any comments about EFD v5?
 

has somebody already compared EFDv5 with other cfd codes like StarCD or Fluent? What are the results of the comparison?

Re: any comments about EFD v5?
 

We have done some testing (v4.3..... something) in the past and compared it with Fluent and CFX. It is plug and play and you can get results but but but but

- numerical scheme? It is almost the same as upwind. No secon order scheme which is required for every swirling flow. And which flow is not swirling...... - Turbulence models? Only k-e. Not sufficient. - geometrical models with a lot of detail? Impossible - multiphase? not available - radiation? not available

A lot of drawbacks. I was not positive. Haven't continued the trial....

Astrid

Re: any comments about EFD v5?
 

OK,I am regret to hear that Astrid get such conclusion :( Actually in EFD.V5,the second-order upwind approximations of fluxes are based on the implicitly treated modified Leonard's QUICK approximations and the Total Variation Diminishing (TVD) method. EFD V5 is suitable for engineering,so I think k-e is ok. And because EFD.V5 is fully embended in Catia V5,so there is no a problem with any detail geometry if you know how to use Catia V5. And Astrid is right that EFD can not handle multiphase. Radiation is ok and it adapts Monte-Caro method which is accurate enough :)

I like EFD.

Re: any comments about EFD v5?
 

Hi Jieremy,

It is good to hear that EFD has improved. We did some testing and it was not in favour of EFD. I would like to see some comparison with experiments of swirling flow. Is that possible?

k-e is ok for a lot of applications but not for heating of water which actually is a very common application. In water Pr=7 which implies that your thermal boundary is much thinner. To solve this correctly at a heated wall you need a very fine grid at the wall. And then k-e is not vallid anymore, and you will be pretty weird results. Your heat transfercoeficient can be a factor of 4 off. Your apparatus can become a factor of 4 times too small. Other packages have suitable models to solve this complex problem. Does EFD have implemented sophisticated things to overcome this?

I agree that engineers won't mind and that they use the results as they come, but I think they need to know these kind of pitfalls. They should not take the results as they come.

Astrid