[RebelLion]

I'm attempting to simulate methane leaks into air and am struggling to finalize my choice of boundary conditions. I keep having issues with methane being partially trapped by the intended atmospheric/open boundary.

I'm currently using the transient incompressible two fluid solver, twoLiquidMixingFoam. I'd like to have a floor at the bottom, the sides and top open to atmosphere and a pipe sticking into the domain on one side.

What BCs would you recommend for p_rgh, U, and alpha at the pipe walls, inlet patch, floor, and atmosphere patches?

Thanks in advance for your help

[wayne14]

Originally Posted by RebelLion

I'm attempting to simulate methane leaks into air and am struggling to finalize my choice of boundary conditions. I keep having issues with methane being partially trapped by the intended atmospheric/open boundary.

I'm currently using the transient incompressible two fluid solver, twoLiquidMixingFoam. I'd like to have a floor at the bottom, the sides and top open to atmosphere and a pipe sticking into the domain on one side.

What BCs would you recommend for p_rgh, U, and alpha at the pipe walls, inlet patch, floor, and atmosphere patches?

Thanks in advance for your help

Hi, twoLiquidMixingFoam is for multiphase flow: 'Solver for mixing 2 incompressible fluids.' Are you sure it is suitable for simulating methane leaks into air?

Regards,
Yan

[RebelLion]

Hi Yan,

Thanks for your post.

You are right that twoLiquidMixingFoam is not ideal. I think it should be a fair approximation, ignoring any effects due to compressibility and should serve our purpose. From what I understand, if I want a compressible solver I should go to reactingFoam and turn off the chemistry. For now I'm using methane/air values for viscosity and diffusion coeffs and it's giving reasonable results, until my methane plume hits the boundary.

My problem is that any backflows are not allowed to exit the domain, they instead only appear to loose the normal component of their momentum and then continue via their tangential component, trapped within the boundaries.

Any other thoughts on this one?

Thanks!