I am a doubting Thomas

[rmw]

I am not a CFD literate person but I have to depend on those who are as a user of CFD analysis.

However, I am looking a CFD for high pressure air entering a large vessel of a heat exchanger that goes against all of my experience (35 years of crawling around in air and flue gas ducts where there was ash and abrasives and the flow patterns were etched in the metal floors and walls-what was left of them).

Should I take the result as gospel or is there a reason to doubt it?

I don't even know enough about it to tell how it was modeled, so sorry about that.

Any guidance?

rmw

 

[btrueblood]

rmw, trust your gut.

I have used CFD from both ends (feeding the computer, and processing the "output"). When feeding it, I learned at the knee of a PhD who lacked a dozen years of experience compared to me.

I would repeatedly startle him and other users, at being able to guess correctly the general result of a proposed CFD run, e.g. the shape/path of streamlines, locations and strengths of eddies, degree of mixing, etc. etc., based on what I started to call my internal CFD processor. What you and I would probably call "gut feel".

If it doesn't look right to you, question the model - what b.c.'s did the analyst assume, what are the flow rates, what does the grid look like in the region of interest and how does the solution change if the grid is refined in that area. Same as you would for a finite element stress analysis. To me, CFD lags well behind in "user friendliness" (i.e. ability of inexperienced users to generate reasonably accurate results) compared to FEA.

 

[btrueblood]

Mm. Another thing, not clear from your post but hinted at - the ability of CFD to track multi-phase (particulates, droplets, etc.) flows is ... not the most robustly understood part of the art. Particles are generally tracked by secondary subroutines, i.e. the flow is first solved, and its affect on a given particle then determined. At least, that's how I've seen it done...it's been 10 or 15 years since then. The reverse (affect of particles on the flow) is generally ignored; though I've seen treatments for particle conveying (very high particulate to gas ratios) that attempt to do so (and my mind fogged out trying to follow the math, it was a warm sunny day and the coeds were walking by the university library windows...)