water absorption of coarse aggregate, % 0.23 ASTM C127-07 Specific gravity of fine aggregate, 25 oC, gr/cm3 2.622 ASTM C128-07a Water absorption of fine aggregate, % 1.04 ASTM C128-07a Specific gravity of filler, 25 oC, gr/cm3 2.708 ASTM C128-07a Los Angeles wearing test , % 28.91 ASTM C535-09 Freezing and thawing test, % 5.467 ASTM C1646-08a Bitumen absorption, % 0.14 ASTM D4469-01 42

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

To determine the optimal bitumen content, Marshall stability tests and flow tests are performed according to the ASTM D 1559-76 specifications. The optimum bitumen ratio was found to be 4.89 % for the 50/70 penetration grade. The crumb rubber selected for this investigation was produced from recapping automobile tires. The bitumen was modified with three different sizes of tire rubber by-products: Type-I (granulated tire rubber that passes a #3/8 mesh), Type-II (515 mm pieces of chipped tire rubber), and Type-III (powdered tire rubber that passes a #40 mesh). All rubbers were obtained by shredding and grinding the tire after removing the fabric and steel belts. Photographs and SEM images compares the crumb rubber gradation and surface texture for three types in Fig. 2 and Fig. 3, respectively. Table 2 shows the physical properties of the virgin binder. The characteristics of the RMA depend on the concentration amount and the polymer type in which the polymer is mixed in concentrations of approximately 0.2-1.0% by weight of aggregates. Higher concentration mixes of polymers can be less economical and may also cause problems related to the material properties (AI and Yi-qiu 2011). Rubber contents of 0.2%, 0.4%, 0.6%, 0.8%, and 1.0% by weight of aggregate were blended with bitumen for each rubber size at a mixing temperature of approximately 160 °C. To verify the repeatability of the tests, three specimens are prepared using an identical procedure (premixing the rubber with bitumen using a mixer at 500 rpm for 2 minutes) from each mix. Visual observation of the flowable bitumen produced revealed that the crumb rubber was well mixed in the fluid state without clumping. A total of 90 modified specimens were used to determine the optimum rubber amount for each type. In addition, 3 control specimens with no additives were prepared for comparison with the modified specimens.

Figure 02. Tire rubber after the shredding and grinding process

(a)

(b)

43

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

(c) Figure 03. Scanning electron micrographs of waste tire: a) Type-I, b) Type-II, c) Type-III Table 02. Properties of bitumen Properties

Related Standards o

Penetration at 25 C, 1/10mm o

Ductility at 25 C, cm Loss on heating, % o

Specific gravity at 25 C, gr/cm o

Softening point, C o

Flash point, C o

Elastic recovery, % (25 C)

3

57.3

ASTM D 5-06e1

>100

ASTM D 113-99

0.17

ASTM D 6-95

1.035

ASTM D 70-03

48.0

ASTM D 36-09

308

ASTM D 92-02b

2.95

ASTM D 6084-06

The average values obtained from the Marshall tests are summarized in Fig. 4 and Fig. 5. The solid line in each figure shows the boundary value for the control specimen. Fig. 4 shows the variation in air voids and unit weight against the rubber content. Note that increasing the rubber content tends to increase the void ratio. Furthermore, the void ratio at a given rubber content tends to increase with the increasing size of the rubber particles. This tendency is also coincident with the unit weight.

44

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

8.00

23.60 Control

Air voids (%)

Type-II

23.80

Type-III

6.00

24.00

4.00 24.20

2.00

Unit weight (kN/m3)

Type-I

24.40 2

4

6

8

10

Rubber content (%) Figure 04. Variation in air voids and unit weight with respect to rubber content Fig. 5a shows the effect of rubber content on Marshall stability, which has a general tendency to decrease as rubber content increases. The initial increase of the Marshall stability with increase of rubber content for Type-III can be attributed to the size effect of the rubber. The Marshall quotient has a tendency to decrease with increasing rubber content for all modified specimens except Type-III, where the Marshall quotient value is almost same with increasing rubber content (Fig. 5b). It is seen that the addition of 0.4% rubber have the most significant impact on the Marshall characteristics of the specimens, and these contents are therefore used for the subsequent creep and resonant column tests.

25000

Marshall Stability (N)

(a)

20000

Control 15000

Type-I Type-II Type-III

10000 2

4

6

8

10

Rubber Content (%)

45

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

4.00

0

3.00

2000

2.00

4000 Control Type-I

1.00

6000

Type-II

Marshall Quotient (N/mm)

Marshall Flow (mm)

(b)

Type-III

0.00

8000 2

4

6 8 Rubber Content (%)

10

Figure 05. Effect of rubber content on a) Marshall stability and b) Marshall flow and quotient 3.1 Creep Tests and Results For practical purposes, the method of load application for creep testing in a laboratory environment can be classified into two types: static and dynamic loading. The static loading condition simulates a heavy vehicle, such as a truck, standing on a pavement specimen and applying a static stress by waiting at a red light. Static creep tests were performed in the laboratory to evaluate the response of an asphalt specimen for such a condition. In a static creep test, the measured data are the deformation time, which is the length of time the pavement can resist the static load until flow occurs. On the other hand, repetitive loading simulates the driving of a heavy vehicle over a specimen of pavement. This condition can be reproduced by applying a dynamic load to an asphalt specimen. The output data for a repetitive test are the number of load cycles the pavement can tolerate before it fails. This test is destructive, so the asphalt specimen can be tested only once. For both tests, the specimens were prepared according to EN12697-25A specifications, with a diameter of 100 mm and height of approximately 63.5 mm. The testing temperature was set to 50 C, and the specimens were kept in a climatic cabinet for 24 hours. 3.2.1 Static creep test To enable perfect coupling between the specimen and top platen a static axial stress of s=5 kPa was applied for ten minutes. After the application of pre-stress, the specimens were axially loaded to a value of 500 kPa in approximately 1 h. The corresponding static creep stiffness plotted against the loading time is shown in Fig. 6. One of the outcomes seen in Fig. 6 is that the static creep stiffness is not appreciably affected by the duration of loading after a certain time. However, adding rubber to the mixture gradually changes the strength properties of the specimens. This fact can be explained by considering the structure of the specimen. An asphalt specimen is composed of an assembly of aggregates and bitumen, where inter-granular forces are transmitted through points of contact. When rubber is added, the resulting mixture is not always homogeneous at all contact points due to the size and shape of the rubber. Furthermore, Type-I and Type-II rubber types have relatively coarse-grained rubber crumbs, which have considerably smaller specific surface areas for a given size particle than in Type-III, and do not fully dissolve in the bitumen mixture, thereby increasing heterogeneity and void ratio. Reacted asphalt rubber materials have drastically different properties compared to unreacted asphalt rubber. The resolving of rubber increases the viscosity of bitumen and causes binding and reinforcement effects. Type-III fine powder rubber, which spread and dissolved homogeneously into the mixture, stuck to aggregate surfaces better, as seen in Figure 6.

46

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Type-I

Creep Stiffness (MPa)

600

Type-II 500

Type-III Control

400 300 200 100 0 0

1000

2000

3000

4000

5000

Loading Time(sec)

Figure 06. Static creep stiffness curves with respect to rubber type and time Static loading causes slip-down movement of aggregates, which reduces the volume of the specimen by repacking the aggregate into a denser state. When slip-down occurs, the aggregates fill the gaps in the void and do not move in the direction of loading. It is for this reason that the increase in strain is generally observed at an early stage of loading in each test, as seen in Fig. 7. In Control specimen, the accumulated strain value is 0.46% at the end of the test. In rubber modified specimens, however, the slip-down movement occurs rather easily due to the rubber between the aggregates. The accumulated strain values are approximately ¼ of the control specimen.

Accumulated Strain (%)

0.6 0.5 0.4

Control Type-III Type-II Type-I

0.3 0.2 0.1 0 0

1000

2000

3000

4000

5000

Loading Time(sec)

Figure 07. Variation of accumulated strain with respect to rubber type and time

47

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians 3.2.2 Dynamic Creep Test In a dynamic creep test, a repeated uni-axial stress is applied to an asphalt specimen for a number of load cycles while axial strain is measured in the same direction as the loading using Linear Variable Differential Transducers (LVDTs). The applied dynamic load used in this test was a sequence of rectangular pulses. The pulse duration was 0.5 seconds, and the rest period before the next pulse was 1.5 seconds. A static axial stress of s=5 kPa was applied for ten minutes to the top platen of the specimen for proper bedding, as in a static creep test. The deviator stress repeated loading was 500 kPa, and the testing temperature was set to 50 ºC. The failure criterion was defined as 5 % of axial strain or complete failure, whichever occurred first. The characteristic change in dynamic creep stiffness can be interpreted as follows. As the number of cycles increases, the dynamic creep stiffness decreases. It is interesting to note that the dynamic creep stiffness tends to decrease with an increasing number of cycles only during the first 200 cycles; thereafter, the dynamic creep stiffness reduction becomes negligibly small (Figure 8). Progressive reduction in dynamic creep stiffness is obvious for all test specimens. According to Fig. 8, number of cycles for failure for unmodified and modified specimens is not same due to elastic behaviour of rubber. Failure occurred at the 2500 th cycle for Type-II and Type-III whereas 1400th cycle for the Control group. Accordingly axial strains become considerably large in modified specimens. However, failure does not occur within the same strain. The loading continued until the magnitude of axial strain increased to a level of approximately 7.5 % for Type-II and Type-III, as shown in Fig. 9. The results of the tests indicate that the shear stiffness of the Control specimens gave the lowest value compared to those in the Type-I, Type-II and Type-III, which can be attributed to the fact that when rubber and asphalt are mixed at high temperatures, such as 145-170 oC, the rubber particles may swell. Swelling has been postulated to occur as a result of physical and chemical interactions between rubber particles and asphalt as well as the reaction between the asphalt and the rubber, which results an increase in viscosity of the asphaltrubber mixture as stated in the previous paragraphs. Furthermore, the imperfect coupling between the rubber and aggregates due to swelling causes bigger void ratios, which produce somewhat larger axial displacements than the Control specimen.

Creep Stiffness (MPa)

1000 Control Type-I Type-II Type-III

100

10

1 0

500

1000

1500

2000

2500

3000

Pulse Count Figure 08. Variation in creep stiffness with rubber type and pulse count

48

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

9

Accumulated Strain (%)

7.5 6 4.5 3

Control Type-I

1.5

Type-II Type-III

0 0

500

1000

1500

2000

2500

3000

Pulse Count Figure 09. Variation in accumulated strain with rubber type and pulse count Fig. 10 shows normalized creep stiffness plotted versus axial strain. It is interesting to note how the dynamic creep stiffness tends to change continuously with the axial strain. The normalization is made by taking the dynamic creep stiffness, CSd, at any time divided by the initial dynamic creep stiffness value, CSo. The deformation characteristic features for the Type-I, Type-II, Type-III and Control specimens are more vividly witnessed in this plot. It is apparent from the Fig. 10 that the dynamic shear stiffness decreases suddenly to approximately 99 % of the initial value in the Control specimen and 97% in the Type-I and Type-II specimens, for a value of 1% axial strain. The axial strain at the time of failure is 4.2% for the Control specimens and 8.5 % and 6 % for the Type-I and Type-II specimens, respectively. Adding rubber to the asphalt shifts the curves to the left and increases the strain value at failure. The same results are observed for Type-III, but the reduction from the original value is only 50 % for 1% axial strain, and does not occur as suddenly as in the other rubber-modified specimens. Moreover, the strain rate for Type-III at the time of failure is almost the same as the other specimens. The gradual shift of the curves to the left implies the increase in elastic response due to the rubber content. The graph also indicates that the rate of dynamic creep stiffness reduction with axial strain becomes greater as the void ratio decreases, implying that the aggregates are more rigidly bonded together in the Control specimen than in the modified specimens.

49

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

1

Dynamic Creep Stiffness Ratio

Control, Air voids=3.65% Type-I, Air voids=3.34% 0.8

Type-II, Air voids=3.05% Type-III, Air voids=2.95%

0.6

0.4

0.2

0 0

2

4

6

8

10

Axial strain (%) Figure 10. Variation in dynamic creep stiffness with strain 3.3 Resonant Column Tests and Results Characteristics of the specimens under cyclic loading were studied in two parts. The first involves the determination of the maximum shear modulus, which is evaluated on the order of 10-4-10-3 % strain with a resonant column (RC) test. The RC device is the most commonly used laboratory test to measure low-strain dynamic properties of soils, concrete, and rocks. The test data analysis is described in detail by Drnevich (1985) and uses the ASTM D 4015 standards. The RC test configuration is a fixed-free system where the specimen is fixed at the bottom and free to rotate at the top at its fundamental frequency via a drive system. By measuring the motion of the free end, the velocity of the propagating wave and the degree of material damping can be derived. The shear modulus is then obtained from the derived velocity and the density of the specimen. The test specimen is a solid cylindrical specimen with an approximate diameter-height ratio of 2. The bottom is fixed to the base of the apparatus. Sinusoidal torsional excitation is applied to the top of the specimen by an electric motor system. A torsional harmonic load of constant amplitude is applied over a range of frequencies, and the response curve (strain amplitude) is calculated. The output angular acceleration at the top of the specimen is recorded by an accelerometer. The frequency of the cyclic torque is automatically and gradually changed until the first resonance of torsional vibration is obtained. The shear wave velocity is obtained from the first-mode resonant frequency, and the shear modulus is then calculated using the shear wave velocity and the specimen density. The shear modulus and damping ratio were measured under a range of shear strains. The power is shut off at resonance (that is, forced vibration is removed), and the material damping is determined from the free vibration decay. The entire system is placed into a Perspex chamber to apply a uniform confining pressure on the specimen using air pressure. A membrane covers the setup to prevent diffusion of air into the specimen. Identical fresh specimens were prepared using the same procedure as for the Marshall stability tests. After the 300 mm diameter cylindrical asphalt specimen had cured, it was cored into a standard size with a diameter of 70 mm for the resonant column test. The height of the specimens was approximately 140 mm. The specimens were fixed onto the bottom pedestal using cyanoacrylate-based fast-acting adhesive. Because the strength and rigidity of the adhesive are higher than those of the asphalt, they have almost no effect on the testing data. After the adhesive was cured, the RC device was set up. Each specimen was tested in sequence with stepwise increased confining pressure. At each confining pressure, cyclic torque was applied to measure the shear modulus, G, and the damping ratio, D. The vertical pressure on the subgrade under a road is between 50-150 kPa when a car or loaded truck axle passes over it. Thus, the tests were conducted by employing four confining pressures of c=0, 50, 100 and 150 kPa. After the adjustment of each confining pressure in each test, the cell pressure was maintained for 30 minutes to allow for volume change of the specimen before the test started.

50

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

3.4. Modulus and Damping Characteristics of Rubber-Modified Asphalt Due to its highly elastic nature, the responses of rubber-modified mixtures are expected to show more elastic behavior with increasing rubber content under cyclic loads. The results from the test specimens are for shear strains less than approximately 0.0006 %. It was impossible to achieve higher strains due to the torque limitation of the RC device. This limitation is satisfactory because ground vibrations produced by vehicles are assumed to induce strains in the lowamplitude range levels (i.e., less than 0.001 %). The calculated shear modulus with respect to shear strain is shown in Fig. 11. As shown in Fig. 11, the Type-III specimens have the largest shear modulus values compared to the Control specimens and the other rubber-modified specimens due to lower air void ratios and perfect coupling between the rubber and aggregates (Figures on the left side of Fig. 11). The difference in air void ratios of rubber-modified specimens could have contributed to the increase in shear modulus; however, for the Type-I and Type-II specimens, which had the similar air void ratios, the stiffness did increase. The shear modulus of rubber-modified specimens was somewhat lower and the damping ratio was considerably higher than those of the Control specimens at corresponding confining pressures. Thus, it can be concluded that adding a certain amount of rubber to an asphalt mix can slightly decrease the shear stiffness, whereas significantly increases damping. Increasing the confining pressure from 0 to 150 kPa increased the shear modulus by approximately 20%. The initial shear modulus increases noticeably in all cases with an increase in confining pressure; however, the rate of increase becomes small after the first incremental stage (from 0 to 50 kPa), and diminishes after 100 kPa. The results agree with the characteristic properties of asphalt obtained from other tests such as the Marshall stability test. The damping ratio is an important characteristic of a material because it indicates how much vibration energy is absorbed during a vibration cycle. If a material has a high damping ratio, attenuation of vibration will also be high. On the other hand, it is not easy to define true material damping, but it is common to express the damping of real materials in terms of their equivalent viscous damping ratios. The viscous damping ratio, D, is measured in the resonant column test from the shape of a free vibration decay curve. This curve is measured using the accelerometer mounted on the resonant column drive plate. A sinusoidal wave is applied to the soil, after which the excitation is shut off so the resulting free vibrations can be measured. The value of the damping ratio obtained in the same test series is shown on the right side of Fig. 11 plotted against the shear strain levels. In all of the figures, the damping ratio increases slightly with increasing shear strain, irrespective of whether the specimen has been modified, and also independently of rubber size. It is also obvious from the figures that the damping ratio increases due to the confining stress becoming more pronounced with increasing rubber content and decreasing rubber particle size. 6.0

Control

Control

1300

Damping ratio, D (%)

Shear modulus, G (MPa)

1500

1100

900

150 kPa 100 kPa

700

5.0

4.0 150 kPa 100 kPa

3.0

50 kPa

50 kPa

0 kPa 500 0.00001

0.0001

0.001

Shear Strain (%)

0 kPa 0.01

2.0 0.00001

0.0001

0.001

0.01

Shear strain (%)

51

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians 7.0

Type-I

Type-I

6.0

1300

Damping ratio, D (%)

Shear modulus, G (MPa)

1500

1100

900

150 kPa 100 kPa

700

5.0

150 kPa

4.0

100 kPa 50 kPa

3.0

50 kPa

0 kPa

0 kPa 500 0.00001

0.0001

0.001

2.0 0.00001

0.01

0.001

7.0

1500

6.0

1300

1100

150 kPa

900

100 kPa

5.0

150 kPa

4.0

100 kPa 3.0

50 kPa

700

50 kPa

0 kPa 500 0.00001

0.01

Type-II

Type-II Damping ratio (%)

Shear modulus, G (MPa)

0.0001

Shear strain (%)

Shear strain (%)

0.0001

0 kPa

0.001

2.0 0.00001

0.01

0.0001

1500

0.001

0.01

Shear strain (%)

Shear strain (%) 9.0

Type-III

Type-III

1300

Damping ratio, D (%)

Shear modulus, G (MPa)

8.0

1100

900

150 kPa 100 kPa

700

7.0 6.0 5.0

150 kPa

4.0

100 kPa

50 kPa

50 kPa

3.0 0 kPa 500 0.00001

0.0001

0.001

Shear strain (%)

0 kPa 0.01

2.0 0.00001

0.0001

0.001

0.01

Shear strain (%)

Figure 11. Effect of confinig stress and rubber type on shear modulus and damping ratio 4. CONCLUSIONS The performance characteristics of rubber-modified asphalt mixtures with for the same type of rubber having different sizes and texures were investigated using static creep, dynamic creep and resonant column tests. The mix design was performed based on the Marshall method and the optimum bitumen content was determined. Specimens were prepared at a proportion of 0.4%, by weight of aggregates. The results can be summarized as follows. The aggregates in the asphalt concrete are extremely stiff, and therefore dissipate very little energy in particle deformation. In contrast, the rubber consumes energy through deformation of the rubber particles themselves. It is seen that, no matter the size of the crumb rubber, the static and dynamic stiffness decrease with any proportion of rubber in the asphalt. However, modified asphalt improves longevity compared to the Control specimens. Creep stiffness is dependent on the type of loading, regardless if it is static or dynamic. Furthermore, creep stiffness is highly dependent on the level of axial strain at which the creep value is determined. Creep stiffness is also strongly influenced by the rubber particle size and texture, and the particle size affects the shear modulus and damping ratio. 52

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians The shear modulus decreased by 20% in the Type-I and Type-II specimens compared to the Control specimen. However, the greatest increase in shear stiffness, 10% compared to the Control specimen, occurred in the Type-III specimens. On the other hand, the damping increased by 30 % in the Type-I and Type-II specimens and 40% in the Type-III specimens. The shear modulus of the specimens decreased slightly with shear strain in all tests, as expected. The shear modulus of the Control specimens was somewhat higher than that of specimens Type-I, Type-II and Type-III at the corresponding confining pressures. The weak interaction between the rubber particle surface and the asphalt changes under static or dynamic loads. Increased specific surface increases the reaction rate with hot asphalt, as in the case of the Type-III specimens. It can be concluded that the rubber-modified asphalt concrete mitigates vibrations generated by traffic loading and results in reduced damage from cyclic straining. The use of polymer-modified bitumen, compared to penetration-grade bitumen, provides improved longevity and marked whole-life cost benefits, increasing the sustainability of pavements. As the crumb rubber content increases, the bulk density of asphalt concrete decreases due to the low specific gravity of the crumb rubber. This rubber also offers other benefits when used in asphalt concrete, such as thermal insulation, which delays the decrease of bitumen’s internal heat during transportation from the plant to the site. However, these and other technical benefits must be explored in future studies. ACKNOWLEDMENTS This work was supported in part by Scientific Research Project Unit of Eskisehir Osmangazi University through project No:200915006. REFERENCES Airey, G.D., (2004), Styrene butadiene styrene polymer modification of road bitumens, Journal of Materials Science, 39, 951-959. Airey, G.D., (2003), Rheological properties of styrene-butadiene-styrene polymer modified road bitumens. Fuel, 82, 1709-1719. Al, A.H., Yi-qiu, T., (2011), The Effect of Plastomers Polymer S and Concentration on Asphalt and Moisture Damage of SMA Mixtures, Al-Rafidain Engineering, 19, 1-11. Alonso, S., Medina-Torres, L., Zitzumbo, R., Avalos, F., (2010), Rheology of asphalt and styrene-butadience blends, Journal of Materials Science, 45, 2591-2597. ASTM D4015, (2007), Standard Test Methods for Modulus and Damping of Soils by Resonant-Column Method, On line at: http://www.astm.org/Standards/D4015.htm. Azizian, M.F., Nelson, P.O., Thayumanavan, P., Williamson, K.J., (2003), Environmental impact of highway construction and repair materials on surface and ground waters: Case study: crumb rubber asphalt concrete, Waste Management, 23, 719-728. Borah, J., Wang, C., (2012), Morphological and flame retardant behaviors of rubber-modified asphalt, Key Engineering materials, 501, 532-537. BS EN 12697-25, (2005), Bituminous mixtures, Test methods for hot mix asphalt, Cyclic compression test, On line at:http://www.ipcglobal.com.au/products/standard/bs-en-12697-25.html. Chen, F., Qian, J., (2003), Studies of the thermal degradation of waste rubber, Waste Management, 23, 463-467. Drnevich, V.P., (1985), Recent developments in resonant column testing, Proc. of Eichart Commemorative Lectures, ASCE, 79-107. Fawcett, A.H., Nally, T.M., (2003), An Attempt at Engineering the Bulk Properties of Blends of a Bitumen with Polymers, Colloid Polymer Science, 181, 275–286. Highway Technical Specifications (2006), General Directorate of Highways, Item No. 170/2, Ankara, Turkey. Leung D.Y.C., Wang, C.L., (1998), Kinetic study of scrap tire pyrolysis and combustion, Journal of Analytical and Applied Pyrolysis, 45, 153-169. Lu, X., Isacsson U., (2001), Modification of road bitumens with thermoplastic polymers, Polymer Testing, 20, 77-86. Mastral, A.M., Murillo, R., Garcia, T., Navarro, M.V., Callen, M.S., Lopez J.M., (2002), Study of the viability of the process for hydrogen recovery from old tyre oils, Fuel Processing Technology, 75, 185-199. Metcalf, J.B., Gopalakrishnan, K., Waters, M.D., (2000), An initial investigation of the use of a rubber waste (EPDM) in asphalt concrete mixtures, Waste Management, 1, 940-952. Mull, M.A, Stuart, K., Yehia, A., (2002), Fracture resistance characterization of chemically modified crumb rubber asphalt pavement, Journal of Materials Science, 37, 557-566. Navarro, F.J., Partal P., Martínez-Boza F., Valencia C., Gallegos C., (2002), Rheological characteristics of ground tire rubber-modified bitumens, Chemical Engineering Journal, 89, 53-61. Omran, A., El-Amrouni, A.O., Suliman, L.K., Pakir, A.H., Ramli, M., Aziz, H.A. (2009), Solid waste management practices in Penang State: A review of current practices and the way forward, Environmental Engineering and Management Journal, 8, 97-106. Pasetto, M., Baldo, N., (2010), Recycling of steel slags in road foundations, Environmental Engineering and Management Journal, 9, 773-777. 53

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians Pierce, C.E., Blackwell, M.C., (2003), Potential of scrap tire rubber as lightweight aggregate in flowable fill, Waste Management, 23, 197–208. Roschen, T., (2000), Report on the status of rubberized asphalt traffic noise reduction in Sacramento County, Report prepared for Sacramento County Public Works Agency. Sienkiewicz, M., Kucinska-Lipka, J., Janik, H., Balas, A., (2012), Progress in used tyres management in the European Union: A review, Waste Management, 32, 1742-1751. Stastna, J., Zanzotto, L., Vacin, O.J., (2003), Viscosity function in polymer-modified asphalts Journal of Colloid and Interface Science, 259, 200-207. TNRCC, (1999), The many uses of crumb rubber. Texas Natural Resource Conservation Commission, Waste Tire Recycling Program, Office of Permitting, On line at: www.tceq.state.tx.us/assets/public/compliance/tires/docs/crumb.pdf Yoon, S., Prezzi, M., Siddiki, N.Z., Kim, B., (2006), Construction of a test embankment using a sand–tire shred mixture as fill material, Waste Management, 26, 1033-1044. Zhang, S.L., Xin, Z.X., Zhang, Z.X., Kim, J.K., (2009), Characterization of the properties of thermoplastic elastomers containing waste rubber tire powder, Waste Management, 29, 1480-1485. Zhong, X.G., Zeng, X., Rose, J.G., (2002), Shear Modulus and Damping Ratio of Rubber-modified Asphalt Mixes and Unsaturated Subgrade Soils, Journal of Materials in Civil Engineering, 14, 496-502.

54

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

MODELING OF TRAVELLING TIME FOR PLANNING VERTICAL PROFILE OF HIGH-SPEED RAILWAYS

B. A. NOOR, Y. SIRONG

[email protected] , [email protected]

Collage of Civil Engineering, Southwest Jiaotong University ABSTRACT Recently, high-speed railway has become a hot research topic of railway development in China. Many countries in the world have built relatively complete high-speed railway networks. Each of them enjoys efficient traffic organization system and method, as well as a theory of station distribution layout. One of the important operational indicators of railways is running speed and running time of train interval, and it also an important indicator to evaluate and design the railway location and estimate operational expenses. How to align the distribution layout of the stations developed during the high-speed railway’s construction process with unique traffic organization model, sectional passing capacity, and travel speeds of various types of trains, has attracted great attentions at present. This paper calculates travelling time between two following stations depending on changing of gradient of the vertical profile. From the results, we get a time consumption model for six standrad vertical profiles. Each model is a quadratic function for particular profile. The model results proved that travelling time increases with the increasing of gradient value. Keywords: High-speed railway, Time consumption model, Vertical profile. INTRODUCTION In the last half century, the world has enjoyed rapid development of high-speed railway (HSR). Recent HSR infrastructure is the result of specialized development techniques in this domain which results into expanded the passenger traffic system. [2] The high-speed train is a type of passenger rail transportation that operates significantly faster than the normal speed of rail traffic [4] and plays exceptionally an important role in our daily life. According to the specific definition of the European Union, high-speed railway refers to the transformed railway tracks whose operating speed can be 200 km/h or more and the new specialized railway tracks whose speed can be 250 km/h or more [4]. As the social and economic benefits are one of the features of high-speed railways beside many features like high-speed, large passing capacity, high efficiency and less pollution, therefore, many studies on efficient energy management have been carried out in metro and rail transit systems [5], [7]. The reduction of energy consumption is also seen as one of the key objectives for the development of sustainable mobility by use of high-speed train. The completion of a high-speed train project will lead to a huge increase in electricity consumption in spite of the comparatively lower mean energy consumption per passenger kilometer [3]. However a high-speed train project requires a very heavy investment for a country, mainly due to the cost of civil works. Therefore, high investment must be balanced by a shorter trip time and lower energy consumption. In this context, more than 40 years ago, a simplified method of calculation of energy and time consumption and operating expenses were derived by Wang Di, et. al [8] in the form of mathematical models and graphs. The calculation work was done in cooperation with engineers of the Third Design Institute of Chinese Railway Ministry and was proved satisfactory in practical designs. The theoretical idea and the main procedure are still useful today. Referring to large amount of design data, the profiles between passing sidings or intermediate stops may be classified into six types of imitation. See Figure (1).

55

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Figure 01. Imitating types of railway profiles. THEORETICAL STUDY Gradient value: The maximum gradient (imax) is calculated by using the formula (1) [13]: (1) where, Pk: required power per ton (kW/t). Vmax: Maximum speed of train (km/h). g: Gravity acceleration (g = 10 m/s2). w0: Basic resistance of maximum speed of running train (N/kN), and for calculating this resistance for China Railway High-speed (CRH) Electrical Multiple Unit (EMU), this paper will use the formula (2) [13]: w0=0.66 + 0.00245V + 0.000132V2 (2) Time consumption: Time consumption of train running between station A and station B is calculated by using the formula (3) [13]: TA/B = (ti.Li) + ts + tp (min) (3) where: ts, tp: extra time consumption of starting and parking of a train. In general, for electric traction and diesel traction: ts = 1-3 min, tp= 1~ 2 min. Li,: Length of slope (km) ti: Time consumption per kilometer on a gradient section (min/km), for ascending, ti= 60 / Vi, Vi is the balancing speed and can be calculated from unit force graph. But as the braking system of high speed train is activated by using computer-controlled comprehensive braking mode, and it doesn’t has the concept of train converted braking ratio and brake lining converted friction coefficient which is in general reffered as train brake calculation, and also doesn’t need to consider the train basic resistance and gradient resistance, so the effective braking distance can be directly calculated by a given deceleration. [14] When the train is getting off a slope, ti is calculated from the formula (4): S=Vmax.t (4) where, S: Horizontal distance of the gradient section (m), V: Maximum speed (km/h). But when the train is getting off a slope and there is a station at the end of the gradient section, the running time is calculated according to the formula (5) of uniformly variable motion: = . + . . (5) 2 where, S: Braking distance (m) V: Speed (km/h) a: Deceleration (m/s ) Conventions: 1- It is assumed that the distance between two following stations is (L = 30 km) and this distance will divide for each gradient section as shown later in each profile. 2- The additional time for starting and parking of the train is considered to be 3 min: ts + tp = 3min 3- This study chooses the CRH - EMU type-3 and Figure (2) shows its unit force curve. From main technical parameters of CRH3 - EMU, it was found that the values of Pk and Vmax which belongs to CRH3 are: Pk = 21.05 kW/t, Vmax = 350 km/h.

56

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Figure 02. Unit force curve of CRH3 – EMU Now, by changing the speed of CRH3 (EMU), this paper will get different values of gradient as shown in Table (1). Table 1: Values of gradient for each speed of CH3 – EMU V (km/h)

350

300

250

200

w0 (N/kN)

17.69

13.28

9.52

6.43

i (‰)

3.96

11.98

20.79

31.46

RESULTS Figures (3), (5), (7), (9), (11) and (13) show the vertical profile between two stations A and B with the distance of each changing in gradient for profiles No.1, No.2, No.3, No.4, No.5 and No.6 respectively. Figures (4), (6), (8), (10), (12) and (14) show the relationship between running time and maximum gradient for profiles No.1, No.2, No.3, No.4, No.5 and No.6 respectively. Also, Table (2) shows time consumption model for each profile. Table 2: Time consumption model for the sixth profiles

DISCUSSIONS A set of vertical profiles, as shown in this study, were assumed in the past years. This series of patterns were drafted from a number of exiting railways for imitation of practical projected lines. Till now, no objectionable case has been discovered. However, it could not be regarded as the only way of imitation. For other countries and even for certain cases in China, different set of typical profile is not excluded. In the planning of a projected line, an overall inspection of the general route on the contour map is always necessary. If the designer pays attention to the towns, village and 57

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

industrial and mining points, a preliminary scheme of arrangement of the station and train stops may be sketched. Of course, the real railway profile is impossible to be the same as the six typical patterns which have been mentioned at the introduction of this paper. But the comparison of many times with the sections of existing railway lines, the imitation method shows high degree of similarity. Based on the theoretical analysis result, the following conclusions are drawn: 1- Time consumption model is a quadratic function for all profiles. The R-squared values for all models are more than 99%. 2-The results show that the time consumption is similar for the profiles which have symmetrical sketch with a horizontal axis, whereas unsymmetrical sketch of different profiles have different values of time consumption.

Figure 03. Vertical profile No.1

Figure 04. The curve t = f ( i ) of profile No.1

Figure 05. Vertical profile No.2

Figure 06. The curve t = f ( i ) of profile No.2

Figure 07. Vertical profile No.3

Figure 08. The curve t = f ( i ) of profile No.3

58

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Figure 09. Vertical profile No.4

Figure 10. The curve t = f ( i ) of profile No.4

Figure 11. Vertical profile No.5

Figure12. The curve t = f ( i ) of profile No.5

Figure 13. Vertical profile No.6

Figure 14. The curve t = f ( i ) of profile No.6

3- Profiles No.2 and No.3 have the same model of time consumption. Also, Profiles No.4 and No. 5 have the same model, Table (2). 4- Time consumption increases with the increasing of gradient value for each profile. 5- The percentage of increasing in time consumption between the first and the last value of gradient for each profile is 33.6% for profile No.1, 16.3% for profiles No.2 and No.3, 14.9% for profiles No.4 and No.5 and it is 20% for profile No.6.

59

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

References: [1] American Railway Engineering Association(1938), Proceedings vol.39, pp. 518-531. [2] Gong Danqing (2007), “Design for network of high-speed rail system station spacing”, Traffic Control Office of Beijing Railway Bureau, Beijing, China. [3] High-Level Group (1995), “Relevance of the trans-European high-speed train network for the European Union: study results”, in High Speed Europe, Office for Official Publications of the European Communities, pp. 5455. [4] International Union of Railways (2009), “General definitions of high-speed railways”. [5] M. Ashiya and M. Yasuda (1994), “Total system simulation of electrical railway of power consumption study”, in Computers in Railways IV-Vol. 1: Railway Design and Management. Berlin, Germany: Springer-Verlag, pp. 429-436. [6] P. Firpo and S. Savio (1994), “Optimal control strategies for energy management in metro rail transit systems”, in Computers in Railway IV, Berlin, Germany: Springer-Verlag, vol. 2, pp. 91-99. [7] R. A. Uher (1987), “Rail traction energy management model”, in Computers in Railway Operations. Berlin, Germany: Springer-Verlag, pp.39-60. [8] Wang Di (2000), “Synthetic optimization in engineering decisions with railway examples”, Southwest Jiaotong University, Chengdu, China. [9] Wang Di, Ma Wei (1988), “Synthetic optimal decision in railway projecting”, Rail International, International Railways of Central America. [10] Wang Di (1962), “Simplified calculation of railway operating expenses, energy and time consumption”, Journal of Tangshan Institute of Railway Technology. [11] William W. Hay (1982), “Railroad engineering – Second edition”, Chapters 4, 5, 9 and 12, Wiley, NewYork, USA. [12] Yao Lingkan (1994), “Study on the aided decision-making system for the railway alignments in debris flow regions”, Doctoral Thesis, Southwest Jiaotong University, Chengdu, China. [13] Yi SiRong (2009), “Railway location design – third edition”, In Chinese, Southwest Jiaotong University Press, Chengdu, China. [14] Zhang ZhongYang, Ma JinFa (2006), “Discussion of 200 km/h EMU train braking distance calculation method”, In Chinese, Journal of Zhengzhou Railway Vocational & Technical College, HuNan, China.

60

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

ANALYZING THE EFFECTS OF NORTHERN MARMARA HIGHWAY AND 3rd BOSPHORUS BRIDGE PROJECT AND 3rd AIRPORT PROJECT ON ISTANBUL S. BACARAN1, M. GÜRSOY2

[email protected], [email protected]

1

2

Res. Assist.., Y ld z Technical University, Faculty of Civil Engineering Assoc. Prof. Dr., Y ld z Technical University, Faculty of Civil Engineering

ABSTRACT The degree of development of transportation infrastructure, due to the high costs, is one of the indicators of socioeconomic development of a country. In addition to huge amount of economic resources requirement, transportation investments are strongly interrelated with the industry, population, transportation and natural environment. These relations might cause both a variety of negative results and some several positive outcomes as well. Investments are expected not to ruin the natural environment, contribute to the industry, improve the population balance, and ameliorate the transportation services and not to cause any problems if it is the first investment in the region. In this study, both positive and negative effects of Northern Marmara Highway and 3rd Bosphorus Bridge project and 3rd Airport project in Istanbul, which are amongst the most expensive transportation projects, are analyzed by utilizing the Delphi Technique. In this technique, 3 round 5 point-likert scale questionnaires are asked to be answered individually from academicians and experts in transportation field. At the end of each round of questionnaires, answers for each question are assessed statistically to determine whether a common answer is provided. For questions that no common answer was obtained, a new round of questionnaire was asked to be answered to reach a common decision. In the study, impacts of the above-mentioned investments are explored under five main titles; economic and financial, industrial, demographic, transportation and natural environment. Assessment of the results demonstrated that 3rd Airport of Istanbul is estimated to have mostly positive effects on industry, whereas, in terms of population it will have both positive and negative effects. Northern Marmara Highway and 3rd Bosphorus Bridge project, however, is predicted to increase the population and urban sprawl and cause both positive and negative outcomes about transportation. Key words: economy, environment, human, population, industry, transportation. INTRODUCTION Transportation can generally be defined as mobilization of people and goods to required place in required time. According to this definition, the influence of performance of the transportation system on the economic and social life in the city is enormous [1] [2]. Transportation is grouped under five sub-systems as transport of rail, road, sea, air and pipeline. Among these, rail and road are two systems that can be substituted for each other [3]. Given the investment costs of transportation projects, the issue of which of these subsystems will be given priority is of great importance in terms of economy. Transport sub-systems form a unity as infrastructure, vehicles, management and all its elements. Necessary transportation policy should be determined in order to achieve purpose of having a healthy transportation, compatible with social and economic development and supporting it, the cheap, fast, organized, safe and non-polluting; in short, with lowest cost to the country [4]. People are willing to use their time more efficiently by reducing the time they spend for transportation. From this point of view, the road transport would be insufficient. Airline constitutes a non-optional kind of transport for passengers of intercontinental or at distances exceeding 800 km. However, at distances around 600 to 800 km, experienced loss of time caused by transportation on the long way between airport and city center and also advancements in railway technology (speeding and even exceeding of 300 km/h high-speed railways) have proved superiority of railway against airlines. In addition to providing a fast transport, being much safer and having in general less damage to the environment than others, high-speed railways show that it will play an important role in the future transportation system [4]. Transportation is an intermediate service needed as a derivative of society's economic, social and cultural activities. That is, demand for transport is not spontaneous; instead, it occurs as a result of socio-economic organization and activities. It 61

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

should be produced as much as needed where required when required. If it is produced less, related socio-economic activity fails, if it is produced more, it would be wasted. Instead of seeking solutions for problems in transportation after it showed up, taking care of rational planning and implementation of physical improvements at the city, region and country level, will be more rational and consistent approach [4]. Transport in terms of planning is considered, on the one hand, as ensuring necessary compliance between different modes of transport, on the other hand, as ensuring of an effective integrity between development of economy and transportation services. Transportation investments are those that required to be assessed not only by the criterion of profit, but also by criteria such as economic development, social, political, security, and mass transportation [3]. New roads / new bridges / tunnels / multi-level junctions encourage the use of automobile, pull the traffic from public transport, create quickly their demands and after relief of very short-term, create a situation that more heavy and problematic than the former. Moreover, as distance for transport increases, urban structure tends to a disorganized pattern away from being compliant for public transport [5]. The traffic condition of Istanbul, application area of this study, not only in Turkey but also in the world, is one of the worst one. Due to unplanned growth of the city and its population growing faster than expected, the transport conditions are worsening in Istanbul. Because of the uncoordinated growth, staying in the daily commuting traffic becomes longer, public transport infrastructure and road infrastructure become inadequate. In addition to that, being center of attraction for various organizations on international scale, and drawing significant amounts of tourist by its historical attractions, its airports have already reached occupancy limit. And waiting for takeoffs and landings at airports have begun to emerge. Istanbul Metropolitan Municipality (IBB), Ministry of Environment and Urbanization (CSB) and the General Directorate of Highways (KGM) are working to resolve these difficulties in Istanbul with the new projects. North Marmara Highway and 3rd Bridge Project, and 3rd Airport Project of Istanbul, are two of the most remarkable projects recently. However, as mentioned above transport projects should be evaluated from many perspectives. The necessity of such large projects, optimal cost, location or route to be built, any potential harm to the environment and other effects that may occur should be investigated carefully. These two projects have caused serious concerns among the public. However, public institutions responsible for the project failed to address the concerns of the public. In this study, the reality of these concerns has been investigated in a sense by consulting opinion of academics and engineers working in the field of transportation. The positive and negative effects of projects / investments, subjected in this study that would be on the city were investigated by the Delphi technique. In this technique, specialists and academics working in the field of transport were conducted a third round questionnaire, unaware of each other. At the end of each survey round, the responses evaluated by statistical analysis and each item was confirmed whether a consensus was reached. The questions, unable to reach a joint decision at the end of each survey round were asked again trying to reach a compromise. North Marmara Highway and Istanbul 3rd Bosphorus Bridge Following benefits of this project are listed by KGM [6]: • Fuel saving will be enabled by decreasing the traffic density within the city and the current bosphorus bridges. • Cars will be able to make transit pass uninterruptedly, safely and comfortably, • Population density of the city will be reduced with the new residential areas to be built in the north Istanbul, • Emerging new commercial zones in north Marmara and in north Istanbul, the whole region will get more dynamic along with the neighboring provinces, • It will support the work of urban transformation to be held against earthquakes, • Connecting Asia and Europe, Turkey’s transportation alternatives and commercial capacity will increase with this bridge having both highway and railway network. • Time cost in our import and export will be reduced following the removal of the transportation restriction for loaded vehicles. • Uninterrupted interurban and urban railway transportation from Edirne to zmit will be conducted via the railway passing from the bridge, and this railway system will be integrated with Marmaray and stanbul Metro, and the Atatürk Airport, Sabiha Gökçen Airport and the 3rd Airport will be connected to each other. Istanbul 3rd Airport In order to meet air traffic to occur in the long term in Istanbul and its surroundings with potential of being financial and commercial center of Southeastern Europe and the Middle East as of its growing population and location, as it was evaluated in the final report of the 10th Transportation Council, construction of 3rd airport in conventional sizes is planned in Istanbul Region. This project is expected to increase the trade volume of Istanbul, and cargo transport as well. The airport will have 6 main runways, 4 aprons and each runway will have 19 landing or takeoff capacity per hour when the project is completed. Length of runway will be between 3500 m and 4100 m and apron capacity will be 500 aircraft. It is planned to have capacity of 150 million passengers per year [7]. 62

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

But the public have major concerns about this project, especially because of the risks to the environment. It is said that forest land and water resources in location of the project will be destroyed and wildlife will suffer. Additionally, it is considered that the damage to nature will increase with new buildings to emerge along the project`s route. MATERIALS and METHODS Delphi Technique Delphi technique has been developed by two researchers named Olaf Helmer and Norman Dalkey in the 1950s for the purposes of interpretation especially on military matters [8]. Researchers often face situations in which people that have a say in the solution of a problem approach to problem situations with different perspectives and conflict with each other’s opinion from time to time. These differences of opinion may arise on issues such as what would be appropriate target to be found in a training program, which qualifications should the product have, whether an expected event / project worth doing, what might be priorities, what kinds of qualifications should be held by person who supposed to fulfill of a specific task, etc. Delphi technique is used as a tool to find a compromise (consensus) in an environment where differences of opinion exist regarding similar conditions. Delphi technique especially should be used in situations of decision-making in political or emotional environment or in situations when there is likelihood of decisions being affected by powerful groups / individuals (quite effective). Expressed as a tool for reconciliation, Delphi is a technique by which expert opinion is obtained on a problematic situation in a systematic way. Using the Delphi technique, the consensus of individuals and groups with different point of views to a problem situation is aimed without coming face to face [9] [10]. In general, the Delphi technique has three main features. These are anonymity of the participants, statistical analysis of group responses and controlled feedback [10]. Anonymity of the participants; seen as the key to success of the Delphi. During the study, to whom the proposed ideas belong is kept confidential. Coming forward of ideas rather than individuals is provided in this way. Unconditional approval to the opinions of well known, respected person within the group is blocked in this way. Obtaining different and new ideas from everyone without any reservations is assured with the feature of anonymity in participations [9]. Statistical analysis of group response; analyzed statistically after each round of Delphi survey. The meaning of the statistics used in this analysis should be well known by the participants [9]. In this context, the person who carried out the study should inform participants clearly about meaning the results of the statistical analysis applied [10]. Controlled feedback; Successive surveys are used in the Delphi technique. After completing the statistical analysis of survey, the result that is general trend of respondent is forwarded to participants within the next round. In this way, individuals reconsider their thoughts comparing them with the results and different views and approaches communicated to them [10]. Implementation of Delphi Technique First, the problem, subject of study should be identified, and be described with a few words so that all participants can understand [9]. There should be a participant group composed of at least seven experts having knowledge and perspectives on the topic being studied as a result of their experience and qualifications. Ideal group size consists of 10-20 experts. Expert groups may be made up of university faculty of relevant branches and employers in related fields [9]. In the first round of Delphi survey, there should be open-ended questions about problem put forward as a result of research on the subject. Participants are asked to write their thoughts on open-ended questions of the first round in itemized form [9]. In the second Delphi survey, the responses obtained from the first round of the survey are written down as guidelines, itemized and not open-ended and by adding opinions achieved as a result of the literature survey by the group who carry out the study. All items prepared in the second round survey are forwarded to participants and they are asked to indicate at what level they agree with each item on 5 or 7 point Likert scale. After collecting the answer to the second round of the survey, 1st quarter, median, 3rd quarter and width values are calculated for each item [9]. Third Delphi survey is the same as the second round of survey. Within this tour, each item is accompanied by 1st quarter, median, 3rd quarter and width values calculated as a result of the second round and by relevant participant's response to that item. In the third round of the survey, participants are asked to review responses of the second round taking statistical value into account [9]. The values, calculated in the second round are also calculated for answers at the end of the third round and evaluated by comparing each other. If there is a decrease for any item in the third round about width values, calculated in the second round, it can be mentioned about a move towards reconciliation for that item. According to Zeliff and Heldenbrand [11] the items having interquartile width less than 1.2 are considered as items compromised on. The number of repetitions may be two or three to provide mobility towards reconciliation or even ten. But in general, the results are indicated to be satisfactory at the end of the fourth round [12] [9]. While preparing Delphi survey in this study, what kind of possible pros and cons would be in projects were considered by investigating both projects carefully. Later, reservations of public regarding projects were investigated and consequently it was decided that the positive and negative aspects of projects to be evaluated in five main topics, 63

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

economic and financial, industry, population, transport and the environment. At a later stage a participants list in the transport sector, composed of mainly academics (25 persons) was prepared. Then, the project files which identifying shortly and concisely North Marmara Highway and the Istanbul 3rd Bosphorus Bridge Integrated Project, and Istanbul 3rd Airport project and open-ended questions for both projects was prepared in five main topics. As an example in open-ended questions: "Write down in itemized manner what kind of positive and negative effects of Istanbul 3rd Airport project would be on Istanbul in terms of industry. And the survey containing presentation file and open-ended questions was sent to each expert in participant group via e-mail. The seven experts responded to the first round of survey. Responses of experts were evaluated and organized in items in order to comply with the format of the second survey. In addition to these, items generated by literature search were added. The items prepared in the second questionnaire were aligned to the 5 point Likert scale. And participants were asked to score each item from 1 to 5 (1 = strongly disagree, 2 = disagree, 3 = undecided, 4 = agree, 5 = strongly agree). 13 participants responded to the second round of the survey. The responses were compiled and 1st quarter, median, 3rd quarter and width values were calculated for each item. Having smaller width value means achieving reconciliation. In this study, it is accepted that items having width value below 1.2 were agreed upon [8]. Then, the third, final round of the survey were conducted and sent to participants, respondents of the second round. The third round was similar to that in the second round of the survey. The 1st quarter, median, 3rd quarter and width values, calculated at the end of the second round, were added into third round of the survey, and participants were asked whether they changed their answers given in the second round by considering of those values. Seven participants responded to the third round survey. The responses were compiled again and 1st quarter, median, 3rd quarter and width values were calculated for each item. Agreed and disagreed items were identified. RESULTS At the end of the third, final round of the Delphi Survey, 18 of 19 items, asked for Istanbul 3rd Airport project were found to be compromised. The results emerged as 15 "Agree (4)", and as 3 " Undecided (3)" of those 18 items. The possible impact of the project on Istanbul: Economically and financially, it was examined by 5 items and 4 of those were agreed upon, Industrially, it was examined by 5 items and all of those were agreed upon, In terms of population, it was examined by 3 items and all of those were agreed upon, In terms of transport, it was examined by 6 items and all of those were agreed upon, Environmentally, it was examined by 1 item and found to be agreed upon. More detailed examination of possible impact of Istanbul 3rd Airport project on Istanbul in terms of economic and financial reveals that participants were compromised at the level of `I agree` on such matters that recognition of Istanbul would increase, that larger investments would be drew at national and international scales, that tourism and international trade would develop due to facilitation of transport to Istanbul and that consumption of petroleum and its products would increase. They were compromised at the level of ‘undecided’ on that freight transport would improve in Istanbul. They have failed to reach a consensus on that unemployment would reduce increasing employment. More detailed examination of possible impact of Istanbul 3rd Airport project on Istanbul in terms of industry reveals that participants were compromised at the level of `I agree` on such matters that international cooperation in the field of science and technology would develop, that airline's share in import and export would increase, that improvements would be in aviation technology and aircraft industry and finally, that significant level of growth would be observed in some sectors (construction, automotive, transportation, logistics, etc.). More detailed examination of possible impact of Istanbul 3rd Airport project on Istanbul in terms of population reveals that participants were compromised at the level of `I agree` on such matters that Istanbul would become a center of attraction, thus both the labor force and the population would increase, unplanned and unregulated urbanization would gain prevalence, and a population distribution toward north would occur due to workforce needs in the north. More detailed examination of possible impact of Istanbul 3 rd Airport project on Istanbul in terms of transportation reveals that participants were compromised at the level of `I agree` on such matters that, national and international accessibility would increase, it would be an important support to be a logistics base, it would increase Istanbul's potential of being hub and air transportation would become easier, thus, the attraction of Istanbul would increase. Additionally, they were compromised at the level of `undecided` on matters of that the amount of travel and lengths of travel between the two sides of the city would increase. More detailed examination of possible impact of Istanbul 3 rd Airport project on Istanbul in terms of environment reveals that participants were compromised at the level of ‘strongly agree’ on such matters that, watersheds, forests, habitats of birds and of other wildlife would be lost in this area due to growth of the city towards the north with the project, and the environment would definitely be affected adversely. At the end of the third, final round of the Delphi Survey, 3 of 31 items, asked for integrated projects of the North Marmara Highway and the Istanbul 3rd Bosphorus Bridge were found to be disagreed. The results emerged as 5 "Strongly Agree (5)”, as 13 "Agree (4)", as 7 "Undecided (3)”, and as 3 "Disagreed (3)" of those 31 items. The possible impact of the project on Istanbul: 64

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Economically and financially, it was examined by 11 items and 8 of those were agreed upon, Industrially, it was examined by 5 items and all of those were agreed upon, In terms of population, it was examined by 5 items and all of those were agreed upon, In terms of transport, it was examined by 8 items and all of those were agreed upon, Environmentally, it was examined by 2 items and found that all of those were agreed upon. More detailed examination of possible impact of integrated projects of the North Marmara Highway and the Istanbul 3rd Bosphorus Bridge on Istanbul in terms of economic and financial reveals that participants were compromised at the level of `strongly agree` on such matters that rents would occur due to appreciation of land around highway and the external costs would emerge due to negative impacts on environmental. They were compromised at the level of `I agree` on such matters that it would provide a benefit on national scale in terms of transport costs due to railway transport costs would be far cheaper than by road, that having railroad crossing over 3rd bridge would lead to an increase of investment in the logistics industry, that Istanbul's internal and external trade volume would increase, and that consumption of petroleum and its products would increase. They were compromised at the level of ‘undecided’ on such matters that plans of moving heavy industry in Istanbul out of the city and of turning it into a financial center would become inactive and it would increase Istanbul`s attraction for heavy industry, and that freight transport would increase with the decrease in shipping price due to the railway to cross over 3rd Bosphorus bridge. They disagreed on such matters that Istanbul would attract major investments on national and international scale, that it would make a positive contribution to the economy by reduction in share of the road in passenger and freight transportation in Turkey due to railroad crossing over 3rd Bosphorus bridge, and that economic gains would be achieved due to the use of domestic products and manpower in the process of the project's construction. More detailed examination of possible impact of integrated projects of the North Marmara Highway and the Istanbul 3 rd Bosphorus Bridge on Istanbul in terms of industry reveals that participants were compromised at the level of ‘I agree’ on such matters that the new sectors would emerge and significant growth would be observed in some sectors (construction, automotive, transportation, logistics, etc.). They were compromised at the level of ‘undecided’ on such matters that plans of moving large industrial branches out of the city would not be performed because of the growth to take place in some sectors, that it would cause a certain level reduction in goods prices through the use of rail transport in exports and imports, and that it would increase the competitiveness. They were compromised at the level of ‘disagreed’ on matter of that it would accelerate production of completely domestic (national) car. More detailed examination of possible impact of integrated projects of the North Marmara Highway and the Istanbul 3rd Bosphorus Bridge on Istanbul in terms of population reveals that participants were compromised at the level of ‘strongly agree’ on such matters that it would cause the city to grow northward, and that it would accelerate population growth in the both sides. They were compromised at the level of ‘I agree ’ on such matters that unplanned and unregulated urbanization would gain prevalence and that it would lead to a production increase in Istanbul, thus, the population would increase within regions close to production. They were compromised at the level of ‘undecided’ on matter of that population distribution of Istanbul would change significantly. More detailed examination of possible impact of integrated projects of the North Marmara Highway and the Istanbul 3 rd Bosphorus Bridge on Istanbul in terms of transportation reveals that participants were compromised at the level of ‘I agree’ on such matters that travel length would increase in Istanbul, that traffic would get lighter a little at the FSM bridge and in the city because of taking truck traffic out of the city, that the transition between the two sides of the city would increase, and that vehicle ownership would increase in Istanbul. They were compromised at the level of ‘undecided’ on such matters that traffic congestion problems would be experienced in medium and long-term, that imbalance in transportation system in favor of road transport would increase in context of Black Sea and European connections, and that number of fatal accidents in Istanbul would decrease. Additionally, they were compromised at the level of `disagreed` on matter of that the lengths of travel in Istanbul would decrease. More detailed examination of possible impact of integrated projects of the North Marmara Highway and the Istanbul 3rd Bosphorus Bridge on Istanbul in terms of environment reveals that participants were compromised at the level of ‘strongly agree’ on such matters that forests, watersheds, wildlife habitats in this area would be severely damaged, that this damage would take more than all benefits that the project would bring. Also, they were compromised at the level of `I agree` on matter of that the distance, especially for commercial vehicles would increase and consequently, fuel consumption and carbon dioxide emissions released into the environment would increase. DISCUSSIONS When study results are evaluated, the authors consider that; Istanbul 3 rd Airport Project will be a hub airport between Asia and Europe by having many more runways, by having a large cargo section and by its location. Thus, it is considered that the accessibility of Istanbul will increase and it will attract major investors into Istanbul. It is thought that investments to be held in Istanbul will lead to increase of the population and indirectly though, consumption of oil and oil products in our country which highly dependent on foreign energy is expected to increase further. On the environment which is one of the biggest concerns relevant to the project, the authors believe that migratory birds, watersheds and wildlife will suffer. 65

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

When study results are evaluated, the benefit of North Marmara Highway and Istanbul 3 rd Bosphorus Bridge Integrated Project, considered only based on the bridge, is great because of its connecting Asia to Europe in terms of international passenger and freight transport thanks to the railway to pass over. As a result, railway with its advantage of economic freight transport can reduce highway`s share a little. But due to the vast majority of highway route included in projects is unbuilt forest land, it is thought that rent will occur in the land on this route and that deforestation will happen. Thus, Istanbul will grow towards the north with the project, new employment areas will emerge and Istanbul's population will increase rapidly. This project is expected to ease Istanbul`s traffic for a while by taking heavy vehicle traffic out of the FSM bridge. It can be said that this project will bring more damage to wildlife, watersheds and forests than Istanbul 3rd Airport project. Besides, because of the increase in length of travel, the fuel consumption and fuel emissions released into the atmosphere are expected to increase. The implementation of this survey to the managers in freight and passenger transport sector with wider participation will provide a broader and objective perspective. However, due to time constraints and difficulty of access to participants willing to fill out a questionnaire, this suggestion is left to further studies. REFERENCES 1) Tanga, K.X., Watersb, N.M. (2005). The internet, GIS and public participation in transportation planning. Progress in Planning, Vol. 64, pp.7-62. 2) Karacasu, M. (2007). Kent içi toplu tasima yatinmlann n de erlendirilmesinde karar destek modeli (ELECTRE Yöntemi) kullamm . 7. Ula rma Kongresi, stanbul. 3) Alt nok, S. Türkiye’de Ula rma Politikalan, Karayollan Ve Demiryollann n Mukayesesi. SÜ IIBF Sosyal ve Ekonomik Ara rmalar Dergisi, pp.73-87 4) Evren, G. Türkiye Ula rma Politikalanna Ele tirel Bir Bak . 10. Ulasim ve Trafik Kongresi-Sergisi, stanbul Teknik Üniversitesi 5) Evren, G. Ula rmada Plans zhgin ve Yanh ta Isrann Simgesi: 3. Köprii. Okan Üniversitesi 6) Karayollan Genel Müdürlugu, Kuzey Marmara Otoyolu ile stanbul 3.Bo az Köpriisii. http://www.google.com.tr/url?sa=t&rct=i&q=&esrc=s&source=web&cd=l&ved =0CCgQFiAA&url=http%3A%2F%2Fwww.kgm.gov.tr%2FSiteCollectionDocu ments%2FKGMdocuments%2FDuvurular%2FKGM%2520Katalog.pdf&ei=yg YfU5a6B8ee4wS19oD4Dw&usg=AFOiCNEsRHtC7EUVRWrWZSSFLsseeGiow&bvm=bv.62788935,d.bGE&cad=ria 7) T.C. Ula rma Denizcilik Ve Haberle me Bakanhgi Altyap Yatinmlan Genel Müdürlugu, stanbul Bölgesi 3. Havalimam. http://www.google.com.tr/url?sa=t&rct=i&q=&esrc=s&source=web&cd=l&ved =0CCYOFiAA&url=http%3A%2F%2Fwww.csb.gov.tr%2Fdb%2Fced%2Fedito rdosva%2Fnihai ced istanbul.pdf&ei=uwcfU7CVHqHk4wSHqIDIBw&usg=A FQiCNHGGWiilpfVSmq90tBrALADle7vwA&bvm=bv.62788935,d.bGE 8) Dalkey, N., & Helmer, O. (1962). An Experimental Application of the Delphi Method to the Use of Experts (No. RAND/RM-727-PR). RAND CORP SANTA MONICA CA. 9) Sahin, A. E. (2001). Türkiye’de lkogretim Okulu Müdürlugiinün Bir Meslek Olarak Mevcut Durumu: Bir Delphi Çah mas . Hacettepe Üniversitesi E itim Fakültesi Dergisi, No: (20); 215-220. 10) Arslan, E. S. (2010). Kültürel Peyzaj Kavram Kapsam nda Bir Degerlendirme: Ulus-Tbmm Tarihi Aks (Ankara), Ankara Üniversitesi Fen Bilimleri Enstitüsü. 11) Zeliff, N. D. ve Heldenbrand, S. S. (1993). What Has Being Done In The International Business Curriculum?. Business Education Forum, 48 (I),23-2S. 12) Erlfmeyer, R., Erffmeyer, E. ve Lane, i. (1986). The Delphi Teehnique: An Empirieal Evaluation of the Optimal Number of Raunds. Group & Organization Management, n (1-2), 120-129.

66

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

MODELLING OF THE EFFECTS OF HYDRATED LIME ADDITIVES ON HOT MIX PAVEMENTS USING A FUZZY LOGIC APPROACH M. S. Yard m1, B. De er2, G. Ayd n2 [email protected], [email protected], [email protected] 1

ld z Technical University Civil Engineering Faculty, Esenler, Istanbul, Turkiye Tel: (212) 383 51 83 2Y ld z Technical University Civil Engineering Faculty, Esenler, Istanbul, Turkiye Tel: (212) 383 51 85

ABSTRACT In this study, it is aimed to determine the effect of different rates of hydrated lime on the Marshall Stability and other mixture parameters for different asphalt contents, in which, hydrated lime is used as a modifier instead of filler material. To determine the effect of hydrated lime, 15 different design experiment sets are prepared and at each set filler material is decremented by rates of 0.5% from 6.8% to 0% and hydrated lime is incremented by the same rates as the filler. A fuzzy logic model is developed to provide an estimate of the experimental results. Experimental data from the previous study are used for both training and testing of the fuzzy logic model. Since this model provides an acceptable estimate of the experimental results, it seems possible to use it to determine the amount of hydrated lime to be added to the mixture, to produce a mixture whose parameters are within specification ranges. Keywords: Hot mix asphalt, hydrated lime sensitivity, asphalt contents of mixture, Marshall Stability, fuzzy logic. INTRODUCTION Hot mix asphalts (HMA) get damages due to the tensions created under the effect of traffic loads on the highway pavement and also temperature changes. These damages mean shortening of the pavements’ service life and increasing maintanence costs. Thus, it is of crucial importance to produce better quality mixtures, which have more resistance to these tensions, in other words, have a higher stability. In order flexible pavements to function well; major defects such as permanent deformations, fatigue cracking, low temperature cracking, stripping due to moisture and low durability should be minimized [1]. For this purpose, one of the methods on which people are studying is modification of the conventional mixtures with additives [2], [3], [ 4]. In this study, we used hydrated lime (HL) as such an additive. HL reacts with aggregate and strengthens its bond with the bitumen. It also reacts with highly polar molecules to inhibit the formation of water-soluble soaps that promote stripping [5], [6]. HL addition to HMA has various benefits, depending on the properties of aggregate, binder and other additives (if any). In some sense, HL can be defined as a multifunctional additive [2], [3], [7], [8]. In this study, we conducted Marshall Stability test [9] on the samples. We prepared the samples by decreasing the filler in the mixture at a predetermined rate and adding HL at the same rate. With the test, we determined, for each different HL content, the optimal bitumen content and the other parameters of the mixture properties. Afterwards, for each optimal bitumen content and respective HL content, we produced 3 different specimens, adding up to 45 specimens. Finally, we determined the mixture properties once more and compared them to the previous ones. After the experimental studies, we developed a fuzzy logic model to provide an estimate of the experimental results. We used experimental data from the previous study for both training and testing of the fuzzy logic model. MATERIALS AND METHODS We applied sieve analysis, specific gravity and water absorption tests to our aggregates. At this stage, we did not repeat stripping resistance, percent wearing loss, weather impacts durability (frost loss) and flatness index tests; since examinations of the oven reports showed that they were in appropriate intervals during the material supply. To determine the effect of hydrated lime, we prepared experiment sets of 15 different Marshall Designs. At each set, we decremented filler material by sensitivity rates of 0.5% from 6.8% to 0% and incremented HL by the same rates as the filler. For each HL increment rate, we conducted a Marshall Stability Test, consisting of 18 samples to get the optimized asphalt content. Materials The properties of aggregate, bitumen and HL in our hot-mixture are indicated below. 67

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Aggregate In this study, aggregates were from sfalt A. . Ümraniye Asphalt Factory (Plant). The source of these materials is the limestone quarries in Ömerli, stanbul. The main element oxide of these limestones consists of 52.9% CaO. In their mineralogical composition, around 75% calcite is found. The production level calcite is generally gray-dark coloured [10]. We determined the particle size distribution of the aggregate with sieve analysis; according to KGM specifications for Typeof-Wear-1 [11]. Size distribution was found as 15% type 2 aggregate, 40% type 1 aggregate and 45% stone powder (Figure 1). Finally, the rates of coarse, fine and filler materials in the mixture were found as 51.3%, 41.9% and 6.8%. Physical properties of the aggregates are given in Table 1.

Figure 01. Aggregate distribution on gradation chart

Table 01. Physical properties of aggregates Aggregate Type Coarse Aggregate Fine Aggregate Filler Aggregate Mixture

Apparent Specific Gravity (g/cm3)

Volume Specific Gravity (g/cm3)

Water Absorption (%)

2.729

2.698

0.42

2.754 2.735

2.693

0.82

2.740

2.699

Bitumen We used 50-70 penetration bitumen from Izmit Oil Refinery (Tüpra ). The test results of the asphalt cement and respective specification values are in Table 2. We determined the optimum bitumen contents for every different HL rate using the Marshall Method.

68

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Table 02. Characteristics of the bitumen which is used in Properties Specific gravity (g/cm3), at 25 oC o

Flash point (Cleveland) ( C) Penetration (0.1 mm), at 25 oC, 100 g, 5 sec Ductility (cm), at 25 oC, 5 cm/min Thin film heating loss (%), at163 oC, 5 hr Penetration percentage after heating loss (%) Ductility after heating loss (cm) Softening point (oC)

Standards

Test Results

ASTM D 70

1.009

ASTM D 92 ASTM D 5 ASTM D 113 ASTM D 1754 ASTM D 5 ASTM D 113 ASTM D 36

345 59.8 > 100 11 65.9 65.4 48.5

Specification

> 230 50-70 > 100 < 80 > 54 > 50 45-55

Hydrated Lime (HL) The HL used in this study is produced in Bart n Lime Factory in northern region of Turkey, with production code S-KK 80-T (Table 3). Table 03. Properties of hydrated lime Test Standard Limit Properties Standards Results Value Chemical Properties Total CaO (%) TS EN 459-1; TS 32 EN 459-2 85.78 > 80 MgO (%) Total CaO+MgO (%)

TS EN 459-1; TS 32 EN 459-2 TS EN 459-1

3.52 89.3

<5 >80

Loss on Ignition (%)

TS 32 EN 459-2

22.51

SO3 (%)

TS EN 459-1; TS 32 EN 459-2

1.47

< 2.0

CO2 (%) Physical Properties Fineness over 90 microns (%) Density(kg/m3)

TS EN 459-1; TS 32 EN 459-2

3.89

< 7.0

TS EN 459-1; TS 32 EN 459-2

6.0

< 9.0

TS 32 EN 459-2

472

< 600

Marshall Stability Tests To observe the effect of each HL content, we decreased filler content from 6.8% with 0.5% sensitivity. We replaced the missing filler by HL at the same amount. For every HL content, we made a design according to Marshall Method [9] and found the optimum AC%. Hence, we produced a total of 270 Marshall specimens, 18 for every 15 different HL content. Afterwards, we had these specimens broken in the Marshall Test Device and found their Marshall stability (MS) and Flow (F) values (Table 4). Using the weight and dimension values of the specimens, we calculated their “bulk specific gravity (Dp)”, “air voids (Va)”, “voids in the mineral aggregates (VMA) ” and “void filled with asphalt (VFA)” values (Table 4). Then, for each optimal bitumen content and respective HL content, we produced 3 different specimens, adding up to 45 specimens. Finally, we determined the mixture properties once more and compared them to the previous ones. The Fuzzy Logic Model In this part of a study, we develop a fuzzy logic model to relate the input variables (AC% and HL%) to the output variables (Dp, VMA, MS, Va, F and VFA). This is to propose a fast and practical method to select design values for the input variables such that the values of the output variables stay within the specification limits. Due to very limited space available, we do not enter into theory of fuzzy logic. Readers who are not familiar with fuzzy logic may refer to [12] to gain some preliminary information.

69

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Table 04. Marshall design values HL Marshall Design Content No. (%) 0 1

Apparent Specific Gravity (Filler+HL) (g/cm3) 2.806

Optimum AC% Content

Dp

Marshall Stability (MS)

Flow (F)

VFA

Va

VMA

(%) 5.20

(g/cm3) 2.409

kN 11.60

mm 4.40

(%) 73.00

(%) 4.10

(%) 15.27

2

0.5

2.766

5.28

2.398

11.64

4.85

71.60

4.45

15.70

3

1

2.76

5.25

2.400

11.60

4.75

72.00

4.40

15.58

4

1.5

2.702

5.47

2.371

8.58

4.13

68.90

5.18

16.63

5

2

2.519

5.47

2.391

9.78

4.35

74.50

3.82

15.48

6

2.5

2.588

5.59

2.373

8.41

4.40

71.60

4.60

16.29

7

3

2.609

5.38

2.377

9.25

6.05

70.05

4.80

16.20

8

3.5

2.586

5.60

2.374

8.75

4.27

72.00

4.50

16.26

9

4

2.571

5.35

2.371

10.25

4.55

70.03

4.84

16.25

10

4.5

2.505

5.54

2.355

8.70

4.10

69.00

5.18

16.65

11

5

2.448

6.16

2.342

8.41

5.10

73.00

4.80

17.32

12

5.5

2.544

6.33

2.358

8.05

5.40

76.50

4.20

17.44

13

6

2.501

6.23

2.366

8.30

6.52

77.00

3.80

16.78

14

6.5

2.488

5.75

2.324

8.12

4.11

66.00

6.10

17.94

15

6.8

2.419

6.10

2.35

8.70

4.60

74.00

4.42

17.04

We started the modeling process by dividing the input ranges into different categories, whose respective parameters are given in Tables 5 and 6. Meanings of Pi (i = 1, 2, 3, 4) are illustrated in Figure 2, for trapezoidal and triangular type membership functions (MFs). It is drawn only for illustrative purposes and does not reflect any modeling work we did.

Category Name Almost zero Very very little Very little Little Medium Much Very much Very very much

Table 05. Categories for HL% Membership P1 P2 Function Type Trapezoidal -2.52 -1 Triangular 0.3 1 Triangular 1.2 2 Triangular 2.3 3 Triangular 3.3 4 Triangular 4.3 5 Triangular 5.3 6 Trapezoidal 6.2 7

P3

P4

0 1.7 2.8 3.7 4.7 5.7 6.6 7.77

0.75 8.17

We chose the input MFs considering our experimental sets. If any two different sets fall into exactly the same categories, they fire exactly the same rules. In our previous trials, we had seen that, if two different inputs fire exactly the same rules, they generate very similar (if not exactly the same) outputs. This is not desirable, as the experimental results indicate, asphalt mixture’s parameters are quite sensitive to AC% and HL%; even small changes in those inputs cause significant changes in the outputs. Thus, categorization was done such that any two different inputs in our set did not fall into exactly the same categories.

70

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Category Name Little Medium Much Very much

Table 06. Categories for AC% Membership P1 P2 Function Type Trapezoidal -2.52 -0.28 Triangular 4.25 5 Triangular 5.25 6 Trapezoidal 6.1 6.7

Trapezium

P3

P4

3.75 5.75 6.6 7.05

4.75 9.86

Triangle

Figure 02. Illustration of the meanings of Pi Next stage in the modeling is to generate the rule sets. Since one input has 8 categories and the other has 4, there are 8 x 4 = 32 different combinations. We generated at least one rule for every different combination. Some combinations needed more than one rules. There were a total of 38 rules. Sample rules are given in Table 7. Table 07. Sample rules for our Sugeno type Fuzzy Logic Model If If If If If

AC % is little medium much very much little

HL% is Dp VMA MS Va F VFA AND almost zero THEN MF1(1) MF1(2) MF1(3) MF1(4) MF1(5) MF1(6) AND little THEN MF12(1) MF12(2) MF12(3) MF12(4) MF12(5) MF12(6) AND much THEN MF21(1) MF21(2) MF21(3) MF21(4) MF21(5) MF21(6) AND

very little THEN MF29(1) MF29(2) MF29(3) MF29(4) MF29(5) MF29(6)

AND

medium

THEN MF5(1) MF5(2) MF5(3) MF5(4) MF5(5) MF5(6)

The fuzzy logic model we proposed is of Sugeno type. Output MFs (MF1(1), etc.) are either linear functions of the inputs or constants. Since the model is a multiple – output one, it was not possible to benefit from the ANFIS approach for calibration [13]. Instead, we carried out a manual calibration procedure. For every rule, we collected and organized all relevant data that would fire this rule. For every different output variable, we employed a separate multiple linear regression approach using Microsoft Excel, so we obtained a linear relation between the inputs (AC% and HL%) and the output. Finally, the rules we generated started to appear in the form of “If AC% is little AND HL% is medium THEN Stability is a*(AC%) + b*(HL%) + c AND VFA is d*(AC%) + e*(HL%) + f AND ……”, where a, b, c, d, e and f are real (not necessarily nonzero or nonnegative) constants. For some of the rules, there were too few (either 2 or 1) data values that fire them to perform a linear regression. For these rules, we used constant output membership functions. We paid careful attention not only to model calibration but also to model validation, which can be defined as the model’s ability to perform well on data which it had not see during the calibration process. Only doing a accurate calibration is not sufficient for a model; it is also expected to perform well on data which it had not seen before. For this purpose, we did not use all of our experimental data to calibrate the model. We used all series with optimal bitumen amount for model validation, and the rest of our data (mostly used for determining that optimal bitumen amount) for model calibration. RESULTS In this part, we present the outcomes of our fuzzy logic model in a comparative way with the Marshall Test results, which was performed to see the effect of hydrated lime on the properties of the asphalt mixture. A total of 270 specimens were produced, with HL content ranging between 0% - 6.8%, to determine the optimal bitumen amount. We 71

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

used the test results on these specimens to calibrate our model. After determining the optimal bitumen amount, another 45 specimens of 15 different designs, but, this time, with the optimal bitumen amount for each design, were produced. We used the test results on these specimens to test and validate the model. A screenshot of our model (formed using the Fuzzy Logic Toolbox of Matlab) is given in Figure 3.

Figure 03. Screenshot of our model, formed using Fuzzy Logic Toolbox of Matlab Comparative results of model testing are given in Figure 4. The highest R2 value is obtained for MS. However, for all mixture parameters, our model produced an acceptable fit. In order to use the model to prepare a mixture, which complies to the specification, only R2 may not suffice. It is also necessary to know “how wrong the model can be.” To measure this, percentage errors are calculated for all test data values. Maximum percentage errors are given in Table 8. When determining a mixture parameters using this model, these values can be taken into account, to be on the safe side with respect to specification limits.

Table 08. Maximum percentage errors of output variables Output Variables V VFA MS Va F Max Error (%)

0.88

4.06

10.23

15.26

11.09

VMA 4.89

72

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Figure 04. Comparison of fuzzy logic model outcomes and experimental results DISCUSSION In this study, we present a fuzzy logic model to determine the effect of hydrated lime addition on the hot mix asphalt parameters. This model provides an acceptable estimate of the experimental results. Since the percent errors in the estimation of the model is within tolerable bounds, it seems possible to use it to determine the amount of HL to be added to the mixture, to produce a mixture whose parameters are within specification ranges. This study, which is for determination of the effect of HL on hot mix asphalt parameters, represents the introductory state of such work. Complete analysis of this effect requires many more mixture and material tests to be conducted. We believe that, this modeling study and its underlying experimental study, although mainly limited to Marshall Test, will shed light on and guide us through our future work. In the study, we could not use ANFIS for calibration of the model. This is because, ANFIS, in its standard from that comes with MATLAB, does not support multiple outputs. When researching the relevant literature, we have seen that there are some studies to extend the system to support multiple outputs. However, we could not implement it due to lack of time and experience. Our plans for future work include making use of such extensions to obtain better trained and validated models. ACKNOWLEDGMENTS We thank YTÜ Scientific Research Projects Coordination Unit for their support to the Project with ID 29-05-01-KAP01 and also sfalt A. . for their support on the experiments.

73

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

REFERENCES [1] Tunç, A. (2004). “Flexible Pavement Materials Handbook/Esnek Kaplama Malzemeleri El kitab (Turkish)”, Asil Yay n ve Da m Ltd. ti, Ankara, Turkey, 2004. [2] Little, D. N. and Epps, J. A., (2001), “The Benefits of Hydrated Lime In Hot Mix Asphalt”, Report for National Lime Assocation (updated by Sebaaly, P. E., 2006), http://www.lime.org/BENEFITSHYDRATEDLIME2006.pdf. [3] NLA (National Lime Association). (2006), “Hydrated Lime: A Solution For High Performance Hot Mix Asphalt”, Fact Sheet on Lime’s Multifunctional Benefits in Asphalt, http://www.lime.org/Aasphalt.pdf. [4] Kim, Y. R., Lutif, J. S, Bhasin, A. ve Little, D. N., (2008), “Evaluation of Moisture Damage Mechanism and Effects of Hydrated Lime in Asphalt Mixtures Through Measurements of Mixture Component Properties and Performance Testing”, Journal of Materials in Civil Engineering, 20: (10), pp. 659–667. [5] Petersen, J. C., Plancher, H. ve Harnsbergen, P. M., (1987), “Lime Treatment of Asphalt to Reduce Age Hardening and Improve Flow Properties”, The Association of Asphalt Paving Technologists, Vol. 56, pp. 632-653. [6] NLA (National Lime Association). (2003). Lime the versatile chemical hydrated lime: a solution for high performance hot mix asphalt, fact sheet. [7] Little, D. N. (1996), “Hydrated Lime as a Multi-Functional Modifier for Asphalt Mixtures”, Presented at the HMA in Europe Lhoist Symposium, Brussels, Belgium. [8] European Lime Association. (2011). “Hydrated Lime a Proven Additive For Durable Asphalt Pavements, Cr cal terature Rev ew”, Report to the European Lime Association, December. [9] ASTM D 1559-89, (1989). “Test Method for Resistance of Plastic Flow of Bituminous Mixtures Using Marshall Apparatus (withdrawn 1998)”. [10] Zarif, . H., Tu rul, A. ve Dursun, G. (2003). “ stanbul’daki Kireçta lar n Agrega Kalitesi Yönünden De erlendirilmesi”, stanbul Üniv. Müh. Fak. Yerbilimleri Dergisi, 16: (2), pp. 61-70, stanbul. [11] KGM (2006). “Turkish Highway Technical Specifications/ Karayolu Teknik artnamesi”, General Directorate of Highways, pub. No. 267, Ankara, Turkey, 2006. [12] Zadeh, L. A. (1965). “Fuzzy Sets” Information and Control, 8, 338-352. [13] http://www.mathworks.com/help/fuzzy/anfis.html. Date of access: 25.02.2014.

74

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

STATISTICAL ANALYSIS OF VEHICLE DELAY MEASUREMENTS CONSIDERING DIFFERENT TIME DURATIONS Y. S. Murata,*, Z. Cakicia a

Pamukkale University, Engineering Faculty, Dept. of Civil Engineering, 20070, Denizli, Turkey * Corresponding Author. E-mail: [email protected] ; Tel: 0 (258) 296 33 57

ABSTRACT Vehicle delay is one of the important performance parameters for signalized intersections and it is used both design and evaluation purposes. Because the vehicles are observed individually so the measurement of delay takes a long time. Besides, these measurements should also be made quite sensitive. Field observation studies constitute a large amount of this delay measurement process. If the studies are made without video-camera, even if any small problem regarding to vehicle record, field observation should be repeated. This situation causes both the waste of time and needs more labor. So, the vehicle delay records which are made by the video cameras are more precise and decreases waste of time and intense labor. The aim of this study is to get reliable analysis approach related to the measurement of the vehicle delay in less time. The average delay observation values are considered three periods of the time as fifteen minutes, thirty minutes and hourly. To investigate effect of observation time periods on the results of delay, statistical tests are applied. Vehicle delay observation values obtained from Pekdemir Intersection in Denizli, Turkey are used for investigations. Kolmogorov-Smirnov and Wilcoxon Signed Rank tests are used for the evaluation of these time durations. As a result of the statistical analyses, thirty-minute delay observation results are the most closed to hourly (peak hour) delay observation results, fifteen-minute delay observation results are not closed to hourly (peak hour) delay observation results. Consequently, use of thirty-minute delay observation instead of hourly delay observation is proved in this study. Keywords: intersection, vehicle delay, kolmogorov-smirnov test, signalization, wilcoxon signed rank test INTRODUCTION Delay is the waste of time that approaching vehicle to intersection because of the other vehicles, the geometric properties of intersection and the control systems like traffic signs and trafic lights. Vehicle delay is an important performance parameters at signalized intersection approaches. This parameter has been used for determining level of service and evaluating the operation of signalized intersections. In the field, measurement of delay should also be quite sensitive to approximate the real observation values. Therefore, these measurements takes a long time. For many decades, various calculation methods have been developed for exactly prediction of the vehicle delay by researchers. Webster [1], Akcelik [2] and Transportation Research Board (Highway Capacity Manual) [3] methods are the common methods used in calculation. Delay observation study takes long times and it is a difficult process. On the other hand, it should be made precisely as much as possible to obtain closer values. These measurements can be made by video camera and at least two observers are required for an intersection approach. Generally, the observation period is one hour that the traffic volumes are the highest. In the field studies, each vehicle has been considered separately and delay value has been obtained for each vehicle. Inaccurate measurements of delay can cause inadequate assignment of signal duration for each intersection approaches and improper design. Thus, these measurement should be made quite sensitive (Ban et al., [4]). These sensitive measurements can be made only using video-camera. If the video-cameras are not used and observations are made only at the field, even if any small problem regarding to vehicle record, the observations should be repeated. This situation causes both the waste of time and the more labor. Therefore, a new method which is more efficient than the others, takes less time and more closer to real observation is needed. Components of Delay and Delay Observation: The vehicle delay at signalized intersection consists of three part that is referred as the stopping, acceleration and deceleration delays (Dion et al., [5]). The stopping delay indicates time process that the vehicle spend because of stopping during the red or green signal period at signalized intersection. When the acceleration delay of the vehicles can be defined as required time to accelerate after the red signal turn the green signal, the deceleration delay of the vehicles can be defined as the time between that drivers start to reduce their velocity while approaching to intersection and that drivers stop in front of the signal while red signal is shown. Figure 1 illustrates trajectory diagram of a vehicle near a 75

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

signalized intersection approach. Deceleration delay (A), demonstrated in the Figure 1.

Stopping delay (B), Acceleration delay (C) are also

Figure 01. Components of Vehicle Delay For the estimation of delay at signalized intersections, many studies have been done by transportation researchers for many years. The 1985 Highway Capacity Manual delay formula was examined and a calibration process was improved by Akcelik, [6]. Akgungor, [7] focused on the new time-dependent delay calculation method considering various analysis time period. Murat, [8] studied on vehicle delay modeling using artificial neural networks and fuzzy logic optimization technique and compared the result with each other. Su et al., [9] investigated the effects on delay considering proportion and position of heavy vehicles at signalized intersection in China. Murat et al., [10] produced a new delay formula depending some variables (Average Entering Time to Intersection, Red Signal Time, Number of Vehicles in Queue, Average Discharging Headway), at the end of the study, the results obtained with the formula were compared with field observations and Akcelik delay formulation results. In this study, thirty-minute delay observation results at the peak hour were compared with hourly (peak hour) delay oservation results. In the next part (Materials and Methods), field studies and samples for delay measurements were presented. Besides, Delay analysis, sample cases of average delays and comparisons of delay observations at Pekdemir Intersection were introduced. Some information about the statistical analysis were also given in the Materials and Methods part. The statistical tests were demonstrated in the Results part, in the Discussion part, the results were evaluated considering Kolmogorov-Smirnov and Wilcoxon Signed Rank Tests and some recommendations about delay measurement were presented. MATERIALS AND METHODS Field Studies: In the field, delay observation studies are made as defined in the following: 1. Determining of Reference Point 1 and Reference Point 2: When Reference point 1 is selected as the middle point of intersection, Reference point 2 is selected as a point existing behind of hundred meters from traffic light. 2. Observers or Video Cameras: At least two observers or a video-camera is needed for an intersection approach when the delay observation is made. Observers must be situated between Reference Point 1 and Reference Point 2. Thus, both entry time (at Reference point 2) of the vehicle to intersection and exit time of (at Reference point 1) the vehicle from the intersection can be precisely recorded. If the video camera is used instead of observers, it is needed that video camera must be placed at a location covering both reference point 1 and reference point 2. The locations of observers or video-camera and reference points at an intersection is depicted in the Figure 2. Vehicle delay is computed in the following Equation; (Equation 1) texit : Exit time of the vehicle from the intersection - Reference Point 1 (sec.) tentry : Entry time of the vehicle to intersection - Reference Point 2 (sec.) Vehicle Delay = t t (1)

76

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians Video Camera Reference Point 1

Traffic Light

Reference Point 2

Observer

A

A

B

B

100 m

Figure 02. The Locations of Observers, Video-Camera and Reference Points at an Intersection Sample cases for delay calculation is given in Table 1. Table 01. Sample Cases for Delay Calculation tentry (sec.)

texit (sec.)

Vehicle Delay (sec.)

2 21 27 28 32 38 40 44 46

62 64 67 70 73 75 77 83 87

60 43 40 42 41 37 37 39 41

Analysis: In this study, vehicle delay observation values were obtained from the Pekdemir Intersection in Denizli, Turkey. Observatios were made at between 08:00 and 09:00, between 12:00 and 13:00 and also between 17:00 and 18:00 by video-cameras as an hourly. Besides, Observations consist of two days of week including a day in the weekday and in the weekend. Each vehicle was examined individually. The average vehicle delay was calculated for each lane. The average vehicle delay based on the lane was calculated in the following equations (Equation 2, Equation 3); t t Average Vehicle Delay for an hour = (2) k t t Average Vehicle Delay for thirty minutes = (3) z Where; texit : Exit time of the vehicle from the intersection (sec.) tentry : Entry time of the vehicle to intersection (sec.) k : The total number of vehicle on the lane for an hour (veh.) z : The total number of vehicle on the lane for thirty minutes (veh.) This study was carried out on six different time period and on three approaches (eight lanes) at the Pekdemir intersection. The sample cases of average delays are given in Table 2. Fifteen-minutes delay observation results are not given in the table beceuse of not close to real delay values (deviations are more than 15%). Each thirty minutes delay observation results and hourly delay observation results are given in the Table 2. Forty eight sample cases were examined and evaluated in this work. First thirty minutes observation results and second thirty minutes observation results were compared with hourly observation results and First thirty minutes observation results and second thirty minutes observation results were compared with each other. The comparison of the results can be seen in Figure 3, Figure 4 and Figure 5.

77

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Approach Name

Mugla

Mugla Hasan Tekin Ada Hasan Tekin Ada Tali

Tali

Table 02. Sample Cases of Average Delays at Pekdemir Intersection Average Average Time Delay Delay Lane Period (sec./veh.) (sec./veh.) 0 – 60 min. 0 – 30 min. Left 40.14 41.51 Weekday Middle 24.79 23.71 17:00 – 18:00 Right 27.37 28.35 Left 39.91 42.81 Weekend Middle 27.54 27.79 12:00 – 13:00 Right 33.23 34.29 Left 51.75 56.45 Weekday 08:00 – 09:00 Right 67.05 67.66 Left 58.24 58.84 Weekend 08:00 – 09:00 Right 70.87 75.70 Left 67.37 74.61 Weekday Middle 68.57 65.12 17:00 – 18:00 Right 17.26 17.86 Left 61.03 59.35 Weekend Middle 54.47 50.95 12:00 – 13:00 Right 19.12 18.20

Average Delay (sec./veh.) 30 – 60 min. 38.73 25.86 26.48 37.22 27.29 32.01 48.17 66.46 57.68 66.36 62.16 71.18 16.53 63.07 57.82 20.00

The Comparison between First Thirty Minutes and an Hourly Observations Average Delay (sec./veh.)

90 80 70 60 50 40 30 20

Average Delay (sec./veh.) : 0 - 60 min. Average Delay (sec./veh.) : 0 - 30 min.

10 0 1

6

11

16

21 26 Sample Cases

31

36

41

46

Figure 03. The Comparison between First Thirty Minutes and an Hourly Observations The Comparison between Second Thirty Minutes and an Hourly Observations

Average Delay (sec./veh.)

90 80 70 60 50 40 30 20 Average Delay (sec./veh.) : 0 - 60 min.

10

Average Delay (sec./veh.) : 30 - 60 min.

0 1

6

11

16

21 26 Sample Cases

31

36

41

46

Figure 04. The Comparison between Second Thirty Minutes and an Hourly Observations 78

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

The Comparison between First Thirty Minutes and Second Thirty Minutes Observation

Average Delay (sec. / veh.)

90 80 70 60 50 40 30 20

Average Delay (sec./veh.) : 0-30 min.

10

Average Delay (sec./veh.) : 30 - 60 min.

0 1

6

11

16

21 26 Sample Cases

31

36

41

46

Figure 05. The Comparison between First Thirty Minutes and Second Thirty Minutes Observations Statistical Tests: As seen on Figure 3, Figure 4 and Figure 5, it can be simply understood that vehicle delay measurements considering different time durations are quite smilar to each other. These similarities are examined statistically considering Kolmogorov-Smirnov and Wilcoxon Signed Rank Test made by using SPSS (Statistical Product and Service Solutions) software. Kolmogorov-Smirnov test used for test of the normality in the statistics is one of the most useful and general nonparametric methods for comparing two samples. H0 and Ha hypotheses of this test can be written in the following: H0: Distrubutions of the data are suitable for normal distrubution. Ha: Distrubutions of the data are not suitable for normal distrubution. If significant value is the greater 0.05, H0 hypothesis is accepted. Thus, it can be said that data are suitable for normal distrubutions. If significant value is less than 0.05, H0 hypothesis is rejected (Kalayci, [11]). Wilcoxon Signed Rank Test is a non-parametric statistical hypothesis test used when comparing two related samples or matched samples. This test can be used as an alternative to the Paired student’s t-test. As Paired student’s t-test is used for data suited to normal distrubution, Wilcoxon Signed Rank Test is used for data not suited to normal distrubution. Data are paired and come from the same population in this test. If significant value is equal to 0.05 or less than 0.05, it can be said that a meaningful difference exists between two data sets statistically according to the Wilcoxon Signed Rank Test (Kalayci, [11]). RESULTS In this work, vehicle delay measurements considering different time durations were investigated statistically. Firstly, Kolmogorov-Smirnov distribution function test was applied on hourly and thirty minutes delay observations. Conformity of delay observation values for the normal distribution was examined. Results of Kolmogorov-Smirnov distribution function test are given in Table 3. Table 03. Results of the Kolmogorov-Smirnov Distribution Function Test Kolmogorov-Smirnov Statistics df Sig. An Hourly Observation Values .129 48 .045 (0-60 min.) Thirty Minutes Observation Values .136 48 .026 (0-30 min.) Thirty Minutes Observation Values .153 48 .007 (30-60 min.) Finally, Wilcoxon Signed Rank Test which is one of non-parametric hypothesis tests was applied on comparisons between first thirty minutes and hourly observation values, between second thirty minutes and hourly observation values and between first thirty minutes and second thirty minutes observation values. Results of the Wilcoxon Signed Rank Test for comparison between first thirty minutes and hourly observation values, results of the Wilcoxon Signed Rank Test for comparison between second thirty minutes and hourly observation values, result of the Wilcoxon Signed Rank Test for comparison between first thirty minutes and second thirty minutes observation values are shown in Table 4.

79

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Table 04. Results of the Wilcoxon Signed Rank Test Wilcoxon Signed Rank Test - Test Statisticsa First Thirty Minutes Second Thirty Minutes First Thirty Minutes An Hour An Hour Second Thirty Minutes -.231b -.210b -.200b

Z Asymp. Sig. .817 (2-tailed) a. Wilcoxon Signed Ranks Test b. Based on Negative Ranks

.833

.841

DISCUSSION In this study, an efficient and time saving method for measuring vehicle delay is searched. One of the main purpose of this study was to save observation time and labor. On the other hand, to get a reliable, efficient and valid observation method. For this aim, instead of hourly observations, fifteen minutes and thirty minutes time periods are considered and validation of these periods are investigated. Kolmogorov-Smirnov distribution function test and Wilcoxon Signed Rank Test are applied for validation search. As can be seen on Table 3., significant values for different time durations are less than 0.05. According to these results, it can be said that distributions of observation values for each time durations are not suitable for normal distribution. Consequently, it is determined that parametric hyphotesis tests can not be applied to the observation values. Thus, Wilcoxon Signed Rank Test is applied. As seen on Table 4, Because of significant values is greater than 0.05, it is concluded that a meaningful differences do not exist between two data set for three comparison and these data sets resemble to each other. Namely, There is no meaningful differences between these time periods. Thus, any of them can be used for measuring purpose. As a result, use of thirty-minute delay observation instead of hourly delay observation is proved considering Wilcoxon Signed Rank Test. New methods and technologies such as plate recognition can be used for automated measurement of vehicle delay in future. Similar investigations should be made for following headways of vehicle and saturation flow measurements at signalized intersection. REFERENCES [1] Webster, F. V.,(1958). Traffic Signal Settings, Road Research Technical Paper, No 39, Road Research Laboratory, Her Majesty Stationary Office, London, UK. [2] Akcelik R. (1981). Traffic Signals: Capacity and Timing Analysis Research Report 123. Australian Road Research Board, Melbourne, Australia. [3] TRB. (1985). Special Report 209: Highway Capacity Manual, Trasportation Research Board, National Research Council, Washington D.C., USA. [4] Ban, X., Herring, R., Hao, P., Bayen A. M. (2009). Delay Pattern Estimation for Signalized Intersections Using Sampled Travel Times, Transportation Research Board of the National Academies, Vol 2130/2009 pp 109-119. [5] Dion, F., Rakha, H., Kang, Y. S. (2004). Comparison of Delay Estimates at Under-Saturated and Over Saturated Pre-Timed Signalized Intersections, Transportation Research Part B-Methodological, Vol 38/2 pp 99-122. [6] Akcelik R. (1988). The Highway Capacity Manual Delay Formula for Signalized Intersections, ITE Journal, pp 2327. [7] Akgungor A. P. (2004). Mathematical Modeling of Delay Estimation at Signalized Intersections I: A New TimeDependent Delay Model for Various Analysis Time Periods, Technology, Vol 7/3 pp 369-379. [8] Murat Y.S. (2006). Comparison of Fuzzy Logic and Artificial Neural Networks Approaches in Vehicle Delay Modeling. Transportation Research Part C-Emerging Technologies, Vol 14/1 pp 316-334. [9] Su, Y., Wei, Z., Cheng, S., Yao, D., Zhang, Y., L , L. (2009). Delay Estimates of Mixed Traffic Flow at Signalized Intersections in China, Ts nghua Science and Technology, Vol 14/2 pp 157-160. [10] Murat Y.S., Kutluhan S., Cakici Z. (2014). Investigation of Cyclic Vehicle Queue and Delay Relationship for Isolated Signalized Intersections, Procedia - Social and Behavioral Sciences, Vol 111 pp 252-261. [11] Kalayci S. (2006). Multivariate Statistical Techniques with SPSS Applications, 2nd ed., 2006.

80

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

ASSESSMENT OF QUALITY INDEX FOR SUBWAY WITH CONCRETE SLAB P. Rakhshani1, M. W. Khordehbinan2 [email protected], [email protected] 1

Department of Civil Engineering, Sanandaj Branch, Islamic Azad University, Sanandaj, Iran, 2 Department of Engineering, Sama College, Sanandaj, Iran,

ABSTRACT Increasing demand for using subway lines, limitation of temporal, financial and human recourses in maintenance of lines and ascending trend of decline in this infrastructure necessitate using systems of line maintenance management as a tool for gaining maximum efficiency, the main goal of which is determining an optimal way for planning and organizing rail line repair and maintenance actions so that costs minimize. Evaluation index for expressing line status as simple and comprehensible information is one of important components in railway maintenance management system. It can indicate trend of line status change by quantifying and processing information over the time, it can also be applied for predicting future status. In addition, by defining limit values for status indexes it is possible to define required threshold for performing repair and maintenance operation and evaluating status development following each operation. Concrete pavement of subway lines is a qualitative process in terms of maintenance parameters which be expressed as a quantitative parameter. Therefore, a specific method is provided in this paper by developing structural and geometrical quality index models in order to calibrate line component quality index model in a field study. In the line under study, line component quality indexes indicated that quantifying qualitative status of line components using proposed index has good results. Keywords: Subway, Maintenance management system, Concrete rail line, Repair and maintenance, Qualitative evaluation index. 1 INTRODUCTION Repair and maintenance issues are introduced as soon as structures are constructed and put into operation. Pavement and infrastructure of railway tracks are no exception and they decline over the time and with repeated use of the track and loadings as well as due to environmental factors. Experience has shown that the best railway tracks in terms of pavement material and operation will become such tracks with minimum technical condition if they are not kept under appropriate maintenance. Thus importance of maintenance and repair becomes more evident in railway tracks compared to ordinary structures (Sadeghi, 2005). Track evaluation stage or simply evaluation is one of the main steps of track management system. It can be said that the main pillar of track management system is the way of making track quality index in such a way that track quality can be expressed quantitatively (Sadeghi, 2005). To this end, various studies were carried out regarding offering a quality index in terms of track condition quality index and track geometry quality index separately. In order to achieve the ultimate goal, failure types in track and their utilization in indexing systems should be investigated so that the method for defining track condition index is obtained. On the other hand, track structure condition index and track geometric condition index are specified using data provided from visual and mechanical inspections, respectively. Most importantly is that development of evaluation models is based on visual and mechanical inspections for concrete tracks (slab tracked). Hence, rating and determining repair and maintenance strategy of concrete tracks can be done as the same as ballast lines. Track geometry is one of factors affecting the level of convenience and track safety. Irregularities developed in geometric characteristics increase vertical and horizontal accelerations imposed on the passenger, they also increase contact forces between rail and wheel. Geometric irregularities should be controlled so as to control imposed accelerations and contact forces. Control and evaluation of track geometric characteristics require correct understanding of geometric parameters (Akbari, 2005). In this work, development of model for evaluating quality of structure and railway concrete pavement based on visual inspection is stated. To this end, initially common failures in slab tracks are defined and are classified according severity level. Then, influence of failure severity groups with various densities on track quality is statistically analyzed in order to convert track qualitative condition into significant quantities. Deduct and correction curves are defined using this influence level which in fact shows track quality deduct value (DV). Finally, final model of track structure quality index based on deduct density is developed which has been used in all evaluation models around the world, for example for roads and airports pavement by Shahin (1976) and railway ballast lines by Uzarski (1996). 81

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

In addition, geometric quality evaluation method development and evaluation of railway concrete pavement based on mechanical inspections fitted to Iran’s railway tracks condition is discussed. There are different methods for analyzing data given by track measuring machine, each of which can state track geometric index quantity. Standard deviation parameter is commonly used in Asian and most European countries for stating track geometric index quantity. For development of geometric quality index in this work, in addition of using standard deviation parameter, average parameter is also used regarding statistical analysis on EM80 machine data used in Tehran subway and statistical investigations and having normal distribution in frequency distribution curves and an index is defined for each geometric parameter. Then, different tolerances in railway concrete tracks in several cities in the world are compared and parameter weighting is used in order to integrate geometric parameters and achieve a single equation for total geometric index.

2 REVIEW OF TRACK QUALITY INDICES NUMERICAL 2.1 Track Structure Quality Index Track structure condition index which is derived from visual inspections and represents qualitative condition of track elements was developed originally by Uzarski in 1993 in U.S. Army Research Institute. Failures were classified according type and severity of them for defining algorithm of this index and then the influence of each type of failure on reduction of track quality was studied using an expert team on railroad and performing statistical analyses. Process of corrected deduct value calculations follows ASTM D5340-93. In this method, first, track elements are introduced and three groupings are presented. Then, failures related to each group are listed and their severity levels are given. Finally the results is given as deduct curves (Uzarski, 1993). 2.2 Track Geometric Quality Index Track geometry represents a situation which is occupied by the rail or path axis in space. ‘Track Geometry’ term is applied to various parameters which describe track path and arrangement. By drawing track geometry in various planes it is possible to determine and measure track geometric parameters and use for track geometry description (Sadeghi, 2005 and Akbari, 2005). Authors and research centers have given different recommendations regarding this index. In India railway, TGI index is used for evaluating track geometric quality. In this method which is based on statistical analysis and standard deviation, weighting average is used for integrating various parameters. Considering impact of each parameter (distortion, direction, track width, profile) on motion index, different weighting coefficients are obtained for different parameters. This method is known as Track Geometry Index (TGI) and it is calculated from following equitation (Mundrey 2003):

2UI

TGI

TI

6 AI 10

GI (1)

Where, TI is distortion index, GI is track width index, AL is direction index, and UI is vertical drop (profile) index. One of the indices based on statistical analysis is index Q. It is used for track condition evaluation in national Sweden railroad. The main criterion in calculating this index is based on weighting between statistical parameter of standard deviation and different parameters of track geometry. Of course, the ratio of deviation value to allowed value is also considered in this index for determining track condition. Index Q is calculated by following relationship (Anderson 2002): Q 150 100

H

2

H Lim

S S Lim

/3

(2)

Where, H is standard deviation average of left and right rail drop, S is standard deviation average of transverse slope, track width and horizontal distortion, H lim and Slim are allowed values for H and S, respectively according track classification. Index J which is used in Poland railroad is based on statistical analysis and different parameters are specified by weighting standard deviation and it provides a quantitative evaluation of the track condition. This coefficient can be obtained from following relationship (Madejski and Grabczyk 2002):

J

SZ

SY

SW 3 .5

0.5 S e

(3)

82

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Where, J is geometric parameters index, Sz is standard deviation of left and right rail drop, Sy is direction standard deviation, Sw is distortion standard deviation, and Se is track width standard deviation. Index ITGI which was suggested by authors in 2005 for Iran’s railway ballast lines (Sadeghi, 2005) is based on using parameters of average and standard deviation considering characteristics of normal distribution and weighting different geometric parameters using allowed tolerances for Iran’s ballast lines for different groups of tracks. ITGI relationship is suggested as follows (Sadeghi, 2005):

ITGI

100 K

a GI 2

a 2

GI

b AI

a a 2

c PI

d CI

(4)

b c d

Where, GI x 3SDG is positive track width index, GI x 3SDG is negative track width index, AI x 3SD A is direction index, PI x 3SD P is profile index, and CI x 3SDC is transverse balance index. K factor for class A tracks is -4.50, for class B tracks is -4.18, and for class B tracks is -3.79. a’, b, c, and d are weighting coefficients of geometric parameters which are obtained according allowed tolerances table. 3 DEVELOPMENT OF TRACK STRUCTURE QUALITY INDEX MODEL Deduct density model was used in this work in order to develop model for evaluation of railway concrete pavements. As mentioned, this model has been used for developing all evaluation models which are based on visual inspections. In all of these models, failure extent depends on three parameters: 1. failure type, 2. failure severity, 3. failure extent which is stated as percent and failure density. Each of these factors affects greatly on recognition and quantifying condition of track structure, thus they should be entered into condition index mathematical model. In this model it is assumed that condition index is assessed by summing failure types with their severity and density using weighting factors (Rakhshani 2009): p

QI

100 F (t , N )

DV

(5)

i 1

Where, QI is quality index, DV is deduct value, P is number of total failure severity level, F is correction factor for several failures which is a function of total deduct values (t) and deduct number (N). As mentioned, DV is deduct value which is itself a function of failure type, failure severity, and failure extent. It is expressed as follows (Rakhshani 2009): DV

f (T , S , D )

(6)

Where, T is failure type, S is failure severity and D is extent or density of the failure. As it can be observed, type of failure, severity of failure and failure density are considered separately or collectively in this model and track qualitative condition is expressed quantitively. For developing a model with above suggested structure for evaluating railway concrete track quality following steps are necessary: 1- Accurate and comprehensive definition of common failures in main elements of railway concrete tracks, 2- Examination and determining influence of each defined failure types with its severity and extent on quality and service of railway concrete track elements, 3- Developing diagrams which show reduction of quality in concrete track elements for different topes, severities and extents of failure. 3.1 Definition and Classification of Common Failures Railway concrete tracks usually consist of elements such as rail, fasteners, sub rail plate and a concrete slab. Some of these elements are considerably different regarding material, decline and application. Thus all elements of the track cannot be evaluated using one specific index. In this study, main elements of the railway tracks are classified into three major groups including: rail (including rail, connection plates (if any), sub rail plates, and seams), 2. Fasteners, and 3. Concrete slab. Quality index is evaluated for each group separately (Rakhshani 2009). Types of concrete slab failures were studied according standards and bylaws of concrete structures, and types of common failures in different elements of the railway tracks were defined regarding views of 24 national railway experts. In this step, influence of different types of failures on concrete track quality should be assessed using statistical analysis (Rakhshani 2009).

83

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

However, regarding that most failures have almost identical impact on track quality, all failures were classified into three levels of severity using opinions by evaluators and based on previous experience in development of visual indices as well as by studying different standards. Three levels of failure severity include: low severity (L): failures which do not affect train activity, medium severity (M): failures which may limit train activity, and high severity (H): failures which may limit train speed or block the track (Sadeghi, 2005 and Uzarski, 1993). Various types of failures are assigned into one of these three levels of severity as shown in Table 1. In this work, considering effect of failures on geometric parameters, track quality is evaluated for investigating failures related to concrete slab. Therefore, failures of this group have been studied and classified according above table. To this end, influential concrete section failures in zones are investigated by zoning concrete slab section in railway tracks and considering a 60cm *60cm space around fastener as shown in Figure 1 (Rakhshani 2009).

Figure 01. Zoning of slab section around fastener (m) Of course, it should be noted that often rail seams are welded together in concrete railway tracks, However, yet rail seam and connecting plates are used in these tracks in place of arcs and in parts which higher longitudinal forces (Rakhshani 2009). Table 01. Failure classification according severity level (Uzarski, 1993 and Ponnuswamy 2005 and ACI 201.,1R-92.) Track Severity Common Structural Failures Structure Level Rail failure: Cleavage, Bump, Surface cracks, Overflow, Slough, Lateral Bending, Chipping and Jagging crest, Rail burning. Low (L) Failure in the seam: Inappropriate size or type of connecting bolts, Inappropriate size or type of connecting plate, Loose bolts, Loose plates, Lost /Bent /Cracked /Fractured bolt. Rail failure: Crushed crest, Base corrosion, Rail vertical abrasion, Rail lateral abrasion, End abrasion (V-shaped flaw), Crest horizontal crack, Crest vertical Medium crack, Welding failure, Crack around bolt holes, Corrugation, and Combined Rail Group (M) crack. Failure in the seam: cracked connecting plate, 4cm 2cm Sub rail plate failure: when it is cracked or broken in inappropriate place Rail failure: Fracture in part of rail, Crest horizontal crack> 4cm, Crest vertical crack> 4cm, Cracks around bolt holes > 1.5 cm, Combined crack > 20 cm percent High (H) of the surface, Jon crack > 1.5 cm. Failure in the seam: All bolts are failed in seam place, Broken or lost connecting plate, Rail seam > 4 cm. Failure in slab: Scaling up to 6 mm depth, Cracking in concrete up to about 0.8 mm width, Spalling in concrete up to less than 12 mm depth, Pop-outs in concrete Low (L) up to about 10 mm diameter, Efflorescence on concrete surface, Staining concrete surface, Damage to seam with good condition, Hollow areas. Failure in slab: Scaling between 6-25 mm depth, Cracking in concrete between Concrete Medium 0.8-3.2 mm width, Spalling in concrete between 12-25 mm depth, Pop-outs in Slab (M) concrete between 10-50 mm diameter, Damage to seam with average condition. Failure in slab: Scaling more than 25mm, Cracking in concrete to width more than 3.2mm, Spalling in concrete to depth more than 25 mm, Damage to seam High (H) with poor condition, Pop-outs in concrete with diameter 50-75 mm (holes with diameter more than 75mm is considered as puncture). Low (L) Failure in fastener: Installation or Placement at inappropriate place Medium Fastener Failure in fastener: Loose or Bending fastener (M) High (H) Failure in fastener: Breaking or Losing 84

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

3.2 Impact of Failures and Related Curves After defining different types of failures in different levels of severities, effect of each group of failure severity on track quality is investigated. To this end, initially a method should be specified for quantifying failures. That is, structural failures within a specific length of track should be expressed in quantitative manner so that it is possible to determine track quality sensitivity for different values and severities of failure. To this end, a specified segment of track is divided into units. Failures density of track elements is given as follows: the number of units with failure divided by total units. In this work, 30m length segments were specified for track and failure density is calculated for this segment. Each unit is considered as distance between two consecutive fasteners in segments. For example, assuming 60 cm distance between fasteners, each 30m segment consists of 50 units. For instance, following concrete slab inspection, if there are 50 units in one segment, and 5 units show one types of failures with medium severity, then concrete slab failure density with medium severity in this segment is 5/50 = 0.1. Failure density for rail and fasteners are calculated in similar way (Rakhshani 2009). The main part of developing concrete track quality index is investigation of influence of failure severities in different elements on track structure quality. To this end, statistical analysis and data collection from evaluator team is necessary. In this study, a 24-member group of railway experts were utilized. In this step, evaluator team members were used for quantitative assessment of effect of different failure types on track structure quality. For this purpose a series of sheets were prepared and each sheet showed one type of failure with specific density and severity. These sheets were given to evaluator team members and they were asked to rate each sheet. This rate represents deduct value of track quality according which the effect of the failure with its density and severity are specified. In fact, the rate given to each sheet by evaluator denotes that according to his opinion, the failure with specific density and severity (specified in the sheet) distorts that part of concrete track quality from 100 to deduct value (100 - DV). Table 2 gives an example of rates given by evaluators for rail failures with low severity (Rakhshani 2009).

Failure density 10 30 50 70 100 Failure density 10 30 50 70 100

A 14 28 45 45 47 M 14 32 42 48 47

Table 02. Example of data provided by evaluators Rail evaluation for failures with low severity Deduct values provided by evaluators B C D E F G H I J 9 15 23 18 21 16 17 26 20 26 30 40 32 40 38 35 43 37 48 44 57 45 56 52 53 51 49 47 45 60 48 56 56 51 58 52 49 50 62 55 63 59 58 62 58 Deduct values provided by evaluators N O P Q R S T U V 15 23 17 15 24 19 16 18 20 31 30 37 35 43 37 34 33 33 47 42 48 49 46 40 37 36 36 49 51 54 51 56 55 49 56 56 51 54 62 56 65 55 65 63 63

K 14 33 44 50 60

L 18 42 50 56 65

W 21 39 42 53 56

X 17 36 39 54 60

Following collection sheets, evaluations were studied accurately and the rates which had distance to average value as much as 2 times of standard deviation were excluded for reevaluation. It was for this reason that evaluators are given another chance to correct their mistakes in rating. At the end, collected data were evaluated again. In this step, assuming normal distribution for statistical data, rates which were higher than average as much as 3 times of SD or were smaller than average as much as 3 times of SD were excluded from statistical population and evaluations were focused on remaining data (Rakhshani 2009). On the other hand, when two or three types of failures with different severities occur in one element of the track, influence of each type of failure on track quality is reduced. Thus, for investigating it and obtaining deduct value for this case, other sheets were prepared and given to evaluators. Two or more types of failures in one element of the track were shown in these sheets simultaneously. For example, in one sheet representing a segment of the track, one type of rail failure with low severity and one type of rail failure with medium severity were shown at the same time. The overall process of gathering information from the evaluators is the same as one described above. Specific sheets were prepared from combinations of failures and evaluators were asked to rate sheets. Correction factors were obtained by comparing these rates with rates previously given by evaluators for individual types of failures. As mentioned in suggested model’s equation, correction factor is a function of total deduct values for each failure and the number of failures. Finally, results obtained from analyzing data provided by evaluators were converted into related curves. 85

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

RL

Rail,N=2

Rail,N=1

RM

RH

100 90

100 90 80 70 60 50 40 30 20 10 0

80

Deduct Value

Corrected Deduct Value

Rail,N=3

70 60 50 40 30 20 10

0

20

40

60

80

100

Sum Of Deduct Values

120

140

160

0 0

10

20

30

40

50

60

Density Failur

70

80

90

100

Figure 02. Deduct curves for rail failures Figure 03. Correction curves for rail failures Figure 2 shows deduct curves for concrete track structure using average of given rates which are drawn using track structure quality reduction versus different densities. In addition, values for correction factors are given in Figure 3, where N is the number of failures. For example, if there is a failure density value for low density failures in rail and a failure density value for medium severity failure in rail for one segment, first, deduct value per density should be obtained using Figure 2, then deduct values are added and using Figure 3 curve, corrected deduct value is calculated by multiplying this correction factor by sum of deduct values. Finally, rail quality index for this segment is given as 100 minus corrected deduct value (Rakhshani 2009). 3.3 Repair and Maintenance Planning Ultimate goal of railway track repair and maintenance systems is achieving safe track performance and increased convenience of passenger as well as reasonable repair and maintenance cost. The model suggested in this study for railway concrete track quality evaluation can serve as a useful tool for this purpose. To this end, specific parts known as ‘managerial parts’ should be defined for tracks so that track repair and maintenance planning is done for them. For this purpose, classification of concrete tracks into following managerial parts is suggested: 1. direct concrete track with maximum 1,000m length, 2. concrete track in arc, 3. concrete track in tunnel, 4. concrete track in bridge with opening over 4 m, 5. concrete track in station, 6. concrete track in railroad switch. In effect, visual inspection of the whole track is not possible along a path. Thus, following specifying managerial parts in respective path, parts are divided into 30 m segments. A sample of segments is selected by repair and maintenance experts. Visual inspection is performed in selected segments and track structure index is calculated along selected segments. Track structure quality index along selected managerial part is obtained by averaging these indices. Quality indices should be able to express track structure condition in qualitative and quantitative manner. Suggested index value for this study is considered according scale 0 to 10. In order to plan for repair and maintenance, numerical values of indices should be related to repair and maintenance needs. Considering opinion of evaluators and by studying available standards, qualitative condition of concrete track was classified into 5 groups: very good, good, medium, poor, very poor. Table 3 presents defined relationship between index numbers and qualitative titles.

Numerical range 91-100 71-90 51-70 31-50 0-30

Table 03. Repair and maintenance strategy (Sadeghi, 2005 and Uzarski, 1993) Qualitative Description title Very low flaws; Undamaged track performance; No need for immediate Very good operation; Routine and preventive repair and maintenance is needed. Average failure; Damaged track performance but not serious damage, Routine Good and repair and maintenance, Trivial repair and sometimes optimizations are needed. Serious failure; Damaged track performance but not serious damage, Routine and Medium repair and maintenance, Trivial repair and sometimes optimizations are needed. Severe failure in low percent of the path; Damaged track performance and Poor critical repair and sometimes reconstruction is needed. Failure in high percentage of the path or the whole path; Halted performance of Very poor track, and Critical repair and sometimes renovation is needed.

86

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

4 SUGGESTED MODEL APPLICATION In order to investigate feasibility and reliability of suggested model, field surveys were performed and track measuring machine data from parts of Tehran subway tracks for visual and mechanical inspections were used. Tehran subway tracks are among the main concrete tracks used in Iran. This tracks with high traffic volume daily transport many passengers to different points of the city. In Tehran subway, the system similar to Rahda system has been used, except that concrete slab is not reinforced and due to high thickness there is no need for armature and it is implemented as weighting. Rais type UIC54 and Vossloh fasteners have been used in these tracks. Field inspections were performed in parts of these tracks shown in Figure 4 (Rakhshani 2009).

Figure 04. Schematic map of inspected locations

Left Rail

Right Fasteners

Right Rail

Failure Severity Level

Lef Fasteners

1

3 5 7

9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49

1

3

5

Unit Number

F ailure Severity L evel

Failure Severity L evel

4.1 Visual Inspections and Data Analysis Several segments of different points in Tehran subway were selected for visual inspection, and field studies were conducted in those points. In this study, 7 30m segments of different points in Tehran subway were selected. They were selected according availability and relative variety of failures and structural quality in segments. An evaluation form considering 30m length segment was designed for inspecting respective paths. Designing evaluation forms for field activities is necessary, and the more accurately and simply they are designed, the more data with higher reliability and in higher speed will be obtained. As mentioned, each 30m segment consists of 50 units and there is one unit distance between two consecutive fasteners. In fact, qualitative condition of rail, concrete slab, and fasteners is represented well in every 30 cm (the distance between two consecutive fasteners). Figure 5 shows data taken from one inspected segments in diagrammatic form (Rakhshani 2009).

7

9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49

Unit Number

Concrete Slab

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49

Unit Number

87

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Figure 05. Data taken from one segment Using data recorded in inspection forms, density of failures of track elements are calculated. Table 5 gives indices calculated in 7 inspected segments. Using failure density values obtained for concrete track elements and using diagrams developed in this study, quality indices for rail, concrete slab, and fasteners are calculated (Rakhshani 2009).

Table 05. Calculated indices for track elements Segment No. Rail Calculated Concrete slab indices Fastener

1 81 85 96

2 80 89 95

3 96 82 96

4 87 92 98

5 91 87 92

6 97 96 94

7 84 90 93

By comparing quality index values and suggested repair and maintenance table, repair and maintenance strategy can be specified. One of advantages of visual evaluation method suggested in this study is that it allows diagnosis of main reasons of reduced quality in different parts of the track. Regarding that indices for rail, concrete slab and fasteners are calculated separately, by comparing these values it is possible to identify lower quality elements and plan more accurately for repair and maintenance so as to increase track quality. An example of concrete slap failures can be seen in Figure 6 (Rakhshani 2009).

Figure 06. Damaged and puncture type failure

4.2 Mechanical Inspection and Data Analysis EM 80 machine was used for recording track geometry in Tehran urban subway. Following visual based inspection, in order to investigate track geometric condition, parts of data recorded by this machine were analyzed. After processing taken geometric data, geometric indices were calculated. Sampling by this system is done in high speed and sampling rate is one sample per 25cm. Considering methodology of calculation of track geometric quality index, parameters needed for calculating index include track width, longitudinal balance (profile), crowbar and distortion. Measuring machine outputs for a 30m segment on diagram is given in below. Statistical characteristics of these diagrams are given in Table 6 (Rakhshani 2009). Geometric quality indices in the paths under test are calculated using data recorded by machine based on methodology presented in this work. Following table presents indices calculated for 4 sample segments.

Number Average SD Variance Range Minimum Maximum

Table 06. Statistical characteristics of calculated data TWS ALR ALL GAU RP 121 121 121 121 121 -0.04 0.46 0.01 0.14 -0.19 0.50 0.56 0.50 0.65 0.51 0.25 0.31 0.25 0.42 0.26 2.03 2.69 2.03 2.40 1.84 -0.97 -1.17 -.0.97 -1.23 -1.12 1.06 1.52 1.07 1.17 0.72

LP 121 0.17 0.58 0.33 2.25 -1.06 1.19

88

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Segment No. 1 2 3 4

Table 07. Values of calculated geometric indices Calculated indices OTPI TI AI GI 1.80 1.54 1.82 2.09 1.59 1..31 1.66 1.87 1.36 1.73 1.21 1.43 1.62 1.76 1.49 1.57

PI 1.81 1.53 1.02 1.61

By comparing values of calculated geometric indices and allowed values (0 < OTGI < 3.02) it is observed that respective segments meet necessary requirements for train passing (Rakhshani 2009). 5 CONCLUSION Development of subway concrete track quality index was studied in visual and mechanical inspection areas. Visual inspection in respective path and then calculation of track structure condition indices (95 for fastener, 89 for concrete slab, and 88 for rail) indicated that quantifying qualitative condition of Iran’s subway concrete tracks using used methodology and as expected had acceptable results in this study and classification of failures and their severities for diagnosis and covering types of failures is appropriate. Subway concrete tracks are classified into groups of rail, fasteners, and concrete slab, which are different in terms of their material, decline and application. In this work, quality index was evaluated separately for above groups. In fact, using quality index of separate groups, it is possible to show tolerance for passing traffic and needs of groups to repair and maintenance. In mechanical evaluation and statistical analysis on obtained results from track measuring machine it was found that frequency distribution of geometric parameters follows normal distribution properties. Considering this fact in developing geometric indices in addition to standard deviation, average parameter was also considered. Geometric index result for respective track is between 0 and 3.02 considering allowed value, which meets necessary requirements for train passing. 6 REFERENCES ACI 201.1R-92., Guide for Make a Condition Survey of Concrete, ACI Manual of Concrete Practice. Akbari, B., (2005), Track geometric condition algorithm for Iran railway maintenance management system, B.S. thesis, Railway Engineering College. Anderson, M., (2002), Strategic planning of track maintenance, Department of Infrastructu, Borlänge, Sweden. Madejski, J., and Grabczyk, J., (2002), Continuous geometry measurement for diagnostics of tracks and switches, Proceedings of the International Conference on Switches Delft University of Technology, Delft, The Netherlands. Mundrey, J.S., (2003), Railway track engineering, Tata McGrew-Hill Publishing. Ponnuswamy, S., (2005), Highway and rail transit tunnel inspection manual, Federal Highway Administration. Rakhshani, P., (2009), Model for determining railway concrete pavement quality index, MA thesis, Tarbiat Modares University, Tehran. Sadeghi, J., (2005), Research project report of implementing Iran railway repair and maintenance system, Railway Engineering College. Uzarski, D.R., (1993), Condition indexes for low volume railroad track: condition survey inspection and distress manual, USACERL. Uzarski, D.R., (1993), Development of condition indexes for low volume railroad track, USACERL.

89

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

ASSESSMENT OF ANALYTICAL TECHNIQUES OF FLEXIBLE PAVEMENTS

M. Keymanesh1, E. G. Salehabadi2, I. Khajehhassani3, M. W. Khordehbinan2 [email protected] 1, 2 3

Department of Civil Engineering, Payame Noor University, PO BOX 19395-3697 Tehran, Iran. Department of civil Engineering, Islamic Azad University, SouthTehran Branch, Tehran, Iran.

ABSTRACT By development of the range of using Finite Element Method in road-construction industry during recent years, pavement construction engineers have tried to adapt this method to analyze pavements. Finite Element Method is able to analyze stability, time- dependent problems and those problems with non- linear properties for the material; although, this method is widely used for several sciences, but one of its weak points is in that it takes a lot of time for analysis on pavement as well as its requirement for advanced code-writing. In the present paper, flexible pavement have been analyzed by means of two techniques: Finite Element Method and Theory of Multi-layer System. Eventually, from statistical viewpoint, the results of analysis on these two techniques have been compared by significance parameter and correlation coefficient. Results of this paper indicate that results of analysis on finite elements are most appropriately complied with results came from theory of multi-layer system and there is no significant difference among the mean values in both techniques. Keywords: Flexible pavements, Finite Element Method, ABAQUS, KENLAYER 1. Introduction Multi-layer System: Boussinesq is one of the first researchers, who introduced some formulae based on a concentrated load applied on an elastic half space that was used for analysis of single-layer isotropic pavement [Boussinesq, 1885]. Gazetas introduced a formula (foundation stiffness) that was correspondent to Boussinesq’s equation for linear anisotropic materials [Gazetas, 1982]. For the first time in 1942, Burmister proposed equilibrium equations for twolayer pavement by assuming homogeneity in materials. Several years later by invention of computer, these formulae were generalized for three-layer and then to multi-layer pavement systems [Hildebrand, 2002]. The basic assumptions which were used in equations by Mr. Burmister included [Huang, 2005]: - Each layers is homogeneous, isotropic and linearly elastic. - The materials is weightless an infinite in areal extend. - Each layer has a finite thickness but the lowest layer (subgrade) is infinite in thickness. - A uniform pressure is applied on the surface over a circular area. -Continuity conditions are satisfied at the layer interface. In this study, KENLAYER [KENLAYER, 2011] computer program is used. The backbone of KENLAYER is the solution for an elastic multilayer system under a circular loaded area. The solutions are superimposed for multiple wheels. Thus can be applied to single, dual, dual-tandem, or dual tridem wheels. Finite Element Method Engineers and physicists usually describe a physical effect by means of system of ordinary and/or partial differential equations that apply to certain region (limit) and boundary and primary apt conditions. In fact, a differential equation with its needed boundary and primary conditions is a perfect mathematical model of an effect. In order to find distribution of the given variables which their relation is expressed in differential equations by the given dominant equation, the aforesaid equation should be solved so that to obtain numerical values of any related quantity at the given point. Finite Element Method is a numerical instruction to solve physical problems described by differential equations. This method has two characteristics that make it distinct from other numerical techniques: - In this method, an integral formulation is employed to create an algebraic equations system. - At this technique, smooth functions are continually used to estimate unknown values. Instead of calculation of analytical response of equation for all continuous points in Finite Element Method, the approximated response of equation is computed only in limited number of concerned points in the range. In fact, by application of such points, continuous range of this model is converted into a continius region [Huang, 2005]. These points are simply called nodes. Some part of a continuous region that is limited among some nodes is called element. 90

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Although, arising of this technique is ambiguous, but its privileges are obviously visible. This technique is able to analyze stable problems, time- dependent problems and those problems with non- linear properties for material(s). Some advance computerized programs have been written in these fields that are comprehensive and independent from a certain problem and/ or specific to certain user. Moreover, such users may use subsidiary programs to create network in order to identify form geometry and graphic analysis on results. Finite Element Method is basis for many ComputerAid- Designs (CADs). Since different properties of pavement may be modeled in this method; therefore, this technique is a very appropriate technique for pavement analysis. In this paper, ABAQUS 6.11 i.e. [Abaqus, 2011] power software is used for modeling of pavement. pavement damage: Deterioration of pavement structures may occur due to adverse effects of environmenal factors, traffic loading, construction dificien, and/or poor maintenance strategies [Huang, 2005]. Two failure mechanisms were considered in this paper, covering the most serious load-associated pavement distresses i.e. fatigue cracking and subgrade rutting. a) Fatigue Cracking: Fatigue cracking is caused by repeated axle load applications, usually lower than the strength of the material. It is a progressive localized damage due to fluctuating stresses and strains in the material and a build-up of irrecoverable strains [Hsu and Tseng, 1996]. Fatigue cracking usually starts at the bottom of the HMA layers, which represents the location of the greatest tensile strain in case of fully-bonded conditions between the different HMA layers. Fatigue cracking may also start at the bottom of the individual HMA layers if unbonded or friction conditions exist. In addition, the following transfer function was used to determine the number of cycles to cause Fatigue Cracking [Adu-Osei, 2000]: Nf = 0.0796 ( t)-3.291 (E)-0.854 (1) where, Nf = Number of repetitions for fatigue cracking; t = tensile strain at the bottom of the HMA layers in microstrain; and E is the resilient modulus of HMA in psi. b) Subgrade Rutting: Subgrade rutting is a longitudinal depression in the wheel path that occurs due to excesive densification of subgrade materials and/or lateral movement of the portion of the pavement layers above the subgrade causing surface depressions in the wheelpath. In the model presented by asphalt institute, the strain on top of the subgrade layer is considered as critical strain. This model includes [Adu-Osei, 2000]: Nf = 1.365 10-9 ( c)-4.477 (2) Nf = number of repetitions for subgrade rutting failure; and c = compressive strain on top of the subgrade. 2. Numerical modeling Throughout this investigation, the finite element commercial code ABAQUS 6.11-1 is used. Numerical simulation is performed by the finite element solver ABAQUS/Explicit (ABAQUS user's manual 2011). The simulation procedure is done in several steps. First the exact geometry and dimensions and other data of experimental model from reference (Al-Qadi et al. 2004). were extracted to produce the numerical model. Second, the simplest materials model (linear Elastic behavior) is used to introduce materials behavior of pavement structure. Third, quasi static loading is also performing to acquire real simulation of experimental model. 2.1. Materials model To define the materials response of the pavement structure correctly, First elastic behavior is used. For acquire this purpose according to the materials consist of the pavement structure (See Fig. 1(a)), the 5 layer of pavement structure plus the soil layer is simulated. In Fig. 1 (b) all 6 layer of pavement structure is shown. It's noticeable that in this simulation the soil layer is modeled sufficient high to avoid boundary condition effect in bottom of the model. Finally, in every step, by considering one of three parameters as variable such as elasticity modulus layers of pavement, elasticity modulus layers of subgrade and value of stress applied due to passing vehicles, the pavement has been analyzed in 100 stages separately by ABASQUS and KENLAYER software. In the models which its variable is value of stress due to passing vehicles, properties of pavement materials and subgrade layer correspond to table 1 and also in the models which its variable is subgrade elasticity modulus, properties of other layers correspond to table 1 and value of stress due to vehicles is 750 kp. Table 01:Back calculated Pavement Moduli Layer Tickness Young’s Modulus (mm) (MPa) Surface Mix(SM) 38 4230 Base Mix (BM) 150 4750 Layer

Poisson’s Ratio 0.33 0.30 91

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

OGDL Cement Stabilized Base Aggregate Subbase Layer Subgrade

75

2415

0.30

150

11000

0.25

175

310

0.35

Infinite

262

0.35

(a) (b) Figure 01.(a) Layers of pavement structure and (b) All 6 layers of simulated pavement structure 2.2. Type of element and mesh pattern The pavement structure is meshed using an 8-node continuum linear brick reduced integration element (C3D8R element), that is shown in Fig. 2. Also the structured meshed with variable thicknesses, depending on the layers model is shown in Fig. 3. At the loading regiqon of the model, stress concentrations are happened and therefore for acquire accuracy in results fined mesh and smaller elements must be used. The large size of the elements in the model is 1.231 m, and in the loading region is 0.02053 m (fine mesh in loading region and coarse mesh far from it). The fine meshed in loading region is shown in Fig. 4. The total number of element is 108630 and the mesh convergence study is executed to find this optimum number of element.

Figure 02. C3D8R element

Figure 03. Meshed model of pavement structure 92

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Figure 04. Fine mesh in loading region

2.3. Loading and Boundary condition Symmetric boundary condition is used in simulation and soil layer is modeled sufficient high to avoid boundary condition effect in loading direction. For simplify simulation the tire and contact between tire and pavement structure is neglected and uniform pressure at 3 regions and whit time dependency is defined as tire loading. In Fig.5 the loading and boundary condition is shown.

Figure 05. Loading and Boundary condition

2.4. Solution method In order to model a dynamic phenomenon like Transient Rolling of Tires with ABAQUS, it is possible to solve the problem as dynamic or quasi static solution by Explicit or Implicit algorithm available in ABAQUS (ABAQUS user's manual 2011). In the present simulation, explicit algorithm available in ABAQUS/Explicit is used, that is useable for both dynamic and quasi static solution. This method is chosen, because the governing equations are solved with the full advantage of its computational efficiency and its inherent effectiveness at solving dynamic models (ABAQUS user's manual 2011). Explicit algorithm which is used in ABAQUS/Explicit, obtained the values for dynamic quantities at t + t, which is based entirely on available values at time t. The central difference operator, which is the most commonly used explicit operator for stress analysis applications, is only conditionally stable. The stability limits being approximately equalled to the time for an elastic wave to cross the smallest element dimension in the model (ABAQUS user's manual 2011). 3. Numerical Results As two important criteria, vertical compressive strain on top of subgrade layer of pavement and horizontal tensile strain atthe bottom of asphalt layer are considered in construction of asphaltic pavements. These parameters are obtained from analytical analysis of selected pavement structures modeled by ABAQUS and KENLAYER computer programs. Statistical evaluation of these parameters were conducted by means of Minitab and SPSS software. 3. 1. Statistical Analysis by Minitab At first step, Minitab statistical software was employed to examine and analyze the relationship between the results obtained from Finite Element Method and layered elastic theory technique. The estimated (fitting) formula is shown in Figs 6 and 7. From these figures it may be concluded that the fitting formula is appropriate and has the acceptable accuracy from statistical point of view. 93

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Figure 06. Horizontal tensile strain at the bottom of asphalt surface layer.

Figure 07. compressive strain on top of subgrade.

In order to determine the significance of results from testing, hypothesis of two dependent means (T-Test) has been adapted. The results of T-Test are given as follows:

Since parameter p is greater than 0.05 (P>0.05) at confidence level of 95%,so one can conclude that there is no significant difference between the means values of the result obtained from these two analytical methods. As a result, hypothesis of equality is rejected; in other words, both methods are appropriately incompliance with each other. 3. 2. Statistical Analysis by SPSS SPSS statistical software was employed to examine and analyze the relationship among the results came from Finite Element Method and techniques of layers theory. In order to determine the significance of results from testing, hypothesis of two dependent means (T-Test) has been adapted. The results of T-Test are given as follows: values of longitudinal tensile strain in center of load at the bottom of HMA layer were analyzed in KENLAYER and ABAQUS software statistically. The results of T-test in table 2 show that there is not significant difference between averages T (197.388) = -2.31, P>0.05}. 94

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Table 02. Result of Independent T-Test of Longitudinal tensile Strain at the bottom of the HMA layer Longitudinal.Strain N Mean (m/m) Std. Deviation Std. Error Mean T df Sig.

KENLAYER

100

0.0000237263

0.00001466253

0.00000146625

ABAQUS

100

0.0000242198

0.00001550299

0.00000155030

-0.231

197.388

0.817

The results of table 3 show that there is a not significant difference between values of vertical compressive strain in center of load on top of the subgrade gained by KENLAYER and ABAQUS software T (398.938) = 2.781 , P>0.01}and values of vertical compressive strain resulted in pavement analysis in two soft wares is the same. Table 03. Result of Independent T-Test of Vertical compressive Strain on top of the subgrade Longitudinal.Strain N Mean (m/m) Std. Deviation Std. Error Mean T df Sig.

KENLAYER

200

0.00005106

0.000032578

0.000002304

ABAQUS

200

0.00004400

0.000032175

0.000002275

2.781

398.938

0.030

Finally, it was began to analyze flexible pavement operation according to two scale of fatigue cracking and permanent rutting in ABAQUS and KENLAYER software for this purpose, it was used hypothesis of two dependent means (Ttest).The results of table 4 for fatigue cracking shows that there is not significant difference between number of repetition for fatigue cracking gained by ABAQUS and KENLAYER software T (35.972) = -1.00 , P>0.05}.According to table 5, the results of number of repetitions for subgrade rutting failure have not significant difference with each other T (47.970) = 0.065 , P>0.05}, hypothesis of equality is rejected; and the values are the same. In other words, the two methods are very similar to each other. In figure 8, it is shown averages of rutting failure number of repetitions for fatigue cracking gained by ABAQUS and KENLAYER software Table 04. Result of Independent T-Test of Fatigue Cracking Fatigue Cracking N Mean Std. Deviation Std. Error Mean T df Sig. KENLAYER

25

9.8291E8

1.51627E9

3.03254E8

ABAQUS

25

1.6433E9

2.93315E9

5.86629E8

Subgrade Rutting

-1.000

Table 05. Result of Independent T-Test of Subgrade Rutting N Mean Std. Deviation Std. Error Mean T

KENLAYER

25

1.7671E10

3.97806E10

7.95612E9

ABAQUS

25

1.6952E10

3.87901E10

7.75803E9

0.065

35.972

0.324

df

Sig.

47.970

0.949

95

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

KENLAYER

Fatigue Cracking

ABAQUS

Subgrade Rutting

Figure 08. Mean Value 4. Conclusions The coclusions of this analytical research study may be categorized as follows: There is not significant difference between average of longtitude tensilestrain value in center of loadat the bottom of HMA layer in linear elastic analysis by ABAQUS and KENLAYER software. In linear elastic analysis of flexible pavement, there is not significant difference between values of vertical compressive strain in center of load on top of the subgrade layer in ABAQUS and KENLAYER software. There is not significant difference in damage analysis for value of fatigue cracking based on vertical tensile strain at the bottom of the HMA in ABAQUS and KENLAYER software. KENLAYER software can be a good replacement for ABAQUS software in linear elastic analysis. The results of ABAQUS and KENLAYER software don’t have any significant differences in value of permanent rutting based on vertical compressive strain on top of the subgrade and hypothesis of averages being same is failed, in other words, the two methods have a suitable similarity.

5. References [1] Abaqus. 2011. Abaqus/Standard User’s Manual Version 6.11.1. [2] Adu-Osei, A. (2000) "Characterization of Unbound Granular Base". Ph.D. Dissertation, Texas A&M University, College Station, TX. [3] Al-Qadi, I.L., Elseifi, M.A.andYoo, P.J. (2004) "Pavement Damage to Different Tires and Vehicle Configuration",Michelin Americas Research and Development Corporation Ichelin Americas Research and Development Corporation 515 MichelinRoad PO BOX 1987Greenville, SC 29602-1987 the Roadway Interastructure Group Virginia Tech Transportation Institute350 Transportation Research Plaza Blacksburg, VA 24060. [4] Kenlayer. 2011. User’s Manual. [5] Minitab. 2009. User’s Manua Version 15l. [6] CurveExpertr. 2010. User’s Manual Version 3.1. [7] Boussinesq. J.1885. “Application des Potentielsal’etude de l’equilibreet du Mouvement des Solids Elastiques “. Gauthier-Villars. Paris. [8] Gazetas .A.M. 1982,“Stresses and displacements in cross-anisotropic soils.” J. Geotech.Engrg.Div,108(4)¸ 532-553. [9] Hildebrand. G. (2002)¸ “Verification of Flexible Pavement Response from a Field Test” Report121 of Danish Road Institute¸ Denmark. [10] Hsu, T.W., and Tseng, K.H. (1996). “Effect of rest periods on fatigue response of asphalt concrete mixtures.” Journal of Transportation Engineering, American Society of Civil Engineering, Vol. 122, No. 4, 316-322. [11] Y.H. 2005. “ Pavement Analysis and Desing”, 2st Edition¸Prentice Hall ¸ Englewood Cliffs¸ NJ.

96

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

A MODEL APPLICATION TOWARDS SOLUTION OF CAR PARKING PROBLEM IN TURKIYE: IZMIR ALSANCAK MULTI-STOREY FULLY AUTOMATED CAR PARK

M. S. Yard m1, M. A rikli2

[email protected], [email protected] 1

ld z Technical University Civil Engineering Faculty, Esenler, stanbul, Turkiye Tel: (212) 383 51 83 Otomatik Otopark A. . 2 Karaa aç Cad. No: 14 Sütlüce, 34445, Beyo lu- stanbul, Turkiye Tel: (212) 255 79 75

ABSTRACT One of the problems affecting the parking sector in Turkiye is the fate of parking fees collected under the Turkish Parking Legislations in the places where there is an insufficient amount of parking capacity. zmir Metropolitan Municipality is the first in Turkiye to consciously allocate this money directly to construction of car parks. One product of this pioneering approach is construction of a multi-storey automated car park. This type of car parks are intelligent systems, where users themselves do not enter the parking lot. Instead, they leave their cars in a reception hall; then the cars are automatically placed into the available lots by means of automated transport devices. In this proceeding, general information will be given about operating techniques of automated car park systems, as well as recent developments and current situations related to them. Then, zmir Alsancak Multi-storey Fully Automated Car Park will be introduced. Some perspectives will be presented about the contributions of it towards solution of car parking problems. Keywords: zmir, multi-storey fully automated car park, parking fees, parking problem. INTRODUCTION Car parking problem is the generic name to the situations that arise when drivers are unable to find a parking place for their cars when they reach to their destinations. Parking problem is defined as the source of many problems in transportation, city planning, economics, safety, social justice, health and environment. Generally, in the central regions of the cities or compact urban campus-type areas where intense urban activities take place, there is a land problem. Available lands are either too small to erect big buildings or too expensive when costs of the alternative investments are considered. Therefore, multi-storey automated car parks (MACP) on small unused lands emerge as a viable alternative. These are intelligent systems, where users themselves do not enter the parking lot. Instead, they leave their cars in a reception hall; then the cars are automatically placed into the available lots by means of automated transporters. One of the problems affecting the parking sector in Turkiye is the fate of parking fees collected under the Turkish Parking Legislations in the places where there is an insufficient amount of parking capacity. In one research study, it was revealed that, approximately 60% of cities in our country, the parking fees were declared to be used in the construction of the parking facility. The status of the remainder is unknown. As of 2012, the municipality having biggest amount of money generated by parking fees in its account is the zmir Metropolitan Municipality. zmir Metropolitan Municipality is the first in Turkiye to consciously allocate this money directly to construction of car parks. One product of this pioneering approach is construction of a MACP. In this proceeding, general information will be given about operating techniques of MACP systems, as well as recent developments and current situations related to them, both in the world and in Turkiye. Then, zmir Alsancak MACP, in which some new techniques are employed for the first time, will be introduced. Some perspectives will be presented about the contributions of it towards solution of car parking problems in Turkiye. MULTI-STOREY AUTOMATED PARKING SYSTEMS (MAPS) Mechanical parking systems were first introduced using freight elevators nearly the time of World War I in the USA. After a few decates a series of other patents were granted. “Some of these early systems were vertical elevator lift modules that placed cars on upper levels of a structure to be moved by attendant and others mechanical devices that could move vehicles into ‘slots’ in a framework built around a central corridor [1]”. “The first generation of mechanical systems were very primitive and capital intensive. Some of them had elevators and rotating floor levels in order to align a free spot with the elevators. The next step in the evolution of parking systems was the concept of automation. The idea was to process the car automatically without requiring the presence of a driver [2]”. 97

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

“MAPS first appeared in Europe as early as the 1900’s and in the North America in the 1920’s. The need for MAPS then is the same as it is now: maximize the value of available land by condensing parking. The 1950’s saw the peak of the industry in North America with a number of high profile systems built but demand for the systems fell off shortly after that time. Although demand in other parts of the world, notably Japan, Korea and parts of Europe, continued to increase for automated parking systems , there was no real interest in North America until the 21st century. Since the turn of the century there have been around 15 systems installed in North America and the rate of installation is increasing [3]”. “There are generally two types of MAPS designs. One type consists of separate horizontal transport devices at each level and separate vertical lifts. A vehicle enters the transfer compartment and the patron vacates the stall and activates the storage process. A door opens on the side toward the storage vault and a motorized carrier slides under the vehicle and picks it up with arms that fold onto the wheels, or the car is on a pallet that is lifted slightly and removed from the compartment. The vehicle is then carried horizontally over to a lift platform and placed on the lift. The lift moves vertically to the desired storage level and the transport device carries the vehicle into the transport aisle and down to a storage space where it is deposited in a stall. The retrieval process works in the reverse order. The second type is called a stacker crane, which is a motorized device that runs horizontally on rails on the ground floor with the lift mechanism built into the same piece of equipment. All vertical and horizontal movements are accomplished with this single device. Because multiple transport devices can be accommodated with the former system, it has more redundancy and more throughput capacity than the stacker crane system [4]”. MAPS are becoming widespread with each passing day. Known major manufacturers of the country are: China, USA, India, Germany, Italy, Japan, South Korea, Netherlands, Switzerland, and Turkiye. “The Emirates Financial Towers in Dubai, UAE, currently has the largest automated parking facility that can store 1,191 cars and occupies a net internal area of 27,606.14 m² (297,150 ft²). The modular parking system is capable of multiple, simultaneous rapid pallet movements and is programmed to control a peak capacity of 360 cars per hour. The facility was completed on 26 June 2011 [5]”. Compared to conventional parking, MAPS has several advantages (Table 1), although the greatest advantage is the space used. MAPS require 30–50% less volume to park the same number of cars. On the other hand, for some types of parking facilities higher construction and operation costs are the disadvantages of the system (Table 1-Table 3). Table 01. Disadvantages and Advantages of Automated Parking Systems [4] Disadvantages Higher construction cost than above-grade, stand-alone, efficient conventional garage. Higher maintenance and operating cost than conventional garage. Added cost to replace electrical components, computer hardware and software at the end of rated life. Queuing during peak usage in some applications.

Advantages Double the parking capacity in the same volume as a conventional garage. Half the volume for the same capacity as a rampaccess garage. Smaller footprint to enhance development opportunities. More secure, unoccupied storage vault. Less lighting, no ventilation of vehicle emissions, industrial stairs, no pedestrian elevators Enhanced user experience (automated valet parking) No attendants required.

“Construction costs are site specific so it is difficult to compare the cost of an automated garage to the cost of a conventional garage (Table 2). When excavation is required, the advantages of MAPS can result in construction costs being significantly cheaper compared to building a conventional garage. When there aren’t any real site restrictions, an MAPS will cost approximately 40 percent more to construct [6]”. “However, operating costs for a conventional garage are considerably higher with greater needs for maintenance, security, cleaning, snow and salt removal. Table 3 is the comparison of operating costs of a proposed 892 car garage to be located in upper Manhattan. When all factors are considered, the cost of operating an MAPS is less than half that of a conventional garage [6]”. Table 02. Sample Construction Cost Comparison [4] Unit Building Automated Efficiency Total Cost Type Cost/Sq. Cost per System Configuration (Sq. Ft./Stall) per Stall Ft. Stall Cost Stand-alone, above grade

Conventional

$50

$320

$16,000

$0

$16,000

Automated

$45

$225

$10,125

$16,000

$26,125

Cost Ratio

1.1

1.4

1.6

-

0.6 98

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Below building, above grade

Conventional

$75

$450

$33,750

$0

$33,750

Automated

$65

$225

$14,625

$16,000

$30,625

Cost Ratio

1.2

2.0

2.3

-

1.1

Below building, below grade

Conventional

$105

$450

$47,250

$0

$47,250

Automated

$85

$225

$19,125

$16,000

$35,125

Cost Ratio

1.2

2.0

2.5

-

1.3

Table 03. Sample Comparison of Expenses [4, 6] Expense Item Payroll and Benefits Insurance Expense

Conventional

Automated

Expense Ratio

(1)

(2)

(1)/(2)

$850,000

$145,000

5.9

$95,000

$50,000

1.9

Utilities Expense

$165,000

$200,000

0.8

Repairs/Maintenance

$145,000

$50,000

2.9

Bank Fee Expense

$100,000

$100,000

1.0

Marketing Expense

$20,000

$20,000

1.0

Support Service Expense

$75,000

$35,000

2.1

Other Operating Expense

$150,000

$75,000

2.0

Real Estate Tax Expense

$150,000

$150,000

1.0

Security Camera System

$30,000

$30,000

1.0

$1,780,000

$855,000

2.1

Total Expenses

Expense per parking space $1,996 $959 2.1 Turkiye’s first MAPS was put into service in stanbul in 2001. The last 15 years has been carried out in Turkiye all multistorey fully automated car park are listed below. Table 04. All Multi-storey Fully Automated Car Park in Turkiye Name Milta Milli Reasürans T.A. . Parkpoint Maçka Suits Residance zmir Metropolitan Municipality Total

Vehicle Capacity Area (m2) Year Location 612 800 2001 li, stanbul 276 320 2009 li, stanbul 76 500 2012 Be ikta , stanbul 280 616 2014 Konak, zmir 1,244 2,236

BACKGROUND OF THE PARKING SPACE PRODUCTION IN TURKIYE Solution of car parking problems are considered with two basic perspectives: The first one is controlling the demand for parking and the second one is managing and meeting the already existing demand. In the first approach, measures against the problem are taken before the problem actually emerges. The second approach, involving creating additional capacity by building car parking facilities, is a solution method to meet the existing demand. In the recent and modern car park planning approaches, this is avoided at the beginning of the solution process. On the contrary, solving the problem by management and operational practices is being attempted [7]. If any additional capacity is strictly required to be created, it is strongly recommended to do this cleverly. As a matter of fact, number of vehicles in the roads tend to increase, due to the nature of highway traffic. If this increase is not met with an increase in parking capacity, a vicious cycle, which turns roads into parking places, is created. Meeting the car parking demand in an uncontrolled way does not mean that the problem is solved, either. There are two reasons for not having sufficient parking places to meet the demand. The first is, failure to implement the rule “Every building has to contain parking place it needs”, which is present in the “Car parking legislation”, for several reasons. This can also be defined as “New buildings being unable to meet the parking demand that they have generated.” The second is the reluctance of private investors to invest in parking lots. The most important reason for investors’ reluctance is the reluctance of the drivers to use the commercial lots, since they can park their cars on the roadsides and sidewalks free of 99

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

charge and without facing any problems. Lack of precautions against this undisciplined parking behaviour continues to affect the process in a negative manner [8]. In the car parking legislation, the following subjects are mentioned: Liability of the municipalities to provide public parking lots if there is not one within or in the parcel of the building (Article 9), requiring parking fees for that (Article 7), depositing the parking fees into accounts in public banks (Article 10), compulsory allocation of the collected fee to land acquisition for the parking lots and the construction of them (Article 11). Also, it is required by legislation that, if necessary, municipalities should contribute to the process using their own resources and take immediate action to complete the construction of already – started lots and acquiring lands for the new ones (Temporary article 4).[9]. Despite these articles in the legislation, parking problems are still getting worse. There were a number of applications to Turkish Parliament’s Petition Commission, requesting the fates of the parking fees collected by the municipalities to be investigated and necessary measures to be taken for an adequate parking service to be provided. These applications were evaluated and a report was prepared by the commission (TBMM decision no 8, 05.03.2012) [10]. The main issues subject to complaints are the allocation of the collected parking fees to non – parking – related areas by the municipalites, failure of the local authorities to provide an adequate parking service, negligance of the municipalities to conform to parking legistation, which in turn leads to traffic and environmental problems. Also, it was requested that, municipalities should be obliged to obey the legislation and thus consumers’ damnification level should be reduced [10]. In 64 provinces (79%), the money is in used in some way, whereas in the remaining 17 (%21), it was either not used or not collected, or even no account was set up (Table 5). In 59.3% of the provinces, it was directly used for parking lot construction. As of 2012, the largest deposit (TL 44,2 M) exists in zmir’s accounts. It is followed by Eski ehir (TL 23,7 M) and Manisa (TL 14,9 M) (Figure 1). According to the information presented to the commision, assuming 25 m2 per parking place, the largest generation of parking places was achieved in stanbul. It is followed by Konya and Kahramanmara (Figure 2). Table 05. Usage of Parking Accounts in Turkiye [10] What Used For Quantity Ratio Used in the construction of parking facilities 48 59.3% Parking facilities and other municipal services 2 2.5% Used for other municipal services 9 11.1% Used in the payment of municipal 4 4.9% Used in the expropriations 1 1.2% Any unused amount within the account 7 8.6% No parking fees collected 3 3.7% No parking account opened 3 3.7% There is no information 4 4.9% Total 81 100.0% Used somehow 64 79.0% Others 17 21.0% Parking Space 12000 10000 8000 6000 4000 2000

Figure 01. Amount within the parking accounts

Çanakkale

zmir

Sakarya

Kayseri

Bursa

Eski ehir

Manisa

K.Mara

Konya

stanbul

Kocaeli

Bilecik

Van

Bursa

Konya

Samsun

Malatya

0

Manisa

Eski ehir

zmir

(Million TL) 50 45 40 35 30 25 20 15 10 5 0

Figure 02. Approximately generated parking capacity

AS A MODEL APPLICATION IZMIR ALSANCAK MACP Since the parking legislation is not correctly implemented, parking space generation is not a desired level in Turkiye. However, for the first time in the last years, as a good example approach to other local administrations, zmir Metropolitan Municipality allocated a portion of the collected money to construct a MACP, with a capacity of 280 vehicles. 100

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Importance of zmir Alsancak MACP Project This project is expected to create a widespread effect, due to its way of using a constrained area, its technology and its style of meeting the parking demand of the region. All the stages of the project, starting from the acquisition of land, act as a model application. It is also expected to pioneer the effective usage of the money collected in car parking accounts of, expecially in the metropolitans. In this model application, technology is entirely native. So, there were great opportunities to reduce the invesment costs. It is thought that, if this application is correctly advertised, the inertia and undisciplined behaviours related to parking fees can be overcome. The project is applied onto a 938 m2 land, which is located on the air E ref Avenue and used to be a petrol station (Figure 3). The region within which it is located carries residence, trade and fun functions together. In the surrounding road network, congestion is observed due to high level of traffic. Due to these facts, Alsancak is one of the regions with highest parking demand in zmir. In attempt to solve the problem, previously, two underground parking lots, Alsancak Atatürk Sports Hall and Kültürpark, had been taken into service. Izmir Metropolitan Municipality acquired the land of the petrol station in November 2009, with a 4.75 M price. In March 2013, a turnkey tender was concluded and the company, who uses native technology and did all of the MACP in Turkiye, won the tender with TL 13.25 M price. After the contract and project stage (Figure 4), construction works were initiated, aiming to finish the work within 12 months. As of March 2014, the end is approached. Test procedures are being carried out.

Figure 03. The location of MACP

Figure 04. Perspective of the design model

Technical Specifications of zmir Alsancak MACP Project The system is designed entirely by Turkish engineers and uses %100 native technology. All hardwares, softwares and construction works are natively facilitated. The land has 616 m2 of available land (Figure 4). If conventional methods had been used, 150 vehicles could have parked there (100 m3/space). However, under the same conditions, MACP can provide a capacity of 280 cars (54 m3/space) (Table 6). The underlying soil has been strenghtened by jet-grout method. Radial base has been constructed using C25-30 concrete and designed to carry 350 tonnes of steel construction on top of it. Unit vehicle mass is assumed to be 2,500 kg. Ground depth is 2.0 m and building height is 24.5 m (Table 6). The building has 11 storeys. 7 of them are allocated for vehicles with height up to 1.6 m, whereas the remaining 4 are allocated for vehicles with height up to 2.0 m (Figure 5). Table 06. Design Features of zmir Alsancak MACP. Items Dimensions Land area

938 m2

Residential area

616 m2

Building height

24.50 m

Ground depth

2.00m

Construction volume

15,092 m3

Volume per parking space

54 m3

Total parking spaces

280

Car dimensions

5.25 m x 1.90 m x 2.00 m

Car weight

2,500 kg

101

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Figure 04. Normal floor plan

Figure 05. MACP profile model

This system provides vehicle-carrying steel park palets to be transferred onto the conveyors by means of elevators, using dynamical position controlled transfer method. Vertical lift takes the vehicles from a transfer cabin and carries them onto the conveyor line on the storey, in which the vehicle will be parked. In the entrance hall, there are the lobby, reception cabins and necessary technical rooms (Figure 6). 4 vertical lifts can be accessed from here. Vehicles are delivered in ready to leave position here. Main body of the structure consists of HE200A – HE300A type structural steel profiles (Figure 7).

Figure 06. Ground floor plan, cabins and turned tables

Figure 07. Construction of steel main framework

These profiles come from the factory ready for bolted connection, and connected directly in the place. Building covered with aluminum panels and special laminated and reflective glasses (Figure 8). An integrated management system, which controls the system components and related electronic equipment, is developed. Communication between the units is provided with advanced industrial hardwares and softwares.

Figure 08 Covered MACP building DISCUSSION The fully MAPS implemented in zmir is a good example for the other provinces, for its technology, nativeness and its financial model. On the other hand, potential for implementation of MAPS in Turkiye is directly related to private investors’ interest on the subject. However, private investors are still reluctant, due to the lack of discipline in parking. The reason for this is the perception of the actors (local administrations, drivers, residents and contractors) related to parking problems [7]. To solve the problem, urban transportation strategies for taking the parking behaviour should be developed, and also effective control should be enforced. All new buildings should be constructed with parking places to meet their own demand. Parking fees should be set to realistic values to bear all the direct and indirect costs. Last but not least, favorable conditions should be set up for the investors, who consider to invest into construction of parking lots [8]. MAPS, everyday 102

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

getting more and more widespread in the world, constitute a good technological potential for solving the parking problems in central regions of the cities and compact urban campus-type areas in Turkiye. All Technical potential for MAPS – related technologies are present in our country. This is the result of a 15 – year experiences, which provides grounds not for a new import items, but for development of an automated parking industry. Acknowledgments We thank YTU Scientific Research Projects Coordination Unit for their support to the Project with ID 2011-05-01-KAP02.. REFERENCES [1] Beebe, R. S. (2001). “Automated Parking: Status in the United States-Advantages and Criteria”, World Parking Symposium III, St. Andrews, Scotland. http://www.worldparkingsymposium. ca/parkinglibrary/download/137/00000137_ d010010wx.pdf. Access date: 03 January 2014. [2] Abboud, N. W. (1994). “Automation of the Parking Industry: A Strategic and Manegarial Overview” M.Sc. Thesis in Civil an Environmental Engineering at the MIT, USA. [3] FATA Automation Inc. (2013). “History”, http://automatedparking.com/history/Access date: 12.27.2013. [4] Monahan, D. (2012). “Man vs Machine-is Robotic Parking Right For Your Project?”, The Parking Professional Magazine, International Parking Institute, September 2012, pp. 34-37. [5] Guinnesss World Records. (2013). “Largest Automated Parking Facility”, http://www.guinnessworldrecords.com/world-records/11000/largest-automated-parking-facility, Access date: 12.27.2013. [6] Schwartz, S. I. (2009). “The Garage of the Future Must be Green” NPA Parking Magazine, National Parking Association, March 2009, pp. 32-36. [7] Yard m, M. S. (2013). “Türkiye’de Otopark Alg Üzerine Bir De erlendirme/ An Evaluation on The Perception of Parking in Turkey”, 3. Uluslararas Ula m ve Araç Park Alanlar Yönetimi Sempozyumu/3. International Transportation and Parking Areas Management Symposium, s. 143-153, 30-31 May 2013, stanbul. [8] Yard m, M. S. ve A rikli, M. (2005). “Otomatik Otoparklar ve Türkiye'deki Otopark Probleminin Çözümü çin Uygulama Potansiyeli", 6. Ula rma Kongresi, MO stanbul ubesi, stanbul, Bildiriler Kitab , s. 363-371. [9] Otopark Yönetmeli i/ Turkish Parking Legislation. (1993). Resmi Gazete Tarihi: 01.07.1993, Say : 21624, http://www.mevzuat.gov.tr/Metin.Aspx?MevzuatKod=7.5.4886&sourceXml Search=&MevzuatIliski=0, Access date: 12.27.2013. [10] TBMM. (2012). “Belediyelerce Toplanan Otopark Bedellerinin Ak betinin Ara lmas ve Söz Konusu Bedellerin Amaç D Kullan n Önlenmesi Hakk nda Rapor”, http://www.tbmm.gov.tr/komisyon /dilekce/belge/kararlar/d24/gkcetvel9.pdf, Access date: 12.27.2013.

103

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

EXAMINING THE USE OF CEBEC DOLOMITE LIMESTONE, GÖLCÜK SANDSTONE AND KARATEPE BASALT IN STONE MASTIC ASPHALT MIXTURES

M. S. Yard m1, F. Arslan2

[email protected], [email protected]

1

2

ld z Technical University Civil Engineering Faculty Esenler, stanbul, Turkiye Tel: (0212) 383 51 83 Republic of Turkey General Directorate of Highways, 1st Regional Directorate Kagithane, stanbul, Turkiye Tel: (0212) 312 17 00

ABSTRACT The paper compares the performance of the Stone Mastic Asphalt (SMA) mixes with a focus on aggregates. Three types of aggregates, namely Cebeci dolomite limestone, Gölcük sandstone and Karatepe basalt, were considered near stanbul. Basalt stone is difficult and high cost to quarrying and its reserves are limited around stanbul. Usability of alternative types of aggregates in SMA mixes is considerable subject to be investigated. The purpose of the study was to collect data on mix design, quality control procedures (scope Marshall Test, core sampling) and evaluation performance (including rutting, indirect tensile strength) of three SMA pavements that had been constructed in stanbul. In addition, the current situation of these sections has been observed and is presented in this study. In this study, primarily some important material properties of coarse aggregate were determined. The results showed that basalt stone except for the flatness index and stripping strength provides higher values than Cebeci dolomite limestone and Gölcük sandstone. However, in terms of resistance to permanent deformation, Marshall Test results of designed and manufactured SMA features indicate that, dolomite limestone SMA mix and sandstone SMA mix an alternative mixes are an alternative to basalt SMA mix. Keywords: Aggregate, basalt, dolomite limestone, rutting, sandstone, stone mastic asphalt. INTRODUCTION SMA is an asphalt mixture containing a gap-graded aggregate mixture with high contents of coarse aggregate fractions, filler and binder. Very frequently, a stabilizing additive (drainage inhibitor), which prevents the drain down of the binder from the aggregate, is needed mastic is the second largest component of SMA; it is approximately 20-25% by weight of the mixture and 30-35% by volume. About 35-40% (v/v, refer to ratios by volume) of the compacted coarse aggregates is made up of voids, and after filling the aggregate with mastic, 3% to 5% (v/v) of empty space will be left. Mastic consists of fine aggregate (passive aggregate), filler, stabilizer and bituminous binder as well as modifier contributions [1], [2], [3]. SMA is considered a premium paving material and is expected to have a service life 20-30% longer than that of the conventional dense-graded hot-mix asphalt. This longer service life is achieved by increased durability and resistance to permanent deformation. The latter is due to stone-on-stone contact of the coarse aggregates [4], [5]. In Turkey, the SMA is used on motorways that have the heavy volumes of traffic, some urban intersection areas, bus rapid transit ( stanbul Metrobus) corridor as a wear layer. From 2008 until 2012, the quantity of SMA, which was produced under the responsibility of General Directorate of Highways (KGM), is 3,598,000 tons out of a total of 94,211,000 tons Hot Mix Asphalt (HMA). At this point, it should be noted that SMA is mostly applied to a thickness of 4 cm in Turkey [2], [3]. Many research reports and engineering case studies [6], [7] have shown that the use of SMA on road surfaces can yield better rut-resistance and durability. Since the strength of SMA relies heavily on the stone-on-stone aggregate skeleton, it is imperative that the mixture is designed and placed with a strong coarse aggregate skeleton. Due to physical properties, in applications worldwide, magmatic rocks are generally preferred to use in SMA. In Turkey, the KGM provided that relevant specifications, the use of alternative types of aggregates in their applications, subject to the permission of the administration [5], [8]. Within the last five years, the use of SMA in the Turkey has continued to grow. However, no testing has been performed on a routine basis during design and/or production stages to ensure that SMA mixtures have an adequate 104

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

coarse aggregate skeleton. The work reported in this paper, the coarse aggregates named basalt, dolomite limestone and sandstone was examined that effects of the SMA mixture. MATERIALS AND METHODS In this study, sections under the responsibility of the KGM, three diversified SMA paving mixes projects were investigated with a focus on aggregate. Turkey’s procedure is as follows: SMA materials that will be used are tested in the laboratory. In order to determine mix formulation, Marshall Mix design is performed with constituent materials with established suitability. A target composition may be expressed in two ways. Input target composition (expression of a mix formulation in terms of the constituent materials, the grading curve and the percentage of bitumen added to the mixture) will usually be the result of a laboratory mix design and validation. Output target composition (expression of a mix formulation in terms of the constituent materials and the mid point grading and soluble binder content to be found on analysis) will usually be the result of a production validation [9]. In this investigation, analyzed mixtures are referred to in short as: DL Mix (SMA with dolomite limestone); S Mix (SMA with sandstone); B Mix (SMA with basalt stone). Within this scope, in the first place, modified binders and aggregates experiments were tested by KGM and Marshall Laboratory mix designs were compared with each other. In the second place, with regard to quality control tests of output mix were done by subcontractor. In the third place, physical and performance tests were done in order to measure the performance pavements and check the suitability specification were evaluated and the advantages of mixtures were hold down. Aggregate The aggregates (dolomite limestone, sandstone and basalt) used in this study were respectively provided by Quarry Cebeci-Su in stanbul, Quarry Gölcük as Elmaslar in zmit and Quarry Karatepe in gorlu. Physical properties for the aggregates are shown in Table 1. Table 01. Some Important Properties of Aggregates Values Dolomite KTS Properties Sandstone Basalt limestone Limits Los Angeles Abrasion Loss (%) 20.60 12.30 9.8 Max. %25 Soundness of aggregate by Na2SO4 Flatness Index (%)

1.00 14.10

Polished Stone Value (PSV)

62.00

Stripping Strength (%)

1.00 10.50 -

21.5 -

Max. %8 Max. %25 Min. 50

50-60

60-70

3

50-60

Min. %60

Apparent Specific Gravity (g/cm )

2.745

2.687

2.936

-

Volume Specific Gravity (g/cm3)

2.705

2.657

2.841

-

Water Absorption (%)

0.55

0.43

1.1

Max. %2

Bitumen The designed bitumen content in the SMA mixtures was adjusted from 6 to 7 % by mass of total mix with 0.35% of fiber content by mass of total mix. Tüpra zmit (PEN 50-70) bitumen was modified by 4.5% Styrene-ButadieneStyrene polymer (SBS) (Table 2). The Polymer Modified Bitumen (PMB) for the SMA mixes was tested by the Dynamic Shear Rheometer (DSR) and evaluated as PG70-16 (DL Mix), PG76-16 (S Mix) and PG70-16 (B Mix). Table 02. Basic Properties of Binders Polymer Modified Bitumen (PMB) DL Mix S Mix Penetration grade Penetration, (0.1 mm), at 25 °C, 100 g

B Mix

Standards

PMB 70-16 38.8

PMB 76-16 37.00

PMB 70-16 45

TS EN 1426

Softening point (°C)

62.10

73.70

68

TS EN 1427

Ductility (cm), at 25 °C, 5 cm/min

100+

100+

100+

TS EN 13589

Flash point (Cleveland) (°C)

310

260

260+

TS EN ISO 2592

1.025

1.026

1.010

TS EN 15326

Specific gravity (gr/cm3)

105

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

RESULTS Stone Mastic Asphalt generally requires more binder content than those of conventional dense graded mixtures, and addition of fiber enhances holding capacity of binder within the aggregate skeletal [10]. The thickness of SMA layers was designed as 4 cm. Input target composition (Marshall mix design) (Table 3-4) (Figure 1) procedure was performed to prepare three mixes. These mix designs were given to the contractor. Output Target Composition was prepared by the contractor (Table 3-4) (Figure 1). Table 03. Input- Output Target Composition Ratios Sieve Size Input Target Output Target Composition Composition % (Design) % (Hot silo grading control) DL S Mix Mix

B Mix

DL Mix

S Mix

B Mix

19-12.5 mm (3/4 "-1/2 ")

5.1

4.7

7.3

8.2

7.6

6.7

12.5-4.75 mm (1/2 "-No: 4)

61.9

63

58.7

57.2

58.7

57.7

4.75-0.075 mm (No: 4-No: 200)

22.2 21.9 23.5

25

23.9

25.1

0.075-0 mm (No: 200-0)

10.8 10.4

10.5

9.6

9.8

10.5

Coarse Aggregate Ratio (%)

67

66

65.4

66.3

64.4

67.7

Fine Aggregate Ratio (%)

22.2 21.9 23.5

25

23.9

25.1

Filler (%)

10.8 10.4

10.5

9.6

9.8

10.5

Total Mix (%)

100

100

100

100

100

100

Figure 01. Aggregate gradations for design and output mixes

Properties

Optimum Bitumen Contents Bulk specific gravity, (Dp), gr/cm3 Marshall Stability (kg) Air Voids (Vh)% Voids Filled with Bitumen (Vf) % Voids in Mineral Agg. (VMA) % Flow, mm

Table 04. Marshall Average Results DL S B DL S Mix Mix Mix Mix Mix Input mix (Design) 6.5 6.5 6.6 2.402 2.358 2.494 894 839 800 3.01 3.21 3.33 82 81 80.5 16.7 16.8 17.06 3.04 3.04 3.9

Output mix 6.43 6.42 2.39 2.36 1.103 886 3.65 3.07 78.64 81.47 17.09 16.59 3.10 3.23

B Mix

6.62 2.49 884 3.43 80 17.21 3.67

KTS Limit Values

±0.3

(%) 2 - 4 (%) min. 16

106

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Wheel Tracking Test for measuring the resistance to permanent deformation (rutting) (Table 5). Also, Indirect Tensile Strength (ITS) for measuring the tensile strength of diametrical specimen in dry state, and also after warm water conditioning for obtaining the Tensile Strength Ratio (TSR). Results of wheel tracking tests with French Laboratory Rutting Tester (FLRT) are summarized in Table 5.

Rutting Tests

Extraction n

Table 05. Mix Extraction and Wheel Tracking Results Mixes Type DL Mix S Mix* B Mix Coarse Aggregate Ratio (%) 65.8 65.2 55.5 Fine Aggregate Ratio (%) 24 24.4 34.4 Filler (%) 10.2 10.4 10.1 Bitumen Contents 6.6 6.48 5.9 Optimum Bitumen 6.5 6.5 6.35** Dp design 2.402 2.358 2.395** Dp sample 2.319 2.354 2.36 % Compression 96.5 98.1 98.5 Load Cycle (N) Rut Depth (mm) % 1000 cycles 0.46 1.56 1.86 3000 cycles 1.04 2.35 2.25 5000 cycles 1.2 2.54 2.45 10000 cycles 1.32 2.94 2.73 30000 cycles 1.45 3.63 3.05 50000 cycles 1.66 4.01 3.29 *Output mix samples extraction average values ** Design include in this study was changed.

Typical comparative results of the tensile strength ratio test performed on three types gradation mixes are illustrated in Figure 2.

Figure 02. Tensile Strength Ratio Results Current Road Surface Condition In this part of study, based on observation, the examination of the current road surface which under traffic conditions was given. Observations, sometimes from inside the car, sometimes stopping vehicle on available places have been made (Figure 3-5). Rutting length of problem sections was determined with miles indicator. Therefore, total rutting lengths are approximate range.

Figure 03. TEM Motorway KM: (54+400)-(59+180) (Avc lar- nal direction), current road condition as of the date of 11.01.2013 (with DL Mix; 10 cm thick binder+SMA surface) 107

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Figure 04. Anatolian Motorway KM: (47+500)-(69+400) (Ankara- stanbul direction), current road condition as of the date of 11.06.2013 (with S Mix; 11 cm thick binder+SMA surface) There is a serious level of rutting, in some sections have used S Mix (Table 6). Table 06. Sections Have Seen Rutting Problem Gebze Intersection -Körfez Intersection Motorway (Ankara- stanbul Direction) Km range

Rut problem of condition

Rutting length (m)

Sections of the problem

69+400-69+000

low

400

Hamzadere Viaduct

48+500-47+900

high

600

After Dilovas Intersection

47+500-46+800

moderate

700

Dilovas Intersection

Figure 05. Hasdal Intersection approach KM: (0+660), Seyrantepe KM: (0+942)- (1+289), current road condition as of the date of 10.26.2013 (with B Mix; 11 cm thick binder+SMA surface) DISCUSSION • For all samples the abrasion loss of is under 25% that is the critical value (Table1). The exceeded ratio of flatness decreases the contact pressure. Aggregate used in the projects, has the highest rate of flatness is basalt (Table1). In general, the quarries, basalt are produced more flat than the other aggregates. The causes of these gaps need to be explored during production. It was seen some differences between the design and output features of SMA mixtures. With the experiments conducted in the laboratory becomes easy to achieving required properties and optimum results. However, the difference in the workarea conditions and laboratory conditions, constantly changing material input, affects the properties of the mixture of productions, therefore, learning and assessment of the properties of productions produced in the workplace becomes even more important. At output target composition, the same rate with the highest ratio of coarse aggregate in mixtures, DL mix and S mix. (Table 3). The best values of the air voids are DL mix and B mix (in output mix) (Table 4). Although S mix has higher coarse aggregate ratio than B mix, S mix has less VMA value, due to the fact contain a lower content of bitumen and air voids value. • Stability is the maximum resistance showed by asphalt briquettes to deformation. B Mix gives low stability values (Table 4). This value of stability should not be considered inadequate by evaluating Marshal Stability value. Likewise, this mixture yielded very good result in wheel tracking test. In pavement construction, very high stability and low flowing value is not acceptable; in this type of mixtures cracking occurs due to heavy traffic load. Also, the marshal method does not reflect exactly the mechanical properties of SMA. In some studies [11], SMA mixtures gives lower marshal stability values than conventional asphalt mixtures. It has founded that the marshal method does not reflect exactly the mechanical properties of SMA [11]. Recently, new methods such as superpave method are used in mix 108

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

design. The debates are ongoing about which of these methods are better and still there is not a clear consensus among experts. The studies with superpave method and gyratory compactor should be emphasized and the advantages to each other of these two methods should be revealed. The roads in Turkey are constructed according to Marshall Design method. • According to wheel tracking test results; each of the three SMA mixes is lower than technical specification limits (maximum 6% value) which is consisting of 30.000 periods. Prior to wheel tracking test, this mixture was extracted. Content of the mixture of bitumen was less than optimal except for DL Mix. Although coarse aggregate ratio was 55%, B Mix achieved performs well in wheel tracking test. This result (Table 5), indicates higher resistance to rutting. • As regard with the ITS values, B Mix have higher moisture damage resistance according to the other mixtures. Between dolomite DL Mix and S Mix were no significant differences (Figure 2). • Considering the road of current situation, on TEM Motorway (02) Avc lar-K nah sections (Annual average daily traffic; AADT value 33,089 vehicle/day; rutting was not observed (Figure 3). Although many of excavation trucks are used the road, rutting was not observed on Hasdal - Kemerburgaz - Yass ören State Road (AADT values 72,260 vehicle/day heavy vehicles) (Figure 5). Sandstone Mix was used on Anatolian Motorway (O4) Gebze-Körfez Section. There are AADT value 66,492; vehicle/day. In these sections no observed rutting expect for Dilovas intersection or ramp (Figure 4). Rutting length was observed on the heavy vehicles lane. As seen Table 6, approximately 1,700 m rutting of this section is still the responsibility of the contractor, because, rehabilitation work is still continued. Problematic sections will be excavated and manufactured with basalt SMA. In this context, to encourage contractors making higher quality production, to take responsibility ensuring many more years and to pave the way for controlling heavy vehicle load is needed. Furthermore, in practice, weight control system (the legal axle load of 10 tons) and punishment system is not effective. Particularly in climbing lane with a volume of heavy vehicles whether to use sandstone in the mix SMA, should be evaluated according to the results after the renewal sections and the monitoring. In SMA applications in Turkey, stone to stone contact isn’t being evaluated. The breakdown of coarse and fine aggregate could not be performed, the stone to stone contact cannot be determined exactly. Good stone to stone means: Voids in the Coarse Aggregate in the aggregate (VCAag) should be less than or equal to Voids in the Coarse Aggregate in mix (VCAmix) or [12]. Therefore the calculation of VCAag and Voids in the Coarse Aggregate in mix VCAmix should be included in SMA Mix calculations and the work needs to be undertaken in this regard. The above-mentioned materials were evaluated; Dolomite limestone SMA Mix and Sandstone SMA Mix an alternative mixes are an alternative to Basalt SMA Mix. REFERENCES [1] Blazejowski, K. (2011). Stone Matrix Asphalt: Theory and Practice. pp. 1-72, CRC Press, Boca Raton, FL. [2] Yard m, M. S. & Arslan, F. (2013). Türkiye’de Ta Mastik Asfalt Kaplama Kullamm ve Literatür Üzerine Bir De erlendirme. 6. Ulusal Asfalt Sempozyumu, 27-28 Kas m 2013, Ankara. [3] Arslan, F. (2014). Ta Mastik Asfalt Kansimlarda stanbul Çevresindeki Cebeci - Dolamitli Kireçtasi, Gölcük Kumtasi ve Karatepe -Bazalt Kullan n ncelenmesi. Yüksek Lisans Tezi, YTÜ Fen Bilimleri Enstitüsü, stanbul. [4] Celaya, B. J. & Haddock, J. E. (2006). Investigation of Coarse Aggregate Strength for Use in Stone Matrix Asphalt. FHWA/IN/JTRP-2006/04, Joint Transportation Research Program, Indiana Department of Transportation and Purdue University, West Lafayette, Indiana. [5] Brown, E. R. & Haddock, J. E. (1997). A Method to Ensure Stone-On-Stone Contact in Stone Matrix Asphalt Paving Mixtures. NCAT Report 97-02997, National Center for Asphalt Technology of Auburn University, Auburn, Alabama. [6] Chiu, CT. & Lu, LC. (2007). A Laboratory Study of Stone Matrix Asphalt Using Ground Tire Rubber. Construction and Build Materials, 21(5), pp. 1027-1033. [7] Brown, E. R. & Cooley, L. A. (1999). Designing Stone Matrix Asphalt Mixtures for Rut-Resistant Pavements. National Cooperative Highway Research Program, Report No: 425, Transportation Research Board, National Research Council, Washington, DC. [8] Republic of Turkey General Directorate of Highways (KGM) Turkish Highway Technical Specifications (KTS) (2006). Edition No. 276, pp. 531-536, Ankara. 2006. [9] EN 13108-5 (2006). Bituminous mixtures - Material Specifications - Part 5: Stone Mastic Asphalt, CEN/CENELEC, Brussels. [10] Hafeez, I., Kamal, M. A., Mirza, M. W. & Aziz, A. (2012). Investigating the Effects of Maximum Size of Aggregate on Rutting Potential of Stone Mastic Asphalt, Pak. J. Eng. & Appl. Sci., 10, pp. 89-96. [11] Ta demir, Y. (1998). A Study on Stone Mastic Asphalt Mixtures (Turkish)/ Stone Mastik Asfalt Kansimlann Etüdü, Unpublished Master’s Thesis, stanbul Technical University, Graduate School of Science, Engineering and Technology, stanbul. [12] NAP A, (1999) “Designing and Constructing SMA Mixtures- State-of-the-Practice”; Quality Improvement Series 122. 109

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

STUDY ON GROUND TREATMENT FOR SUBWAY SHALLOW TUNNEL CROSSING AN URBAN RIVER BY SHIELD TUNNELING METHOD

Z. Xiaojun, G. Bo, W. Jinghe, L. Jianguo [email protected]

Key Laboratory of Transportation Tunnel Engineering, Education Ministry Southwest Jiaotong University No.111 of north section 1 of the 2nd ring road Chengdu, 610031 China Tel:+86028-87634687Fax:+86028-87600612

ABSTRACT According to the geological and environmental conditions of surrounding soil around a shallow tunnel used to cross a river in Chengdu subway lines, the design and construction technique of ground treatment such as uplift piles, cast-inplace reinforced concrete slabs and grouting technique have been summarized in the paper. The riverbed above shallow subway tunnel is reinforced with covering slabs and uplift piles; meanwhile, the surrounding rock is reinforced with cement and sodium silicate grout by means of high pressure grouting which is conducted inside the tunnel. The practice of the scenario presented in the paper turns out to be successful and shows that the treatment technique can be used to guide the design and construction of ground reinforcement for shallow tunnel used to cross urban rivers in subway lines by shield tunneling method under similar situations. Keywords: Cement and sodium silicate grout, ground treatment, subway shallow tunnel, shield tunneling method, uplift pile, uplift slab INTRODUCTION The rapid development and construction of subway systems in Chinese cities has encountered many tough technical problems because of complicated ground geology and urban surroundings. According to the general planning and extension of subway networks, the main subway lines are usually set to go underground in densely populated area in central cities, but in suburb and sparsely populated area, they are usually set to go from underground to free surface or elevated on the ground with viaducts [1, 2, 3]. The longitudinal gradient of main lines also varies from place to place due to different ground buildings and topography. Underground subway lines may usually penetrate different strata along its extension in different cities. In order to diminish the influence of underground tunneling on normal traffic flow and ground subsidence, shield tunneling method is generally used to drive shallow running tunnels in subway lines in cities both in China and other countries in comparison with drill and blast method [4, 5]. The major advantage of using shield tunneling is that it can greatly decrease the disturbance of surrounding rock and induce little ground subsidence during its tunneling in strata. Nonetheless, if the ground depth is extremely thin, then the risk of shield tunneling also gets enhanced higher and special measures must be taken to eliminate or diminish the soil collapse and unexpected ground settlement. For instance, the entrance and exit line from depot in Chengdu subway line 3 is built by means of shield tunneling method. The double lines go underground from the depot. According to the comprehensive layout of the departure line from depot and the receiving line to depot in Chengdu subway line 3, the two transfer lines must cross a river name Hong Qi in Chengdu city. The river is 31m wide with a water depth from 1.5m to 2.5m; its flow velocity approaches 1.2m/s. The aerial view of the relationship among subway lines and the river is illustrated in Fig.1. Since the two subway lines belong to underground transfer ones, and their longitudinal gradients vary from 2‰ to 36‰. This leads to the different overburden depth of bored tunnel in each transfer line when they cross the river. So the most difficult problem encountered in the design and construction of the transfer line is to keep the stability of riverbed and the normal service of the river current during tunneling with earth pressure balance shield machine

110

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Figure 01. Aerial view of the relationship among subway lines and river Geological Conditions of the Surrounding Rock It has been proved from a large amount of engineering practice that geological condition of surrounding rock around tunnels mainly affects the design and construction scheme of subway tunnels and stations [1, 2, 3, 5, 7]. According to the geological reconnaissance of the subway line 3, the strata through which the two subway lines pass mainly consist of clay and silty gravel in the Quaternary system, and including completely weathered, strongly weathered and medium weathered mudstone in the Cretaceous period. The obtained physical and mechanical parameters of the surrounding

Table 01. Physical and mechanical parameters of surrounding rock

It is seen from the value of soil cohesion and friction shown in Table 1 that the shear strength of surrounding rock is medium and low; its self-stability is also very poor [4, 5, 6, 7]. Therefore, in order to enhance rock strength, reinforced concrete support such as lining and prefabricated segment to maintain its stability must be timely erected as quick as the surrounding rock has been excavated [8, 9]. According to the geology of subway line 3 in Chengdu city, the ground water falls into two categories, i.e., perched water and bedrock fissure water. The former mainly exist in the 111

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Quaternary clay and silty boulder, but the latter briefly occurs in the Cretaceous mudstone. After geological reconnaissance, the permeability coefficient of clay and silty gravel shows that they posess very low water permeability, but the one of mudstone proves that crevice water can flow and permeate freely in it. Since there is little ground water in the strata, then it does not apparently influence the underground construction of the tunnel. To sum up, the conventional method, namely, drill and blast method is not suitable for the weak rock, and then shield tunneling method can only be used to drive the tunnels in the two transfer lines [7, 8, 9]. The detailed illustration of both the surrounding rock and tunnel transection are still shown in Fig.2.

Figure 02. Transection A-A of shield tunnels and geological profile (Unit: m) Fig. 2 also illustrates the transection A-A as shown in Fig.1. Since the two tunnels must cross an urban river which is now under normal service, so the safety of both shield tunneling and river service becomes the key task to be solved during tunneling practice. It is seen from Fig.2 that the overburden depth T of the two shield tunnels varies from 1.6m to 5.3m. This depth explicitly exceeds the limit for shield tunneling, namely T<1D, D denotes the diameter of shield jumbo [1, 2, 5, 6]. Therefore, any mistake and ignorance in the design and construction of the shallow tunnel may lead to serious casualty. So, the ground treatment and driving parameters of the shield jumbo must be carefully anatomized in combination with the situation of the river and bored tunnels. Ground Treatment and Driving Parameters of Shield Machine In order to lessen the risk in shield tunneling of the two tunnels, ground treatment must be carried out so as to raise the shear strength of surrounding soil. According to the properties of surrounding rock, including the thickness of riverbed, the reinforcement of riverbed and surrounding rock are both carried out in terms of the calculation of tunneling procedure by Finite Element Method. Since the thickness of riverbed above the two tunnels is different, one varies from 1.95 to 5.8m, the other changes from 1.6 to 5.3m. Meanwhile, in view of the normal erosion induced by river current, the riverbed must be thickened with reinforced concrete slabs cast in place. In addition, the anti-buoyancy of the concrete slabs must be also taken into account. This aim can be realized by means of setting up uplift piles beside the tunnel. The detailed design of the uplift slabs and piles for thickening the riverbed is depicted in Fig.3.

Figure 03. Uplift piles and slabs around the two tunnels (Unit: m) 112

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

The stabilization procedure shown in Fig.3 is merely used to thicken the riverbed depth; its anti-buoyancy is attained with the action of bored piles set beside the two tunnels The length of the uplift piles approaches 15 meters, the spacing between two piles is kept within 5m in the tunnel’s longitudinal direction. The red dotted points in Fig.1 denote the planar layout of piles, and the riverbed protected with concrete slabs is also shown in Fig.1. In addition to the uplift piles, reinforcement of the surrounding rock is also carried out. According to site experimental study, the shallow roof of the surrounding rock around the two tunnels is prone to collapse during underground tunneling; therefore, the ground soil above the two tunnel crowns requires reinforcing either. In consideration of the porosity and void ratio of the Quaternary clay, grouting with high pressure is implemented with steel pipes. According to the availability of grouting and its efficiency, cement and sodium silicate grout are adopted and squeezed into the pores of the clay with high pressure. The grouting for reinforcing surrounding rock is conducted inside the tunnels separately. The detailed scheme of soil grouting is shown in Fig.4.

Figure 04. Grouting scheme in the surrounding rock above tunnel (Unit: m) Since the depth of riverbed only varies from 1.6m~5.3m, then the length of grouting scope in the clay varies from 1.5m to 3.5m. The soil treatment is only conducted within the scope of 120°above the tunnel crowns as shown in Fig.4. The grouting pipe is 42mm in diameter and with thickness of 3.5mm. The designed proportion of cement and sodium silicate grout are adopted for soil grouting are as follows. The ratio of cement to ash is 0.8:1~1:1, the proportion of cement grout over silicate grout varies within 8(10):1. The grouting volume is controlled within 0.5 to 1.0 m3. And the grouting pressure in the steel pipe is set to be within 0.3~0.35MPa. This grouting task is carried out with steel pipes inside the two transfer tunnels as shown in Fig.4. After the uplift piles, covering slabs and soil grouting have been completed and their strength has risen to the designed grade, then the excavation by shield machine begins. The driving parameters of shield machine may certainly affect the ground settlement and stability of the riverbed. In order to eliminate the influence of tunnel excavation on the surrounding rock, these parameters have been adjusted and determined according to experience obtained from actual practice of shield tunneling in the construction of subway line 1 and 3 in Chengdu city, they are summarized as follows. (1) The driving speed of shield machine is controlled within 2~2.5cm/min (2) The total propulsion of the shield machine is controlled within 8000 ~12000kN; (3) The revolution of cutter head is set to be within 1~1.3rpm; (4) The torsion of cutter head is controlled within 1500~3000kN·m; (5) The earth pressure in the soil chamber is controlled within 0.04~0.07MPa; (6) Synchronized grouting volume is set to be within 4.5~5.0m3, and the grouting pressure is set to be within 0.2~0.3MPa; (7) The earth discharge from the screw conveyor of the shield jumbo is controlled within 45~46m3. The practice of shield tunneling has been smoothly carried out to drive the cross river tunnels in the two subway lines according to the above stated scenario. During the whole tunneling process, the surrounding rock has not apparently been affected and its stability has been preserved. The successful construction of the two tunnels has verified and confirmed the feasibility and efficiency of the scenario presented in the paper. 113

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

CONCLUSIONS The treatment of surrounding rock and reinforcement for riverbed above shallow subway tunnel has been discussed, and the driving parameters of shield machine have also been summarized in the paper. The main conclusions are drawn as follows. (1) When subway tunnels cross urban rivers with extremely shallow depth in the main line, the riverbed usually requires reinforcing in order to keep its safe service. The most effective way is to thicken the depth of riverbed with reinforced concrete slabs cast in place. The thickness of concrete uplift slabs can still be mainly set within 0.7~0.8m. But the thickness of covering slabs on the riverbed can be controlled within 0.3~0.5m. (2) Bored piles can be used to resist the buoyancy of concrete slabs in the river when its current flow becomes full. The dead end of uplift piles in the deep soil must be extended into medium weathered mudstone; its embedded length should be kept at least 2.0m. (3) Grouting with cement and sodium silicate grout is an effective way to ameliorate the soil mechanical properties. This technique can well be carried out in tunnels during tunneling with shield machine. The driving parameters of shield machine will also affect the stability of surrounding rock and can be adjusted concurrently according to the geological conditions of surrounding rock around shallow subway tunnels. (4) The parameters of driving shield machine in weak soil presented in the paper can be referenced while designing and building the cross river shield tunnel in subway lines. ACKNOWLEDGEMENTS The author greatly wishes to express sincere thanks to the National Natural Science Foundation of China (No.51378436) and the Fundamental Research Funds for the Central Universities (SWJTU11ZT33) for their financial support. References 1. Zhou Xiaojun, Zhou Jiamei. Urban Metro and Light Rail Transit (in Chinese). Chengdu: Southwest Jiaotong University Press, 2012. 2. National Standard of China. Code for Design of Metro (GB50157-2003) (in Chinese). Beijing: China planning press, 2003 3. Zhou Xiaojun, Hu Hongyun, Wang Xiaofeng. Design of a Subway Station Crossing Urban Trunk Road by Open Cut and Tunneling method. The 4th international Conference on Digital Manufacturing & Automation, 2013, IEEE computer society, pp:457-460. 4. J. C. Jaeger, N. G. W. Cook., R.W. Zimmerman. Fundamentals of Rock Mechanics, Fourth Edition. Blackwell Publishing Ltd, USA, 2007 5. Xia Mingyao, Zeng Jinlun. Manual to design of undergound works (in Chinese). Beijing: China building industry press, 2001 6. Zhou Xiaojun, Wang Jinghe, Hu Hongyun. Design of Enclosure for Subway Station Foundation Pit in Southward Extension Line of Chengdu Metro Line One. Advanced Materials Research 2013,Vols: 671-674, pp:1122-1125. 7. Zhou Xiaojun, Yu Heran, Hu Hongyun, Wang Xaiofeng, Cao Yaodong. Design and Construction of Shaft for Xijiang Shield Tunnel in the Pipeline Project of Guangdong Provincial Natural Gas Grid (in Chinese). Chinese Journal of Underground Space and Engineering. 2013, Vol.9,No.4, pp:0884-0895. 8. Zhou Xiaojun, Wang Xiaofeng. Design of shallow depth tunnel with asymmetric twin arches and its construction procedure in poor rock mass. Applied Mechanics and Materials, 2013, Vols:353-354, pp:1321-1324. 9. Zhou Xiaojun, Yang Changyu. Study on structural design and driving process of a subway station with large transection by tunneling method. Applied Mechanics and Materials, 2013, Vols: 353-354, pp:1325-1328.

114

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

ASSESSMENT BALLAST LAYER THICKNESS FOR DIFFERENT TRACKS M. Keymanesh1, M.Khordehbinan2, S.Rezaei2, E. Ghasemi2 [email protected], [email protected], [email protected] 1

2

Assistant of professor, Payamenoor University of North Tehran, Tehran, Iran. Ph.D. Candidate of civil engineering, Payamenoor University of North Tehran, Tehran, Iran

ABSTRACT Appropriate thickness of ballast layer can avoid track geometric disturbance in vertical direction by providing a firm, steady and even load bearing area for sleepers and transferring load imposed by track in a compressive stress tolerable for bed. It also can limit unallowable settlements. Researches indicate that ballast layer thickness role in increase of track axial load and passing velocity has been not adequately analyzed. Thus in this study, criteria of pavement design and railway track construction are studied, ballast behavior is analyzed and in order to increase axial load and passing velocity in current railway tracks, a practical model is presented regarding ballast layer thickness analysis based on criteria defined in international valid codes. In this study, finite element method is used for analysis of ballast layer thickness in fixed track condition and increase of track axial load and passing velocity. To achieve this goal, usual condition of track system is modeled by Kentrack software, then model is loaded under different loading conditions and ballast layer thickness change is analyzed with the aim of track condition maintenance. Finally, obtained results are analyzed and different tables of appropriate thickness for ballast layer are provided in order to maintain track system quality and increase track axial load and passing velocity of train. Keywords: Ballast, Axial load, Passing Speed, Track

1 Introduction Ballast layer should have the ability of decreasing stresses imposed on bed layer so that it doesn’t exceed allowed tolerable range. Thus different factors can have impact on ballast layer structural role which are mainly related to technical and general characteristics of used materials (Sadeghi, 2009). Studies indicate that a limited number of railway codes have explicit criteria related to minimum allowed thickness value for ballast layer or a method for its determination. Among these, AREMA Code and Code 719 of International Railway Union and can be referred. Minimum value required for ballast layer thickness is given in AREMA Code and some calculation relations are also presented based on limiting pressure imposed on bed layer for appropriate thickness of ballast layer. This code states explicitly minimum thickness of ballast layer in main track should not be less than 300mm (Khordehbinan, 2009). The other standard which has good criteria related to ballast layer thickness is Code 719 of International Union of Railway (UIC). Criteria of this standard are related to determining minimum thickness required for ballasted layer. Differences in various railway codes and references are mainly due to two factors: different ideas about allowed amount of normal stress imposed on bed layer resulting in differences in required minimum thickness for this layer, and different ideas about integrality or non integrality of ballast and sub ballast layers’ structural performance and thus defining common or separate thickness for each of these layers. The aim for structural analysis of ballast layer is to determine loading pattern imposed on it as result of sleeper-ballast interaction and then, calculate internal stresses generated in this layer. After determination of internal stress levels in ballast layer, design stage of the layer is initiated. Generally ballast layer design can be studied in two areas: structural design and geometric design. The point which is considered in structural design of ballast layer is avoidance of stresses available in ballast layer exceeding from values allowed by railway codes which have been proposed based on different and many researches. Achieving this goal requires providing suitable clod-crushing strength by material grains in ballast layer and appropriate thickness selection for these layers in order to reduce stresses imposed on bed layer. Design process of ballast layer in different tracks is as figure1 algorithm. Most often this fact that most of track problems can be traced to subgrade rather than in track upper structure is ignored in railway networks. Modern policies of railway networks (higher velocities, more axial loads) lead more stresses in bed. Railway supporting system constitutes a compressed and massive area in paths which should be subject to change as less as possible. In addition, any change in bed should be restricted to areas in which there are special problems. Decision making between improvement and development of bed and ballast layer thickness increase should be considered technically and economically. Many factors can be regarded as effective in ballast layer structural role as 115

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

elasticity provider of track which is mainly related to technical and general characteristics of used materials. There are different factors in ballast thickness determination such as type and spacing of sleepers, ballast type, rail dimensions, bed type, axial load, Bougie axes arrangement and passing velocity, thus regarding variability of above mentioned factors and their interactions as well as dynamic nature of forces makes determining ballast thickness a difficult task and requires extensive study (Selig, 1994).

Initial selection of ballast materials

Selection of material gradation Determination of abrasion resistance and material hardness Determination of material resistance against environmental factors Modification or substitution of material

Geometric design of ballast layer

Control and design of ballast material mechanical characteristics

Ballast layer thickness

Width of ballast layer shoulders

Slope of side pitched roof Stress analysis of ballast layer

(Calculation of maximum normal stress in terms of layer depth) Design criteria control

Utilization criteria:

Structural criteria (allowed stresses)

1. Avoiding ballast dirtiness 2. Limiting ballast layer settlement

No

Do materials have required conditions?

Yes

Presentation of finial ballast layer design

have design criteria been met?

n

Yes

Figure 01. Ballast layer design algorithm (Sadeghi, 2009). Above facts necessitate ballast layer thickness analysis including all mentioned parameters so that track failures such as progressive gradual shear failure of bed, high settlement of bed due accumulative plastic strain. Researches show that although stress distribution by ballast can be calculated via linear analysis, Boussinesq, Eisenman methods or any other accurate method or there are various theoretical, half experimental and experimental methods for determining required ballast layer thickness, there is no specific method in railway for ballast layer thickness determination based on increase of velocity and axial load having general acceptance, and recommendations proposed in initial railway designs are yet used. Sometimes in some track ballast is more than needed which is uneconomical and in many tracks, thickness is less than needed which this leads to track settlement, rail and conveyor gear depreciation and their useful life reduction as well as increase restoration and maintenance costs. In researches performed so far just minimum and maximum needed ballast layer thickness have been mentioned (Sadeghi, 2009). In this study finite element method is used as a method having least result deviation with regard to real values. Track system is analyzed using KENTRACK software. Model is analyzed with diffract thickness of ballast layer based on variable parameters such as load characteristics and passing velocity, then in order to better understanding of obtained results for determination of optimal ballast thickness as a function of velocity and axial load increase results are given in some tables and diagrams. Different relations have been proposed for calculating dynamic impact factor by different institutes and individuals based on above mentioned 116

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

factors. Regarding dominant condition of pavement (including train speed and axial load) load factor of each axis is calculated independently for different loads by AREMA method.

Figure 02. Track models & Finite Element Method. 2 Numerical Analysis of Track System KenTrack software has been developed for determining strains, stresses and deflections in bed as pavement design principles. It originally was used for pavement design of highways. Regarding the fact that pavement design in railway and highway is similar, this software with some modification in loading and layers characteristics was used for railway pavement design. KenTrack software was developed in Kentucky University in 1980 in order to railway bed analysis and follows DOS operating system. This software is based on two theories: finite element and multilayer systems. Strains and stresses are calculated by using finite element method and multilayer system (which accomplishes analysis of all railway bed types). Structural system of ballasted railway pavement is consisted of two parallel rails and a series of sleepers placed on a multilayer system. Connection’s elastic behavior between rail and sleeper is modeled by a linear spring. Sleepers can have identical section in their lengths or may have different sections; in latter case sleeper can be regarded as yield of some different section. Therefore rail effect can be expressed by springs. Bed layer behavior is assumed as linear. Because of track fit in longitudinal direction and track section, only one fourth of track is analyzed. Figure 3 indicates model of this track(Jerry, 2006).

Figure 02. Track model in KenTrack software. Rail types used in this study is UIC60 which is common rail section in Iran and characteristics of which are given in Table 1. It should be mentioned that elasticity modulus of steel rail is 2.03× 105 MPa and its Poisson’s ratio is 0.3. Sleepers spacing is considered 60cm in calculations.

Mass & section of Rail (kg/m) 60.34

Table 01. Section properties of UIC60 rail. Area Section Total Width of Foot Width of (mm2) Height (mm) (mm) Head (mm) 7686

172

150

72

Horizontal Moment of Inertia (cm4) 3055

Ballast is consisted of coarse-grained materials, which has non linear behavior in loading and unloading and its behavior can be assumed either linear or nonlinear in analysis. Ballast elasticity modulus is a function of its loading, 117

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

stress and rest pressure coefficient in which K1 and K2 parameters are indicative of materials’ non linear behavior. In performed analyses ballast and sub ballast layers’ behavior are assumed as non linear and K1 and K2 coefficients are considered as 600 and 0.5 kg/cm2 respectively. Poisson’s ratio and lateral stress factor of ballast and sub ballast layer are considered as 0.35 and 0.8 respectively. Ballast layer thickness for Iran railway tracks should be determined in 3050cm based on code. Sub ballast layer thickness is assumed as 15 cm and fixed regarding Iran tracks. Track bed in this study has always linear elastic materials. In this bed, lower initial layers are generally rocky and uncompressible and its Poisson’s ratio and lateral stress factor is 0.4 and 0.5 respectively. Bed quality is medium and its elasticity modulus is assumed as 500 kg/cm2(Jerry, 2006). Results of analysis for different passenger and cargo tracks are given in figure 4. Regarding ballast layer thickness range, minimum thickness is proposed as 30cm and maximum thickness is proposed as 50cm. if ballast thickness higher than 50cm is needed, sub ballast layer thickness must be increased or bed quality must be improved (Khordehbinan, 2009). 60 Speed 60 km/h 55

Speed 80 km/h Speed 100 km/h

Depth of Ballast (cm)

50

Speed 120 km/h 45

40

35

30

25 20

21

22

23

24

25

26

27

28

29

30

Axle Load (ton)

Figure 04. Numerical Analysis (Freight Tracks). 60 Speed 100 km/h

55 Speed 120 km/h

D epth o f B allast (cm )

50 Speed 140 km/h

45 40

Speed 160 km/h

35

Speed 180 km/h

30

Speed 200 km/h

25 Speed 220 km/h

20 Speed 240 km/h

15 12

13

14

15

16

17

18

19

20

21

22

23

Axle Load (ton)

Figure 05. Numerical Analysis (Passenger Tracks). In fixed track system condition, percentage of ballast layer thickness increase required for increasing passing velocity in different track are presented as average in Table 2.

118

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Table 02. Passing velocity of Train increase in terms of ballast layer thick increase Passing velocity increase level (km/h) 20 40 60 80 100 120 140

Average percent of required increase of ballast layer thickness (%) Passenger tracks Freight tracks 7 8 14 17 22 28 30 41 56 73 -

Add Percent Depth of Ballast (%)

Results indicate that by steady increase of passing velocity, stress on bed surface in higher axial loads is increase nonlinearly and thus increase percent of ballast layer thickness for bed condition maintenance is increased. Figure 6 diagram indicates this.

80

Passenger Track

70

Freight Track

60 50 40 30 20 10 0 20

30

40

50

60

70

80

90 100 110 120 130 140

Add Speed Train (km/h)

Figure 06. Passing velocity of Train increase. Thickness percent in fixed track system condition and increase percent of ballast layer thickness required for increasing axial load in different track are given as average in Table 3. Table 03. Axial load increase in terms of ballast layer thick increase Axial load increase Average percent of required increase of ballast layer thickness (%) level (Ton) Passenger tracks Cargo (Freight) tracks 1 7 5 2 15 10.5 3 23 16 4 32 22.5 5 41 30 6 51.2 37 7 62.5 44 8 74 53 9 86 61 10 99 70

By keeping track condition, increase of axial load in fixed passing velocity condition increases ballast layer thickness progressively.( Figure 7)

119

ICESA 2014

Add Percent Depth of Ballast (%)

Internat ional Civil Engineering & Architecture Symposium for Academicians

95 85 75 65 55 45 35 25 15 5

Passenger Track Freight Track

1

2

3

4

5

6

7

8

9

10

Add Axle Load (ton)

Figure 07. Axial load increase. 5 Conclusions Technical and economical underlayment of railway tracks in order to providing track standard quality for achieving safe and comfortable traveling on tracks in one hand, and modern policies of railway networks for higher velocities and more axial loads on the other hand, result in increasing bed stresses. Deciding to improve and develop bed with increasing ballast layer thickness is the best option economically and technically. In this study regarding different parameters in ballast thickness determination, a good model is provided for determining ballast thickness based on load characteristics and passing velocity. In this study, track system was modeled by using KenTrack software based on relationship between variable parameters of ballast layer thickness and velocity and axial load parameters. Then axial load and velocity of vehicles used in tracks with 12 to 23 tons for passenger tracks with variable velocity of 100240km/h and 20 to 30 tons axial load for cargo tracks with 60-120km/h velocity were analyzed in model. Quantitative and qualitative effect of these two parameters on fixed parameter of bed area stress indicates that for keeping track system condition under a quasi static loading cycle, average percent of ballast layer thickness increase linearly increased by 5-99% per 1-10t increase of axial load and it increases by 7-73 % for 20-140km/h increase of passing velocity of Train. References Al Shaer, A., Duhamel, D., Saba, K., Foret, K., Schmitt, L., (2008), Experimental settlement and dynamic behavior of a portion of ballasted railway track under high speed trains, Journal of Sound and Vibration, Vol. 316, pp. 211–233, April. Beena Sukumaran, Associate Professor, (2002), Suitability of using California bearing ratio test to predict resilient modulus, Civil & Environmental Engineering, Rowan University. Boresi, Arthur P., Schmidt, Richard J., (2003), Advanced Mechanics of Materials", 6th Edition. John Wiley & Sons, New York, NY: Chap. 5, 10, USA. Esveld C., (2001), Modern railway track, 2nd ed. The Netherlands: MRT Publication, Netherlands. Jerry G. Rose, P.H.D., P.E.Professor of Civil Engineering. University of Kentucky,( 2006) Comparisons of Railroad Track and Substructure Computer Model Predictive Stress Values and In-Situ Stress Measurements, USA. Jerry G. Rose, Ph.D., P.E.Professor of Civil Engineering, University of entucky, (2006) KENTRACK - A Railway Track bed Structural Design Program, USA. Kaewunruen, S., Remennikov, A.M., (2007), Field trials for dynamic characteristics of railway track and its components using impact excitation technique, NDT & E International, Vol.40, pp 510-519. Manual for Railway Track Engineering, Vol.1, Chapter 1, part 2, (2006), Roadway and Ballast- ballast", American Railway Engineering and maintenance of way Association, USA. Sadeghi J., (2009), Fundamentals of analysis and design of railway ballasted track”, Iran University of Science and Technology Publication, Iran. Sadeghi, J., (2008), Experimental Investigation on the accuracy of current practices in analysis and design of railway track sleepers", Canadian Journal of Civil Engineering. Vol. 35, Canada. Selig E. T., Waters J. M, (1994), Track geotechnology and substructure management", Chapter 10-2, University of Massachusetts, USA. Sadeghi J. and Khordehbinan M., (2009), Investigation on Influences of Rail Support System on Railway Track Structural Behavio”, Journal of Transportation Engineering, Vol-1, PP 79-88, Iran. UIC CODE, 719 R, (1994), Earthworks and trackbed construction for railway lines", International Union of Railways, 2nd Edition. 120

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

RISK ACCIDENT ON URBAN ROADS RELATED WITH OVERTAKING MANEUVER

A. AFEZOLLI, I. KONDI [email protected], [email protected]

Polytechnical Univeristy of Tirana, Albania, Civil Engineering Faculty

ABSTRACT Safety conditions are the most important elements of planning and verification of transport infrastructures. The urban road safety level is strongly influenced by mutual interferences of vehicular traffic. In this contest, it would be of a very special interest, the manoeuvre of overtaking, because, when the traffic exceeds some fixed values, it causes the separation of traffic flows in two lanes. Negligence and disregard of traffic rules, mainly in urban road segments, has led to considerable decrease of the level of infrastructure security, and consequently the increase of accidents’rate. It is confirmed that the increase in well-being levels, is associated with the growing demand of the movement of vehicles and persons. Taking into account that one of the main causes of accidents is directly linked with the wrong overtaking maneuver, this paper aims to provide a synthesis of the reasons, which encourage this wrong maneuver, its dependence from human element, and recommend real ways of correcting that and increasing its level of security.[1] This paper is dedicated to the analyzis of the flow of vehicles on the street, especially referring to the problem of overtaking. Interaction between the vehicles that reduces the average velocity of flow, might conditionate the vehicle movement and overload attention, which the driver must have. Trend inhomogeneous on distribution of flows in the two lanes is due, consistent with the value of the flow inversion, condition limit is reached, for which the time interval that separates a rapid vehicle of two consecutive parades, results to be low by a time limit. As a result of this scenario, the rapid vehicle does not come anymore to the overtaking lane of normal motion, but continues the moving along the overtaking lane, causing her overstocking. Limiting the use of road infrastructure by slow vehicles, for some periods of the day, presents a condition for security improvement, related mainly with low probability of forming long group of vehicles and is tried to supply vehicle flows with the same characteristics of velocity. Key words: traffic, maneuver, vehicle, volume, flow, lane, accident INTRODUCTION Optimal operation of a road infrastructure provides traffic to flow in certain conditions and above all that, the user should be guaranteeing a high degree of security. Starting from the study of the overtaking maneuver on the street, should be studied on the dangerousness of overtaking maneuvers, driven by the change of traffic flow at road infrastructures. Materials and Methods Direction of a vehicle is expressed through a series of perceptions and reactions. At every moment, driver should exercise a control over the speed of momentum (choosing between the desired speed and how implemented) and the trajectory as a function of perception of safety and comfort.But, except to be focused before viewing, it is necessary to observe the space around the vehicle, using the look "back-side" view. (Figure 1) For these reasons happens that, the direction of a vehicle in roads with lot of traffic, causes an increase in fatigue, because of the grow information perceived and that must be processed by the driver. This situation might create moments of rejection, such that could lead him in performing wrong maneuvers. [2] Analysis of accidents in urban roads has shown that, the context in which accidents are more likely manifested is related to those physical parts of the road with high traffic volume.

121

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Figure 01. View plan during the guide Reality shows clearly that, probably an accident should appear in case when, a vehicle that is traveling in the slow lane passes through overtaking lane, or for negligence, or as a consequence of inflammation, taking into consideration that, the desired speed is limited by vehicles that traverse the movement slow lane at low speed. Case study realised by me for the problem of this paper is part of the junction between December 21 Road to the intersection with the new ring of Tirana city. (Figure 2)

Figure 02. Satellite view of the studied zone The chosen road segment for the study of vehicle has a length of 1,053 m. The cross-section consists of two lane carriageway of 4, and 3.5 m and a jetty of 2 m for each sense of movements separated by a separator, the central traffic of 2.2 m. [2] The maximum speed allowed, according to the Albanian Highway Code, is 40 km/h. Recordings were made in the running carriageway towards the East - West. The purpose of traffic movement of vehicles in this segment is home-working type, characterized by three peak periods, one in the morning, one at noon (lunch) and a later afternoon. Usually there is a peak by about 3100 vehicles/h. Measurements were carried out by two cameras with digital zoom. By exploiting differences between points registration quota placed on the roofs of the overpass and the carriageway, it becomes possible obtaining images with a better perspective. Cameras are positioned to target run-line separation lanes; two sections were recorded along the carriageway in 500 meters distance between them. [3] 122

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Cameras are synchronized between them in order to have the same time base. Start of records has been done through an operating system capable of simultaneously operating in two cameras. Measurements actions were calibrated through the use of an integrated topographic station, in order to put the coordinates of points in relation to the images, with the coordinates of the points of corresponding objects, referred to a Cartesian system orthogonal terrain. To implement a complete analysis of the behaviour that flows along the road segment, assets was recorded about 8 hour traffic during peak periods (morning, noon, afternoon, evening). Simultaneous analysis of trajectories of vehicles allowed the evaluation of the time sections time of analysis, identified by a characteristic element (eg. a vehicle with a specific color used as the reference element). To study the parameters of vehicle flow motion, some counted operations have been performed. Particularly, are counted vehicles that are within the "frame section analysis", their distribution through lanes, and their moving speed. With increasing of density "k, it increases also the interference between vehicles, reducing motion maneuverability, which brings a reduction in the average velocity of vehicle flow, expressed clearly with stated traffic fluency equation q = kv (Figure 3).

Figure 03. Velocity variation related to flow change The relationship between the distribution of vehicles "q", that in special case is detemined from Da = V*t, where t is the average time of reaction, that can be taken equal to 1.5 sec, is clearly expressed in figure 4. The diagram notes that lanes (of motion and overtaking) are charged in a different ways with the increase of cash flow. [4] By comparing the data obtained from the simulation, with experimental data obtained through observations carried out in different places of urban roads, is revealed full compliance of the data. Experimental data evidence that the absorption capacity of the two lanes flow, reverses for a total value equal to 1980 vehicles/hour; over this limit left lane gets flows with consistency (Figure 5).

Figure 04. Theoretical relationship between flow, density and velocity

123

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

3000

Flow of vehicles per lane

(veh/h)

Counting interval every 5 minuta 2500

y = 0.6332x - 214.43 R² = 0.9804

2000 1500

y = 0.3011x + 319.79 R² = 0.9447

1000 500 0 0

1000

-500

2000

3000

4000

5000

Total flow of vehicles (veh/h) Figure 05. Vehicle load in urban roads lanes

Measurement performed covers a wide range of time, where are presented different case studies. From measurements are extrapolated 55 case studies identified through a single flow element (example: a machine of a type or color). Each case study lasts 20 seconds. [4] Due to used device, it was possible motion tracking of the vehicles involved from the beginning to the last section. Case studies are extrapolated from a set of measurements basen on, in reference to the referring vehicle, which represent the "head", could never be crossed. Crossing speed of flow vehicles is achieved through the velocity of all vehicles involved in the case study. The same discussion is made for vehicles passing in both lanes, the slowdown and overtaking ones. Measurements were made for the time interval of 5 minutes and then were reported hourly. Each of the 55 case studies belongs to a time interval of 5 minutes investigation. This thing allows us to extend the considerations of accurate observations (5 minutes) in a greater interval, such that characterize the conditions of motion. From the made measurement was possible the identification of the average velocity of all vehicles held in the segment over the control. Velocities were taken from a provisional value of switching to all community vehicles in each of lanes. (Figure 6)

Average flow on lane every 5 min

3000 y = 0,7324x - 350,2 R² = 0,9317

2500 2000 Overtaking lane

1500 y = 0,1923x + 519,94 R² = 0,4736

1000

Motion lane

500 0 -500

0

500

1000

1500

2000

2500

3000

3500

4000

Average total flow every 5 min Figure 06. Loads of vehicles in both lanes, according to the total load change Results and Conclusions The graph in Figure 7, we have putted in relation intermediate flow in 5 minutes and the total flow of traffic that will be taken, if he was kept constant during the one hour, minimum intermediate distance, measured during the case studies.

124

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Figure 07. Passing Velocities on vehicles flow in both lanes Different slopes of the parts shown in Figure 7 clearly indicates that in the first major results the possibility of increasing the flow of vehicles, through approaching the "waves", while with increasing of vehicles flow, it lows the slope of the segment, which means that results difficult the of "waves" densification of vehicles to the detriment of freedom of maneuver. Beyond a threshold value (which results from surveys about 2000 vehicles/hour) in the overtaking lane vehicles move in close way, while vehicles traveling in the slow lane, begin to move in the distance. [5] When a driver will perform an overtake manoeuver, being shifted from the slow lane into that of overtaking, with increasing of traffic flow, will always face more difficulties, that will make him to take some risks, attempting to perform the maneuver in lack of safety distance to the preceding vehicle, or introducing with force to the flow of vehicles left, even when the distance between them does not allow. By the graph is evidently shown that curves, that intrapolate the distance between as temporary, have a similar trend, although they relative to data obtained indicate temporary full-time very low, coming in decline in the overtaking lanes; this indicates that the overtaking lane results so overloaded, such it is impossible the introduction to the slow lane. The driver, who is traveling in the slow lane and aims to overtake, with difficulty, could find a "suitable space" to be inserted in the left lane, only with the condition of assuming great risks. Riskiness on urban roads becomes higher when the flow vehicle increases significantly. For the reduction of this riskiness, mainly characterizing stages of overtaking, it is important to keep tight control on the development of traffic, trying to eliminate the disturbing elements that favor the formation of "waves" of vehicles. The argument addressed in this paper has allowed highlight as the average value of the density of vehicles flow, or even better the average distance between vehicles. It is just a mathematical abstraction that fails to present properly the quantities associated with quality of urban road traffic. [6] Discussions The difference of velocity between the right and the left lane increases by changing the flow of vehicles, which means that with the increase of flow, lanes could be specialized in different movement speeds. Much greater the speed difference is, more greater is the incentive that encourages drivers to guide with high speed and dare to be displaced from the slow lane to it faster. Analysis performed on the way of the appearance of driving vehicles load inside the road, show that the vehicle flow never travels uniformly; the movement is made in the form of waves, characterized by strong charge period, followed by periods of lower loads. The average value applied in the case of cash flow, gives some distorted information from reality, misunderstanding the correct functioning. [7] If the movement couldn’t become with waves and that during the entire hour will be held deflow conditions measured over 55 case studies, the flow that can be passed during road section would be too great. With a simple processing, comparing the vehicle flows drawn, taken from the minimum separation of each case study, and intermediate flow value of 5 minutes, it will be noted that transition volumes will be very high and not comparable with human possibility, which needs a freedom in movements to exercise correctly the management activity. [5]

125

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

References [1] Atti del Simposio Internazionale: Road Development and safety - 1989 [2] AA.W Highwaw Capacity Manual- Special Report 209- Ed Transportation Research Board – 1994 [3] G. Battiato, B.K. Larsen. La pianficazione della manutenzione stradale. Atti del XXI Convegno Nazionale Stradale – Trieste – Giugno 1990 [4] F.Bella, M.R. De Blasis. Lo studio dei deflussi per la siccurezza d’esercizio [5] A. Ranzo, G. Castisani – Aspetti critici nella percesione dei dipositivi di segnalamento [6] G. Tesoniere. Strade Ferrovie Aeroporti – Il progetto e le opera d’arte – 1990 [7] B. Festa, M. Marino, P. Giannattasio. La distribuzione transversal dei veicoli nelle carreggiate stradali

126

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

UTILIZATION OF RFID DATA TO EVALUATE CAMPUS COMMUTE CHARACTERISTICS

O. Altintasi, H. T. -Yaman [email protected], [email protected]

Department of Civil Engineering, METU, Ankara, Turkey,

ABSTRACT Analyzing travel behavior of METU campus users via traditional survey approach requires great effort. However using Radio Frequency Identification (RFID) System installed at all the campus entry gates provided a cheaper and easier approach to determine basic characteristics of the campus commute behavior. The RFID data with traveler details enabled the study of arrival and departure car-based commute behavior of academic personnel, administrative personnel and students, separately. The results revealed that campus travel demand is mainly active between 07:00 to 22:00. While majority of the commuters arrive during 08:00-09:00, the evening peak is distributed over a much longer period of 15:00 to 19:00. Administrative personnel have sharper evening departures between 17:00-18:00, while academic ones show a more scattered pattern lasting longer. Car-traveler students mostly arrive later during 09:00-10:00, and start exiting the campus as early as 15:00 lasting until late evenings. Stay time of vehicles on campus revealed that 43% of all trips to campus lasted less than 15 minutes, especially during morning and evening peaks, suggesting a high number of RFID card holders pass through the campus for pick-up or drop-off. A small reverse commute pattern suggested the work trips originated in the housing units on campus. Keywords: Commute travel; Traffic surveys; Travel behavior; Travel demand 1 INTRODUCTION Big university campuses may generate a travel demand like a small city, where a strong commute behavior is inevitable. Middle East Technical University (METU), Ankara, Turkey has a large campus serving a population over 30,000 people: University has 21,000 students and approximately 6,000 of them stay in dormitories. 4,300 people work as faculty and administrative personnel at METU, however, only 350 of who reside in on-campus housing units. The research and development (R&D) park on campus (called Technopolis) with a population of 3,000 is another region triggering commute travel. Besides the research and academic activities, the campus also have K-12 educational units (a nursery, an elementary and a high school) for the children of campus workers that affect the commute travel behavior, as well. METU campus was originally developed outside the city limits at a distance of approximately 13 km to the city center. Currently, it is surrounded by the city and access is controlled at 3 main gates (Gate A1, A4 and A7) as shown in Figure 1. Campus is accessible by municipality or privately operated bus and minibus services departing from the city center. Municipality buses also provide morning and evening commute services for students, once a day and to only limited number of neighborhoods. However, private car is an ever-increasing commute mode, which is controlled by a Radio Frequency Identification System (RFID) system at the entry gates; stickers are granted to faculty and administrative personnel directly, and for students in limited numbers. Additionally, RFID gate is located to control the traffic from Technopolis to campus (see Figure 1). Determination of travel behavior of the METU campus users requires traditional survey approaches,which would need a big amount of resources. However, integration of a series of data from different campus data sources (RFID and video recordings) provided a cheaper and easier approach to determine basic characteristics of the campus commute behavior, which is the main focus of this study.

127

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Figure 01. The location and the layout of METU Campus The layout of this paper is as follows: after a brief review of literature on commute travel studies in university campuses in Section 2, the RFID data analysis approach is summarized in Section 3, followed by the major commute behavior findings summarized in Section 4 and followed by the discussion and conclusions in the last section. 2 LITERATURE REVIEW Analyzing travel behavior of campus users has been focused of more detailed studies. Zhou (2012) studied the commute behaviors of university students in Los Angeles. He stated that improvement of multimodal transportation system and discounts in public transit may greatly change the travel behavior of students. Miralles-Guasch and Domene (2010) determined the travel pattern and transportation challenges of a university travelers in Barcelona through an online survey, where the lack of adequate infrastructure, the marginal role of walking and cycling and longer time involved using public transport were detected as the main barriers to shift from private car to non-motorized modes. Akar et al. (2012) also examined the travel patterns of the campus community at the Ohio State University. The result showed students were more likely to travel by alternative modes than faculty and staff members. Limanond et al. (2011) studied travel behavior of 130 students who live on campus in a rural university, and found that males and females had similar travel pattern; students owning a car preferred driving and non-car owners preferred riding with a friend and bus. Another study conducted by Gilhooly and Low (2005) investigated the primary school travel behavior in Midlothian, Scotland. The survey performed among 1008 primary school children and 776 of their parents to understand their travel behavior and the reasons motivating these travel choices. They stated that travel behavior was significantly influenced by age and the distance from school. As an evaluation of parking management on Beijing University of Aeronautics and Astronautics, China, Huayan et al. (2007) obtained inflow and outflow of vehicles during the day and calculated the average parking stay time of vehicles and draw conclusions about the travel behavior of university travelers. 3 MATERIALS AND METHODS In literature, most of the studies were performed via travel surveys to analyze the commute travel behavior; however, we have not found any study using new technologies, such as RFID, as such these technologies are not designed to study travel demand, but provide useful travel information as a byproduct. The location of the RFID readers in a region would inherently shape the capability of the commute behavior evaluations; if the whole region is access controlled like METU campus, it would provide a very rich data that could portray the behaviors of the majority of the commuters. If the region is partially controlled in terms of vehicular traffic, it is not possible to account for all the travelers, but get some sampling, which would inherent more bias based on the locations of the control points. On the other hand, traditional commute behavior studies also rely on sampling in their surveys, which may inherent some sources of bias, 128

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

themselves. Furthermore, RFID data-based evaluations would present “revealed” behaviors as opposed to “stated” ones in the survey-based approaches. Thus, it is impossible to capture everything about commute trips via either surveybased or RFID-based approaches, it is important to understand the limitations and sources of bias in the natures of the different approaches. Existence of RFID control system for private car users at all campus access gates made it possible to capture the mobility of all the car-based commuters. As the RIFD sticker included the license plate number, sticker type, and system stored the entry location, time for every reading (see Table 1), it was possible to obtain campus entry and exit information for all the movements of different commuter types (academic and administrative personnel, students, Technopolis workers, and parents of elementary-high school children). Tabulation of the entry and exit movements hourly (or in 15-minute intervals) easily provides the daily campus profiles. Further processing the RFID data for every unique license plate enabled the detection of all the recorded movements of the vehicles, leading to matching of entry-exit pattern that would define a “trip” (see Table 1). Even though it was not needed in the campus daily profile calculation, minor gate controlling the access to/from Technopolis is also assumed as a gate in the trip detections so that the trips destined to and originating from Technopolis region could be identified separately. Moreover, calculation of a “stay time” for a trip, which is simply the difference of the exit and entry times, sheds more light into the commuter behavior of the campus user. As the trips and their stay times can be time-stamped, it is possible to see the change in the stay times in a day. Table 01. Analysis result of RFID gate data

It should be kept in mind that, not all the trips detected from an RFID data processing should be regarded as “commute trips” directly and must be checked for the purpose of it. However, the fact that the METU grants RFID data to only campus commuters, disregards any concern on the matter for this study. Secondly, although majority of the commute trips are “work” trips, which are expected to happen in the morning and in the evening, it is not possible to generalize it for every campus user, as some would be working the late evening and night shifts in different units. Thirdly, it is not possible to know the true origin or destination of these trips from RFID data, but, the loss of this information does not necessarily endanger the study of basic characteristics of the campus commute behaviors. RFID data would have some reliability issues, as well. Due to some reading errors, it is possible to miss an entry or an exit of a vehicle, which would leave unmatched recording. But, not all unmatched recording would be due to reading errors. It is also possible that a campus traveler would make arrive to the campus with a car and leave campus with another mode for a specific purpose, or not leave the campus before midnight which would be the end of the daily data processing. Similarly, an exit with a missing entry could be the closing end of a car trip from a previous day. 4 RESULTS To study the commute characteristics, a whole week RFID data from Nov 21-25, 2011 was processed. To give an idea of the scale, the statistics from Wednesday of the control week is summarized in Table 2. The analysis result showed that 91 % of the vehicles tracked by RFID systems were private cars. It shows that the majority of the motorized vehicle demand in the METU campus was due to the private car access. To validate the results from RFID data, video recordings of the security cameras at the entry gates were processes for another Wednesday showing that there were slightly almost 3000 more vehicle entries, some of which can be due to daily fluctuations but the majority is due to the vehicles visiting the campus. In the video recordings, the difference between the total entries (15280) and exists (14828) was very small suggesting the majority of the vehicles did not stay overnight. However, this difference in the RFID data was 5074 movements due to significantly lower number of recorded exits. Further calculation of the entry-to-exit ratios for each gate from RFID and video data showed that majority of this difference was caused by a reading error in the exit lanes of the gate A1 (see Table 2).

129

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Table 02. METU campus vehicle activity statistics

In the video recordings, the difference between the total entries (15280) and exists (14828) was very small suggesting the majority of the vehicles did not stay overnight. However, this difference in the RFID data was 5074 movements due to significantly lower number of recorded exits. Further calculation of the entry-to-exit ratios for each gate from RFID and video data showed that majority of this difference was caused by a reading error in the exit lanes of the gate A1 (see Table 2). Exist behavior in this study is underestimated; however, this systematic error happened throughout the whole observation day and affect all the cardholder types similarly. Thus, it does not create any bias on the time of the day or the cardholder type analysis. 4.1 Campus daily travel profile The RFID gate data can simply be used to get the daily demand profile of METU campus and to detect the peak hours. As there were very small variations between the weekdays (Altintasi and Tuydes-Yaman, 2013), the average entry and exit profiles are used to illustrate the travel demand profile (see Figure 2). Both entry and exit profiles support the existence of the urban commute behavior among the METU campus travelers, where majority lived in the city and arrived METU as their workplace, while only a small reverse commute pattern is suspected due to the small on-campus housing population. The characteristics of the daily commute profile are as follows: A major entry demand was observed between 07:00-09:00; during which hourly arrivals reach up to 3000 in this time interval. Then, the campus entry demand gradually decreased until 11:00, after a small peak around 12:00, entry demand was almost constant at a rate of 1000 vehicles/hour until 18:00, after which it diminished significantly until midnight. The exit profile of METU campus showed a few exits until 07:00, after which, the number of exiting vehicles increased and first peak was seen around 08:00-09:00 with 1500 exits. This is expected to be the result of commuters residing in the on-campus housing units, and car travelers doing a quick drop-off for a campus user. The second peak was seen at noon times (12:00-13:00) with 1200 exits. The majority of the exit demand was in a scattered nature, starting from 15:00 to 19:00, which revealed a highest volume over 2300 during 17:00-18:00. After 19:00, the number of exiting vehicles gradually decreases.

130

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Figure 02. Average daily travel profile of METU campus on 21-25 November 2011 4.2 Commute behavior of different campus travelers Availability of the cardholder information in the RFID data revealed more about the commute behaviors of the different campus users. As the commute patterns of academic and administrative personnel and students have already been analyzed in detail in Altintasi and Tuydes-Yaman (2013), it is briefly summarized for the sake of continuity: • Administrative personnel showed a sharper morning arrival (see Figure 3) and evening departure times (see Figure 4), parallel to definitions of their work hours between 08:30-17:30. • Academic personnel had also a sharp morning arrival peak, and exit profile shows a small peak at 08:00-09:00. A major exit demand was observed during the evening hours, which started from 15:00 and continued until 21:00. • Entry profile of student travelers showed that travel to the campus started to increase after 08:00; morning peak was seen around 09:00-10:00 (see Figure 3). After which, entries gradually decreases during the day. The exits of students started from 08:00 gradually increased throughout the day (see Figure 4). After that, they had again scattered evening peak that start from 15:00 to 20:00. • Private car usage among Technopolis workers was a major component of the campus car usage. Their movements suggest a much clearer “work” trip purpose with sharper and departure time windows. Their entries start as early as 06:00 and continue until 09:00, after which it is dropped significantly (see Figure 3). The second peak during the lunch time definitely indicates a strong “lunch trip” demand that they also showed a typical urban travel demand profile. Lastly, parents of children were active only at 08:00-09:00 and 15:00-16:00.

Figure 03. Entry profiles of major traveler groups in a day

131

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

Figure 04. Exit profiles of major traveler groups in a day 4.3 Stay times of private car commuters The stay time analysis results in Table 3 of the METU campus users showed that in the overall evaluations, 42.8 % of the trips by the RFID cardholders were “short stay”, that is less than 15 minutes. A second major traveler group stayed between 1-5 hours (22.9 % in a day), which is more like a half workday time. Only 16.1 % had long stays of 5-10 hours suggesting that they may be academic or administrative personnel. The low percentage of the long stays in spite of the large number of personnel may be due to the fact that some of the personnel had multiple trips in and out of the campus. Table 03. Campus stay time of vehicles by entry time

* Out of 12194 entries, there were a total 6344 trips with matched entry and exit times 132

ICESA 2014 Internat ional Civil Engineering & Architecture Symposium for Academicians

During the early morning hours (07:00 to 10:00), more than 40 % of the entering vehicles stayed short, 20-30 % stayed almost the whole working hours 5-10 hours, while another major group stayed a half-day (1-5hours). The further investigation of the RFID cardholder among these drivers revealed that the major short stays were performed by Technopolis workers (46.0 %), academic and students had almost same ratio (17.5 %) and the parents of the children at the elementary and high school were 11.7 %. The short stay ratios were predominant in the evening hours (between 17:00-20:00), reaching up to above 60 %. The traveler distributions in these short stays had 35.7 % as Technopolis workers, 28.1 % by academics, 24.0 % by students and 8.2 % by administrative personal. Only 4.2 % of the total short trips were performed by the parents of the children, which is proportional to the number of sticker provided 5 DISCUSSION AND CONCLUSIONS For a regional commute behavior study, even for campuses, traditional surveys are generally performed. However, availability of RFID control for all the private car commuters at all the campus entry gates in the METU campus provided a cheaper alternative of data for commute behavior analysis. The RFID data provided the daily travel profile of the campus. The license plate and time stamp in the RFID data made it possible to detect “trips” in and out of the campus. Furthermore, RFID cardholder type information made it possible to study commuter arrival and departure times of the different campus users, such as academic and administrative personnel, students, etc. Campus stay time analysis, which is a revealed behavior, showed that there had been a great number of vehicles were staying in campus less than 15 minutes. Furthermore, the relatively small reverse peak-hour flows confirmed the existence of commuting of family members of academic personnel living on campus In METU campus, it was able to detect over 6300 trips from approximately 12200 RFID entry records in a day, which makes almost a 50 % sampling; this is much more than any sampling ratio aimed in a traditional survey. However, the lack of purpose for the trips identified from RFID data does not enable the identification of all components of commute behaviors clearly and create some room for speculation. REFERENCES Akar G., Flynn C., and Namgung M. (2012). Travel choices and links to transportation demand management: Case study at Ohio State University. Journal of the Transportation Research Board No. 2319, Transportation Research Board of the National Academies, 77–85. Altintasi O. and Tuydes-Yaman H. (2013). Commute behavior of METU campus travelers. In Proceedings of the 18th international conference of Hong Kong society for transportation Studies ( Lam WWY and Loo BPY (eds)). Hong Kong, pp. 89-95. Gilhooly P. and Low D.J. (2005). Primary school travel behavior in Midlothian, UK. Proceedings of the Institution of Civil Engineers – Municipal Engineer 158(2), 129-136. Huayan S., Wenji L., and Haijun H. (2007). Empirical study of parking problem on university campus. Journal of Transportation Systems Engineering and Information Technology 7(2), 135-140. Miralles-Guasch, C., and Domene E. (2010). Sustainable transport challenges in a suburban university: The case of the Autonomous University of Barcelona. Transport policy 17, 454- 463. Limanond T., Butsingkorn T., and Chermkhunthod C. (2011). Travel behavior of university students who live on campus: A case study of a rural university in Asia. Transport Policy 18, 163-171. Zhou J. (2012). Sustainable commute in a car-dominant city: Factors affecting alternative mode choices among university students. Transportation Research Part A 46, 1013-1029.

133