This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright Author's personal copy International Journal of Mineral Processing 112–113 (2012) 13–18 Contents lists available at SciVerse ScienceDirect International Journal of Mineral Processing journal homepage: www.elsevier.com/locate/ijminpro Comminution of forest biomass by modified beater wheel mill in a power plant B. Csőke a, J. Faitli a,⁎, G. Mucsi a, G. Antal a, F. Bartók b a b University of Miskolc, Institute of Raw Materials Preparation and Environmental Processing, Miskolc, Hungary AES Co. Ltd., Berente, Hungary a r t i c l e i n f o Article history: Received 2 April 2011 Received in revised form 23 November 2011 Accepted 15 February 2012 Available online 23 February 2012 Keywords: Biomass comminution Beater wheel mill Mill–classifier cycle Sampling a b s t r a c t Today, renewable energy sources – e.g. forest biomass – are of great importance, not only domestic but also industrial – f.i. wood fired power station – utilization is wide spread as well. However, the comminution of such fibrous texture materials requires relatively high energy and special grinding stress. In the thermal power station of AES Co. Ltd., Berente mainly wood biomass is burned as fuel. The size reduction of biomass is achieved by beater wheel mills, which were designed originally for coal. The aim of the research carried out by the University of Miskolc was to increase the capacity of the fuel supplying system. Systematic industrial and laboratory experimental series were carried out under different conditions and results are presented in this study. During the on-site industrial tests four fuel supplying systems (beater wheel mill, air classifier, heat exchanger, pipes) were equipped with a complete data acquisition system containing different sensors. Three sampling points were built 1) from the feed, 2) from the rejected coarse material after the air classifier and 3) from the final product (isokinetic) from the pneumatic transport pipe. Biomass moisture significantly influenced comminution energy consumption, especially for finer size reduction. © 2012 Elsevier B.V. All rights reserved. 1. Introduction The protection of the environment and economical operation at the same time is a real challenge for companies that are providing power by burning fossil fuels. The interest in producing energy from biomass has increased in recent years, owing to depleting fossil energy supplies and climate change caused by carbon emissions (Reményi, 2007). In the field of biomass-based bioenergy, mechanical size reduction plays a crucial part in producing biomass with appropriate particle size for further processing. The trade of CO2 quota and present legislation had led the thermal power station of AES Co. Ltd., Berente, Hungary to change their fuel into wood biomass from coal. Their original technology was designed for coal and beater wheel mills were applied to comminute the raw fuel and supply the furnaces. This technology is widely applied in coal power plants. The University of Miskolc has carried out an on-site and laboratory research resulted in significant capacity increase. This paper reports about the results of this research (Anon, 2008). The energy requirement for biomass comminution and the resulting particle size are important factors to study grindability characteristics, select equipment, and assess overall efficiency. Esteban and Carrasco (2006) investigated electrical energy demand of biomass comminution. Three different forest biomasses, namely poplar chips, pine chips and pine bark, have been selected for size reduction ⁎ Corresponding author. Tel.: + 36 30 9654420. E-mail address: ejtfaitj@uni-miskolc.hu (J. Faitli). 0301-7516/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.minpro.2012.02.003 study. Four different types of open circuit processes have been designed and assayed utilizing a pilot plant comprising two hammer mills, one screener and one dynamic air separator, in order to obtain the desired fineness for the product. Miao et al. (2011) examined the grinding characteristics of biomass in a Retsch SK100 hammer mill and an SM2000 knife mill. The results showed that the specific energy consumption of biomass comminution and the aperture sizes of the milling screens were related in power-law forms. Biomass moisture significantly influenced comminution energy consumption, especially for finer size reduction. Given a specific milling screen, the hammer mill was found to be more energy efficient than the knife mill. This was mainly attributed to the higher motor speed and axial feeding mechanism of the hammer mill. 2. Background 2.1. The fuel supplying system with a beater wheel mill The circulated flue gas of a wood biomass furnace is used to transport the fuel. Flue gas is discharged from the furnace, then the pre-crushed (b30 mm) wood transported by a belt conveyor is fed into the flue gas pipe through a simple hole or through a lamella feeder. Afterwards, a heat exchanger considerably decreases the flue gas temperature, the flue gas with the fed and crushed wood gets into the beater wheel mill horizontally, after a 45° pipe section at about 160 °C temperature. The beater wheel mill has two basic roles; it is a ventilator and a mill at the same time. This is a considerable disadvantage of this technology because the two functions cannot be controlled separately; on the other Author's personal copy 14 B. Csőke et al. / International Journal of Mineral Processing 112–113 (2012) 13–18 2.2. Characteristics of a closed-circuit system General characteristic of a closed-circuit system is shown in Fig. 2. If the B mill feed is increased, the V capacity of the system increases as well up to a limit value, because the quantity of material in the mill increases. At the point of Bopt optimal charge the mass flow rate of the final product reaches the Vmax value, and in this point recycling material D(B) = K(B) − V(B) (K − mill product), therefore Dopt is the mass flow rate of the recycling material (Kolostori et al., 1979). If the mill feed is higher then the optimal Bopt the mill gets into an unsteady range, the V mass flow rate of the final product and the milling speed decrease, the grinding fineness of mill product, the quantity of D recycling material and B mill feed drastically increase and finally the mill is blocked. This short explanation illustrates how important the proper mill–classifier regulation is. 3. Experimental 3.1. The data acquisition and sampling system The created control and regulation system of a fuel supplying unit is suitable only for the daily operation, but not sufficient if we require finding the way to improve the mill performance. Therefore, complete fluid mechanics data acquisition- and crushed wood sampling systems were installed into each of the four fuel supplying systems. For the determination of the reduction ratio and other data necessary Mill characteristic K(B) V Blockage V K [t/h] Unsteady range hand one machine is sufficient. The mill is a beating–impacting type of mill. Comminution mainly takes place by impacting and by beating in smaller extent. The original armor plate of the mill was modified for wood comminution; ribs were built to improve the shearing effect. The mill and the air classifier (Fig. 1.) operate in a closed-circuit grinding system (Beater Wheel Mill). The product of the mill–classifier cycle is transported by the flue gas, through a heat isolated pipe into the burner in the furnace. One furnace is supplied by four fuel preparing systems, each with a beater wheel mill and a burner. The technology is equipped with a computer control and regulation system. Each fuel supplying system of a furnace is monitored by different transducers; the following parameters are measured on-line: mass flow rate of the feed, electrical current of the driving motor, and temperatures before and after the mill–classifier cycle. Several other parameters are monitored online, but those four and the moisture content of feed measured by sampling periodically are important for the regulation of the fuel supply. If the motor current increases into dangerous levels; the system reduces the feed mass flow, this is the basic principle of regulation. V(B) Bopt B,[t/h] Fig. 2. Characteristic of a closed-circuit system (mill characteristic). to evaluate the mill performance, samples from different points were taken. The whole cross section of fresh wood feed was sampled at the feeder. After the feeder, a conveyor belt transports the crushed wood and coal in appropriate percent into the flue gas pipe. A specially designed sampling vessel was built, where the whole cross section of the material falling down from the belt conveyor was sampled. This sample represents the input of the mill because the heat exchanger does not influence the particle size structure of the material. The next stage of the technology, – after the mill – is the classifier. In Fig. 1, the point of material recycling can be seen, where the coarser particles get into the mill. A window on the mill, and an adequate shaped sampling vessel was built where the whole cross section of the back fed material was sampled. The most difficult sampling task was to solve representative sampling from the product of the mill– classifier cycle. The output of the classifier goes up to a vertical 450 mm diameter pipe section that is connected later into the burner but at a height of about 20 m higher. The sample is taken from this vertical pipe section. For this reason an isokinetic (the velocity of sampling equals to the velocity of flue gas in the pipe on the same axis) sampling system has been developed (Fig. 3). The sampling pipe was connected into a Venturi pipe where the velocity of sampling was measured. The flue gas sample from the sampling pipe was led into a sucked vessel, in which a heat resistant filter was installed. The maximum flue gas temperature was about 160 °C. In the outer and inner sides of the sampling pipe static pressure sensing holes were made. If the static pressure at a height at the outer side equals the one at the inner side the flow velocities have to be equal as well, because of the Bernoulli equation. That means sampling is isokinetic. The revolution number of the ventilator Air classifier outlet Air classifier Feed Mill inlet Recycled material Armor plates Rotor Fig. 1. The beater wheel mill and air classifier. Parts: rotor, armor plates, air classifier opening, air classifier plates, recycled material, feed, blow bar. Author's personal copy 15 B. Csőke et al. / International Journal of Mineral Processing 112–113 (2012) 13–18 Heat resistant filter Ventilator Holes for measuring static pressures to check isokinetic sampling. Fig. 3. Schematic of the isokinetic sampling system. shaft has to be regulated. If this revolution number setting is correct, the static pressures in the two holes are equal. A fluid mechanics computer data acquisition system was temporarily installed. The distribution of static pressure along the fuel supplying system was measured by piezo electric pressure transducers (0–2000 Pa range). The static pressure was measured between the mill and the classifier as well. The velocity profiles of different pipe sections were measured by a Prandtl tube and a hand pressure instrument during static investigation. During dynamic research, flow velocities at the center point of different pipe sections were measured by some Prandtl tubes and piezo pressure sensors. The outputs of the amplifiers of different sensors were connected into a computer AD card, the measuring software was written in C++. 3.2. Industrial and laboratory measurements To obtain general information about the four fuel supplying systems the investigation started with systematic load tests. It means that after the complete instrumentation of a unit, tests started by setting up a low load (typically 4 t/h) and after the unsteady state, continuous computer monitoring and systematic sampling were performed. Each sampling was repeated three times to check reproducibility. After the complete measurement of a selected load, it was increased to the next level and tests were accomplished by the same protocol. This procedure was continued until the maximum load (typically 5–7.4 t/h depending on the unit and surrounding conditions) was reached, namely the beater wheel mill was blocked. During the laboratory measurements the milling of wood biomass utilized in the power station was carried out in cutting and hammer mills under different conditions (moisture content) in order to discover the relation between specific grinding work and moisture content of wood feed. Three experiments were performed with different moisture contents (0, 17 and 36%) in both mills. The moisture content of the samples was adjusted as follows: 0% of moisture content was produced in a drying cabinet until constant weight. 17% was reached similarly by continuous control of the moisture content until the required value. The highest value was the initial moisture of the sample. Additionally the effect of mill sieve was investigated as well. The apparatuses applied for grinding experiments were laboratory size cutting- and hammer mills equipped with digital energy meter to register grinding work. Both mills were operated with bottom discharge (gravity fall) and horizontal axis. During the grinding the operating parameters that affected the milling were constant, in this way the effect of the moisture constant of the biomass could be monitored properly. Wgrinding ¼ ∫ðPðtÞ−P0 Þdt ð1Þ where P(t) is the measured electric instantaneous power and P0 is the no-load power. Hence Wspec the specific grinding work is as follows: Wspec ¼ Wgrinding : m ð2Þ where m is the weight of material being ground. 4. Results and discussion 4.1. Industrial measurements Load tests of the four units had resulted in several data, some observation follows. The design and physical construction of the four units were the same, however, they performed differently. In respect to fluid mechanics, the static pressure distribution along the technology line, flow rate of fuel laden gas (0–10,000 m 3/h), and fuel concentration (0–350 g/m 3) were similar, but grinding fineness was different. The rate of size reduction concerning 50% particles was between r50 = 2.9–4.5, for 80% particles it was r80 = 2.2–3.3. Particle size distributions of feed, recycling from classifier and product of a given working condition and unit are shown in Fig. 4. The quality of the feed material was almost constant during the load tests, for example moisture content was 35% with less then 1% deviation based on 36 analyses. Moisture content of the product in the pipe before the burner was 18–21%. Concerning the particle size distribution the aim of the power plant is to comminute the wood fuel below 5 mm; the operation of the furnace was designed for such fuel (fluidized bed). It was found that even the feed had already contained about 20% b 5 mm and the recirculated material from the Mill No. 7/2, before research 100 Particle size distribution, F(x) [%] Venturi pipe The consumed electric power during the experiments was measured by the microcomputer controlled digital energy meter Carlo Gavazzi WM1-DIN for the characterization of the grindability. Besides the immediate power, the work required for the grinding, the electric current (I), the voltage (U) and cosφ can be measured as well by this device. The digital energy meter recorded the electric work in the cumulated form, in this way the grinding work could be calculated as the difference of the initial and the final value. The measured grinding work can be determined as follows: 80 60 Feed material x50 =11,07 mm, x80=16,78 mm F(x<5 mm)=12,84 % Product x50 =2,92 mm, x80 =5,23 mm F(x<5 mm)=78,03 % Recirculated material x50 =5,58 mm, x80 =9,82 mm F(x<5 mm)=43,55 % 40 20 0 0,1 1 10 100 Particle size, x [mm] Fig. 4. Particle size distributions of feed, recirculated from classifier and product of unit 7/2, load 4.2 t/h, before modification. Author's personal copy 16 B. Csőke et al. / International Journal of Mineral Processing 112–113 (2012) 13–18 40 8 2500 360 35 2000 6 Load [ t / h ] 20 1000 15 320 Load [ t / h ] 1500 Motor current [ A ] Flow velocity [ m/s ] 10 4 280 2 Motor current [ A ] 25 Static pressure [ Pa ] Flow velocity vIK [ m/s ] 30 240 500 5 Static pressure [ Pa ] 0 0 200 0 42000 44000 46000 48000 50000 Time [ sec ] Fig. 5. Dynamic examination of a fuel supply system. classifier into mill contained about 30–40%, smaller then 5 mm fraction. The obvious conclusion would be the insertion of a separate classifier to cut fines from the feed and more importantly this data indicated that the applied classifier did not work properly. Another observation was that with increasing load the particle size distribution of the product firstly starts to be finer, then after its peak it becomes coarser. Supplementing the on site industrial data a laboratory experiment with a hammer mill was carried out to examine the effect of moisture content onto grindability of wood. All other milling parameters were kept constant and only the moisture content was changed by drying and wetting different moisture content samples that were produced and comminuted in a laboratory hammer mill. As was expected, 100% of the completely dried wood sample was milled below 8 mm after 90 s. After 90 s milling, coarse (>8 mm) fraction was removed and milling was continued in the case of wet samples. Even some parts of the material of some samples were not recovered completely, because fine wet particles stick onto the machine. In the concluding results of the load tests, many suggestions were formulated: 1) in the industry there are beater wheel mills with a separate pre comminution machine, a hammer mill built into the door of the mill. 2) It was also suggested that blades of the ventilator– mill machine should be made from a harder material with designed cutting edge. According to our suggestions the aperture of the recycling material from the classifier into the mill was reconstructed to be bigger. The gap between the rotating plates and the armor of the spiral shaped house was constructed to be narrower. The structure of the classifier was considerably modified. After the modifications, newer load and dynamic tests had been carried out to check the effect of changes. During these tests a quite interesting phenomenon was discovered. In the beginning it was mentioned that the regulation strategy of such a feedback technology Mill No. 7/2, after modification Hammer mill, sieve: 8 mm Hammer mill, sieve: 12 mm Cutting mill: sieve: 7 mm Cutting mill: sieve: 20 mm 80 60 10 Feed material x50=11,82 mm, x80=18,89 mm, F(x<5 mm)=11,53 % Product x50=2,39 mm; x80=3,82 mm; F(x<5 mm)=87,74 % Recirculated material x50=4,38 mm; x80=7,97 mm; F(x<5 mm)=55,98 % Median of product, mm Particle size distribution, F(x) [%] 100 40 20 0 0,1 1 10 100 Particle size, x [mm] 0 17 34 Moisture content, % Fig. 6. Particle size distributions of feed, recirculated from classifier and product of unit 7/2, load 4.2 t/h, after modification. Fig. 7. Median of product as function of moisture content. Author's personal copy 17 B. Csőke et al. / International Journal of Mineral Processing 112–113 (2012) 13–18 Table 1 Median of the laboratory mill products. Table 2 Specific grinding work of the laboratory mills. Median of the product, mm Hammer mill Specific grinding work, kWh/t Cutting mill Hammer mill Cutting mill Moisture content, % Sieve size: 8 mm Sieve size: 12 mm Sieve size: 7 mm Sieve size: 20 mm Moisture content, % Sieve size: 8 mm Sieve size: 12 mm Sieve size: 7 mm Sieve size: 20 mm 0 17 34 3.58 3.66 3.77 4.88 5.54 5.42 3.41 4.14 3.65 6.44 9.59 8.43 0 17 34 16.25 17.50 18.75 6.25 6.50 7.50 8.75 11.25 12.50 1.25 2.50 3.75 is very important. Operators in the control room of the plant set the capacity of fuel supply in accordance to the contracted electrical power. The fuel supply system is operated at safe levels and it means at low levels because sometimes after many hours of stable operation at a given load an accidental mill plug can happen. If the current of the motor increases the control system automatically decreases or ceases the feed. Therefore, dynamic tests had been carried out to examine the system behavior before blockage. Fig. 5 shows the main parameters before a point where the automated system stopped the feed because of high electric current. At this time fluid mechanics parameters were monitored as well. In Fig. 5, parameters are shown as a function of time; the feed is shown by the thick line. The feed was increased gradually and at the point of about 48,000 s, the current suddenly increased, and the feed was stopped. The current increase showed the blockage condition at the last moment, but the trend line of fuel laden gas velocity or static pressure at a given position shows it much earlier. Trends of the velocity and static pressure are practically identical. Conclusion of the observed phenomenon is that a control signal should be the easily measurable static pressure and if at a given load its trend is decreasing later blockage should be expected. As a result of the shortly described modifications at the end of the research, capacity has been increased by about 15% and product fineness has been improved considerably as well. Fig. 6 shows the particle size distributions of different material flows after modifications for the same unit (Mill 7/2) and load of this data was showed in Fig. 4 earlier. Finer product and recirculating material can be seen well. 4.2. Laboratory measurements The effect of moisture content of biomass and mill sieve size on the material fineness characterized with median of the particle size distribution of the ground material is presented. The experimental results can be seen in Fig. 7, where the moisture content of wood is plotted as function of median (see details in Table 1). It can be observed that the mill product is coarsening with increasing moisture content in both mill types. However, while almost linear relation is found in the case of the hammer mill (moderate increase), the curves of the cutting mill have the maximum point at 17% of moisture content. The most significant difference was observed at 20 mm mill sieve size. The above mentioned phenomena can be explained by the different grinding stresses of the mills. Namely the hammer mill operates with impact and shear forces, and the cutting mill mainly with shearing and cutting. Consequently, based on the experimental results the cutting mill is more sensible for the moisture content of the mill feed than the hammer mill. On the other hand, using the cutting mill finer product particle size can be reached using the appropriate settings. Concerning the correlation between the specific grinding work and moisture content of the biomass, it was observed that the energy demand increases as function of moisture content in the case of the applied mills (Fig. 8). However, a more significant difference was found in the cutting mill (Table 2). In Fig. 9 the average specific grinding work is demonstrated for the same circumstances of the industrial beater wheel mill as well as the laboratory cutting- and hammer mills. It can be seen that only a slight difference was found between the cutting mill and the beater wheel mill (6.58%) compared with the difference of the industrial mill and the laboratory hammer mill, where a remarkable difference (40.13%) was measured in grinding energy. 5. Conclusions During the industrial tests four fuel supplying systems (beater wheel mill, air classifier, heat exchanger, pipes and burner) were equipped with complete data acquisition and sampling systems containing different sensors and specially designed sampling places and devices. Based on the results of systematic load tests many suggestions have been taken to modify the technology. One of these recommendations was a separate pre comminution machine, where a hammer mill is built into the door of the mill. Secondly, it was suggested that blades 20 Hammer mill, sieve: 8 mm Hammer mill, sieve: 12 mm Cutting mill: sieve: 7 mm Cutting mill: sieve: 20 mm 16 14 12 10 8 6 4 18 Specific grinding work, kWh/t Specific grinding work, kWh/t 18 16 14 12 10 8 6 4 2 2 0 0 5 10 15 20 25 30 35 Moisture content, % Fig. 8. Specific grinding work of laboratory mills under different conditions. cutting mill (lab) beater wheel mill (industrial) hammer mill (lab) Mill type Fig. 9. Specific grinding work of different mills. Author's personal copy 18 B. Csőke et al. / International Journal of Mineral Processing 112–113 (2012) 13–18 of the mill rotor might be made from harder material with designed cutting edge. The structure of the classifier was considerably modified. Experiments with dynamically loading the fuel supply system has resulted in a great observation: the trend of the easily measurable static pressure or flow velocity at a given place is suitable to predict the blockage of the system much earlier than the monitoring of the driving motor electric current. The introduction of the above mentioned modifications has resulted in significant capacity increase and finer product. Based on the laboratory experiments, it can be established that the cutting mill is more sensible for the grinding circumstances – namely the moisture content of the mill feed and the mill sieve size – than the hammer mill. Acknowledgments The described work was carried out as part of the TÁMOP 4.2.1.B 10/2/KONV 2010/0001 project in the framework of the New Hungarian Development Plan. The realization of this project is supported by the European Union, co-financed by the European Social Fund. References Examination of possibilities of capacity increase of mills of the Berente Power Plant. Final report. University of Miskolc. Beater Wheel Mill. (prospectus) ALSTOM Power Boiler GmbH. Esteban, L.S., Carrasco, J.E., 2006. Evaluation of different strategies for pulverization of forest biomasses. Powder Technol. 166, 139–151. Kolostori, J., Petho, Sz., Hilger, M., 1979. Korszerű cementipari technológiák (Up-to-Date Technologies of Cement Industry). Budapesti Műszaki Egyetemi Kiadó, Budapest. Miao, Z., Grift, T.E., Hansen, A.C., Ting, K.C., 2011. Energy requirement for comminution of biomass in relation to particle physical properties. Ind. Crop. Prod. 33, 504–513. Reményi, K., 2007. Megújuló energiák (Renewable Energies). Akadémiai Kiadó, Budapest.