[clmr90]

Good day for everyone,


My case is the flow simulation of a filter holder (the carcass) where I need to calculate the pressure loss at the end. Sadly I had several problems using different configurations. It started as a way of helping a professor from my uni but now I am taking it as a personal challenge.
First I used a pressure inlet for the entrance of the flow and a mass outflow for the exit (0.1104kg/s as the engine needed). I am entering 98800 pa as operating pressure and locating the point of reference near the entrance (there are some losses from the tubes that conduct to the air to the inlet, obtained experimentally from the people that I am helping). Also, I had turbulence intensity as 5% and 0.066 for the hydraulic diameter. I started using the k-e realizable with enhanced wall functions but I could not get any sign of convergence at all after 400 iterations and the warning of high turbulence viscosity could not dissapear. Interestingly enough my max y+ was about 2421 which I though it was strange so I tried using the other turbulence models such as standard k-e with the same mesh (about 2.4 million cells) and k-omega standard and sst with a more refined one (3.1 million cells). Again, I didn’t get convergence (specially the continuity equation, like a flat salt lake) but I didn´t get any high turbulence viscosity warning or high y+ (about 80 for the std k-e and 50 for the k-omega’s ). My first question is why would that happen with k-e realizable and not with the others.


After that I tried changing my boundary conditions as pressure inlet and pressure outlet with a target mass flow and after 4000 thousand iterations I got a very small convergence with the standard k-e and the k-omega. Finally I changed to mass flow inlet and outflow and with that I got a better convergence but just around 3*10-2 with the continuity equation.
So I am still thinking what to do improve or change. I was thinking of using the RSM since there was a lot of rotation and maybe it can qualify as a swirl flow. Also, since my total pressure never converged to a stable value and always changed (sometimes very rapidly) I was thinking of a transient simulation.



So I hope I can get your feedback in what to do next. You can see the geometry and the flow from the last simulation in the pictures attached.


Thank you