Flash HP/VT algorithms question

[Johan]

Hello,

I am impressed by the way you offer us the possibility to look into your program and especially access to your source codes.

However, i might have some questions about the algorithms you used for flash calculations, especially the boston-britt algorithm and the boston-fournier algorithm.

Is it the other ways to calculate HP and VT flashes without using first a calculation with the PT flash ? Do you have some litterature references ?

Thanks in advance,

Johan

[Daniel Medeiros]

Hi Johan,

You mean during a simulation, using a Material Stream? If yes, you can specify the flash type in the property grid. You can also do different flash calcs using the Excel interface, or even use the thermo library and build a program to do the calcs.

I think that in the Technical Manual you can find the literature references for the flash algorithms.

The Nested Loops algorithm is based on the code available in this book -> http://www.amazon.com/Introduction-Chemical-Engineering-Thermodynamics-Mcgraw-Hill/dp/0073104450

The IO flash algorithms are based on the original papers by Boston/Britt and Boston/Fourier.

The rest of the algorithms are my own, and the phase stability code is based on a procedure by Michael Michelsen -> http://www.tie-tech.com/more.php?id=32_0_1_0_M13

Regards,
Daniel

[Johan]

thanks for the references.

What i also want to know is that if there is any algorithm/procedure for HP flash (VT flash as well) in which you don't need to use the PT flash.

i will start to look your codes.

[Daniel Medeiros]

The flash algorithms are independent, there's a procedure for each one of them (TP, TV, PV, PH, PS) in the corresponding code files.

[Johan]

Have you take a look of this algorithm ? : http://kmlinux.fjfi.cvut.cz/~mikysjir/papers/FPE-2013-VTFlash.pdf

Let me know,

Johan

[Daniel Medeiros]

Thanks Johan, I'm taking a look at it right now.

Daniel

[Johan]

It is a "new" way to formulate an efficient VT algorithm using the same logic as the gibbs energy minimization.

I'm also looking for a PH Algorithm and especially an efficient way to maximize the entropy (function of H and P)

[silverShaded]

The best way i've found so far for flash algorithms is to modify (ie simplify) Russels column algorithm to work for a single stage. Its very fast and the same algorithm with a few options built in will solve any combination of PTESQ and solve adiabatic flashes with heat loss as a constraint. Works great and very easy to program in.

[Daniel Medeiros]

The main problem with the IO algorithm is the R/S convergence loop. It is the single source of instability of the convergence process. Fortunately I was able to make it a little more stable by using a minimization approach. Newton is not suitable here.

The current implementation of the Nested Loops algorithm is, in my opinion, much more stable and reliable and work nicely in almost all cases, even for difficult three-phase calculations.