Development of a Chemical Quasi-Equilibrium Model of Biomass Waste Gasiﬁcation in a Fluidized-Bed Reactor by Using Aspen Plus

: In the delicate context of climate change, biomass gasiﬁcation has been demonstrated to be a very useful technology to produce power and hydrogen. Nevertheless, in literature, there is a lack of a ﬂexible and fast but accurate model of biomass gasiﬁcation that can be used with all the combinations of oxidizing agents, taking into account both organic and inorganic contaminants, and able to give results that are more realistic. In order to do that, a model of biomass gasiﬁcation has been developed using the chemical engineering software Aspen Plus. The developed model is based on the Gibbs free energy minimization applying the restricted quasi-equilibrium approach via Data-Fit regression from experimental data. The simulation results obtained, considering di ﬀ erent mixes of gasifying agents, were compared and validated against experimental data reported in literature for the most advanced ﬂuidized bed technology. The maximum discrepancy value obtained for hydrogen, with respect to experimental data, is of 8%, and all the other values reached by the developed simulations, considering both organic and inorganic compounds, are in good agreement with literature data. 1.1–1.3 Nm 3 / kg.


Introduction
Every year, a great amount of agro-industrial, municipal and forestry residues are treated as waste; instead, they can be recovered and used to produce thermal and electrical energy by biological or thermo-chemical conversion processes [1,2]. Contrary to biological methods, thermo-chemical ones are more viable allowing the treatment of a wide range of feedstock in shorter residence time [3,4]. Among thermo-chemical processes, biomass gasification is one of the most efficacious conversion technologies because of lower investment costs while maintaining the ability for high-rate fuel gas production [5,6]. This process utilizes oxidizing agents (oxygen, air, steam or a mix of them) at high temperature (in the range of 750-1000 • C) to produce a fuel gas, called syngas, mostly rich in hydrogen, carbon monoxide, carbon dioxide, methane and steam along with several unwanted by-products [3]. A good quality syngas is characterized by low level of N 2 , high level of H 2 , low level of contaminants and high heating values (LHV) [7,8]. Further studies have demonstrated that the fluidized-bed is a promising type of gasifier, which ensures high reaction rates and conversion efficiencies thanks to good mixing and gas-solid contact [9,10]. Process and system simulation models have obtained great interest in the prediction of performance, giving a good description of both chemical and physical phenomena In the present work, the authors have developed, and validated against experimental data, a quasi-equilibrium model of air/steam/oxygen biomass-gasification that includes organic and inorganic products. In this way the model is able to predict syngas composition and contaminants, and the authors have evaluated the effect of several variables including gasification temperature and steam to biomass ratio (S/B) on the gas composition for different gasifying agents.

Biomass Characteristics
Using biomass waste from agricultural means zero costs for the feedstock, avoiding fuel vs. food competition and having less life cycle impacts. In order to select the most suitable biomass waste to feed the gasifier, the following criteria have to be taken into account [1]: Availability of biomass on a significant scale (tons/year); 2.
Low heat value (LHV), which has to be high, so biomass with lower humidity is preferable; 3.
Chemical composition, which has to be low in sulfur, chlorine and ash; 4.
Size and shape of biomass, which have to be uniform in order to ensure homogeneous and efficient gasification and bulk density, which has to be comparable with that of the gasifier bed, even if the latter can be adjusted via pretreatment and feeding systems.
Following the previous criteria, among the different agricultural biomass waste, hazelnut shells, which represent an abundant agricultural sub-product in regions of moderate climate [23,24], have been selected for this study. The characteristics of hazelnut shells are reported in Table 1.

Aspen Plus Modelling
The simulation of the biomass gasification process, carried out on Aspen Plus, is based on mass-energy balance and chemical equilibrium among all processes. The following assumptions were considered for the simulation: The tars considered are toluene (1-ring), naphthalene (2-ring) and benzene; • The inorganic contaminants considered are hydrogen sulphide, hydrogen chloride and ammonia.
The Aspen Plus flow sheet of the developed model is shown in Figure 1 while all the units are described in Table 2.

RGIBBS GASIF
Gibbs free energy reactor-simulates drying, pyrolysis, partial oxidation and gasification and restricts chemical equilibrium of the specified reactions to set the syngas composition by specifying a temperature approach for individual reactions In this simulation, the biomass is defined as a nonconventional component, its ultimate and proximate analyses specified according to Table 1. The Peng-Robinson equation with Boston-Mathias (PR-MB) modification, has been used to evaluate all physical properties of the conventional components in the gasification process. HCOALGEN and DCOALGEN models are selected for the evaluation of the enthalpy and density of both biomass and ash, which are non-conventional components.

Description of Aspen Plus Flow-Sheet
The power plant is mostly composed of a gasification reactor producing hazelnut shell-derived syngas. The stream BIOMASS, representing the hazelnut shell feed, goes firstly in the DECOMP block that is a RYield reactor, used to simulate the decomposition of the unconventional feed into its conventional components (carbon, hydrogen, oxygen, sulfur, nitrogen and ash, by specifying the yield distribution according to the biomass ultimate analysis in Table 1). Considering that the DECOMP block creates N, Cl and S as elemental components, which are known to produce principally HCl, NH 3 and H 2 S, and the results of experimental fractional conversion model are closer to the experimental data with respect to restricted chemical equilibrium, the product out of DECOMP is moved to the RStoic block to simulate the production of H 2 S, HCl and NH 3 by the following reactions [5]: The fractional conversion considered for S, Cl 2 and N 2 is equal to 1 [12]. Deriving stream S2 goes into a separator SEP, which separates the stream S 2 in three sub-streams: Volatile part VOLATILE, char part CHAR and a stream composed by HCl, NH 3 and H 2 S, called INORG. Then, VOLATILE stream is divided in two sub-streams: VOL and H 2 ; the former, after mixing with the oxidizing fluid, goes into the gasifier, GASIF, and the latter is used to simulate tar production in the RYield block TARPROD;. The block TARPROD is necessary since we are in steady-state conditions, and it is not possible to simulate tar formation inside the gasifier GASIF. CHAR stream is split in two sub-streams: TOGASIF that represents the char reacted in the gasifier and S3 that represents the un-reacted char; the latter is then fed to TARPROD where it reacts with hydrogen from the H 2 stream. The tar is assumed to be a formation of toluene, benzene and naphthalene. The quantities of these tars are set in accordance with literature [29] and proportions of about 60%, 20% and 20% for benzene, toluene and naphthalene, respectively, are maintained [30]. The stream S6 is the stream that represents the real output of the gasifier; in fact, it is made of the union of GASRAW, INORG and TAR streams.

Gasification Model
The reactions considered in the gasification process are reported in Table 3. Table 3. Gasification reactions [5].

Reaction Reaction Name Heat of Reaction Reaction Number
Heterogeneous reaction Table 3 are the chemical reactions considered in this work for the gasification process where the oxygen comes from biomass composition, as showed in ultimate analysis, and from air/oxygen stream if air/oxygen gasification is considered. To simulate the gasification process on Aspen Plus, a RGibbs reactor, called GASIF in Figure 1, has been used. This reactor was modelled with the restricted quasi-equilibrium approach, which allows to describe syngas composition more accurately than equilibrium models, as explained in the Introduction. Therefore, the reactions within the reactor (listed in Table 3) are conducted at their QET, rather than at the actual temperature of the gasifier. In order to be more rigorous, a Data Fit of experimental data (about hazelnut or almond shells, since they have very similar characteristics [5], using steam and/or oxygen or air as gasification agent with a gasification temperature from 600 • C to 870 • C and S/B from 0.33 to 1) has been implemented on Aspen Plus (Table 4). In this way, RGibbs gasifier evaluates the chemical equilibrium constant for each reaction at the gasifier temperature, and by that, it provides the equilibrium syngas composition. Since the Data-Fit was made from literature data obtained from experiment carried out with several gasifying agents (steam, air, oxygen or a mix of these) and silica sand as bad material, the model can simulate almost all oxidizing gasification agents (i.e., CO 2 excluded).

Results and Discussions
In order to demonstrate the feasibility of the developed model with different mix of oxidizing agents, several comparisons with experimental data have been conducted. The comparisons showed that the results predicted by the model are reasonable and near to the real ones.

Steam Gasification Results and Model Validation
The developed simulation model has been validated using experimental data of Rapagnà and Latif [32] and Karatas and Akgun [38] for steam gasification. Rapagnà and Latif used a lab-scale fluidized-bed reactor as gasifier, at 1 bar and 800 • C and hazelnut shells as biomass feedstock. The lab-scale reactor used by Karatas and Akgun was a fluidized-bed, at 1 bar and 800 • C, and the biomass feedstock used was walnut shells, which have very similar characteristics with respect to hazelnut shells. The comparison of the operative conditions and results of the present model simulation and literature data are reported in Table 5. The discrepancy of the simulative results against experimental values is shown in Table 6.  The comparison of the resulting composition values with the literature ones, reported in Table 5, shows a comparable product syngas composition (the difference is below 10% for the main gas, i.e., H 2 , below 38% for CO and CO 2 and below 15% for CH 4 ). The under or the over prediction of CH 4 is an ordinary issue in the simulative modelling since tar is not considered in the equilibrium models and it is simulated apart from the gasifier block. The underproduction of CO and the overproduction of CO 2 might be caused by the fact that the steam kinetic factors were not taken into account in the simulation. Indeed, steam lowers the gas speed and thus causes an increase in residence time and favors the dissociation of H 2 O. Therefore, the conversion of char and heavy hydrocarbons into light ones (i.e., CO) was favored. The production of tar and inorganic contaminants is not evaluated by Rapagnà and Latif and Karatas and Akgun; however, the results obtained for toluene, benzene and naphthalene agree with other literature experimental sources [25,29,39]. Zhou et al. reported that the concentration of NH 3 in the gasification product is between 500 to 30,000 ppm, depending on the nitrogen content of the biomass feedstock and the gasifier conditions [40]. The concentration of H 2 S reported in literature ranges from 1000 to 14,000 ppm in the raw syngas [41], and the concentration of HCl is around 750 ppm [41].

Effect of Steam to Biomass (S/B) Ratio
A sensitivity analysis is carried out in order to evaluate the effect of the S/B ratio on the hazelnut shell-derived syngas composition that comes from the gasifier at the fixed temperature of 800 • C. In Figure 2, the trend of the dry molar fraction of each component is reported against S/B ratio. The concentration of H 2 and CO 2 rises with the rise of S/B ratio, and the concentrations of CO and CH 4 decrease with S/B ratio. Increasing steam favors the water-gas reaction and steam methane reforming reactions resulting in an increase of H 2 and CO concentrations; however, the CO concentration decreases with increasing S/B ratio due to the water-gas shift reaction, which reduces CO reacting with steam and increases H 2 and CO 2 concentrations. Similar trends were reported in literature references [42,43].

Effect of Gasification Temperature on Syngas Composition
A sensitivity analysis is carried out in order to evaluate the effect of the gasification temperature on the hazelnut shells-derived syngas composition, keeping the S/B ratio fixed at 0.8. In Figure 3, the trend of the syngas composition as a function of the temperature in the range 770-880 • C is shown. It can be observed that H 2 and CO concentrations increase with the increase of temperature, due to the endothermic reactions R5 (water-gas) and R8 (steam methane reforming). On the other hand, the increase of temperature causes the decrease of CO 2 , CH 4 and H 2 O. The decrease of CO 2 production depends on reaction R7 (water-gas shift) that is exothermic, being favored at low temperatures, and for this reason, the higher temperature means higher CO and lower CO 2 production. Similar trends were observed by [5,44].

Air-Steam Gasification Results and Model Validation
The developed simulation model has been validated using experimental data of Lv et al. [45] for air-steam gasification. The experimental set-up was a fluidized-bed reactor, operated at 1 bar and 800 • C, and the biomass feedstock used was pine sawdust, which presents very similar characteristics with respect to hazelnut shells, except for the inorganics that are not considered. The comparison of the operative conditions and results of the present model simulation and literature data are reported in Table 7. The discrepancy of the simulative results against experimental values is shown in Table 8.
The comparison of the resulting composition values with the literature ones, reported in Table 7, shows a comparable product syngas composition (the difference is 6% for the main gas, i.e., H 2 , below 18% for CO and CO 2 and 80% for CH 4 , but it has already been mentioned previously that CH 4 is under-or over-predicted). Tar production was not evaluated by Lv et al., but as for the case mentioned previously in the text, the results obtained for toluene, benzene and naphthalene agree with other literature experimental sources [25,29,39], and the same goes for the inorganic compounds' concentration.

Steam-Oxygen Gasification Results and Model Validation
Experimental data is rarer regarding steam-oxygen; therefore, the developed simulation model has been validated using experimental data of Barisano et al. [46] for steam-oxygen gasification that used olivine instead of silica sand as bed material and almond instead of hazelnut shells as biomass. Almond shells have very similar characteristics respect to the hazelnut shells, and the olivine is well known to have a WGS catalyst effect (so the H 2 and CO 2 contents will be higher meanwhile CO lower). The experimental set-up used by Barisano Table 9. The discrepancy of the simulative results against experimental values is shown in Table 10.
The comparison of the resulting composition values with the literature ones, reported in Table 9, shows a comparable product syngas composition (the difference is 16% for the main gas, i.e., H 2 , below 16% for CO and CO 2 and 58% for CH 4 , but it has already been mentioned previously that CH 4 is under-or over-predicted). Thus, even if olivine was used as bed material, the difference of H 2 , CO and CO 2 are similar to the previous ones. Tar production was not evaluated by Barisano et al., but as for the case mentioned previously in the text, the results obtained for toluene, benzene and naphthalene agree with other literature experimental sources [25,29,46] and the same goes for the inorganic contaminants concentration. Table 9. Comparison of operating conditions and steam-oxygen gasification results (stream S10) of the present study with literature.

Conclusions
In the last decades, biomass gasification technology has been established as a viable technique for the conversion of many solid biomass residues in valuable syngas that can generate electricity and heat via cogeneration devices such as combustion engines, turbines or fuel cells or produce biofuels or chemicals. In order to optimize gasifier design and its operation with minimal time and costs, models that do not require specific information on the dimensions, capacity and structure of the reactor but that at the same time can give accurate descriptions of syngas composition are needed. Quasi-equilibrium temperature models seem to provide an answer to this problem, but there is no model in literature that encompasses air/steam/oxygen gasification including inorganic (hydrogen sulphide, hydrogen chloride and ammonia) and organic (toluene, benzene and naphthalene) contaminants. In order to develop a model that can be applied to almost all the combinations of oxidizing agents, a Data-Fit of experimental data carried out with several gasifying agents (steam, air, oxygen or a mix of these) has been taken from literature. The model has been validated using the most abundant lignocellulosic waste (here hazelnut shell) and the most used oxidizing agents (S/B ratio at 0.4 and 0.8; steam at 0.5 and ER at 0.2; S/B ratio at 0.4 and oxygen to biomass ratio at 0.36) showing good correlation between simulative and experimental data. The maximum value of discrepancy for the hydrogen concentration, which is the main component of the gasification gas, is of 16.3%. Moreover, the sensitivity analysis has showed similar trends to the ones reported in literature. Thus, the model can be used to simulate composition and contaminants of different gasification processes without taking into account specific information on the dimensions, capacity and structure of the reactor but, nevertheless, being able to have results not too different from experimental ones.