Biomass steam gasification for production of SNG – Process design and sensitivity analysis

Applied Energy 97 (2012) 451–461

Contents lists available at SciVerse ScienceDirect

Applied Energy journal homepage: www.elsevier.com/locate/apenergy

Biomass steam gasification for production of SNG – Process design and sensitivity analysis Thomas Gröbl ⇑, Heimo Walter, Markus Haider Vienna University of Technology, Institute for Energy Systems and Thermodynamics, Getreidemarkt 9/302, A-1060 Vienna, Austria

a r t i c l e

i n f o

Article history: Received 22 June 2011 Received in revised form 11 January 2012 Accepted 14 January 2012 Available online 10 February 2012 Keywords: Biomass Gasification Methanation Simulation Sensitivity analysis

a b s t r a c t A process design for small-scale production of Substitute Natural Gas (SNG) by steam gasification of woody biomass is performed. In the course of this work, thermodynamic models for the novel process steps are developed and implemented into an already existing model library of commercial process simulation software IPSEpro. Mathematical models for allothermal steam gasification of biomass as well as for cleaning and methanation of product gas are provided by applying mass balances, energy balances and thermodynamic equilibrium equations. Using these models the whole process is integrated into the simulation software, a flowsheet for an optimum thermal integration of the single process steps is determined and energy savings are identified. Additionally, a sensitivity study is carried out in order to analyze the influence of various operation parameters. Their effects on amount and composition of the product gas and process efficiency are evaluated and discussed within this article. Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction In increasingly industrialized societies the energy consumption has risen steadily in the last decades. Besides, the dependence on fossil primary energy sources is very strong. Because of limited fossil fuel resources and environmental aspects the share of renewable and sustainable energy of the overall energy demand must be increased. Therefore political strategies have been developed and targets have been defined by the European Commission [1]. It is commonly agreed, that one promising option to reach these targets is the use of biomass as a renewable energy carrier. Biomass is carbon neutral, which means that during growth process as much CO2 is captured via photosynthesis as is released during thermal utilization and consequently net carbon emissions equal zero. Furthermore, biomass is independent from short-term availability fluctuations, unlike hydro, wind and solar power and is therefore suitable for stabilization of renewable energy systems [2]. In recent years, research and development strongly focuses on gasification of solid biomass for production of so called secondary generation fuels (SGF), like Substitute Natural Gas (SNG), Fischer–Tropsch fuels (BtL) or bio hydrogen (BioH2). Especially the production of SNG seems to be a reasonable way because the possibility to feed the SNG into an already existing natural gas infrastructure is given. Further the SNG process is characterized by high conversion efficiency [3].

⇑ Corresponding author. Tel.: +43 1 58801 302313. E-mail address: [email protected] (T. Gröbl). 0306-2619/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.apenergy.2012.01.038

The production of SNG via gasification of solid biomass and subsequent methanation of the synthesis gas requires a gasifier generating a producer gas with a suitable composition (low nitrogen, high hydrogen). Therefore, mostly allothermal steam gasification technologies are used. Due to the spatial separation of endothermic gasification and exothermic combustion there will be no mixing of the combustible producer gas with the flue gas respectively atmospheric nitrogen. Thus, a virtually nitrogen free, medium calorific value producer (respectively synthesis gas) with high hydrogen content, well suited for SNG production, can be generated using a cheap and technically simple gasification agent without the need for an air separation unit. In principle, also autothermal gasification using pure oxygen as gasifying agent would be suitable. However, air separation for generation of oxygen is expensive and therefore in general economically viable only for large-scale applications. Allothermal steam gasification therefore constitutes the technology of choice for small to middle scale applications, which are mainly considered for biomass due to its local distribution and availability limitations. Several projects deal intensively with the production of SNG, e.g. the EU-project ‘‘BioSNG’’ [4]. In the combined heat and power (CHP) plant in Güssing (Austria) [5,6] a 1 MW pilot and demonstration unit for the production of BioSNG was installed and successfully proven. The process of the Güssing CHP plant, as well as other present concepts, is based on atmospheric gasification and commercial cold gas cleaning. Thus, it has to be operated in the power range of 10–25 MW to become economically profitable. The combination of pressurized gasification and hot gas cleaning entails a clear process simplification and significant efficiency

452

T. Gröbl et al. / Applied Energy 97 (2012) 451–461

Nomenclature A cp p d

0

hf ;298 j Kp L LHV _ C;fuel m _ C;res m _ pyr;waf m _ res;waf m p0 Pel,C PEH pdeq pi pi

filter surface (m2) isobaric heat capacity (J kg1 K1) sauter-diameter (m) standard enthalpy of formation at 298,15 K (J kg1) variable for the operation parameters () thermodynamic equilibrium constant () length of the packed bed (m) lower heating value (kJ kg1) total carbon input into the reformer (kg s1) residue carbon leaving the reformer (kg s1) pyrolysis char input into the reformer (kg s1) residue char leaving the reformer (kg s1) standard pressure (bar) electric power consumption (kW) heating rate of excess steam (kW) logarithmic deviation from thermodynamic equilibrium () partial pressure of species i (bar) partial pressure of species i at thermodynamic equilibrium (bar)

increase by which a decentralized SNG production is allowed in small arrangements [7]. Furthermore, the pressurized operation is favorable for subsequent methanation. Pressurized gasification is enabled by the application of the Biomass Heatpipe Reformer (HPR), an innovative allothermal steam gasification technology [8,9]. Main parts of the HPR are a combustion chamber and a reformer, both designed as bubbling fluidized bed reactors. Heat is transferred from the combustor into the allothermal reformer using so called Heatpipes. Heatpipes are enclosed metal pipes containing an alkali metal working fluid (e.g. Na, K). This working fluid is evaporated in the area of the exothermic combustion chamber fluidized bed and then condensed in the area of the endothermic gasification fluidized bed. The resulting elevated heat transport density enables complete decoupling of combustion and gasification and therefore pressurized operation of the reformer. This allows compact reactor design and integration, what leads to economic and energetic efficiency comparable to the performance of large systems [10]. A 500 kWth demonstration plant of the Biomass Heatpipe Reformer has been in operation in Pfaffenhofen (Germany) over the past two years showing stable operation [10]. Further, a commercial sized CHP application with a thermal power input of 1.3 MWth is currently under construction in Grassau (Germany). The objective of the work presented in this article is the performance of a process design for production of SNG using a combination of pressurized gasification with the HPR, hot gas cleaning and methanation. Therefore, mathematical models for the single process steps are required. Mathematical models can be divided into thermodynamic equilibrium and kinetic rate models. Kinetic rate models provide detailed information on the analyzed process and can give accurate results. But however, they are computationally intensive and always contain parameters which make them applicable only for one particular reactor design. In contrast, thermodynamic equilibrium models are computationally less intensive, independent of reactor design and therefore more convenient for process studies. Moreover, equilibrium models predict thermodynamic limits of performance under different conditions. This makes these models a useful tool for process design and optimization. Hence, many applications of equilibrium models are reported in literature [11–15]. A detailed review and analysis of biomass gasification models is given in [16].

R s sj,rel T V_

a b DG0R

Dp

g gchem gges gth k

mi qf r v

w

universal gas constant (J mol1 K1) filter thickness (m) relative sensitivity coefficient () temperature (K) flow rate (m3 s1) viscosity coefficient (m2) inertia coefficient (m) Gibbs free enthalpy of reaction at standard pressure (J mol1) pressure drop (Pa) dynamic viscosity (Pa s) cold gas efficiency (%) overall thermal utilization rate (%) thermal process efficiency (%) air excess ratio () stoichiometric coefficient of species i () fluid density (kg m3) steam excess ratio () superficial velocity (m s1) porosity ()

The presented article is organized in three parts. In the first part, thermodynamic equilibrium models developed for the SNG process will be described. A validation of the newly developed HPR model is performed within the second part. In the third part, results of process simulation are shown. Energy savings have been identified and an optimum thermal integration for the overall process has been determined. Furthermore, a comprehensive sensitivity analysis has been carried out to show the influence of the most important operation parameters.

2. Process scheme A flow chart of the SNG production process is shown in Fig. 1. Gasification occurs in the Biomass Heatpipe Reformer (HPR), which is a patented development of the Technical University of Munich [17]. As gasification agent superheated steam is used. Based on the allothermal operation, the HPR produces a nearly nitrogen-free, high calorific value syngas. Because of the ideal H2/CO ratio and the pressurized operation, the generated gas is well suited for subsequent methanation. Methanation catalysts are very sensitive against catalyst poisoning. Thus, an intensive gas cleaning is required. In a first step, solid particles are separated with the help of a cyclone. The separated particles are fed back into the combustion chamber of the HPR. In the next step, tars are removed from the syngas by catalytic reforming on a nickel based catalyst. Before the alkali elements can be removed, it is necessary to cool down the syngas below the condensation temperature of the alkali elements. After cooling the alkalis are removed in a filter cartridge. Sulfur leads to poisoning and deactivation of the methanation catalyst and therefore it is captured by chemical adsorption on a copper-oxide sorbent. The methanation itself is carried out in a fixed-bed reactor with a nickel based catalyst. Since the methanation reaction is strongly exothermic, cooling of the reactor is inevitable. After methanation, the raw-SNG has to be dried and cleaned from carbon dioxide and hydrogen to comply with the demands of the natural gas grid. While gasification of the biomass is endothermic, syngas and flue gas cooling as well as methanation are exothermic. Therefore, an optimal process design and thermal integration plays a major role in the performance and efficiency of the overall process.

453

T. Gröbl et al. / Applied Energy 97 (2012) 451–461

Fig. 1. Flow chart of the SNG process.

3. Process modeling For process modeling the commercial process simulation software IPSEpro [18] is used. IPSEpro is a steady state, equation oriented simulation tool for modeling and analyzing processes in energy- and chemical engineering as well as related areas. With IPSEpro’s Model Development Kit (MDK) new component models can be built up or already existing ones can be modified. Consequently, user defined models can be prepared and implemented in an existing model library. For the present work, the IPSEpro Pyrolysis and Gasification Process Library (PGP-Lib) [19,20] is used as basic library. Different component models for simulation of the SNG production process are developed and implemented into this library. 3.1. Model description In this section the mathematical models developed for the SNGprocess will be described. For the remaining process units (pumps, compressors, heat exchangers. . .) standard models from the PGpLib are used. 3.1.1. Gasification – The Biomass Heatpipe Reformer Fig. 2 illustrates the arrangement of the models used for simulation of the gasification unit (HPR) and the cyclone for particle separation. The Biomass Heatpipe Reformer is subdivided into a combustor and a reformer. In the combustion chamber, a part of the biomass feedstock is burnt together with ungasified char remaining from

the reformer in order to supply the required energy for the endothermic gasification process. The heat released in the combustion chamber is transferred into the allothermal reformer via the Heatpipes. In the model, the transferred heat is calculated using the enthalpy difference of the flue gas as indicated in Fig. 2. 3.1.1.1. Combustion chamber. For modeling the combustor the standard combustion chamber model of the PGP-Library is extended by two additional inlets for ungasified char and separated particulates. The model handles mass and energy balances and in case of the excess air coefficient k > 1 complete conversion is assumed. The energy balance implemented in each kind of chemical reactor model is formulated using total thermodynamic enthalpies of all streams including their possible loadings. The total thermodynamic enthalpy additionally includes the enthalpy of formation and is calculated according to Eq. (1). By this, heats of reaction are basically considered. 

0

h ðTÞ ¼ hf ;298 þ

Z

T

cp ðTÞdT

ð1Þ

298:15

3.1.1.2. Reformer. To determine temperature and composition of the syngas leaving the reformer, elementary mass balances, an energy balance and thermodynamic equilibrium equations are implemented in the reformer model. Since the gas phase is more likely to react to thermodynamic equilibrium than the solid phase, only homogeneous reactions are considered. Table 1 shows a summary of these reactions. The amount and composition of ungasified char leaving the gasification zone has to be defined by the user. This solid residuum does not take part at the equilibrium calculation and is fed into the combustion zone. Further, the tar content in the product gas as well as the composition of tar has to be specified by the user too. It is well known, that the biomass gasification process is not completely governed by thermodynamic equilibrium. Thus, thermodynamic equilibrium of the considered reactions is not

Table 1 Homogeneous gasification reactions. Reaction equation

Fig. 2. Scheme of the gasification process.

CO + H2O M CO2 + H2 CH4 + H2O M CO + 3H2 C2H4 + 2H2O M 2CO + 4H2 C2H6 + 2H2O M 2CO + 5H2 C3H8 + 3H2O M 3CO + 7H2 N2 + 3H2 M 2NH3 NH3 + CO M HCN + H2O

Name (R1) (R2) (R3) (R4) (R5) (R6) (R7)

Water gas shift reaction Methane steam reforming Ethylene steam reforming Ethane steam reforming Propane steam reforming Ammonia synthesis Cyanide formation

454

T. Gröbl et al. / Applied Energy 97 (2012) 451–461

stringently defined within the model. Instead, a logarithmic deviation from equilibrium is calculated (Eq. (2)). According to [19,20], the logarithm of the ratio of the actual partial pressure product and the equilibrium constant is implemented as a model parameter.

2Q  ti 3 pi

6 i p0 7 7 pdeq ðpi ; TÞ ¼ log10 6 4 K p ðTÞ 5

ð2Þ

The thermodynamic equilibrium constant Kp(T) of a certain reaction is calculated directly from thermodynamic data according to the following equation: DG0 ðTÞ R RT

K p ðTÞ ¼ e

¼

Y p mi i

i

p0

ð3Þ

with the Gibbs free enthalpy of reaction DG0R ðTÞ at standard pressure and a temperature T, the partial pressures of the species at thermodynamic equilibrium pi and the standard pressure p0. The required data is already included in the used PGP-Lib for all reactions given in Table 1. For each reaction given in Table 1, the parameter pdeq is calculated. Thereby it is possible, either to describe thermodynamic equilibrium of certain reactions or to calculate to what extent thermodynamic equilibrium is achieved with respect to particular chemical reactions for a given drain gas composition of the corresponding components. If pdeq < 0, the actual state of the drain gas is still on the side of the reactants and further reaction in direction of the products is thermodynamically possible. pdeq > 0 means, that the actual state of the drain gas is on product side and thus the reaction can only proceed towards the reactants. If pdeq = 0, thermodynamic equilibrium is fulfilled. In the model the variable pdeq of a certain reaction can optionally be set by the user, or calculated as a result for a given drain gas composition as stated above. If set, calculated species of the drain gas can be fitted to experimental data by adjusting the corresponding parameter pdeq. The substance classes available within the IPSEpro PGP-library, and therefore considered in the model, are summarized in Table 2. For the given species, all the required thermodynamic data are included in the PGP model library and can therefore be used within the model. A gas stream can carry loadings of organic (char and/or tar) and inorganic (dust) substances. Because organic substances are not clearly defined on a molecular basis, the characterization is based on the elementary analysis. Inorganic substances are described on a molecular base. The enthalpy of formation of these optional loadings is taken into account within the energy balance. As can be seen in Table 2, the substance class Gas consists of 18 components. In the reformer model, the oxidized gas species in the syngas N2O, NO, SO2 and O2 are set to zero because of the reducing atmosphere during allothermal gasification. Furthermore, elementary mass balances for the elements C, H, O, N, S and Cl, and thermodynamic equilibrium relations for (R1–R7) in form of Eq. (2) are implemented within the model. This, together with the relation that the sum of all components mole fractions equals 1, provides 18 equations for calculation of the 18 syngas components. Table 2 Substance classes considered within the model. Substance class

Species/elements

Gas

Ar, C2H4, C2H6, C3H8, CH4, CO, CO2, H2, H2O(g), H2S, HCl, HCN, N2, N2O, NH3, NO, O2, SO2 C, H, O, N, S, Cl K2O, MgO, CaO, SiO2, Mg2SiO4, Fe2SiO4, MgCO3, CaCO3, CaSO4, CaOH2 H2O(l), H2O(g)

Organic Inorganic Pure water/ steam

Concerning the heterogeneous fuel conversion rate, an equilibrium approach is not reasonable because of the large deviation from thermodynamic equilibrium at typical gasification conditions (temperature, mean residence time) due to the comparatively low reaction rates of heterogeneous reactions. Hence, in the present work amount and composition of residue char are specified within the model based on empirical data. To quantify the fuel conversion rate, the parameters carbon conversion (Eq. (4)) and fixed char conversion (Eq. (5)) are implemented in the reformer model. _ C;res and m _ res;waf constitute the mass flow of carbon Whereby m _ C;fuel and m _ pyr;waf are the toand residue char leaving the reformer, m tal carbon and pyrolysis char (fixed char) input.

Carbon conv ¼



1

 _ C;res m 100% _ C;fuel m

  _ res;waf m 100% Fixed char conv ¼ 1  _ pyr;waf m

ð4Þ

ð5Þ

It is commonly known that the conversion rate is strongly influenced by the residence time of char in the reformer. The complete decoupling of combustion and gasification within the HPR allows the variation of the char residence time in a wide range by adjusting the level in the reformer (height of the reformer fluidized bed). Generally, the higher the level, the longer the char residence time. Consequently, the char conversion rate can also be varied in a wide range during plant operation. Regarding this fact, the specification of the conversion rate is quite reasonable, especially for performing a sensitivity analysis. Such an empirically modified approach has also been used in previous literature by other authors [12,13]. Another approach to define the conversion rate could be the inclusion of kinetics by implementation of kinetic rate expressions for various heterogeneous gasification reactions. This would lead to significantly more complex model equations and thus make the model less convenient for parametric studies. However, if constructive design analyzes should be carried out, the implementation of kinetics is essential. 3.1.2. Tar reforming Due to the fact that tar is defined as additional loading of the gas stream in IPSEpro, two models are required to simulate the tar reforming process. In the first model tar and water steam are converted into the gas species CO and H2 according to the following equation:

Cx Hy Oz þ ðx  zÞH2 O $ xCO þ

y 2

 þ x  z H2

ðR8Þ

To determine the composition of the drain gas stream elementary mass balances for the elements C, H and O including the gas components H2O, H2 and CO have been implemented. Furthermore, the model contains equations for mass conservation of the components which are not affected by tar conversion. The tar conversion rate is set as an input parameter by the user. Analogous to the reformer model, the energy balance considers enthalpies of formation since it is calculated using total thermodynamic enthalpies (Eq. (1)). In the second model the reformation of the preserved gas mixture is carried out. Therefore, elementary mass balances, an energy balance and thermodynamic equilibrium relations for (R1–R3) in form of Eq. (2) are implemented in the model. Similar to the reformer model, either thermodynamic equilibrium of the implemented reactions (Eq. (2)) or the outlet composition of the corresponding components may be defined. The pressure loss in the packed bed is calculated according to the Ergun equation [21].

Dp ð1  wÞ2 gv 1  w qf v 2 ¼ 150 þ 1:75 2 p L d d w3 w3 p

ð6Þ

T. Gröbl et al. / Applied Energy 97 (2012) 451–461

3.1.3. Alkali removal For separation of the alkali elements a model of a filter cartridge was developed. The separation efficiency has to be specified according to the manufacturer data sheet. The pressure drop is calculated according to the equation for gas filtration with a filter element:

_ g qf V_ Vs Dp ¼ þ A a bA

! ð7Þ

with the specific permeability coefficients a and b, the filter thickness s and the filter surface A. In order to determine the number of required filter elements, an equation for the surface filter load as well as for the specific filter load was additionally implemented in the model. 3.1.4. Sulfur capture As a result of the reducing atmosphere during allothermal gasification, sulphur is given completely in the form of H2S. The removal of H2S by using a copper-oxide sorbent can be described with the following thermodynamic equilibrium equation:

CuO þ H2 S $ CuS þ H2 O:

ðR9Þ

Besides this equation, mass and energy balances are solved within the model. The capture rate can optionally be calculated according to thermodynamic equilibrium for reaction (R9), or defined as an input parameter. In case of a user defined capture rate, the implemented model checks to what extend thermodynamic equilibrium is reached in the drain gas according to Eq. (2). The pressure loss in the packed bed is calculated according to Eq. (6). Since the species CuO and CuS are not considered within the PGP-Lib by default, thermodynamic data for calculation of the total enthalpies has to be provided to the model. Therefore total thermodynamic enthalpies are calculated according to Eq. (1). Data for 0 standard enthalpies of formations hf ;298 are taken from literature [22] and for description of the isobaric heat capacities cp(T) polynomial functions according to [23] are implemented. Additionally, the thermodynamic equilibrium constant Kp(T) has to be calculated for

455

reaction (R9). This is done by Eq. (3), whereby values for the Gibbs free enthalpy of reaction DG0R ðTÞ are taken from literature [22]. 3.1.5. Methanation The methanation model is quite similar to the second tar reformer model described above. Elementary mass balances, an energy balance and thermodynamic equilibrium relations in form of Eq. (2) are implemented in order to calculate the composition of the drain gas. Here, the water gas shift reaction (R1) and the methane steam reforming reaction (R2) are taken into account. For simulation of the reactor cooling, an additional water stream is considered within the energy balance. 3.1.6. General heat losses Heat losses towards the environment can be defined in each of the described models. Moreover, the relative heat losses related to the thermal power of the fuel or the sensible enthalpy entering the model are determined. 3.1.7. Raw-SNG upgrading The raw-SNG upgrading is not taken into consideration in the present study. A concept study will be conducted at a later time. 3.2. Process integration The models of the single process steps are combined in the IPSEpro Process Simulation Environment (PSE) to build up an overall model from biomass to raw-SNG. A reasonable combination of the models for an optimal thermal integration has been a major challenge of this work. A flowsheet of the complete process can be seen in Fig. 3. The high temperature level of the waste heat urges to connect a useful heat production with the Bio-SNG production. A part of the heat produced in the process is used for air preheating as well as for generation and superheating of process steam. The remaining heat is used for steam generation (excess steam). The way excess steam is used is not further considered in this study. It may be

Fig. 3. Detailed flowsheet of the SNG process.

456

T. Gröbl et al. / Applied Energy 97 (2012) 451–461

delivered to a district heat system for example. However, the resulting coupling (polygeneration) improves the economic efficiency of the arrangement substantially.

4. Validation of the HPR model In order to verify the validity of the modeling assumptions and simplifications, the results obtained with the model are compared to reference values. As reference values for validation, measurements from the aforementioned pilot plant in Pfaffenhofen are used [10]. Additionally, the model results are compared to measurements from a lab-scale gasifier, which has been developed for experimental analysis of the methanation process [24–26]. 4.1. Lab-scale device An important parameter for operation of the Biomass Heatpipe Reformer is the steam excess ratio r, which is defined as the quotient of fed steam to stoichiometric steam demand for complete conversion. Thereby, the fed steam includes the steam brought into the reformer as fluidization and gasification agent and the water content of the biomass feedstock. At the reference case the labscale gasifier is operated with a steam excess ratio of 4. The thermal power of the biomass feed amounts 1 kW. The reformer temperature is 750 °C and the system pressure is 2 barabs. The fixed char conversion was determined to be 90%. An ultimate and proximate analysis of the wood pellets used for operation of the labscale gasifier is given in Table 3. Experiments showed that the methane (CH4) content in the syngas does not correlate with thermodynamic equilibrium [24]. Therefore, the methane steam reforming reaction (R3) cannot be described in a proper way by defining thermodynamic equilibrium within the model. Instead, the CH4 content in is set to the measured value. Table 4 compares the composition of the syngas derived by the model with measurements from the lab-scale gasifier. The high amount of nitrogen (N2) can be attributed to the N2 stream required for pressurizing the fuel supply lock hopper system (0.15 kg/h). As can be seen, the model results show very good concordance with the measured values. A reason for this is the long gas residence time in the lab-scale gasifier.

Concerning the subsequent methanation step it was shown by the work of Kienberger [24] that thermodynamic equilibrium is reached for the drain gas stream of the lab-scale methanation reactor for different conditions. The measured gas composition corresponds to that of thermodynamic equilibrium. Therefore, the assumption of thermodynamic equilibrium for the reactions (R1 and R2), made within the methanation model, is valid and allows realistic description of this process step.

4.2. Pilot plant For validation of the model results, the same operation parameters have to be defined within the model as are used for operation of the pilot plant. The required parameters are taken from the work of Gallmetzer et al.[10]. A fuel analysis of the wood chips used in the pilot plant as well as for further simulation is given in Table 5. The total thermal power input (reformer + combustor) is set to 500 kW. The reformer is operated at a temperature of 820 °C, a pressure of 4 barabs and a steam excess ratio of 3.5. Here, it should be noted that r only includes the steam brought into the reformer as fluidization respectively gasification agent. Contrary to the conventional definition, the water content of the biomass feedstock is not considered within r. The heat losses of the reformer and the combustion chamber are defined with 1% and 15% of the respective thermal power input. The conversion of fixed char was determined in the pilot plant based on a carbon balance over the reformer. The fixed char conversion is specified to be in the range of approximately 60–80%. For simulation a value of 70% is assumed. Table 6 compares composition and LHV of the syngas derived by the model with measurements from the pilot plant. As before, the CH4 content is set to the measured value. The relatively high amount of N2 again comes from pressurizing the fuel feedstock with N2 (5 kg/h). Generally, the calculated results show good concordance with measured values. Slight deviations can be attributed primarily to differences in the fixed char conversion and the steam excess ratio. The thermal power of the syngas yields 351.7 kW, which corresponds to a cold gas efficiency for the HPR of approx. 70.3%. The cold gas efficiency of the pilot plant was found to be up to 70% for the power range of 500–550 kW.

Table 3 Ultimate and proximate analysis of the wood pellets used in the lab-scale gasifier [24]. Ultimate analysis

Proximate analysis

C H O N S Water Ash

Fixed carbon Volatile matter Water Ash

0.4739 kg/kg 0.0540 kg/kg 0.3845 kg/kg 0.0036 kg/kg 0.0006 kg/kg 0.0690 kg/kg 0.0142 kg/kg

C H O N S Water content Ash content Volatiles LHV

0.2632 kg/kg 0.6533 kg/kg 0.0690 kg/kg 0.0142 kg/kg

Table 4 Syngas composition – comparison model results/lab-scale gasifier.

CH4,dry COdry CO2,dry H2,dry N2,dry H2O

Table 5 Fuel data.

Model (vol%)

Lab-scale gasifier [24] (vol%)

6.1 13.1 17.2 37.7 25.9 27.6

6.1 13.1 19.9 37.6 23.3 27.0

50.41 wt% (waf) 6.74 wt% (waf) 42.69 wt% (waf) 0.15 wt% (waf) 0.01 wt% (waf) 19.55 wt% 1.33 wt% 80 wt% (wf) 14780 kJ/kg

Table 6 Composition and LHV of the syngas – comparison model results/pilot plant.

CH4,dry COdry CO2,dry H2,dry N2,dry LHV

Model

Pilot plant [10]

10.0 vol% 16.5 vol% 21.7 vol% 48.3 vol% 3.5 vol% 10.87 MW/Nm3

9–12 vol% 20 vol% 22–24 vol% 40–45 vol% 3–4 vol% 10–11 MW/Nm3

457

T. Gröbl et al. / Applied Energy 97 (2012) 451–461

The comparison of the calculated values with measured values for two different reference cases confirms the validity of the model assumptions and simplifications. 5. Results and discussion 5.1. Input streams and standard settings A fuel analysis of the wood chips used for simulation is given above in Table 5. The ambient temperature is assumed to be 15 °C, which means that water and air are supplied to the process at a temperature of 15 °C. A non-exhaustive list of required settings for the design operation case is summarized in Table 7. In the reformer the homogeneous gas phase reactions are calculated according to thermodynamic equilibrium equations, as explained within the model description. For simulation thermodynamic equilibrium is expected for reactions (R1–R7) (pdeq = 0) but not for (R2). As determined in experiments [10,24], the methane content in the syngas does not correlate with thermodynamic equilibrium. Measurements given in open literature show methane contents in the range between 8 and 12 vol% of dry product gas [5]. This value also correlates with measurements from the pilot plant [10]. The reason for this is that pyrolysis of solid biomass leads to high contents of methane, which do not react completely to thermodynamic equilibrium. To take this into consideration, the parameter pdeq for the methane steam reforming reaction (R2) is set to a constant value which results in 10 vol% methane in dry product gas (pdeq,R2 = 0943). As already mentioned, heterogeneous fuel conversion is not calculated by the model. The conversion rate of the fixed char is assumed to be 70% for the design operating case. Thus, 30% of the amount of fixed char entering the gasification zone is leaving the reformer without taking part at the equilibrium calculation. Furthermore, it is assumed that the ungasified char is composed of pure carbon only. The tar content in the producer gas as well as the composition of tar is also taken from measurements and set as an input parameter. Values for the fixed char conversion as well as for the amount and composition of tar are taken from measurements at the demonstration plant in Pfaffenhofen. Within the tar reforming (R1–R3) and methanation model (R1 and R2) complete thermodynamic equilibrium of the gas phase is assumed, i.e. pdeq is set to 0 for the corresponding reactions. Reactor dimensions for tar reforming, sulfur capture and methanation have been determined concerning to a suitable space velocity and a length/diameter ratio of approximately 3. For the packed bed a porosity of 0,4 and a hydraulic diameter of 3 mm has been assumed. 5.2. Results for the design operation case The syngas composition obtained with the described models and the given parameters for the base design case is shown in Table 8. The thermal power input of 1300 kW corresponds to a total biomass mass flow of 316.6 kg/h and yields syngas thermal power (LHV  gas yield) of 1025.5 kW at the outlet of the reformer. The required heat transferred from the combustion chamber into the reformer amounts 263.4 kW. Table 9 gives the composition of the raw-SNG after the methanation. Because of the strongly exothermic methanation reaction, a cooling power of 200.80 kW is required to reach the outlet temperature of 250 °C. The thermal power of the raw-SNG amounts 884.08 kW. Values for thermal and electric power are summarized in Table 10. It shows, that a thermal power of 448.47 kW is generated by cooling of the product and flue gas and the exothermic methanation. 168.17 kW of the thermal power are used for generation

Table 7 Summary of required settings. Combustion chamber Air excess ratio k Relative heat loss Pressure combustion zone Pressure drop

1.3 5% 1 bar 0.1 bar

Reformer Steam excess ratio r Fixed char conversion Relative heat loss Pressure gasification zone Pressure drop Temperature product gas Temperature flue gas Char content in product gas Tar content in dry product gas CH4 content in dry product gas Thermodynamic equilibrium reactions considered Logarithmic deviation from equilibrium (R1, R3–R7) Logarithmic deviation from equilibrium (R2) Thermal power of biomass (reformer + combustor)

2 70% 0.5% 5 bar 0.1 bar 800 °C 950 °C 2 g/Nm3 5,5 g/Nm3 dry 10 vol% drya R1–R7 0 0.9426 1300 kW

Particulate removal product gas Separation efficiency Relative heat loss Pressure drop Temperature difference gas/solid

90% 2% 0.02 bar 0 °C

Tar reforming Tar conversion Char conversion Relative heat loss Thermodynamic equilibrium reaction considered Logarithmic deviation from equilibrium (R1–R3)

100% 100% 1% R1–R3 0

Product gas cooling (co-current) Pressure drop product gas Pressure drop water Product gas temperature at the outlet

0.01 bar 0.1 bar 450 °C

Alkali removal Separation efficiency Relative heat loss Temperature difference gas/solid

98% 5% 0 °C

Sulfur capture Capture rate Relative heat loss Methanation Temperature raw-SNG Pressure drop cooling water Thermodynamic equilibrium reaction considered Logarithmic deviation from equilibrium (R1 and R2) Temperature raw-SNG after H2O condensation

250 °C 0.1 bar R1 and R2 0 90 °C

Air preheater (counter-current) Air preheating temperature Pressure drop flue gas Pressure drop air

600 °C 0.01 bar 0.1 bar

Steam superheater (counter-current) Pressure drop flue gas Pressure drop steam Pinch point

0.01 bar 0.1 bar 15 °C

Particulate removal flue gas Separation efficiency Relative heat loss Pressure drop Temperature difference gas/solid

90% 2% 0.02 bar 0 °C

100% 1%

Flue gas cooling Flue gas temperature at the outlet Motors

150 °C

gmech gel

0.98 0.96

Compressors, pumps

gmech gs

0.98 0.7

a The CH4 content is defined via the parameter pdeq for the methane steam reforming reaction (R2).

458

T. Gröbl et al. / Applied Energy 97 (2012) 451–461

Table 8 Composition and LHV of the raw-syngas.

CH4 CO CO2 H2 H2O Others LHV

Humid

Dry

8.05 vol% 22.10 vol% 12.26 vol% 38.03 vol% 19.50 vol% 0.06 vol% 9.774 MJ/Nm3

10 vol% 27.45 vol% 15.23 vol% 47.24 vol% – 0.08 vol% 12.141 kJ/Nm3

Table 9 Composition and LHV of the raw-SNG.

CH4 CO2 H2 H2O Others LHV

Humid

Dry

33.34 vol% 28.05 vol% 0.49 vol% 38.03 vol% 0.09 vol% 11.986 MJ/Nm3

53.8 vol% 45.26 vol% 0.79 vol% – 0.15 vol% 19.343 MJ/Nm3

Table 10 Thermal and electric power. Product gas cooling – evaporator Methanation Air preheater-flue gas Steam superheater-flue gas Evaporator-flue gas Economizer-flue gas Excess steam (for export) Compression of combustion air Suction draught flue gas Feed water pump

50.23 kW 200.80 kW 112.01 kW 11.76 kW 54.74 kW 18.92 kW 280.31 kW 4.13 kW 3.10 kW 0.17 kW

and superheating of process steam as well as for internal air preheating. The way excess steam (280.31 kW) is used is not further considered in this study as already mentioned afore. The total electric power input amounts 7.09 kW. To evaluate the process performance the parameters cold gas efficiency, thermal efficiency and overall thermal utilization rate, including the heating rate of the excess steam PEH, are used. Cold gas efficiency and thermal efficiency are calculated for both, the gasification process in the HPR and the entire SNG-process. The overall thermal utilization rate, however, is only calculated for the whole SNG process (Table 11). A promising cold gas efficiency of 78.9% was calculated for the conversion process of wood to raw-SNG. It has to be considered, that this value does not include CO2 separation of the raw-SNG. Therefore, the final value will be slightly lower. In case of combining the SNG production with district heat generation an overall thermal utilization rate of approximately 89.0% can be achieved. By cooling the raw-SNG to 90 °C and considering the heat of H2O condensation, the overall thermal utilization rate can be further

increased to a value of approximately 93.5%. It is obviously seen from these results, that thermal integration has a determining influence on process performance. The use of process heat, either inside the process or external in form of excess steam, is therefore a key factor for the efficient operation of such a plant. 5.3. Influence of steam excess ratio r The steam excess ratio constitutes an important operation parameter for the HPR, since steam is used for both, as gasification and fluidization agent. The effect of r the on composition and heating value of dry syngas is illustrated in Fig. 4. An increase in feed steam shifts the equilibrium of the water gas shift reaction (R1) towards the products H2 and CO2, whereas CO declines. Consequently, the H2/CO ratio can be adjusted by varying r. Because a higher partial pressure of H2O favors methane decomposition according to reaction (R2), the CH4 content drops with rising r. The decrease of the lower heating value (LHV) of dry product gas can be attributed to the decrease of the C/H ratio. The rising water content leads to an almost linear decrease of product gas thermal power, whereby the cold gas efficiency drops from 80.8% to 74.8%. For r = 4, the H2/CO ratio before the methanation unit (after tar reforming) has a value of 3, which is stoichiometrically required for the methanation reaction (reverse reaction of methane steam reforming reaction (R2)). If using a methanation catalyst, that allows water gas shift reaction concurrently to the methanation reaction, a complete conversion of the reactants is possible also for non-stoichiometric H2/CO ratios [24]. The composition and LHV of the dry raw-SNG by contrast, show only little dependence on r. This can be explained due to the fact that thermodynamic equilibrium of the methane steam reforming reaction is almost entirely at the side of CH4 at 250 °C. Additionally, the water gas shift reaction (R1) is considered within the methanation model. Thus, in case of substoichiometric H2/CO ratio, the missing H2 is generated by the shift reaction (R1) from the resulting water of the methanation reaction. In case of excess H2, CO2 is converted together with excess H2 to CO. Therefore, a complete conversion of the reactants CO and H2 is assured. In analogy to the gasification process, the cold gas efficiency of the whole process decreases from 70.7% to 63.0%. As the demand on process steam increases, the amount of excess steam and thus the overall thermal utilization rate decreases from 92.8% to 79.0%. Although a higher steam excess ratio leads to lower efficiencies, it cannot be chosen arbitrarily low. A minimum steam feed is required for fluidization. Furthermore, higher steam content enhances tar reforming during gasification as well as in the subsequent tar reforming unit according to reaction (R8), and reduces the risk of carbon decomposition during methanation [24,27]. Higher steam to biomass ratio also leads to higher biomass conversion efficiency, since the higher steam partial pressure enhances the heterogeneous char conversion [28]. A challenge for future work will therefore be the determination of a steam excess ratio which results in suitable operation conditions (high char conversion, low tar content, low carbon deposits on the methanation catalyst) accompanied by a high cold gas efficiency.

Table 11 Efficiency of the process. Parameter

b

Definition

Cold gas efficiency

gchem ¼

Thermal efficiency

gth ¼

Overall thermal utilization rate

gges

Heat of H2O condensation in raw-SNG included.

_ PG LHV PG m _ F LHVF m

 100 P  100 P _ PG LHVPG þPEH  m P el;C ¼  100 _ F LHVF m _ PG LHVPG  m Pel;C _ F LHVF m

HPR

SNG

78.9%

68.0%

78.3%

67.4%

–

89.0%/93.5%b

T. Gröbl et al. / Applied Energy 97 (2012) 451–461

Fig. 4. Syngas composition and LHV versus steam excess ratio.

Fig. 5. Syngas composition and LHV versus gasification temperature.

5.4. Influence of gasification temperature Fig. 5 shows the effect of gasification temperature on product gas composition and LHV. A strong influence on the CH4 content can be seen. As the methane steam reforming reaction (R2) is favored at higher temperatures, the CH4 content decreases whereby H2 and CO increase. In case of assuming thermodynamic equilibrium for this reaction, no methane would be produced above approximately 800 °C, as reported by Schuster et al. [11]. Because of thermodynamic equilibrium of the water gas shift reaction (R1) the CO2 content decreases at the same time due to the rising partial pressure of H2. Generally, a higher gasification temperature leads to a decrease in the H2/CO ratio. The composition of the raw SNG remains unchanged. For heating the product gas at the higher gasification temperature, more biomass has to be fed into the combustion chamber, by which the cold gas efficiency drops from 83.1% to 75.7% for gasification and from 76.7% to 61.9% for the entire process. The overall thermal utilization rate nearly stays constant (89.6–88.6%), because of the increasing sensible heat of product and flue gas in case of rising gasification temperature. Nevertheless, the choice of gasification temperature is limited as the char conversion rate decreases and the tar content in the syngas increases with decreasing temperature [29].

459

Fig. 6. Syngas composition and LHV versus fixed char conversion rate.

decreases, more steam is used for gasification of the biomass. The decreasing H2O content in the product gas shifts the equilibrium of the reactions (R1 and R2) towards the reactants, whereby CO and CH4 increase. Simultaneously, H2 and CO2 decrease. Due to the higher C/H ratio the lower heating value rises (Fig. 6). The ungasified char is fed into the combustion zone and therefore lowers the amount of biomass required. At a certain char conversion rate (7%) no additional biomass is required in the combustion chamber. At this point the HPR is operated at the so called char limit. For lower conversion rates, excess char has to be discharged from the process. A fixed char conversion of 10% results in a cold gas efficiency of 77.9% for the HPR or 75.4% for the entire process. The overall thermal utilization rate amounts 77.4%. With rising conversion, the LHV of the product gas and therefore the cold gas efficiency increases for both, the gasification and the whole process. For complete conversion, the cold gas efficiency reaches 79.3% for gasification and 68.7% for the entire process. The overall thermal utilization rate is heightened to 89.6%. With regard to the process efficiency, the gasifier should therefore be operated at a high fuel conversion rate, even if the residue char is used as a feedstock for the combustor. In the HPR this could be achieved quite easily by ensuring a correspondingly long residence time, as already explained in the model description section (3.1.1). 5.6. Influence of biomass water content As expected, a rising water content of the biomass leads to a drop in the cold gas efficiency for both, the gasification and the whole process. The reason for this is the increasing energy required for vaporization of the water brought into the reforming zone with the biomass feedstock. An increase in the water content from 0 to 35 wt% results in a decrease of the cold gas efficiency from 82.0% to 74.4% for the HPR and 70.7–64.2% for the whole SNG process. As a consequence of this, the biomass feedstock should be supplied as dry as possible. The implementation of a fuel drying system could further enhance process efficiency. Furthermore, it is important to note that rising water content results in a decrease of fluidization steam at constant steam excess ratio r. In order to ensure adequate fluidization of the reformer fluidized bed, the water content must not exceed a certain value for a given r. 5.7. Influence of air preheating and steam superheating temperature

5.5. Influence of fixed char conversion rate The fixed char conversion affects mass and energy balance of the HPR. As the amount of solid char (pure carbon) leaving the reformer without taking part at the equilibrium reactions

An increase of both, the air preheating and steam superheating temperature has a positive effect on the efficiency of the HPR. Due to the fact that the flue gas stream provides only a certain amount of sensible heat, these two temperatures, however, behave in opposite directions (Fig. 7). It is therefore necessary to examine

460

T. Gröbl et al. / Applied Energy 97 (2012) 451–461

efficiency. In this context, furthermore it is of great interest for operation of the HPR that air preheating stronger impinge on the efficiency than steam superheating. With a relative sensitivity coefficient of 0.13 the air preheating temperature is the most influencing operation parameter which can be modified in a wide range for plant operation. The influence of fixed char conversion on contrary is comparatively low. 6. Conclusions

Fig. 7. Steam superheating temperature and cold gas efficiency versus air preheating temperature.

Table 12 Linear sensitivity coefficient of cold gas efficiency. Operation parameter

sj,rel – HPR

sj,rel – SNG

r

0.044 0.378 +0.015 0.055 +0.130

0.069 0.888 +0.025 0.055 +0.130

Gasification temperature Fixed char conversion Biomass water content Air preheating temperature

which of the two effects predominates. I.e. does the air preheating or the steam superheating stronger impinge on the efficiency of the HPR? Calculation results show that for varying the air preheating temperature from 200–700 °C, the cold gas efficiency of the HPR advances from 70.3% to 80.5%. Concurrently, the cold gas efficiency of the complete SNG process increases from 60.6% to 69.4% (Fig. 7). This is because of the much higher relative heat loss (heat loss in relation to fuel thermal power input) of the combustion chamber compared to the reformer (see Table 7). Therefore, process optimization with regard to the cold gas efficiency should primarily be achieved by increasing (respectively maximizing) the air preheating temperature. The performed sensitivity study clearly shows the influence of the most important operation conditions. Even if the actual performance of the HPR may deviate from the predicted (equilibrium) conditions for certain combinations of operating parameters, the overall trends stay essentially unchanged. In order to provide a comparable measure of the influence of various operating parameters on process performance, relative linear sensitivity coefficients according to Eq. (8) are calculated.

Acknowledgments

Dgchem

sj;rel ¼

gchem Dj j

A simulation model for the conversion of solid biomass into raw-SNG was developed using commercial process simulation software IPSEpro. Based on this model, mass balances, energy balances, product gas compositions and process efficiencies were determined. A process flowsheet for an optimum thermal integration was provided and the influence of different operation conditions of the used Biomass Heatpipe Reformer was analyzed. A promising cold gas efficiency of approximately 79% for the HPR and 68% for the entire process can be expected for the design operation case. This clearly demonstrates the potential of decentralized small-scale SNG production by the combination of pressurized gasification, hot gas cleaning and methanation. Results from the sensitivity analysis show an increase of the cold gas efficiency of the HPR with decreasing steam excess ratio, gasification temperature and biomass water content. An increase of fixed char conversion and air preheating temperature, by contrast, has a positive effect on the HPR cold gas efficiency. Among the analyzed parameters, the strongest influence on the cold gas efficiency of the HPR and the entire SNG process can be attributed to the gasification temperature and the air preheating temperature. The study confirms that thermodynamic equilibrium models form a useful tool for process design and optimization as well as for sensitivity studies of various operating parameters. However, pure equilibrium models have their limitations, as total equilibrium conditions may not be achieved in a real gasifier. Especially the heterogeneous char conversion is usually controlled by non-equilibrium factors, but also homogeneous reactions like the methane steam reforming do not reach thermodynamic equilibrium. This has to be considered by empirical data, if available, as done in the present work, or by implementation of kinetic rate expressions. A reasonable modification of the HPR model for future work could be the inclusion of a kinetic rate approach for various heterogeneous gasification reactions in order to predict the non-equilibrium fixed char conversion. Additionally, the implementation of the logarithmic deviation from thermodynamic equilibrium for the methane steam reforming reaction as a function of temperature, instead of a constant value, could possibly further improve the quality of the model results.

ð8Þ

Here, j describes the reference value of a certain operation parameter and Dj the according variation. As a reference point, the above determined design operation case is used. The relative sensitivity sj,rel is calculated for a variation of ±10% around the reference value for each parameter for both, the HPR as well as for the entire SNG process. The results are summarized in Table 12. Among the analyzed parameters, the strongest influence on the cold gas efficiency of the HPR and the entire SNG process can be attributed to the gasification temperature, since increasing temperature leads to a higher biomass feed into the combustor which lowers both, the mass flow and the LHV of the syngas. But, as already mentioned in Section 5.4, the choice of the gasification temperature is restricted. The results further show that the air preheating temperature also has a considerable influence on the

The authors would like to thank the Austrian Research Promotion Agency FFG for funding the project ‘‘DistributedSNG – Decentralized Conversion of Biomass into Substitute Natural Gas’’ (FFG 821863) and the project partner agnion (www.agnion.de) for the constructive and pleasant working relationship. References [1] European Commission. Combating climate change – the EU leads the way; 2007. . [2] Budzianowski W. Opportunities for bioenergy in Poland: biogas and solid biomass fuelled power plants. Rynek Energii 2011;94(3):138–46. [3] Karl J. Distributed generation of substitute natural gas from biomass. In: Proceedings 16th European biomass conference and exhibition; 2008. p. 2515–9. [4] Rehling B, Hofbauer H, Rauch R, Tremmel H, Aichernig C, Schildhauer T, et al. Bio-SNG – first results of the 1 MW pilot and demonstration unit at Güssing; 2009.

T. Gröbl et al. / Applied Energy 97 (2012) 451–461 [5] Bolhár-Nordenkampf M, Rauch R, Bosch K, Aichernig C, Hofbauer H. Biomass CHP plant güssing – using gasification for power generation. In: Proceedings of the 2nd RCETCE conference, Phuket, Thailand; 2003. p. 567–72. [6] Hofbauer H, Rauch R, Fürnsinn S, Aichernig C. Energiezentrale Güssing. Berichte aus Energie und Umweltforschung; 79/2006. [7] Karl J, Gallmetzer G, Hochleitner T, Kienberger T, Schweiger A, Kröner M. Small-scale generation of substitute natural gas. In: Proceedings of the 1st international conference on polygeneration strategies, Vienna/Austria; 2009. [8] Agnion web site http://www.agnion.de. [9] Karellas S, Karl J, Kakaras E. An innovative biomass gasification process and its coupling with microturbine and fuel cell systems. Energy 2008;33:284–91. [10] Gallmetzer G, Ackermann P, Schweiger A, Kienberger T, Gröbl T, Walter H, Zankl M, Kröner M. The agnion Heatpipe-Reformer – operating experiences and evaluation of fuel conversion and syngas composition. In: Proceedings of the 3rd international conference on polygeneration strategies, Vienna/Austria; 2011. p. 13–22. [11] Schuster G, Löffler G, Weigl K, Hofbauer H. Biomass steam gasification – an extensive parametric modeling study. Bioresource Technol 2001;77:71–9. [12] Duret A, Friedli C, Maréchal F. Process design of Synthetic Natural Gas (SNG) production using wood gasification. J Clean Prod 2005;13:1434–46. [13] Damiani L, Trucco A. Biomass gasification modelling: an equilibrium model, modified to reproduce the operation of actual reactors. In: Proceedings of ASME Turbo Expo, Orlando, Florida, USA; 2009. [14] Li X, Grace JR, Watkinson AP, Lim CJ, Ergüdenler A. Equilibrium modeling of gasification: a free energy minimization approach and its application to a circulating fluidized bed coal gasifier. Fuel 2001;80:195–207. [15] Buragohain B, Mahanta P, Moholkar V. Thermodynamic optimization of biomass gasification for decentralized power generation and Fischer– Tropsch synthesis. Energy 2010;35:2557–79.

461

[16] Puig-Arnavat M, Bruno J, Coronas A. Review of biomass gasification models. Renew Sustain Energy Rev 2010;14:2841–51. [17] Technical University of Munich. Vorrichtung zur Vergasung biogener Einsatzstoffe. Patent DE 199 26 202 C 1; 2001. [18] SimTech web site http://simtechnology.com. [19] SimTech. Pyrolysis and Gasification Process Library. User manual; 2008. [20] Pröll T, Hofbauer H. Development and application of a simulation tool for biomass gasification based processes. Int J Chem React Eng 2008;6:A89. [21] VDI-Wärmeatlas. Pressure loss for flow through fixed beds. Chapter Laf; 2006. [22] Barin I. Thermochemical data of pure substances. 3rd ed. VCH; 1995. [23] Binnewies M, Milke E. Thermochemical data of elements and compounds. Wiley-VCH; 1999. [24] Kienberger T. Methanierung biogener Synthesegase mit Hinblick auf die direkte Umsetzung von höheren Kohlenwasserstoffen, Doctoral thesis. TU Graz; October 2010. [25] Kienberger T. Methanation with an allothermal smart lab-scale gasification system. In: Proceedings 17th biomass conference, Hamburg/Deutschland; 2009. [26] Kienberger T. Tests on methanation with tar and sulphur loaded syngas. In: Proceedings of the 1st international conference on polygeneration strategies, Vienna/Austria; 2009. [27] Kopyscinski J. Production of synthetic natural gas in a fluidized bed reactor, Doctoral thesis, ETH Zuric; December 2009. [28] Maschio G, Lucchesi A, Stoppato G. Production of syngas from biomass. Bioresour Technol 1994;48:119–26. [29] Kumar A, Jones D, Hanna M. Thermochemical biomass gasification: a review of the current status of the technology. Energies 2009;2:556–81.