CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   Main CFD Forum (https://www.cfd-online.com/Forums/main/)
-   -   Roe-Solver Instability at Mach=1 (https://www.cfd-online.com/Forums/main/120409-roe-solver-instability-mach-1-a.html)

 Dwayne July 6, 2013 08:34

Roe-Solver Instability at Mach=1

Hey Guys,

can somebody tell me the reasons why Roes solver become instable at Mach=1? The literature says that the solver becomes a central-differences-method at Mach=1 because the right side of the Roe-flux vanishes, is this true? Isn't it only one eigenvalue of the jacobian that becomes zero?

Thanks for help!

 immortality July 7, 2013 04:00

hi
What software you use?

 Dwayne July 7, 2013 06:36

Quote:
 Originally Posted by immortality (Post 438202) hi What software you use?
I use Matlab to calculate a one-dimensional shock-tube-problem for the Euler-Equations.

 immortality July 7, 2013 07:50

how?is there such program in Matlab or you have written it?

 Dwayne July 7, 2013 08:06

Quote:
 Originally Posted by immortality (Post 438264) how?is there such program in Matlab or you have written it?
I have written it.

 Dwayne July 8, 2013 13:00

Nobody any idea?

 praveen July 9, 2013 21:13

At sonic point one eigenvalue becomes zero. This leads to loss of dissipation in one of the eigenfields. It manifests as an entropy violating jump inside a sonic rarefaction. Look for "entropy fix" to overcome this problem.

 Dwayne July 10, 2013 05:57

Quote:
 Originally Posted by praveen (Post 438785) At sonic point one eigenvalue becomes zero. This leads to loss of dissipation in one of the eigenfields. It manifests as an entropy violating jump inside a sonic rarefaction. Look for "entropy fix" to overcome this problem.
Ok i understand, thank you!

 All times are GMT -4. The time now is 22:11.