Use a pressure solution as input for the advection-diffusion algorithm

[MTJKoning]

Hi OpenPNM team,

I am simulating an packed bed reactor in terms of mass transport. For that i am using a pore network model, and have looked into OpenPNM before.

Now i wanted to implement a non-linear pressure-flow solution, which OpenPNM could not possibly provide, so i made my own solver, and now i have the non-linear solution for the pressure-flow field throughout all my pores. I want to use this as input for the advection-diffusion or the reactive-transport algorithm in OpenPNM, but I can't find anywhere how the pressure is actually used in the solving method. I see how in the settings of the advection-diffusion algorithm the pore.pressure is used, but if i look at the advection diffusion equation that is solved for, it used a velocity (calculated from the flow rate) and not pressure directly. Is it maybe so that upon running the advection diffusion algorithm, it calls the stokes algorithm and calculates a linear flow field from that?

I hope you could explain to me how I can use my own non-linear pressure field as input for the advection diffusion in OpenPNM.

Kind regards,

Menno Koning

[jgostick]

Hi @MTJKoning
Openpnm uses the pressure field and the conduit hydraulic conductance values to find the velocity in each throat, then uses those to compute the direction dependent 'advective-diffusive' conductance values. These are then used inside the AdvectionDiffusion algorithm. So basically AdvectionDiffusion is just a diffusion simulation where the conduit conductances are adjusted by the local fluid velocity in each throat. If can compute the velocity values for each throat (which I assume you can since you know the pressure field), then you can also link into this workflow, but you'll have to create your own pore-scale model to compute the advective-diffusive conductance. Our version is here. You'll want to focus on lines 65 and 66 as this is where the velocity/flow is calculated.

[ma-sadeghi]

@jgostick It might be a good idea for the algorithm to optionally accept conduit flow rates. It's aligned with OO principles, i.e., separation of responsibilities. AdvectionDiffusion implicitly assumes StokesFlow, but it doesn't have to.