I am trying to model oldroyd-B fluids in comsol, but there is some problems. I try to add extra stresses due to non-Newtonian fluid to N-S equations but this will influence my convergence heavily. nobody has tried to model this in COMSOL? without adding this extra stress in N_S equations the convergence is quite good and is matching the analytical results. I have also read COMSOL manual PDF which models this but that PDF also ignores adding extra stresses! and i my self think that simulations was wrong.
I think you should go for the weak form, and working on the convergence for this type of formulation. Adding extra terms in the NS module (I think) is not a good idea, since the corresponding solver is tuned for that equation. This could be a good starting point
https://www.researchgate.net/profile/WB_Zimmerman/publication/267946264_Stabilised_Finite_Element_Modelling_of_Oldroyd-B_Viscoelastic_Flows/links/54bb98ec0cf29e0cb04bdfdc/Stabilised-Finite-Element-Modelling-of-Oldroyd-B-Viscoelastic-Flows.pdf
F
Here is another way to do VE flow in COMSOL from Professor Bruce Finlayson, emeritus professor at Universtiy of Washington. It uses DEVSS. I did my dissertaion on VE flow modeling with finite elements and Finlayson was my advisor. I am currently working on VE flow using a log conformation formulation for free surface flows.
thank you Mr. rekha. it seems that this problem for a simple channel do not converge for Weissenberg numbers than wi=~1.
- Badges
- Science topic