[AshwaniAssam]

Dear all,

I was simulating the inviscid flow over circular bump (10% height) for M=0.675 (transonic regime). The problem I am facing is that I am not getting a proper match for the mach number plot along the lower bump wall at downstream. The plot is attached below. I have used subsonic outflow boundary condition (Whitfield'84 et al). I am using total pressure boundary condition at inflow. The walls have slip condition. it is a density-based solver, with roe scheme. The spatial accuracy is 2nd order, with minmod and Venkatkrishnan limiter. Can anyone please help. Did somebody also got such difference?

Also, I want to ask as far as I have seen most transonic bump cases has been reported using pressure-based solver. Is that one reason for such difference.

I am also attaching the Mach contours.

Thanks for going through this post.

M675.jpg

machcnt.jpg

[Martin Hegedus]

I suspect it's the numerical dissipation which is adding additional entropy. I would suggest running at a lower Mach number such that at shock does not exist. That will remove the entropy generated by the shock. Then try adding more grid points. I suspect the "wake" will disappear to some extent. However, to some degree it will always be there. To remove it you might need to go to a higher order less dissipative method.

[AshwaniAssam]

Thanks for your reply. I would run the lower mach number and report for the same. We are understanding that since we are using AUSM i.e. an upwind scheme we are getting such dissipation downstream. I have ran the similar cases with MacCormack Scheme (Central scheme), there I got a better match at the downstream. Is it the something same with pressure based solver, because of which dissipation is less their?

[Martin Hegedus]

I'm not familiar with pressure based schemes, but how one iterates towards steady state should not effect the amount of dissipation at steady state. If the pressure based scheme uses a different discretization method, then yes the amount dissipation could be different.