CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Phoenics (http://www.cfd-online.com/Forums/phoenics/)
-   -   relaxation factors adjust (http://www.cfd-online.com/Forums/phoenics/51630-relaxation-factors-adjust.html)

zhujianguo July 12, 2003 04:40

relaxation factors adjust
 
i am engaging with the numerical simulation of combustion with phoenics 3.4. the model are as follows: pressure and solution;enthalpy(total),kemodl,radiosity,3-gas mixing.the convergence process is not satisfied,i want to ask for help.how to set the relaxtion factors for every variable. the iteration is in accessory. i set the relaxation factors: P1 -1.000E-01 U1 8.000E-04 V1 8.000E-05 W1 8.000E-06 KE 2.000E-04 EP 2.000E-04 H1 6.000E-03 PRPS -1.000E-01 FUEL -3.456E-02 TMP1 -1.000E-02 PROD -1.000E-01 OXID -3.456E-02 MIXF 2.000E-03 WCRT -1.000E+00 VCRT -1.000E+00 UCRT -1.000E+00 SRAD 1.200E-03 RHO1 -3.000E-01 are these relaxation factors appropriate? how to adjust them. thanks.

CHAM Support July 15, 2003 11:11

Re: relaxation factors adjust
 
For combustion, it is advisable to set DENPCO=T which activates a form of the pressure correction that is suitable for strong variations in density.

The relaxation of density is often important, say 0.1 linear or even lower. For example, unless something is wrong, the case should converge if the density is relaxed very heavily, i.e. linear at 1e-3.

The pressure relexation should not be less than 0.1 but 0.3 is more advisable unless the grid is highly distorted. Your velocity inertial relaxation factors look too small unless the flow speed is quite high or the BFC is very non orthogonal.

It is best to use linear relexation on KE and EP, say 0.3 with the k-e source term linearisation option of KELIN=3. Your MIXF inertial relexation also looks too heavy because it will prevent the combustion from developing. There should be no relaxation on TMP1, PROD, OXID, FUEL (presuming that this is not solved because the flame sheet model is used), but rather these should be limited using VARMIN and VARMAX so as to prevent non-physical values occuring during the course of convergence. Likewise, a VARMIN of say 1.e-6 can be set on DEN1 to control the density.

It is best to confirm that the case does converge: (a) without radiation; and (b) without combustion (set heat of combustion to zero)?


All times are GMT -4. The time now is 07:06.