Discussion Forum

Problem with the computation in High mach number flows

Hello everyone, I try to compute one on my fluid computation on Comsol but unfortunately I have the same error each time. When I use the laminar flow physics I can find a solution but when I use the high mach number flow physics my problem appears. This is the pop up message (Sorry the message is in french because I'm french but basically it says that "during the computation there is a non-integer power of a negative number") :

Etape ségrégée 2 Essai de calcul d'une puissance non entière d'un nombre négatif. - Fonction: ^ Echec dans l'évaluation de la variable. - Variable: comp1.hmnf.mu - Défini par: ((((comp1.T/comp1.hmnf.Tmu_ref)^1.5)comp1.hmnf.mu_ref)(comp1.hmnf.Tmu_ref+comp1.hmnf.S_s))/(comp1.hmnf.S_s+comp1.T) Echec dans l'évaluation de la variable. - Variable: comp1.hmnf.time_step_inv - Défini par: max(sqrt(emetric(comp1.u,comp1.v)((2^gmg_level)^2)),comp1.hmnf.mu/(((0.25comp1.hmnf.l_mix_lim)^2)comp1.hmnf.rho)) Echec dans l'évaluation de l'expression. - Expression: nojac(comp1.hmnf.time_step_inv/comp1.hmnf.locCFL) Echec dans l'évaluation de l'expression. - Expression: d((-comp1.hmnf.rhonojac(comp1.hmnf.time_step_inv/comp1.hmnf.locCFL)(comp1.ep-nojac(comp1.ep))test(comp1.ep))*(dvol),{test@2}) - Caractéristique: Solveur stationnaire 1 (sol1/s1)

I have join a screenshot of my geometry, it's a basic one but I need it. Eventually, my initial value is the mach on x-axis : M=4. The temperature doesn't change during the fluid flow.

1 Reply Last Post 17 Oct 2020, 11:54 GMT-4

Hello Jonathan Pressoir

If you do not hold an on-subscription license, you may find an answer in another Discussion or in the Knowledge Base.

Posted: 2 months ago

Hi Try load ramping. Initially take all values at the least. Like Inlet pressure = 1[atm], T_inlet=200[K], Mach=0. Then find a solution. Add another study and then ramp up any of your parameter in that study. Take the last solution from that study and then ramp up another parameter. Continue like this.

This is how I fixed this up with my analysis. Same trouble I faced for a very long time. Maybe this works for you.