Title | Large Eddy Simulation of a 15MW Tangentially Oxy-Fired Pulverized Coal Boiler: Ignitor |
Creator | Pedel, Julien; Wu, Yuxin; Schmidt, John; Thornock, Jeremy; Smith, Philip J. |
Publication type | report |
Date | 2012-09-06 |
Abstract | Oxy-coal combustion, in which an O2/CO2 mixture replaces air, is one of the few possible capture technologies to enable CO2 sequestration for existing coal-fired boilers. Burning coal with relatively pure oxygen, together with recycled flue gases, can produce a highly concentrated (up to 95% CO2) flue gas stream, which makes carbon sequestration more economical. One issue of interest in rapidly implementing a strategy to retrofit existing air-fired units is to understand how replacing air by a O2/CO2 mixture changes the kinetics, aerodynamics and heat transfer of the flame. This study presents Large Eddy Simulations (LES) of a commercial scale pulverized coal boiler in oxycombustion conditions. Simulations were performed using Arches, a massively parallel LES tool developed at the University of Utah on Kraken, one of the fastest academic computer in the world on 10,000+ processors. The simulation was applied to a 15 MW boiler with tangential fired burners located in the corners under oxy-combustion conditions with an overall PO2 = 40%. Particles are tracked in an Eulerian framework with the Direct Quadrature Method of Moments (DQMOM). Coal devolatilization and char combustion are modeled and the Discrete Ordinates method is used to solve radiation. Results show that LES coupled with DQMOM has the potential to predict oxy-coal flame characteristics and to be an important tool in the retrofitting or design process of oxy-coal burners and boilers. |
Type | Text |
Format | application/pdf |
Rights | This material may be protected by copyright. Permission required for use in any form. For further information please contact the American Flame Research Committee. |
OCR Text | Show Large Eddy Simulation of a 15MW Tangentially Oxy-Fired Pulverized Coal Boiler: Ignitor Julien Pedel, Yuxin Wu, John Schmidt,Jeremy Thornock, Philip J. Smith the Institute for Clean and Secure Energy, University of Utah, Salt Lake City, 84108, U.S.A Abstract Oxy-coal combustion, in which an O2/CO2 mixture replaces air, is one of the few possible capture technologies to enable CO2 sequestration for existing coal-fired boilers. Burning coal with relatively pure oxygen, together with recycled flue gases, can produce a highly concentrated (up to 95% CO2) flue gas stream, which makes carbon sequestration more economical. One issue of interest in rapidly implementing a strategy to retrofit existing air-fired units is to understand how replacing air by a O2/CO2 mixture changes the kinetics, aerodynamics and heat transfer of the flame. This study presents Large Eddy Simulations (LES) of a commercial scale pulverized coal boiler in oxy-combustion conditions. Simulations were performed using Arches, a massively parallel LES tool developed at the University of Utah on Kraken, one of the fastest academic computer in the world on 10,000+ processors. The simulation was applied to a 15 MW boiler with tangential fired burners located in the corners under oxy-combustion conditions with an overall PO2 = 40%. Particles are tracked in an Eulerian framework with the Direct Quadrature Method of Moments (DQMOM). Coal devolatilization and char combustion are modeled and the Discrete Ordinates method is used to solve radiation. Results show that LES coupled with DQMOM has the potential to predict oxy-coal flame characteristics and to be an important tool in the retrofitting or design process of oxy-coal burners and boilers. Keywords: Oxy coal combustion, coal ignition, LES, DQMOM 1. Introduction Coal is one of the major energy resources due to its abundant reserves and low price. Since coal has a high carbon content, CO2 capture in coal combustion has been concerned worldwide. Among all CO2 capture technologies, Oxy-Fired pulverized coal boiler is regarded as the most appealing technology to retrofit the existing coal power plants. In this technology, pure oxygen is first separated from air and then mixed with the recycling flue gas to serve as the oxidizer for the coal boiler. Since there is no N2 introduced from the oxidizer into the furnace, the CO2 concentration in the flue gas is high enough to be sequenced in an economic way. Due to the variation of physical/chemical properties between CO2 and N2, coal combustion in an Oxy-Fired pulverized boiler performs different characteristics. One chanllenge is the delay of coal ignition and unstable flames occurs easily in an Oxy-Fired boilerChen et al. [1], Molina and Shaddix [2], which requires new burner design to ensure that the similar flame and heat flux distribution can be acquired in an Oxy-fired boiler. As the development of CFD technologies as well as computing capacities, it's possible to simulate coal combustion flames inside a real boiler with Large Eddy Simulation (LES) method. Since LES calculates the major turbulence that dorminates the flow field without any additional model, it has been proved to provide more accurate results for oxy-coal combustion progressChen and Ghoniem [3]. Especially, LES can capture the instantaneous flames, which make it an appealing method to predict the heat flux distribution inside the furnace. In this article, a multi-phase model were Preprint submitted to Elsevier August 17, 2012 proposed to simulation coal combustion with LES method. Direct Quadrature Method of Moments (DQMOM) model was adopted to calculate dispersion and combustion of coal particles. Based on LES code Arches, ignation of the burner flames were simulated for a 15MW tangentially Oxy-Fired pulverized coal boiler. 2. Model description 2.1. LES method LES directly solves the transport equations for the large eddies and models only the smallest eddies. Unsteady turbulent motions are evaluated and the number of model parameters is greatly reduced. In this work, simulations were performed using ARCHES, a LES tool resulting from a ten-year partnership with the Department of Energy and the University of UtahSpinti et al. [4]. It is a massively parallel code that solves conserved quantities (mass, momentum, energy and scalar) spatially and temporally in a turbulent flow field, allowing for detailed and accurate simulations of fires and flames. the gas-phase governing equations are solved using filtered continuity and momentum equations for incompressible flows in a finite volume formulation. The transport equations including mass balance equation @ug¯,i @xi = ¯ Sm momentum equation @ug¯,i @t + @ @xj (¯ug,i¯ug,j) = 1 ⇢g @ ¯p @xi + 2⌫ @ @xj ¯ Sij @⌧ij @xj + ¯ Sui + g mixture fraction equation @⇢ ¯ fi @t + @ @xj ! ¯ug,i⇢ ¯ fk " = @ @xi ⇣ ⇢D @ @xj ¯ fk ⌘ + ¯ Sfk and Enthalpy equation @¯h @t + @ @xj ! ¯ug,i¯h i " = @ @xi ⇣ $ ⇢cp @ @xj ¯h i ⌘ @ @xj q + S¯h where ¯ Sm, ¯ Sui, ¯ Sf and ¯ Sh are the mass, momentum, mixture fraction and enthalpy source terms contributed by coal particles. q is the radiative flux, which is calculated with the Discrete Ordi-nate method. The ¯# represent the anisotropic box filter of a scalar #, where the scalar variable is simply Favre averaged over the intervals of a cell. In order to solve the stress tesinons in momen-tum equation, the Dynamic coefficient Smagorinsky subgrid scale model was adopted to close the momentum equation. Two mixture fractions were adopted to represent the mixture scalars of the reacting gas phase, which are f1 = mp mt and f⌘ = mcoal mt The meaning of f1 and f⌘ are mass fraction of the elements coming from the primary air and from the coal separately. The radiation heat transfer is very importnant for good simulation, and Discrete Ordinate model was adopted to calculate the readiation heat flux q. The set of filtered equations are discretized in space and time and solved on a staggered, finite volume mesh. The staggering scheme consists of four offset grids. One grid stores the scalar quantities and the remaining three grids store each component of the velocity vector. 2.2. DQMOM Model The DQMOM method was implemented into ARCHES to solve the dispersed phase in an Eule-rian manner in this simulation. In DQMOM model, the quadrature weights and weighted abscissas were tracked to represent the the transport equation of the Number Density Function (NDF) of the coal particles, which is represented by @f(⇠;x,t) @t + @ @xi (hui|⇠; x, ti f (⇠; x, t)) + @ @⇠j (hGj |⇠; x, ti f (⇠; x, t)) = h (⇠; x, t) 2 where f (⇠; x, t) is the particle numbers in a unit volume. And these particles can be clasified by their position and typical physical properties (called internal coordinates, ⇠) such as particle diameter, velocity and components. Gj is he velocity of the NDF in internal coordinates, and is defined by Gj = d⇠j dt the birth/death term h (⇠; x, t) represents the generation or destory of the coal particles within the domain. Considering that f (⇠; x, t) is represent by a set of delta functions of internal coordinates, or f (⇠; x, t) t Pn ↵=1(w↵ (x, t) % (⇠ h⇠ (x, t)i↵)) The NDF transport equation can finally be transformed into a set of transport equations @w↵ @t + @ @xi (huii↵ w↵) @ @xi ⇣ Dx,↵ @ @xi w↵ ⌘ = a↵ @⇣n,↵ @t + @ @xi (huii↵ ⇣n,↵) @ @xi ⇣ Dx,↵ @ @xi ⇣n,↵ ⌘ = bn,↵ where w↵ is the weights of an internal coordinate variable ⇠↵, and ⇣n,↵ = w↵ h⇠ni↵is the weighted abscissa for the nth internal coordinate and huii↵ is the average internal coordinate velocity. The right hand side terms a↵ and bn,↵ are obtained by solving a linear system of equations. In this simulation, a total of seven internal coordinates were chosen to represent particle char-acteristics in the DQMOM formulation: mass of raw coal ↵c, mass of char ↵h, particle diameter dp, particle enthalpty hp and three particle velocity components ux,uy and uz. 2.3. Modeling of Coal particles 2.3.1. Particle dispersion The aggregation/breakage phenomena are neglected and the only physical process modeled is the drag force, which can be described by the Stokes drag law. The momentum equation for the particle is dup,i dt = f ⌧p (ug,i up,i) + gi(⇢p⇢g) ⇢p where g is the gravity force acting on the particle, ug and up are the gas and particle velocities, ⌧p is the particle relaxation time, and ⌧ p = ⇢pdp/ (18μg). Given the particle Reynolds number, Rep = ⇢pdp |up ug| /μg, and the drag force coefficient f can be calculated by f = 8>< >: 1 Rep < 1 1 + 0.15Re0.687 p 1 < Rep < 1000 0.0183Rep Rep > 100 2.3.2. Particle Reactions and heat transfer Coal reactions were divided into two steps, the coal devolatilization step and char reaction step. In this study, a modified one-step devolatilization model proposed by YamamotoYamamoto et al. [5] was used for simplification. Only one char reaction with oxygen was considered in the simulation. And the semi-empirical nth order Arrhenius model of char combustion was adopted to predicte the char reaction rateMurphy and Shaddix [6]. The particle is heated by convection, radiation, and reaction enthalpy changes: dhp dt = Qrad + Qconv + rjhj whereQrad and Qconv represents energy transfer due to radiation, convection and conduction between the gas phase and the particle, and the third therm represents both the amount of energy lost by the particles due to lost mass, and the enthalpy released during devolatilization and oxidation reactions. 3. Results and Discussion 3.1. Simulation cases The simulation case is for the 15MWth tangentially fired Boiler Simulation Facility (BSF) conducted by AlstomKluger et al. [7]. The structures of the BSF are shown in Fig.1. 3 (a) Front View (b) top view Figure 1: Alstom BSF structures Lusatian dried lignite is adopted in the test and the coal analysis are listed as Table.1 Table 1: Analyses of Lusatian dried lignite in the BSF Proximate analyses VM FC Ash Moisture HHV, Btu/lb wt% 46.97 34.11 5.22 13.70 9008 Ultimate analyses C H O N S wt% 53.96 3.9 22.01 0.6 0.62 The test matrix for BSF Campaign 6 included 16 air-fired and 21 oxy-fired tests. Selected tests points were run for extended duration to match previous conditions from both pilots with respect to recycle rate/global oxygen concentration, recycle take-off location/recycle treatment, thermal conditions/furnace outlet, fuel/oxygen stoichiometry, and oxygen in the flue gas at the economizer. Test conditions during these tests are summarized in Table2. 4 Table 2: Summary of Lusatian lignite test conditions Test# Air/Oxy Firing Rate(MMBTU/hr) Staging Level Global O2(Vol%) Outlet O2(Vol%) Carbon in Ash, % HFOT Temp 103 Air 40 low 20.6 2.4 0.17 2069 109 Air 45 Moderate 20.7 2.9 0.18 2088 127 Oxy 40 Low 22.5 2.8 0.29 2093 134 Oxy 45 Low 27.0 2.7 0.18 2264 Test134 were selected as the simulation case since that most information can be found in Kluger et al. [8] and Kluger et al. [7]. The simulation domain is just limited in the furnace, the resolution is 2cm and the total cubical mesh number is 16,126,320 so that burner inlet boundary conditions are same as the real case. The simulation took1536 processors 120 hours to acquire the first 8s start-up case for the BSF. Based on this simulation, the ignation of the BSF were simulated and the coal flames were acquired. 3.2. coal ignition inside the oxy-fired boiler The real coal ignition inside BSF and the structure of the burners are shown in Fig2Kluger et al. [7]. Since it's a tangential-fired boiler, the flames from each corner meet at the central furnace and forms a tangential flame circle. The advantages of such design is that each flames from a corner is supported and strengthen by the flames from the other three corners, so that good combustion performance can be achieved under a variaty of coal types. Fig.3 shows the predicted temperature profile inside the BSF furnace when the coal combustion flames are fully developed. Comparing Fig.2 and Fig.3, it shows that LES simulation captures the characteristics of the tangential fire. (a) burner view (b) burner structures Figure 2: Alstom BSF tangential burner 5 (a) temperature profile of the coal burners (b) secondary air and OFA Figure 3: Temperature profile inside the furnace Fig.4 shows a snapshot of the flow field during the development of the burner injections. The entrainment of the jet and the dissipation of the injection are clearly captured by LES simulations. Figure 4: Velocity flow field In order to study the coal ignation development, Fig.5 shows the snap shots of the O2 concentra-tion at different time steps. At very beginning, oxygen are injected into the furnace and tangential flow was not formed before 0.3s. But the injector develops very fast and in 2 seconds, the flow field 6 are almost developed. At the primary and secondary flow height, the oxygen are consumed quickly and the O2 concentration at the central furnace is very low. This is a reaonable result and it also shows the importance of the OFA burners. Since OFA provide additional oxygen, the unburned coal particles contiune reacting with O2 and thus improve the combustion efficiency while decreaing the NOx formation. (a) t=0.1s (b) t=0.3s (c) t=2s (d) t=5s Figure 5: Development of the coal flame represented by O2 concentration Fig.6 shows the O2 concentration inside the furnace when the combustion flow is developed. Comparing Fig.6a and Fig.6b, it shows that O2 concentration profile varies at differen directions. At the corners, O2 concentration is high while at the same time, O2 concentration at the middle furnace width and depth is very low due to the limited diffusion and quickly O2 consumption. 7 (a) overall O2 concentration (b) O2 cross-section Figure 6: O2 concentration inside the furnace 4. Conclusion Oxy-fired coal boiler is an appealing clean coal techology for CO2 sequence. This study presents Large Eddy Simulations (LES) of a pilot scale pulverized coal boiler in oxy- combustion conditions. Simulations were performed using Arches, a massively parallel LES tool developed by ICSE sim-ulation group in the University of Utah. The simulation was applied to the Alstom 15 MW BSF under oxy-combustion conditions with an overall PO2 = 40%. Particles are tracked in an Eulerian framework with the Direct Quadrature Method of Moments (DQMOM). Coal devolatilization and char combustion are modeled and the Discrete Ordinates method is used to solve radiation. Results show that LES coupled with DQMOM has the potential to predict oxy-coal flame characteristics and to be an important tool in the retrofitting or design process of oxy-coal burners and boilers. Acknowledgement: The material in this article is based upon work supported by the Depart-ment of Energy under Award Number DE-NT0005015. [1] Lei Chen, Sze Zheng Yong, and Ahmed F. Ghoniem. Oxy-fuel combustion of pulverized coal: Characterization, fundamentals, stabilization and cfd modeling. Progress in Energy and Com-bustion Science, 38:156-214, 2012. [2] Alejandro Molina and Christopher R. Shaddix. Ignition and devolatilization of pulverized bitu-minous coal particles during oxygen/carbon dioxide coal combustion. Proceedings of the Com-bustion Institute, 31:1905-1912, 2007. [3] Lei Chen and Ahmed F. Ghoniem. Simulation of oxy-coal combustion in a 100 kwth test facility using rans and les: A validation study. Energy and Fuels, inprocess, 2012. [4] J. Spinti, J. Thornock, E.Eddings, P. Smith, and A. Sarofim. Heat transfer to objects in pool fires. WIT Press, Southampton, UK, 2008. [5] K. Yamamoto, T. Murota, T. Okazaki, and M. Taniguchi. Large eddy simulation of a pulverized coal jet flame ignited by a preheated gas flow. Proceedings of the Combustion Institute, 33(2): 1771-1778, 2011. 8 [6] J.J. Murphy and C.R. Shaddix. Combustion kinetics of coal chars in oxygen enriched environ-ments. Combustion and Flame, 144(4):710-729, 2006. [7] Frank Kluger, Thomas Wild, and Patric Monckert. 15mwth and 30mwth oxy-combustion pilot testing of lignite from vattenfalll's lusatia open cast mine. In 37th International Technical Conference on Clean Coal and Fuel Systems, volume 37, 2012. [8] Frank Kluger, Patric Moenckert, and Georg-Nikolaus Stamatelopoulos. Alstom's oxy-combustion technology development-update on pilot plants operation. In 35th International Technical Conference on Clean Coal and Fuel Systems, volume 35, clear water, Florida, USA, June 2010. 9 |
ARK | ark:/87278/s6x069pv |
Setname | uu_afrc |
ID | 14338 |
Reference URL | https://collections.lib.utah.edu/ark:/87278/s6x069pv |