[YADAVRAJA]

hi,
i am extremely new to ansys ,i want to analysis an simple shock tube but i am not able to model it ,like a simple shock tube with a rupturing diaphragm please tell how to model rupturing of diaphragm in a shock tube

[urosgrivc]

Transient pressure Boundary condition comes to mind,
-Tabular data
or
-function
at the inlet

[ghorrocks]

It is normally much simpler than what urosgrivc suggests.

Normally you set up a shock tube simulation as:
1) Apply the high pressure boundary condition to the oriface location. No need for time dependant conditions as time will start at exactly the time of rupture.
2) Start the domain with the initial condition applied (low pressure)
3) simulate from there.

I have done many simulations of this case as a benchmark for compressible transient flow with shocks so I am quite familiar with it.

Remember that there is an analytical solution for shock wave flow so you have an exact answer to compare to.

Make sure you look at the effect of convergence tolerance, mesh size, advection scheme, transient time differencing and other compressible flow options to see what difference they make. This exercise is very instructive and will help you learn a lot about the basics of CFD.

In my experience you should be able to get the post-shock pressure and shock velocity accurate to less than 1% if you do things right.

[urosgrivc]

Mr. Glenn
Do I understand corectlly?
To put the pressure to 100% in zero time.
Does this produce any erors conected with large pressure gradients.
Should the inlet face have inflation layers than? just a thought.

I thoughth that with function you are able to capture the rise of presure to the 100%max than fall of pressure as the air tank has a limited size.
(would need to calculate or predict rise and fall times obwiously) or simulate them.
I probably didnt understand what the purpose of the simulation is exactly.
Aither that or the membrane pop takes no time at all and the tank is indefinete

[ghorrocks]

To put the pressure to 100% in zero time.

Yes, you start with a pressure discontinuity. That is what a shock wave is.

Does this produce any erors conected with large pressure gradients.

Yes, it gives the solver a hard time but it can handle it.

Should the inlet face have inflation layers than? just a thought.

Generally no, in my experience a high quality mesh evenly spread throughout the domain is more important. Don't forget the shock moves so inflation at a fixed point does not help much.

The shock tube simulation is a standard benchmark simulation for shock waves and compressible flow. The standard benchmark case is a step change in pressure (ie a shock) with no ramping. There are analytical solutions for this so it is a really good benchmark as you have an exact answer to compare to.