Runs entirely in your browser. Stiff-grid, positive-sequence model; illustrative parameters, not validated against PowerWorld.
Inputs and disturbances
Reference, frequency and phase steps persist. Only the voltage dip clears. Presets reset the scenario; controller parameters stay as entered.
Initial-condition offsets
Zero offsets start from an initialized equilibrium. These values perturb controller states without changing the electrical operating point specified above.
Controller parameters
Imax = 0 disables current limiting. In direct-E mode, consider Kpqmax = 0.1 and Kiqmax = 10; mode selection does not change gains automatically. All quantities are per unit unless a unit is shown.
Inspect up to eight signals from the last completed run. Selection changes do not rerun the simulation. Angles delta and theta_g are radians; angle_relative_deg is degrees.
How to explore the model
Power step: follow Perror → droop → f_inv → angle_relative_deg → P. Grid frequency: smaller Mp gives a larger power response. Voltage dip: compare requested Iraw with actual I, and Edroop with Enet. P/Q limits: these act through feedback, so brief excursions are possible.
The model has nine states and a transient radial current cap. It does not include PWM, DC-side dynamics, unbalanced phases, islanding or an external network solver. Terminal grid voltage is imposed, not controlled by this inverter.
The JavaScript solver uses adaptive RK4 step doubling, ends steps at disturbance boundaries, and projects bounded states after accepted steps. Numerical differences from bdsim are documented in VALIDATION.md. Large gains or severe disturbances can produce unstable responses.
Drag a graph to zoom, double-click to reset. Click legend entries to hide signals. CSV and JSON downloads describe the last completed run, even if you subsequently edit the controls.