Simulation of Olive Kernel Gasification in a Bubbling Fluidized Bed Pilot Scale Reactor ()
1. Introduction
The traditional approach necessary to establish comercial plant technology is based on comprehensive experimental investigations, progressing from a laboratory scale test unit to a pilot scale plant, before building a full-scale commercial demonstration plant. For process optimisation, an extensive investigation of the plant behaviour depending on various operating parameters is required for each scale up step. To support this optimisation procedure, mathematical models are helpful to reduce the temporal and financial efforts.
Pre-condition is a reliable simulation tool, which includes the mathematical formulation of all important chemical and physical processes by describing their dependency on operating parameters and their interdependencies [1]. However, only limited research has been performed to model those complex systems including also the hydrodynamics of biomass particles.
The development of numerical models for fluidizedbed gasification (FBG) documented in the literature [2] are devoted mainly to coal and less to biomass. Even though, biomass in comparison to coal is made up not only of lignin but also of cellulose and hemi cellulose, each one having its own thermal behaviour and making the modelling more difficult, the modelling approaches used for coal can serve for biomass gasification modelling as well. Ma et al. [3] developed a steady-state model which involved instantaneous devolatilization of coal at the top of the gasifier (freeboard region) and char combustion and gasification in the fluidized bed. Chejne and Hernandez [4] developed a comprehensive mathematical model to predict the behaviour of coal combustion and gasification on stacks in non-stationary operation. Sett and Bhattacharya [5] developed a mathematical model for the behaviour of a fluidised-bed charcoal gasifier, taking into account fluidised-bed hydrodynamic conditions, chemical reaction rates, mass diffusion rates, and the conservation of mass and energy. The model was solved numerically using an IBM 3083 mainframe computer. Ross et al. [6] modified their one-dimensional numerical isothermal model of a fluidised-bed coal gasifier in order to simulate the performance of a laboratory-scale gasifier.
A new numerical model based on the two-fluid model (TFM) including the kinetic theory of granular flow (KTGF) and complicated reactions has been developed to simulate coal gasification in a bubbling fluidized bed gasifier (BFBG) [7]. The collision between particles was described by KTGF. The coal gasification rates were determined by combining Arrhenius rate and diffusion rate for heterogeneous reactions or turbulent mixing rate for homogeneous reactions. The flow behaviours of gas and solid phases in the bed and freeboard could be predicted, which were not easy to be measured through the experiments. Hamel and Krumm [2] have developed a mathematical model for simulation of gasification processes of solid fuels in atmospheric or pressurised bubbling fluidised beds incorporating bed and freeboard hydrodynamics, fuel drying and devolatilization, and chemical reaction kinetics is presented. The model has been used to simulate four bubbling fluidised bed gasifiers, described in literature, of different scales from atmospheric laboratory scale up to pressurised commercial scale, processing brown coal, peat and sawdust. The gasifiers have been operated within a wide range of parameters using air, air steam or oxygen steam as gasification agent, operating with or without recirculation of fines at operating pressures up to 2.5 MPa. The simulation results for overall carbon conversion, temperature and concentrations of gaseous species agreed sufficiently well with published experimental data.
Corella and Sanz [8,9], at the University of Zaragoza and Madrid (Spain) started to study the modelling of fluidized bed biomass gasifiers in the mid-1980s. More recently Corella et al. [10], discussed the reaction network existing in a CFB biomass gasifier and the problems associated with the accuracy of the kinetic equations needed for the existing complex reaction network. Furthermore, he presented a model for bubbling fluidized bed (BFB) biomass gasifiers, gasifying with pure steam [10]. That model identified the four main, for modelling purposes, chemical reactions among the reaction network existing in the gasification process. With only four kinetic parameters, the model predicted quite well the BFB gasifier. More recently, Corella and Sanz [8,9] have presented a whole model for CFBBGs. Such model is 1-dimensional for steady state. The model has a semi rigorous character because of the assumptions that had to be introduced by lack of accurate knowledge of some parts of the modelling.
De Souza-Santos [11] developed a comprehensive mathematical model and commercially available computer program performing a comprehensive simulation of fluidized-bed equipment (CSFB Version 3.5), to use as a tool for engineering design and operation optimisation, by predicting the behaviour of a real unit during steadystate operation. His model is regarded as complete and it includes the conservation equations for the emulsion phase and bubbles, empirical equations for hydrodynamics, and it is also includes a through mass balance which considers that both drying and volatilisation are not instantaneous. Jiang and Morey [12] developed one-dimensional, steady state, numerical model for a fluidized bed biomass gasifier. The gasifier model consisted of a fuel pyrolysis model, an oxidation model, a gasification model and a freeboard model which were validated with experimental results.
Haggerty and Pulsifer [13] used the reaction model together with three different reactor models (a) plug-flow, (b) complete-mixing and (c) bubble-assemblage, where the bubble-assemblage model represents a valid alternative when modelling fluidized-bed gasifiers. Marias et al. [14] developed a mathematical model for the fluidizedbed incineration process using the waste composed of wood, cardboard and polyvinyl chloride. Aarsen’s model treated an isothermal fluidized bed as two separate zones, the oxidation zone at the bottom and the gasification zone at the top [15]. Fuel pyrolysis took place at such a fast rate that only pyrolysis yields, which were assumed to vary with bed temperature, were needed. Mukosei et al. [16] examined the problem of the mathematical simulation of heterogeneous processes in a fluidized-bed reactor. These works and other more recent ones such as the work by He and Rudolph [17] proposed a new approach to the modelling of gross gas-solids flow through the riser in a circulating fluidized-bed system. This approach differs from the previous ones, which are found to be theoretically incorrect based on a fundamental analysis of the riser process hydrodynamics.
Panopoulos et al. [18] used the simple way of approaching the biomass gasification modelling by predicting thermodynamic equilibrium composition through Gibbs free energy minimisation calculations for the C, H, and O atoms of the fuel and the gasification agent mixture. Samuilov et al. [19] developed a mathematical model for the gasification in carbon dioxide of a single carbon particle. The porous structure of the particle, diffusion processes, the gasification processes on the pore surface according to the Langmuir-Hinshelwood model, and reactions on active carbon sites were taken into account. Deen et al. [20] reviewed the use of discrete particle models (DPMs) for the study of the flow phenomena prevailing in fluidized beds.
The ASPEN PLUS process simulator has been used by different investigators to simulate coal conversion; examples include methanol synthesis [21,22], indirect coal liquefaction processes [23], integrated coal gasification combined cycle (IGCC) power plants [24], atmospheric fluidized bed combustor processes [25], compartmented fluidized bed coal gasifiers [26], coal hydrogasification processes [27], and coal gasification simulation [28]. However, the work that has been done on biomass gasification is limited. Mansaray et al. [29] used ASPEN PLUS to simulate rice husk gasification based on material balance, energy balance, and chemical equilibrium relations. Because of the high amount of volatile material in biomass and the complexity of biomass reaction rate kinetics in fluidized beds, they ignored the char gasification and simulated the gasification process by the assumption that biomass gasification follows Gibbs equilibrium. In a typical atmospheric fluidized bed gasifier, feed, together with bed material, are fluidized by the gasifying agents, such as air and/or steam, entering at the bottom of the bed. The product gas resulting from the gasification process is fed to a gas-solid separator (i.e., cyclone) to separate solid particles carried by exhaust gas.
2. Modelling Approach
Mathematical modelling is an alternative means to study fluidisation and fluidised bed biomass gasification. Modelling can not only account for the fundamental hydrodynamic behaviour of fluidisation and the coal gasification process, but also serve as a predictive tool to assist with the design, optimisation and scale-up of fluidised bed gasifiers. Therefore, in recent years mathematical modelling of fluidised beds has increasingly attracted more attention.
2.1. Assumptions
The following assumptions were considered in modelling the gasification process:
● Process is steady state and isothermal.
● Biomass devolatilization takes place instantaneously and volatile products mainly consist of H2, CO, CO2, CH4, and H2O.
● All the gases are uniformly distributed within the emulsion phase.
● Particles are spherical and of uniform size and the average diameter remains constant during the gasification ,based on the shrinking core model
● Char only contains carbon and ash.
● Reactions of N and S have not been taken account
● The reactions reached chemical equilibrium
2.2. Reactions
2.2.1. Combustion Zone
The combustible substance of a solid fuel is usually composed of elements carbon, hydrogen and oxygen. In complete combustion carbon dioxide is obtained from carbon in fuel and water is obtained from the hydrogen, usually as steam. The combustion reaction is exothermic and yields a theoretical oxidation temperature of 1450˚C. The main reactions, therefore, are:
(1)
(2)
2.2.2. Reduction Zone
The products of partial combustion (water, carbon dioxide and uncombusted partially cracked pyrolysis products) now pass through a red-hot charcoal bed where the following reduction reactions take place.
(3)
(4)
(5)
(6)
(7)
Reactions (3) and (4) are main reduction reactions and being endothermic have the capability of reducing gas temperature. Consequently the temperatures in the reduction zone are normally 800˚C - 1000˚C. Lower the reduction zone temperature (~ 700˚C - 800˚C), lower is the calorific value of gas.
2.2.3. Pyrolysis Zone
Wood pyrolysis is an intricate process that is still not completely understood 14. The products depend upon temperature, pressure, residence time and heat losses. However following general remarks can be made about them. Up to the temperature of 200˚C only water is driven off. Between 200˚C to 280˚C carbon dioxide, acetic acid and water are given off. The real pyrolysis, which takes place between 280˚C to 500˚C, produces large quantities of tar and gases containing carbon dioxide. Besides light tars, some methyl alcohol is also formed. Between 500˚C to 700˚C the gas production is small and contains hydrogen. Thus it is easy to see that updraft gasifier will produce much more tar than downdraft one. In downdraft gasifier the tars have to go through combustion and reduction zone and are partially broken down.
2.3. Hydrodynamic Parameters
The following assumptions were made in simulating the hydrodynamics:
● Fluidized bed reactors divided into two regions bed and freeboard
● The fluidization state in the bed is maintained in the bubbling regime
● The volume fraction of solids decreases as height increases, corresponding to the coalescence of bubbles in the bed and the returning of solid particles to the bed in the TDH zone
● Volumetric flow rate of gas increases along with height, corresponding to the production of gaseous products
● The mixing of solid particles, consisting of ash, char particles, and bed material, is perfect
● The reactor is divided into a finite number of equal elements with constant hydrodynamic parameters
● The fluidized bed is one-dimensional; any variations in conditions are considered to occur only in the axial direction.
2.4. Global Model of Fluidized Bed Gasifier
The proposed gasification dynamic model represents two stages: instantaneous devolatilization of straw and combustion of the char at the bottom of the gasifier and the gasification in the fluidized bed. Thus, a two-phase representation of the fluidized bed incorporates the phenomena of jetting, bubbling, slugging and dynamic mass and energy balances. Figure 1 shows the fluidized bed domain divided into three regions: jet, bubbles and slugs. The following assumptions were made regarding to the gas low divisions and the low patterns of gas and solids:
1) the fluidized bed consists of a dilute phase (jets, bubbles and = or slugs) and an emulsion phase2) the emulsion phase is divided into an interstitial gas phase and a solid phase 3) mass and heat exchange take place between the dilute phase and interstitial gas and between the interstitial gas and the solids4) the fluidizing gas enters the bed through nozzles in a jet form. The jets degenerate into bubbles, which rise through the bed and grow by coalescence with other bubbles to form slugs5) slugging occurs if the bubble diameter becomes larger than one third of the reactor diameter, (Slugging occurs in improper fluidization 6) the dilute gas and the interstitial gas are in plug low and the dilute gas is free of solids7) the gas behaves ideally, and 8) the produced gas consist of CO, CO2, H2, H2O, CH4 and N2.
2.5. Mass and Energy Balances
The olive kernel feed flow is calculated by the following equation [30]:
(8)
For the calculation of the air that is demanded for the stoichiometric combustion of biomass is used [30]:
(9)
where the coefficients a, b, c, d represent the molecular fractions H/C, O/C, N/C and S/C . Therefore we can calculate the air flow for several values of equivalence ratio (ER) according to previous equation [30]
With the assumption that gasification char is consisted of fixed carbon and ash, is calculated the solid mass flow from the equation [30]:
(10)
Then the overall mass balance is formed for the gasification reactor and the mass flow for the gas gasification is estimated as following [30]:
(11)
Figure 1. The fluidized bed gasifier zones.
Correspondingly the overall energy balance is described as below [30]:
(12)
Es represents the thermal content of biomass feed flow which is estimated from the following equation [30]:
(13)
For the air it is assumed that enters the reactor at environmental temperature, so its enthalpy content is considered zero (Eair, gasif = 0) .
Eth represents the thermal energy and the enthalpy of gas product and it is given by the following equation [30]:
(14)
where
(15)
The following equation estimates the enthalpy of each component of gas gasification in gasification temperature
(16)
Finally Eloss represents the energy losses from the system due to ash and char removal and the losses to environment from the walls of the gasifier. Therefore:
(17)
It is assumed that the energy losses from the gasifier walls constitute the 10% of the overall losses. The energy losses due to char gasification removal is calculated as below [30]:
(18)
2.6. Air Equivalence Ratio
Equivalence ratio (ER) constitutes a significant factor of biomass gasification. The wise choice of it discourages the propulsion of the oxidation reactions. It is defined by the following equation [31]:
(19)
Aπ represents the mass ratio between air and biomass that has been used in laboratory and Aθ represents the stoichiometrically demanded mass ratio of air for the total combustion of the same quantity of biomass.
The alteration of air flow effects the process since it leads it to the two terminal situations:
1) High air equivalence ratio (ER > 1) approaches the process of combustion giving as gasification product mainly CO2
2) Low air equivalence ratio (ER < 0.25) benefits the process of pyrolysis giving primarily as product syngas (CO + H2).
2.7. Producer Gas LHV
Temperature and ER also significantly affect the heating value of producer gas. Main research target is the production of a producer gas enriched in CO, H2, and CH4. The presence of these combustible species contributes to the production of a medium to high heating value gas, suitable for further exploitation in internal combustion engines (ICE) and turbines for power production. LHV calculation was made using the following equation [32]:
(20)
where, CO, H2 and CH4 are the molar ratios of the species in the producer gas as measured in our gasification experiments.
3. Experimental
3.1. Biomass Characteristics
In the laboratory olive kernel was used as biomass feeding and its characteristics are illustrated in the Table 1. Olive kernel is the final solid residue emerging from olive oil production industries. Olives are processed in an oil extracting plant to recover the oil content. After a first residue drying, residual oil is hexane extracted generating a residual solid product with a moisture content of around 10 - 12 wt% called olive kernel. Traditionally, such kind of residue is sold as fuel for combustion in small boilers and specially designed furnaces due to its significant calorific value (HHV ~21 MJ/kg).
Nowadays, there are several boilers fuelled with olive kernels destined for residential and greenhouse heating or in some cases to generate steam for electricity production in conventional steam turbines. However, a more efficient and environmentally friendly alternative with respect to CO2 and CH4 mitigation to the atmosphere, is olive kernel gasification; aim is to produce a high calorific value gas to fuel a gas engine or even a gas turbine in a combined cycle of heat/power production.
3.2. Gasifier Characteristics
3.2.1. The Bench Scale FBG Unit
A 5 kWth, bench scale, bubbling fluidized bed gasifier (FBG) was designed and manufactured for biogenic material gasification. It is an atmospheric air blown FBG unit that consists of four sections: 1) the biomass feeding section, 2) the air supply, control and preheating section, 3) the gasification facility, 4) the gas sampling and off line analysis section.
Producer gas was sampled using airtight gas sampling bags and analyzed with a GC system for quantitative determination of CO, CO2, H2, CH4, C2H4 and C2H6.
3.2.2. The Olive Kernel Feeding Section
The FBG unit is equipped with a feeding hopper suitable for feeding of low density biomass. Feeding takes place through a system of a constant speed screw feeder and a variable speed screw feeder (motorized via inverter), separated by a rotary valve in order to avoid the hot gas backflow that could pyrolyse olive kernels and cause serious tar fouling problems. Additionally, a small proportion of fluidization air (~1 l/min) is supplied for backflow prevention. The motorized screw feeder is able to control the biomass feeding, based on a series of calibration curves and the maximum mass flow rate of olive kernel could reach 120 g/min. Olive kernel pass through the metal to metal rotary valve and is introduced into the reactor by the second screw feeder, into the hot zone of the reactor, 120 mm above the air distributor.
Table 1. Olive kernel characteristics.
3.2.3. The Air Supply, Control and Preheating Section
Laboratory’s air supply unit provides the FBG unit with the primary and secondary air, necessary for the fluidization and gasification process, always in hypo-stoichiometric amounts obtaining the desired ER ratio. It consists of a central compressor, a main valve to control the air flow rate and an air flowmeter with maximum flow capacity of 200 l/min. The primary air for gasification process is introduced from the bottom of the FBG and a uniformly perforated plate (151 holes of 1mm diameter) is used in order to retain bed fluidization. Primary air is preheated up to 350˚C, by passing through a stainless steel tube that is kept in touch with the hot zone of the reactor inside the furnace. In the beginning of each operation, while the reactor is still cold, air preheating is obtained via an electric resistance placed at the bottom of the FBG. Secondary air contributes in producer gas temperature increase, which results in further tar cracking improving carbon conversion efficiency.
3.2.4. The 5 kWth FBG Test Unit
Figure 2 shows the bench scale FBG reactor with nominal capacity of 5 kWth. Bed and freeboard sections are made of stainless steel with diameters of 60 mm and 90 mm respectively; the total reactor height is 1400 mm. A more detailed analysis about the reactor design and construction can be found in a previous published work [33]. FBG design enables the easy dismantling and thorough cleaning, in order to avoid not only memory effects, but also possible tar fouling under the hot flow experimental conditions. The gas cleaning system consists of a cyclone of 10 μm cut point. Limited experimental work has done using high temperature ceramic filters and various types of metallic foam for particulate abatement and tar elimination.