Three Phase Grid Inverter

Three-Phase Grid Inverter

Overview

Use this active subskill for three-phase grid-connected inverter work. It contains the populated knowledge in this repository: T-type/NPC three-level SVPWM, state-vector table checks, gate mapping, From/Goto routing risks, and simulation-backed waveform validation.

Use the root simulink-power-electronics workflow first, then load this subskill when the model is in scope.

Resource Map

• Read references/svpwm-three-level.md for T-type or NPC three-level SVPWM logic, sector handling, and gate mapping.
• Read references/table7-state-vectors.md when auditing or rebuilding the 36 small-sector state-vector table.
• Read references/vsg-control-notes.md for VSG (Virtual Synchronous Generator) control implementation, PI tuning experience, power calculation verification, and feedforward formulas specific to grid-connected inverters.
• Use scripts/svpwm_diagnostics.m when MATLAB/Simulink can run and the model exposes state, gate, and voltage signals.
• Use scripts/print_table7_state_vectors.py when a deterministic JSON or Markdown copy of the reference table is needed.
• Also read the root ../../references/control-algorithm-debugging.md for general control algorithm debugging methodology and signal-tracing procedures.

In Scope

• three-phase grid-connected inverters
• two-level inverter diagnostics when the same gate-routing and waveform evidence workflow applies
• T-type or NPC three-level bridges
• SPWM, SVPWM, and closely related discontinuous PWM checks
• phase/line voltage balance, current balance, DC offset, and phase-sum checks
• controller-to-plant gate ordering and legal switching states
• VSG (Virtual Synchronous Generator) grid-forming inverter control
• PI controller tuning, anti-windup mechanisms, and feedforward design
• S-Function based control algorithm implementation and debugging
• Power calculation verification (P/Q) with Clarke/Park transform coefficients

Out Of Scope

Use a different subskill or the root workflow for:

• motor-drive torque/speed control
• DC-DC converter regulation
• rectifier/PFC-specific behavior
• battery or BMS modeling
• HVDC/FACTS studies
• MMC or flying-capacitor balancing not covered by the existing three-level inverter notes

Workflow Additions

1. Confirm grid-side reference frame, phase sequence, measurement polarity, and analysis window before interpreting waveforms.
2. Identify the active modulation path: SPWM, SVPWM, disabled reference path, or commented subsystem.
3. Log controller state vectors, final gate vectors, plant-side gate inputs, and phase/line voltage or current measurements.
4. For three-level SVPWM, compare every large sector and small sector against table7-state-vectors.md; partial sector-I checks are not enough.
5. For T-type/NPC gate mapping, confirm the physical switch order before applying N/O/P assumptions.
6. Validate after startup transient unless startup behavior is the target.

Evidence Checks

• controller gate vector equals plant-side gate input
• no illegal switch state appears in any bridge leg
• phase and line RMS values are balanced in the selected steady-state window
• phase sum, DC offset, and current polarity match the model objective
• sector/state-vector counts are plausible for the modulation method and operating point

Scripts

For a conventional three-level SVPWM model with accessible logging points, add this subskill's scripts directory to the MATLAB path and call:

cfg = struct();
cfg.model = "YourModel";
cfg.svpwmSubsystem = "YourModel/SVPWM";
cfg.stateBlock = "YourModel/SVPWM/Subsystem1";
cfg.gateBlock = "YourModel/SVPWM/Subsystem3";
cfg.voltageBlocks = ["YourModel/Voltage Measurement2", "YourModel/Voltage Measurement3", "YourModel/Voltage Measurement4"];
cfg.stopTime = "0.10";
report = svpwm_diagnostics(cfg);

For the bundled table reference:

python3 subskills/three-phase-grid-inverter/scripts/print_table7_state_vectors.py --format markdown
python3 subskills/three-phase-grid-inverter/scripts/print_table7_state_vectors.py --format json