Add RCR-Windkessel model

[LasNikas]

Lumped parameter model to simulate vascular systems

[Copilot]

Pull Request Overview

This PR introduces the RCR-Windkessel lumped parameter model for simulating vascular systems at open boundaries. The model captures the relationship between pressure and flow in pulsatile systems using three components: characteristic resistance (R1), peripheral resistance (R2), and compliance (C).

Key changes:

• Added RCRWindkesselModel struct and associated calculation methods
• Integrated pressure model support into the OpenBoundarySystem and BoundaryZone infrastructure
• Added comprehensive validation tests using ferret vascular system data

Reviewed Changes

Copilot reviewed 13 out of 13 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
src/schemes/boundary/open_boundary/pressure_model.jl	Core implementation of the RCR-Windkessel model including pressure and flow rate calculations
src/schemes/boundary/open_boundary/system.jl	Extended OpenBoundarySystem to support pressure models with new fields and initialization logic
src/schemes/boundary/open_boundary/boundary_zones.jl	Added pressure_model field to BoundaryZone with validation logic
src/schemes/boundary/open_boundary/dynamical_pressure.jl	Modified pressure reference to support pressure model computations
src/schemes/boundary/open_boundary/mirroring.jl	Updated boundary quantity handling to work with pressure models
src/callbacks/update.jl	Modified update callback to pass timestep to open boundary updates
test/schemes/boundary/open_boundary/pressure_model.jl	Comprehensive tests including pulsatile flow validation against ODE solution
docs/src/systems/boundary.md	Added documentation section for pressure models
docs/src/refs.bib	Added references for Gasser (2021) and Westerhof (2008)
Other files	Minor updates for file inclusion and exports

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[LasNikas]

/run-gpu-tests

[efaulhaber]

I don't understand why this has to be so complicated.
It looks like this model can only be used if reference_pressure = nothing is passed to the BoundaryZone, so it is essentially just a complex function for reference_pressure. It is also only used where a reference_pressure function would be used (imposed_pressure) plus it is updated. These are the only two occurrences.

So why do you need this extra struct and extra field in the boundary zone? Why can't you just make this a functor? Basically, just add a field pressure to the struct and make it a functor as

function (model::RCRWindkesselModel)(x, t)
    return model.pressure[]
end

And then the only thing you have to change in the open boundary system and boundary zones files is that you add an update. This update function updates reference_pressure for each boundary zone. If it's ::RCRWindkesselModel, do the update, otherwise do nothing.

[LasNikas]

I'm not sure I understood correctly. What happens if some BoundaryZones use a pressure_model while others have a prescribed pressure (as a function or scalar)?
Is this then type stable?

[LasNikas]

Also, IIRC, storing RefValues in a GPU array is not allowed.

[efaulhaber]

I'm confused. What do Zhang et al. have to do with this? You listed different references.

[LasNikas]

When I implemented this, I based it on this literature because it was also mentioned in the SPH context. It's a simple ODE. I don't really care which literature I cite here...
The other references are purely in a biomechanical context where Zhang is in a SPH context.

[efaulhaber]

Why Euler? You could just use the default, which is an adaptive method that doesn't even need a time step.

[LasNikas]

and then? It doesn't makes a difference here, since I only needed this for the pressure reference values