Hi all

I have to setup up a 1D shock wave in TascFlow or CFX 4.4 Any help will be greatly appreciated

Thanks in advance ant

Hi Ant,

I suggest creating a converging-diverging duct. Specify Total conditions at the inlet and a static pressure at the outlet which will result in sonic conditions at the throat.

Don't specify mass flow. If you do so and the mass flow is greater than the choke condition, the solver will blow as the boundary conditions are unphysical. For TASCflow: I recommend using the modified linear profile with PAC terms on (iskew=3, lpac=t).

Robin

Hi,

But wouldnt that simply set up a standing shock in the diverging section of the throat?

How would you generate a moving normal shock, such as in a shock tube?

cheers N Menon

Yes, that would set up a standing normal shock wave.

You could set up a shock tube by running a transient simulation and setting the initial conditions with a step function to provide the initial pressure difference.

On the other hand, if you wish to solve a steady state standing shock wave as though you were following it through the shock tube (ie moving walls and appropriate b.c.'s), you will have much more difficulty as the location of the shock would be indeterminate and could prove to be unstable. The advantage to my first suggestion is that the location of the shock would be fixed.

Robin

Hi all,

There is another type of 1D shock wave which is commonly modelled, that is the transient shock wave problem. I did an extensive validation of the CFX4.3 code on the transient shock wave problem as part of my PhD thesis. The following is a quick summary of my results from my thesis.

As the original posting did not specify which type of 1D shock problem ant was modelling, I thought I would discuss the transient type of problem (The steady state problem has already had some comments).

The set-up for this problem is a long thin (1D) domain, initially with two regions at different temperatures and pressures. All boundaries are walls, so there is no fluid which enters or leaves the domain. At time zero the gases are allowed to interact and a shock wave goes through the low pressure gas, a rarefaction wave through the high pressure gas and a contact surface at the interface of the two gasses. An analytical solution exists for some classes of these problems, so an exact answer is known.

In my experience CFX did an excellent job of modelling this problem, with the post-shock conditions predicted with accuracies less than 0.1%. The location of the shock was predicted quite well, something like within 1 or 2%.

When using 1st order temporal differencing the shock was smeared across 4 nodes and had no overshoots. When 2nd order temporal differencing was used the shock was resolved sharper but had 10-20% overshoots. I did not find much difference between the various spatial differencing schemes.

It was possible to reduce the overshoots using the high mach number version of the code (as discussed in the manual), but as I was validating the code for use on another problem where the high mach number code was inappropriate I mainly used the default options.

This problem can be entirely set up in the command file, but I used some user fortran. If you want to about how I set it up I can dig my thesis up again and send you the relevant bit.

Regards, Glenn Horrocks

Hi Glen,

Could you send me the command file if you still have it with you. I am battling to set up a multi region simulation as u suggested in response to the original question.

regards Nandu

I want to simulate shock wave in an explosive. can I use your software? I would thank you, if quide me. Regards, maryam.

Hi Maryam,

This is not a public site and has no affiliation with CFX. I suggest you contact CFX directly for your query. You will find contact information on the CFX web site at www.cfx.aeat.com.

Regards, Robin