Innovatively investing in Europe’s Northern Periphery for a sustainable and prosperous future
European Union European Regional Development Fund
Final report WP 4
Pelletizing and Combustion
Håkan Örberg, Sylvia Larsson, Björn Hedman
SWEDISH UNIVERSITY OF AGRICULTURAL SCIENCES UNIT OF BIOMASS TECHNOLOGY AND CHEMISTRY
Contents Summary Preface Acknowledgements
1 Introduction and background 1.1 Pellet production and utilization 1.2 Set up for energy pellets production 1.2.1 General 1.2.2 Hopper and conveyors 1.2.3 Shredder 1.2.4 Pneumatic transport and separation of unwanted pieces 1.2.5 Dust separation 1.2.6 Fine mill 1.2.7 Conditioner 1.2.8 Pelletizer 1.2.9 Cooling and sieving 1.3 PELLETime 2 Experimental setups 2.1 Upgrading to pellets 2.2 Measurements at pelletizing 2.3 Sampling and chemical analysis of raw materials 2.3.1 Ash fusion temperature 2.4 Combustion tests 3 Results and conclusions
2 4 4 4 4 4 4 6 6 6 7 7 7 8 8 9 11 11 12 12 13 13
3.1 Quality of pellets and raw materials 3.2 Pelletizing properties and correlations with pellet quality 3.3 Improvement of pelletizing technology 3.4 Chemical characteristics of raw material 3.4.1 Ash contents 3.4.2 Net calorific value (NCV) 3.4.3 Chlorine (Cl) 3.4.4 Ash fusion temperature 3.4.5 Potassium, calcium and silica 3.5 Combustion tests
15 15 17 19 20 20 21 21 21 22 32
4 References
38
Summary The use of energy pellets for heat and power production is increasing world-wide. The dominating feed stock for energy pellets is saw dust or shavings from existing wood industry. Other raw materials than wood residues will be needed in efforts to increase pellet production and both energy crops and agricultural residues will likely be used as feedstock.
Willow and reed canary grass had considerably higher ash contents than what is regarded for class 1 pellets standard (≤ 0.7 %), which means that they probably only can be used in combustion facilities dimensioned for high ash contents. NCV varied from the lowest value 17.3 MJ/kg in reed canary grass up to 19.3 MJ/kg in pine. None of the analyzed materials has a risky high chlorine value.
Within the PELLETime project, the evaluation of some interesting raw materials from Northern Periphery Programme (NPP) area has been completed by pelletizing trials and combustion trials. From each country samples of approximately 500 kg was sent to the SLU pilot research plant Biofuel Technology Centre (BTC) to be analysed and treated before pelletizing. The samples were:
The ash fusion temperatures (AFT) of larch with bark and reed canary grass from Orkney indicates possible sintering problems in combustion. The ratios between potassium, calcium and silica are important factors for the ash melting characteristics in biomass. In reed canary grass from Orkney, the concentration of potassium is much higher than the concentration of calcium, giving an explanation of the lower AFT.
• Willow from plantations on Orkney Islands (Scotland), stored 8 months and not stored • Birch with bark from thinnings in northern Finland. • Pine with bark, contaminated and uncontaminated and debarked. From thinning in northern Finland. • Larch, with and without bark from Iceland • Reed canary grass, delayed harvest, from Sweden and Scotland (Orkney Islands)
Willow from Orkney also has high contents of potassium, but its relatively high AFT can be explained by its even higher contents of calcium. The K/Ca ratios of pine are almost the same in spite of the variation in concentrations, and so are AFTs also. For Larch, AFTs are lowered by mixing in bark in spite of a decreasing K/Ca-ratio.
At BTC, the materials were chipped, dried, shredded and milled in several steps before pelletizing took place. During the shredding operation, 10 samples from each assortment were taken out and split down to a general sample for proximate analysis (moisture content, ash content and net calorific value (NCV)), ultimate analyses (percent of C, H, N, S, Cl, K, Ca, Si and other elements) and ash fusion temperatures.
The materials were pelletized at different moisture contents and specific energy consumption (kJ/kg), mechanical durability of pellets (%) and pellet bulk density (kg/m3) were measured. Except for birch, all materials were successfully pelletized and good pellet qualities were achieved. Pellets produced from larch, willow and reed canary grass are all denser and have higher durability than pine pellets. Durability for pine pellets and for larch pellets show a clear decreasing tendency with increasing admixture of bark. For larch pellets, the durability also increases
with higher moisture content in the range between 9.0 and 13.3 %. Raw material moisture content is negatively correlated to specific energy consumption within the moisture content intervals tested. There is a raw material specific moisture content optimum for maximum mechanical durability within the range of 5 and 16%. Specific energy consumption is negatively correlated to moisture content. Pellet bulk density is negatively correlated to raw material moisture content A development of pelletizing technology, based on temperature control, for materials prone to discontinuous production behavior has been performed, and the results from these tests give indications for temperature control to be very interesting as a new process controlling parameter. Most pellets qualities tested for combustion (pine and larch, with and without bark) were easily combusted with fairly low CO-emissions. In combustion of willow however the heat evolved slowly, resulting in low CO2 emissions and frequently re-occurring COpeaks. This inefficient combustion can be the result of too hard pellets. For larch with added bark, the sintering tendency (slag forming tendency) is very high. Approximately 80 % of the ash part of bottom ash sintered.
Authors Håkan Örberg Sylvia Larsson Björn Hedman ............................................................. Swedish University OF Agricultural Sciences Unit of BIomass Technology and Chemistry www.btk.slu.se
Preface The PELLETime project aims to increase production and use of fuel pellets in peripheral Northern areas.
Raw material storage
Separation of metal and rocks
Shredding
Moisture adjustment
Acknowledgements Financial support of European Union, Northern Periphery Programme (NPP) and the Faculty of Forest Sciences of the Swedish University of Agricultural Sciences (SLU) is gratefully acknowledged.
Steam conditioning
Fine milling
Separation of metal and rocks
Buffer storage
1 INTRODUCTION AND BACKGROUND Pelletizing
1.1 Pellet production and utilization Energy pellets are used world-wide for heat and power production and the international trade of pellets is under tremendous increase. It is expected that the total world production will have a tenfold increase from 50 million tons/year (10 TWh) today to 500650 million tons/year (100-130 TWh) 2020. Countries like Sweden and Finland have well established pellet industries and production in countries with large raw material sources, such as Canada and the Baltic countries is increasing fast. The dominating feed stock for energy pellets is saw dust or shavings from existing wood industry. In the efforts to increase the use of renewable energy, other raw materials than wood residues will be used for pellet production. In countries with wide areas of agricultural land, both energy crops and agricultural residues will likely be used as feedstock for pellets. Local production and consumption of fuel pellets can be based on raw materials presently not dominating,
such as trees from thinnings, energy wood, short rotation crops (SRC), and other energy crops like reed canary grass (RCG), etc. It is also considered that profits from small scale pellets production and local consumption is better distributed in the society compared to large scale production and export. Some small scale producers also have the opportunity to use their pellets in their own heat plants which may give them the possibility to use more unexplored feedstock that presently are used by large pellets producers.
Cooling
Sieving
Pellets
Figure 1. Flow chart for the production of energy pellets
Low specific density of milled material also causes problems in pelletizing. The feeding system of industrial pelletizers is not developed completely for all type of raw materials. The production capacity of the different components in the upgrading system
has to be dimensioned to guarantee a continuous feeding to each component. Buffer volume has to be built in to even out flow variations. A complete set up for production of energy pellets can be seen in Figure 1.
In some parts of the NPP areas like Scotland (Orkney and Shetland) the production and use of pellets are almost non-existent due to lack of wood material. In these areas, a production of i.e. willow can be an important feed stock for pellets. In Iceland a considerable planting program of larch and spruce has started to produce possible raw materials for small scale pellets production.
1.2 Set up for energy pellets production 1.2.1 General The complete system for upgrading biomass into pellets includes a series of production steps. Some of these steps are only partly examined and tested in full scale or medium scale for other materials than sawdust. The planning of a handling system must be considering the fact that “dry” material (12-16 % dry material (DM)) includes a risk of fire accidents and dust explosions. For this reason a shredder with slow speed rotation tools must be used. Dust generation from shredders, transport equipment and pelletizing equipment is a general problem and has to be controlled in order to avoid influence on
personal health. Dust has to be collected and filtered from ventilation air. When shredded and milled, the flowing properties of straw type materials are poorer than compared to i.e. sawdust. When the material does not flow and fall freely, there is a great risk for bridging of the material during transport and eventually the material can get stuck. This must be avoided by transport equipment that actively helps the material to flow. In this aspect transport by air stream (pneumatic transport) is advantageous.
1.2.2 Hopper and conveyors
1.2.5 Dust separation
The conveyor is transporting chips or bales in to a shredder or a bale separator. For transportation of chipped material, chain conveyors are functional. For transporting whole round or square bales, band conveyors are more reliable. The size of conveyors has to be dimensioned for a number of bales that can supply the rest of the system for a period without operator. For a full scale factory, a number of band conveyors can be mounted in parallel. Band conveyors are tolerant to variation in loading capacity. They can also be complemented with detector and automatic stop if big metal pieces are hidden in bales.
Straw materials are generating great amounts of dust when shredded and transported. This dust causes environmental problems and must not be exposed to operators since it also causes personal health risks. The whole upgrading system must be set under pressure and dust transported away by air flow to a filter (Figure 4). The filtrated air must not be brought back in to the factory localities but out into open air. To save energy, filtrated air could pass through a heat exchanger connected to the heating system of the factory facilities. The dust fraction that is collected in the filters is pushed out from the filters by air pressure shocks. The dust fraction can be used for briquette production or be combusted as a powder fuel.
For shredded or coarse milled material that is nonfree flowing material, chain conveyors operate safely. These conveyors have to be covered in order to avoid dust spreading. It is also recommended that wearing components are made of plastic or rubber in order to avoid ignitions.
1.2.3 Shredder Different shredders with a wide range of capacity are available on the market. Capacity is strongly dependent on sieve size and diameter of cutting cylinder. Shredders with a capacity of up to 10-15 tons/h are available. A slow rotating speed (80-100 rpm) of the cutting cylinder is necessary to avoid ignition generation when dry feedstock is used. Figure 2. Shredder and bale reducer
1.2.4 Pneumatic transport and separation of unwanted pieces
Figure 3. Principal function of pneumatic transport and separator (“Liftsep” Fransson recycling)
The risk of dust explosions has to be considered and “fire fly systems”, rapid extinction equipment that is able to limit damages in case of fire caused by dust explosion, might be necessary in some sections.
1.2.6 Fine mill Before pelletizing a second step of milling the material is needed. This is usually done with a hammer mill (Figure 5). In the mill an exchangeable screen is limiting the particle size in the outgoing flow. Milling uses a lot of electrical energy and with finer screens more energy is needed. A general recommendation is that the openings in the screen should not be less than 2/3 of the pellets diameter.
1.2.7 Conditioner
After coarse milling the material can be transported pneumatically (Figure 3), which is both efficient and flexible. This transport system is suitable to combine with sorting system to separate and get rid of unwanted components like rocks or metal pieces.
Some biomass materials have long and stiff fibres. This makes the pelletizing process more difficult. In order to make fibres softer and easier to compress and deform, they can be treated with superheated stem in a conditioner (Figure 6).
Coarsely milled material can be led further in the production chain through pipes and the flexibility of pipes makes it easy to change the flow direction if needed.
In the conditioner, hot steam is let into the material flow during intensive mixing. This method is generally used in energy pellet production and is cost efficient since the total specific energy consumption is lowered by the conditioning process. The conditioner is normally mounted directly to the pelletizer. In some applications an extra conditioner is mounted earlier in the material flow in order to achieve a resting period after the steam addition.
From the cyclone, material is fed into a fine mill. The fine mill is intolerant to metal pieces or small stones, so these have to be eliminated by the use of a separation table. The separation can be complemented with strong natural magnets in order to collect small metal pieces. The capacity of these components must be correlated to following machineries.
Figure 4. Dust filter, automatically cleaned with pressurized air shocks.
Figure 5. Hammermill
Figure 6. Industrial pelletizer with conditioner on top (Münch)
1.2.8 Pelletizer
1.2.9 Cooling and sieving
1.3 PELLETime
Several industrial pelletizers for energy pellets are available on the market with a production capacity 0.5-5.0 tons/h. Machines are normally optimized for woody material and raw material like straw or energy grass have different characteristics for pelletizing compared to wood. This fact gives for example that no production capacity for willow, straw and reed canary grass can be guaranteed. Further development of special pre-treatment equipment is demanded. Typical specific energy consumption for pelletizing is lowered by larger machine size. For an industrial 5 ton pelletizer normal electric consumption is 45-50 kWh/tons pellets measures for wood.
When pellets are produced, temperature will rise to 90-100°C in the pellets, due to pressure and friction. To ensure durable pellets, the temperature has to be lowered before storing or packing in sacks. The cooling procedure can be carried out by special pellet coolers. It is of great importance that cooling capacity is corresponding to pellet production capacity and temperatures to avoid production of dangerous gases in the pellets storage.
The evaluation of some interesting raw materials from NPP area has been completed by pelletizing trials and combustion trials. From each country samples of 200 – 1000 kg of each species was sent to SLU pilot research plant (Biofuel Technology Centre) BTC in Umeå, Sweden, to be analysed and treated before pelletizing. A survey of the different materials being evaluated is shown in Table 1.
Pelletizers and conditioners can be complemented with a facility for dosing additives into the pellets. Additives can be used up to 2 % of DM in order to improve pelletizing or combustion characteristics.
After cooling process, fine particles and dust attached to pellets have to be reduced by sieving. Different sieving or screening machines have been developed for the pellets industry. Dust mixed in with the pellets product will cause disturbances in combustion and will easier absorb moisture compared to solid pellets.
COUNTRY
Scotland (Orkney Islands)
Willow from plantations, age
AMOUNT DELIVERY DATE 500 kg DM
Sep-08
500 kg DM
May 2010
4-5 years, stored 8 month Scotland (Orkney Islands)
Figure 7. Centrifugal screener (Münch)
MATERIAL
Willow from plantations, age 4-5 years, not stored
Northern Finland
Birch with bark, from thinnings
Northern Finland
Pine debarked, small trees from thinnings, contamination avoided during transports
Northern Finland
Pine with bark, small trees from thinnings, contamination avoided during transports
Northern Finland
Pine with bark, small trees from thinnings, contamination not avoided during transports
500 kg DM 500 kg DM
Oct 2009
500 kg DM
Oct 2009
500 kg DM
Oct 2009
Iceland
Larch
1000 kg DM
Iceland
Larch bark
200 kg DM
Sweden
Reed canary grass, delayed harvest
1000 kg DM
Reed canary grass, delayed harvest
500 kg
Scotland (Orkney Islands)
Oct 2008
Sep-2008 and Nov-2009
Nov 2009 June 2008 and June 2009
May 2010
Table 1. Received materials for evaluation.
Compared to Sweden, commercial production and combustion of RCG is better developed in Finland, where the estimated production area in 2008 was 20 000 ha [3]. Also in Finland RCG is co-combusted with wood and peat in large CHP plants.
Willow (Scotland) Willow (Salix) is a potential energy crop for production in some areas in Scotland where shortage of forest is a reality caused by climate or cultural conditions. Upgrading willow by drying and densification to pellets or briquettes is an alternative for local production and residential heating and an alternative to fossil fuel heating. Pelletizing of willow has been carried out in several projects in Sweden [1, 2] and reported results have been acceptable considering pelletizing characteristics.
Larch (Iceland) Since the 1970s, larch (Larix), Sitka spruce (Picea sitchensis) and lodgepole pine (Pinus contorta) has been planted in Iceland. The first plantations has been harvested and especially for larch there is an interest to investigate local upgrading of whole or parts of trees into pellets for combustion in small scale residential heating. Although Iceland in wide areas is well supplied with geothermal heat, other areas do not have this advantage. A small scale production of pellets from larch and larch residues could be an alternative for some areas in Iceland.
Small scale production and combustion of RCG briquettes has been successfully performed in the research plant BTC at SLU, Umeå, Sweden for more than ten years. RCG pelletizing have proven to be more challenging, but new promising technology for the pelletizing of straw and RCG has been developed and is presented in this project.
Pine, small trees (Finland) Small trees from thinning in pine stands has in Sweden and Finland rarely been collected as a raw material for pulp or energy wood. Instead they have mainly been used for log firing. In small scale production of biomass for energy, even small trees from thinning have a rising value. Up to a certain level, bark can be included in the biomass without problems in combustion, which is studied in this project. An improved technology for multi tree harvesting of small trees in thinnings has contributed to the possibility of using this assortment for energy production.
Birch (Finland)
2
EXPERIMENTAL SETUPS
2.1 Upgrading to pellets The pelletizing properties of materials sent to BTC (Biofuel Technology Centre) were evaluated. Some of the pellets qualities were made from materials as received whereas others were made by mixtures of wood and bark. Raw materials tested were: • Pine – pure stem wood and stem wood with bark. In addition there were two qualities of samples with bark – one where contamination had been avoided during transports and one where contamination was not avoided. • Larch – pure stem wood, and stem wood/bark mixtures with up to 15% bark content • Willow • Reed canary grass – delayed harvest • Birch The materials, delivered from the NPP project partners, were chipped wood (larch, pine, birch), willow and in the case of reed canary grass, round bales (140x120 cm) when they arrived to SLU-BTK in Sweden. The bales have been handled by tractor with a front loader from the storage area into the plant. For upgrading to pellets, an experimental set up (BTC, Umeå, Sweden) with shredding, milling and pelletizing was used (Figure 8).
All assortments were dried down to 12 % water content. After drying, materials were placed on a band conveyor for transport into the shredder (Lindner 2000, 5000 kg/h). In the shredder, which is operating with cutting elements toward corresponding counter elements and a sieve for particle size control, the material was milled down to 15-20 mm. In this operation the material was transported directly to the mill passing silos, mixing screw, gravity separator and cyclone. In the gravity separator possible heavy particles like gravel or metals were isolated. In the fine mill (Hammer mill Bühler DFDZK-1, 55 kW), the particle size was further reduced by passing sieve diameter Ø 4 mm before transportation to the pellet press. The pelletizing experiments were carried out in a small scale pelletizer with a nominal production capacity of 300 kg/h (SPC 300). The pellets press had a steady mounted ring die and a rotating shaft for holding two press rollers. The ring die had three rows with Ø8 mm holes and with an active press channel length of 55 mm.
In small scale forestry, birch has the best value for log firing in stoves and small boilers. By thinning operations, small trees of birch could be collected in the same way as small pine trees in a multi tree harvester. In some cases it could be interesting to use birch wood for mixing with other raw materials. The hardwood characteristics of birch are disadvantageous in pelletizing operation.
Reed canary grass (Sweden and Scotland) For more than ten years, reed canary grass (RCG) has been an option for farmers willing to produce energy crops in agricultural fields in Sweden. However, the market has been more doubtful and not many combustion facilities have been using this fuel. Some disadvantages is a relatively high ash content and a low bulk density. Recently, some large combined heat and power plants (CHP) in Västerbotten in northern Sweden have contracted some 500 ha of RCG for co-firing with other fuels like wood and peat. These mixtures have been successfully combusted under normal full scale conditions.
10
Figure 8. Experimental set up for shredding, milling and pelletizing
11
2.2 Measurements at pelletizing The materials were pelletized at different moisture contents and the following responses were measured: • Specific energy consumption (kJ/kg) • Mechanical durability of pellets (%) • Pellet bulk density (kg/m3) Specific energy consumption was defined as the amount energy needed for pelletizing one kg of pellets. Specific energy consumption was calculated from the pelletizers motor current (minus the current required for idle running) over a specific amount of time when the pellet production (kg/h) was measured. Raw material shavings were sampled repeatedly from production line during pellet production and the moisture content was determined by heating a sample overnight at 105°C. Portions of 5-10 kg of produced pellets were sampled and used for determination of share of fines (< 3.5 mm), bulk density and mechanical durability. The mechanical durability was measured by tumbling a specific mass of pellets for a certain time period in a tumbler. After tumbling, pellets are sieved and the percentage
of remaining pellets from the original mass equals the durability measure. Bulk density is measured by weighing a known volume of pellets. It should be pointed out that bulk density primarily is not a quality measure of the pellets, but a parameter that is of interest for storage and transportation. Bulk density is highly dependent on the length of the pellets and can tell something of their properties only when comparing qualities of similar lengths. The determinations were made according to standard methods [4, 5, and 6].
2.3 Sampling and chemical analysis of raw materials In order to create a representative sample from each biomass material the sampling has been done carefully. During the first milling (shredding) operation and in downstream, 10 different samples were taken out, mixed and split down to a general sample (Figure 9). The general sample was used for proximate analysis (moisture content, ash content and net calorific value (NCV)), ultimate analyses (percent of C, H, N, S, Cl, K, Ca, Si and other elements) and Ash fusion temperature (AFT), in SLU and other laboratories.
Original sample
Initial melting temperature
Spherical melting temperature
Hemispheric melt temperature
Flow melting temperature
Figure 10. The principles of ash melting phases according to ASTM D 1857-68
2.3.1 Ash fusion temperature Ash fusion temperatures (AFT) have been analyzed according to standard ASTM D 1857-68 and the different temperature levels are defined as: IT: Initial melting temperature (Shrinkage temp) SF: Spherical melting temperature (Deformation) HF: Hemispheric melt temperature FT: Flow melting temperature A cone of ash is formed by pressing it in a mold and is the placed in an oven where heat is raised slowly. The deformation temperatures of the cones are registered as showed in Figure 10 and Figure 11 below. Figure 11. Cones of ash in different stages of melting
Sampling for chemical fuel characterisation
2.4 Combustion tests Pellets of seven qualities, larch (pure wood), larch with 5% bark, larch with 15% bark, pine without bark, pine with bark (contaminated during transport and uncontaminated) and willow, were tested in combustion. The pellets were combusted using an underfed burner cup technology (EcoTech 30 kW) (Figure 12). Pellets are transported to the cup from underneath in the centre of the ring-formed cup. Primary air is supplied trough slots in the ring. The upper part of the ring made of cast iron is rotating during operation to make the ash to be lifted out of the cup. Secondary air is supplied through a pipe on top of the burner cup.
1 general sample
Lab 1
Lab 2
Lab 3
Figure 9. Routine for sampling Figure 12. EcoTech 30 kW burner with underfed burner cup.
12
13
In Figure 13, the experimental set up for the combustion test is shown. Emissions in flue gas were collected through a heated ceramic filter and a heated sample tube and measured by an array of detectors. NO, NO2 and NOx were measured by electrochemical detection, O2 by paramagnetic resonance and SO2, CO and CO2 by infrared detection. These values were logged continuously through the combustion.
After every combustion experiment, all the deposits from the burner cup and the bottom of the boiler (in the following denominated “bottom ash”) were collected for analyses. Unburned share of bottom ash was determined by heating in 550 °C according to standard method [7]. The ash was also sieved to separate ash from slag and all melted particles greater than 3 mm was removed from the ash as slag.
Particles in the flue gas (dust) were determined with isokinetic sampling on a filter followed by gravity measuring.
3
RESULTS AND CONCLUSIONS
3.1 Quality of pellets and raw materials Except for birch, all materials were successfully pelletized and good pellet qualities were obtained. For birch, the press channel length of 55 mm used for the other materials was too long and production conditions were impossible to sustain due to high wear on the machinery. Instead, a die with a press channel length of 30 mm was tested, but that channel length was too short, and durable pellets could not be produced. No further
testing on birch feedstock was performed at SLU. Pine pelletizing was done to produce pellets for combustion tests, but no data was gathered during pelletizing. Analysis of bulk density and durability for some of the tested materials are shown in Table 2.
Fluegas pipe
Boiler tubes
Heated ceramic filter
Pellet fuel feeder
Heated sample tube
Air supply fan
Emission analyser
Figure 13. Experimental setting for combustion tests.
14
Burnercup
BULK DENSITY, kg/m3 DURABILITY, %
Larch stemwood
699.6
98.1
Larch 5 % bark
679.6
97.1
Larch 15% bark
673.7
96.4
Pine without bark, clean transport
517.5
80,4
Pine with bark, clean transport
552.1
71.7
Pine with bark, clean transport
531.4
67.1
Larch with <5% bark, low moisture
672.8
96.8
Larch with <5% bark, medium moisture
643.8
94.7
Larch with <5% bark, high moisture
571.8
89.2
Larch stemwood
699.6
98.1
Larch with 5% bark
679.6
97.1
Larch with 15% bark
673.7
96.4
Willow
698.1
90.2
Table 2. Bulk density and durability for some tested materials
15
Since bulk density depends highly on the length of pellets, cautious comparisons can only be made between measurements where the pellets length distributions can be assumed to be similar. In this study, this usually is the case when tests have been done on similar material on same occasion. Therefore, in Figure 14 and Figure 16), bulk densities are shown as relative values, compared only to others of the same species.
Figure 14. Bulk density of pine, relative values.
Bulk density for pine pellets with bark is somewhat higher than without (Figure 14). No effect of contamination is visible. These results are however from single samples and thus impaired by uncertainty. Durability for pine pellets (Figure 15) and both bulk density and durability for larch pellets show a clear decreasing tendency with increasing admixture of bark (Figure 16, Figure 17). And for larch pellets, the durability also increases with higher moisture content in the range between 9.0 and 13.3 %. Pellets produced from larch, willow and reed canary grass all have higher durability than pine pellets (Table 2, Table 3, Figure 17). High material density entails high energy density, which of course is positive. High durability is also usually associated with high surface hardness, which can cause problems with oxygen supply in the interior of the pellet during combustion. Moisture content (%)
Figure 15. Durability for pine pellets. Single samples
RCG1 RCG2 RCG3 RCG4 RCG5 RCG6
13.6 15.4 13 13.6 15.6 14.1
Bulk density (kg/m3)
684 651 677 648 613 642
Durability (%)
92.6 98.3 96.5 98.2 97.3 97.7
Table 3. Moisture content, bulk density and durability of some of the tested reed canary grass samples.
3.2 Pelletizing properties and correlations with pellet quality Raw material moisture content (%) was negatively correlated with specific energy consumption (kJ/ kg) for all materials within the moisture content intervals tested (Figure 18). When pelletizing larch, energy consumption was lower for larch with bark compared to stemwood. This is probably due to a higher content of lubricating resins in the bark, compared to stemwood. Willow pellet production showed the highest energy consumption and larch pelletizing the lowest. Addition of bark to the larch stem wood appears to further lower the specific energy consumption. Trends for willow and larch were clear, whereas the correlations in reed canary grass pelletizing were ambiguous. Uneven pellet production in reed canary grass pelletizing skews the results. Strong correlations between raw material moisture content and mechanical durability were found for all materials tested (Figure 19). Larch and willow pellets durability showed linear correlations with raw material moisture contents within the moisture content ranges tested. Larch was negatively correlated, and willow positively correlated. Durability of reed canary grass pellets can be approximated with a second order correlation with moisture content. The maximum durability was achieved at approximately 14-15% moisture content. The pattern of a durability maximum at specific moisture contents is most likely also true for willow and larch, but the maximum durability for these materials is outside the moisture content range tested in this experimental series. The moisture content optimum at maximum mechanical durability is generally lower for woody materials, compared to straw materials, and this pattern is confirmed in this study. The high moisture content optimum for willow, suggests that short rotational woody biomass behave similar to straw biomass in a compaction situation. Durability of larch pellets with bark was slightly higher than for stem wood larch pellets at 11% moisture content.
Figure 18. Specific energy consumption (kJ/kg) for pelletizing raw materials at different moisture contents (%)
Figure 19. Mechanical durability (%) of pellets pelletized at different moisture contents (%)
Pellet bulk density is negatively correlated to raw material moisture content for all materials. Correlations for reed canary grass are not clear, most likely due to uneven pellet production.
Figure 16. Bulk density of larch, relative values.
16
Figure 17. Durability for larch pellets with similar moisture content (7.0-7.6%). Average and standard deviation (four samples for larch stem wood, three samples for larch with 5% and 15% bark respectively).
Figure 20. Pellet bulk density (%) of pellets pelletized at different moisture contents (%)
17
No clear tendencies were found between specific energy consumption (kJ/kg) and mechanical durability (%) for reed canary grass and larch (Figure 21). The correlation found for willow may very well be a hidden effect of higher moisture content at lower specific energy consumption.
Figure 21. Mechanical durability (%) of pellets pelletized at different specific energy consumptions (kJ/kg).
Figure 22. Bulk density (kg/m3) of pellets pelletized at different specific energy consumptions (kJ/kg).
Bulk density was positively correlated to specific energy consumption (Figure 22) but this is also an effect of the negative correlation between moisture content and specific energy consumption. At lower moisture contents more energy is required to produce the pellets, and the higher compression energy results in a higher bulk density. However, whether it is the moisture content or the energy consumption, or both, resulting in a higher bulk density is not possible to say from this experimental series. Correlations between bulk density and mechanical durability were found for larch and willow. The positive correlation found for larch is due to the fact that both durability and bulk density increase with decreasing raw material moisture content. To reach higher overall pellet quality, a lower raw moisture content than 9% is recommended for larch pelletizing. Willow pellets have a negative correlation between durability and bulk density. Thus, no general advice can be given on how to improve pellet quality. However, the bulk density and durability of willow pellets are extremely good, and no further optimization is needed. For reed canary grass, no clear trend between durability and bulk density is found. Pellet durability is substantially lowered at moisture contents below 13% (Figure 19), whereas the bulk density is rather stable over a wider moisture content range. Thus, reed canary grass pellets should be produced at a moisture content of approximately 14-15%.
3.3 Improvement of pelletizing technology Raw materials, such as reed canary grass, straw, and some woody materials are prone to cause problems with a discontinuous production pattern. The industrial process becomes intermittent, pellet quality deteriorates, and, in severe cases, no pellets at all are produced. A discontinuous raw material feeding to the die, i.e. feed layer loss, due to unfavorable raw material specific frictional conditions will give rise to this kind of production behavior. In many cases, adjustments of the raw material moisture content or steam treatment are enough to improve pelletizing properties, but not for all raw materials. A development of pelletizing technology for materials prone to feed layer losses has been performed at the Biofuel Technology Centre, Umeå, Sweden, based on temperature control. A pelletizer die was equipped with a temperature control system where the die surface temperature could be regulated within the range of 30 to 75°C. A series of pelletizing trials, using reed canary grass as raw material, was performed, and the results from these tests give indications for
temperature control to be very interesting as a new process controlling parameter. The pelletizer motor current curves are helpful to illustrate pellet production behaviours. In Figure 24. pelletizer motor current (grey line) at three different die surface temperatures (red line) when pelletizing reed canary grass with 12% moisture content. In Figure 24 motor current curves from reed canary grass pelletizing is shown. Raw material settings are the same in all three cases, but pelletizing is performed at three different die surface temperatures. At the higher temperatures, pellet production is intermittent, and the current fluctuates and sometimes reaches very high levels due to shocks in the process. At the lowest die temperature, production is continuous and the current show a smooth pattern. Another advantage with temperature control is the immediacy of response when die temperatures are regulated. The regulation mechanism is fast and exact, and at some settings, there was an immediate response in production behavior from discontinuous to continuous when the die was cooled.
Some general conclusions on pellet quality and specific energy consumption are: • Bulk density is negatively correlated to moisture content at raw material moisture contents above 9% • There is a raw material specific moisture content optimum for maximum mechanical durability within the range of 5 and 16%. • Specific energy consumption is negatively correlated to moisture content Figure 24. Pelletizer motor current (grey line) at three different die surface temperatures (red line) when pelletizing reed canary grass with 12% moisture content.
Figure 23. Mechanical durability (%) and bulk density (kg/ m3) of pellets.
18
19
3.4 Chemical characteristics of raw material All general samples were analysed with the same methods and the ash contents and the ash chemical composition can be seen in Table 4.
Description
Pine without bark, clean transport
Pine with bark, Clean transport
Pine with bark, Contamin ated
Larch without bark, Iceland
Larch with Larch with Willow 5 % bark, 15% with bark, Iceland bark,Icela Orkney nd
Willow with bark, Orkney
Birch, Finland
Reed canary grass, Sweden
12,0
Reed canary grass, Orkney
10,0
Moisture %
5.6
8.8
7.1
10.8
10.8
11.1
6.5
4.2
8.5
9.1
11
Ash %
0.3
0.4
0.4
0.3
0.9
0.7
1.4
1.3
0.8
7.5
2.4
NCV MJ/kg DM
19.2
19.3
19.2
19.0
19.0
19.2
18.6
18.4
18.9
17.3
17.8
NCV MJ/kg sample
18.0
17.4
17.7
16.7
16.7
16.8
17.2
17.5
17.1
15.5
15.6
S % of DM
<0.01
<0.01
<0.01
<0.01
0.01
0.01
0.04
0.03
0.01
0.11
0.06
Cl % of DM
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
0.05
0.02
<0.01
0.03
0.08
C % of DM
50.6
50.6
50.8
50.2
50.6
50.4
49.1
49.1
49.6
45.5
47.6
H % of DM
6.2
6.1
6.2
6.1
6.2
6.0
6.0
6.0
6.1
5.7
5.8
N % of DM
<0.1
<0.1
<0.1
<0.1
0.1
<0.1
0.6
0.4
0.2
1.0
0.3
O % of DM
42.8
42.8
42.5
43.3
42.2
42.8
42.8
43.2
43.3
40.2
43.8
Al % of DM
0.003
0.006
0.007
0.002
0.01
0.005
0.005
0.01
0.045
0.0007
Ca % of DM
0.068
0.093
0.088
0.048
0.074
0.088
0.33
0.28
0.153
0.38
0.103
Fe % of DM
0.0007
0.0012
0.0018
0.002
0.076
0.009
0.019
0.009
0.008
0.004
0.0037
0.05
0.058
0.06
0.046
0.063
0.045
0.185
0.19
0.111
0.25
0.385
K % of DM Mg % of DM
14,0
0.015
0.021
0.021
0.021
0.031
0.037
0.074
0.05
0.03
0.073
0.086
P % of DM
0.0087
0.011
0.012
0.0099
0.019
0.026
0.105
0.1
0.035
0.14
0.085
Si % of DM
0.0013
0.0025
0.0074
0.025
0.108
0.163
0.029
0.022
0.056
0.37
0.704
IT �C
1530
1530
1530
1260
1160
1140
>1500
>1500
>1500
1530
1080
ST �C
1530
1530
1530
1270
1180
1180
>1500
>1500
>1500
1530
1360
HT �C
1530
1530
1530
1290
1200
1250
>1500
>1500
>1500
1580
1400
FT �C
1530
1530
1539
1320
1240
1250
>1500
>1500
>1500
1590
1480
Ratio K/Si
35.4
23.2
8.1
1.8
0.6
0.3
6.4
8.6
2.0
0.7
0.5
Ratio Ca/Si
52.3
37.2
11.9
1.9
0.7
0.5
11.4
12.7
2.7
1.0
0.1
Ratio K/Ca
0.7
0.6
0.7
1.0
0.9
0.5
0.6
0.7
0.7
0.7
3.7
8,0 Willow Orkney
6,0
Reed canary grass Orkney
4,0 2,0 0,0 Ratio K/Si
Ratio Ca/Si
Figure 25. Comparison of potassium, calcium and silica rations between willow and reed canary grass from Orkney
3.4.2 Net calorific value (NCV)
3.4.4 Ash fusion temperature
To compare the net energy value in different biomass samples the net calorific value (NCV) in the dry matter is usually used. Variation between different biomass materials is mostly depending on variations in ash contents. Higher ash contents are lowering the NCV. In the PELLETime samples, pine has the highest NCV (19.3 MJ/kg) and reed canary grass with high ash content has the lowest NCV (17.3 MJ/kg).
When the ash fusion temperature (AFT) for a fuel according to ASTM D 1857-68 is IT > 1350 °C there are normally no problems with sintering ash when combusted. For the tested materials in PELLETime, resulting values for larch with bark and reed canary grass from Orkney indicates possible sintering problems in combustion.
Table 4. Proximate and ultimate analysis of biomass and ash in PELLETime materials
3.4.3 Chlorine (Cl) 3.4.1 Ash contents To fulfill European standard for class 1 pellet, the ash contents must be ≤ 0,7 % of dry matter (d.m.). As showed in Table 4, this has been obtained for pine and larch without bark. For the willow and the reed canary grass it will not be possible to reach class 1 quality for pellets. This means that there will
20
be difficulties to use this kind of pellets in ordinary small scale pellet burners. All combustion facilities must be dimensioned for high ash contents if these kinds of fuels are going to be used. This, for example include ash cleaning equipments of convection tubes and transport equipment of bottom ash.
Ratio K/Ca
The presence of chlorine in biomass might cause complications when combusted. High values of chlorine might increase the risk of forming low temperature melting ash together with potassium. It might also form hydrochloric acid and contribute to corrosion. Chlorine is also a necessary factor for formation of dioxins during certain conditions of combustion and flue-gas cooling. The analyzed levels of Cl in the PELLETime materials cannot be considered to be cause for any concern.
3.4.5 Potassium, calcium and silica The chemical factors behind ash melting characteristics in biomass are complex, but ratios between potassium, calcium and silica are considered important factors. In general, a high concentration of potassium in relation to silica and calcium is lowering the AFTs and will cause problems with sintering ash and slag formation. On the other hand a high concentration of calcium will raise the AFTs and consequently avoid sintering ash. This is the case with willow from Orkney, which is rich in potassium but even richer in calcium (Figure 25).
21
Considering reed canary grass from Orkney the concentration of potassium is much higher than the concentration of calcium, giving an explanation of the lower AFTs according to the ASTM method. When looking the different treatments of pine, ratios between K and Ca (K/Ca) are almost the same in spite of the variation in concentrations of the elements. This explains why the AFTs are high for all different treatments.
In some of the species the levels of both K and Ca are low and when ratios between them are based on single values, they are impaired with large uncertainties. This may be an explanation to why in the case of Larch from Iceland, the results of mixing bark into the pure stem wood lowers K/Ca-ratio despite the decreasing AFTs.
3.5 Combustion tests Most pellets qualities tested for combustion (pine and larch, with and without bark) were easily combusted with fairly low CO-emissions, especially for larch pellets (Table 5, Figure 26, Table 6, Figure 27). There is also a tendency that admixture of bark increases combustion efficiency (c.e.) expressed as CO2/COratio. This is especially apparent in the case of larch,
whereas with the case of pine, there are too few fuels with added bark to draw clear conclusions. In the case of larch, the increasing c.e. may be a result of lower durability, facilitating oxygen supply in the interior of the pellets. A somewhat higher NCV in pellets with added bark may also contribute.
Pine without bark, clean transport
Pine with bark, clean transport
500 450 400 350 300 250 200 150 100 50 0
Pine without bark, clean transport Pine with bark, clean transport Pine with bark, contaminated
CO, ppm
CO2, %
NO2, ppm
NO, ppm
NOx, ppm
02, ppm
SO2, ppm
CO2/CO
Figure 26. Average values of logged data during pine pellets combustion
Pine with bark, contaminated
Larch
Larch
Larch
+5%bark
+15%bark
Total amount fed fuel, kg
36
43
42
Fuel flow, kg/h
6.0
7.1
7.0
Total ammount fed fuel, kg
48
45
33
Bottom ash (incl. unburnt), g
132
159
241
Fuel flow, kg/h
8.0
7.1
5.4
Bottom ash, % of fuel
0.4
0.4
0.6
Bottom ash (incl. unburnt), g
148
184
198
Unburnt, share of bottom ash, %
45.5
54.6
68.7
Bottom ash, % of fuel
0.3
0.4
0.6
Ash, share of bottom ash, %
54.5
45.4
31.3
Unburnt, share of bottom ash, %
65.2
18.9
18.0
Unburnt, share of fed fuel, %
2.8
3.3
8.7
Ash, share of bottom ash, %
34.8
81.0
81.7
Slag, g
1.0
0.4
0.1
Unburnt, share of fed fuel, %
0.2
0.1
0.1
Slag, share of all ash, %
1.4
0.5
0.1
Slag, g
7.6
123
131
Slag, share of fed fuel, %
0.003
0.001
0.0001
Slag, share of all ash, %
14.7
82.8
81.3
Slag, share of fed fuel, %
0.016
0.272
0.4037
Average values of logged data during sampling
CO, ppm
471
381
283
Average values of logged data during sampling
CO2, %
9.8
9.0
8.2
CO, ppm
152
146
294
NO, ppm
78
72
71
CO2, %
8.6
9.1
9.3
NO2, ppm
4.4
6.5
7.3
NO, ppm
84
113
127
NOx, ppm
82
79
78
NO2, ppm
7.4
0.4
0.3
O2, ppm
10.4
11.2
12.2
NOx, ppm
92
113
126
SO2, ppm
0.0
0.0
0.0
O2, ppm
11.9
11.3
11.0
CO2/CO
233
269
341
SO2, ppm
0.0
0.0
0.2
42
45
32
CO2/CO
580
954
1288
48
46
51
Dust, mg/Nm3 (dry gas, 10% CO2)
Table 5. Data and results from combustion of pine pellets.
Dust, mg/Nm3 (dry gas, 10% CO2)
Table 6. Data and results from combustion of larch pellets
22
23
1400 1200
90,0 80,0
1000
70,0
800
60,0
Larch 600
Larch+5%bark
50,0 %
Larch+15%bark
400
40,0 30,0 20,0
200
10,0 0,0
0 CO, ppm CO2, %
NO, ppm
NO2, ppm
NOx, ppm
02, ppm
Figure 27. Average values of logged data during larch pellets combustion
SO2, ppm
CO2/CO Figure 28. Sintered ash in combustion, percentage share of all ash.
0,45 0,40 0,35 0,30 0,25 %
0,20 0,15 0,10
The ash contents of the combusted fuels are, with the exception of willow, low. For these low-ash fuels the combustion results in amounts of bottom ash well below 1%, including unburned material. The different pine treatments all have low levels of slag formation (Figure 28, Figure 29), which is in line with the previously mentioned high AFTs (Table 4). Since ash contents can vary greatly, in heat production the sintered ash as share of fed fuel is of more interest than its share of ash. For pine with its low ash content, these shares become rather insignificant (Table 5, Figure 29). Pine with added bark show an even lower tendency for slag formation than pine without bark. This can be a result of increased calcium contents due to the added bark (Table 4). The lowering of Ca/Si ratio with added bark is caused by a simultaneous increase of the Si-content, which is too low to have any major effect on the sintering tendency. For larch with added bark, the sintering tendency is very high. Approximately 80% of the ash part of bottom ash sintered (Figure 28) (between a quarter and a half percent of the fed fuel (Figure 29)) and large part of it was forming a solid residue on top of the burner cup (Figure 30). The high sintering tendency of larch with added bark is in line with the lowering of AFTs, mentioned in paragraph 3.4.5
0,05 0,00
Figure 29. Sintered ash in combustion, percentage share of fed fuel.
Figure 30. Slag in burner cup from combustion of larch with 15% bark
24
25
Willow however was not as easily combusted (Table 7). The heat evolved slowly, resulting in low CO2 emissions and frequently re-occurring CO-peaks. Two combustion tests were done where in the second test the boiler and combustion zone were preheated by combustion of commercial wood pellets (Solett). This led to a somewhat better, however not satisfying
combustion performance. The CO2/CO –ratio of the latter was only 58, compared to between 233 and 954 for the other materials. The inefficient combustion of willow can be the result of too hard pellets. NCV for willow is also somewhat lower than for pine and larch (Table 4).
Willow combustion 1 Willow combustion 2
Total ammount fed fuel, kg
30
28
Fuel flow, kg/h
11.9
12.3
Bottom ash (incl. unburnt), g
1015
714
Bottom ash, % of fuel
3.4
2.6
Unburnt, share of bottom ash, %
62.2
60.0
Ash, share of bottom ash, %
37.8
40.0
Unburnt, share of fed fuel, %
2.1
1.5
Sintered ash, g
53.2
18.4
Sintered ash, share of all ash, %
13.9
6.5
Sintered ash, share of fed fuel, %
0.179
0.067
Average values of logged data during sampling CO, ppm
4109
1244
CO2, %
7.2
6.1
NO, ppm
78
62
NO2, ppm
5.1
3.3
NOx, ppm
83
65
O2, ppm
12.9
14.3
SO2, ppm
9.8
0.4
CO2/CO
20
76
Geometric mean
CO, ppm
3878
1024
CO2/CO
19
58
176
176
Dust, mg/Nm3 (dry gas, 10% CO2)
Table 7. Data and results from combustion of willow pellets
26
27
References [1] Martinsson,L. Österberg, S. 2004. Pelletering med skogsbränsle och Salix som råvara.
En undersökning av pelleterbarheten. Rapport 876. Värmeforsk, Stockholm. ISSN 0282-3772.
[2] Näslund, M. 2003. Teknik och råvaror för ökad produktion av bränslepellets. Energidalen, Sollefteå. [3] Timo Lötjönen,T. Phkala, K. et al. Energy from field energy crop- a handbook for energy producers. [4] SS-EN 14774-2:2009 Solid biofuels – Determination of moisture content – Oven dry method – Part 2:
Total moisture – Simplified method
[5] SS-EN 15103:2010 Solid biofuels - Determination of bulk density [6] SS-EN 15210-1:2010 Solid biofuels - Determination of mechanical durability of pellets and briquettes - Part 1: Pellets [7] SS187187:1 Fasta bränslen - Bestämning av halten oförbränt i fasta restprodukter
Contact person PELLETime Sweden: Håkan Örberg, SLU BTK
[email protected] www.btk.slu.se Project websites: www.pelletime.fi
Innovatively investing in Europe’s Northern Periphery for a sustainable and prosperous future
28
European Union European Regional Development Fund Graphic design: Olga Pletcheva, Images: SLU, NKUAS
