****** TO LOAD CASE 105:TYPE L(N105) ***** GROUP 1. Run title and other preliminaries TEXT(YX TURBULENT BACKWARD-FACING-STEP FLOW TITLE mesg( PC486/50 time last reported as 1.0 hr DISPLAY Numerical Schemes validation example: 2-d x-y, Cartesian, steady, elliptic simulation The problem considered is turbulent flow over a backward facing step, as studied experimentally W.D.Moss et al (1st Symp. on Turb. Shear Flows, Univ. Park, Penn., USA, Vol.II, p13.1, 1977). The Reynolds number based on step height H is 5E4 and the expansion ratio is 1.1. The inlet is specified as a uniform inflow located 12H upstream of the step. The outlet plane is located 30H downstream of the step. This Q1 uses quadratic-upwind differencing for the velocities, and Koren's bounded scheme for the turbulence variables. ENDDIS The configuration is similar to that of Kim et al (ASME J.Fluids Engng, Vol.102, p302, 1980 ), as studied numerically throughout the PHOENICS Library (see e.g. library cases T103 &T305) and by many others (see e.g. S.Thangam & C.G.Speziale, AIAA J, Vol.30, No.5, p1314, 1992). The present geometry has a smaller expansion ratio and slightly higher Reynolds number. The measured reattachment length for this case is Xr/H=5.5, as measured from the step. The default computation using higher- order differencing schemes predicts a reattachment length of Xr/H=4.74. The PHOENICS default of hybrid differencing predicts Xr/H=4.43. No grid refinement studies have been conducted. PHOTON USE P 0.20443E+04 0.15633E+04 CR gr ou z 1;use patgeo;msg vectors;vec z 1 sh msg pressand then to end pause ENDUSE REAL(HIN,HSTEP,LEIN,LEUP,LERC,LEDN,RENO,VIN,KEIN,EPIN) REAL(RLXFAC);INTEGER(NX1,NX2,NX3,NX4,NY1,NY2,NY3,NXS);CHAR(SCHM) ** All dimensions are based on: kg, m, sec HIN=0.762;HSTEP=0.076;RENO=50000.0;ENUL=1.6E-5 VIN=RENO*ENUL/HSTEP KEIN=0.5*1.E-5*VIN*VIN;EPIN=KEIN**1.5*0.1643/(0.09*HIN) GROUP 2. Transience; time-step specification GROUP 3. X-direction grid specification NX1=10;NX2=10;NX3=30;NX4=15;NXS=NX1+NX2 LEIN=7.0*HSTEP;LEUP=5.0*HSTEP;LERC=10.0*HSTEP;LEDN=30.0*HSTEP NREGX=4;IREGX=1;GRDPWR(X,NX1,LEIN,1.0) IREGX=2;GRDPWR(X,NX2,LEUP,-1.3);IREGX=3;GRDPWR(X,NX3,LERC,1.2) IREGX=4;GRDPWR(X,NX4,LEDN-LERC,1.4) GROUP 4. Y-direction grid specification NY1=10;NY2=10;NY3=10 NREGY=3;IREGY=1;GRDPWR(Y,-NY1,HSTEP,1.3) IREGY=2;GRDPWR(Y,NY2,HSTEP,1.2) IREGY=3;GRDPWR(Y,NY3,HIN-HSTEP,1.5) GROUP 5. Z-direction grid specification NZ=1;ZWLAST=0.01 GROUP 6. Body-fitted coordinates or grid distortion GROUP 7. Variables stored, solved & named SOLVE(P1,U1,V1) SOLUTN(P1,Y,Y,Y,N,N,N);TURMOD(KEMODL);STORE(ENUT) ** use arithmetic averaging for vels as V2.1.3 & V2.2 default of harmonic averaging produces unphysical U1 velocities in near-wall cells of recirculation zone. SOLUTN(U1,P,P,P,P,P,N);SOLUTN(V1,P,P,P,P,P,N) ** conjugate-gradient solver CSG3=CNGR;RLXFAC=XULAST/(NX*VIN) GROUP 8. Terms (in differential equations) & devices MESG( Enter required convection scheme MESG( Default: QUICK for momentum; KOREN for k and eps MESG( The alternative is: MESG( HYB - Hybrid differencing for all variables READVDU(SCHM,CHAR,HOC) CASE :SCHM: OF WHEN HYB,3 + MESG(Hybrid-differencing scheme + DIFCUT=0.5;RLXFAC=XULAST/(NX*VIN) WHEN HOC,3 + MESG(QUICK for momentum; KOREN for k and e + SCHEME(QUICK,U1,V1);SCHEME(KOREN,KE,EP) + RLXFAC=0.1*XULAST/(NX*VIN) ENDCASE GROUP 9. Properties of the medium (or media) RHO1=1.2 GROUP 10. Inter-phase-transfer processes and properties GROUP 11. Initialization of variable or porosity fields FIINIT(U1)=VIN;FIINIT(P1)=1.3E-4;FIINIT(KE)=KEIN;FIINIT(EP)=EPIN ** Initialization of variables in blocked region ** since using arithmetic averaging on vels, it is necessary to intoduce wall friction patches around blockage separately, remove - from patch limits. CONPOR(STEP,0.0,CELL,#1,-#2,#1,-#1,#1,#1) CONPOR(STEP,0.0,CELL,#1,#2,#1,#1,#1,#1) GROUP 12. Convection and diffusion adjustments GROUP 13. Boundary conditions and special sources PATCH(IN1,WEST,1,1,NY1+1,NY,1,NZ,1,LSTEP) COVAL(IN1,P1,FIXFLU,RHO1*VIN) COVAL(IN1,U1,ONLYMS,VIN);COVAL(IN1,V1,ONLYMS,0.0) COVAL(IN1,KE,ONLYMS,KEIN);COVAL(IN1,EP,ONLYMS,EPIN) ** Wall boundary conditions PATCH(TWALL,NWALL,1,NX,NY,NY,1,NZ,1,LSTEP) COVAL(TWALL,U1,LOGLAW,0.0) COVAL(TWALL,KE,LOGLAW,LOGLAW);COVAL(TWALL,EP,LOGLAW,LOGLAW) PATCH(BWALL,SWALL,#3,NX,1,1,1,NZ,1,LSTEP) COVAL(BWALL,U1,LOGLAW,0.0) COVAL(BWALL,KE,LOGLAW,LOGLAW);COVAL(BWALL,EP,LOGLAW,LOGLAW) ** Outlet boundary PATCH(OUT,EAST,NX,NX,1,NY,1,NZ,1,LSTEP) COVAL(OUT,P1,10.0,0.0);COVAL(OUT,KE,ONLYMS,SAME) COVAL(OUT,EP,ONLYMS,SAME) ** Step wall boundary conditions use explicit patches & not CONPOR - arguments because of need to use arithmetic averaging on velocities PATCH(STEP-WW,WWALL,NXS+1,NXS+1,1,NY1,1,1,1,1) COVAL(STEP-WW,V1,LOGLAW,0.0);COVAL(STEP-WW,KE,LOGLAW,LOGLAW) COVAL(STEP-WW,EP,LOGLAW,LOGLAW) PATCH(STEP-SW,SWALL,1,NXS,NY1+1,NY1+1,1,1,1,1) COVAL(STEP-SW,U1,LOGLAW,0.0);COVAL(STEP-SW,KE,LOGLAW,LOGLAW) COVAL(STEP-SW,EP,LOGLAW,LOGLAW) GROUP 14. Downstream pressure for PARAB=.TRUE. GROUP 15. Termination of sweeps ** About 2000 sweeps are required for complete convergence when using higher-order differencing. Only 400 sweeps are required when using hybrid differencing. LSWEEP=200 GROUP 16. Termination of iterations LITER(U1)=10;LITER(V1)=10;LITER(KE)=10;LITER(EP)=10 GROUP 17. Under-relaxation devices RELAX(P1,LINRLX,0.8);RELAX(ENUT,LINRLX,0.6) RELAX(U1,FALSDT,RLXFAC);RELAX(V1,FALSDT,RLXFAC) RELAX(KE,FALSDT,RLXFAC);RELAX(EP,FALSDT,RLXFAC) GROUP 18. Limits on variables or increments to them GROUP 22. Spot-value print-out TSTSWP = -1;IYMON=NY1/2;IXMON=NX1+NX2+NX3/2 GROUP 23. Field print-out and plot control NPLT=50;ITABL=3 GROUP 24. Dumps for restarts