Problems in transient supersonic air flow
 

i wanted to simulate the transient supersonic shock wave of a body crossing Mach 1. It's a 10 sec transient, I tried timesteps of 10 ms, 1 ms, and overnight for 0.1 ms because it stopped with errors. They all fail after writing a couple of transients (every 0.1 s) at the same point, just after 0.2 sec.
My body is just a cylinder and the geometry is like a plate (modeling a 2-D flow).
Is it always so tricky this kind of simulations or am I doing something wrong? Any suggestions?

Transonic flows are always tricky. But you should be able to get this to work.

Mesh quality is very important. Make sure the mesh is as good as you can make it.

I think so too. The mesh is not specially fine but is not coarse at all, I focused on a surface inflation but may is necessary some refinement in the shock wave area. But.... holidays first, this will be three weeks later :).

I would start with a coarse mesh with no inflation - this should minimise the boundary layer complications and be easier to solve. Once a coarse mesh is running well you can start refining the mesh to resolve flow details.

But again, I stress that mesh quality is going to be very important. Make sure your mesh does not contain high/low angle elements, the volume ratios are good and the aspect ratios not too excessive.