Modeling of particle transport and combustion phenomena in a large-scale circulating fluidized bed boiler using a hybrid Euler-Lagrange approach

The constantly developing fiuidized combustion technology has become competitive with a conventional pulverized coal （PC） combustion. Circulating fluidized bed （CFB） boilers can be a good alternative to PC boilers due to their robustness and lower sensitivity to the fuel quality. However, appropriate engineering tools that can be used to model and optimize the construction and operating parameters of a CFB boiler still require development. This paper presents the application of a relatively novel hybrid Euler-Lagrange approach to model the dense gas-solid flow combined with a combustion process in a large-scale indus- trial CFB boiler. In this work, this complex flow has been resolved by applying the ANSYS FLUENT 14.0 commercial computational fluid dynamics （CFD） code. To accurately resolve the multiphase flow, the original CFD code has been extended by additional user-defined functions. These functions were used to control the boiler mass load, particle recirculation process （simplified boiler geometry）, and interphase hydrodynamic properties. This work was split into two parts. In the first part, which is referred to as pseudo combustion, the combustion process was not directly simulated. Instead, the effect of the chemi- cal reactions was simulated by modifying the density of the continuous phase so that it corresponded to the mean temperature and composition of the flue gases, In this stage, the particle transport was simu- lated using the standard Euler-Euler and novel hybrid Euler-Lagrange approaches, The obtained results were compared against measured data, and both models were compared to each other. In the second part, the numerical model was enhanced by including the chemistry and physics of combustion. To the best of the authors＇ knowledge, the use of the hybrid Euler-Lagrange approach to model combustion is a new engineering application of this model, In this work, the combustion process was modeled for air-fuel combustion. The simulation results were compared with experimental data.     

Virtual experimentation through 3D full-loop simulation of a circulating fluidized bed

Eulerian granular multiphase model with a drag coefficient correction based on the energy-minimization multi-scale （EMMS） model was used to simulate a semi-industry scale circulating fluidized bed （CFB）. Three-dimensional （3D）, time-dependent simulation of a full-loop CFB revealed that the axial profiles of cross-sectionally averaged solid volume fraction, and the radial profiles of solid axial velocity and solid volume fraction were in reasonable agreement with experimental data. Based on this agreement, database derived from experiments not yet accomplished was replenished with such simulations, and fluid regime diagrams and pressure balance around the CFB loop were derived accordingly. This work presents an integrated viewpoint on CFB and unfolds a fresh paradigm for CFB modeling, which can be expected to help resolve certain issues long in dispute but bard for experiments.     

Virtual experimentation through 3D full-loop simulation of a circulating fluidized bed

Eulerian granular multiphase model with a drag coefficient correction based on the energy-minimization multi-male (EMMS) model was used to simulate a semi-industry scale circulating fiuidized bed (CFB).Three-dimensional(3D), time-dependent simulation of a full-loop CFB revealed that the axial profiles of cross-sectionally averaged solid volume fraction,and the radial profiles of solid axial velocity and solid volume fraction were in reasonable agreement with experimental data.Based on this agreement,database derived from experiments not yet accomplished was replenished with such simulations, and fluid regime diagrams and pressure balance around the CFB loop were derived accordingly. This work presents an integrated viewpoint on CFB and unfolds a fresh paradigm fur CFB modeling, which can be expected to help resolve certain issues long in dispute but hard for experiments.     

Particle Measurement Sensor for in situ determination of phase structure of fluidized bed

Based on three-dimensional (3D) acceleration sensing, an intelligent particle spy capable of detecting, transferring, and storing data, is proposed under the name of Particle Measurement Sensor (PMS). A prototype 60-mm-dia PMS was tested to track its freefall in terms of velocity and displacement, and served as a particle spy in a fluidized bed delivering the in situ acceleration information it detects. With increasing superficial gas velocity in the fluidized bed, the acceleration felt by PMS was observed to increase. The variance of the signals, which reflect the fluctuation, increased at first, reaching a maximum at the gas velocity (Uc) which marks the transition from bubbling to turbulent fluidization. Through probability density distribution (PDD) analysis, the PDD peak can be divided into the emulsion phase peak and the bubble phase peak. The average acceleration of emulsion and bubble phase increased, while the variance of both phases reached a maximum at Oc, at the same time. However, the difference between the variances of two phases reached the maximum at Oc. Findings of this study indicate that PMS can record independent in situ information. Further, it can provide other in situ measurements when equipped with additional multi-functional sensors.     

DEM simulation of polydisperse systems of particles in a fluidized bed

Numerical simulations based on three-dimensional discrete element model (DEM) are conducted for mono-disperse, binary and ternary systems of particles in a fluidized bed. Fluid drag force acting on each particle depending on its size and relative velocity is assigned. The drag coefficient corresponding to Ergun’s correlation is applied to the system of fluidized bed with particle size ratios of 1:1 for the mono-disperse system, 1:1.2, 1:1.4 and 1:2 for the binary system and 1:1.33:2 for the ternary system b...     

Particle convective heat transfer near the wall in a supercritical water fluidized bed by single particle model coupled with CFD-DEM

Supercritical water fluidized bed (SCWFB) is a promising reactor to gasify biomass or coal. Its optimization design is closely related to wall-to-bed heat transfer, where particle convective heat transfer plays an important role. This paper evaluates the particle convective heat transfer coefficient (h_pc) at the wall in SCWFB using the single particle model. The critical parameters in the single particle model which is difficult to get experimentally are obtained by the computational fluid dynamics-discrete element method (CFD-DEM). The contact statistics related to particle-to-wall heat transfer, such as contact number and contact distance, are also presented. The results show that particle residence time (τ), as the key parameter to evaluate h_pc, is found to decrease with rising velocity, while increase with larger thermal boundary layer thickness. τ follows a gamma function initially adopted in the gas–solid fluidized bed, making it possible to evaluate h_pc in SCWFB by a simplified single particle model. The theoretical predicted h_pc tends to increase with rising thermal gradient thickness at a lower velocity (1.5 U_mf), while first decreases and then increases at higher velocity (1.75 and 2 U_mf). h_pc occupies 30%–57% of the overall wall-to-bed heat transfer coefficient for a particle diameter of 0.25 mm. The results are helpful to predict the overall wall-to-bed heat transfer coefficient in SCWFB combined with a reasonable fluid convective heat transfer model from a theoretical perspective.     

Modeling and simulation of circulating fluidized bed reactors applied to a carbonation/calcination loop

A fluid dynamic model for a gas-solid circulating fluidized bed （CFB） designed using two coupled riser reactors is developed and implemented numerically with code programmed in Matlab. The fluid dynamic model contains heat and species mass balances to calculate temperatures and compositions for a carbonation/calcination loop process. Because of the high computational costs required to resolve the three-dimensional phenomena, a model representing a trade-offbetween computational time requirements and accuracy is developed. For dynamic processes with a solid flux between the two reactor units that depends on the fluid dynamics of both risers, a dynamic one-dimensional two-fluid model is sufficient. A two-fluid model using the constant particle viscosity closure for the stress term is used for the solid phase, and an algebraic turbulence model is applied to the gas phase. The numerical model implementa- tion is based on the finite volume method with a staggered grid scheme. The exchange of solids between the reactor units constituting the circulating fluidized bed （solid flux） is implemented through additional mass source/sink terms in the continuity equations of the two phases, For model validation, a relevant experimental analysis provided in the literature is reproduced by the numerical simulations, The numerical analysis indicates that sufficient heat integration between the two reactor units is important for the performance of the circulating fluidized bed system, The two-fluid model performs fairly well for this chemical process operated in a CFB designed as two coupled riser reactors. Further analysis and optimization of the solution algorithms and the reactor coupling strategy is warranted.     

Two and three dimensional modeling of fluidized bed with multiple jets in a DEM-CFD framework

Fluidized beds with multiple jets have widespread industrial applications. The objective of this paper is to investigate the jet interactions and hydrodynamics of a fluidized bed with multiple jets. Discrete element modeling coupled with in-house CFD code GenlDLEST has been used to simulate a bed with nine jets. The results are compared with published experiments. Mono dispersed particles of size 550 ~m are used with 1.4 times the minimum fluidization velocity of the particles. Both two and three dimensional computations have been performed. To the best of our knowledge, the results presented in this paper are the first full 3D simulations of a fluidized bed performed with multiple jets. Discrepancies between the experiment and simulations are discussed in the context of the dimensionality of the simulations. The 2D solid fraction profile compares well with the experiment close to the distributor plate. At higher heights, the 2D simulation over-predicts the solid fraction profiles near the walls. The 3D simulation on the other hand is better able to capture the solid fraction profile higher up in the bed compared to that near the distributor plate. Similarly, the normalized particle velocities and the particle fluxes compare well with the experiment closer to the distributor plate for the 2D simulation and the freeboard for the 3D simulation, respectively. A lower expanded bed height is predicted in the 2D simulation compared to the 3D simulation and the experiment. The results obtained from DEM computations show that a 2D simulation can be used to capture essential jetting trends near the distributor plate regions, whereas a full scale 3D simulation is needed to capture the bubbles near the freeboard regions. These serve as validations for the experiment and help us understand the complex jet interaction and solid circulation patterns in a multiple jet fluidized bed system.     

Three-dimensional dynamic characterization of square-nosed slugging phenomena in a fluidized bed

Slugging represents one of the major regimes in fluidization, which occurs in small diameter beds with large bed height-to-diameter ratio or in large diameter beds with internals that resemble multiple small diameter fluidized beds. Slug types include round-nosed slug, wall slug and square-nosed slug. Studies of the slugs have been mainly focused on round-nosed or wall slugs known as half slug, typically occurring in Geldart group A particle fluidization. The square-nosed slug typically occurring for Geldart group D particles appears to be regarded as simple in its structure. The Electrical Capacitance Volume Tomography (ECVT) imaging of the square-nosed slugging phenomena conducted in this study reveals otherwise. That is the structure of the square-nosed slug is, in fact, complex, particularly with respect to its dynamic variation in fluidization. More broadly, this study examines experimentally the hydrodynamic characteristics of the square-nosed fluidization regime. Specifically, simultaneous measurements from multiple ECVT sensors provide non-invasive, continuous, 3-dimensional imaging of the entire flow region of the slugging bed and hence enabling the dynamic characterization of the evolution of the slugs. The analysis of the 3D images reconstructed for real-time gas–solid volume fraction profile of the slugging fluidized bed indicates that there are three different zones, namely, the bottom fluidization zone, the gas slug zone, and the solid slug zone, co-existing in the bed. The three zones present different hydrodynamic characteristics during the slug evolution. It is found that varying the gas velocity of the slugging bed mainly varies the maximum length of the gas slug zone, while it only has a minor effect on the lengths of the bottom fluidization zone and solid slug zone. It also has an insignificant effect on the solid volume fraction of the three zones.     

Experiments on the local heat transfer characteristics of a circulating fluidized bed

The role of particle diameter in the heat transfer of a gas–solid suspension to the walls of a circulating fluidized bed was studied for particles of uniform size. This work reports and analyzes new experimental results for the local bed to wall heat transfer coefficient, not including the radiation component, in a long active heat transfer surface length laboratory bed, which extend previous findings and clear up some divergences. The research included determining the effects of extension and location of the heat transfer surface, circulating solids mass flux and average suspension density. An experimental set-up was built, with a 72.5 mm internal diameter riser, 6.0 m high, composed of six double pipe heat exchangers, 0.93 m high, located one above the other. Five narrow sized diameter quartz sand particles − 179, 230, 385, 460 and 545 μm − were tested. Temperature was kept approximately constant at 423 K and the superficial gas velocity at 10.5 m/s. The major influence of suspension density on the wall heat transfer was confirmed, and contrary to other authors, a significant effect of particle size was found, which becomes more relevant for smaller particles and increasing suspension density. It was observed that the extension of the heat transfer surface area did not influence the heat transfer coefficient for lengths greater than 0.93 m.The heat transfer surface location did not show any effect, except for the exchanger at the botton of the riser. A simple correlation was proposed to calculate the heat transfer coefficient as a function of particle diameter and suspension density.     

Modeling and simulation of circulating fluidized bed reactors applied to a carbonation/calcination loop

A fluid dynamic model for a gas-solid circulating fluidized bed (CFB) designed using two coupled riser reactors is developed and implemented numerically with code programmed in Matlab. The fluid dynamic model contains heat and species mass balances to calculate temperatures and compositions for a carbonation/calcination loop process.Because of the high computational costs required to resolve the three-dimensional phenomena, a model representing a trade-off between computational time requirements and accuracy is developed. For dynamic processes with a solid flux between the two reactor units that depends on the fluid dynamics of both risers, a dynamic one-dimensional two-fluid model is sufficient.A two-fluid model using the constant particle viscosity closure for the stress term is used for the solid phase, and an algebraic turbulence model is applied to the gas phase. The numerical model implementation is based on the finite volume method with a staggered grid scheme. The exchange of solids between the reactor units constituting the circulating fluidized bed (solid flux) is implemented through additional mass source/sink terms in the continuity equations of the two phases.For model validation, a relevant experimental analysis provided in the literature is reproduced by the numerical simulations. The numerical analysis indicates that sufficient heat integration between the two reactor units is important for the performance of the circulating fluidized bed system.The two-fluid model performs fairly well for this chemical process operated in a CFB designed as two coupled riser reactors. Further analysis and optimization of the solution algorithms and the reactor coupling strategy is warranted.     

Three-dimensional multi-phase simulation of the mixing and segregation of binary particle mixtures in a two-jet spout fluidized bed

This study presents a three-dimensional numerical study of the mixing and segregation of binary particle mixtures in a two-jet spout fluidized bed based on an Eulerian–Eulerian three-fluid model. Initially, the particle mixtures were premixed and packed in a rectangular fluidized bed. As the calculation began, the gas stream was injected into the bed from the distributor and jet nozzles. The model was validated by comparing the simulated jet penetration depths with corresponding experimental data. The main features of the complex gas–solid flow behaviors and the mechanism of mixing and segregation of the binary mixtures were analyzed. Moreover, further simulations were carried out to evaluate the effects of operating conditions on the mixing and segregation of binary particle mixtures. The results illustrate that mixing can be enhanced by increasing the jet velocity or enlarging the difference of initial proportions of binary particle mixtures.     

Experimental study and modeling of particle drying in a continuously-operated horizontal fluidized bed

Multistage fluidized beds are frequently used for product drying in industry. One advantage of these fluidized beds is that they can achieve a high throughput, when operated continuously. In this study, γ-Al₂O₃ particles were dried in a pilot-scale horizontal fluidized bed, without considering any comminution effects. For each experiment, the particle moisture content distribution and residence time distribution were determined. To take into account particle back mixing in our experiments, a one-dimensional population balance model that considers particle residence time was introduced into a fluidized bed-drying model. Experimental particle residence time distributions were reproduced using a tank-in-series model. Subsequently, the moisture content distribution was implemented, as a second dimension to the population balance in this model. These two-dimensional simulations were able to describe the experimental data, especially the spread in the residual particle moisture distribution, much more accurately than one-dimensional simulations. Using this novel two-dimensional model, the effects of different operating parameters (process gas temperature, solid feed rate, superficial air velocity) on the particle moisture content distribution were systematically studied.     

Statistical and frequency analysis of the pressure fluctuation in a fluidized bed of non-spherical particles

In this paper, the pressure fluctuation in a fluidized bed was measured and processed via standard devia- tion and power spectrum analysis to investigate the dynamic behavior of the transition from the bubbling to turbulent regime. Two types （Geldart B and D） of non-spherical particles, screened from real bed materials, and their mixture were used as the bed materials. The experiments were conducted in a semi- industrial testing apparatus. The experimental results indicated that the fluidization characteristics of the non-spherical Geldart D particles differed from that of the spherical particles at gas velocities beyond the transition velocity Uo The standard deviation of the pressure fluctuation measured in the bed increased with the gas velocity, while that measured in the plenum remained constant. Compared to the coarse particles, the fine particles exerted a stronger influence on the dynamic behavior of the fluidized bed and promoted the fluidization regime transition from bubbling toward turbulent. The power spectrum of the pressure fluctuation was calculated using the auto-regressive （AR） model; the hydrodynamics of the flu- idized bed were characterized by the major frequency of the power spectrum of the pressure fluctuation. By combining the standard deviation analysis, a new method was proposed to determine the transition velocity Uk via the analysis of the change in the major frequency. The first major frequency was observed to vary within the range of 1.5 to 3 Hz.     

Statistical and frequency analysis of the pressure fluctuation in a fluidized bed of non-spherical particles

In this paper, the pressure fluctuation in a fluidized bed was measured and processed via standard deviation and power spectrum analysis to investigate the dynamic behavior of the transition from the bubbling to turbulent regime. Two types (Geldart B and D) of non-spherical particles, screened from real bed materials, and their mixture were used as the bed materials. The experiments were conducted in a semi-industrial testing apparatus. The experimental results indicated that the fluidization characteristics of the non-spherical Geldart D particles differed from that of the spherical particles at gas velocities beyond the transition velocity U_c. The standard deviation of the pressure fluctuation measured in the bed increased with the gas velocity, while that measured in the plenum remained constant. Compared to the coarse particles, the fine particles exerted a stronger influence on the dynamic behavior of the fluidized bed and promoted the fluidization regime transition from bubbling toward turbulent. The power spectrum of the pressure fluctuation was calculated using the auto-regressive (AR) model; the hydrodynamics of the fluidized bed were characterized by the major frequency of the power spectrum of the pressure fluctuation. By combining the standard deviation analysis, a new method was proposed to determine the transition velocity U_k via the analysis of the change in the major frequency. The first major frequency was observed to vary within the range of 1.5 to 3 Hz.     

Simulation of mass balance behavior in a large-scale circulating fluidized bed reactor

We determine using a compound model the influence of the mass of granular matter on the behavior of a supercritical circulating fluidized bed (CFB) reactor. Population balance enables a stationary-regime modeling of the mass flow of granular matter inside a CFB unit in a large-scale. The simulation includes some important dynamic processes of gas-particle flows in fluidized bed such as attrition, fragmentation, elutriation, and fuel combustion. Numerical calculations with full boiler loading were performed of operational parameters such as furnace temperature, furnace pressure, feeding materials mass flows, and excess air ratio. Furthermore, three bed inventory masses were adopted as experimental variables in the simulation model of mass balance. This approach enables a sensitivity study of mass flows of granular matter inside a CFB facility. Some computational results from this population balance model obtained for a supercritical CFB reactor are presented that show consistency with the operational data for large-scale CFB units.     

Effect of static bed height in the upper fluidized bed on flow behavior in the lower riser section of a coupled reactor

To study olefin reduction by using an auxiliary reactor for FCC naphtha upgrading, a large-scale cold model of a riser-bed coupled to an upper fluidized bed was established. The effect of static bed height in the upper fluidized bed on particle flow behavior in the lower riser was investigated experimentally. A restriction index of solids holdup was used to evaluate quantitatively the restrictive effect of the upper fluidized bed. Experimental results show that, under the restrictive effect of the upper fluidized bed, the riser could be divided into three regions in the longitudinal direction: accelerating, fully developed and restriction. The axial distribution of solids holdup in the riser is characterized by large solids holdup in the top and bottom sections and small solids holdup in the middle section. Overall solids holdup increased with increasing static bed height in the upper fluidized bed, while particle velocity decreased. Such restrictive effect of the upper fluidized bed could extend from the middle and top sections to the whole riser volume when riser outlet resistance is increased, which increases with increasing static bed height in the upper fluidized bed. The upper bed exerts the strongest restriction on the area close to the riser outlet.     

Application of CFD-DEM to the study of solid exchange in a dual-leg fluidized bed

The CFD-DEM model was developed to simulate solid exchange behavior between two half beds in a bench-scale two-dimensional dual-leg fluidized bed （DL-FB）. Power spectrum density （PSD） analysis was applied to obtain the dominant frequency （F） of the simulated differential particle number （APLR） between the two half beds. Effects of fluidization velocity （u） and bed material inventory （H） on the solid exchange behavior were studied using the CFD-DEM model. Not only snapshots of the simulated particle flow patterns using the OpenGL code but also the dominant frequency of APLR was similar to the experimental results. The simulation results show that higher fluidization velocity assists the exchange of more particles between the two half beds, but the dispersion of clusters on the bed surface into single particles decreases the cluster exchange frequency. A greater bed material inventory results in more intense cluster exchange. The cluster exchange frequency decreases with an increase of the bed material inventory.     

3D full-loop simulation and experimental verification of gas-solid flow hydrodynamics in a dense circulating fluidized bed

Because of their advantages of high efficiency and low cost, numerical research methods for large-scale circulating fluidized bed （CFB） apparatus are gaining ever more importance. This article presents a numer- ical study of gas-solid flow dynamics using the Eulerian granular multiphase model with a drag coefficient correction based on the energy-minimization multi-scale （EMMS） model. A three-dimensional, full-loop, time-dependent simulation of the hydrodynamics of a dense CFB apparatus is performed. The process parameters （e.g., operating and initial conditions） are provided in accordance with the real experiment to enhance the accuracy of the simulation. The axial profiles of the averaged solid volume fractions and the solids flux at the outlet of the cyclone are in reasonable agreement with experimental data, thereby verifying the applicability of the mathematical and physical models. As a result, the streamline in the riser and standpipe as well as the solids distribution contours at the cross sections is analyzed. Computational fluid dynamics （CFD） serves as a basis for CFB modeling to help resolve certain issues long in dispute but difficult to address experimentally. The results of this study provide the basis of a general approach to describing dynamic simulations of gas-solid flows.