Biomass and Bioenergy 91 (2016) 69e82 Contents lists available at ScienceDirect Biomass and Bioenergy journal homepage: http://www.elsevier.com/locate/biombioe Research paper Evaluation of different biomass gasification modeling approaches for fluidized bed gasifiers Guilnaz Mirmoshtaghi*, Hailong Li, Eva Thorin, Erik Dahlquist** €lardalen University, Box 883, SE-721 23 Va €sterås, Sweden School of Business, Society and Engineering, Ma a r t i c l e i n f o a b s t r a c t Article history: Received 17 November 2015 Received in revised form 28 April 2016 Accepted May 2016 To develop a model for biomass gasification in fluidized bed gasifiers with high accuracy and generality that could be used under various operating conditions, the equilibrium model (EM) is chosen as a general and case-independent modeling method However, EM lacks sufficient accuracy in predicting the content (volume fraction) of four major components (H2, CO, CO2 and CH4) in product gas In this paper, three approachesdMODEL I, which restricts equilibrium to a specific temperature (QET method); MODEL II, which uses empirical correlations for carbon, CH4, C2H2, C2H4, C2H6 and NH3 conversion; and MODEL III, which includes kinetic and hydrodynamic equationsdhave been studied and compared to map the barriers and complexities involved in developing an accurate and generic model for the gasification of biomass This study indicates that existing empirical correlations can be further improved by considering more experimental data The updated model features better accuracy in the prediction of product gas composition in a larger range of operating conditions Additionally, combining the QET method with a kinetic and hydrodynamic approach results in a model that features less overall error than the original model based on a kinetic and hydrodynamic approach © 2016 Elsevier Ltd All rights reserved Keywords: Biomass gasification Fluidized bed gasifiers Kinetic Empirical Equilibrium model Generality Introduction Because of environmental and economic incentives, such as increasing energy prices and fossil fuel depletion, countries are changing their energy profiles toward more renewable and sustainable resources Thermochemical gasification of carbon-based solid and liquid materials, which results in product gas consisting of H2, CO, CO2, CH4 and some light hydrocarbons, has been used and developed for nearly two hundred years [1] This technology can convert renewable resources such as biomass or black liquor to energy products that substitute for fossil-based fuels Among the existing types of gasifiers, the fluidized bed gasifier has many advantages, such as easy scale-up, flexibility regarding feedstock type and size, uniform temperature distribution and high carbon conversion efficiency; therefore, it is suitable for the gasification of biomass Biomass gasification in fluidized bed gasifiers is * Corresponding author ** Corresponding author E-mail addresses: Guilnaz.mirmoshtaghi@mdh.se (G Mirmoshtaghi), Erik dahlquist@mdh.se (E Dahlquist) http://dx.doi.org/10.1016/j.biombioe.2016.05.002 0961-9534/© 2016 Elsevier Ltd All rights reserved quite a complex process, which means that the operating parameters are influenced by a large number of variables Therefore, process modeling and simulation of the gasification process is more cost effective than performing experiments According to the reviews by Puig-Arnavat [2]and Gomez Barea [3]and the study by Radmanesh [4], there are two major approaches to model gasification in fluidized beds: equilibrium modeling and dynamic modeling considering the kinetics and hydrodynamics of the bed Dynamic modeling gives a better interpretation of the real case However, this approach requires detailed information on the geometry and design of the reactor, which makes it dependent on measurements and estimation of these inputs for any further analysis of gasification process [5] Due to the complex and quite fast flow regime of different phases in the gasifier, measuring and calculating residence time is necessary for developing a correct dynamic model However defining this parameter close to reality is an issue which has been studied during years [3] In contrast, equilibrium modeling (EM), which is based on thermodynamic analysis, does not require information on the dimensions, capacity and structure of the gasifier and therefore is suitable for concept studies, preliminary design and optimization of 70 G Mirmoshtaghi et al / Biomass and Bioenergy 91 (2016) 69e82 the process [5e7] EM has been applied to the gasification process in different waysdfor example, the entire gasification mechanism is considered to be at equilibrium [6], or only the pyrolysis stage is assumed to be at equilibrium [7,8] EM is mostly applicable when the operating temperature is high and the retention time is longer than the time required for complete gasification However, the model may not provide accurate results at low operating temperatures in the range of 750e900 C [3] EM also has limitations in predicting the amount of light hydrocarbons and unconverted solid carbon Several studies have been performed on how to improve the accuracy of EM in gasification modeling Some examples are related to the gasification of coal in a fluidized bed [9] and an entrained flow bed [10], whereas other examples are of biomass gasification in a downdraft gasifier [11] and fluidized bed gasifiers [6] [12,13] In 2001, Kersten [12] reviewed and compared different quasiequilibrium models for biomass gasification in fluidized bed gasifiers He studied two methods: implementing empirical corre€pfer model [14] and using the quasilations in the Schla equilibrium temperature (QET) in the Gumz model [15] This method is explained more in Section 2.1 Kersten concluded that the Gumz model with QET yields better results Li and his colleagues in different studies [6,16] have investigated different methods to improve the accuracy of EM for biomass air gasification in circulating fluidized beds (CFB) They found that adding empirical correlations for light hydrocarbons (mainly CH4) and carbon conversion is a successful method for improving EM Recently, Lim and Lee [13] also developed a quasi-equilibrium model for fluidized bed gasifiers They built their model based on 43 experimental datasets, which were gathered from different CFB [17,18] and bubbling fluidized bed (BFB) [19,20] gasifiers They concluded that to achieve a higher level of accuracy in quasi-equilibrium models, the empirical parameters in the correlations for improving EM models should be adjusted to the experimental data from the same plant that is modeled Other attempts have been made to improve the accuracy of equilibrium models by considering reaction kinetics In these studies, the pyrolysis step is assumed to be at equilibrium, whereas char gasification and part of the homogeneous reactions in the gasification are considered to be kinetically controlled For example, Bilodeau et al [8], Nikoo et al [21] and Wang et al [22]included different reaction kinetics and, in some cases, hydrodynamics of the bed to improve the results of EM According to the literature mentioned above and as Gomez and Leckner described in their review paper [3], the modification of EM (which is called pseudo-equilibrium in Ref [3]) for the modeling of fluidized bed gasifiers can be categorized into three groups: Modifying the equilibrium temperature by the QET method, Using quasi-equilibrium by adding empirical correlations for specific components and Introducing the kinetics for specific reactions and adding hydrodynamics of the bed Gomez and Leckner [3] evaluated the capability of different modified EMs to predict the composition of the product gas at different operating conditions to measure the “generality” of those models They concluded that pseudo-equilibrium models give the most accurate results for gas composition, whereas tar and char content cannot be predicted as generally as other components According to the mentioned studies, although the gasification system is quite complex and dependent on many interrelated and independent variables, the “generality” characteristic for a model is one of the major concerns in the field of gasification modelingemostly, whether it is aimed to be used further in process design and simulation level As discussed above, EM is independent of the gasifier size and type, which makes it suitable as a basis for developing a general model However, addressing the limitations of this model to improve the accuracy of prediction results in some “non-generality” factors Therefore, a systematic study to evaluate the advantages and disadvantages of different modification methods and mapping the barriers and complexities that result in this “non-generality” would be essential for any further development of any possible generic model This is one of the major novel contribution of this study to the field of biomass gasification modeling The investigation of further possibilities in improving the modeling of biomass gasification is another part of this study All three modeling approaches presented above are included and are based on the results of this investigation along with new models suggested in this paper Methodology In this study, three equilibrium-based models from the literature, one model for each modeling approach described in Section 1, have been selected for evaluation Two of the models have also been further modified The same set of experimental data have been used for the evaluation of all models The simulation tool ASPEN PLUS has been used for the evaluation As a steady state simulation tool, ASPEN Plus has been widely used to implement EM to model biomass gasification in fluidized bed gasifiers [8,23] owing to its powerful database of thermodynamic and chemical properties [24] According to Puig-Arnavat [2], ASPEN PLUS is chosen for modeling of gasifiers and further gasification processes to avoid complexity when principal gasification reactions and some fundamental physical characteristics are included The models evaluated in this study are called MODEL I, MODEL II and MODEL III, corresponding to the three different modeling approaches described in Section MODEL I, MODEL II and MODEL III are first replicated in ASPEN PLUS and verified by comparison with the model results in the original presented studies Experimental data from different BFB and CFB gasifiers have been collected from the literature and used to evaluate the model performance, with the aim to test whether the models are also valid for experimental conditions other than those for which they were originally validated The input data used for the simulations are biomass ultimate/proximate analysis, temperature, pressure, and biomass, air and steam flowrates (see Section 2.4) The detailed information on ultimate and proximate analysis of different biomasses used in this study can be found in the referred papers for each case, respectively To choose the suitable experimental data as the input for evaluating the models, major input parameters have been compared The parameters are gasifier type (CFB or BFB), equivalence ratio (ER) (is a dimensionless index for the ratio of the mass of air input to the stoichiometric amount of air needed for full combustion [25]), temperature, load (as an index for the size and residence time of the gasifier) and the mass ratio of steam to the moisture and ash free mass of biomass (S/B) These parameters are combinations of major operating parameters (ER, temperature, S/B) and variables that can limit the “generality” aspect of the model (gasifier type, load) The cases with input parameters in different ranges than those of the original validated experiments have been chosen for the evaluation step In the first part of the paper, the overall accuracy of the models in predicting the four major components in the product gas (H2, CO, CO2 and CH4) is the focus The main results in this part of the paper are from studying the sensitivity of the models to the major input parameters, which were mentioned earlier, and analyzing the variation of accuracy in different cases In the second part of the paper, based on the discovered limitations and “bottlenecks” in the existing modified EMs found and discussed in the first part, MODEL II and MODEL III are further G Mirmoshtaghi et al / Biomass and Bioenergy 91 (2016) 69e82 modified and evaluated The new models, as extensions of MODEL II and MODEL III, are called MOD-MODEL II and MOD-MODEL III The major evaluation criteria are prediction accuracy and generality of the models Generality is evaluated for a model in terms of its ability to be applied to both BFB and FB cases in a larger range of operating parameters 2.1 MODEL I-quasi equilibrium temperature (QET) To improve EM accuracy in predicting product gas composition, Gumz [15] proposed using QET instead of the actual operating temperature of the reactor QET is the temperature, different from the operating temperature, at which the specific chemical reaction is assumed to reach equilibrium [2,12] One way to find QET is to determine a new temperature at which the difference between the measured content of product gas components and the values calculated by the equilibrium model would be at a minimum level Since this is an empirical method, the new temperature would be the result of curve-fitting It is important to know that the difference between the operating temperature and the best-fit temperature determined by this method is called “degree of approach to equilibrium” and this is actually the value which is considered in “restricted equilibrium” sheet under “temperature approach” part [15,16,26] In ASPEN PLUS, the Gibbs reactor (RGIBBS), which is based on the minimization of Gibbs energy, has an option for “restricting the equilibrium” The “degree of approach to equilibrium” can be used in this option either for the whole unit or for a specific reaction Using QET for CO-shift, methane reforming and ammonia formation reactions, Doherty et al [27] developed a model (MODEL-I) to simulate biomass gasification in a CFB gasifier in ASPEN PLUS This model was used to test the effect of air preheating on the gas yield and composition The detailed information on the model development is given in Ref [27] More details are given in the supplementary document 2.2 MODEL II-empirical correlation Hannula and Kurkela proposed another model (MODEL II) based on using empirical correlations to increase the accuracy of EM in predicting the composition of light hydrocarbons in product gas [28] Hydrocarbons, ammonia and total carbon conversions were correlated to ER, which is an indicator of the amount of oxygen entering the gasifier for partial oxidation reactions These correlations were derived by fitting correlations to the empirical results from measurements on a pressurized CFB gasifier The correlations are implemented by the CALCULATOR block in ASPEN PLUS The results from this block are used to adjust hydrocarbon formation in a stoichiometric reactor (RSTOIC) In Ref [28], all correlations and blocks used for this model are listed More details are given in the supplementary document 2.3 MODEL III-kinetic and hydrodynamic CH4 conversion, char gasification and tar formation reactions are kinetically controlled Sotudeh et al [29] proposed a model by combining EM with kinetic models for coal gasification Similarly, Nikoo and Mahinpey modified the EM to simulate the biomass gasification in a BFB [21] (MODEL-III) This model includes the kinetics of combustion and steam gasification reactions together with fluidized bed hydrodynamics derived from the simple twophase theory [21] After the decomposition of the solid biomass in the model, volatiles and fixed carbon are handled in separate steps in which volatile combustion and gasification are considered at equilibrium However, the char gasification step is modeled 71 including the kinetics of the reaction As described in detail in Ref [21], RGIBBS and RCSTR reactor modules were used for the respective gasification stages Char gasification in bed and freeboard were modeled externally as a “user model” and were linked to the flow sheet in ASPEN PLUS Another essential point to be mentioned is that the CH4 content in the product gas for this model is tuned by a nonlinear regression with specific temperature points: 700, 750, 800, 850 and 900 C In fluidized bed gasifiers, the gas velocity is a key parameter when considering the hydrodynamics of the bed Gas velocity is derived from the flowrate of the oxidizing agent and the crosssectional area of the bed MODEL III describes a specific gasifier with an internal diameter (ID) of 40 mm The applicable velocity of the oxidizing agent and consequently of the biomass is therefore limited to a specific range This means that the diameter of the bed must be provided to use MODEL III correctly The main limitation of this model is the requirement of the bed diameter or any type of similar factors that is equivalent to the diameter Therefore, to use MODEL III for gasifiers that operate at other flowrates, the equivalent flowrate is calculated assuming the equality of load between that gasifier and the one in the original model Table summarizes the characteristics of EM-based models studied in this work Additional information are given in the supplementary document 2.4 Experimental data used for model evaluation The experimental data by which each model has been validated are listed in Table Basically each model has been validated with exactly the same data used by the authors that developed that model However due to further use of the input data presented in this table for evaluation of models, it is important to report all of the experimental data used for validation of the original models Further, the experimental data that have been used for the evaluation of all models are listed in Table Each “set point” is a combination of one set of ER value, S/B ratio, temperature, pressure and load Since in some cases (gasifiers) the combination of these operating parameters are changing at the same time (not one parameter at the time), it is not scientifically correct to report the results versus variation of only one of the operating parameters Therefore in such kind of cases, set point is used to show which set of operating parameters are used as an input to the model 2.5 Model evaluation To evaluate the original and modified models, the results for the volume percentage of H2, CO, CO2 and CH4 in the gasification product gas have been compared with experimental data, and the errors are calculated based on following equations The relative error for each component at each data point is calculated by yie À yip relE i ¼ yie (1) where yie and yip are the experimental and predicted results of the volume percentages, respectively To evaluate the total performance of the model in predicting the product gas composition, the overall error (OEi) at each set point is also calculated and considered as the basis for comparison OEi ¼ n X ðrelE i yie ị (2) iẳ1 To evaluate the performance of different models in a range of set points, two other indexes are required The first is the average 72 G Mirmoshtaghi et al / Biomass and Bioenergy 91 (2016) 69e82 Table Modified EMs simulated in ASPEN PLUS for biomass gasification Models name MODEL I MODEL II MODEL III Reference to models Oxidizing agent(s) Temp range ( C) ER range S/B range Flowrate (kg/h) Modification methods from EMs [27] [28] [21] Air Air/Steam Air/Steam 730e815 856e955 700e900 0.22e0.53 e 15.5e48.5 Assigning temperature lower than equilibrium temperature in RGIBBS block for specific components and reactions 0.28e0.39 0.08e0.28 18e58 Using the empirical correlation for nonequilibrium components in product gas by calculator block 0.19e0.27 0e4.04 0.445e0.512 Including reaction kinetics and hydrodynamics of the bed as an external FORTRAN subroutine Including an empirical nonlinear regression with specific temperatures (700, 750, 800, 850 and 900 C) for CH4 concentration and “n” represents the number of set points in the analyzed dataset overall error (OE), and the other is the variation width (VW) These are calculated as follows: OE ¼ n X !, OE i n VW ¼ OE imax À OE imin (3) (4) i¼1 OEimin and OEimax are the minimum and maximum values in the analyzed dataset, respectively OEi is the overall error at each data point taken from Equation (4), Table Experimental data used for validation in original paper and partly for model evaluation in this study Test rigs Set point no ERa S/Bb Temp ( C) Pressure (kPaa) Load (Mg.me2$he1) Ref 0.53 0.45 0.40 0.52 0.37 0.43 0.34 0.35 0.4 0.38 0.22 0.26 0.3 0.46 0 0 0.004 0.03 0 0 0 0 740 718 766 815 772 787 718 730 752 789 701 728 739 805 165 119 119 119 119 119 119 119 119 119 119 119 119 119 3.051 2.918 3.427 3.330 3.684 3.227 4.283 4.507 3.952 4.051 6.174 5.819 5.275 1.981 [17] CFB-EXP I 10 11 12 13 14 0.28 0.39 0.43 0.32 0.3 0.34 0.31 0.31 0.39 0.39 0.074 0.26 0.21 0.16 0.15 0.16 0.156 0.13 0.17 0.16 822.5 900 885 840 855 890 860 885 900 930 400 400 400 400 400 400 400 500 500 500 1.181 0.712 0.807 0.917 0.983 1.013 0.939 1.005 0.785 0.858 [30] CFB-EXP II 10 0.19 0.21 0.23 0.25 0.27 1.56 1.56 1.56 1.56 1.56 800 800 800 800 800 100 100 100 100 100 0.408 0.408 0.408 0.408 0.408 [31] 10 0.22 0.22 0.22 0.22 0.22 2.7 2.7 2.7 2.7 2.7 700 750 800 850 900 100 100 100 100 100 0.354 0.354 0.354 0.354 0.354 [31] 11 12 13 14 15 0.22 0.22 0.22 0.22 0.22 1.35 2.02 2.7 4.04 800 800 800 800 800 100 100 100 100 100 0.354 0.354 0.354 0.354 0.354 [31] BFB-EXP III a b Equivalence Ratio ¼ Air (kg)/air stoichiometry (kg) Steam (kg)/biomass (kg dry ash free) G Mirmoshtaghi et al / Biomass and Bioenergy 91 (2016) 69e82 73 Table Experimental data used for model evaluation ERa S/Bb Temp ( C) Pressure (kPaa) 0.17 0.19 0.2 0.22 0.26 0 0 686 690 720 919 974 100 100 100 100 100 2.070 1.951 1.951 1.919 1.672 [32] CFB I 0.32 0.34 0.34 0.37 0.43 0 0 861 855 822 850 805 100 100 100 100 100 1.790 1.783 1.904 2.796 2.548 [33] CFB II 0.22 0.25 0.28 0.30 0 0 688 726 784 779 100 100 100 100 0.438 0.438 0.632 0.438 [34] CFB III 0.38 0.38 0.38 0.38 1.13 1.45 0.93 1.22 775 770 780 780 100 100 100 100 1.942 11.942 2.219 2.219 [18] CFB IV 0.26 0.28 0.33 0.34 0.44 0 0 800 800 800 800 800 100 100 100 100 100 0.245 0.227 0.193 0.187 0.145 [35] BFB I 5 10 0.09 0.18 0.28 0.37 0.18 0.18 0.18 0.18 0.18 1.4 1.4 1.4 1.4 1.4 1.7 1.7 1.7 1.7 1.7 850 850 850 850 850 750 800 850 900 950 100 100 100 100 100 100 100 100 100 100 0.241 0.241 0.241 0.241 0.241 0.273 0.273 0.273 0.273 0.273 [36] 11 12 13 14 15 0 0 1.1 1.4 1.8 2.7 4.7 800 800 800 800 800 100 100 100 100 100 0.370 0.289 0.225 0.161 0.088 [36] 0.19 0.24 0.27 0.32 0 0 800 800 800 800 100 100 100 100 0.271 0.438 0.438 0.199 [20] Test rigs BFB II BFB III a b Set point no Load (Mg.me2$he1) Ref [36] Equivalence Ratio ¼ Air (kg)/air stoichiometry (kg) Steam (kg)/biomass (kg dry ash free) Results 3.1 Model verification In Table 4, the results of replicated models and originally published results from the literature are shown for the major components The results show a good agreement between the original models and the replicated ones The small differences between the results of replicated and original model are expected since not all the detailed information for the model set up in ASPEN plus (for example, selection of property model) is given in the published material of each original model This can therefore cause slightly different results that can still be accepted In the following subsections, the results from simulations of the replicated models are compared with experimental data 3.2 Evaluation of the original EM-based models To form a common basis for comparison of the different models, the same set of experimental data has been used for the evaluation of all models The overall error has been calculated for all tested cases, and a descriptive statistical analysis has been made on the set of OEis The results of these tests are illustrated and analyzed further in Section 3.2.1 The experimental setups chosen for evaluation of the models are listed in Table Moreover, set points to of CFB-EXP II and to 15 of BFB-EXP III from Table are also used for model evaluation These points are highlighted with italic outlined numbers in Table 3.2.1 Evaluation of original models in different operating conditions Using MODELS I, II and III for different ranges of input data results in variation of the overall error This variation illustrates the impact of changing the operating parameters on the accuracy and performance of each model in predicting product gas concentration Gasifier type (CFB and BFB) In Fig 1, the average overall error and the variation of this value in the dataset for different gasifier types (CFB and BFB) are shown 74 G Mirmoshtaghi et al / Biomass and Bioenergy 91 (2016) 69e82 Table Verification of original models replica Input parameters Results NO.a Temp ( C) ERb S/Bc Pressure (kPaa) Model name 718 930 905 890 980 700 750 800 850 0.34 0.39 0.35 0.36 0.4 0.22 0.22 0.22 0.22 0.2 0.09 0.14 0.12 2.7 2.7 2.7 2.7 120 500 500 500 500 100 100 100 100 MODEL MODEL MODEL MODEL MODEL MODEL MODEL MODEL MODEL a b c I II II II II III III III III Original model volume fraction (%) Replicated model volume fraction (%) H2 CH4 CO CO2 H2 CH4 CO CO2 5.5 10 10.5 10.2 39 39 40 42 7.7 6.5 3.6 10.5 10 9.3 16.8 12 16 15.5 17 36 36 36 36 14.6 13.5 14 13 11 14 10 14 14.5 5.5 7.7 9.7 10.1 9.6 38.1 38.7 39.4 40.5 6.9 4.3 6.3 5.3 10.3 9.8 9.3 8.8 16.1 11 15.4 14 17 36.6 36.6 36.3 35.5 14.6 14.1 13.1 13 11 15 10 15 15.1 NO refers to [17], NO to refer to [28], NO to refer to [21] Equivalence Ratio ¼ Air (kg)/air stoichiometry (kg) Steam (kg)/biomass (kg dry ash free) Of the 63 tested set points, 24 points represent CFB gasifiers and 39 points represent BFB gasifiers Comparing the average values of overall error shows that MODEL I has almost the same level of inaccuracy in predicting the volume fraction of different components in the product gas for both BFB and CFB gasifiers; therefore, it can be concluded that the gasifier type does not affect the level of prediction accuracy for MODEL I Conversely, MODEL II and MODEL III show lower levels of OE for CFB and BFB, respectively This is because both of these models were originally built for these gasifier types Therefore, without any further analysis, it can be expected that MODEL III, which is based on the kinetics and hydrodynamics of the bed/freeboard of a specific BFB, is a more suitable approach for BFB gasifiers, whereas MODEL II, with the implemented empirical correlations taken from one specific CFB, has a better prediction capability for CFBs The other input parameter that is essential for air gasification is the ER value To investigate the impact of this parameter on the performance of each model, three groups of set points with fixed ER values, as shown in Table 5, have been analyzed based on average overall error (OE) and variation width (VW) OE identifies the level of accuracy, whereas VW shows the extent to which the changes of other parameters at that fixed ER can affect the accuracy of the model The smaller the width, the lower the level of impact on the model performance by the variation of other parameters In Tables 5e8, the minimum OE and VW are in boldface, and the maximum values for these factors are shown by outlined numbers According to the results in Table 5, MODEL III, which modifies EM by including the reaction kinetics and bed hydrodynamics, has the lowest level of OE and the smallest VW compared to the other models for ER values 0.18 and 0.22 These ER values actually represent set points 6e10 from BFB II and 6e15 from BFB-EXP III, respectively Because these points are taken from BFB gasifiers, this can be the main reason for the better performance of MODEL III compared to the other models However, in the case of ER ¼ 0.38, the most accurate model is MODEL I, whereas the smallest VW is from MODEL II This means that MODEL II, which is modified using correlations with ER for non-equilibrium components, is less affected by the variation of the other parameters when ER is greater than 0.3 Actually, based on the dataset used in this paper, ER values equal to or greater than 0.3 are more typical for CFB gasifiers; therefore, less variation of results in this ER range for MODEL II is expected However, better performance by MODEL I, which modifies EM by the QET method, can be connected to a temperature level less than or equal to 800 C, which is connected to set points analyzed for ER ¼ 0.38 Temperature CFB gasifiers 60 Overall Error (%) Overall Error (%) 60 50 40 30 50 40 30 20 20 10 10 0 MODEL I MODEL II MODEL III Fig Average overall error level and variation for prediction of gas composition from BFB and CFB gasifiers Equivalence ratio (ER) value BFB gasifiers 70 70 MODEL I MODEL II MODEL III G Mirmoshtaghi et al / Biomass and Bioenergy 91 (2016) 69e82 Table Average overall error (OE) and variation width of overall errors of original models at fixed ER (À) OE MODEL I MODEL II MODEL III 75 Table Average overall error (OE) and variation width of overall errors of original models at fixed S/B (À) VW OE 0.18 0.22 0.38 0.18 0.22 0.38 37.76 42.26 32.37 40.54 63.69 18.37 23.74 35.83 56.98 25.78 28.79 11.96 42.86 50.52 20.08 15.65 4.63 6.67 MODEL I MODEL II MODEL III VW 1.4 1.56 1.7 2.7 1.4 1.56 1.7 2.7 37.76 42.29 32.37 33.016 63.09 17.18 45.65 53.81 36.34 44.23 66.2 17.81 5.2 7.44 1.72 45.23 69.84 18.7 25.78 18.6 11.96 34.13 31.72 20.08 The minimum OE and VW are boldface, whereas the maximum values for these factors are shown by outlined numbers The minimum OE and VW are boldface, whereas the maximum values for these factors are shown by outlined numbers The other important parameter that affects the performance of a gasifier to a large extent in reality is temperature A general model should be predictive at different temperatures, so to find the nongenerality connected to a specific temperature point or range, three fixed temperature values are considered in this part of the study Table shows the OE and VW in the errors of the original models when the temperature is fixed Based on the results in this table, MODEL II shows the largest VW for all fixed temperature cases This means that MODEL II, which is based on correlating nonequilibrium components to ER value, is more sensitive to the variation of other parameters when applying fixed temperature values Conversely, MODEL I at temperatures of 780 and 800 C shows the smallest VW, which is equivalent to the larger impact of these fixed temperatures than the variation of the other input parameters on the performance of the model This can be observed because the quasi-temperature (as described in 2.1), which is set for nonequilibrium reactions in this model, is chosen by assuming an equilibrium temperature of 800 C In the case of 850 C, which is from a BFB gasifier and is also one of the specific temperatures for which MODEL III is tuned, this model shows the lowest OE and VW This means that for BFB gasifiers with temperatures higher than 800 C, the variation of other parameters affects the performance of MODEL III less than the other models In steam gasification or the cases in which steam is also used as one of the oxidizing agents, the steam-to-biomass ratio (S/B) is important This parameter is defined as the ratio between steam input mass flow and the biomass flow in kg/h A general model, therefore, is expected to reflect the impact of variation in this parameter and other operating parameters In Table 7, the OE and VW of the three original models in predicting the product gas composition at a fixed S/B for four different tests at two BFB gasifiers are shown Based on the results in Table 7, MODEL II shows the largest OE compared to the other models This can be attributed to the fact that MODEL II was originally developed for air gasification in a CFB gasifier and that the non-equilibrium components are not correlated to oxidizing agents other than ER value, which can result in inaccuracy in this model when steam is one or the only oxidizing agent Therefore, “not being connected to steam flow” can be considered as a source of non-generality in this model Moreover, the largest VW for MODEL II with S/B ¼ 1.4 and 1.56 shows a larger impact by variation in other input parameters than the fixed S/B on the performance of MODEL II Conversely, MODEL III shows the smallest OE and VW for all set points in this analysis compared to other models This means that the performance of MODEL III is less affected by the variation of other parameters than by the fixed S/B This finding is observed because in MODEL III, CO-shift is considered as one of the kinetically controlled reactions; thus, steam flowrate and the ratio between steam flow and input biomass (S/B) affects the prediction of product gas composition in this model In the cases of air gasification without any steam flow, as illustrated in Fig 2, MODEL III has the largest OE and VW whereas MODEL II shows the smallest OE and VW The results shown in this figure are based on 23 set points from CFB and BFB gasifiers As discussed earlier, a large VW indicates a higher impact by changes from other input parameters than by the fixed value (in this case, zero steam flow), and a small VW indicates a lower impact by other parameters than by the fixed value This means that, at least for the air gasification cases analyzed in this study, the approach of correlating the non-equilibrium components to the ER value, as in MODEL II, is more suitable than using kinetics for combustion and steam gasification considering the hydrodynamics of the bed, as in MODEL III Table Average overall error (OE) and variation width in overall errors of original models at fixed temperature ( C) Steam to biomass ratio (S/B) In addition to the operating parameters discussed above, the size of the gasifier, which affects the residence time for gas and char in the reactor, is also important The load, which is defined as biomass mass flowrate per cross-sectional area of the gasifier, not only is an index for the size of the reactor but also shows the input rate of the feedstock The OE and VW presented in Table are for the cases in which the load is fixed for the number of set points According to Table 8, for loads less than 0.430 Mg.me2$he1, which basically represents BFB gasifiers in the analyzed dataset, Table Average overall error (OE) and variation width in overall errors of original models at fixed load (Mg.me2$he1) OE 780 800 850 780 800 850 MODEL I MODEL II MODEL III 16.08 33.94 59.27 32.36 57.51 30.84 45.65 53.82 36.34 0.6 1.34 0.74 37.22 67.01 42.48 45.23 69.84 18.7 MODEL I MODEL II MODEL III OE MODEL I MODEL II MODEL III Load VW The minimum OE and VW are boldface, whereas the maximum values for these factors are shown by outlined numbers 0.241 0.273 0.354 0.407 0.438 1.942 1.951 2.219 45.65 53.81 36.34 37.76 42.3 32.37 40.55 63.7 18.37 33.016 63.09 17.18 38.36 38.78 52.8 17.46 22.33 60.23 31.4 37.73 54.7 16.08 33.94 59.27 45.23 69.84 18.7 25.78 18.6 11.96 53 50.52 20.08 5.2 7.44 1.72 10.69 15.29 22.03 8.02 1.27 13.82 0.05 0.44 3.44 0.6 1.34 0.74 The minimum OE and VW are boldface, whereas the maximum values for these factors are shown by outlined numbers 76 G Mirmoshtaghi et al / Biomass and Bioenergy 91 (2016) 69e82 No Steam/Biomass 70 60 Overall Error (%) 50 40 30 20 10 Fig Regression equation for modifying MODEL II and developing MOD-MODEL II MODEL I MODEL II MODEL III Fig Average overall error level and variation for no steam/biomass MODEL III shows the smallest OE and VW whereas MODEL II shows the largest OE and VW except for load ¼ 0.273 Mg.me2$he1 Small VW in this range of loads (for BFB gasifiers) leads to variation of other operating parameters being less effective on the accuracy of MODEL III than the application of fixed loads Conversely, MODEL II is more significantly affected by the changes in other parameters when the load is fixed for values less than 0.430 Mg.me2$he1 Because MODEL III considers the hydrodynamics of the bed and freeboard, the approach is quite sensitive to the size of the reactor and the biomass and gas flowrates In a fixed-size gasifier, which is close to the original case of MODEL III, the model has better performance and predictive capability However, MODEL II was originally developed for a pilot-scale CFB gasifier with a larger volume Therefore, this model is less affected by a fixed load compared to other parameters These results can also be related to the discussed factor of gasifier type, which significantly affects the performance of MODELS II and III In the case of MODEL I, at fixed loads greater than 0.430 Mg.me2$he1, the VW (and mostly OE) are the smallest compared to the other models This can be interpreted in terms of the changes in other operating parameters than these fixed loads having a lesser effect on the performance of MODEL I Because the temperatures for the set points taken for fixed loads greater than 0.430 Mg.me2$he1 are all less than 800 C and this temperature is assumed to be the operating temperature in the original model, the model performance is better than that at temperatures above 800 C According to the discussed behavior of three modified equilibrium models (MODEL I, II and III) for a range of major operating criteria (gasifier type, fixed ER, temperature, S/B and load), the Table Overview of equilibrium-based models for biomass gasification in fluidized bed gasifiers The italic and bold text indicates the limitations that have been the focus of the modified models Equilibrium model (EM) Special criteria Limitations Overprediction of H2, CO, Underprediction of CH4, tar and CO2 Minimization of the system Gibbs free energy MODEL I MODEL II MODEL III Special criteria Limitations Special criteria Limitations Special criteria Limitations Restricting equilibrium to temperature lower than operation temperature “approach equilibrium” (QET method) -Large inaccuracy when steam is one or the only oxidizing agent -The prediction is only correct for a reactor temperature of approximately 800 C -No prediction of light hydrocarbons and tar other than CH4 -Trial-and-error method to find temperature approach Using empirical correlations relating ER value to carbon conversion, hydrocarbon conversion and NH3 -The model is valid only within the range of the data used for fitting -Not applicable for ER > 0.44 -Not accurate for ER < 0.3 mostly for BFB gasifiers -Not accurate for different values of load Using kinetic equations together with hydrodynamics for char gasification (heterogeneous) reactions -Large inaccuracy for CFB gasifiers -Inflexible to large loads because it has been developed for a BFB gasifier with ID ¼ 0.04 m -No prediction of light hydrocarbons and tars other than CH4 -Cl is not considered in the ultimate analysis of the feedstock -Functions only for specific temperatures: 700e750e800 e850e900 C and not the temperatures in between MOD-MODEL II MOD-MODEL III Special criteria Special criteria Using empirical correlations for CH4 content based on regression of experimental Restricting equilibrium to temperature lower than operation temperature data from different tests for both BFB and CFB gasifiers in addition to MODEL II “approach to equilibrium” for methanation reaction in addition to MODEL III G Mirmoshtaghi et al / Biomass and Bioenergy 91 (2016) 69e82 (a) 77 (b) Fig -a Comparison of MODEL II for BFB III [20] (set points to 4) with MOD-MODEL II after applying improvement ideas -b Comparison of MODEL II for CFB III [34] (set points to 4) with MOD-MODEL II after applying improvement ideas 78 G Mirmoshtaghi et al / Biomass and Bioenergy 91 (2016) 69e82 Fig MOD-MODEL III flow sheet in ASPEN PLUS following results can be obtained: MODEL I, with a quasi-temperature approach that restricts COshift and steam reforming reactions, shows an OE between 35 and 45% in all cases Fixing the temperature at approximately 800 C has a greater effect on the performance of this model than the other operating conditions Moreover, large loads, which are equivalent to CFBs at temperatures lower than 800 C in the experimental data used for validation, also have a positive impact on the accuracy of this model MODEL II, with the approach of correlating non-equilibrium components to ER values, does not show larger sensitivity for any of the analyzed operating parameters However, it functions better for CFB gasifiers with ER equal to or higher than 0.3 MODEL III, with the approach of including the kinetics for char combustion and steam gasification reactions and hydrodynamics of the bed and freeboard, is accurate only for specific operating parameters and therefore is quite sensitive to any changes in the major input data Therefore, this model is inaccurate for a wide range of input data MODEL III shows higher accuracy only if it is applied to BFB gasifiers with steam flow and a smaller load when ER is less than 0.3 and the temperature is greater than 800 C The results of observations made in this section are summarized in Table Moreover, some preliminary ideas for the further improvement of EM using these three approaches are given The new models are called MOD-MODEL II and MOD-MODEL III and are discussed below 3.3 Evaluation of the modified EM-based models for different ER values According to the abovementioned points, MOD-MODEL II is developed based on the correlations derived from a wider range of experimental data, and MOD-MODEL III is built as the combination of two approaches used in MODEL I and MODEL III The new models are further described briefly in this section along with the visual results of testing new models for the cases with a wide variation of input data Finally the accuracy of prediction through modified models is compared with that of original models and discussed 3.3.1 Description and results of MOD-MODEL II As discussed earlier in the paper, one of the limitations of MODEL II is the strong impact of gasifier type on the accuracy of this model in the prediction of product gas composition Therefore, including equations that distinguish between CFB and BFB gasifiers is one step towards “generality” for this model However, MODEL II is more accurate for ER values greater than or equal to 0.3 (which is more typical for CFB gasifiers) This “non-generality” factor can also be tackled using empirical regression with a wider range of ER The new dataset that is used for the definition of new empirical equations is collected from experimental studies of different BFB [20,31,35e41] and CFB gasifiers [17,32,33,42e45] with a wider range of ER, temperature, S/B and load compared to the empirical equations used in the original model A new equation, Equation (5), which relates CH4 content to the ER value, is obtained from Fig and used to substitute the corresponding equation in the original model In this equation CH4 content is calculated as volume fraction of the product gas Implementing this equation in the CALCULATOR block, the “execution sequence” of result should be before RSTOIC to adjust the “molar extend” of the respective reaction As given in the setup explanation of RSTOIC in ASPEN plus, molar extend is defined as the number of moles generated for any component divided by its stoichiometric coefficient All these terms are explained in ASPEN plus unit operation user guide Equation (5) is derived from 92 set points from tests on CFB gasifiers and 44 set points from tests on BFB gasifiers CH4 ¼ 1:207 Â ERÀ0:92 (5) In Fig 4-a, the results of MOD-MODEL II are compared with those of the original MODEL II and the experimental data for a BFB gasifier from Kim et al [20] This gasifier is operated isothermally at 800 C while ER is varied from 0.19 to 0.32 The circular points are experimental results, and the dashed lines represent MODEL II and MOD-MODEL II For all components, the accuracy of the modified model is improved compared to the original model, and the average overall error for the modified model is 30.9% compared to 44% for the original model Fig 4-b illustrates the comparison about MOD-MODEL II and the original MODEL II for the CFB gasifier tested by Miao et al [34] This gasifier is operated in low temperature (688e784 C) while ER value is varied in a small range of 0.22e0.3 As mentioned before, circular points are experimental points and dashed lines of different size represent MODEL II and MOD-MODEL II As shown in Fig 4-b, there is not remarkable difference between the two models 3.3.2 Description and results of MOD-MODEL III According to Table and the limitations highlighted for MODEL I and MODEL III, combining MODEL I and MODEL III is suggested as another modification strategy to improve the accuracy of prediction for wider range of input data and consequently improve the generality of the original models by overcoming the limitations G Mirmoshtaghi et al / Biomass and Bioenergy 91 (2016) 69e82 (a) 79 (b) Fig -a Comparison of MODELS I and III for CFB I [32] (set points to 5) with MOD-MODEL-III after applying modifications -b Comparison of MODELS I and III for BFB II [36] (set points to 5, ER variation) with MOD-MODEL-III after applying modifications 80 G Mirmoshtaghi et al / Biomass and Bioenergy 91 (2016) 69e82 Fig Average overall error level of modified models compared to original models for CFB gasifier, BFB gasifiers and all gasifiers The new modified model is called MOD-MODEL III As shown in Fig 5, the devolatilization and char gasification parts are taken directly from MODEL III, whereas the volatile gasification mainly controlled by steam and restricting the equilibrium for methane formation reaction by the QET method is added based on the approach in MODEL I Methane formation reaction ”temperature approach” is set to À350 C Applying this modification has been evaluated by experimental data from CFB gasifier by Wu and et al [32] and BFB gasifier by Turn and et al [36] The results are shown in Fig 6-a and -b The overall accuracy and the generality of MODEL III has been improved in both cases In the case of CFB I, the overall error is reduced from 50% to 40% while in the case of BFB II, it is reduced from 36% to 8% Moreover, the predictive capability of MODEL III has been improved, especially for CH4 and H2 at high temperatures and high ER conditions This can be seen in Fig 6-a at set points and 5, also in Fig 6-b at ER values around 0.2 and more On the other hand, in the case of CFB I, the accuracy in predicting CO and CO2 content is decreased in MOD-MODEL III, but the trend of changes in value of contents by variation of ER and temperature, follows the experimental results The link between CO, H2 and CO2 content is CO-shift reaction Additionally, since H2 and CH4 are related by the methanation reaction, which is assigned as restricted equilibrium by the QET method, the amount of these two components are accurately predicted However, even though H2 content of product gas, as one of the effective components in the CO-shift reaction, is well predicted, CO is still overestimated and CO2 is underestimated The lack of accuracy for these two components compared with the performance of each original model (MODEL I and III) can be connected to the kinetic and hydrodynamic equations used for MODEL III However, the reason why MOD-MODEL III shows worse results than MODEL III for these components can be explained by referring to the COshift reaction Based on this reaction, when H2 amount is reduced (MOD-MODEL III predicts lower content of H2 than MODEL III), the reaction is shifted to the left side, which results in more CO and less CO2, which can be clearly observed from the MOD-MODEL III results compared to MODEL III In the case of BFB II, as shown in Fig 6-b, the accuracy of the modified model is also improved Fig summarizes the average overall error of each model in predicting the content of major components in the product gas It clearly shows that the accuracy of MODEL III has been improved in the case of CFB gasifiers, BFB gasifiers and all the gasifiers by MODMODEL III However in case of CFB gasifiers the most accurate model is MODEL II which has not been improved substantially in MOD-MODEL II This shows that changing the conversion equation to the Equation (5) does not improve the results for CFB gasifiers Based on the results for BFB gasifiers, the accuracy of MODMODEL II is slightly better than the original model Generally a small improvement in accuracy of MODEL II has been obtained by developing MOD-MODEL II, as shown in the case of all gasifiers Generally it can be concluded that applying some minor modification and changes in the presented original models can improve the accuracy of models especially in the case of MODMODEL III which shows the most accurate prediction compared to all other models This further adds up to the generality of the model to give more accurate prediction in different operating condition and gasifier design Conclusions Using the equilibrium approach as the basis for modeling gasification enables the model to be independent of the detailed design G Mirmoshtaghi et al / Biomass and Bioenergy 91 (2016) 69e82 information of the gasifier However, improving the accuracy of EM by different approaches can introduce limitations and sources of non-generality to EM In this study, the performance of three EM modification approaches have been studied To investigate the effect of different operating parameters on the predictive potential of the models, different experimental cases have been selected for validation The quasi-temperature modeling approach (MODEL I) shows the least sensitivity to gasifier type compared to other models The modeling approach based on including empirical correlations with ER values (MODEL II) shows the highest inaccuracy and higher sensitivity to all tested parameters, and the approach of including kinetics and hydrodynamics (MODEL III) shows the best accuracy but only for specific and limited input data By including more experimental data for the determination of the empirical correlations, as performed in MOD-MODEL II, better accuracy in the prediction of product gas composition at a larger range of operating conditions is reached This is mainly the case for BFB gasifiers Additionally, combining the QET method used in MODEL I with the kinetic and hydrodynamic approach in MODEL III results in a modified model (MOD-MODEL III) that shows lower overall error than not only MODEL III but also the other models presented in this study However, CO and CO2 content of product gas are not well predicted in all of the conditions [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] Acknowledgments [17] This work is carried out in the Swedish Gasification Centre consortium The Swedish Energy Agency and the academic and industrial partners are gratefully acknowledged [18] Appendix A Supplementary data [19] Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.biombioe.2016.05.002 [20] Nomenclature Number of samples À Relative error À The experimental results Volume fraction (%) The predicted results Volume fraction (%) The overall error À Average overall error À Variation width Equivalence ratio (kg) air/(kg) stoichiometric air Steam to biomass ratio (kg) steam/(kg) dry and ash free biomass CFB Circulating fluidized bed BFB Bubbling fluidized bed QET Quasi equilibrium temperature RGIBBS Gibbs reactor in ASPEN plus RSTOIC Stoichiometric reactor in ASPEN plus RCSTR Continuous steering tank reactor in ASPEN plus n relE i yie yip OEi OE VW ER S/B [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] References [31] [1] G.J Stiegel, R.C Maxwell, Gasification technologies: the path to clean, affordable energy in the 21st century, Fuel Process Technol 71 (2001) 79e97 [2] M Puig-arnavat, J.C Bruno, A Coronas, Review and analysis of biomass gasification models, Renew Sustain Energy Rev 14 (9) (2010) 2841e2851, http://dx.doi.org/10.1016/j.rser.2010.07.030 [3] A Gomez Barea, B Leckner, Modeling of biomass gasification in fluidized bed, Prog Energy Combust Sci 36 (2010) 444e509 [4] R Radmanesh, J Chaouki, C Guy, Q Hc, Biomass gasification in a bubbling fluidized bed reactor: experiments and modeling, AICHE J 52 (12) (2006) [5] B Buragohain, P Mahanta, V.S Moholkar, Thermodynamic optimization of [32] [33] [34] [35] 81 biomass gasi fi cation for decentralized power generation and Fischer -Tropsch synthesis, Energy 35 (6) (2010) 2557e2579, http://dx.doi.org/ 10.1016/j.energy.2010.03.003 X Li, J Grace, C Lim, A Watkinson, A Ergüdenler, Equilibrium modeling of gasification: a free energy minimization approach and its application to a circulating fluidized bed coal gasifier, Fuel 80 (2001) 195e207 S.S Sadaka, A.E Ghaly, M.A Sabbah, Two-phase biomass air-steam gasiÿcation model for uidized bed reactors: part III d model validation, Biomass Bioenergy 22 (2002) 479e487 J.F Bilodeau, N Therien, P Proulx, S Czernik, D.D.G Chimique, A mathematical model of fluidized bed biomass gasification, Can J Chem Eng 71 (1993) 549e557 V Gururajan, PA, JA, Mathematical modeling of fluidized bed coal gasifiers, Trans.IChemE 70a (1992) 211e238 H Yoshida, F Kiyono, H Tajima, A Yamasaki, Two-stage equilibrium model for a coal gasifier to predict the accurate carbon conversion in hydrogen production, Fuel 87 (2008) 2186e2193 Z.A Zainal, A Rifau, G.A Quadir, K.N Seetharamu, Experimental investigation of a downdraft biomass gasifier, Biomass Bioenergy 23 (2002) 283e289 S.R.A Kersten, A van der Drift, W Prins, W.P.M van Swaaij, Interpretation of biomass gasification by “Quasi”-equilibrium models, in: 12th, European Biomass Conference; Biomass for Energy, Industry and Climate Protection, 2002 Amsterdam Y Lim, U Lee, Quasi-equilibrium thermodynamic model with empirical equations for air e steam biomass gasification in fluidized beds, Fuel Process Technol 128 (2014) 199e210, http://dx.doi.org/10.1016/j.fuproc.2014.07.017 P Schl€ apfer, J Tobler, Theoretische und praktische Untersuchungen über den Betrieb von Motorfahrzeugen mit Holzgas schweizerische gesellschaft fur das stadium der motorbrenstoff, d Gesellschaft, Bern: Selbstverl, 1937 W Gumz, Gas Producers and Blast Furnaces, John Wiley & Sons, Inc., New York, 1950 X.T Li, Biomass Gasification in Circulating Fluidized Bed, University of British Columbia, 2002 X.T Li, J.R Grace, C.J Lim, A.P Watkinson, H.P Chen, J.R Kim, Biomass gasification in a circulating fluidized bed [Internet], Biomass Bioenergy 26 (2) (2004) 171e193 Available from: http://linkinghub.elsevier.com/retrieve/pii/ S0961953403000849 X Meng, W de Jong, N Fu, A.H.M Verkooijen, Biomass gasification in a 100 kWth steam-oxygen blown circulating fluidized bed gasifier: effects of operational conditions on product gas distribution and tar formation, Biomass Bioenergy 35 (7) (2011) 2910e2924, http://dx.doi.org/10.1016/ j.biombioe.2011.03.028 mez-barea, F.B Vidal, P Ollero, Air e steam gasification of M Campoy, A Go biomass in a fluidised bed: process optimisation by enriched air, Fuel Process Technol 90 (5) (2009) 677e685, http://dx.doi.org/10.1016/ j.fuproc.2008.12.007 Y Doo Kim, C Won Yang, B Jong Kim, K Su Kim, J Woo Lee, J Hong Moon, et al., Air-blown gasification of woody biomass in a bubbling fluidized bed gasifier, Appl Energy 112 (2013) 414e420, http://dx.doi.org/10.1016/ j.apenergy.2013.03.072 M.B Nikoo, N Mahinpey, Simulation of biomass gasification in fluidized bed reactor using ASPEN PLUS, Biomass Bioenergy 32 (12) (2008) 1245e1254 Y Wang, C.M Kinoshita, Kinetic model of biomass gasification, Sol Energy 51 (1) (1993) 19e25 J Corella, A Sanz, Modeling circulating fluidized bed biomass gasifiers A pseudo-rigorous model for stationary state, Fuel Process Technol 86 (2005) 1021e1053 Aspen Plus User Guide, 2006 Cambridge,MA G Mirmoshtaghi, H Li, E Dahlquist, E Thorin, Bio-methane production through different biomass gasifiers, in: ICAE2013, 2013, pp 1e10 D Bacon, J Downie, J Hsu, J Peters, Modeling of fluidized bed gasifiers, in: R.P Overend, K TM, L.K Mudge (Eds.), Fundamentals of Thermochemical Biomass Conversion, Elsevier Applied Science, London, 1985, pp 717e732 W Doherty, A Reynolds, D Kennedy, The effect of air preheating in a biomass CFB gasifier using ASPEN Plus simulation, Biomass Bioenergy 33 (9) (2009) 1158e1167, http://dx.doi.org/10.1016/j.biombioe.2009.05.004 I Hannula, E Kurkela, A semi-empirical model for pressurised air-blown fluidized-bed gasification of biomass, Bioresour Technol 101 (12) (2010) 4608e4615 http://www.ncbi.nlm.nih.gov/pubmed/20153170 R Sotudeh-Gharebaagh, R Legros, J Chaouki, J Paris, Simulation of circulating fluidized bed reactors using ASPEN PLUS, Fuel 77 (4) (1998) 327e337 E Kurkela, P Stahlberg, J Laatikainen, Pressurized Fluidized Bed Gasification Experiments with Wood, Peat, and Coal at VTT[ Espo, Finland] in 1991e1992 1: Test Facilities and Gasification Experiments with Sawdust Espo, 1993 P.M Lv, Z.H Xiong, J Chang, C.Z Wu, Y Chen, J.X Zhu, An experimental study on biomass air e steam gasification in a fluidized bed, Bioresour Technol 95 (2004) 95e101 W Jianzhi, X Bingyan, L Zhenfang, Z Xiguang, Performance analysis of a biomass circulating fluidized bed gasifier, Biomass Bioenergy (2) (1992) 105e110 A Van Der Drift, J Van Doorn, J.W Vermeulen, Ten residual biomass fuels for circulating fluidized-bed gasification, Biomass Bioenergy 20 (1) (2001) 45e56 Q Miao, J Zhu, S Barghi, C Wu, X Yin, Z Zhou, Model validation of a CFB biomass gasification model, Renew Energy 63 (2014) 317e323 I Narvaez, A Orio, M.P Aznar, J Corella, Biomass gasification with air in an 82 [36] [37] [38] [39] [40] G Mirmoshtaghi et al / Biomass and Bioenergy 91 (2016) 69e82 atmospheric bubbling fluidized bed Effect of six operational variables on the quality of the produced raw gas, Ind Eng Chem Res 5885 (95) (1996) 2110e2120 S Turn, C Kinoshita, D Ishimura, An experimental investigation of Hydrogen production, Int J Hydrogen Energy 23 (8) (1998) 641e648 W.A Wan, A Karim Ghani, R Alipour Moghadam, M.A Mohd Salleh, A.B Alias, Air gasification of agricultural waste in a fluidized bed gasifier: hydrogen production performance, Energies (2009) 258e268 U Arena, L Zaccariello, M.L Mastellone, Gasification of natural and waste biomass in a pilot scale fluidized bed reactor, Combust Sci Technol 182 (2010) 625e639 V Skoulou, G Koufodimos, Z Samaras, A Zabaniotou, Low temperature gasification of olive kernels in a 5-kW fluidized bed reactor for H2-rich producer gas, Int J Hydrogen Energy 33 (22) (2008) 6515e6524, http:// dx.doi.org/10.1016/j.ijhydene.2008.07.074 R Radmanesh, J Chaouki, C Guy, Q Hc, Biomass gasification in a bubbling [41] [42] [43] [44] [45] fluidized bed reactor: experiments and modeling, AICHE J 52 (12) (2006) 4258e4272 P Subramanian, A Sampathrajan, P Venkatachalam, Fluidized bed gasification of select granular biomaterials, Bioresour Technol 102 (2) (2011) 1914e1920, http://dx.doi.org/10.1016/j.biortech.2010.08.022 K.N Sheeba, J.S Chandra, S Jaisankar, Energy for sustainable development air gasification characteristics of coir pith in a circulating fluidized bed gasifier, Energy Sustain Dev 13 (2009) 166e173 E Kurkela PS and JL Pressurized Fluidized Bed Gasification Experiments with Wood, Peat, and Coal at VTT 1: Test Facilities and Gasification Experiments with Sawdust,” VTT Espo, Finland; ~ ez, A Cabanillas, J.M Sa nchez, Gasification of leached orujillo P García-Iban (olive oil waste) in a pilot plant circulating fluidised bed reactor Preliminary results, Biomass Bioenergy 27 (2) (2004) 183e194 Lars Waldheim, C Fredriksson, Evaluation of the Pilot Plant Test on Cane Trash, 2002