[Roger9318]

Hi everyone !!!



According with the literature, the algorithm PISO is one of the most suitable algorithm for transient analysis. I found many publications related with this topic. these publications helped me a lot (really really helped me). However I'm still working on that. The PISO algorithm has been already developed in an in-house FORTRAN code based on the finite volume technique. However the program diverge also I have doubts and that´s the reason for this post. Hopefully you can (in some way) help me.



Which are the main difference between the iterative PISO and noniterative PISO? I have an idea and I represented in the attached flow charts images.


flow chart iterative piso.jpg

NONiterative.jpg


Another of my concerns is related with the following term, which appears in the source term of the second pressure correction equation and in the correction of velocities. Is it correct how I represented?

relacion.jpg


The PISO algorithm is implemented in a forzed convection(channel) problem. After that, I’m going to develop in a problem with natural convection but before do that the PISO algorithm needs to be done


How did you do to reach the convergence?

Accordig with your experience which are the better parameters (Δt-step time, under relaxation factors, solver -LbL-x,lbl-y lbl adi, etc..) to reach the convergence? I know it depends on the problem but which were the best for you?


Thank you in advanced

Best regards,

[LuckyTran]

In the non-iterative PISO, you choose a time-step small enough (i.e. Courant number << 1) such that you always converge to your tolerance criteria. It's the same as the iterative PISO if you were to limit the number of iterations to 1. The difference is whether or not you are converged in 1 iteration.

You can imagine that with a large enough time-step, I run the non-iterative PISO algorithm, my solution is unstable, diverges, and doesn't even converge at all.

In my opinion, if you are using PISO it should be the non-iterative PISO and should not have any underrelaxation factors. If you need to do either, then an iterative SIMPLE is a better choice. Note* you can also have a non-iterative SIMPLE the same way as you have non-iterative PISO, but you learn quickly why no one uses such an algorithm.