Discussion Forum

I am working on flow over cone(axis symmetric) for mach number M=5.75
The approach to the problem begins with a half cone geometry of 8* half-angle and a rectangular box(15-times larger than maximum dimension of half cone)

I am using COMSOL 4.1 with non-isothermal flow module(laminar)

The rectangular box has following boundary conditions

1>inlet-velocity,constant temperature
2>outlet-pressure,outflow
3>outer wall opposite to half-cone(slip)
4>wall containing half-cone(slip-axis symmteric)
5>half-cone surface(no slip)

Study Type
-stationary solver

Approach
I started with inlet velocity 20m/s and kept this as initial condition for inlet velocity 50m/s and so on in steps of 30m/s.
I refined the mesh as and when required as per the physics and the way of meshing was well taken care. As i increased inlet velocity, i was slowly decreasing outlet pressure also. The results were satisfactory till 250m/s-inlet and outlet pressure-0.4[atm].

Issue
After 250m/s with all possible and careful configuration of solver setting and meshing the results are distorted for
inlet-260m/s and after this it doesnot converge properly. Maximum i have gone upto 280/m/s-inlet with step increase reduced to 5m/s.

How can i curb this problem ?Do i need to change CFL number?

Additional information
>>Relative tolerence(0.001)
>>used both Pardiso or iterative(krylov-preconditioning) with fully coupled
>>cone is spherically blunted
>>automatic damping [varied both i)step for update(2-5) ii)minimum dampin factor(10^-4 to 10^-8)]
>>configuration of computer 8GB RAM quad core also 32GB RAM server was also used
>>Ck=0.5-0.75 was used in consistent stabilization
>>inconsistent stabilization was never used as it would affect the physics