motor-fcs-mpc — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited motor-fcs-mpc (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Three-phase 2-level voltage-source inverter + PMSM (SPMSM / IPMSM via parameterization). Inner-loop = FCS-MPC current control in dq frame. Outer-loop (optional) = speed PI providing iq_ref. Generalizes across machine sub-types by parameterizing Rs, Ld, Lq, ψf, Pn, J, F — no topology change between sub-types.
Layered on motor-pmsm-base. All base discipline applies (Goto TagVisibility, Vdc/BEMF rule, Visual 4-check, broken-FOC defense).
add_block, write a numbered plan: parameter table, design-decision choices (design_decisions.md), build-script structure. Get user approval.set_param(mdl, 'InitFcn', sprintf(...)). Model must Run from .slx double-click in fresh MATLAB session. See crit_conditions.md §J-CRIT.Rs/Ld/Lq/ψf/Pn/Tsc as numeric literals at build time. Never put MPC params in an external .m file (drifts from plant). See crit_conditions.md §K-CRIT.ch.SampleTime='-1', ch.ChartUpdate='INHERITED', leave Inputs DataType/Size at Inherit. Rate is locked by ZOH @ Tsc on every chart input + ZOH @ Tsc on chart output. See crit_conditions.md §G-CRIT.LimitOutput='on', limits [-iq_max, +iq_max] with 1.5·Pn·ψf·iq_max ≥ 1.3·TL_max. Without saturation, runaway iq_ref collapses the cost-weight ratio and MPC abandons d-axis. See outer_pi_saturation.md.| Phase | Action | Reference |
|---|---|---|
| 0 | Validate inputs + sanity grid | base/pre_build_grid.md |
| 1 | Plant layer (powergui Discrete @ Ts, DC, UB Inverter, PMSM Salient-pole, TL Step) | (this skill) |
| 2 | Measurement layer (Bus Selector → Goto/From) | (this skill) |
| 3 | Outer speed PI (optional, RPM domain, mandatory saturation) | outer_pi_saturation.md |
| 4 | Inner FCS-MPC chart | algorithm_pseudocode.md + crit_conditions.md §G/§A/§K |
| 5 | Logging (To Workspace @ Tsc) | base/measurement_logger.md |
| 6 | Solver (Variable-step Auto, MaxStep ≤ Ts; powergui handles SPS at Ts) | (this skill) |
| 7 | InitFcn injection | base/plant_modeling.md §InitFcn |
| 8 | Self-tests + acceptance | acceptance_criteria.md |
If issues arise, consult crit_conditions.md (A/D/G/H/J/K) and anti_patterns.md (15 common mistakes).
Ask the user before starting. Defaults in parameter_defaults.md.
| Group | Parameter |
|---|---|
| Machine | Rs (Ω), Ld, Lq (H), ψf (V·s), Pn, J (kg·m²), F (N·m·s) |
| Power stage | Vdc (V) |
| Sampling | Ts (plant solver, ~1–5 μs), Tsc (control period, ~20–50 μs; Tsc/Ts ≥ 10) |
| Outer loop | Kp_w, Ki_w (RPM domain; recommend pi_design.m), iq_max (1.5·Pn·ψf·iq_max ≥ 1.3·TL_max) |
| MPC | control objective (ripple-priority / balanced / torque-priority), λ_d, λ_q per objective, id_ref (SPMSM: 0; IPMSM: ask) |
| Scenario | StopTime, ramp_time, ω_ref_target (RPM), TL_step_time, TL_value (< 1.5·Pn·ψf·iq_max) |
| ✅ Use | ❌ Skip |
|---|---|
| Build / port / extend FCS-MPC simulation in Simulink | FOC, DTC, sensorless, scalar V/Hz, BLDC trapezoidal |
| Single-vector / 7-vector / 8-vector MPC, predictive current control | Multi-step horizon (N>1), delay compensation, weak-field, MTPA optimization |
| Generalizing FCS-MPC to a new PMSM machine parameter set | Pure theory questions about MPC math |
MATLAB perf (matlab-performance-optimizer) or unit tests (matlab-test-creator) |
| Sub-type | Parameter constraint | Strategy |
|---|---|---|
| SPMSM | Ld == Lq | id_ref = 0 (no reluctance torque) |
| IPMSM mild saliency | Lq > Ld, Lq/Ld ≤ 1.5 | id_ref = 0 workable; MTPA gives ~5% torque-per-amp gain |
| IPMSM strong saliency | Lq/Ld ≥ 2 | id_ref from MTPA solver (out of scope; ask user) |
Topology does not change — same blocks, same wiring, same chart algorithm, same CRIT conditions. Only parameters and id_ref strategy differ.
Out-of-scope sub-types: SynRM (ψf ≈ 0), IM, BLDC trapezoidal — different prediction equations.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.