Format And Type Fonts CCHHEEMMIICCAALL EENNGGIINNEEEERRIINNGG TTRRAANNSSAACCTTIIOONNSS VOL. 39, 2014 A publication of The Italian Association of Chemical Engineering www.aidic.it/cet Guest Editors: Petar Sabev Varbanov, Jiří Jaromír Klemeš, Peng Yen Liew, Jun Yow Yong Copyright © 2014, AIDIC Servizi S.r.l., ISBN 978-88-95608-30-3; ISSN 2283-9216 DOI:10.3303/CET1439159 Please cite this article as: Boriouchkine A., Sharifi V., Swithenbank J., Jämsä-Jounela S., 2014, A study on the combustion dynamics of a biomass fuel bed in a biograte boiler, Chemical Engineering Transactions, 39, 949-954 DOI:10.3303/CET1439159 949 A Study on the Combustion Dynamics of a Biomass Fuel Bed in a BioGrate Boiler Alexandre Boriouchkine* a , Vida Sharifi b , Jim Swithenbank b , Sirkka-Liisa Jämsä- Jounela a a School of Chemical Technology, Aalto University, P.O. Box 16100, 00076 Aalto, Finland b Department of Chemical and Biological Engineering, The University of Sheffield, Mappin Street,Sheffield,S1 3JD, UK alexandre.boriouchkine@aalto.fi The main objective of this research was to study fuel bed combustion dynamics of a BioGrate boiler with a mechanistic model. First, the fuel specific pyrolysis reaction rates were experimentally determined for the model. Second, the model was validated and finally, it was used to investigate the effects of the primary air flows on drying, pyrolysis and char consumption rates occurring inside the fuel bed. The research results are presented and the role of the dynamic behaviour of the reactions on improving the efficiency of the biomass combustion process discussed. 1. Introduction Biomass is receiving increasing attention as an alternative energy resource to fossil fuels due to increasing environmental concerns related to energy production. However, biomass is associated with varying fuel composition; in addition, it is highly dependent on storing conditions as well as on local and seasonal factors. Such variations affect power production efficiency and increased pollution. As a consequence a fuller understanding of the factors which affect the efficiency of power production is particularly important for the further development of these types of energy systems. Mathematical models have shown to be an excellent tool in assessing the complex physical phenomena of biomass combustion. Shin and Choi (2000) utilized modelling to determine the effect of process parameters on the combustion of municipal waste. The results indicated that a low air flow rate limits the combustion rate, whereas, an excessively high flow rate results in flame extinction. Van der Lans et al. (2000) described straw combustion with a two dimensional steady-state model and studied the effect of the combustion parameters on the burning of the fuel bed. Goh et al. (2000) have developed a model to simulate an incinerator bed; the results from the study indicated that fuel mixing can be modelled with swap probability method. The model developed by Yang et al. (2002) showed that air channelling can increase the concentrations of hydrocarbons in flue gas due to poor mixing. Similarly, a study by Kær (2004) suggested that poor mixing of flue gases and secondary air results in high CO concentrations and in unburnt carbon in fly ash. Miljkovic et al. (2013) constructed a model for straw bale combustion to obtain information such as a temperature profile of the fuel bed, a combustion rate and the produced chemical species. Hallett et al. (2013) found through experimental and modelling work that a volume-surface mean diameter can be used to describe non-uniformly sized particles. Combustion in a conical grate boiler was described mathematically by Boriouchkine et al. (2012) to provide information on the combustion characteristics of woody biomass, with specific emphasis on the effects of moisture content, particle size and air flow on the combustion in a BioGrate boiler. To summarize, these studies focused solely on combustion under steady-state conditions to evaluate the effect of several parameters, like the fuel moisture content and the air flow on the burning behaviour of the biomass fuel bed. However, very little research has been done on the dynamics of the burning fuel bed which is a particularly important issue, as during the operation, a boiler is subject to several dynamic disturbances (Kortela and Jämsä-Jounela, 2012), such as changes in a power demand. Consequently, the 950 performance of a boiler is largely dependent on the behaviour of a burning fuel bed under changing conditions. Due to the importance of fuel bed combustion dynamics, this paper focuses on studying the dynamic behaviour of fuel combustion and, especially, the behaviour of drying, pyrolysis and char combustion reactions as well as the temperature inside the bed. The analysis is done by inducing changes to the primary air flow of different magnitudes and amplitudes, and at different conversion stages to gain insight into the time-dependent combustion behaviour. The simulations considered fuel combustion in a BioGrate boiler which is a conical grate boiler that operates under concurrent combustion conditions. 2. Dynamic Model of a BioGrate boiler furnace In a BioGrate boiler, biomass reacts through three main reaction pathways which can occur either in parallel or sequentially with respect to each other: drying, pyrolysis and char conversion and the mechanistic modelling of the BioGrate boiler considers these three stages of the biomass combustion. The burning fuel bed in a BioGrate boiler is modeled in one dimension using the walking grate assumption, since temperature gradients are more significant in a vertical than in a horizontal direction. Furthermore, since BioGrate boilers operate under concurrent combustion the combustion front propagates upward in the direction of the airflow. 2.1 Modelling of the solid phase mass conservation The solid phase reactions considered by the model include drying, pyrolysis as well as char oxidation and gasification which in general can be described by Eq(1). jsjs rt ,,  (1) where  s,j is the mass concentration of each solid component (moisture, volatiles and char) while rs,j is the reaction rate of a respective component. The rate of drying, Eq(2), is defined by the energy available for evaporation. For numerical stability, the evaporation rate is multiplied by a conversion factor  s,H2O/max( s,H2O). Although this lowers the maximum local evaporation rates, as more energy is transferred to the consequent layers, the overall evaporation in the fuel bed rate remains the same as without the factor. where Ts is the temperature of the solid phase and Hvap is the vaporization enthalpy. The accurate modelling of fuel combustion requires that pyrolysis rate equations are determined for the simulated fuel. In this study, pyrolysis rate equations were estimated for debarking residue produced from Norway spruce. The equations were estimated from mass loss curves obtained from thermogravimetric analysis (TGA) performed in a Perkin Elmer TGA 4000 apparatus under nitrogen atmosphere (20 mL/min) and with the heating rate of 80 K/min. The rate equations were then estimated by fitting the simulated mass loss curve to the one obtained from TGA with sequential quadratic programming. The fitting results suggested that in order to obtain satisfactory fit, the pyrolysis had to be described by two parallel subreactions with the assumption of first-order kinetics. This suggested that a fuel sample comprises two components with distinct decomposition behaviors which are likely to be lignin and holocellulose. The initial mass of each component was estimated along with the rate equations and for the first volatile component it was 15.39 w% and for the second one 84.61 w% of the initial fuel sample. The obtained reaction rate constants are given in Eq(3) and (4). Char is consumed in three different chemical reactions: oxidation, Eq(5), (Branca and Di Blasi, 2003) with the associated CO/CO2 ratio, Eq(6), (Evans and Emmons, 1977) and reduction with H2O, Eq(7) (Senneca, 2007) and CO2, Eq(8) (Matsumoto et al., 2009).  )/(105.114exp101.1 36 , sCs RTk  (5) )/3390exp(3.4/ 2 s TCOCO  (6) )/)/)15.373()max((,0max( 2,2,2, vapspOHsOHsOHs HdtKTCr   (2) ))/(1037.1exp(1016.3 52 ,1 spyr RTk  (3) ))/(1084.7exp(1021.2 44 ,2 spyr RTk  (4) 951    )1ln(1011)/(10136exp1099.9 34 2, XXRTk sOHs  (7)  XRTk sCOs )/(10260exp101.1 39 2,  (8) where X is the conversion stage of char. For each heterogeneous char reaction, an effective reaction constant - given by Eq(9) - is then calculated as follows: where vp is the area to volume ratio of a particle, kr,i is the reaction constant and kc,i is the mass transfer coefficient. 2.2 Modelling of the gas phase mass conservation The gas phase modelling, presented in Eq(10), describes convection of each gaseous specie (O2, N2, CH4, H2, CO, CO2, Tar, H2O), combustion of CO (Eq(11)) (Babushok and Dakdancha, 1993), H2 (Eq(12)) (Pomerantsev, 1986) and CH4 (Eq(13)) (Pomerantsev, 1986), and gas formation in the pyrolysis reaction, Yg,irs,pyr. where g,i is the mass concentration of a gaseous specie, b is the bed porosity, vg is the velocity of the gas, rs,pyr is the pyrolysis reaction rate, while Yg,i is the mass fraction of the forming gaseous pyrolysis product. 2.3 Modelling of the solid phase energy conservation The energy equation for the solid phase describes heat conduction, heat exchange between the phases, energy consumed in the drying and pyrolysis reactions and energy gained in char combustion (Eq(14)):   4 ,, )()( saasolidreactionssgpconvseffssss TkIIkQTTvkxxTktTC     (14) where kconv is the convective heat transfer coefficient, ks,eff is the effective heat conduction coefficient of the solid phase, ka is the absorption coefficient, I + is the radiative flux in the direction of the fuel layer surface, I - is the flux in the direction of the grate, Tg is the gas phase temperature, Ts is the solid temperature and Qreactions,solid is the energy produced or consumed in the solid phase reactions (drying, pyrolysis and char reactions),  is Stefan-Boltzmann constant. The radiative heat transfer inside the bed is described with a two-flux model given by Eq(15) and (16), where the absorption and scattering coefficients are given by Eq(17) (Shin and Choi, 2000) where dp is the fuel particle diameter.   iricpiricpieff kkvkkvk ,,,,,  (9) pyrsigigbiggbig rYrxvt ,,,,, )()(   (10) 5.0 2 5.0 2 314 , ))/(105.125exp(103.1 OHOCOgCOg CCCRTr  (11) 2 14 2, ))/(129exp(1014.2 OgHg CRTr  (12) 2 12 4, ))/(8.103exp(106.5 OgCHg CRTr  (13) 4 21)( sassa TkIkIkkdxdI   (15) 4 21)( sassa TkIkIkkdxdI   (16) 0),ln(/1  sbpa kdk  (17) 952 2.4 Modelling of the gas phase energy conservation The energy continuity equation, Eq(18) of the gas phase considers the heat exchange between the gas and solid phases, the energy received through gas convection, and the energy gained from carbon monoxide, methane and hydrogen oxidation. where Hg is the overall enthalpy of the gas phase and Qreactions, gas is the energy released in the oxidation of CO, CH4 and H2. 3. Model validation with the estimated kinetic parameters The model with the estimated kinetic parameters was evaluated against ignition front propagation velocities presented in Saastamoinen et al. (2000) for fixed bed combustion of spruce wood chips. The validation showed that the model predictions were similar to the measured values presented in Saastamoinen et al. (2000) (Table 1). This shows that the model predictions are valid in a range of moisture contents and airflows. 4. Dynamic behavior of a fuel bed Combustion of a biomass fuel bed provides energy for boiler operation and thus boiler dynamics depend to a large extent on the fuel combustion dynamics. In this study, the combustion dynamics were analyzed with the mechanistic model by changing the rate of primary air flow fed into the burning fuel bed and by analyzing the effects on drying, pyrolysis and char combustion. In simulations, the bed density was 150 kg/m 3 on a dry basis, moisture content 55 w.%, the fuel layer was 0.5 m high, the freeboard temperature 900 °C and the fuel speed 1.5 mm/s. The air flows were varied between 3 m 3 /s and 5 m 3 /s which is the nominal air flow range used in BioGrate boilers. The dynamic responses shown in Figures 1 and 2 indicate a significant effect of the air flow on the amount of the released gas and especially on the amount of released water. The effect of the air flow on the drying rate is also confirmed by the reaction rates presented in Figures 3 and 4 which show that the increase in the air flow notably increases the drying rate. Furthermore, the results suggest that water evaporation is only affected by the current air flow rate and not by the previously used flows as the simulations with two different air flows indicate that evaporation rate dependent only on the air flow rate, but not on the pattern of changes in the flow. The dependence of drying on the air flow was particularly well demonstrated by the simulation with a gradually increasing air feed which showed that the combustion time was significantly shortened compared to the other cases. Such behaviour indicates that the increase in the gas flow also improves the overall heat exchange due to increased convection which, in turn, enhances drying. Similar to drying, the pyrolysis reaction rates also showed the dependence on the air flow, however, the influence was slightly smaller. For instance, the decrease in the air flow from 5 to 3 m 3 /s, which occurred between 0.5 and 1 m from the grate center (Figure 3), decreased the drying rate by 40 % whereas the pyrolysis of the first component was decreased by 30 %. Changes in pyrolysis rates also affected the calorific value of the released gas (Figure 5). Furthermore, the calorific value also demonstrated the dependence on the location of the fuel on the grate and consequently its stage of conversion. Close to the grate center, the effect of air flow on the calorific value was minor: the change in air flow from 3 to 5 m 3 /s   )()( , sgpconvgasreactionsgggg TTvkQxHvtH   (18) Table 1: Comparison between the measured combustion front propagation velocities and the ones predicted by the model Moisture content Air flow Saastamoinen et al. (2000) mm/s Model prediction mm/s 20 mm/ 12.5 mm 10.8 0.07 0.42 0.468/0.476 0.15 0.58 0.519/0.571 0.23 0.47 0.476/0.529 18.8 0.07 0.34 0.380/0.391 0.15 0.47 0.420/0.484 0.23 0.39 0.414/0.426 33.4 0.07 0.28 0.257/0.244 0.15 0.25 0.268/0.292 0.23 0.27 0.277/0.260 953 which occurred between 0.5 and 1 m increased the heat content by 20 % (from 0.208 to 0.250 MW/m 2 ). In contrast, the step changes which occurred between 2.0 and 2.5 m from the centre resulted in a 70 % (from 0.150 to 0.254 MW/m 2 ) increase. The step changes during the char combustion phase also showed a dramatic increase in the calorific value of the gas mainly due to the increased CO production. 1 2 3 4 0.02 0.04 0.06 0.08 A ir f lo w , m 3 /( m 2 s ) Distance from the grate center, m 1 2 3 4 0 0.02 0.04 0.06 G a s r e le a s e r a te k g /m 2 s - 1 O 2 H 2 O CH 4 CO CO 2 H 2 C 3 H 8 Tar 0.5 1 1.5 2 2.5 3 3.5 0.02 0.04 0.06 0.08 A ir f lo w , m 3 /( m 2 s ) Distance from the grate center, m 0.5 1 1.5 2 2.5 3 3.5 0 0.01 0.02 0.03 0.04 0.05 G a s r e le a s e r a te k g /m 2 s - 1 O 2 H 2 O CH 4 CO CO 2 H 2 C 3 H 8 Tar Figure 1: Gas phase composition during the simulation with steps in the airflow Figure 2: Gas phase composition during the simulation with a gradual increase in the airflow 0.5 1 1.5 2 2.5 3 3.5 4 4.5 0 0.02 0.04 E v a p . ra te , k g /m 2 s -1 0.5 1 1.5 2 2.5 3 3.5 4 4.5 0 0.05 0.1 0.15 P y r. r a te 2 , k g /m 2 s -1 0.5 1 1.5 2 2.5 3 3.5 4 4.5 0.01 0.02 0.03 0.04 P y r. r a te 1 , k g /m 2 s -1 Distance from the grate center, m 0.5 1 1.5 2 2.5 3 3.5 0 0.02 0.04 E v a p . ra te , k g /m 2 s -1 0.5 1 1.5 2 2.5 3 3.5 0 0.1 0.2 P y r. r a te 2 , k g /m 2 s -1 0.5 1 1.5 2 2.5 3 3.5 0.01 0.02 0.03 0.04 0.05 P y r. r a te 1 , k g /m 2 s -1 Distance from the grate center, m Figure 3: Reaction dynamics during the simulation with steps in the airflow Figure 4: Reaction dynamics during the simulation with a gradual increase in the airflow 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 0 0.2 0.4 0.6 0.8 1 G a s H H V , e x c l. t a r, M W /m 2 Distance from the grate center, m Figure 5: Calorific values of gas under different air flows 954 5. Conclusions Simulations showed that combustion dynamics are largely dependent on the air flow. All reactions demonstrated a quick response to the changed conditions and as a result, the calorific value of the flue gas quickly reacted to the changes in air flow. This implies that boilers which operate under concurrent combustion conditions can quickly respond to changes in the power production demand. The results also indicated that, the power output of a boiler can be best controlled by regulating air flow 1.5 m from the grate center and especially during the char combustion phase. Furthermore, the dynamics of the system results in a quick response to the fuel bed combustion on the control actions. Based on the simulated behavior of drying, pyrolysis and char reactions, it can be argued that moisture evaporation controls the pyrolysis rate by absorbing most of the energy produced by char combustion thus preventing the temperature from reaching the point where pyrolysis could start. Furthermore, the increased air flow influences the pyrolysis rate less significantly since drying consumes most of the extra energy generated by char oxidation. However, the results indicated that drying rate can be effectively controlled by the air flow and thus the combustion of fuels with high moisture content, which are extensively used in BioGrate boilers, can be intensified by using higher air flow rates. References Babushok V.I., Dakdancha A.N., 1993, Global kinetic parameters for high-temperature gas-phase reactions, Combust. Explo. Shock+ 29, 464–489 Boriouchkine A., Zakharov A., Jämsä-Jounela S.-L., 2012, Dynamic modeling of combustion in a BioGrate furnace: The effect of operation parameters on biomass firing, Chem. Eng. Sci., 69, 669-678 Branca C., Di Blasi C., 2003, Global Kinetics of Wood Char Devolatilization and Combustion, Energ. Fuel., 17, 1609-1615 Evans D.D., Emmons H.W., 1977, Combustion of wood charcoal, Fire Safety J., 1, 57-66 Goh Y.R., Lim C.N., Zakaria R., Chan K.H., Reynolds G., Yang Y.B., Siddall R.G., Nasserzadeh V., Swithenbank J., 2000, Mixing, Modelling and Measurements of Incinerator Bed Combustion, Process Saf. Environ., 78, 21-32 Hallett W., Green B., Machula T., Yang Y., 2013, Packed bed combustion of non-uniformly sized char particles, Chem. Eng. Sci., 96, 1-9 Kær S.K., 2005, Straw combustion on slow moving grates a comparison of model predictions with experimental data, Biomass Bioenerg., 28, 307–320 Kortela J., Jämsä-Jounela S.-L., 2012, Fuel-quality soft sensor using the dynamic superheater model for control strategy improvement of the BioPower 5 CHP plant, Int. J. Electr. Power Energy Syst., 42, 38- 48 Matsumoto K., Takeno K., Ichinose T., Ogi T., Nakanishi M., 2009, Gasification reaction kinetics on biomass char obtained as a by-product of gasification in an entrained-flow gasifier with steam and oxygen at 900–1000 °C, Fuel, 88, 519–527 Miljković B., Pešenjanski I., Vićević M., 2013, Mathematical modelling of straw combustion in a moving bed combustor: A two dimensional approach, Fuel, 104, 351-364 Pomerantsev V., Ed., 1986, Fundamentals of applied combustion theory, Energoatomizdat Publishing House, Leningrad, USSR. (in Russian) Saastamoinen J.J., Taipale R., Horttanainen M., Sarkomaa P., 2000, Propagation of the ignition front in beds of wood particles, Combust. Flame 123, 214-226 Senneca O., 2007, Kinetics of pyrolysis, combustion and gasification of three biomass fuels, Fuel Process. Technol., 88, 87–97 Shin D., Choi S., 2000, The combustion of simulated waste particles in a fixed bed, Combust. Flame, 121, 167–180 van der Lans R.P, Pedersen L.T, Jensen A., Glarborg P., Dam-Johansen K., 2000, Modelling and experiments of straw combustion in a grate furnace, Biomass Bioenerg., 19, 199-208 Yang Y.B., Goh Y.R., Zakaria R., Nasserzadeh V., Swithenbank J., 2002, Mathematical modelling of MSW incineration on a travelling bed, Waste Manage., 22, 369-380