Hi all,

When I simulate a subsonic flow (Mach about 0.5)in a channel with an solid parallelepiped inside, the flow rate weighted total pressure is increasing for sections just upstream of the obstacle.

Is anyone has a explanation?

Thanks in advance. Karim.

The flow-rate averaged total pressure for the entire channel should not increase - something is wrong with the models you use or the numerical quality of your solution.

However, you should get a small total pressure rise on stagnating streamlines approaching the solid - this is a physical effect caused by the viscous forces.

A turbulence model which behaves badly in stagnating flows (common problem with many k-eps models) can enhace this effect and give an unphysically large rise in total pressure in the stagnation regions. You could try to switch turbulence model and see if that helps. I'm not sure which models Star-CD has available but most "non-linear k-epsilon" models behave better in stagnating flows, the Kato-Launder fix is another solution. There are also a bunch of "realizability" corrections which helps in stagnating flows.

Thanks for your answer Jonas. I dont have an increase of the total pressure for the entire channel(fortunately) but only at the approach of the stagnation area (same thing by switching off the turbulence model) . I dont known any physical effect linked by the viscous forces which increase the total pressure. For me, the diminution of total pressure along the flow is deduced by the energy conservation equation. So without chimical reaction, how the total pressure can be increase in the downstream direction??? Do you have any reference on this?

Thanks in advance Karim

The stagnation streamline will, through viscous forces, extract energy from surrounding streamlines, thus giving a local increase in total pressure in the stagnation region and a corresponding decrease in areas around. As long as your turbulence model behaves well this effect is usually quite small though, since viscous forces are small.

Total pressure on a streamline is only conserved if you have an inviscid flow.

You could try to run an inviscid simulation - then the total pressure should be constant. However, numercial dissipation might of course still give you something similar, but then you know if the cause is numerical problems.