Replace ClosedLoopHeartPulmonary with individual blocks

[mrp089]

Problem

We currently have a closed-loop block ClosedLoopHeartPulmonary that combines four heart chambers and valves with a pulmonary and coronary circulation. @menon-karthik mentioned that it might become unstable for certain parameter choices and not easily adaptable to the monolithic structure.

Solution

Have a test case using individual blocks for chambers, valves, and vessels that reproduces the behavior of ClosedLoopHeartPulmonary (that should be implemented already). It would be nice to also add some documentation and plots on this closed-loop model.

Additional context

A challenge might be the handling of valves. The ClosedLoopHeartPulmonary block has four "hard" valves that explicitly set flows to zero when closed. This behavior might be harder to replicate with individual valve blocks.

Code of Conduct

• I agree to follow this project's Code of Conduct and Contributing Guidelines

[mrp089]

@claude and I implemented the decomposed version in this branch. Aortic flow and pressure have a difference of <0.1%.

These are the PV loops:

The valve flows actually look more stable with the tanh valves:

@menon-karthik, are there more places where we need to change the code to remove ClosedLoopHeartPulmonary? Does it make sense to remove closed_loop_blocks entirely? Or should we introduce a "compound block" that combines several existing blocks?

[menon-karthik]

This is great!

I think removing closed_loop_blocks entirely might be the way to go. Compound block would be redundant because, in theory, any 0D model is a compound block. It will make the distinction between internal variables and other variables confusing because normal variables will become internal variables in a "compound block".

For now, I would suggest keeping the older closed loop stuff. Partly because I have an existing project that we have not transitioned away from this yet. I will work on removing it after we are done using it.

Does this sound good?

[mrp089]

Sounds good! With #219, we have a decomposed closed-loop model to work with.