[dpitz]

Hello,

I've been using Polyflow to perform simulations of a PTT fluid in axisymmetric geometries. My problem is that Polyflow has only the exponential PTT model implemented, and I would like to perform simulations using the linear PTT model, in order to compare my results with some data that I have from some authors.

My first attempt was to create an UDF for the eps parameter, in order to eliminate the exponential term from the constitutive equation of the exponential PTT model.

Here is the line from my UDF where I rewrite the eps parameter:

(bind ?R (* ?n0 (/ 1 (* ?lbd ?tr)) (log (+ 1 (* ?lbd ?reps ?tr (/ 1 ?n0))))))

where R is the modified eps paramater (that's the parameter responsible for "transforming" the exponential ptt model into the linear ptt model), n0 is the zero shear-rate viscosity, lbd is the relaxation time, reps is the "real" eps parameter and tr is the trace of the stress tensor.

I've tested my UDF file in qa.exe, and it works fine.
Unfortunately, I couldn't obtain any good results from my simulations using this UDF; most of them diverge, and some converge to a wrong result (I have only simulated cases of low Reynolds and Deborah numbers).
I think the difficulty to obtain convergency is due to the presence of the log term, but I couldn't find another way to do what I need.

I hope someone have an alternative suggestion to implement the linear PTT model in Polyflow, or an explanation for the instability of the numerical solutions using the UDF that I've created.

Thanks in advance!

[Amir]

Hi,
I also want to simulate viscoelastic flow . I think OpenFOAM is more powerful than polyflow isn't it ?