POREĐENJE D I 3D MODELIRANJA TRANSPORTNIH PROCESA PRILIKOM SAGOREVANJA BALIRANIH POLJOPRIVREDNIH OSTATAKA A. M. Erić *, S. Đ. Nemoda *, M. S. Komatina **, D. V. Dakić ***, B. S. Repić *, M. R. Mladenović * * Univerzitet u Beogradu, Institut za nuklearne nauke Vinča, P.O. Box 5, 111 Beograd, Srbija ** Univerzitet u Beogradu, Mašinski fakultet, Kraljice Marije 16, 111 Beograd 35, Srbija *** Univerzitet u Beogradu, Mašinski fakultet, Inovacioni centar, Kraljice Marije 16, 111 Beograd 35, Srbija Apstrakt: U radu su prikazani rezultati poređenja D i 3D CFD modeliranja sagorevanja balirane poljoprivredne biomase u ložištu koje radi po principu cigaretnog sagorevanja. U obzir su uzeti složeni fizičko hemijski procesi koji se odvijaju u bali i prostoru oko nje. Bala poljoprivredne biomase je tretirana kao porozna sredina, pa proračuni uključuju zakone održanja toplote i supstancije u poroznoj sredini. Prostor oko bale je tretiran kao fluidna sredina u kojoj su primenjeni dobro poznati načini modeliranja transportnih procesa. Modeli su opisani setom parcijalnih diferencijalnih jednačina kojima se definišu procesi prenošenja toplote i supstancije u poroznoj i fluidnoj sredini. Cilj istraživanja je analiza rezultata dobijenih korišćenjem D i 3D CFD modela kako bi se sagledao uticaj pojednostavljenja i pretpostavki uvedenih u D model, na tačnost dobijenog rešenja. Verifikacija razvijenih modela je izvršena pomoću složenih eksperimentalnih istraživanja sprovedenih na eksperimentalno-industrijskom kotlu od 1,5 MW koji služi za zagrevanje plastenika u Poljoprivrednom kombinatu Beograd. Rezultati pokazuju da se D model sa zadovoljavajućom tačnošću može koristiti za opisivanje kompleksnih transportnih procesa koji se dešavaju prilikom sagorevanja balirane biomase u razvijenom ložištu. Ovo saznanje može biti od velike koristi prilikom korišćenja numeričkih simulacija u cilju optimizacije rada ložišta, jer korišćenje D modela u značajnoj meri smanjuje potrebne resurse i vreme proračuna. Ključne reči: balirana biomasa, sagorevanje, D i 3D modeliranje
COMPARISON OF D AND 3D MODELING OF TRANSPORT PHENOMENA WITHIN BALED AGRICURTURAL RESIDUES COMBUSTION A. M. Erić *, S. Đ. Nemoda *, M. S. Komatina **, D. V. Dakić ***, B. S. Repić *, M. R. Mladenović * * University of Belgrade, Vinca Institute of Nuclear Sciences, P.O. Box 5, 111 Belgrade, Serbia ** University of Belgrade, Faculty of Mechanical Engineering, Kraljice Marije 16, 111 Belgrade 35, Serbia *** University of Belgrade, Faculty of Mechanical Engineering, Innovation Centre, Kraljice Marije 16, 111 Belgrade 35, Serbia Abstract: This paper presents a comparison of results of D and 3D CFD modeling of baled agricultural biomass combustion in the furnace which operates on the cigarette combustion principle. The complex physical and chemical processes occurring in the bale and the area around it are taken into account. The bale of agricultural biomass is treated as a porous medium, so calculations include laws of the conservation of mass and heat transfer in porous media. The area around the bale is presented as a fluid environment in which are applied well-known transport processes modelling. The models are described by a set of partial differential equations which define the momentum, heat and mass transfer processes in the porous and fluid system. The aim of this investigation was analysis of results obtained using D and 3D CFD models for investigating the influence on the accuracy of the final modelling results due to approximations that are introduced in the D modeling. The verification of the developed numerical models was performed through a comprehensive experimental research conducted on an experimental-industrial plant i.e. on the 1.5 MW boiler for heating the greenhouses in the Agricultural Corporation in Belgrade. The results have shown that the D model with satisfactory accuracy can be used to describe the complex transport processes during combustion of baled biomass in the proposed furnace. This knowledge can be of great benefit to the use numerical simulation for the furnace operate optimization because the use of a D model significantly reduces the required calculation time and the resources required for the simulation. Key words: baled biomass, combustion, D and 3D modeling
1. INTRODUCTION In the world currently about 8% of primary energy comes from fossil sources that are primarily oil, coal and natural gas. As these sources of energy are the non-renewable, it is clear that their depletion is expected in the near future (next 5 years). In addition, fossil fuels have a major impact on the process of global warming, and there is a major initiative of the European Commission to reduce their use. Therefore the European Union has set targets by. year to reduce emissions of greenhouse gases by %, increasing the share of renewable energy to % of final energy consumption and increase energy efficiency by %. As in the European Union and Serbia there is significant biomass energy potential the development of its use is an important step towards achieving the stated goals. For these purposes, a 1.5 MW hot water boiler was constructed and installed in the Agricultural Corporation Belgrade. The boiler is based on baled agricultural straw combustion and it is used for heating 1ha greenhouses. Combustion in the boiler is carried on "cigarette" principle [1, and 3] where.7 1.. m straw bales are used as fuel. Figure 1. Boiler scheme The scheme of the boiler is shown in Figure 1. Baled straw is gradually introduced into the combustion chamber by a hydraulic system. Combustion process takes place only on the part of the bale, which is located in the furnace. Combustion air is divided into three parts. The first part of the air is introduced surrounding the straw bale, which is located in the furnace. The second part of the air is passed through the ash storage located under the grate, and the third part of the air is introduced through secondary air inlet directly into the char combustion zone. This inlet facility has the ability to translational and partially rotational movement. By translator movement furnace
power is regulated (allowing the smaller or larger amounts of biomass involved in the combustion process), and by rotational movement the ash is removed from top of the bale. Combustion process takes place only at the top of the bale and combustion of unburned biomass completes within the fluidized ash area, at the bottom of the furnace. Flue gases produce in the primary zone after entering the secondary and tertiary zone where the combustion process is completed, enter the part of the heat exchangers and finally leave the plant through the chimney. The proposed technology of combustion is a relatively new and insufficient researched, therefore is necessary to continue development activities. Experimental studies on a semi-industrial plant can be complicated and not enough detailed, so it is appropriate to develop a sufficiently accurate CFD model that can simulate the processes occurring in the combustion chamber. Numerical simulation of such systems involves modelling the momentum, heat and mass transfer process of the combustion of baled biomass which have a porous media structure [4]. In the development of the proposed mathematical model of combustion except porous media thermophysical properties [4, 5, 6], it is very important to know the input (boundary) values, and output parameters because of its verification. Numerical simulation of complex processes can be carried out using D and 3D mathematical model. Of course, the choice between D and 3D modeling approach depends primarily on whether the case has a two or three-dimensional character. A two-dimensional mathematical model is more suitable from the standpoint of saving time and resources needed to simulate the operating regimes, but require certain simplifications that may affect the accuracy of the results. For the observed simulation in the cigarette burning furnace may be applied D approach with the certain approximations. Which model will be used during the further researches depends on the accuracy that can be achieved by D model compared to the 3D model. This paper presents a comparison of results of D and 3D CFD modeling of baled agricultural biomass combustion.. D AND 3D MODEL DESCRIPTION The proposed numerical model takes into account momentum, heat and mass transfer processes occurring during combustion of baled biomass, which have a porous structure, with surrounding fluid. Therefore the processes occurring in the boiler needed to be modelled both with respect to processes occurring in the porous media (biomass bale) and processes occurring in the fluid media (Figure ). Based on the literature review [1, 3], in order to accurately describe changes of three primary parameters (velocity, concentration and temperature), the momentum, heat and mass transfer equations have been defined. In order to eliminate the factors that were deemed to have very little
3 1 5 1 impact on the analyzed processes certain assumptions and simplifications have been adopted. The first assumption was that the model is stationary with constant position of the combustion front. Porous layer i.e. the biomass fuel was assumed to be continuously fed into the combustion chamber, whereby all processes were considered stationary due to the fact that the biomass feed rate was equal to the moisture, volatile and coke residue source rate. Fuel feeding process was simulated by volumetric sources of different fuel components present in the porous zone (moisture, volatiles and coke residue), defined based on the known boiler heat output. Combustion process in porous and fluid media was assumed as homogenous with Arrhenius equation utilized to describe the constant rate of chemical reactions. Volatile composition of the biomass was presented by three gases: propane, carbon dioxide and water vapour [4]. Energy conservation in the porous medium was modelled by solid-and-liquid-phase-temperature-balance model i.e. by one-equation, single-phase model. Transport processes are modelled assuming that the flow regime in both environments is turbulent, and turbulence is modelled using widespread and generally accepted k-ε model. 1 Outlet 5 Inlet Inlet 4 Inlet 3 Inlet 1 5 1 1 Figure. Modeled furnace scheme Outlet Y Z X Inlet, m1 Inlet, m4 D model Inlet, m3 Figure 3. Geometry of the models 3D model Inlet, m3
The geometry of modelled area in D and 3D models is presented in Figure 3..1. Velocity field Description of the velocity field is based on continuity and moment equation. The velocity through the porous media substantially corresponds to the velocity in the fluid environment [4, 8, and 9], so that it can be defined the continuity equation which applies to the porous and the fluid environment in the well-known standard form. f ρ W S (1) c where S c is the source which is defined by boundary conditions and W is the velocity. To define a velocity field it is also necessary to define the momentum equation, which is different for the fluid and porous media. The momentum equation for a surrounding fluid has the well-known form, but the momentum equation within the porous media includes an additional term representing the pressure drop due to fluid flow through porous media. The pressure drop in momentum equation ( in Eq.) is usually defined by Forchheimer's equation that for higher velocities implies a nonlinear dependence between fluid velocity and pressure gradient [4, 8 and 9]. In accordance with the said momentum equation for the system porous media - free flow of fluid, has the following form: eff μ ρ K K 1 eff f ρ f W W μ W p W W W where the turbulent stresses are determined by k-ε model. Turbulent viscosity coefficient in the porous area is assumed that is equal to the coefficient of turbulent viscosity in the free fluid zone. The Forchheimer s equation coefficients K 1 and K are experimentally determined [5, 6 and 1]. ().. Species concentration fields The character of combustion has been observed as a quasi-homogeneous, i.e. boundary conditions sources (in each computational point) of moisture and volatiles were determined experimentally with calculating the homogeneous combustion process of the volatiles and coke residue. Transport equations of chemical species, which defines the species concentration field, as in the case of defining momentum conservation equation for a fluid flow, has a common form with diffusion, convection and source term. Macroscopic examination of transport processes in porous media enables the use of an effective diffusion coefficient, and thus the chemical species transport equation can be expressed as:
ρfw Yk ρf D Y εr (3) eff k k where D eff is the effective diffusion coefficient. Due to negligible turbulent range and low values of dispersed turbulent diffusion, the effective diffusion coefficient (in Eq. 3) can be equated with the effective diffusion coefficient in the fluid flow. Assuming that the chemical reactions take place only in the cavities of porous layer, the source term of this equation is multiplied by the porosity ε. The form of the source term R k depends on the combustion process rate, as well as the stoichiometry of the process. For the homogeneous reaction according [11] the modified Arrhenius's expression for the conversion components rates is often used. Source term of the chemical species conservation equations presented in the Table 1. Table 1. Source term of the chemical species conservation equations. Species Source, R k CH R 3 8 CH 3 8 M M M CO CO CO CO 3 R R R R 3 8 M M M O HO C C H CO CO C C3H8 CO C 7 M M 1 M 1 M M M M M M 4 M RC HO CH 3 8 O O O O R R R R C3H8 CO C CO C3H8 CO C CO R CH 3 8 CO M CO R CO R CO M CO For the purposes of the volatiles combustion modelling in accordance with the assumptions related on their composition, the two-stage reversible reaction of the propane combustion was adopted and accordingly the following model can be set: 7 C H O 3CO 4H O 3 8 1 CO O CO 1 CO CO O E1 d C H 3 8 RT,1 1,65 g R k e C C H O 3H8 o,1 3 8 dτ E RT,5,5 g o, d CO R k e CO O H O CO dτ E3 d CO RTg R k e CO CO o,3 The char combustion is also modelled by two-step reaction: 1 C O CO dτ E4 RTg o,4 d C R k e C O C dτ
.3. Temperature field In a similar way as in the case of species concentration fields the energy conservation equation in porous media has been defined. In this case, the porous media is also considering as a fluid where the characteristics of porous media are introduced through the effective thermal conductivity. This consideration corresponds to single phase model [4] that implies thermal equilibrium between the solid and gaseous phases. The expression of energy conservation equation for porous media can be defined as: μ t ρcpw T λeff T ρ Dm, k cp, kt Yk ε R H q k k r ρσ (4) t The effective thermal conductivity of the fluid flow through porous media is analyzed in detail in the literature [4, 11 and 1] where is highlighted the importance of two terms: stagnant thermal conductivity and dispersed thermal conductivity. Neglecting a disperse part of effective thermal conductivity (because at high temperatures the stagnant thermal conductivity is dominant) the effective heat transfer coefficient is defined as follows: λ λ ε λ 1ε eff f s where λ s is the coefficient of thermal conductivity of the porous solid matrix layer, or in this case, soybean residue, whose value is determined by experimental research. The last term in the energy conservation equation is the radiation energy transfer within the gas flow which is defined by P1 model which is part of FLUENT 6 CFD package. Partial differential equations that constitute the mathematical model used for simulation the combustion of baled soybean residue are non-linear and mutually coupled. The numerical procedure of solving the equations has been performed by using the control volume method, including the collocated numerical grid for momentum equations, hybrid numerical scheme (the combination of upstream and central differencing) and SIMPLE algorithm for solving the equations. The iteration process stabilization is done by sub-relaxation technique. The calculation procedure and the numerical method are described in more details in [7]. (5) 3. RESULTS AND DISCUSSION In order to compare the results obtained by the model experimental research are carried out. Experimental studies were conducted on industrial-demonstration facility and included measuring the air mass flows at the entrances, flue gas temperature on the exit of the fluidized bed of ash, fuel mass flow, the flue gas temperature at the outlet cross-section of the modeled area and the dry flue
gas composition on the outlet cross-section of the modeled area. Special attention is paid to the temperature field determination in the central cross section of bales of biomass residue, i.e. in the intense combustion zone. For this experiment, four thermocouples were installed in height of the central vertical plane, according to the scheme of Figure 4. This experiment was conducted in the furnace stationary regime and temperatures were measured in the function of thermocouples position. Fig. 4. Measurements scheme During the experimentation goal was to achieve power of 1.5 MW for optimum combustion. This was achieved by proper arrangement of mass flow of air and fuel at the entrance, which are presented in Table and used as boundary conditions in proposed models. Table. Boundary conditions on the front sections of the model Mass flow, [kg/s] Mass flow of Unit Inlet 1 Inlet Inlet 3 Inlet 4 fuel, [kg/s] Value.3486.3843.3647.455.11143 Results obtained by experimental measurements and results obtained by numerical analysis using the models are shown in Table 3. Table 3. Results from experimental measurement and D and 3D models Value Unit Experiment D model 3D model Temperature on outlet o C 889 896 893 CO % vol 7.46 7.46 7.5 Concentration O % vol 11.7 11.96 11.85 (in dry flue gas) CO ppm vol 548 55 53 NO ppm vol 161 17 175
Temperature, [ o C] Temperature, [ o C] Temperature, [ o C] Temperature, [ o C] Table 3 shows that the results obtained by both models are in very good agreement with measured outlet temperature and the composition of dry flue gas. Also has been observed a satisfactory agreement between modeled and measured temperature profile along the bale middle cross section (Figure 5). 11 1 9 7 5 3 1-1 -1-4 6 8 1 1 14 16 18 Position, [cm] y=15cm T1 experiment T1 D model T1 3D model 11 1 9 7 5 3 1 1 9 7 5 3 1-1 -1-4 6 8 1 1 14 16 18 Position, [cm] y=75cm T experiment T D model T 3D model 1 9 7 5 3 1 11 1 9 7 5 3 1-1 -1-4 6 8 1 1 14 16 18 Position, [cm] T3 experiment T3 D model T3 3D model 11 1 9 7 5 3 1-1 -1-4 6 8 1 1 14 16 18 y=45cm y=15cm Figure 5. The temperature along the central cross-section of bale 1 13 1 11 1 9 7 5 3 1 Position, [cm] T4 experiment T4 D model T4 3D model 1 13 1 11 1 9 7 5 3 1 From the Figure 5 the satisfactory agreement of the measured and modelled temperature fields within bale can be observed, especially in positions y=75cm and y=45cm, while in the positions y=15cm and y=15cm there are some discrepancies in the range of 1-15cm from the end of combustion zone. Higher temperature in the planes closest to the upper and lower surface of bales, in the positions just before entering the combustion chamber, can be explained by the bale and air input ducts are not complete seal which leads to the diffusion of oxygen, where the temperature in this region is high enough to create a favorable conditions for combustion. This phenomenon can be overcome by construction changes in the air supplying system. Temperature profiles in the middle cross-section obtained by D and 3D model are shown in Figure 6.
Y Y 1.5 1.5.5 1 1.5.5 X T, [K]. 1933.3 1866.7 1. 1733.3 1666.7 1. 1533.3 1466.7 1. 1333.3 166.7 1. 1133.3 166.7 1. 933.3 866.7. 733.3 666.7. 533.3 466.7. D model 3D model Figure 6. Temperature profile in the middle cross-section Z X T [K] 3 1 19 1 17 1 15 1 13 1 11 1 9 7 5 According analyzes presented in this paper it can be concluded that the D numerical model of baled biomass combustion can be used with sufficient accuracy for the analysis and optimization process of the furnace. Use D model greatly reduces the time and numerical calculation resources in comparison with 3D model. 4. CONCLUSION The aim of this study was to analyze the results obtained using D and 3D CFD models to investigate the impact of approximations introduced in the D modeling on the accuracy of the final results of modeling. The verification of the developed numerical models was performed through a comprehensive experimental research conducted on an experimental-industrial plant i.e. on the 1.5 MW boiler for heating the greenhouses in the Agricultural Corporation in Belgrade. The mathematical models are established in such a way that is enabled the simulation of combustion in stationary operation regimes in furnace. By the proposed models it is possible to calculate the three groups of (D and 3D) fields: velocity, temperature and components of combustion reaction (chemical species). The velocity distribution in the analyzed area is described by the set of Navier -Stokes equations- with the turbulent Reynolds-stress resolving using standard k-ε turbulence model. Fluid flow in porous environment also has been described by the set of Navier-Stokes equations modified taking into account the specifics of porous media. The temperature distribution has been determined by energy equation and transport equations of chemical species. The character of combustion has been observed as a quasi-homogeneous, i.e. boundary conditions sources of the moisture and volatiles were determined experimentally.
The results have shown that the D model with satisfactory accuracy can be used to describe the complex transport processes during combustion of baled biomass in the proposed furnace. This knowledge can be of great benefit to the use proposed numerical simulation for the operating optimization of considered furnace because the use of a D model significantly reduces the required calculation time and the resources required for the simulation. ACKNOWLEDGMENT The paper was developed through activities carried out within the scope of the project III411of Serbian Ministry of Science and Technological Development: Development and improvement of technologies for energy efficient and environmentally sound use of several types of agricultural and forest biomass and possible utilization for cogeneration. REFERENCE [1] Bech, N., Wolff, L., Germann L., 1996, Mathematical Modeling of Straw Bale Combustion in Cigar Burners, Energy & Fuels, No 1, pp. 76-83. [] Mladenović R., Erić A., Mladenović M., Repić B., Dakić D., 8, Energy production facilities of original concept for combustion of soya straw bales, 16 th European Biomass Conference & Exhibition From Research to Industry and Markets, Proceedings on DVD- ROM, ISBN 978-88-8947-58-1, Valencia, Spain, -6 June. [3] Miltner M., Miltner A., Harasek M., Friedl A., 7, Process simulation and CFD calculations for the development of an innovative baled biomass-fired combustion chamber, Applied Thermal Engineering 7 1138 1143. [4] A. Erić, thermo mechanical processes connected to baled soybean residue combustion in the pushing furnace, Ph.D. theses (in Serbian), Mechanical Engineering Faculty in Belgrade, 1. [5] A. Eric, D. Dakic, S. Nemoda, M. Komatina, B. Repic, Experimental mehod for determining Forchheimer equation coefficients related to flow of air through the bales of soy straw, International Journal of Heat and Mass Transfer 54 (11) 43 436. [6] A. Erić, D Dakić, S Nemoda, M. Komatina, B. Repić, Determination of the stagnant thermal conductivity of the baled soybean residue, an original work (in Serbian), Contemporary agricultural engineering, (1), Vol. 36, No. 4, p.p. 334-343. [7] Patankar S.V., Numerical Heat Transfer and Fluid Flow, Hemisphere, New York, 198.
[8] S. Nemoda, G. Živković and M. Komatina, Numerical Simulation of Reacting Fluid Flow in Porous Media Applied on The Biomass Combustion Research, The First International Conference on Computational Mechanics, (CM 4), Belgrade, 4. [9] M. H. J. Pedras, M. J. S. de Lemos, Thermal dispersion in porous media as a function of the solid fluid conductivity ratio, International Journal of Heat and Mass Transfer 51 (8) 5359 5367. [1] A. Eric, D. Dakic, S. Nemoda, M. Komatina, B. Repic, Experimental determination thermo physical characteristics of balled biomass, Energy 45 (1) 35-357. [11] M.Kaviany, Principles of Heat Transfer in Porous Media, Second Edition, Springer. [1] D. A. Nield, A. Bejan, Convection in Porous Media, Springer Science Business Media, Inc. 6.