Chemical reactions with FEniCS

[boris-martin]

Description of the case

Added a new tutorial for partitioned advection-diffusion-reaction of chemical species in a incompressible fluid.

• Fluid participant solves the NS equations and sends its velocity to Chemical participant
• Chemical participants computes diffusion, production and reaction of chemical species (A + B -> C with sources of A, B) as well as advection through the fluid, reading velocity from the Fluid participant.

Coupling is unidirectional, serial explicit.

Due to tight time stepping constraints on the fluid flow and high cost of the non-linear chemical reactions, different time scales are used: the fluid solver is subcycling.

This test case was made for showcasing Linear Cell Interpolation data mapping.

Todo-list

• [ ] Writing README
• [x] add run.sh and clean.sh scripts
• [ ] add some measurements: the chemical case keeps track of total quantities of A, B and C species over time (through FEniCS built-in support of integrals)
• [ ] General cleaning, formatting, etc
• [x] Investigate better discretization choices (e.g. more accurate schemes, like mid-point rule for velocities, choose between 2nd order and 1st order elements for chemistry)

[uekerman]

Nice :+1: Alternative names could be channel-chemistry or channel-transport-reaction. It is a channel flow with no-slip walls, right? This would put the case close to the existing channel-transport.

[boris-martin]

Yes it's a channel. I imagine channel-transport-reaction is the most appropriate since it describes what is coupled. On the other hand, no reference to chemistry at all doesn't seem right.

[uekerman]

You could say that "chemistry" is already covered by "reaction"? :grin:

[uekerman]

Just saw right now: PR should point to develop, not master.

[precice-bot]

This pull request has been mentioned on preCICE Forum on Discourse. There might be relevant details there:

https://precice.discourse.group/t/iceberg-ahead-and-new-precice-release-v2-5/1133/1