pydynamicestimator.tests.test_inverter_lcl_static

Phase 5: validate the quasi-static LCL_static filter (filter quantities as device-private algebraics) against the dynamic LCL filter.

LCL_static is NOT byte-identical to dynamic LCL – it is a singular-perturbation reduction that removes the fast LCL dynamics. So the validation is:

1. Self-consistency / machinery (rigorous): LCL_static runs as a DAE (line_dyn=False), initializes, and is exactly reproducible. This proves the algebraic-port plumbing (var_sym + DAE integration + the init routing the algebraic filter values into dae.yinit) is correct, with no reference to the dynamic model. LCL_static supports BOTH init methods (sequential default and joint) and they reach the same operating point – checked here.

2. Quasi-static reduction (physics): comparing slow quantities between LCL_static and dynamic LCL (same static network), they are machine-identical BEFORE a disturbance and re-converge AFTER the fast transient settles; they differ only during the fast transient (a few % for a fault, negligible for a gentle load step). Validated for both a bus fault and a small load step.

Attributes

TS
T_END
SLOW

Functions

_build_fixture(root, filter_name, disturbance)	Copy the inverter fixture into root, select filter_name on every
_run(root, filter_name, disturbance)
_dev_state(sim, idx, state)	Trajectory of a named state for the device with the given idx.
test_lcl_static_self_consistent(tmp_path)	LCL_static runs as a DAE, initializes (sequential by default), and is
test_lcl_static_both_init_methods_agree(tmp_path, ...)	LCL_static initializes correctly under BOTH the sequential init (routing the
test_lcl_static_quasi_static_reduction(tmp_path, ...)	The slow quantities of LCL_static and dynamic LCL (same static network)

Module Contents

pydynamicestimator.tests.test_inverter_lcl_static.TS = 0.005

pydynamicestimator.tests.test_inverter_lcl_static.T_END = 5.0

pydynamicestimator.tests.test_inverter_lcl_static.SLOW = [('GFMI4', 'Pc_tilde'), ('GFMI4', 'delta_c'), ('GFMI4', 'Qc_tilde'), ('GFLI7', 'Pc_tilde'),...

pydynamicestimator.tests.test_inverter_lcl_static._build_fixture(root, filter_name, disturbance)

Copy the inverter fixture into root, select filter_name on every converter, and set the disturbance ("fault" keeps the bus fault at t=4.0; "load" replaces it with a small load step at t=2.0 on a load bus).

pydynamicestimator.tests.test_inverter_lcl_static._run(root, filter_name, disturbance)

pydynamicestimator.tests.test_inverter_lcl_static._dev_state(sim, idx, state)

Trajectory of a named state for the device with the given idx.

pydynamicestimator.tests.test_inverter_lcl_static.test_lcl_static_self_consistent(tmp_path)

LCL_static runs as a DAE, initializes (sequential by default), and is exactly reproducible (the algebraic-port machinery is correct).

pydynamicestimator.tests.test_inverter_lcl_static.test_lcl_static_both_init_methods_agree(tmp_path, monkeypatch)

LCL_static initializes correctly under BOTH the sequential init (routing the algebraic filter values into dae.yinit) and the joint init (solving them as private-algebraic unknowns); both reach the same operating point.

pydynamicestimator.tests.test_inverter_lcl_static.test_lcl_static_quasi_static_reduction(tmp_path, disturbance, t_dist, after_atol)

The slow quantities of LCL_static and dynamic LCL (same static network) are machine-identical before the disturbance and re-converge after the fast transient settles – they differ only during the fast transient.