hello,

I'm looking for information on the relative performance characterestics of the SIMPLE-like algorithms when solving compressible flows at all flow speeds.

Is there any article reporting this performance characteristics?

I've got an article from Moulkalled & Darwish, published in the Numerical Heat Transfer in 2000, where a number of SIMPLE-like algorithms are described but they don't report any performance characteristics.

As a general statment (for what its worth):

* pressure based schemes are better than kludged density based schemes at low speeds (unkludged density based schemes do not work).

* pressure based and density based schemes can be made to perform about the same in the transonic/less than Mach 2 region.

* density based schemes work at high Mach numbers but pressure based do not. I am not aware of any kludged hypersonic pressure based schemes but there may be some.

The distinction above is blurring somewhat in modern strongly coupled schemes.

what do you mean by density based schemes. Is it coupled solvers. please answer

Density based schemes are ones where the continuity equation is treated directly as a transport equation for density instead of using a pressure-related equation derived from the momentum equation with the continuity constraint worked in. It is the classical approach for compressible flows.

Hello,

when I asked about relative performance characteristics of SIMPLE-like algorithms I meant behaviour of the different SIMPLE-like algorithms (SIMPLE, SIMPLEC, SIMPLEX, SIMPLEST... and PISO,...) when treating compressible flows.

These algoritms, if I'm not wrong, are all pressure based. As I'm not, at present, working at high Ma numbers, 0,3-2.0 would be my target range, I guess these are the methods that suits me the best.

Apologies, I can see I answered the wrong question.

All comparisons I am aware of are for incompressible flow so no help with the numbers. The little I can add is:

* If you are not dropping below 0.3 Mach number and want to use a non-coupled sequential scheme then I suspect most people would use a density based method since they are generally simpler and more compact.

* My suspicion (based on no hard evidence) is that the ranking of incompressible pressure correction methods is not going to be preserved for compressible flows. My reasoning stems from the observation that one cannot use large time steps for compressible flow like one can for incompressilbe flow because the coupling between the variables is too strong.