I have tried to model a shock tube, but am struggling to get the initial conditions right, because in order to define the different pressures, it seems one has to define 2 fluids, and it seems that one cannot have different fluids in adjacent cells. I have tried defining baffle cells between them and then removing the baffles (using the events module), but this doesn't seem to work either.

Has anyone succeeded in doing this kind of thing without using subroutines? (I don't have the fortran compiler)

Any feedback would be appreciated.

Gareth

Hi, i had the same problem a few years back when also running a shock tube case.

I got around this by defining a dummy model with two fluid(material) types, each with a different initial pressure, separated by a baffle. I ran this for 1 iteration and then smapped the results onto my shock tube model (one material type, no baffle), hence defining the initial conditions. Hope this helps ..

Regards Dean

Hello gareth,

I'd like to know if you are trying to model the shock tube in conjunction with experimental results for shock interaction or for expansion wave interactions. the method suggested by Dean seems to be the best for simulating an entire shock tube. but the driver and driven sections can be modelled seperately using pressure/inlet boundaries depenending on which wave u are trying to model (the shock or the expansion)

cheers nandu

Thanks Nandu

I actually ended up doing what you suggested after a suggestion from someone else. It worked fine. I was trying to get an idea of the abillity of CFD to model shock waves, particularly in shock tubes for a research idea in the area of engine design.

Have you had experience in modelling shock tubes in conjunction with experimental results? What for, and how did the results compare. I assume you are in mech eng at Wits. Is it part of a MSc or PhD? I'd be interested to email further.

Regards

Gareth