|
STAR
Cal STAR — firmware, DAQ server, and engine design
|
Variables | |
| ROOT = Path(__file__).resolve().parents[1] | |
| float | PSI = 6894.757 |
Layer 1 search mechanics and sign-off: the design that is scored is the design that is shipped. L1-01 validation never raises tank pressure or swaps the candidate to make a design "solve". L1-04 the derived throat is solved to tolerance, independent of where CMA proposed it. L1-05 sign-off re-checks every declared geometric limit at face value. L1-07 the block stage searches no derived DOF and uses per-coordinate scaling. L1-08 valley escape fires only on the infeasible plateau, not on a shaping-term floor. DEF-09 the chamber OD floor follows the smallest reachable throat, not the envelope cap.
|
protected |
Run Layer-1 setup up to the worker pool and return (bounds, constants).
Definition at line 170 of file test_layer1_search_and_signoff.py.
Referenced by test_od_floor_admits_the_shape_band_at_low_thrust().
| test_layer1_search_and_signoff.test_a_one_dimensional_block_does_not_crash_cma | ( | ) |
Dropping the derived DOFs can leave a block with one coordinate; CMA_stds on a 1-D problem raised IndexError inside cma (seen on configs/ethalox_6500N.yaml, seeds 3 and 4).
Definition at line 211 of file test_layer1_search_and_signoff.py.
| test_layer1_search_and_signoff.test_blocks_search_no_derived_dof_and_scale_per_coordinate | ( | ) |
L1-07: blocks ran isotropic CMA (sigma = 0.2 x median span) over coordinates whose ranges differ by 1e5 and searched the solved throat, eps and fuel pitch freely.
Definition at line 127 of file test_layer1_search_and_signoff.py.
| test_layer1_search_and_signoff.test_derived_throat_is_independent_of_the_proposed_start | ( | ) |
L1-04: two secant steps from the proposal left +0.4..0.9 % thrust depending on the start.
Definition at line 27 of file test_layer1_search_and_signoff.py.
References max.
| test_layer1_search_and_signoff.test_element_count_scan_crosses_basins_along_constant_injector_area | ( | ) |
Seeds ended at n = 5, 5, 7, 13 with objectives 20 % apart: at fixed jet diameters each count is a different injector, so an integer CMA step never crossed between them. Scored along constant flow area and pitch circle, every count is reachable in one batch.
Definition at line 225 of file test_layer1_search_and_signoff.py.
| test_layer1_search_and_signoff.test_od_floor_admits_the_shape_band_at_low_thrust | ( | monkeypatch | ) |
DEF-09: the OD floor was 0.5 x cap -- a 72.9 mm bore under a 1 kN engine whose D/Dt band (2.2-3.2) wants a 41-60 mm bore.
Definition at line 195 of file test_layer1_search_and_signoff.py.
References _captured_bounds().
| test_layer1_search_and_signoff.test_signoff_enforces_the_declared_element_pitch | ( | ) |
L1-05: 24 doublets in a 127 mm bore is a 22.97 mm pitch against the 6500 N config's 22.5 mm limit; the 4.4e-4 squared miss hid under gate_eps and every seed said VALID.
Definition at line 62 of file test_layer1_search_and_signoff.py.
| test_layer1_search_and_signoff.test_validation_does_not_boost_tanks_to_make_a_design_close | ( | ) |
L1-01: the replay retried at 1.03-1.72x tank pressure, took the first that solved, and stamped the design VALID with the un-raised pressures written to the YAML.
Definition at line 76 of file test_layer1_search_and_signoff.py.
| test_layer1_search_and_signoff.test_valley_escape_ignores_a_constant_shaping_floor | ( | ) |
L1-08: 'best > 100/150/300' read the ~1970-point mass floor as stuck on every 6500 N run (7 sigma kicks) and never fired on blank runs.
Definition at line 156 of file test_layer1_search_and_signoff.py.
| float test_layer1_search_and_signoff.PSI = 6894.757 |
Definition at line 24 of file test_layer1_search_and_signoff.py.
| test_layer1_search_and_signoff.ROOT = Path(__file__).resolve().parents[1] |
Definition at line 23 of file test_layer1_search_and_signoff.py.