
[Sponsors] 
March 25, 2002, 15:12 
Combustion Convergence problems

#1 
Guest
Posts: n/a

I'm doing the following simulation and would like to hear from others who have experience doing similar work:
6 m by 2 m by 1.5 m furnace with refractory lined walls considered adiabatic 0.5 m dia. circular exhaust port in center of roof, 50 Pa pressure outlet 3 burners  0.05 m dia. natural gas (methane) ports, 20 m/s velocity inlet  0.15 m dia. annulus air ports, 30 m/s velocity inlet  locations two at one end on wall, single burner in roof for other half load  bottom surface represents molten metal bath, 1000C const. temperature tramp air  0.2 by 0.7 vent where molten metal enters in real life, 2 m/s inlet The issue is nonuniformity of the temperature distribution in the gas phase above the molten metal. The modelling is focussed on the flames and exhaust gases. The issue I'm having is that I've been having trouble converging the solution for the cold gases without combustion below 1E03 average total residual and experiencing very slow convergence with combustion because I have to set relaxations very low to avoild crashing the simulation. Here are some details on the grid and relaxations I'm using: largest cell dimension 0.0635 m in parts of the domain away from the flames smallest cell dimension 0.0127 m in the vicinity of the burner ports total cell count 144 by 94 by 55 = 744,480 cells expected max velocity in flames about 220 m/s solved for variables: P1, U1/V1/W1, EP/KE, H1, MIX1 models: KEMODL default for turbulence, 3_gases_mixing for combustion, buoyancy 3 gases density, old property table viscosity, 3 gases mixing temperature library reference case: C113 relaxation strategy: clamp down on P1 to at least 1.0E04, even to 1.0E06 for first 1000 iterations, up by factor of 10 each 1000 clamp down on U1/V1/W1 to at least 1.0E05, even to 1.0E06 for first 1000 iterations, up by factor of 2 each 1000 clamp down on EP/KE to at least 1.0E03, even to 1.0E04 for first 1000 iterations, up to 0.1, then 0.5 thereafter set H1/MIX1 to at least 1.0E02, even to 1.0E04 for first 1000 iterations, up by factor of 100 each 1000 model feedback: in first 1000 iterations (critical points at 30, 300, 600) velocity field will degrade to random E+05 values if relaxations for P1/U1/V1/W1 are at 2.0E04 or above (except when I've tried MIGAL for faster convergence on large grid and found I could use 1.0E03 and even higher) nonuniformity between the three flames in size and peak velocities variations from 15% to 50% for identical inlet conditions The best I've done so far is to use 1E06 for first 1000 iterations to overcome the nonuniformity issues between the three flames, only reaching E+07 on residual error and E+03 on mass balance. The people with experience I talk to expect me to reach typical total error residuals of 1E+06 in under 500 iterations and mass balance to 1E01 within 1000 iterations. It has been taking me on the order of 3000 iterations (20 hours on a P4 1.6 GHz with 1 GB RDRAM 3GB SCSIUW swap file) to reach those goals. I will attach the Q1 file in anticipation that I haven't provided enough detail. Thanks in advance for any advice that can be offered on my simulation experience. Art Stretton Houston, Texas P.S. I made up three new dat files which I can email on request RIDGE8.DAT  like a staircase with 50% first step, 25% second step, 25% last step RIDGE8R.DAT  for fitting two RIDGE8.DAT objects together in a corner BABET4.DAT  like BABET.DAT but only a quarter domain with a 30 degree truncation P.P.S. sorry my cut and paste of Q1 got formatted poorly, my attempts to edit combined lines didn't work  8< CombE3b.q1 follows 8<  TALK=T;RUN( 1, 1) ************************************************** ********** Q1 created by VDI menu, Version 3.4, Date 10/08/01 CPVNAM=VDI;SPPNAM=Core ************************************************** ********** Echo DISPLAY / USE settings DISPLAY The case considered is the current roof mounted burner with combustion without the burner chambers. ENDDIS ************************************************** ********** IRUNN = 1 ;LIBREF = 0 ************************************************** ********** Group 1. Run Title TEXT(COMINCO CDF Combustion E ) ************************************************** ********** Group 2. Transience STEADY = T ************************************************** ********** Groups 3, 4, 5 Grid Information * Overall number of cells, RSET(M,NX,NY,NZ,tolerance) RSET(M,144,94,55) * Set overall domain extent: * xulast yvlast zwlast name XSI= 6.188000E+00; YSI= 2.088000E+00; ZSI= 1.380000E+00 RSET(D,CHAM ) ************************************************** ********** Group 6. BodyFitted coordinates ************************************************** ********** Group 7. Variables: STOREd,SOLVEd,NAMEd ONEPHS = T * Nondefault variable names NAME(142) =PRPS ; NAME(143) =EPKE NAME(144) =ENUT ; NAME(145) =FUEL NAME(146) =PROD ; NAME(147) =OXID NAME(148) =RHO1 ; NAME(149) =TMP1 NAME(150) =MIXF * Solved variables list SOLVE(P1 ,U1 ,V1 ,W1 ,H1 ,MIXF) * Stored variables list STORE(TMP1,RHO1,OXID,PROD,FUEL,ENUT,EPKE,PRPS) * Additional solver options SOLUTN(P1 ,Y,Y,Y,N,N,Y) TURMOD(KEMODL) ************************************************** ********** Group 8. Terms & Devices TERMS (H1 ,N,Y,Y,Y,Y,Y) NEWRH1 = T NEWENL = T ISOLX = 0 ;ISOLY = 0 ;ISOLZ = 0 DENPCO = T ************************************************** ********** Group 9. Properties SETPRPS(1, 30) RHO1 = GRND6 RHO1A = 1.604303E+01 ;RHO1B = 2.894200E+01 RHO1C = 2.918912E+01 PRESS0 = 1.013250E+05 TMP1 = GRND7 TMP1A = 1.500000E+03 ;TMP1B = 1.500000E+03 TMP1C = 1.500000E+03 TMP2A = 5.482456E02 ;TMP2B = 4.943350E+07 TMP2C = 0.000000E+00 TEMP0 = 2.731500E+02 CP1 = GRND10 ENUL = GRND10 ENULA = 1.460000E06 ;ENULB = 1.100000E+02 ENULC = 0.000000E+00 DRH1DP = 3.410000E03 DVO1DT = 3.330000E03 PRNDTL(H1 ) = 2.630000E02 PRT (EP ) = 1.314000E+00 ************************************************** ********** Group 10.InterPhase Transfer Processes ************************************************** ********** Group 11.Initialise Var/Porosity Fields RESTRT(ALL) No PATCHes used for this Group INIADD = F ************************************************** ********** Group 12. Convection and diffusion adjustments No PATCHes used for this Group ************************************************** ********** Group 13. Boundary & Special Sources PATCH (CHSO ,VOLUME,0,0,0,0,0,0,1,1) PATCH (BUOYANCY,PHASEM,0,0,0,0,0,0,1,1) COVAL (BUOYANCY,U1 , FIXFLU , GRND2 ) COVAL (BUOYANCY,V1 , FIXFLU , GRND2 ) COVAL (BUOYANCY,W1 , FIXFLU , GRND2 ) BUOYA = 0.000000E+00 ; BUOYB = 0.000000E+00 BUOYC =9.806000E+00 BUOYD = 1.184000E+00 ************************************************** ********** Group 14. Downstream Pressure For PARAB ************************************************** ********** Group 15. Terminate Sweeps LSWEEP = 1000 RESFAC = 1.000000E03 ************************************************** ********** Group 16. Terminate Iterations LITER (U1 ) = 20 ;LITER (V1 ) = 20 LITER (W1 ) = 20 ENDIT (P1 ) = 1.000000E01 ;ENDIT (U1 ) = 1.000000E01 ENDIT (V1 ) = 1.000000E01 ;ENDIT (W1 ) = 1.000000E01 ENDIT (KE ) = 1.000000E01 ;ENDIT (EP ) = 1.000000E01 ENDIT (H1 ) = 1.000000E01 ;ENDIT (PRPS) = 1.000000E01 ENDIT (EPKE) = 1.000000E01 ;ENDIT (ENUT) = 1.000000E01 ENDIT (FUEL) = 1.000000E01 ;ENDIT (PROD) = 1.000000E01 ENDIT (OXID) = 1.000000E01 ;ENDIT (RHO1) = 1.000000E01 ENDIT (TMP1) = 1.000000E01 ;ENDIT (MIXF) = 1.000000E01 ************************************************** ********** Group 17. Relaxation RELAX(P1 ,LINRLX, 1.000000E02) RELAX(U1 ,FALSDT, 2.000000E05) RELAX(V1 ,FALSDT, 2.000000E05) RELAX(W1 ,FALSDT, 2.000000E05) RELAX(KE ,LINRLX, 1.000000E02) RELAX(EP ,LINRLX, 1.000000E02) RELAX(H1 ,FALSDT, 1.000000E+02) RELAX(MIXF,FALSDT, 1.000000E+02) KELIN = 3 ************************************************** ********** Group 18. Limits VARMAX(P1 ) = 1.000000E+20 ;VARMIN(P1 ) =1.000000E+05 VARMAX(U1 ) = 1.000000E+06 ;VARMIN(U1 ) =1.000000E+06 VARMAX(V1 ) = 1.000000E+06 ;VARMIN(V1 ) =1.000000E+06 VARMAX(W1 ) = 1.000000E+06 ;VARMIN(W1 ) =1.000000E+06 VARMAX(H1 ) = 1.000000E+10 ;VARMIN(H1 ) = 1.000000E10 VARMAX(EPKE) = 1.000000E+10 ;VARMIN(EPKE) = 1.000000E10 VARMAX(ENUT) = 1.000000E+10 ;VARMIN(ENUT) = 1.000000E10 VARMAX(FUEL) = 1.000000E+00 ;VARMIN(FUEL) = 0.000000E+00 VARMAX(PROD) = 1.000000E+00 ;VARMIN(PROD) = 0.000000E+00 VARMAX(OXID) = 1.000000E+00 ;VARMIN(OXID) = 0.000000E+00 VARMAX(RHO1) = 5.000000E+02 ;VARMIN(RHO1) = 1.000000E03 VARMAX(TMP1) = 1.000000E+10 ;VARMIN(TMP1) = 1.000000E10 VARMAX(MIXF) = 1.000000E+00 ;VARMIN(MIXF) = 0.000000E+00 ************************************************** ********** Group 19. EARTH Calls To GROUND Station USEGRD = T ;USEGRX = T GENK = T ASAP = T RADIA = 0.000000E+00 ;RADIB = 0.000000E+00 EMISS = 0.000000E+00 SPEDAT(SET,MIGAL,SOLVED01,C,HYDRO) SPEDAT(SET,MIGAL,SOLVED02,C,H1) SPEDAT(SET,MIGAL,SOLVED03,C,MIXF) SPEDAT(SET,MIGAL,REFAC01,R,0.00000E+00) SPEDAT(SET,MIGAL,REFAC02,R,0.00000E+00) SPEDAT(SET,MIGAL,REFAC03,R,0.00000E+00) SPEDAT(SET,MIGAL,LITER01,I,5) SPEDAT(SET,MIGAL,LITER02,I,5) SPEDAT(SET,MIGAL,LITER03,I,5) SPEDAT(SET,MIGAL,LINRLX01,R,5.00000E01) SPEDAT(SET,MIGAL,LINRLX02,R,1.00000E+00) SPEDAT(SET,MIGAL,LINRLX03,R,1.00000E+00) SPEDAT(SET,MIGAL,COEFF01,R,1.00000E+00) SPEDAT(SET,MIGAL,COEFF02,R,1.00000E+00) SPEDAT(SET,MIGAL,COEFF03,R,1.00000E+00) ************************************************** ********** Group 20. Preliminary Printout ECHO = T ************************************************** ********** Group 21. Printout of Variables OUTPUT(P1 ,Y,Y,Y,N,Y,Y) OUTPUT(TMP1,Y,Y,Y,N,Y,Y) ************************************************** ********** Group 22. Monitor PrintOut IXMON = 141 ;IYMON = 47 ;IZMON = 34 NPRMON = 1 NPRMNT = 1 TSTSWP = 1 ************************************************** ********** Group 23.Field PrintOut & Plot Control NPRINT = 100000 ISWPRF = 1 ;ISWPRL = 100000 No PATCHes used for this Group ************************************************** ********** Group 24. Dumps For Restarts NOWIPE = T GVIEW(P,0.000000E+00,1.000000E+00,0.000000E+00) GVIEW(UP,0.000000E+00,0.000000E+00,1.000000E+00) > DOM, SIZE, 6.188000E+00, 2.088000E+00, 1.380000E+00 : DOM, MONIT, 6.115190E+00, 1.050350E+00, 9.595750E01 : DOM, SCALE, 1.000000E+00, 1.000000E+00, 1.000000E+00 : DOM, SNAPSIZE, 1.000000E02 : GRID, RSET_X_1, 4, 1.400000E+00,G : GRID, RSET_X_2, 29,1.300000E+00 : GRID, RSET_X_3, 4, 1.000000E+00 : GRID, RSET_X_4, 4, 1.000000E+00 : GRID, RSET_X_5, 4, 1.000000E+00 : GRID, RSET_X_6, 30, 1.300000E+00 : GRID, RSET_X_7, 8,1.000000E+00 : GRID, RSET_X_8, 55,1.200000E+00 : GRID, RSET_X_9, 6,1.350000E+00 : GRID, RSET_Y_1, 4, 1.400000E+00,G : GRID, RSET_Y_2, 8,1.300000E+00 : GRID, RSET_Y_3, 4, 1.000000E+00 : GRID, RSET_Y_4, 4, 1.000000E+00 : GRID, RSET_Y_5, 4, 1.000000E+00 : GRID, RSET_Y_6, 5, 1.300000E+00 : GRID, RSET_Y_7, 5, 1.050000E+00 : GRID, RSET_Y_8, 6,1.300000E+00 : GRID, RSET_Y_9, 4, 1.000000E+00 : GRID, RSET_Y_10, 4, 1.000000E+00 : GRID, RSET_Y_11, 4, 1.000000E+00 : GRID, RSET_Y_12, 6, 1.300000E+00 : GRID, RSET_Y_13, 3, 1.000000E+00 : GRID, RSET_Y_14, 8,1.300000E+00 : GRID, RSET_Y_15, 4, 1.000000E+00 : GRID, RSET_Y_16, 4, 1.000000E+00 : GRID, RSET_Y_17, 4, 1.000000E+00 : GRID, RSET_Y_18, 8, 1.300000E+00 : GRID, RSET_Y_19, 5,1.500000E+00,G : GRID, RSET_Z_1, 9, 1.200000E+00 : GRID, RSET_Z_2, 19,1.250000E+00 : GRID, RSET_Z_3, 4, 1.000000E+00 : GRID, RSET_Z_4, 4, 1.000000E+00 : GRID, RSET_Z_5, 4, 1.000000E+00 : GRID, RSET_Z_6, 7, 1.200000E+00 : GRID, RSET_Z_7, 8,1.200000E+00 : DOM, RELAX, 5.000000E01 > OBJ1, NAME, WEST : OBJ1, POSITION, 0.000000E+00, 0.000000E+00, 0.000000E+00 : OBJ1, SIZE, 0.000000E+00, 2.088000E+00, 1.380000E+00 : OBJ1, CLIPART, cube : OBJ1, ROTATION24, 1 : OBJ1, VISIBLE, 1.000000E+00 : OBJ1, TYPE, PLATE : OBJ1, ADIABATIC, 0.000000E+00, 0.000000E+00 > OBJ2, NAME, EAST : OBJ2, POSITION, 6.188000E+00, 0.000000E+00, 0.000000E+00 : OBJ2, SIZE, 0.000000E+00, 2.088000E+00, 1.380000E+00 : OBJ2, CLIPART, cube : OBJ2, ROTATION24, 1 : OBJ2, VISIBLE, 1.000000E+00 : OBJ2, TYPE, PLATE : OBJ2, ADIABATIC, 0.000000E+00, 0.000000E+00 > OBJ3, NAME, NORTH : OBJ3, POSITION, 0.000000E+00, 0.000000E+00, 0.000000E+00 : OBJ3, SIZE, 6.188000E+00, 0.000000E+00, 1.380000E+00 : OBJ3, CLIPART, cube : OBJ3, ROTATION24, 1 : OBJ3, VISIBLE, 1.000000E+00 : OBJ3, TYPE, PLATE : OBJ3, ADIABATIC, 0.000000E+00, 0.000000E+00 > OBJ4, NAME, SOUTH : OBJ4, POSITION, 0.000000E+00, 2.088000E+00, 0.000000E+00 : OBJ4, SIZE, 6.188000E+00, 0.000000E+00, 1.380000E+00 : OBJ4, CLIPART, cube : OBJ4, ROTATION24, 1 : OBJ4, VISIBLE, 1.000000E+00 : OBJ4, TYPE, PLATE : OBJ4, ADIABATIC, 0.000000E+00, 0.000000E+00 > OBJ5, NAME, BATH : OBJ5, POSITION, 0.000000E+00, 0.000000E+00, 0.000000E+00 : OBJ5, SIZE, 6.188000E+00, 2.088000E+00, 0.000000E+00 : OBJ5, CLIPART, cube : OBJ5, ROTATION24, 1 : OBJ5, VISIBLE, 1.000000E+00 : OBJ5, TYPE, PLATE : OBJ5, SURF_TEMP, 0.000000E+00, 1.000000E+03 > OBJ6, NAME, EXHAUST : OBJ6, POSITION, 2.909000E+00, 8.190000E01, 1.380000E+00 : OBJ6, SIZE, 4.500000E01, 4.500000E01, 0.000000E+00 : OBJ6, CLIPART, cylinder : OBJ6, ROTATION24, 1 : OBJ6, TYPE, OUTLET : OBJ6, PRESSURE, 4.786300E+01 : OBJ6, TEMPERATURE, 2.500000E+01 : OBJ6, COEFFICIENT, 1.000000E+03 : OBJ6, VELOCITY, 1.026000E+04,1.026000E+04,1.026000E+04 : OBJ6, TURBULENCE, 1.026000E+04,1.026000E+04 > OBJ7, NAME, VENT3 : OBJ7, POSITION, 1.000000E03, 6.500000E01, 1.200000E+00 : OBJ7, SIZE, 0.000000E+00, 7.000000E01, 1.800000E01 : OBJ7, CLIPART, cube : OBJ7, ROTATION24, 2 : OBJ7, TYPE, INLET : OBJ7, DENSITY, 1.183778E+00 : OBJ7, VELOCITY, 2.292000E+00, 0.000000E+00, 0.000000E+00 : OBJ7, TEMPERATURE, 2.500000E+01 : OBJ7, TURBINTENS, 5.000000E+00 > OBJ8, NAME, WRIDGE : OBJ8, POSITION, 0.000000E+00, 1.500000E01, 0.000000E+00 : OBJ8, SIZE, 1.500000E01, 1.788000E+00, 3.000000E01 : OBJ8, CLIPART, ridge8 : OBJ8, ROTATION24, 5 : OBJ8, TYPE, BLOCKAGE : OBJ8, MATERIAL, 117 : OBJ8, ADIABATIC, 0.000000E+00, 0.000000E+00 : OBJ8, INI_TEMP, 2.000000E+01 > OBJ9, NAME, SRIDGE : OBJ9, POSITION, 1.500000E01, 0.000000E+00, 0.000000E+00 : OBJ9, SIZE, 5.888000E+00, 1.500000E01, 3.000000E01 : OBJ9, CLIPART, ridge8 : OBJ9, ROTATION24, 10 : OBJ9, VISIBLE, 1.000000E+00 : OBJ9, TYPE, BLOCKAGE : OBJ9, MATERIAL, 117 : OBJ9, ADIABATIC, 0.000000E+00, 0.000000E+00 : OBJ9, INI_TEMP, 2.000000E+01 > OBJ10, NAME, ERIDGE : OBJ10, POSITION, 6.038000E+00, 1.500000E01, 0.000000E+00 : OBJ10, SIZE, 1.500000E01, 1.788000E+00, 3.000000E01 : OBJ10, CLIPART, ridge8 : OBJ10, ROTATION24, 19 : OBJ10, TYPE, BLOCKAGE : OBJ10, MATERIAL, 117 : OBJ10, ADIABATIC, 0.000000E+00, 0.000000E+00 : OBJ10, INI_TEMP, 2.000000E+01 > OBJ11, NAME, NRIDGE : OBJ11, POSITION, 1.500000E01, 1.938000E+00, 0.000000E+00 : OBJ11, SIZE, 5.888000E+00, 1.500000E01, 3.000000E01 : OBJ11, CLIPART, ridge8 : OBJ11, ROTATION24, 22 : OBJ11, TYPE, BLOCKAGE : OBJ11, MATERIAL, 117 : OBJ11, ADIABATIC, 0.000000E+00, 0.000000E+00 : OBJ11, INI_TEMP, 2.000000E+01 > OBJ12, NAME, ROOF : OBJ12, POSITION, 0.000000E+00, 8.190000E01, 1.380000E+00 : OBJ12, SIZE, 6.188000E+00, 4.500000E01, 0.000000E+00 : OBJ12, CLIPART, cube : OBJ12, ROTATION24, 1 : OBJ12, VISIBLE, 1.000000E+00 : OBJ12, TYPE, PLATE : OBJ12, ADIABATIC, 0.000000E+00, 0.000000E+00 > OBJ13, NAME, NARC : OBJ13, POSITION, 0.000000E+00, 1.269000E+00, 1.043150E+00 : OBJ13, SIZE, 6.188000E+00, 8.190000E01, 3.368500E01 : OBJ13, CLIPART, babet4 : OBJ13, ROTATION24, 9 : OBJ13, VISIBLE, 1.000000E+00 : OBJ13, TYPE, BLOCKAGE : OBJ13, MATERIAL, 117 : OBJ13, ADIABATIC, 0.000000E+00, 0.000000E+00 > OBJ14, NAME, SARC : OBJ14, POSITION, 0.000000E+00, 0.000000E+00, 1.043150E+00 : OBJ14, SIZE, 6.188000E+00, 8.190000E01, 3.368500E01 : OBJ14, CLIPART, babet4 : OBJ14, ROTATION24, 23 : OBJ14, VISIBLE, 1.000000E+00 : OBJ14, TYPE, BLOCKAGE : OBJ14, MATERIAL, 117 : OBJ14, ADIABATIC, 0.000000E+00, 0.000000E+00 > OBJ15, NAME, BNR1GAS : OBJ15, POSITION, 1.521600E+00, 1.018600E+00, 1.380000E+00 : OBJ15, SIZE, 5.080000E02, 5.080000E02, 0.000000E+00 : OBJ15, CLIPART, cylinder : OBJ15, ROTATION24, 1 : OBJ15, TYPE, INLET : OBJ15, DENSITY, 6.560000E01 : OBJ15, VELOCITY, 0.000000E+00, 0.000000E+00,2.267900E+01 : OBJ15, TEMPERATURE, 2.500000E+01 : OBJ15, INLET_MIXF, 1.000000E+00 : OBJ15, TURBINTENS, 5.000000E+01 > OBJ16, NAME, BNR1AIR : OBJ16, POSITION, 1.470800E+00, 9.678000E01, 1.380000E+00 : OBJ16, SIZE, 1.524000E01, 1.524000E01, 0.000000E+00 : OBJ16, CLIPART, pip2colo : OBJ16, ROTATION24, 1 : OBJ16, TYPE, INLET : OBJ16, DENSITY, 1.059000E+00 : OBJ16, VELOCITY, 0.000000E+00, 0.000000E+00,3.104400E+01 : OBJ16, TEMPERATURE, 6.000000E+01 : OBJ16, TURBINTENS, 5.000000E+01 > OBJ17, NAME, BNR2GAS : OBJ17, POSITION, 6.188000E+00, 1.618600E+00, 9.403000E01 : OBJ17, SIZE, 0.000000E+00, 5.080000E02, 5.140000E02 : OBJ17, CLIPART, cylinder : OBJ17, ROTATION24, 9 : OBJ17, TYPE, INLET : OBJ17, DENSITY, 6.560000E01 : OBJ17, VELOCITY, 2.240000E+01, 0.000000E+00,3.548000E+00 : OBJ17, TEMPERATURE, 2.500000E+01 : OBJ17, INLET_MIXF, 1.000000E+00 : OBJ17, TURBINTENS, 5.000000E+01 > OBJ18, NAME, BNR2AIR : OBJ18, POSITION, 6.188000E+00, 1.567800E+00, 8.888500E01 : OBJ18, SIZE, 0.000000E+00, 1.524000E01, 1.543000E01 : OBJ18, CLIPART, pip2colo : OBJ18, ROTATION24, 9 : OBJ18, TYPE, INLET : OBJ18, DENSITY, 1.059000E+00 : OBJ18, VELOCITY, 3.066200E+01, 0.000000E+00,4.856000E+00 : OBJ18, TEMPERATURE, 6.000000E+01 : OBJ18, TURBINTENS, 5.000000E+01 > OBJ19, NAME, BNR3GAS : OBJ19, POSITION, 6.188000E+00, 4.186000E01, 9.403000E01 : OBJ19, SIZE, 0.000000E+00, 5.080000E02, 5.140000E02 : OBJ19, CLIPART, cylinder : OBJ19, ROTATION24, 9 : OBJ19, TYPE, INLET : OBJ19, DENSITY, 6.560000E01 : OBJ19, VELOCITY, 2.240000E+01, 0.000000E+00,3.548000E+00 : OBJ19, TEMPERATURE, 2.500000E+01 : OBJ19, INLET_MIXF, 1.000000E+00 : OBJ19, TURBINTENS, 5.000000E+01 > OBJ20, NAME, BNR3AIR : OBJ20, POSITION, 6.188000E+00, 3.678000E01, 8.888500E01 : OBJ20, SIZE, 0.000000E+00, 1.524000E01, 1.543000E01 : OBJ20, CLIPART, pip2colo : OBJ20, ROTATION24, 9 : OBJ20, TYPE, INLET : OBJ20, DENSITY, 1.059000E+00 : OBJ20, VELOCITY, 3.066200E+01, 0.000000E+00,4.856000E+00 : OBJ20, TEMPERATURE, 6.000000E+01 : OBJ20, TURBINTENS, 5.000000E+01 > OBJ21, NAME, SWRIDGE : OBJ21, POSITION, 0.000000E+00, 0.000000E+00, 0.000000E+00 : OBJ21, SIZE, 1.500000E01, 1.500000E01, 3.000000E01 : OBJ21, CLIPART, ridge8 : OBJ21, ROTATION24, 1 : OBJ21, TYPE, BLOCKAGE : OBJ21, MATERIAL, 117 : OBJ21, ADIABATIC, 0.000000E+00, 0.000000E+00 : OBJ21, INI_TEMP, 2.000000E+01 > OBJ22, NAME, SERIDGE : OBJ22, POSITION, 6.038000E+00, 0.000000E+00, 0.000000E+00 : OBJ22, SIZE, 1.500000E01, 1.500000E01, 3.000000E01 : OBJ22, CLIPART, ridge8r : OBJ22, ROTATION24, 10 : OBJ22, TYPE, BLOCKAGE : OBJ22, MATERIAL, 117 : OBJ22, ADIABATIC, 0.000000E+00, 0.000000E+00 : OBJ22, INI_TEMP, 2.000000E+01 > OBJ23, NAME, NERIDGE : OBJ23, POSITION, 6.038000E+00, 1.938000E+00, 0.000000E+00 : OBJ23, SIZE, 1.500000E01, 1.500000E01, 3.000000E01 : OBJ23, CLIPART, ridge8r : OBJ23, ROTATION24, 19 : OBJ23, TYPE, BLOCKAGE : OBJ23, MATERIAL, 117 : OBJ23, ADIABATIC, 0.000000E+00, 0.000000E+00 : OBJ23, INI_TEMP, 2.000000E+01 > OBJ24, NAME, NWRIDGE : OBJ24, POSITION, 0.000000E+00, 1.938000E+00, 0.000000E+00 : OBJ24, SIZE, 1.500000E01, 1.500000E01, 3.000000E01 : OBJ24, CLIPART, ridge8r : OBJ24, ROTATION24, 22 : OBJ24, TYPE, BLOCKAGE : OBJ24, MATERIAL, 117 : OBJ24, ADIABATIC, 0.000000E+00, 0.000000E+00 : OBJ24, INI_TEMP, 2.000000E+01 STOP 

March 25, 2002, 15:15 
Re: Combustion Convergence problems

#2 
Guest
Posts: n/a

There's one more object file I made up that is mentioned in the Q1 file:
PIP2COLO.DAT  like PIPECOLO.DAT but with 3:1 ratio instead of 2:1 Art Stretton 

March 26, 2002, 17:58 
Re: Combustion Convergence problems

#3 
Guest
Posts: n/a

I have been using the CHEMKIN subsystem to model combustion in a simpler, laminar case, so I don't think my experience would benefit what you are doing very much. The only thing I can think of is the size of your grids. It may not affect your simulation, but I have noticed convergence to be sensitive to the grid, at least witht the CHEMKIN subsystem. You might try coarsening your grid substantially and see how that affects convergence. Just a thought.  PattiMichelle


March 27, 2002, 00:33 
Re: Combustion Convergence problems

#4 
Guest
Posts: n/a

I started off with about 200,000 nodes and couldn't get the convergence errors below E+05, more than doubled it to 450,000 nodes to get to E+03, and hoped to get below E+01 with 750,000 nodes. The gradients from the fire plumes are so steep that about 30% of the volume needs to be pretty finely divided. I was also concerned about the aspect ratio of the cells when the finely divided grid intersects with coarse grid sections (kept it at 4:1 or below). But with the increased grid size and increase model features, runs for 1000 iterations have increased from 1/2 hour to 8 hours. Unfortunately, there has also been a decrease in rate of convergence so it seems to be taking more iterations as well to exponentially increase the total time of execution needed.
What I will probably do is increase my tolerance for aspect ratio to 8:1 or even 10:1 to probably lop at least half of the node count off. Maybe Phoenics is more tolerant of cell aspect ratio than other CFD codes that I've been exposed to. Thanks for your input, PattiMichelle. I looked through all the relevant posts I could find before posting myself, and your name was pretty prominent:^) Here's hoping the combustion experts at CHAM volunteer chime in to give me some guidance with Phoenics issues. 

March 27, 2002, 22:16 
Re: Combustion Convergence problems

#5 
Guest
Posts: n/a

Art,
Have you adjusted your relaxtation parameters. In combustion modeling I have found that in the past the convegence was sensative to the convergance of the density of the gas. Maybe it will converge faster if you play with this relaxation parameter. Hope this helps. Leon 

April 2, 2002, 05:59 
Re: Combustion Convergence problems

#6 
Guest
Posts: n/a

If you need linear relaxation on p1 as low as 1e4 then something is very wrong, either with the setup, phoenics itself or the convergence parameters. The pressure relaxation should be no lower than 0.3 and more likely around 0.5 to 1.0.
As already suggested the density should be stored and relaxed, say 0.3 linear. If very strong linear relaxation (say 1.e3) is applied to the density, then convergence should be achieved but this should not be necessary. If this fails, then I would simplify the whole problem by using as coarse a mesh as possible even if this means simplifying the geometry somewhat. The motivation is to get something that converges. The velocity relexation also appears to be somewhat heavy. From what I can make out in your Q1 file the velocity false time step should be roughly around 1.e3 to 1.e2 s. Linear relaxation of 0.3 to 0.5 should be used on k and e. The enthalpy needs to be relaxed, probably using a false time step around 1 sec. The mixture fraction shouldd employ little or no relexation, which is what you are effectively doing. 

Thread Tools  
Display Modes  


Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
convergence problems using SBF=1  Fonzie  CFX  1  March 23, 2007 09:46 
Problems in convergence  Simon  CDadapco  3  February 14, 2005 06:36 
Convergence Problems  James  FLUENT  1  May 29, 2004 07:07 
Convergence problems  Chetan  FLUENT  3  April 15, 2004 19:13 
Convergence problems  Emilien  FLUENT  3  May 3, 2002 08:43 