APFC Panel Working Explained

Open any industrial LT panel room and you'll find a grey cubicle humming quietly next to the main incomer, its controller display cycling through numbers: 0.97… 0.98… 0.96. That is the APFC panel — Automatic Power Factor Correction — and it exists for one brutally financial reason: utilities penalize low power factor, and this panel keeps it high without anyone touching it.

If your final-year project touches power factor correction, an APFC panel is almost certainly the hardware at its center. This guide explains what it does, how each part works, how the controller thinks, and how to size it — plus the failure modes that make great viva questions.

The problem it solves

Inductive loads (motors, transformers, welders, fluorescent chokes) draw lagging reactive power (kVAr) to magnetize their cores and windings. Reactive power does no useful work, but it still flows through cables, transformers and the utility network, loading everything up. The ratio of useful power to total power is the power factor:

pf = kW / kVA = cos φ

A plant at 0.7 pf draws 43% more current than the same plant at unity pf for the same useful power — bigger cables, bigger transformer, higher losses, and a utility penalty on the bill (or kVAh billing that punishes it automatically). Capacitors supply leading reactive power that cancels the lagging reactive power of the loads. The APFC panel is simply a bank of capacitors switched in and out automatically to track the load.

Anatomy of an APFC panel

Component	What it does	Selection notes
Power capacitors (kVAr steps)	Supply leading reactive power	MPP (metallized polypropylene) self-healing type; rated voltage ≥ system voltage with margin (440V capacitors on a 415V system)
Detuning reactors (optional but recommended)	Series inductance per step to prevent harmonic resonance	7% reactors (tune ~189Hz) standard; 14% (134Hz) where harmonics are severe
Capacitor-duty contactors	Switch capacitor steps	Special contactors with pre-insertion damping resistors/contacts to limit inrush — never use ordinary motor contactors
APFC controller (relay)	Measures pf and switches steps	Microprocessor-based, 6–14 steps, with CT input; displays pf, kVAr, V, I, THD on better units
MCCB / switch-fuse incomer	Protection and isolation	Sized for 1.5–1.6× the total capacitor current (capacitor inrush + harmonics)
Discharge resistors	Drain residual voltage after switch-off	Built into modern MPP capacitors (discharge to <50V in 60s per IS 13340); verify, don't assume
Ventilation	Removes heat	Capacitors lose life fast above 45–50°C; forced ventilation or louvers, and keep the panel out of direct sun

Note: Capacitor-duty contactors exist because switching a capacitor is electrically violent: an uncharged capacitor looks like a short circuit, and inrush can hit 100× rated current. The damping contacts pre-charge through resistors before the main contacts close. Using a standard contactor welds the contacts — sometimes on the first operation.

How the controller thinks

The controller measures voltage (direct) and current (via CT) at the incomer and computes the reactive power deficit. It then switches capacitor steps to drive the power factor toward the target (typically 0.98–0.99 lagging — not unity, and never leading).

Two subtleties that separate understanding from memorization:

1. Target slightly lagging, never leading. A leading power factor (over-compensation) raises voltage, can resonate, and some utilities penalize it too. Controllers are set to hold ~0.98 lag.
2. Switching program and C/k setting. The C/k value = (smallest step kVAr) / (CT ratio × √3 × V) — it tells the controller the smallest reactive change it can resolve. Set wrong, the panel hunts (steps chattering in and out) or ignores small loads. Auto C/k on modern controllers removes the pain, but you should know what it means.

Switching sequence: controllers rotate steps (first-in-first-out or intelligent rotation) to equalize wear, and enforce a discharge delay (typically 30–60s) before re-switching a step — reconnecting a still-charged capacitor to an opposite-polarity half-cycle is a classic capacitor-killer.

Sizing the panel: the worked calculation

Qc = P × (tan φ₁ − tan φ₂)

where P = active power (kW), φ₁ = current angle, φ₂ = target angle.

Worked example: A workshop draws 120kW at 0.72 pf; target 0.98.

• φ₁ = cos⁻¹(0.72) = 43.9°, tan φ₁ = 0.964
• φ₂ = cos⁻¹(0.98) = 11.5°, tan φ₂ = 0.203
• Qc = 120 × (0.964 − 0.203) = 120 × 0.761 ≈ 91 kVAr

Practical panel: 100 kVAr in steps — e.g. 8 steps of 12.5 kVAr, or a mixed bank (25 + 4×12.5 + smaller trim steps like 2×6.25 for fine resolution at light load). Mixed step sizes give finer control; equal steps are simpler. Always include at least one small step so the controller can correct at light load without overshooting into leading.

Sanity checks:

• Capacitor current per step: I = Q/(√3 × V). A 25kVAr step at 415V: 25000/(1.732 × 415) ≈ 34.8A — size that step's contactor, fuse and cable for ~1.5× this.
• Incomer: 100kVAr → 139A → MCCB ~200A frame with appropriate setting.
• With drives/VFDs on site: detune every step (7% reactors) — undetuned capacitors + harmonics = resonance and swollen cans.

Commissioning checklist

• CT installed on the correct phase, correct orientation (P1 toward source), and the controller's CT ratio setting matches the actual CT
• Voltage sensing from the same phases the CT measures (phase association correct — wrong phasing makes the controller correct backwards)
• Target pf set to 0.98 lag (not unity)
• C/k set (or auto-detected and verified)
• Step sizes programmed to match the actual installed steps
• Discharge delay ≥ 30s observed on a test switch
• Panel ventilated; capacitor can temperatures checked after 1 hour at load
• Utility meter / independent analyzer confirms pf improvement and no leading pf at light load

Warning: Capacitors store charge. Even with discharge resistors, verify zero voltage with a rated meter before touching terminals — a resistor can fail open and leave a can charged at hundreds of volts. APFC commissioning is live-panel work: insulated tools, arc-flash awareness, and supervision. Never bypass the discharge delay to "test faster."

Failure modes (and how to answer them in the viva)

1. Swollen / burst capacitors — overheating (poor ventilation, high ambient), harmonic resonance (no detuning), or overvoltage. The pressure interrupter in MPP cans disconnects internally; a swollen can is a dead step.
2. Hunting — steps switching in and out repeatedly: wrong C/k, CT on the wrong feeder (measuring only part of the load), or a target set unrealistically tight.
3. Welded contactor contacts — ordinary contactor used for capacitor duty, or damping contacts failed.
4. Leading pf at light load — fixed (unswitched) compensation too large, or controller holding steps in with no load. Fix: smaller trim steps, or switch the fixed bank with the main load.
5. Controller shows pf but bill doesn't improve — CT measuring only the incomer after the capacitor connection point, so the controller corrects a current it can't see properly; or the penalty is on kVAh/MD, not pf directly.
6. Resonance with the supply — undetuned bank + VFDs: one harmonic order amplified many times, capacitors fail within months. Solution: detuning reactors sized to the measured spectrum.

Where to go from here

• Power Factor Capacitor Sizing Calculation — the Qc math in full detail.
• VFD Basics: Motor Speed Control — the harmonic sources that make detuning reactors necessary.
• DOL vs Star-Delta Starter — contactor-based motor starting, the same switching duty as APFC steps.
• Three-Phase Wiring Basics — the supply system the APFC panel corrects.
• More power-factor topics in the Electrical branch hub.