Battery Inverter 3D Simulator — Bidirectional Converter, P/Q & PLL Interactive

Interactive 3D grid-following battery inverter simulator with a bidirectional DC-to-AC workbench (battery rack, DC-link capacitor, six-switch bridge, AC filter inductors, AC contactor/PLL and three-phase grid), controls for real-power request, reactive-power request, DC-link voltage, grid line-to-line voltage, grid frequency, switching frequency, battery state of charge, a utility-voltage-available toggle and a bridge-enable toggle, time-stepped playback with real-time to 60× fast-forward, live power/current/loss/SOC readouts, waveform and power-balance charts, model equations, six guided experiments (charge from AC, reactive priority, low DC headroom, loss of utility, filter comparison, empty battery), a built-in model-verification bench (checks.js), a timestamped event log with trial-report export, guided lessons and a knowledge-check quiz.

← Renewable Energy Labs
About this tool — how it works & FAQOpen ▾Close ▴

About the Battery Inverter Simulator

This simulator models a grid-following bidirectional battery inverter — the six-switch bridge that converts a battery's stored DC energy into three-phase AC for export, or converts incoming AC into DC to recharge the battery — inside a 100 kVA capability circle. Unlike the standalone Solar Inverter lab, which models a one-way PV string inverter tracking a DC source's maximum power point, this simulator's bridge is fully bidirectional and is driven by direct real-power and reactive-power requests against a battery, not by MPPT-seeking behavior against a solar array.

What the simulator shows

• A real-time 3D bidirectional DC-to-AC workbench — battery rack, DC-link capacitor, six-switch bridge, AC filter inductors, AC contactor/PLL and three-phase grid — with a toggleable enclosure, auto-rotate, expand, hideable labels and selectable components with callouts. • Converter controls: real-power request (−100 to 100 kW), reactive-power request (−100 to 100 kvar), DC-link voltage (500–850 V), grid line-to-line voltage (360–440 V), grid frequency (45–65 Hz), switching frequency (1,000–12,000 Hz), battery state of charge override, a utility-voltage-available checkbox and an enable-switching-bridge checkbox. • Playback controls: pause/resume, advance 0.1 s, advance 1 s, and a speed selector from 10× slow motion through real time to 60× faster, plus a reset-protection button and full laboratory reset. • Live readouts for AC real power (signed, + is export), AC reactive power, apparent power, AC line current RMS, conversion loss and battery state of charge, an operating-sequence narrative, and a status badge. • A Waveforms & power analysis tab with two charts, the underlying model equations, an analysis-scope note and snapshot measurements. • An Experiments tab with six guided scenarios (charge from AC, reactive priority, low DC headroom, loss of utility, filter comparison, empty battery), a Model verification bench running independent deterministic checks against a fresh model instance, and a timestamped event log with trial-report export. • A Learn & assess tab with guided lessons, a knowledge-check quiz and a scope/references section.

How the bridge, PLL and capability circle interact

The six-switch bridge is bidirectional: the same controlled semiconductor paths carry current from battery to grid during export or from grid to battery during charging, with the sign of the AC real-power request determining which direction is active. Real and reactive power share a single 100 kVA apparent-power rating (S = √(P² + Q²)), so a large reactive-power request reduces how much real power the converter can simultaneously deliver — the model gives reactive power priority within that capability circle.

Because this is a grid-following design, the AC contactor/PLL controller needs an existing utility voltage reference to synchronize to; it waits a modeled one second to acquire that reference before closing, and it disconnects immediately if utility voltage disappears rather than trying to energize a dead bus. The two-level bridge also needs enough DC-link voltage headroom to synthesize the requested AC voltage through linear sinusoidal PWM (VLL,rms,max ≈ 0.612·Vdc) — drop DC voltage too low and switching is blocked until headroom is restored.

Scope and what this model excludes

This is an averaged three-phase P/Q converter model with a 0.3 s response and a simplified positive loss map (Ploss = 0.015S + 0.0002S² kW); DC and AC voltages are ideal adjustable fixtures rather than solved network quantities. Waveform displays are fundamental references, not switching-resolved outputs, and the ripple estimate is an order-of-magnitude fixture that decreases as switching frequency increases — it is not a solved harmonic spectrum. EMI, harmonic distortion, protection certification and semiconductor thermal dynamics are excluded. Because this is a grid-following (not grid-forming) model, it explicitly cannot energize an isolated/dead bus — that specific islanding and frequency-formation behavior is the subject of the separate Islanded Microgrid lab.

Frequently asked questions

How is this different from the standalone Solar Inverter lab?

The Solar Inverter lab models a one-way PV string inverter that tracks a DC solar array's maximum power point (MPPT). This Battery Inverter lab instead models a fully bidirectional battery converter driven directly by real-power and reactive-power requests — it can both discharge the battery to export AC power and rectify incoming AC to recharge the battery, which the solar inverter model does not do.

What does "grid-following" mean here, and how is that different from grid-forming?

A grid-following converter needs an existing AC voltage/frequency reference from the utility to synchronize to before it can operate, and it disconnects immediately if that reference disappears. It cannot create or sustain an isolated AC bus on its own. That grid-forming, islanding-capable behavior is modeled separately in the Islanded Microgrid lab.

Why does a reactive-power request reduce available real power?

Real and reactive power both draw on the same 100 kVA apparent-power rating, related by S = √(P² + Q²). At the 100 kVA limit, requesting 80 kvar of reactive power leaves only about 60 kW of real-power capacity, since reactive power is given priority within that fixed capability circle.

What does the model verification bench check?

The Experiments tab includes a Model verification bench that runs independent deterministic checks — covering the 100 kVA capability circle, the DC-power-equals-AC-power-plus-loss balance, charging behavior, immediate shutdown on loss of grid reference, blocked switching under low DC headroom, and the inverse relationship between switching frequency and ripple — against a freshly constructed model, leaving your live experiment state untouched.

Related tools & guides