This question is for those of you who use CFD for reaction models. I am working with Fidap 7.63 on a Dec Alpha workstation, modeling a combustion reaction with turbulent fluid flow. For a converged simulation, the sum of all of the species concentrations for a specific element (actually, for a lot of elements) was greater than 1 (about 1.3). The manuals for this program states that species concentrations are expressed as mass fractions. However, I know that realistically a sum of all mass fractions cannot be greater than 1. Technical suppart at Fidap says that that having a sum greater than 1 is possible (for the software). My question is this: what do the results mean, realistically? How can the species' concentration values be interpreted in order to make sense?

Thanks,

Lisa Eastep

(1). You have to obtain the so-called mesh-independent solution first, before you can evaluate the converged solution. (2). Pick some recently published CFD papers from ASME/Journal of Fluid Engineering or AIAA Journal, and study how other researchers are doing in this area first. (3). Remember that, if you are using a commercial code, there is a very long way to go. (4). To learn CFD, you must learn how to write your own code. Then you will be able to ask the right question. (5). Only the code vendor knows what is in the code, and only you know what you are doing with the code. It is like asking " How come my phone bill last month is so high? " There are only three parties involved: the phone company, the phone and you. (6). The problem has nothing to do with CFD at all, it is the relationship between the code vendor, the code and you. Very simple. To become a CFD problem, you have to be able to obtain at least the mesh-independent solution first.

Dear Lisa,

You are correct in stating that the sum of the mass fractions should be 1. However, it is well-known that the finite element method, upon which FIDAP is based, can produce "wiggles" in areas where there are strong gradients and the mesh is under-resolved. These wiggles are undershoots and overshoots of the actual solution and hence you might see mass fractions of -0.01 or 1.03, etc. A value of 1.3 sounds a little excessive.

If you'll send me your .FDREAD file, I'd be happy to have someone take another look at it.

Regards,

Eric Grald, Fluent Inc.

Hello, Lisa.

It means that the approximation errors are very big. Try to use finest mesh and less Courant number.

Alexey.

Dear Lisa,

Forward and reverse reaction rates are functions of temperatures. When you add these reaction rates in turbulence model, it is obvious that the results may act wild because reaction rates are wild. Many computational techniques are being developed now-a-days to treat detailed chemistry in reactive flow calculations. Theoretically, the sum of mass fractions should be equal to 1. Most reactions rates are obtained experimentally in the laboratory so that there are chances of error in calculations. Species concentrations are obtained using these reaction rates. You are getting the accumulation of errors. When reactions rates are added to the equations, it is hard to converge the equations as much as required to get better results. You are saying that you are working on a converged problem. I would recommend you to try to use smaller convergence criteria. It may be the reason why you have all these errors. There is nothing wrong with the software. Finally, I would say that no turbulence model can be totally trusted because there does not exist any unique turbulence model.