There is already UDF available for Linearized mass transfer for Fluent.
How can use such UDF in CFX?
Is it readily available somewhere or do I have to convert it into a CEL file?, if so can somebody please guide me how to do that?
No u cant do that. CEL is different format. To my knowledge there is no such tool
Yeah, so maybe I will have to write the code in UDF as Expressions in CEL manually.
The formulas have be converted manually to CEL language for corresponding source terms. But it should not be a too difficult task. Source term linearization is available for CEL source terms, so you have to just put in the appropriate formulas in CFX CCL language.
Regards,
Dr. Th. Frank.
As, Dr. Thomas Frank correctly wrote, it is not a difficult task. Only for aggregate some details to his answer:
CFX accepts FORTRAN subroutines also - such an UDF in Fluent (written in C). It is a early feature when some functions or calculations were not possible through CCL, and this features is aailable if needed. But with new releases , more and more functions and calculations has been added,. e.g . an IF statement is possible.
An advice: be aware with DIMENSIONS, instead FLUENT that works with numbers only in the UDF's ( it it presupposes that all numbers are in the correct unities and scales) , CFX work by default with unities in CFX-Pre. The sources must have correct dimension, otherwise a error message will appear. Some functions are non-dimensional and work with numbers as argument. e.g. exp() , step()., e.g. exp(b*t), t, is time in [s] , hence, b must have [s^-1]. Sometimes we need divide for 1[dimension] to "clean" the dimension, e.g. step( (Ho -z)/1[m]) , Ho is a high , for example.
Other: avoid zeros! It is a good practice. If some variable can be zero, use some function such as max() or min() and a very low value such as 1e-20, 1e-15, etc.,e.g x=1/a , it is better if you do x =1.0/ max(a,1e-20) which will avoid a division by zero.
With these cares you will be well successful.
Regards!
- Badges
- Science topic