Pulverized Coal Jet Flame Research Articles

Non-gray gas and particle radiation in a pulverized coal jet flame

In pulverized coal combustion, the participating radiative media includes gas, soot, and coal/char particles; a fundamental understanding of their radiation behavior is essential for the prediction of radiation heat transfer and control of radiation in the furnace. In this study, the individual contributions of gas, soot, and coal particles to the total radiation were examined based on the large eddy simulation of the CRIEPI pulverized coal jet flame. The radiation transfer equation was solved using the discrete ordinate method with state-of-the-art non-gray radiative property models of multi-phase media. The full spectrum correlated k-distributions (FSCK) model was used to calculate the radiative property of the gas-soot mixture. A burnout-based particle radiative property model was applied for coal/char particles. The results showed that the predicted spatial distributions of coal particles, tar, and soot were in good agreement with the experimental data. The gray gas radiation model predicted temperatures approximately 50 K lower in this flame, and the peak soot volume fraction was predicted at approximately 22% lower. For the laboratory scale burner under examination, the individual radiative contributions of gas, soot, and coal/char were found to be 35%, 40%, and 25%, respectively, so that it was necessary to consider all components for accurate prediction. Moreover, these contribution ratios varied with the flame position. Because of the high particle concentration, the individual contribution of coal/char particles accounted for up to 65% in the devolatilization region. Soot radiation played an important role in the pulverized coal flame. In the region where the ensemble-averaged soot volume fraction reached the maximum of 2.7 ppm, the individual contribution of soot was 50% of the total radiation, which accounted for 70% of the total particles (soot and char) radiation.

Application of flamelet/progress-variable approach to the large eddy simulation of a turbulent jet flame of pulverized coals

In this study, the flamelet/progress-variable (FPV) approach was applied to a large eddy simulation of a pulverized coal jet flame. The FPV approach considers the characteristics of the pulverized coal flame, e.g., non-adiabatic system and several types of fuel streams, via additional representative variables. First, the applicability of the FPV approach to a turbulent flame with pulverized coals was confirmed through a comparison of the numerical solutions and experimental data. In this study, the pure pilot case was also investigated to clarify the effects of pulverized coals on the flame. The flame structure changes significantly upon the injection of pulverized coals, and the flame index suggests the coexistence of premixed and diffusion combustion modes even in the downstream region. In particular, the combustion mode fluctuates with time in the middle region of the flame. The fuel gas released from the pulverized coals should increase in this region; therefore, the release and combustion behavior of the volatile matter must be involved in the combustion mode variation. The evaluation of the combustion modes of fuel gas in the coal flame is useful for the design and optimization of pulverized coal combustors with next-generation technologies.

Large eddy simulations and analysis of NO emission characteristics in a laboratory pulverized coal flame

Large eddy simulation of a lab-scale turbulent pulverized coal jet flame is conducted using the extended flamelet/progress variable (FPV) approach, where normalized total enthalpy, char conversion parameter, progress variable, coal mixture fraction and coal mixture fraction variance are selected as the control parameters for a five-dimensional flamelet library. Two ways of fuel-N partitioning between volatiles and char are applied for NO formation in two cases, respectively. The model performance is evaluated by comparing with the experimental data. The results indicate that the present model can well reproduce the pulverized coal flame. Reasonably good agreements on particle dispersion and velocity, temperature as well as major species distributions are obtained. In addition, the distributions of NO obtained with different treatments are analyzed qualitatively as well as quantitatively. The results indicate that the extended FPV approach shows an advantage over the conventional post-processing approach. When reasonable fuel-N partitioning is considered, a more accurate NO prediction can be achieved. Finally, combustion and NO emission characteristics are further discussed.

Identification of the initial particle size distribution for coal combustion simulations

AbstractPulverized coal generally features wide particle‐size distributions (PSD). How to set an initial PSD in coal combustion simulation remains an open question. To answer this, the Gaussian‐quadrature theory was applied to provide a discrete reconstruction of the PSD described by Rosin‐Rammler and Upper‐Limit functions. From the perspective of the moments of PSD, a theoretical analysis was conducted for mass, momentum, and energy exchange terms between gas and particle phases in pulverized coal jet flame. The analysis presented that the first, the second and the third moments of PSD were the most important to capture the flame ignition behavior. The large eddy simulations showed that the proposed method with two or more particle sizes could provide good predictions of flame ignition distances. And Sauter mean diameter (D32) could represent the complete PSD when predicting the ignition behavior of pulverized coal jet flame.

Numerical Investigation on the Detailed Structure of a Coaxial Coal Jet Flame Using Large-Eddy Simulation with Elementary Reactions

A numerical simulation of a pulverized coal jet flame was performed to investigate the physics and flame structure in detail by means of large-eddy simulation with an elementary reaction mechanism....

Assessment of Global and Network Models of Devolatilization for Numerical Analysis of Pulverized Coal Combustion

ABSTRACTMathematical modeling of devolatilization plays a significant role in the prediction of flame development in solid fuel combustion simulations. Commonly implemented devolatilization models available in the literature to simulate pulverized coal combustion are reviewed and evaluated. The effect of the particle heating rate on the onset of devolatilization and on final volatile yield is discussed. Selected devolatilization models are investigated with respect to their implementation in Computational Fluid dynamics (CFD) simulations of methane-assisted pulverized coal jet flame experiments. Differences in volatile release region, particle temperature and species concentration are presented and discussed with comparisons to the experimental data.

Eulerian LES simulation of coal jet flame with a simplified DQMOM model

Eulerian methods are popular for describing particle dispersion and reaction in gas-particle two-phase flow simulations. However, for problems with wide Particle Size Distributions (PSD), such as those associated with coal combustion, the PSD has to be represented by a limited number of particle sizes in order to decrease the computational costs of the transport equations. The objective of this study is to investigate the effects of the number of particle sizes on the Eulerian LES simulation of a pulverized coal jet flame. In this study, experimental work on a Hencken-type entrained flow reactor was carried out to provide the validation data of flame ignition distances. A simplified Direct Quadrature Method of Moments (DQMOM) method was proposed in order to decrease the computation cost in the Eulerian framework. Several simulation cases with different quadrature nodes (particle sizes) were conducted. The flame ignition distance was well predicted when two and more quadrature nodes were adopted in the simulation. But it was overestimated when only one quadrature node was considered. So one particle size is not enough to represent a wide PSD of coal in the Eulerian LES simulation of pulverized coal flames. When 5 particle sizes were adopted in the simulation, the ignition distances and time-variant fluctuations can be well predicted when compared with experimental results under different operating conditions. This study shows that LES with the simplified DQMOM method can describe the ignition distance in gas-solid two-phase reacting flows even with only a few quadrature nodes.

Numerical analysis of particle dispersion and combustion characteristics on a piloted coaxial pulverized coal jet flame

A numerical simulation was performed and validated on a piloted coaxial pulverized coal jet flame by means of large eddy simulation (LES) in which the dynamic Smagorinsky SGS model and the Lagrangian particle tracking model were employed. In the devolatilization process, a postulate substance, CaHbOc, was considered as the devolatilized gas. For gaseous phase reactions, a simple two-step global kinetic mechanism was implemented. For a solid phase reaction, a two-step char oxidation reaction was installed. Results were compared to an experimental result in terms of particle distribution, gaseous temperature, gas composition, and particle velocity. The particle distribution result shows that particle dispersion for the combustion case is narrower than that for the non-combustion case. This phenomena is discussed by observing an axial interphase momentum transfer. The axial momentum of fluids can be maintained as high or similar to the particles by combustion, so momentum transfer is limited in the combustion case. It is also confirmed in the particle and gas velocity results. Time-averaged temperature and mole fractions of gases capture the featured characteristics of the experiment well. From these results, an effect of devolatilized gas combustion on particle dispersion and stream flow are discussed. And it is revealed from instantaneous results that particles tends to form groups following the edge of the devolatilized gas combustion.

Effect of multiphase radiation on coal combustion in a pulverized coal jet flame

The accurate modeling of coal combustion requires detailed radiative heat transfer models for both gaseous combustion products and solid coal particles. A multiphase Monte Carlo ray tracing (MCRT) radiation solver is developed in this work to simulate a laboratory-scale pulverized coal flame. The MCRT solver considers radiative interactions between coal particles and three major combustion products (CO2, H2O, and CO). A line-by-line spectral database for the gas phase and a size-dependent nongray correlation for the solid phase are employed to account for the nongray effects. The flame structure is significantly altered by considering nongray radiation and the lift-off height of the flame increases by approximately 35%, compared to the simulation without radiation. Radiation is also found to affect the evolution of coal particles considerably as it takes over as the dominant mode of heat transfer for medium-to-large coal particles downstream of the flame. To investigate the respective effects of spectral models for the gas and solid phases, a Planck-mean-based gray gas model and a size-independent gray particle model are applied in a frozen-field analysis of a steady-state snapshot of the flame. The gray gas approximation considerably underestimates the radiative source terms for both the gas phase and the solid phase. The gray coal approximation also leads to under-prediction of the particle emission and absorption. However, the level of under-prediction is not as significant as that resulting from the employment of the gray gas model. Finally, the effect of the spectral property of ash on radiation is also investigated and found to be insignificant for the present target flame.

LES of pulverized coal combustion with a multi-regime flamelet model

It is well known that premixed and non-premixed flamelets coexist in the combustion system, especially for multiphase combustion (Luo et al., 2011; Bai et al., 2016). In this work, a multi-regime combustion model for partially premixed multiphase combustion (2PMC) is developed in the framework of large eddy simulation (LES). In this model, the multi-regime combustion mode is decomposed into two pure combustion regimes (i.e. premixed and non-premixed) through combustion regime indicator. Different combustion models are chosen for different combustion regimes. For example, for premixed combustion regime, the flamelet generated manifold (FGM) tabulation method in combination with the artificially thickened flame (ATF) approach is used. For non-premixed combustion regime, on the other hand, the thermo-chemical quantities will be extracted from the non-premixed chemtable generated by the flamelet/progress variable (FPV) approach. The proposed multi-regime flamelet model is then extended to adapt to the pulverized coal combustion (PCC) and applied to a laboratory-scale pulverized coal jet flame. The simulation results show that the percentage of the volume-weighted premixed combustion regime of this studied flame is up to 18%. Quantitative comparisons between the experimental data and the numerical results with the present multi-regime flamelet model, the FPV model, and the Eddy Break Up (EBU) model shows that the proposed multi-regime flamelet model has several advantages. It performs better than the FPV model in the regions where premixed combustion mode prevails and also much better than the EBU model in species concentration prediction.

A DNS study on effect of coal particle swelling due to devolatilization on pulverized coal jet flame

The effects of particle swelling and shrinkage during devolatilization and char combustion on the flame characteristics are investigated by performing DNS of a lab-scale turbulent pulverized-coal straight-jet flame. The obtained characteristics are compared with those in the previous DNS (Hara et al., 2015), in which particle swelling and shrinkage are not taken into account. The results show that the changes in coal-particle diameter because of swelling and shrinkage during devolatilization and char combustion do not affect significantly the gas temperature distribution of the turbulent pulverized-coal straight-jet flame, although it explicitly affects the coal-particle diameter distribution in the flame.

Measurements and CFD modeling of a pulverized coal flame with emphasis on ash deposition

Measurements of fly ash deposition in a 15kW pulverized coal jet flame and CFD-based mathematical modeling have been performed. The deposits have been collected at two ports at particle Stokes numbers in the 0.02–0.34 range and particle kinetic energies not larger than 2×10-9J. Inertial impaction and thermophoresis have been identified as main mechanisms of particle transport towards the deposition surfaces. Deposition rates on air-cooled probes (600–700°C surface temperature) have been measured to be 24% (Port 2) and 79.4% (Port 3) larger than those measured on uncooled probes (1150°C surface temperature) due to the enhanced role of thermophoresis. Complex dependencies of the deposition rate on the probe surface temperature and the probe location have been observed. The CFD-model predictions are able to reproduce these dependencies after adjustments to the particle sticking sub-model. The paper contains estimations of both the impaction and sticking efficiencies.

Direct numerical simulation of a pulverized coal jet flame employing a global volatile matter reaction scheme based on detailed reaction mechanism

A two-step global reaction scheme for the volatile matter of coal is proposed, and the unsteady coal particle and combustion behaviors in a turbulent pulverized coal jet flame are investigated by performing a direct numerical simulation (DNS) employing the proposed global reaction scheme. The two-step global reaction scheme is constructed to take into account the properties of the volatile matter such as transport coefficients, laminar flame speed and unburned gas temperature and to be applicable to various coal types, and it is validated by comparing the results with those obtained by the detailed reaction mechanism which includes 158 chemical species and 1,804 reactions. The validity of the DNS is also assessed by comparing the results with those in the previous experiment (Hwang et al., 2005) [16], and the unsteady coal particle motions and combustion characteristics are examined in detail. The results show that the proposed two-step global reaction scheme for the volatile matter of coal can precisely predict the laminar flame speed and burned gas temperature for various coal types from bituminous to low-rank coals over wide ranges of conditions of equivalence ratios, pressures and unburned gas temperatures. In addition, it can correctly take into account the effects of dilutions by H2O and CO2 which compromise the evaporated moisture from coal and products of char reaction. It is also verified that a lab-scale turbulent pulverized coal jet flame is well predicted by the DNS employing the proposed global reaction scheme. That is, the pulverized coal particles’ velocity and its fluctuation and the characteristics of particle preferential motions are in general agreement with those observed in the experiment. The DNS reveals that in the turbulent pulverized coal jet flame, there appear premix and diffusion flame layers inside and outside, respectively. In addition, the reaction of the volatile matter and O2 in coal-carrier air occurs in the inner premixed flame layer, whereas the reactions of the volatile matter and CO and O2 in surrounding air occur in the outer diffusion flame layer.

J0110102 微粉炭噴流火炎の直接数値シミュレーション : 当量比の影響

A direct numerical simulation (DNS) is applied to a laboratory scale pulverized coal jet flame. Pulverized coal particles are injected with primary air in the double stream burner and the effects of equivalent ratio in the primary stream on the coal jet flame are investigated. In the DNS, the number of grid is about 0.17 billion, and the coal particle motions are tracked by a Lagrangian approach. The results show that the gas temperature of the flame annularly formed downstream the burner increases as the equivalent ratio increases. While, the spatial distribution of the gas temperature inside the flame becomes uniform under the low equivalence ratio condition.

Large-eddy simulation of pulverized coal jet flame – Effect of oxygen concentration on NOx formation

Large-eddy simulation is applied to a laboratory-scale open-type pulverized coal flame generated by a triple stream burner, and the NO production and reduction in oxy-fuel condition are investigated for the first time. Pulverized Cerrejon coal which is classified as bituminous coal is used as a fuel. The results show that regardless of the equivalence ratio, as the O2 concentration increases from 21% to 40%, O2 consumption becomes marked because gas temperature rises and oxidation reaction is enhanced by the higher concentration of O2. Also, NO is formed rapidly due to the oxidation reaction of nitrogen from volatile matter of coal, and its concentration reaches a few hundred ppm further downstream. After the rapid formation, in the case of equivalence ratio larger than unity, NO decreases, because the reducing atmosphere becomes dominant due to the lack of O2. The trend becomes significant as the O2 concentration in the carrier gas increases from 21% to 40%. In the case of equivalence ratio less than unity, on the other hand, NO does not decrease clearly, because the oxidizing atmosphere contributes to the further formation of NO. Present study shows the usefulness of the large-eddy simulations for predicting the characteristics of pulverized coal flames.

Numerical investigation of pulverized coal jet flame characteristics under different oxy-fuel conditions

Pulverized coal combustion in mixture of oxygen and recycled flue gasses, known as oxy-fuel combustion, is considered as one of the several possible alternatives to conventional pulverized coal combustion. Switching from conventional pulverized-coal combustion to oxy-fuel combustion brings significant changes in flame properties among which the most important are ignition properties and flame stability. This paper presents the results of experimental and numerical analysis of ignition phenomena under O2/CO2 mixtures with different oxygen content. The main focus of the presented paper is to suggest novel ignition sub-model which can describe all possible ignition mechanisms. Proposed ignition sub-model together with Large Eddy Simulation (LES) turbulence modeling enables accurate prediction of main flame characteristics: ignition point position, ignition temperature, and flame stability.

Large Eddy Simulation of Pulverized Coal Jet Flame Ignition Using the Direct Quadrature Method of Moments

The goal of this study is to show that large eddy simulation (LES) combined with the direct quadrature method of moments (DQMOM) can accurately predict coal flame ignition mechanisms and more generally be a valuable tool for predicting and understanding coal flame characteristics. A validation study of the ARCHES LES tool is performed. LES is applied to a pulverized coal jet flame ignited by a preheated gas flow. A DQMOM has been developed to track the particle phase in an Eulerian framework. The simulation results are compared qualitatively and quantitatively to experimental observations for different inlet stoichiometric ratios. LESs are able to capture the various combustion regimes observed during flame ignition and to accurately model the flame stand-off distance sensitivity to the stoichiometric ratio. The gas temperature and coal burnout predictions are also examined and show good agreement with experimental data. Overall, this study shows that high-fidelity LESs combined with DQMOM can yield a deeper understanding of complex coal flames and be a valuable tool for the design, retrofit, and scale-up of oxy-coal boilers.

Direct Numerical Simulation of Pulverized Coal Combustion in a Hot Vitiated Co-flow

A compressible direct numerical simulation (DNS) solver for pulverized coal combustion has been developed and used to study a pulverized coal jet flame with a Reynolds number of 28 284 based on the nozzle diameter. An eighth-order center differential scheme combined with an explicit tenth-order filter is used for spatial discretization. The classical fourth-order Runge–Kutta method is used for time integration. The characteristic non-reflecting boundary conditions are used to describe the boundary conditions. A comprehensive model for coal combustion is applied, and the reaction mechanism of CH4 with five species and two-step reactions is adopted for gas-phase combustion. The grid system has been carefully designed to make sure that turbulent scales and chemical reaction scales are reasonably resolved and the point-source assumptions of particles are valid. The simulation is partially validated against the experiment, and the particle behavior is investigated. It is found that, in the upstream region, the reaction rate is quite scattered and a single particle is found inside the burning flame to form an individual particle combustion mode. While in the downstream region, the reaction zone is more continuous, with a large number of particles enclosed, which characterizes the group combustion mode. Conditional statistics with respect to the mixture fraction are also obtained to provide insights into coal combustion and the related models.

Numerical simulation of pulverized coal jet flame employing the TDP model

The effect of the devolatilization model on the coal particle behavior is investigated in detail by performing numerical simulations of a simple pulverized coal jet flame formed by a small jet burner (0.5kg-coal/h) [S. Hwang et al., Energy & Fuels 2005;19:382–92]. As the devolatilization model, widely-used conventional devolatilization model, two competing reaction rate model and newly-proposed tabulated-devolatilization-process model (TDP model) [N. Hashimoto et al., Combust Flame 2012;159:353–366] are used. The results show that the coal particle velocities predicted by the TDP model are in better agreement with the experiments than those by the other models, with a slight increase in computation time. The difference in the mean axial particle velocity between the simulations is caused by the difference in the axial gas velocity, which is ultimately caused by the difference in the volatile matter evolution rate. It is also found that the devolatilization model has great influence on particle velocity prediction compared to the turbulence model, the gas-phase combustion model and the radiation model.

Measurement of Flame Height by Image Processing Method

In order to measure the flame height (Lf), the image processing method was employed. The jet flame image was captured by the common-used digital camera. The resolution (C, with the unit of mm/pixel) of the pixel of the digital camera was calibrated by metric scale. The jet flame image was then successively processed by ROI (region of interest) processing, gray processing, binarization and edge detection. Through the image processing, the pixels where the flame root and the flame tip were located was detected. The flame region expressed in pixels can be easily obtained by the difference between the flame root (X1) and the flame tip(X2). The flame height (Lf) was calculated by multiplying the difference in pixels (Lp= X2-X1) with the resolution of pixel (in millimetres per pixel). Using the proposed method, the flame heights of the pulverized coal jet flame at the high temperature air combustion condition were measured. The effect of experimental parameters on the flame heights of the coal jet flame was discussed.