CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Phoenics (http://www.cfd-online.com/Forums/phoenics/)
-   -   Problems with variant i208 "Flow around Big Ben" (http://www.cfd-online.com/Forums/phoenics/51958-problems-variant-i208-flow-around-big-ben.html)

Alexander September 26, 2005 06:27

Problems with variant i208 "Flow around Big Ben"
 
Good morning for all. I am making the calculation of distribution of hot air plume in the wind, and it was interesting for me to run library case I208: Flow around "Big Ben". I used Q1 file directly from the 3.6.1 PHOENICS library! The velocity of my computer is 3200 MHz. After 80766 iterations error for V1 and W1 velocities components is about 40%. The error for p1, KE, and EP is about 0.3%, and for U1 error is about 2%. It is normally? It seems to me that error for V1 and W1 is very great. It is a library case! Thank you very much in advance for your help, Alexander.


mick September 28, 2005 12:54

Re: Problems with variant i208 "Flow around Big Be
 
I have just run this with v3.6.1, it takes a while but it is convergent. This convergence rate could probably be improved by setting auto-relax off and selecting better values of dtfals.

But for the two secondary velocities to which you refer and from the printout at the bottom of the result file as shown below.

Normalised residual for V1 is

7.790E+01/8.663472E+00 * 100 = 900%

Normalised residual for W1 1.913E+02/7.734980E+01 * 100 = 200 %

so the case could definitely improve from running longer, I would want these to come down to say 40% at least.

But remember that V and W only really have values around the big ben object itself caused by the deflection, so place the probe in a sensible position near this object and check the spot-values whilst the case is running, you should be happy when the spot value is hardly changing.

I dont think there is anything wrong with this case it just needs more sweeps or conwiz (auto-relax) switched off and good dtfals values selected for quicker convergence.

Print out from the result file below .......

& resfac= 1.000000E-03 variable resref (res sum)/resref (res sum)

P1 1.690E-02 1.080E+02 1.826E+00

U1 5.717E-02 5.629E+03 3.219E+02

V1 5.572E-04 1.398E+05 7.790E+01

W1 3.888E-04 4.920E+05 1.913E+02

KE 9.689E-04 8.505E+02 8.240E-01

EP 1.810E-05 5.563E+03 1.007E-01

************************************************** **********

Sources and sinks

Nett source of U1 at patch named: OB3 (OUT ) =-6.587086E+04 Nett source of U1 at patch named: OB4(SKY ) =-2.107056E+03 Nett source of U1 at patch named: OB5 (SSIDE ) =-1.725731E+02 Nett source of U1 at patch named: OB6 (NSIDE ) =-9.824491E+01 Nett source of U1 at patch named: BLIN1 (IN ) = 6.878616E+04 Nett source of U1 at patch named: OB2 (GROUND ) =-1.663947E+02 pos. sum= 6.878616E+04 neg. sum=-6.841512E+04 nett sum= 3.710312E+02

Nett source of V1 at patch named: OB3 (OUT ) =-8.663472E+00 Nett source of V1 at patch named: OB4 (SKY ) =-8.805566E-02 Nett source of V1 at patch named: OB5 (SSIDE ) = 1.308345E+00 Nett source of V1 at patch named: OB6 (NSIDE ) =-3.410851E-01 Nett source of V1 at patch named: OB2 (GROUND ) =-1.222104E-01 pos. sum= 1.308345E+00 neg. sum=-9.214824E+00 nett sum=-7.906479E+00

Nett source of W1 at patch named: OB3 (OUT ) = 7.734980E+01 Nett source of W1 at patch named: OB4 (SKY ) =-9.972672E+00 Nett source of W1 at patch named: OB5 (SSIDE ) =-3.159088E-01 Nett source of W1 at patch named: OB6 (NSIDE ) =-9.849952E-02 pos. sum= 7.734980E+01 neg. sum=-1.038708E+01 nett sum= 6.696272E+01


Alexander September 28, 2005 14:14

Re: Problems with variant i208 "Flow around Big Be
 
Dear Mick, Thank you very much for your message. Now I understand situation clearer. Regards, Alexander



All times are GMT -4. The time now is 22:03.