Gen classical

[abdourahmanbarry]

Description

Implemented the second-order classical generator model as a Phasor Dynamics family of components. This model has been used to simulate a 2-bus example and the results outputted to a .csv file. New CMake files have been added as needed, and existing ones have been updated to incorporate the new additions.

This closes #73

Proposed changes

I have added a test case that sets up the system with a consistent DAE solution through the initialize method of the classical generator model and checks that the residual is zero. Another test case has been added to verify that the residuals are correct for other states different from the initialization. Additionally, the CMake files have been updated, and I have built and verified that all configurations work correctly.

Checklist

• Generator model implementation
• Proper CMake build configuration.
• Documentation in a form of README file with model equations.
• Example of usage.
• All tests pass.
• Code compiles cleanly with flags -Wall -Wpedantic -Wconversion -Wextra.
• The new code follows GridKit™ style guidelines.
• There are unit tests for the new code.
• The new code is documented.
• The feature branch is rebased with respect to the target branch.

[abirchfield]

I would move the example to PhasorDynamics. Other than that it looks good.

[lukelowry]

Small suggestion, but for admittance values use capital letters $g\mapsto G$ and $b\mapsto B$

No other suggestions other than documentation, good job sir.

[pelesh]

Quick comment: @abirchfield and @lukelowry suggested GenClassical as the name for this model. See generator documentation.

Can we agree on that name and use it consistently across the GridKit code?

[abdourahmanbarry]

Thank you all for the feedback. I have incorporated all the suggestions. Please let me know if any further changes are needed.

[pelesh]

Very nice work! A few minor things to address, mainly about designing tests.

Also, there are merge conflicts that need to be resolved. The feature branch needs to be rebased with respect to develop.

[pelesh]

A comment would be helpful to explain why Bus-1 is initialized like this. The numbers are not intuitive.

[abdourahmanbarry]

I am not sure I have good explanation for why the numbers are the way they are. I copied that from one of Adam's examples. Maybe if you have good numbers, I can use those instead.

[pelesh]

The other example was for a different machine. I am not sure it applies here. I would try first Vr = 1 and Vi =0 and see if the solver can converge to the steady-state solution from there.

CC @abirchfield

[abdourahmanbarry]

I used Vr = 1 and Vi = 0, and it looks like the solution is approaching the steady-state solution. See the graph below:

[pelesh]

Rebased to the most recent develop.

[shakedregev]

      Start 13: AdjointSens
13/29 Test #13: AdjointSens ......................***Exception: SegFault  0.23 sec

[nkoukpaizan]

• CPU tests (Ubunutu x86_64) / Build gridkit with Spack (16, gridkit@develop +enzyme ^[email protected] ^sundials@develop) (pull_request)

@shakedregev You might want sundials@develop (or a the specific commit) for this. See #118.

[pelesh]

Closing in favor of #127