Modelling a material flow network of a casting house Alexander Bock [email protected]

Abstract The present contribution shows the process-modelling and analysing of a foundry with the software Umberto. Besides, the special difficulties which appear at this enterprise should be indicated. In addition to that, a suggestion should be discussed to solve the problem of the data capture. 1

Introduction

The modelling and analysis of operational material flow systems has become increasingly important. It is particularly suitable for putting the synthesis by economics and ecology on the operational level (Stappen 2006). And it is especially suited to save the resources and to increase the resources efficiency, which is the cause of a lasting development in the enterprise. The article shall point out the goal to illustrate practical experiences of the production of a material flow model in the case of the BAE batteries GmbH. This should be give a contribution for the energy conservation of the enterprise. An internal analysis shows that the foundry is an important power consumer. Therefore they decide to build up the complex system connections foundry as a model. That should include all necessary raw materials, supplies, auxiliaries and fuels, wastes and sources of energy.

2

Alexander Bock

2

Method

It is tried to lean on the procedural model of Bullinger. Bullinger and Beuker divide a material flow management cycle into a multiplestage model e.g. analysis, control and so on (Bullinger and Beucker (2000). They include into the stage analysis a material flow analysis cycle. This article looks at the stage of the analysis. For the solution of the concept the software Umberto is used. It is based on the concept of the material flow nets, which are developed according to the principles of the Petri - network theory (Möller, Häuslein and Rolf 1997). This represents a formal method to model systems and/or transformation processes (Schmidt and Keil 2002). At first an investigation period was specified (03.10.2005 - 31.12.2005). A second stage used activities, which are necessary to get basic information for energy conservation. In addition to that the concept (structuring of the nets) and the data acquisition of the data is important. Measurements of the casting machines were accomplished, because a lot of data were not present. Primary measuring points were thereby the lead boilers because these produce the largest current consumptions. It is also interesting to know, if there are energetically beneficial to drive down the lead boilers for a special time period. It was badly need to organise the data measurements and discussions and to evaluate existing data bases. With the getting knowledge of the data the model could be realised and balanced.

3

Realisation

The data is divided into raw materials, supplies, auxiliaries and fuels, wastes and sources of energy and products. The different lead alloys are assigned to certain machines. The experience shows that the production of the grids is very efficient. The input quantities of the lead could be determined over the incoming goods for the period of the fourth quarter 2005. A problem is that the single quantity is not directly assigned to the order. Therefore some total quantities had to be scaled on the machines. The consumption of tin could be determined indirectly, because for 100kg lead calcium 700g tin are needed. Similarly it behaved with the fleece bags. One is assigned

Modelling a material flow network of a casting house

3

exactly to one grid. This consumption based on the calculation of the output grid types. It is much more difficult to calculate the consumption of the auxiliary and fuels. All auxiliary and fuels are led only in SAP. Cork flour e.g. that e.g. cork flour (parting agent) is ordered only once in a year, but is used by five machines. Therefore all values consumed in Umberto are based on statements of the master of the foundry. The sources of energy electric power and natural gas could be well determined. A special current tool (electric measuring instrument) acquired data, which was written by a displaying measuring instrument of the enterprise Jumo. This was adjusted in such a way that every two minutes a value of the temperature and the used up ampere-hours are written into his memory. With emissions above all the lead emissions are meant. These were proven and based on a gravity casting machine. But a personal dosimeter was used, which was screwed on a mounting plate. The mounting plate is hung into the boiler departure of the machine. Due to the same design and the relatively similar extent of utilization of the machines a comparable emission is proceeded. All producible grid types are designated as products. 3.1 Concept During the investigation of complex system connections human beings push soon at limits. Without additional aids it is hardly possible, to realise, to understand and to reconstruct various dependence and dynamic procedures in a larger system. This applies equally to the systems of the natural environment (e.g. ecological systems), to the technical created by humans (in the available case a foundry) and economic systems. Therefore we use the characteristics of the system, which are relevant in the context of the respective problem definition and illustrate them in a model of the system. The system in the model is substantially simplified represented by aggregation, abstraction and idealization. The investigation takes place at the model of the system, because complexity of the systems becomes manageable. So the model shows the substantial system properties. Therefore it can be transferred the results, which were obtained with the model, to the system. Which system properties are substantial depends in each case on the question. This applies also to the defini-

4

Alexander Bock

tion of the system border. You define what has to be adding to the system and which can be regarded as system environment. Important is that each model describes a certain scenario in the context of a problem definition. 3.2 Structuring of the nets The illustration 1 shows the backbone network of the foundry. The orange-coloured framework represents the system border. A characteristic comes to the modelling of the grid blend. This is in- and output, because the blend from the last production step of a grid again in the first production step.

Illustration 1: main-net of the foundry

The illustration point out the three main energy consumers with the Transition, T1 foundry, T5 cooling system and T6 exhaust air system in the backbone network. This organization was made, in order to make clear that the foundry consists not only of the production department, but also of other ranges, which cause the energy consumption. It is attending to that in all nets a separation between in and output is clear, in order to be able to keep the nets well arranged. This formalism was broken into all sub-nets inside. The foundry is

Modelling a material flow network of a casting house

5

subdivided in two main production chains. On the one hand the production of positive, and on the other hand the production of negative grids. For the positive production process four die casting machines are used, which are appropriate for the production of certain grid types. The positive production process meant the production of grid types, which apply to the production step “shake”. In contrast to it the negative production process meant the production of grids, which apply to the coming production step “compound”. The nets differ between gravity casting (negative production process) and die casting machines only in the last production step. As you can see in illustration 2, the positive grids are drawn after cutting into tubing bags. That is necessary, because into the “shake” the red lead normally would not remain sticking to the grid. Red lead is a basic module for the building of a battery. It is needed for the construction of the batteries.

Illustration 2: positive production process

3.3 Computation of the nets The computation method works completely locally, continuously related to an individual transition or an individual place. This beginning reduces the qualifications for the application of the computation

6

Alexander Bock

method and permits the execution of the computations, even if the specifications are not yet complete in a material flow net. The computation direction is fixed on the direction of the material stream. The computation toward the material stream was made for the computation of the operating statement. The computation starts from the sub-nets, at which the input lead is the starting point of the computation in each case. This becomes clear by a pink coloured connection to the first Transition. The third picture exemplary shows the computed backbone network of the foundry in a sankey-diagram. The net shows thereby the energy consumption of the main energy consumers, whereby the consumption of the foundry is the largest.

Illustration 3: sankey-diagram of the foundry

4

Results

The results are based on the measurements and a weak point analysis. First results of the analysis are to be regarded. It is determined, who are the three main consumers of the foundry. The consumption by the lighting resounds are neglected. This should be for instance

Modelling a material flow network of a casting house

7

about 1% of the total consumption of the foundry. Interesting is here the portion of the exhaust air system. Approximately a third of the current consumption of the foundry reveals the fears that the installed plant is nevertheless too largely laid out. Since this the plant can drive only the conditions "enterprise" and "non-enterprise". An optimization is to be seen if you use more casting machines in the enterprise, because is the higher the efficiency of the plant. It is further interesting, as the monetary portion of the product grid behaves. The energy consumption at the grid is the second highest cost factor, but it is not to compare with the material costs, which are caused by the lead alloy. In addition, on the other hand it must be understood, that the enterprise does not have influence on the lead price, because these daily fluctuations are subjected to the conditions at the world market. The lead quantity of a grid is fixed to certain tolerances, which may not to fall below or exceed a limit. So the current consumption as next cost factor is much more important for the enterprise, because an influence on the consumption can be exerted here. Further results arose e.g. with the measuring of the warming up phases of the machines. So a preheating time could be based by 7h on the gravity casting machine 3, while the other gravity casting machines lay within the allowable times for the heating phase of the boilers. The reason is to be seen in only five functioning heaters. This affects adversely on the current consumption and the assignment of the machine. Further was to be demonstrated, when it is worthwhile to drive down a machine completely. If we look to the gravity casting with an average preheating time from 4h and 7 heaters a'5,5 KW, we would save energy starting from the 14h, that means that the machine would has approx. two shifts not to be used, when you want to have an energetically advantage. It can recapitulatory be said that it is meaningful to have more heaters at a boiler. So the energy consumption in the warming up phase can be lowered. With all nine functioning heaters a machine uses least electricity. This has on the other hand the advantage that you could redefine switch-on times for heating the machines in the week planning. By a temporally smaller heating duration you would save not only energy costs, but could be also more flexibly to the situation concerning orders react.

8

Alexander Bock Prozentuale Materialkosten an einem negativ gegossenen Gitter ( E PzS 150HS)

0,23%6,95%

Energie, elektrisch Talkum, Erdgas, Reinigungsmittel 360 Bleilegierung PBSB 2,1 SN Cu

92,82% Illustration 4: material costs of a negative produced grid

Switching a boiler off is worthwhile itself, if you does not have to produce two shifts. From several measurements a characteristic number for the current consumption at the boiler in the operating condition could be determined. This lies with about 42.1 % of the energy consumption, which is needed when warming up. From the view with the work on the model it can be said that it would be meaningful to assign the orders to the machines. Thereby you could determine individual lead consumption more qualitatively. An optimization is still possible, if the energy consumption of the product is taken up. You should think about the point after a suitable measuring period whether you can take up the flow coefficients at a machine. So you can be transfer the consumption of all necessary inand outputs to a grid. 5

Outlook

I think that there are possibilities to improve the vague statements to the auxiliary and fuels of the foundry. Some auxiliary materials are

Modelling a material flow network of a casting house

9

ordered e.g. once in a year for the sub-range foundry and used up then in roughly one year (e.g. cork flour). The quantitative employment at the individual casting machines is not documented, therefore the backtracking of the employment of cork flour at an individual grid is only possible over a rough approach estimation. This can be faulted. That means that the admission of certain data at the machine should be still made available. In order to avoid no larger expenditure, I would like to submit the following suggestion. If you would change that the once in a year bought cork flour is divided into all gravity casting machines and in the same course you equip the gravity casting machines with a counter, you could measure after many gone through grids, how much cork flour was used. Thus an average value could be determined for each machine. So you could get more exact statements about the final consumption of the individual operating supplies, if you should use this procedure with every other auxiliary material. After a certain measuring period a relatively good average value would result, which could be inmaintained again as meaningful calculation size in the provided model of Umberto. On the other hand it could be uncovered whether there are machine-specific differences in consumption. The model provided with Umberto could be designed to the individual grid for the flow coefficients and the internal computation algorithm could be made dependent on the flow coefficients. Thus could be more easily to assign the operating supplies to the grids. In my opinion the BDE of the foundry should be considered again. References Bullinger HJ, Beucker S (2000): Stoffstrommanagement und Betriebliche Umweltinformatiomssysteme (BUIS) liefern neue Impulse für das Umweltcontrolling. In: IAO 2000, p 1-18 Schmidt M, Keil R (2002): Stoffstromnetze und ihre Nutzung für mehr Kostentransparenz sowie die Analyse der Umweltwirkung betrieblicher Stoffströme. In: Hochschulausgabe Nr. 103, Pforzheim, Möller A, Häuslein A, Rolf A (1997): Öko – Controlling in Handelsunternehmen. Berlin Heidelberg New York Stappen RK (2006): A Sustainable World is Possible - Der Wise Consensus. Eichstätt