Divide by zero warnings in costs.py

[clmould]

Occurs when running integration or regression tests.

regression/test_process_input_files.py::test_input_file[large_tokamak_eval]
regression/test_process_input_files.py::test_input_file[st_regression]
regression/test_process_input_files.py::test_input_file[large_tokamak_nof]
regression/test_process_input_files.py::test_input_file[low_aspect_ratio_DEMO]
process/costs.py:3028: RuntimeWarning: invalid value encountered in sqrt
    annoam = cost_variables.ucoam[cost_variables.lsa - 1] * np.sqrt(

regression/test_process_input_files.py::test_input_file[large_tokamak_eval]
regression/test_process_input_files.py::test_input_file[st_regression]
regression/test_process_input_files.py::test_input_file[large_tokamak_nof]
regression/test_process_input_files.py::test_input_file[low_aspect_ratio_DEMO]
process/costs.py:3099: RuntimeWarning: invalid value encountered in sqrt
    annwst = cost_variables.ucwst[cost_variables.lsa - 1] * np.sqrt(

[OceanNuclear]

Script to reproduce on linux:

cd tests/regression/
ipython

run test_process_input_files.py
tracked_regression_test_assets = RegressionTestAssetCollector()
Path("/tmp/manual-pytest/").mkdir(exist_ok=True)
i = 0
tmp_path = Path(f"/tmp/manual-pytest/pytest-{i}/")
for file_root in ["large_tokamak_eval", "st_regression", "large_tokamak_nof", "low_aspect_ratio_DEMO"]:
    while tmp_path.exists():
        i += 1
        tmp_path = Path(f"/tmp/manual-pytest/pytest-{i}/")
    tmp_path.mkdir(exist_ok=False)
    input_file = Path("input_files") / (file_root + ".IN.DAT")
    test_input_file(
        input_file,
        tmp_path,
        "vmcon",
        tracked_regression_test_assets,
        reg_tolerance=5.0,
        opt_params_only=False,
    )

[OceanNuclear]

Diagnosis: heat_transport_variables.p_plant_electric_net_mw can occasionally take negative value during the optimization process. (In fact, each input file's run leads to exactly two iterations where heat_transport_variables.p_plant_electric_net_mw<0:

input file used	offending value 1	offending value 2
large_tokamak_eval.IN.DAT	-515845.9632885829	-515845.9632885829
st_regression.IN.DAT	-13.482151687036321	-13.482151687036321
large_tokamak_nof.IN.DAT	-172492.59287270543	-172492.59287270543
low_aspect_ratio_DEMO.IN.DAT	-154429472.70250472	-154429472.70250472

)
A simple fix is proposed in #4065