ADVANCED PROCESS MODELS FOR BIOMASS GASIFIERS E. Biagini1, L. Masoni1, M. Simone2, E. Bargagna2, G. Pannocchia2, C. Nicolella2, L. Tognotti2 1 Consorzio Pisa Ricerche – Divisione Energia Ambiente, Lungarno Mediceo, 40, 56125 Pisa – ITALY 2 Dipartimento di Ingegneria Chimica – Università di Pisa, Via Diotisalvi, 2, 56127 – ITALY
ABSTRACT: The optimization of biomass gasification should be studied with advanced models to evaluate the effect of the operating conditions, quantify the by-products (CO2, CH4, tar) and compare different reactor configurations. In this work different gasifiers (fixed beds, fluidized beds, entrained flow reactors) are modeled with Aspen Plus® according to an originally developed procedure. The innovative feature of the modeling procedure for all cases consists in the separation of the steps of solid fuel gasification (devolatilization, oxidation, gasification of the char, homogeneous reactions and tar cracking) and the development of dedicated sub-models (by adapting conventional blocks of the software or implementing dedicated sub-models). All steps are connected to respect material and heat balances according to the gasifier configuration. In addition a detailed description of the downdraft gasifier is realized as a distributed domain of several CSTRs and solving energy and mass equations for gas and solid phases with gPROMS software. The entrained flow reactor is also developed with Aspen HYSYS® to give a more comprehensive and automatic solution. All gasifier models developed in this work are powerful tools to be integrated in process study and optimization analysis. Keywords: gasification, pyrolysis, fixed bed, dual fluidized bed, fluidized bed
1
INTRODUCTION
Biomass gasification is an attractive process to convert a solid fuel into a gaseous product. Although gasification is a relatively old process, the versatility of the process (with production of syngas, electricity, hydrogen or chemicals) and the multiplicity of technological solutions (fixed beds, moving beds, fluidized beds and entrained flow reactors) make it a current topic of investigation. Process studies should be performed for defining the best process configurations and optimizing the operating conditions [1-2]. The gasification reactor can be designed under very different solutions [3]. The heat needed can be provided by partial oxidation of biomass with air or pure oxygen, or by sand recirculation. Steam may be added to promote gasification. Temperatures, pressures and residence times vary in wide ranges depending on the technological configuration. In spite of all these differences, most process studies in the literature modeled the gasifier as an equilibrium reactor. This approach is indeed fundamental for a preliminary study but hardly suitable for process analysis and optimization procedures. Some issues arise when introducing the equilibrium hypothesis in optimization studies, as detailed below: 1. the relation between gasification temperature and the operating conditions is limited, as essentially two conditions can be set: isothermal or adiabatic conditions. In all cases a realistic thermal profile can not be introduced, nor a heat recovery (for instance in the reactor jacket). In the real reactor, the reactor temperature should arise from a global heat balance dependent on the conditions (above all the oxygen-to-fuel ratio), and this will be a goal of this work; 2. by-products in the syngas (e.g. CH4 and CO2) are generally underestimated even though their value is fundamental for the realistic evaluation of the process efficiency. Residual char is not predicted in equilibrium calculations, while the conditions for complete conversion should be determined to assure high efficiencies and avoid problems in downstream units. Similarly, tar is not predicted in most studies: its quantification is actually fundamental to estimate the process efficiency and the quality of the syngas
produced; 3. gasification is a complex ensemble of chemical and physical phenomena. Each step can be operated under different conditions and the configurations of gasifiers can be compared only by developing a detailed model. This is also the case of reactors that can be hardly represented with an equilibrium reactor (e.g., due to the low temperatures and residence times achieved). For all the above points a “gasifier model” should be developed instead of a “gasification model”. So, the aim of this work is the development of a procedure for modeling different gasifiers and show some examples of gasifier models.
2
MODELLING METHODOLOGY
A general procedure is developed to represent different gasifiers as multizonal models. The main points are summarized here and discussed in the next sections along with some examples for different reactors: definition of the functional scheme of the gasifier; separation of the characteristic steps of solid fuel gasification (devolatilization, oxidation, gasification of the char, homogeneous reactions and tar cracking); development of sub-models of each step (by adapting conventional blocks of the software or implementing specific models); connection of all steps to respect the material and heat balances according to the gasifier configuration. A pyramidal approach is developed dividing the phenomena occurring in the gasifier on different levels. (Figure 1). On the first one the evaluation of the heat and mass transfer phenomena at the molecular level should be based on the operating conditions and allows the initial reactions to be described. The pyrolysis model is the basic step of all gasification models and is known to depend strongly on the operating conditions (temperature, residence time and heating rate) besides the fuel characteristics. The homogeneous reactions are fast and
connected to the previous evaluations. Heterogeneous reactions (involving the solid char particle) are the controlling step of the entire system and are studied on a second level as the consequent transformations involve the particle (via size variation, fragmentation, ash distribution and porosity evolution phenomena). Diffusion of gasifying agents (O2, H2O and CO2 above all), kinetics of char reactions, diffusion of gaseous products should be represented in a realistic model (considering intra-particle phenomena) to take into account the variation of the conditions during the gasification. Gas-particle interactions, gas and solid fluiddynamic, solid-solid interaction (for instance in cogasification applications where different fuels can interact) should be studied on a reactor level by considering the reactor configuration. Also the heat transfer on a macro-scale (e.g., presence of cooling jackets or heat transfer surfaces) can be described only once the geometry of the gasifier is defined. Finally the gasifier model should be validated with experimental data. Lab-scale reactors can be used to validate decoupled sub-models on molecular and particle levels. Pilot-scale and large-scale gasifiers can be used to validate the entire models.
LEVEL 2 PILOT-SCALE GASIFIER LEVEL 1 LAB-SCALE GASIFIER
LEVEL 3 LARGE-SCALE GASIFIER
Validation GASIFIER MODEL
REACTOR LEVEL 3 PARTICLE LEVEL 2 MOLECULAR LEVEL 1
Gas-Solid Interactions
Intra-Particle Phenomena
Heat and Mass Transfer Phenomena
Gasifier Geometry
Particle Evolution
Pyrolysis Description
Heterogeneous Reactions
Gas-Phase Reactions
Figure 1: Pyramidal approach for the development of a gasifier model
based models. The peculiarities and the capabilities of these codes will be discussed in this work.
3
DESCRIPTION OF MAIN SUB-MODELS
All reactive sub-models are represented as Kinetic Reactors (Plug Flow Reactor or Continuous Stirred Tank Reactor depending on the reactor configuration). Different thermal options (adiabatic, constant temperature, thermal profile, constant coolant temperature) can be set for the heat transfer according to the reactor configuration. The list of all reactions is given in Table 1. 3.1 Devolatilization sub-model The first reacting step of the biomass is the devolatilization. It is a thermal decomposition that produces a solid residue (char, that will be the reactant in the following gasification reactions), a condensable organic product (tar) and the main gaseous species (CO, CO2, CH4, H2O, H2, C2H4, N2, NH3, HCN, H2S, COS). No conventional block can represent this step in any commercial codes. Here, a structural model (ABCD Advanced Biomass and Coal Devolatilization model [4]) is used for the biomass devolatilization. As a matter of fact this model can simulate also the devolatilization of different rank coals. The ABCD model gives the yield of macro-products and the speciation of gases once the fuel composition and the operating conditions are given. The ABCD code can be hardly implemented in Aspen Plus® because of the expensive computational cost. Therefore, a User Routine is developed for the devolatilization step (scheme of Figure 2). Basically, it consists of a database and a calculation function. The former is created with the results of off-line simulations of the ABCD model applied to the devolatilization of a specific biomass in a wide range of pressures and temperatures. The calculation function dialogues with the main model by receiving the actual values of temperature and pressure, interpolating the results of the database and returning the speciation of the devolatilization product stream. In all cases the material balance is verified.
DEVO USER ROUTINE
The main sub-models will be described in the following section for the reacting steps. They are adapted for the actual configuration of the gasifier once the particle and reactor levels are defined. Some examples of gasifier models will be discussed in section 4. All models are developed using Aspen Plus® and the model approach will be compared. Fixed bed (downdraft), fluidized beds (circulating and dual beds) and entrained flow gasifiers are studied here. Only in the first case the small scale justifies the feed of biomass alone, while in the other cases a cogasification of biomass with coal can be studied. All gasifier models are developed in view to be linked to other unit models for studying different processes (for example hydrogen production) in a global Aspen Plus® model. Also other codes are used for modeling some gasifiers (gPROMS for a downdraft reactor and Aspen HYSYS® for an entrained flow gasifier). The methodological approach is similar to the Aspen Plus®
(T,P)
MODEL in the main code
Calculation function (interpolation of data and normalization)
devo products
PYROLYSIS DATABASE P 30 600 800 1000 %char %tar
### ###
### ###
### ###
%H2O %CO2
### ###
### ###
### ###
%CO %CH4
### ###
### ###
### ###
%H2
###
###
###
ABCD model off-line simulations: biomass composition T = 600 – 1600 °C P = 1 – 40 bar
Figure 2: Scheme of the devolatilization block
3.2 Homogeneous reactions sub-model The homogeneous reactions are modeled assuming parameters from literature. The general rate expression is the following:
E rj = A jT m exp − j ∏ cin RT i where A and E are the kinetic parameters of j-th reactions, m is the exponent of the temperature, c is the concentration of reactants and n is the order of the reaction with respect to that reactant. For reversible
reactions (e.g., the water gas shift reaction) the last term involves also the concentration of the products. 3.3 Tar reactions sub-model The quantification of the tar is a fundamental issue in gasification models to define the quality of the produced syngas, the downstream cleaning units and the end-use of the process. Therefore it is important to develop a global model that allows one to study the optimal conditions to limit the tar formation or enhance the tar destruction. The tar is formed during the pyrolysis step, while it is destroyed in the subsequent steps: tar cracking (bond scission caused by heat transfer to give light gases), oxidation (if oxygen is available in the reactor) or reforming (with reactions with H2O or CO2 to give partially oxidized products). The tar from biomass is here represented as levoglucosane (C6H10O5) that is the monomer of cellulose (while anthracene C14H10 represents the tar from coal). In all cases, a power law expression is used for the reaction rate as that reported in the previous sub-section. As for the kinetic parameters, they are adapted from [5,6] for the tar cracking. Kinetics of heavy hydrocarbons (with similar molecular weight) are adopted for both tar oxidation and reforming.
3.4 Heterogeneous reactions sub-model Heterogeneous reactions (those involving the char from the previous devolatilization step) are modeled assuming the unreacted core-shrinking model. Both the diffusion and kinetic transfer coefficients are considered for all heterogeneous reactions (of combustion as well as gasification). The general expression is:
rj =
cj ⋅ AS 1 1 + kD kR
where c is the concentration of the j-th species (O2, H2O, CO2, H2), kD is the diffusion coefficients that takes into account the mass transfer to the particle, kR is the reaction coefficient (that has the Arrhenius form) and AS is a surface factor that depends on the external surface area of the particle. This latter parameter varies during the reaction. The shrinking of the particle is modeled by considering the conversion and the ash content of the biomass [7]. In most cases a density constant model is adopted, while only the particle diameter reduces from the initial value to a critical value depending on the ash content of the original fuel. Different sets of kinetic parameters can be found in literature for combustion/gasification systems. Every set was validated and thus can be applied in a specific range of operating conditions. In this work, kinetics of combustion and gasification are divided in two ranges: range 1, for room pressure and relatively low temperatures (800-1200 °C); range 2, for high pressure (20-40 bar) and high temperature (1200-1600°C). Kinetics for the heterogeneous reactions are adapted from [6,8] for range 1 and from [7] for range 2.
Table I: Reaction sets used in the sub-model development Devolatilization reactions fuel → char + tar + ligh gas
Combustion reactions
volatile combustion H2 + ½ O2 → H2O CO + ½ O2 →CO CH4 + 2O2 → CO2 + 2H2O C2H4 + 3O2 →2CO2 + 2H2O char oxidation tar oxidation
C (char) + ½ O2 →CO
C6H10O5(tar) + 17/2 O2 →6CO2 + 5H2O
Gasification reactions char gasification
tar reforming methane reforming water gas shift
4
C (char) + H2O →CO + H2 C (char) + CO2 → 2CO C (char) + 2H2 → CH4
C6H10O5(tar) + H2O →6CO + 6H2 CH4 + H2O → CO + 3H2 CO + H2O → CO2 + H2
DEVELOPMENT OF GASIFIER MODELS
The development of different gasifiers are discussed in this section. The first examples (fixed bed, fluidized beds, entrained flow reactor) are reported for Aspen Plus® applications and compared among them as uniform hypotheses are adopted. The last two examples concerns the downdraft gasifier model developed with gPROMS and the entrained flow reactor model with Aspen HYSYS®. 4.1 Example 1: downdraft gasifier A scheme of the downdraft gasifier is shown in Figure 3a. It is a versatile and proven solution in the range 50-500 kWth. Most reactors are operated with substoichiometric air in commercial units, but also some scientific investigations with steam/oxygen mixtures can be found [9,10]. The internal peak temperature is around 1200 °C, the syngas leaves the reactor at 700-1000°C. In this work we studied a 250 kg/h atmospheric gasifier. The gasifier is modelled with Aspen Plus® according to the scheme of Figure 3a. After the heating of the biomass, a first reactor block (the Yield Reactor DECOMP) represents the devolatilization step, that gives the pyrolysis macro-products (with the speciation of light gases) balancing the moisture and ash content of the biomass. The pyrolysis products are then mixed with the gasifying agents to feed the reactor block (H-REACT) in which all the above equations of combustion, gasification and tar-cracking are modelled in a PFR configuration. The temperature is calculated from the heat balance of the system. The MIXHEAT block collects all heat streams from heater and reactor blocks. A design specification (a calculator tool in Aspen Plus®) is used to iteratively define the reaction temperature in the main reactor, so that the value of the dispersion (the heat stream DISP exiting the MIXHEAT block) converges to the assumed value (5% of the generated heat). Finally, a separator (SEPASH) separates the solid residue from the syngas (according to an efficiency value of 0.9) and a cooler (H-REC) simulates the heat transfer between the syngas and the reactor before the exit.
4.2 Example 2: circulating fluidized bed gasifier The circulating fluidized bed is formed by a riser (where oxidation/gasification reactions occur) and a downcomer (for the recirculation of the sand used as a heat carrier). Fuel, oxygen and steam are fed in the upper sections of the riser. Some syngas (from the downstream units) is also used as a recirculation gas to assure the hydrodynamics of the system. Some modeling aspects of the fast fluidized bed are discussed in this section. The approach is used for simulating both the circulating fluidized bed and the beds of the dual bed gasifier. The high temperature and heating rate allow one to consider the devolatilization step completely separated from the combustion and gasification reactions. The hydrodynamics of the fluidized beds was modeled by following the considerations on a fluidized bed combustor model exposed in [11]. A similar approach was considered also by Corella The reactor bed was divided in two regions (see example in Figure 3b): a dense lower region with a constant suspension density (turbulent fluidized bed); a more dilute upper region with a decreasing suspension density. The upper region is assumed to be axially composed of two zones: the acceleration zone is at the bottom part of the upper region where the solids are accelerated to a constant upward velocity; the fully developed zone is located above the acceleration zone, where the flow characteristics are invariant with height. In the acceleration zone, the axial voidage decreases with the height of the riser. The average value of the voidage between two height values of the riser can be calculated using the expression proposed by Kunii and Levenspiel [13]. The lower region is represented by a single CSTR, while a series of CSTR with decreasing voidage is used to take into account the solid fraction variation in the upper region. In the lower CSTRs both combustion and gasification reactions are considered, while after a certain height only gasification reactions are considered because of the completely oxygen consumption. As discussed before, heterogeneous reactions of combustion and gasification are modelled with the unreacted core-shrinking model, so a mean particle diameter is assumed in each CSTR in order to evaluate the reaction rates. Another important topic is to determine the conditions for the fast fluidization of the beds. The superficial velocity of the gas should be at least higher than the transport velocity [14]. An iterative procedure is implemented to assure the proper superficial velocity by varying the recirculation gas flowrate. Iterations are necessary because variations in the fluidization gas recirculated causes variations in the syngas produced.
4.3 Example 3: dual beds gasifier The dual bed gasifier is formed of two fluidized beds (see Figure 3c): bed1 is fed with the fuel and the hot sand and works as a gasifier; bed2 is fed with air and the cold sand from
bed1 (that contains also some residual char) and work as a combustor to heat the sand. The main pro of this configuration is the use of air (instead of pure oxygen) to get a syngas with no nitrogen. The sand recirculation is used as a heat carrier and variations in the fuel-to-sand ratio can be studied only with a detailed approach. Both beds are modeled with the same approach described in the previous sub-section.
4.4 Example 4: entrained flow gasifier The reactor studied is an entrained flow reactor with quench water (technology of Future Energy, scheme in Figure 3d). The pressure is between 20 and 40 bar. A burner in the top section feeds the fuel and substoichiometric oxygen. Very high peak temperatures are achieved (1500-1600°C), so steam can be added to promote endothermic gasification reactions. A cooling screen shields this part of the gasifier and recovers heat for the production of low pressure steam. A cooling jacket for the production of pressurized hot water envelops the entire reactor. The high gasification temperatures limit the tar in the syngas and make the ash to melt, thus forming a slag on the reactor walls. This slag is cooled and removed by the quench water in the bottom part of the gasifier. In this work we studied a 5000 kg/h gasifier under a fixed pressure of 30 bar. The gasifier is modelled with Aspen Plus® according to the functional scheme of Figure 3d. As in the previous cases, the first step is the devolatilization of the fuel, followed by a separated combustion chamber (modelled as a PFR reactor with the oxidation reactions of table 1) that represents the burner. The high heating rate expected in this reactor allows to justify this hypothesis. The subsequent gasification step includes the other reactions of table 2. The heat produced in the combustion step is used for the other endothermic reactions (pyrolysis and gasification), the pre-heating of the reactants and the heat recovery in the cooling screen. The same iterative method described in 4.1 is applied.
4.5 Example 5: downdraft gasifier with gPROMS An accurate description of the mechanisms and geometry of the fixed bed gasifier described in section 3.1 was done with gPROMS. A distributed model was developed to give an interpretative and diagnostic tool, capable to aid the experimental activity with pilot scale gasifiers, as well as provide indications about the effect of the operating conditions and geometry. The gPROMS model schematizes the gasifier as a 1D domain (only the axial variable distributions are represented) meshed with 300-1000 cells. Figure 3a reports a scheme of the gPROMS model. The simulation of the system requires operational input (gas and solid flow rates, wall temperature), gas and solid compositions and some details about the pyrolysis behaviour of the biomass (devolatilization kinetics, macroproducts distribution and gas species). The model is based on dynamic equations of heat and mass balance. Therefore the ignition behaviour can be simulated as well as variations in the operating conditions. The main outputs of the simulation are the temperatures distribution along the gasifier axis and the
gas species distribution. 4.6 Example 6: entrained flow gasifier with Aspen HYSYS® The entrained flow gasifier described in the previous sub-section 4.4 was modeled also with Aspen HYSYS®. The functional scheme is the same of figure 3d. Aspen HYSYS® has a user-friendly interface, a good thermodynamic package and the possibility of implementing dedicated extensions that can enrich the block library of the software and allow the simulation of unconventional operations (as the devolatilization and gasification steps). The extensions can be written with programming codes (e.g., Visual Basic 6.0) that support the automation, and create Unit Operations that can be integrated in Aspen HYSYS® in an easy, robust and versatile manner. The first extension was created for the devolatilization step. Theoretically the basic approach is the same ABCD model mentioned in the sub-section 3.1. This model is originally written in FORTRAN and is available as an executable file: it requires an input text file and writes the results to an output text file. The Aspen HYSYS® Devolatilization extension developed in this work acquires the data (on fuel characteristics and
operating conditions) to write the input file, runs the executable file and transfers the results from the output file to the main code of Aspen HYSYS® by defining the stream of the Devolatilization Products-1 (see scheme in Figure 4). This is a big step forward that avoids off-line simulations of the ABCD model (with the generation of a DataBase as described in the sub-section 3.1), removes the errors due to interpolation procedures and allows quicker simulations with different fuels (in the previous case we had to build a DataBase for all fuels tested). The second peculiar extension developed in the Aspen HYSYS® model was for the gasification step. Also in this case the unreacted-core shrinking model was implemented (according to the approach described in the sub-section 3.4) in an adiabatic PFR block. It is worth noting that kinetic reactions of solid compounds is not a conventional operation in Aspen HYSYS® [16]. Finally the heat balance is set by algebraically adding the heat streams of all blocks (heaters and reactors) and the heat recovery (cooling screen and cooling jacket) to the block BALANCE-1 (see figure 4). The Adjustment tool (ADJ-1) allows the combustion temperature to be iteratively defined to solve the heat balance.
gPROMS model
Heat streams
biomass inlet
Biomass Mass streams
syngas DRYING ZONE
PFR
Gasifying agent
SOLID PHASE PFR
GAS PHASE PYROLYSIS ZONE
air inlet t=0 IGNITION
OXIDATION ZONE
fixed bed over grate
Dispersion PFR
Dispersion PFR
GASIFICATION ZONE
periodically removed solid
(a)
SynGas
Syngas
CYCLONE
GASIFICATION REACTOR
Sand recirculation Steam Biomass
Ash
Oxygen
Recirculation gas
(b) Figure 3: Sketch of the gasifiers studied and relative scheme for modeling (cont.)
Solid Residues
Syngas
syngas
Exhaust DEDUST
PYROLYSIS/ GASIFICATION REACTOR (BED 1)
gas recirculation
Air
Hot Sand
BED 2
Char +Sand
sand/ char/ ash
GASIFICATION sub-model (n CSTR in series) cation BED 1
Ash
Biomass 20°C
exhaust gases
separation
DEVO sub-model
COMBUSTION sub-model (CSTR) cation air
COMBUSTOR (BED 2)
preheating
biomass
Fluidization Gas (from clean gas section)
hot sand
excess solid
(c) fuel
fuel preheating
gasifying agents preheating
oxygen steam
DEVO sub-model (CPD Db) COOLING SCREEN COMBUSTION sub-model (PFR)
LP steam
GASIFICATION sub-model (PFR)
quench water
hot water
i-th QUENCH (MIXER)
COOLING JACKET
HOMOGENEOUS REACTIONS (PFR)
hot water i-th HEAT RECOVERY pressurized water SEPARATION material streams heat
syngas waste water
(d) streams slag Figure 3: Sketch of the gasifiers studied and relative scheme for modeling
Figure 4: Scheme of the entrained flow gasifier model with Aspen HYSYS®
Table II: Composition of fuels Fuel Poplar wood SA coal
Ultimate Analysis (wt% dry and ash free) C H N S Cl O 51.7 6.47 0.25 0.05 0.01 41.52 81.6 4.84 1.75 1.27 10.54
Proximate Analysis (wt% dry) VM FC ash 85.1 13 1.9 26.7 58.6 14.7
Moisture (wt%) 20 7
RESULTS AND DISCUSSION
Further results can be obtained with this detailed approach. For instance the rate of char conversion due to
12
60
LHV syngas (MJ/Nm3) 10
H2 production
50
8
40
6
30
4
20
tar (gTAR/kg biomass dry)
2
10
0
Hydrogen production (gH2/kg biomass dry)
Some examples are commented here remarking the capability of each approach and the peculiar results with respect to equilibrium or simplified models. It is worth reminding that all models were developed to compare different gasifier configurations and evaluate the effects of the operating conditions on the syngas conversion to optimize the gasifier performance or, in case, the global process. In the first example, the downdraft model developed in Aspen Plus® was used for the gasification of poplar wood (properties in Table 2) with mixtures of oxygen and steam. In all cases the temperature achieved in the gasifier is limited to 1200°C. The results are shown in Figure 6 as functions of the Equivalent Ratio (ER), that is the ratio between the actual oxygen present in the gaseous feed and that needed for the complete oxidation of the fuel. The higher the value of ER, the lower the heating value of the syngas produced in the gasifier (due to the higher oxidation level). So, the conditions for the maximum LHV could be considered for a direct combustion of the syngas (e.g., for power production). The syngas composition should be taken into account for hydrogen production processes. The hydrogen produced in the gasifier shows a maximum for a value of ER near 0.5 (see Figure 5). Above this value the excess oxygen oxidizes the gaseous products and thus a decrease in the production of hydrogen is observed. The syngas contains also CO, CO2 and CH4 (not shown in the figure), so the complete composition should be considered for assessing the performance of the entire process (e.g., the CO can be converted to give additional hydrogen in a downstream water gas shift unit). The tar at the exit of the gasifier is also shown in Figure 5. The high tar content in the syngas for values of ER between 0.35-0.40 makes prohibitive the direct utilization of the syngas in engines or turbines for power production. In this case onerous gas-cleaning units should be installed and a loss in efficiency should be expected. Values of HR slightly higher (0.45-0.50) can be programmed for obtaining a syngas with a lower heating value but with limited tar content. It is worth remarking that this consideration can not be done with an equilibrium model of the gasifier because tar can not be predicted. The gPROMS model of the same gasifier allows one to obtain more detailed results. For instance, the thermal profiles inside the reactor under various conditions are shown in Figure 6. The gas and solid temperatures in the gasifier are reported for two values of the equivalence ratio. The higher the ER, the higher the maximum temperature of both phases. Three different zones can be observed in the thermal profile: 1. the solid phase is heated up by the radiation from the lower hot zone; consequently the gas phase is heated up by convective heat transfer from the solid phase; 2. the temperature peak is caused by the oxidation reactions that mainly rise the gas temperature; 3. after the oxidation reactions the temperature decreases due to gasification reactions and thermal dispersion; in this zone the gas and solid phases reach a thermal equilibrium.
char oxidation and gasification which is an useful information to estimate the reaction front position and plan a sampling position in the reaction bed. Also some dynamic simulation can give interesting evaluations on conditions variations. Pressure drops due to the particle size of the biomass can be estimated. A mixture of biomass and coal (properties in Table 2) is studied for the circulating fluidized bed. A sensitivity study can be performed by varying the biomass-to-coal ratio, the steam-to-fuel ratio or the oxygen-to-fuel ratio (also expressed as equivalence ratio). The example of Figure 7 shows the syngas composition at the exit of the gasifier for fixed values of steam and oxygen (ER = 0.4, steam-to-fuel = 0.2) for the co-gasification of poplar wood and coal. The higher the biomass sharing, the lower the temperature of the gasifier; also the lower the syngas quality (in terms of hydrogen and CO content). Also some more tar can be observed and this can be a consequence of the lower temperature achieved. A mixture of biomass and coal is studied also for the dual beds gasifier. In this case the co-gasification is dictated by the need of residual char that has to be passed from bed1 to bed2 to assure the autothermal feasibility of the system. Some biomass indeed produces a too low quantity of char. This can be observed in Figure 8 where the temperature of both beds are reported as function of the biomass-to-coal ratio. The higher this ratio, the lower the temperature of both beds. Also the residual char from bed1 to bed2 decreases significantly as shown in the figure. Finally, the hydrogen production decreases. Additional analysis can be carried out with this tool in order optimize the system. For instance, the sand-tofuel ratio and the gas recirculation flowrate can be varied to define the optimal conditions.
Specific unit
5
0 0.3
0.35
0.4
0.45
0.5
0.55
0.6
ER
Figure 5: Results of the downdraft gasifier model with Aspen Plus® for the gasification of poplar wood with oxygen/steam mixtures.
Figure 6: Results of the downdraft gasifier model with gPROMS for the gasification of poplar wood (effect of Equivalence Ratio on the temperature of solid and gas phases).
parameters (e.g., efficiency, hydrogen production) considering the critical restrictions (e.g., maximum temperature or tar content). Finally, some example of the results from the entrained flow gasifier developed with Aspen HYSYS® are reported in Figure 10. In this case the optimization results are shown. Starting from the reference values of the operating conditions (peak temperature 1600°C, oxygen-to-fuel ratio 0.8, steam-to-fuel ratio 0.25), they are varied in wide range to maximize the Cold Gas Efficiency:
0.4
T
1100
H2
0.3
1000
CO
0.2
900
CO2
0.1
TAR 0
800 0
0.1
0.2
0.3
0.4
0.5
Biomass-to-Coal ratio
Figure 7: Results of the circulating fluidized bed gasifier model with Aspen Plus® for the gasification of coal/poplar wood blends with oxygen/steam mixtures.
0.4
1100
1050
H2
0.3
1000
0.2
950
900
Tbed2
Temperature (°C)
H2 mole frac in the syngas, char from bed1 to bed2 (kg/kg feed)
CHAR
0.1 850
Tbed1 0
800 0
0.1
0.2
0.3
0.4
0.5
Biomass-to-Coal ratio
Figure 8: Results of the dual bed system model with Aspen Plus® for the gasification of coal/poplar wood blends (bed1: gasifier, bed2: combustor).
LHVsyngas Fsyngas LHV fuel F fuel
where the Low Heating Value and the mass Flowrate of syngas and fuel are compared. For instance, increasing the oxygen-to-fuel ratio between 0.80 to 0.90 and maintaining constant the other operating conditions a maximum in CGE can be found for values around 0.88. Below this value the gasification is not complete, above that excess oxygen oxidized the gaseous products. However, combining all the operating parameters the automatic procedure (implemented by the Hyprotech SQP Optimizer) gave the optimized values reported in Figure 10. The automatic procedures included in the model and powered by the software allow one to define also other targets (for instance the hydrogen production of the entire process) to obtain the optimal conditions. The developed model is a powerful tool for process analysis and optimization.
0.6
2000
0.5 0.4
CO
1800
Char T
1600
H2
0.3
1400
0.2
CO2
0.1
1200
TAR
0 0.2
Peak temperature (°C)
Also in the case of the entrained flow gasifier a biomass/coal blend is considered. The effect of the operating conditions on the performance of the gasifier can be studied. The equivalence ratio, the biomass-tocoal sharing, the ratio between the steam and the fuel blend can be varied to study the syngas composition, the temperature inside all sub-units of the model and the byproducts yields (CH4 and tar) [16]. For example, the syngas produced for the gasification of a 10%wt biomass-to-coal blend in a mixture of oxygen and steam (steam-to-fuel 0.24 kg/kg) is shown in Figure 9. The higher the equivalence ratio, the higher the maximum temperature achieved in the reactor (corresponding to the peak temperature of the combustion unit in the scheme of Figure 3d). The singular point for ER 0.3 can be explained once the residual char is also considered. Some residual char is present for low values of ER (causing loss in efficiency). For high values of ER, no residual char is observed so the excess oxygen consumes the gaseous products. The maximum of the hydrogen production (and also of CO) is just for values of ER around 0.3. Correspondingly the tar production is very low due to the high temperatures achieved. It is worth noting that temperatures higher than 1200°C are crucial for this kind of gasifiers to melt and remove the ash in the slag and minimize the formation of tar. As remarked above, residual char and tar production can be hardly quantified with equilibrium or simplified model, as this is possible only with the implementation of a detailed devolatilization sub-model able to predict the pyrolysis products as functions of the operating conditions. Therefore this gasifier model is a powerful tool to carry out a global analysis by varying all the operating variables and optimizing the performance
CGE =
mole frac
1200
Peak temperature (°C)
Syngas mole frac
0.5
1000 0.25
0.3
0.35
0.4
ER
Figure 9: Results of the entrained flow gasifier model with Aspen Plus® for the gasification of a coal/poplar wood blends (10%wt of biomass) with oxygen/steam mixtures.
6
CONCLUSIONS
A modeling procedure has been developed to provide powerful tools for process analysis and optimization on solid fuels gasification. The basic steps of pyrolysis, combustion, gasification are described in detailed submodels that are combined to represent the reactor configuration. Also heat streams are connected to respect the heat balance of the system. This approach allows one to optimize the operating conditions and compare different configurations (fixed beds, fluidized beds and entrained flow reactors) for a subsequent integration in several processes of current interest (e.g., combined power production, hydrogen production). Different commercial codes were used (Aspen Plus®, Aspen HYSYS®, gPROMS) and peculiar aspects and capabilities were commented.
Peak Temperature (°C) 1350
1400
1450
1500
1550
1600
1650
1700
0.8
optimized values
Cold Gas Efficiency
1750 0.8
1700 T
steam-to-fuel
0.75
0.75
oxygen-to-fuel T
0.7
0.7
0.50
(2005) pag. 1021 [13] Kunii D., Levenspiel O. High velocity fluidization. In: Fluidization Engineering, 2nd ed. (1991) pag. 193. Butterworth-Heinemann, Boston, England. [14] Basu P. Combustion and gasification in fluidized beds, Taylor & Francis Group, LLC 2006 [15] Aspen Technology Inc. Aspen HYSYS® V7.1 Simulation Basis. Ten Canal Park, Cambridge (MA) (2009) [16] Biagini E., Bardi A., Pannocchia G., Tognotti L. Development of an entrained flow gasifier model for process optimization study. Ind.Eng.Chem.Res. 48 (2009), pag. 9028.
0.81
0.65
0.65 0
0.2
0.4
0.6
0.8
1
steam-to-fuel, oxygen-to-fuel ratios (kg/kg)
Energy & Environment Division
Figure 10: Results of the optimization of the entrained flow gasifier modeled with Aspen HYSYS® for the gasification of a coal/poplar wood blends (10%wt of biomass) with oxygen/steam mixtures.
7
REFERENCES
[1] Zheng, L.; Furinsky, E. Comparison of Shell, Texaco, BGL and KRW gasifiers as part of IGCC plant computer simulations. Energy Conversion and Management 2005, 46, 1767–1779. [2] Chiesa, P.; Consonni, S.; Kreutz, T.; Williams, R. Co-production of hydrogen, electricity andCO2 from coal with commercially ready technology. PartA: Performance and emissions. Int. J. Hydrogen Energy, 30:747 (2005) [3] Collot, A.G. Matching gasifiers to coal. IEA Clean Coal Centre 2002, CCC/65. [4] Falcitelli M., Biagini E., Tognotti L. Development of the advanced biomass and coal devolatilization (ABCD) model. Proceedings of the 10th Conference on Energy for a Clean Environment, Lisbon 7-10 July 2009 [5] Rath J., Staudinger G. Cracking reactions of tar from pyrolysis of spruce wood. Fuel, 80 (2001) pag. 1379. [6] Di Blasi C. Modeling Wood Gasification in a Countercurrent Fixed-Bed Reactor, AIChE J. 50 (2004) pag. 2306. [7] Wen, C.Y.; Chaung, T.Z. Entrainment coal gasification modeling. Ind. Eng. Chem. Process Des., 1979, 18, 684. [8] Hobbs, Michael L., Predrag T. Radulovic, and L. Douglas Smoot, “Modeling Fixed-Bed Coal Gasifiers” AIChE Journal, Vol. 38, No. 5, May 1992, pp. 681 [9] Dogru M., Howarth C.R., Akay G., Keskinler B., Malik A.A. 2002. Gasification of hazelnut shells in a downdraft gasifier. Energy 27, 415–427 [10] Lv P., Yuana Z., Maa L., Wua C., Chena Y., Zhu J. 2007. Hydrogen-rich gas production from biomass air and oxygen/steam gasification in a downdraft gasifier. Renewable Energy 32, 2173 [11] Sotudeh-Gharebaagh R., Legros R., Chaouki, Paris J. Simulation of circulating fluidized bed reactors using Aspen Plus. Fuel 77 (1998) pag. 327. [12] Corella J., Sanz A. Modeling circulating fluidized bed biomass gasifiers. A pseudo-rigorous model for stationary state. Fuel Processing Technology 86
Lno Mediceo, 40 56127 Pisa – Italy
C. Pisa Ricerche
for this activity please contact:
[email protected] ph: +390500988447 fax: +390503869660
Chemical Engineering Department Via Diotisalvi, 2 56126 Pisa – Italy
Università di Pisa
for this activity please contact:
[email protected] ph: +390502217840 fax: +390502217866
