Hi, all,

I am trying to predict the results of a tracer test with CFX5.6(the tank I simulatd has 5 baffles, tracer was pulse injected in the inlet within 10 s). Then I found that there are two peaks in the tracer curve. Is this normal or not. If it's normal, how to explain it? If not, what's the possible reason of that?

BTW: Generally, how to improve the accuracy of a tracer simulation by CFX?

Regards!

Jim

Jim,

See my previous post for Bacterial Dose. Same advice applies. Be sure to use the second order differencing in both time and space. One or the other won't do it.

I think your double peak can be explained if some of the flow short circuits (heads straight for the outlet) and some of it gets hung up behind the baffles and released with a time delay from the first "slug" of material. Animate your isosurfaces of tracer material to see if this is really what is happening.

Jeff

Hi, Jeff

Yes, I did see the tracer mass seperated into two parts in the animation result.

But I never saw any literature reported double peak tracer curve(with similar geometry). So I will try to use the second order differencing in both time and space as you suggested. Could you please give more detailed information on how to set the second order differencing in both time and space in CFX5.6?

Thanks a lot for your advice!

Regards!

Jim

Spacial second order is the default (high resolution scheme) as available on the solver control panel in Pre.

Second order in time requires a CCL command to set "Second Order Upwind Scheme" (I think this is what it's called). I can't recall exactly the syntax, but if you search the rules file, you'll find the proper syntax and where to put it in your CCL. Your support person should be able to help you find this as well.

Jeff

Jeff,

Thanks a lot, I will try that!

Best regards!

Jim

If you are using CFX-5.6, it is available in the GUI.

-Robin