[CoG]

Hello everyone,

I have a problem about simulating a filter with a pressure loss boundary condition (use pressure outlet as my boundary condition), a expectable pressure loss is about 8 to 14 psi.
My model having 4.45E+6 cells is meshed by Patran. I can simulate it with 10 psi pressure loss (inlet 10 psi , outlet 0 psi), but get a floating point exception: [overflow] at first few iterations with 11 psi pressure loss (inlet 11psi, outlet 0psi). However, i get the same result when i set inlet and outlet for 11psi and 5psi. It seems that i can not set a pressure more than 11psi.
With another model having 9.94E+6 cells meshed with STAR-CCM+ 5.02, I get the same problem as i set a pressure over than 0.2 psi. :-(
All settings are default values, is there any setting i can try to simulate with a higher pressure ?



Thank you, all!!
CoG

[Maddin]

Ramp the pressure...

[CoG]

Thank you for your answer!!

It works now.

[CoG]

hello,

I set under-relaxaiton factors of pressure & velocity small during first several iterations , then set them for default value. The simulation ran expectably about 320 iteration and all residuals were belows 1E-4, but residuals went up after 350 iterations. All residuals were between 1E+1 and 1E-2 at 6xx iteration.

I have another problem about floating point exception [Invalid operation] now. :-(


I try to ramp the factor, but it doesn't work. It there another way?



CoG

[CoG]

hello,

I have a question about initial condition of velocity and pressure outlet boundary conditions. I simulate a steady pipe flow through a filter with pressure outlet and notice that the pressure contains velocity energy. If I set a velocity initial condition, would it help the simulation converge ?

thanks, CoG.