`bulk_diffusion` uses the interpolated throat diffusivity for pores as well.

I just pushed a fix (at least I assume it’s a fix 😀) to the dev branch without making a PR. Just wanted to check with you guys if you approve of the idea.

Right now, bulk_diffusion uses the interpolated diffusivity for calculating gp1, gp2, and gpt (pore/throat conductances in a conduit). This approach might be reasonable in the absence of Knudsen effect. Even in the absence of Knudsen effect, this approach could potentially fail (introduce error) for cases where the gradient of properties could be high such as in multiscale networks at the interface between small and big pores. So, I think it’s proper to use the corresponding diffusivity when calculating conductances within a conduit (1/2 pore + throat + 1/2 pore).

Also I added throat_diffusivity as argument to bulk_diffusion method (and inevitably changed the name of diffusivity to pore_diffusivity). This change is to allow directly passing throat diffusivity instead of interpolating it (ex. Knudsen effect).

What do you think?

[jgostick]

No, I fixed this in #898, and it will be merged soon.

[ma-sadeghi]

Yup, confirmed. My bad.