[tomhcosta]

Hi,

I'm modeling a 2D shock tube, in fluent, as a rectangle (1meter long by 6cm diameter) with 2 separated sections, a high-pressure one called driver and a low pressure one called driven. Im using density-based solver with transient analysis, inviscid with energy eq. and species transport with hydrogen-ar mixture with both as ideal gas. I set the operating pressure inside to 0 pa. After this i Initialize with 10 bar in the driver section and 1 bar in the driven section and value 1 of hydrogen for the driver section. I run the calculations for 50 time steps of 20 iterations each with a time step of 7e-06s and i do all of this with 4 solver processes and double precision.

When I model it for 10 bar, it all goes smooth and the solution converges but when i change the driver pressure to 20bar or higher the solution diverges a lot and presents results like 0 pressure along the shock-tube. The reason i am doing this is to see the temperate and density behaviour for different pressures inside the shock-tube and then test if the length of the tube will have any effect on them.

I think this is a very simple problem but i just can't seem to get it right. If anyone can help, i would be very grateful.

[AlexanderZ]

did you try to decrease timestep?

you may also decrease under relaxation factors and switch to first order discretization models

if you are using not default material properties, check them again

[tomhcosta]

Originally Posted by AlexanderZ

did you try to decrease timestep?

you may also decrease under relaxation factors and switch to first order discretization models

if you are using not default material properties, check them again
Like
__________________
best regards

Since I am modeling a shock tube, I am using density based solver so there are no urfs saddly.

I can try to reduce the time step but since it is already at 7e-6 do you think it can be reduced more? What would this do?

Haven't tried the first order discretization, will try but can you explain what this does in relation to second order?

Thank you,
Tomás