Best Way to Set Up a Diffusion Problem
 

I am trying to run a simulation where I have ideal air diffusing into ideal helium. I have set up a variable composition mixture and initialized one area of the simulation to have only air and the other to have only helium. When I run the simulation I am not getting accurate results using a well proven kinematic diffusion function. Does anyone have any insight on a better approach to this problem? Thanks.

Have you done the standard checks for simulation accuracy? https://www.cfd-online.com/Wiki/Ansy..._inaccurate.3F

You define the molecular diffusivity. The program will also add in turbulent diffusivity from fluid motion if you are using a turbulence model. Is your flow laminar? Are you using a laminar model or turbulent? If turbulent, additional diffusion is added which possibly shouldn't be there?
Otherwise, Check the FAQs as Glenn suggests. there could be numerical diffusion, convergence not good enough, etc. I've done many diffusion models and I know it works properly and matches the complementary error function equation for diffusion, where you set 50% at the initial interface location.

Thanks, I have looked at the FAQs but haven't found anything that really helps. My flow is laminar as well. I will try looking to see if there is some turbulence variable which may be influencing the results. If I fix it, I'll let you know.