Ash Deposition Modeling Incorporating Mineral Matter Transformations Applied to Coal and Biomass Co-firing

Paper from the AFRC 2013 conference titled Ash Deposition Modeling Incorporating Mineral Matter Transformations Applied to Coal and Biomass Co-firing by Bradley Adams

behavior when firing biomass in coal-fired boilers. The Nalco Mobotec System can enable boilers to switch to biomass co-firing or burn up to 100 percent biomass with no change in boiler load capacity, while meeting strict emission control requirements. As biomass often contains inorganic compounds that can impact deposition/slagging/fouling/corrosion, care has been taken to consider potential ash impacts, offering chemical treatments that have successfully reduced deposition rates and reduced the tenacity of the deposits that do form. In an effort to mitigate risk and to provide guidance related to the magnitude, location, and sintering extent of deposition, REI has worked with Nalco Mobotec to complement their in-house modeling efforts with its unique tools and expertise in this area. Results of CFD-based modeling efforts incorporating REI's Mineral Matter Transformation and Deposit Build-up Model quantitatively illustrated the impact of various pelletized biomass fuels on deposition behavior and fireside corrosion for a 660 MW opposed-wall coal-fired boiler. Three different woods and one combined wood and straw mixture were evaluated. Simulation results indicated that biomass co-firing relative to 100% coal-firing produced: • An increase in deposition rates and deposit sintering, resulting in some reduction in wall heat transfer, • Decreases in water wall corrosion rates, • Modest increases in furnace exit gas temperature (FEGT), • 35-40% decrease in NOx emissions, • Similar CO emissions, • Slight decreases in unburned carbon in fly ash.

AFRC 2013 Industrial Combustion Symposium Ash Deposit Modeling, Adams, et al. 1 Ash Deposition Modeling Incorporating Mineral Matter Transformations Applied to Coal and Biomass Co-firing Bradley Adams*, Kevin Davis, Huafeng Wang, James Valentine Reaction Engineering International Brian Smith, Liming Shi, Ed Pozzobon Nalco Mobotec Inc. Reaction Engineering International (REI) and Nalco Mobotec have teamed to quantitatively evaluate ash behavior when firing biomass in coal-fired boilers. The Nalco Mobotec System can enable boilers to switch to biomass co-firing or burn up to 100 percent biomass with no change in boiler load capacity, while meeting strict emission control requirements. As biomass often contains inorganic compounds that can impact deposition/slagging/fouling/corrosion, care has been taken to consider potential ash impacts, offering chemical treatments that have successfully reduced deposition rates and reduced the tenacity of the deposits that do form. In an effort to mitigate risk and to provide guidance related to the magnitude, location, and sintering extent of deposition, REI has worked with Nalco Mobotec to complement their in-house modeling efforts with its unique tools and expertise in this area. Results of CFD-based modeling efforts incorporating REI's Mineral Matter Transformation and Deposit Build-up Model quantitatively illustrated the impact of various pelletized biomass fuels on deposition behavior and fireside corrosion for a 660 MW opposed-wall coal-fired boiler. Three different woods and one combined wood and straw mixture were evaluated. Simulation results indicated that biomass co-firing relative to 100% coal-firing produced: • An increase in deposition rates and deposit sintering, resulting in some reduction in wall heat transfer, • Decreases in water wall corrosion rates, • Modest increases in furnace exit gas temperature (FEGT), • 35-40% decrease in NOx emissions, • Similar CO emissions, • Slight decreases in unburned carbon in fly ash. INTRODUCTION The management of ash in solid fuel fired boilers has long been a key concern in utility boilers and is a particularly important issue when a significant change in fuel selection is considered for an existing boiler. The heterogeneity of coal and the complexity of its evolution during combustion in a pulverized coal fired boiler have been noted.1 The development of advanced methodologies for predicting this behavior has continued for many years. More recently the improvement of computational fluid dynamics (CFD) modeling tools and cost-effective computational equipment has created an increasingly valuable platform within which more fundamentally sound descriptions of ash behavior can be implemented. The increasingly convenient availability of fuel characterization techniques including computer controlled scanning electron microscopy (CCSEM) and chemical fractionation has served to complement the development and usefulness of such implementations. Efforts to improve the capability of CFD-based * Corresponding author. Tel.: [phone]; E-mail: [e-mail] 1 Borio, R. and A. Levasseur "Overview of coal ash deposition in boilers," ACS Fuel Chemistry Division Preprints 1984, 29(4), 1984.AFRC 2013 Industrial Combustion Symposium Ash Deposit Modeling, Adams, et al. 2 tools to predict ash evolution and deposit build-up during pulverized coal combustion has obvious value for design and risk mitigation purposes involving boilers and REI has been working toward this objective for several years. 2 Recent modeling developments in this area have proven valuable for troubleshooting equipment modifications in coal fired boilers. The variation in ash composition/distribution/morphology for pulverized coal can be quite large. However, a great deal of work has been done to characterize coal ash and its behavior within pulverized coal fired boilers. The behavior of ash from biomass fuels has the potential to vary even more than that of coal. In addition, the interaction of biomass with coal in a co-firing scenario can create additional complications. This paper will focus on the adaptation of the aforementioned modeling approaches and their application to coal-biomass co-firing scenarios involving multiple biomass fuels available in pelletized form. GLACIER CFD CODE The simulation of coal combustion including multi-phase flow and kinetic-rate-limited emissions chemistry is a highly specialized field. Reaction Engineering International (REI) is a recognized leader in this area. With its relevant expertise in solid fuels, turbulence chemistry interactions, and radiative heat transfer, REI has established a reputation and track record of successful CFD modeling of solid fuel combustion in practically every type of equipment available for use in utility and industrial applications. In the power generation industry alone, REI has performed detailed CFD simulations of well over 200 solid-fuel fired boilers including many biomass and biomass co-firing applications. The model used in this evaluation is a CFD-based reacting flow code, GLACIER, developed by REI. The code couples the effects of turbulent fluid mechanics, gas-phase combustion chemistry, turbulent particle dispersion, heterogeneous particle reactions, and convective and radiative heat transfer. GLACIER assumes that the flow field is a continuum field that can be described locally by general conservation equations. The flow is assumed to be steady-state and gas properties are determined through local mixing calculations. The fluid is assumed to be Newtonian and dilatation is neglected. The comprehensive model uses an Eulerian framework and can incorporate either Reynolds- or Favre-averaging. The code couples turbulent fluid mechanics and chemical reactions, while using progress variables to track the turbulent mixing and equilibrium chemistry to describe chemical reactions. Within the model, the rate at which the primary combustion reactions occur is assumed to be limited by the rate of mixing between the fuel and the oxidizer, which is a reasonable assumption for the chemical reactions governing heat release. The thermodynamic state at each spatial location is a function of the enthalpy and the degree of mixing of two mixture fractions, one of which corresponds to the coal off-gas. The effect of turbulence and mixing on mean chemical composition is incorporated by assuming that the mixture fractions are defined by a "clipped Gaussian" probability density function (pdf) having a spatially varying mean and variance. The mean and variance are computed numerically at each grid point and mean chemical species concentrations are obtained by convolution over the pdf. Species concentrations are calculated as properties based on the local stream mixture and enthalpy. Computationally, this is appreciably more efficient than tracking individual species. An exception to this approach is the tracking of NOx species, which is performed using finite-rate chemistry calculations. Particle mechanics are solved by following the mean path or trajectory for a discretized group or ensemble of particles in a Lagrangian reference frame. Particle mass and momentum sources are converted from a Lagrangian to an Eulerian reference frame where they are coupled with gas phase fluid mechanics. The radiative intensity field is solved based on properties of the surfaces and participating medium and the resulting local flux divergence appears as a source term in the gas-phase energy equation. 2 Eddings, E., Davis, K., Heap, M., Valentine, J. and Sarofim, A. (2001), Mineral Matter Transformation During Pulverized Coal Combustion. Dev. Chem. Eng. Mineral Process., 9: 313-327.AFRC 2013 Industrial Combustion Symposium Ash Deposit Modeling, Adams, et al. 3 More detailed information regarding the computational model has been published previously (see http://www.reaction-eng.com/downloads/publications.html). SLAGGING MODEL DESCRIPTION REI's approach in this area is to integrate detailed slagging submodels into our comprehensive multiphase, reacting CFD model in a pseudo-steady state calculation procedure. The CFD simulation provides flame-side information and local deposition rates of entrained solids. The slagging submodel performs the deposit growth, sintering and properties calculation and updates the wall boundary condition for the comprehensive CFD simulation. Total deposition time is discretized into several time steps during which the deposition rate is assumed to be constant and the steady-state comprehensive CFD model can be used. The transient deposition build-up process in the furnace can be addressed with this approach with adequate accuracy and manageable computational time. The two key elements in this approach are the prediction of (1) deposition rate and (2) deposit properties and growth. The deposition rate calculation is presented in detail first, followed by the discussion of the approach used to predict deposit property and growth. Mineral Matter Transformation After devolatilization of coal, the remaining carbon (char) will continue to oxidize over a relatively longer period. The mineral inclusions in the char are heated to high temperatures during this process. The inclusions can melt and coalesce as the char surface recedes. Organically associated elements are largely incorporated into coalescing ash particles. If the char particle remains whole during combustion, then the result will be one ash particle per initial coal particle. However, some coals undergo char fragmentation during combustion, which results in more than one ash particle per coal particle. One of the critical aspects of predicting mineral matter transformations is to accurately represent the minerals as they are distributed within the coal particle. In order to account for the statistical nature of the distribution of minerals and the mineral behavior during combustion, the Random Coalescence Model for prediction of fly ash size and composition is used. In this approach, the coal particles are divided into a number of size bins and the mineral particles (as determined from CCSEM) are distributed into randomly selected coal particles. During the distribution process, the composition and amount of the minerals in each coal particle are tracked. Any mineral inclusion that exceeds the volume of the coal particle is assigned as an excluded mineral. Once the complete distribution of minerals in coal particles has been computed, then the coalescence algorithm is invoked. Char fragmentation estimates are made within the Random Coalescence Model3 that is incorporated in a Mineral Matter Transformation (MMT) model. This MMT model was used in the present work to predict fly ash size and composition distribution. Ash Deposition Rate The initial suite of ash particles (in terms of size and composition) from each individual coal particle is assumed to be known and is estimated by the MMT model, expressed by <n> groups of compositions with a weighting factor for each group. When the coal particle cloud in the CFD model interacts with wall cells, the fraction of particles that stick on the wall surface for a particle was calculated after the method outlined by Walsh and co-workers.4 3Barta, L. E., M. A. Toqan, et al. (1992). "Prediction of Fly Ash Size and Chemical Composition Distributions: The Random Coalescence Model." Proceedings of the Twenty-Fourth International Symposium on Combustion, The Combustion Institute, Pittsburgh, PA. 4 Walsh, P.M.; Sayre, A.N.; Loehden, D.O.; Monroe, L.S.; Beér, J.M.; Sarofim, A.F. "Deposition of Bituminous Coal Ash on an Isolated Heat Exchanger Tube: Effects of Coal Properties on Deposit Growth." Prog. Energy Combust. Sci. 1990, 16, 327-346.AFRC 2013 Industrial Combustion Symposium Ash Deposit Modeling, Adams, et al. 4 Particle viscosity has been found to provide a reasonable representation of the particle properties that affect sticking probability during particle-surface interactions. A viscosity of 105 poise was suggested by Richards et al.,5 based on the relative deposition rates observed for two different coals in a laminar drop tube furnace and is used here. A viscosity model developed by Senior and Srinvasachar6 is used to calculate the viscosity of silicate glass particles as a function of both temperature and composition. The viscosity calculation is based on the ash composition and the particle temperature, the influence of unburned carbon on particle viscosity is accounted for in such a way that when the particle burnout is larger than a critical burnout, the ash tends to dominate the material at the outer surface of the particle, thus the viscosity of ash provides a representative measure for calculating the sticking probability. When particle burnout is less than a critical burnout, the ash tends to be buried in the carbon matrix, thus, a large viscosity is assigned to the particle in calculating the sticking probability. After deposition, the deposit composition in the wall cells of the CFD model is recorded and the ash composition in the coal particle along the trajectory is adjusted based on the ash composition balance by recalculating a weighting factor for ash composition in the coal particles. Deposit Properties and Growth A quantitative description of deposit properties is necessary in order to assess the effect of the ash deposition on heat transfer and to assess deposit strength. Key physical properties related to heat transfer are the emissivity, absorptivity, and thermal conductivity. The viscosity, porosity and surface tension of the deposit are related to sintering and deposit strength. The emissivity and absorptivity of the deposit were found to decrease with temperature and to increase with particle size or iron content prior to sintering. A sintered or fused deposit is found to have a high emissivity of approximately 0.9. A quantitative method for calculation of deposit emissivity and absorptivity was reported by Wall et al.7 The thermal conductivity of ash deposits was reviewed by Gupta et al.,8 who recommended methods for determination of the effective thermal conductivity for different types of deposits. The thermal conductivity of the partially sintered deposit is more complex, and can be estimated by using volume-averaged properties. The thermal conductivity of a completely sintered or molten deposit has a high value, close to the thermal conductivity of the solid deposit. A deposit is characterized by partially sintered layers near the heat transfer surface and highly sintered layers exposed to the fireside. The development of deposit strength is due primarily to viscous flow sintering, which is determined by the surface tension and viscosity of depositing particles. Methods for estimating the surface tension of a slag deposit have been summarized by Mills.9 Deposit properties (porosity, thermal conductivity, emissivity, absorptivity, etc.), which are determined by the deposit morphology, composition, and temperature, are critical in the prediction of heat transfer and deposit growth. Equilibrium is not reached locally in the deposit, and viscous flow sintering is the main mechanism responsible for the change of deposit morphology and properties. Sintering changes the 5 Richards, G.H.; Slater, P.N.; Harb, J.N. Simulation of Ash Deposit Growth in a Pulverized Coal-Fired Pilot Scale Reactor. Energy & Fuels 1993, 7, 774-781. 6 Senior, C., Srinivasachar, S. Viscosity of Ash Particles in Combustion Systems for Prediction of Particle Sticking. Energy & Fuels 1995, 9, 277-283. 7 Wall, T.F.; Bhattacharya, S.P.; Zhang, D.K.; Gupta, R.P.; He, X. "The Properties and Thermal Effects of Ash Deposits in Coal-Fired Furnaces." Prog. Energy Combust. Sci. 1993, 19, 487-505. 8 Gupta, R.P.; Wall, T.F.; Baxter, L. "The Thermal Conductivity of Ash Deposits: Particulate and Slag Structures." Presented at Impact of Mineral Impurities in Solid Fuel Combustion. An Engineering Foundation Conference, Kona, Hawaii, Nov. 2-7, 1997. 9 Mills, K.C. Estimation of Physicochemical Properties of Coal Slags and Ashes. In Mineral Matter and Ash in Coal, K.S. Vorres (Ed), American Chemical Society, Washington, D.C., 195-214, 1986.AFRC 2013 Industrial Combustion Symposium Ash Deposit Modeling, Adams, et al. 5 pore structure of a deposit and causes an increase of deposit cluster size or connectivity. However, the porosity of a deposit appears to decrease significantly only in the outer region of the deposit where the temperature is sufficiently high. Deposit Sintering Viscous flow sintering in the deposit is largely determined by viscous forces and surface tension forces. Surface tension forces cause adjacent particles to sinter together by increasing the width of the neck between particles and by decreasing the distance between the centers of the particles. Viscous forces are required to be low enough not to hinder sintering. The extent of sintering is described by a sintering percentage or the densification parameter. The sintering between particles with uniform size and uniform composition can be predicted by using Frenkel's model10 for the initial stage of sintering, Scherer's model11 for the middle and/or initial stages of sintering, and Mackenzie and Shuttleworth's model12 for the final stage of sintering. All these models predict the relative density as a function of particle size, viscosity, surface tension and time. The influence of particle shape and size distribution on the sintering rate has been discussed elsewhere.13,14 The effect of the non-uniform composition on the sintering rate has also been recognized as important, although it has not been discussed quantitatively. Ash deposits are formed from fly ash particles that come in a wide variety of sizes, shapes, and compositions. Thus, accurate prediction of the sintering rate in an ash deposit is a very difficult task. Also, sintering predictions must be simple enough computationally to permit integration with a comprehensive combustion code for the simulation of deposition behavior in large-scale boilers. Viscous-flow sintering is the principal mechanism responsible for changing the structure of a deposit. Based on experimental observations,15 initial sintering tends to increase the connectivity of the deposit without making significant changes in the deposit porosity. Observations also indicate substantial porosity changes in the later stages of sintering. These observations have been incorporated into the sintering model developed here. Fly ash particles that adhere to the surface and form the ash deposit are assumed to be spherical for the purpose of simplification. This assumption is not expected to have a large effect on the predicted results. Before the calculation of sintering in the deposit, particles that adhere to the surface and form the deposit are divided into a finite number of groups (according to their sizes and compositions) in order to represent the behavior of a large number of particles. Grouping of the particles is required in order to make the sintering calculations computationally feasible. The local overall extent of sintering in the deposit as a function of time was calculated by dividing the deposit into discrete layers. The extent of sintering for each layer during the time step Δti was approximated by the mass average of the extent of sintering for each of the relevant particle groups. High-viscosity particle groups were not allowed to mix and sinter until other groups of equal or smaller size had completely sintered. Once the sintering of a particular particle group was complete, the mass of this group was distributed among the high-viscosity particle groups. This procedure permitted approximation of the local extent of sintering as a function of time, particle size, and composition, and was used to estimate the deposit properties. 10 Frenkel, J. Viscous Flow of Crystalline Bodies Under the Action of Surface Tension. J. Phys. (Moscos) 1945, 9, 385-390. 11 Scherer, G.W. Sintering of Low-Density Glasses: 1, Theory. Journal of The American Ceramic Society 1977, 60, 236-239. 12 Mackenzie, J.K.; Shuttleworth, R. Phenomenological Theory of Sintering. Proc. Phys. Soc. London. 1949, 62 (12-B), 833-852. 13 Chappell, J. S.; Ring, T. A. Particle Size Distribution Effects on Sintering Rates. Journal of Applied Physics 1986, 60(1), 383-391. 14 Coble, R.L. Effects of Particle-Size Distribution in Initial-Stage Sintering. Journal of the American Ceramic Society 1973, 56(9), 461-466. 15 Wang, H.; West, J.; Harb, J.N., Microanalytical Characterization of Slagging Deposits in a Pilot-Scale Combustor. Energy & Fuels 1999, 13, 570-578.AFRC 2013 Industrial Combustion Symposium Ash Deposit Modeling, Adams, et al. 6 One complication of this approach is the influence of unburned carbon in the deposit on the deposit growth and sintering. The effect of unburned carbon on deposit growth is determined by the release rate of unburned carbon in the deposit, which in turn influences the deposit surface sticking probability. The release rate of deposited unburned carbon depends upon the local gas environment where the deposit is formed. The viscosity of a partially burned carbon particle is calculated in such a way that when the particle burnout exceeds a critical burnout, the ash tends to be dominate the outer surface of the particle, thus the viscosity of the ash particle is representative of the particle. When the particle burnout is less than a critical burnout, the ash tends to be contained within the carbon matrix; thus, a large viscosity is assigned to the particle. Deposit properties include deposit porosity, local thermal conductivity, emissivity, and absorptivity, etc. The initial deposit porosity was assumed to be 0.6 in the model, a typical value for a particle-based deposit and consistent with values measured by Anderson et al.16 The local thermal conductivity of an ash deposit depends on the extent of sintering, the deposit porosity and the temperature. The emissivity and absorptivity of the particulate deposit surface were determined by the particle size and composition distribution as per Harb et al.17 The method utilized the scattering albedo and the scattering asymmetry factor from single particle calculations performed with the Mie code of Bohren and Huffman.18 Once the surface deposit begins to sinter, the emissivity is assumed to increase linearly with temperature. These thermal and physical properties of the deposit were then used to predict the deposit growth and heat flux through the deposit as discussed in the next section. Deposit Growth and Heat Flux Deposit growth and heat flux through the deposit are simulated for a time period during which the deposition rate and the incident heat flux are assumed to remain constant. The procedure used for the simulation of deposit growth and heat flux through the deposit has been adapted from Wang et al.19 The simulated period is divided into a finite number of time steps and the amount of deposit, which accumulates during the ith time step, is referred as the ith deposit layer. First, the extent of sintering that occurs during the time step i is calculated for each layer of the existing deposit (layers 1 to i-1) and the deposit thermal conductivities for each of these layers are updated. Next, the mass of the deposit, which accumulates during the time step i, is determined. The thickness and sintering of the new deposit layer is then calculated and the thermal conductivity, emissivity, and absorptivity of the new deposit layer are approximated as discussed in the previous section. Once the thermal resistance of the deposit and the emissivity and absorptivity of the deposit surface have been calculated, it is then possible to calculate the deposit surface temperature and the net heat flux through the deposit. These temperatures at different layers are saved and used to calculate the sintering in these deposit layers at the next time step. The entire process is repeated until the total time specified for deposit growth has been reached. The overall approach to describe slagging is to use the slagging submodel together with a comprehensive CFD combustion model. The comprehensive CFD simulation provides the flame-side information and deposition rate. The slagging submodel performs the deposit growth, sintering and property calculations and then updates the wall boundary condition for comprehensive CFD simulation. Total deposition time is discretized into several time steps during which the deposition rate is assumed to be constant and the 16 Anderson, D.W.; Viskanta, R.; Incropera, F.P. Effective Thermal Conductivity of Coal Ash Deposits at Moderate to High Temperatures. Journal of Engineering for Gas Turbine and Power 1987, 109, 215-221. 17 Harb, J.N.; Slater, P.N.; Richards, G.H. A Mathematical Model for the Build-Up of Furnace Wall Deposits. In the Impact of Ash Deposition on Coal Fired Plants. J. Willamson and F. Wigley (Eds), Solihull, Taylor & Francis press, 637-644, 1993. 18 Bohren, C.F.; Huffman, D.R. Absorption and Scattering of Light by Small Particles, John Wiley & Sons, 1983. 19 Wang, H; Harb, J.N. "Modeling of Ash Deposition in Large Scale Combustion Facilities Burning Pulverized Coal." Progress in Energy and Combustion Science 1997, 23, 267-282.AFRC 2013 Industrial Combustion Symposium Ash Deposit Modeling, Adams, et al. 7 steady-state comprehensive CFD model can be used. The transient deposition build up process in the furnace can be addressed with this approach with adequate accuracy and manageable computational time. The two key elements in this approach are how to predict the deposition rate and how to predict the deposit property and growth. CORROSION MODEL Correlations are used to predict water wall corrosion due to three mechanisms: 1) the presence of near wall H2S, 2) the deposition of unoxidized sulfur containing material such as FeS, and 3) the presence of chlorine. The correlation for H2S corrosion was developed by others while the latter two correlations have been developed by REI during the course of multiple government and industry funded efforts. These correlations have been coupled with GLACIER. Prediction of water wall corrosion is performed in a post-processing manner; GLACIER output files are interpreted and the predicted flow fields and wall conditions are used as input to the correlations. For the prediction of water wall corrosion due to gas phase H2S, the empirical correlations of Kung20,21 for low and high chromium steels are used. A second proprietary correlation for the prediction of water wall corrosion arising from the deposition of unoxidized sulfur containing deposits is used. The corrosion rate for this parallel mechanism is a function of coal sulfur content, the local amount of unoxidized fuel reaching the water wall, the local gas phase stoichiometry, and the local temperature. A third proprietary correlation for the prediction of water wall corrosion arising from gas-phase chlorine attack is used. The corrosion rate for this parallel mechanism is a function of coal chlorine content, local heat flux to the deposit face, and the tube metal temperature beneath the deposit. This three mechanism approach has been validated at both pilot and full scale. SIMULATION DESCRIPTION The coal-fired boiler simulated was a 660 MW opposed-wall unit. Five cases were simulated involving the following fuel mixtures and target generation rate: • Baseline Bituminous coal (Baseline) • 60% Wood Pellet 1 / 40% Baseline coal (WP1) • 60% Wood Pellet 2 / 40% Baseline coal (WP2) • 60% Wood Pellet 3 / 40% Baseline coal (WP3) • 42% Wood Pellet 1 / 18% Straw Pellet 1/ 40% Baseline coal (WP1+SP1) The fuel properties of the five cases are summarized in Table 1. The fuel ash elemental analyses are shown in Table 2. Key fuel properties to be noted related to potential biomass fuel effects include: • Much higher volatile yield • Much lower moisture content • Much lower ash • Lower heating value • Lower nitrogen content • Lower sulfur content • Lower chlorine content 20 Kung, S. C., "Prediction of Corrosion Rate for Alloys Exposed to Reducing/Sulfidizing Combustion Gases," Material Performance, 36(12), 36-40, (1997). 21 Kung, S. C., "Corrosion Studies for Low NOx Burner Technology," EPRI Report TR-108750, 1997.AFRC 2013 Industrial Combustion Symposium Ash Deposit Modeling, Adams, et al. 8 • Much larger particle size distribution • Much higher potassium content in ash • Much lower iron content in ash Table 1. Fuel Properties Table 2. Fuel Ash Elemental Analyses Firing large amounts of biomass also required some modification to operating conditions. Whereas the baseline coal-fired operation had two burner rows out of service, the biomass fuel was distributed amongst all burners. SIMULATION RESULTS The baseline simulation was discussed in detail with plant personnel. Results were considered to be reflective of the actual situation. With a satisfactory baseline simulation to serve as basis for comparison, the biomass simulations were then performed. Table 3 summarizes the overall results from the five cases simulated. These overall results showed that co-firing biomass produced:AFRC 2013 Industrial Combustion Symposium Ash Deposit Modeling, Adams, et al. 9 • A modest increase in FEGT due to a more distributed heat release throughout the furnace (coal baseline had two of the upper burner rows out of service, whereas the biomass was more evenly distributed throughout the burner zone). • Some reduction in wall heat transfer due to lower flame temperatures and incident heat flux in the burner zone. • 35-40% decrease in NOx emissions, primarily due to the reduced nitrogen in the biomass. • Similar CO emissions. • Slight decrease in carbon in fly ash due to less fixed carbon. • An increase in deposited ash (reduced fly ash existing the furnace) due to the increased impaction of biomass particles on furnace walls due to their much larger size. Table 3. Summary of Predicted Model Results for Simulated Cases Figure 1 shows the predicted boiler gas temperatures after four hours of operation for each of the five cases simulated. Temperature profiles are similar, but show lower temperatures in the lower furnace for the biomass cases. Figure 1. Predicted gas temperatures after four hours of operation. Model Exit (after superheat)Revised BaselineWP1WP2WP3WP1+SP1Gas Temperature, oC 871917899928913CO Concentration, ppm dry228143202362232O2 Concentration, vol. % dry2.822.572.642.572.54NOx Concentration, ppm dry245152164167161NOx emission (lb-NO2/MBtu)0.340.200.220.220.21Unburned Carbon in Fly ash, %6.53.54.84.43.4Percent Ash Exiting (as fly ash)89.278.479.467.670.4AFRC 2013 Industrial Combustion Symposium Ash Deposit Modeling, Adams, et al. 10 Figure 2 shows the CO concentration profiles for the simulated cases. CO concentrations in the biomass cases are higher in the lower furnace due to the reduced amount of oxygen fired in this region, creating a lower stoichiometry. Most of the CO has been oxidized by the furnace nose, and at the furnace exit the integrated CO concentration is similar for the cases. Figure 2. Predicted CO concentrations after four hours of operation. Figure 3 shows the NOx concentration profiles for the baseline and biomass co-firing cases. NOx concentrations are much lower throughout the boiler for the co-firing cases because of the much lower nitrogen content in the biomass and straw fuels. The lower stoichiometry in the lower furnace for the biomass cases also increases NOx reburning in this zone. Figure 3. Predicted NOx concentrations after four hours of operation.AFRC 2013 Industrial Combustion Symposium Ash Deposit Modeling, Adams, et al. 11 Figure 4 shows the predicted boiler rear wall deposit thickness after four hours of operation for each of the five cases simulated. Figure 4. Predicted deposit thickness on boiler rear and left wall after four hours of operation. The deposition patterns for the coal-only simulation indicate that areas of peak deposition are on the front and rear walls where burners are unopposed. In addition, deposition is observed on the upper side of the arch/nose. Model results indicate that WP1 will result in the least deposition amongst all of the biomass co-firing cases. The rear wall deposit thickness after four hours of build-up is slightly reduced in comparison to the baseline while the front wall thickness is slightly increased. Deposit thickness patterns are slightly different for WP2, with a slight increase in the area over which noticeable deposition occurs; but overall the predicted deposition is similar to that of WP1 and Baseline. The deposition patterns for WP3 are qualitatively very similar to those of WP2, however the thickness is somewhat increased. The WP1&SP1 co-firing scenario results in the most deposit build-up after four hours. Sintering extent can be used as an indicator of deposit removal during soot blowing. Figure 5 and Figure 6 show the predicted front wall and rear wall, respectively, boiler deposit sintering after four hours of operation for each of the five cases simulated. WP1 and WP2 simulations indicate that the deposits in areas of high build-up are somewhat easier to remove; however, WP1 results indicate that sidewall deposits may sinter extensively if allowed to accumulate. Predictions of sintering extent for WP3 are somewhat mixed, while those for WP1&SP1 indicate that extensive sintering may occur in the same regions where build-up occurs - again indicating that deposition difficulties may be a problem for this mixture of fuels.AFRC 2013 Industrial Combustion Symposium Ash Deposit Modeling, Adams, et al. 12 Figure 5. Predicted deposit sintering on boiler front and right wall after four hours of operation. Figure 6. Predicted deposit sintering on boiler rear and left wall after four hours of operation. Figure 7 shows the net heat flux profiles for the furnace rear and left side walls after four hours of furnace operation. Regions with high particle deposits (see Figure 4) and particularly high sintered deposits (see Figure 6) show reduced heat transfer. This is caused by the increased thermal resistance created by the deposits. Figure 8 shows the wall or deposit surface temperature for the furnace rear and left side walls after four hours of furnace operation. Not surprisingly, the regions with highest levels of deposits (burner zone, nose) also show the highest surface temperatures. These temperatures are the surface temperatures of the deposits, not the water wall surface temperatures. These high temperatures result from the deposits limited ability to conduct heat from the surface to the underlying water wall.AFRC 2013 Industrial Combustion Symposium Ash Deposit Modeling, Adams, et al. 13 Figure 7. Predicted net heat flux on rear and left side wall after four hours of operation. Figure 8. Predicted wall or deposit surface temperatures on boiler rear and left walls after four hours of operation. Figure 9 and Figure 10 show the predicted corrosion rates in nano-meters per hour (nm/hr) for the simulated cases. One nm/hr is approximately 3 mils/yr. Corrosion rates for the coal-only baseline indicated broad areas of concern and peak tube wastage rates on sections of the sidewalls, division panel, and rear wall where corrosion rates exceed 50 nm/hr. When biomass is used to replace 60% of the coal, the corrosion rates are reduced significantly for all cases. The primary reason for this is the reduction in fuel sulfur content. A mechanistic analysis of the baseline corrosion patterns indicates that the most important mechanism involves the deposition of unburned, sulfur rich coal. The impact of this AFRC 2013 Industrial Combustion Symposium Ash Deposit Modeling, Adams, et al. 14 mechanism is obviously reduced by the reduction in fuel sulfur. However, other contributing factors also seem to be positively affected as well, including a moderation of near wall reducing conditions and a slight reduction in peak heat transfer. Figure 9. Predicted corrosion rate in nano-meters/hr for boiler front and right walls after four hours of operation. Figure 10. Predicted corrosion rate in nano-meters/hr for boiler front and right walls after four hours of operation.AFRC 2013 Industrial Combustion Symposium Ash Deposit Modeling, Adams, et al. 15 SUMMARY Proposed biomass co-firing implementations using four different biomass fuels/blends have been evaluated using a unique CFD-based analysis. The CFD analysis incorporated advanced deposit and sintering models to predict impacts of ash behavior after four hours of operation in a 660 MW opposed-wall fired furnace. Based on this analysis, the following observations can be made regarding biomass co-firing: • 60% WP1 co-firing results in deposition behavior similar to that of coal alone; sidewall deposition is very limited, but material that is deposited is prone to sintering; corrosion rates are significantly reduced in comparison to the baseline. • 60% WP2 co-firing results in deposition and deposit sintering behavior similar to that of coal alone; corrosion rates are significantly reduced in comparison to the baseline. • 60% WP3 co-firing results in deposition behavior slightly worse than that of coal alone; corrosion rates are significantly reduced in comparison to the baseline. • 42% WP1 / 18% Straw Pellet co-firing results in deposition and deposit sintering behavior noticeably worse than that of coal alone; corrosion rates are significantly reduced in comparison to the baseline. Overall, the simulation results predicted that co-firing with biomass would produce: • A modest increase in FEGT due to a more distributed heat release throughout the furnace. • Some reduction in wall heat transfer. • 35-40% decrease in NOx emissions, primarily due to the reduced nitrogen in the biomass. • Similar CO emissions. • Slight decrease in carbon in fly ash due to less fixed carbon. • An increase in deposited ash (reduced fly ash existing the furnace) due to the increased impaction of biomass particles on furnace walls due to their much larger size. ACKNOWLEDGEMENTS CFD graphics were produced with the Fieldview software from Intelligent Light (http://www.ilight.com)