[rahul_jayakumar]

Hi,
I would like to model contaminant dispersion in air in a rectangular domain using the passive scalar approach in Fluent. For this I am using a user-defined scalar(UDS). The flow field of air in the domain is already converged. I would like to inject a contaminant like ethylene with a different density than air.

The contaminant is being injected into the domain at a specified mass flow rate (say A kg/s) and mass fraction of 1 through a small hole on the bottom wall. I have also modeled the flow field through this inlet using the mass-flow-inlet bc. For the UDS, the inlet boundary is set to have a constant value of 1 corresponding to the mass fraction of contaminant. Whereas the air inlet,outlet and walls are given with specified flux of 0. The material diffusivity takes into account both laminar and turbulent diffusivity, based on the turbulent viscosity and Schmidt no of 0.7. The objective is to analyse the spread of the contaminant similar to a tracer injection.

My current problem is that I am not able to model the contaminant transport through the domain. After going through several forum discussions(on cfd-online and ansys forums) and literature for similar problems, I had implemented a source term for the UDS equation equal to the density of the fluid. Since most of these discussions were about computing the age of air(in air) this source was set to density of air. But since the density of the contaminant is different, should the source term be modified to the mixture density using the species?

Also, when the contaminant is injected at a specified mass flow rate, is it right to model the flow though the inlet while computing the flow field and using a constant value of concentration at the inlet or should the mass flow rate be incorporated into the concentration value at inlet itself?

Thanks.