Power Factor Formula: How to Calculate Power Factor (With Examples)

The Invoice That Doesn’t Add Up

I was on a factory audit last year — a mid-sized food processing plant, three-phase supply, mix of motor-driven conveyors and refrigeration compressors. The energy manager had pulled me in because their monthly electricity bills kept climbing even though production hadn’t changed. He was convinced there was a metering error.

There wasn’t. The problem was sitting right there in the apparent power figure on the network analyser: their power factor was running at 0.72. If you’re not familiar with what that number means in practice, here’s the short version — it was costing them several hundred euros a month in kVA demand charges, and nobody had measured it properly in years.

Why a Poor Power Factor Hits You in the Pocket

Power factor is one of those things that’s invisible until it becomes expensive. A low power factor means your installation is drawing more current from the grid than its real load actually requires. That excess current heats cables, stresses switchgear, reduces the available capacity of your transformers and — critically — many industrial and commercial tariffs include a kVA demand charge or apply a power factor correction penalty. Get below 0.9 or 0.95 (depending on the network operator), and the bills go up. It also affects your ability to size a solar PV system accurately: if you’re exporting apparent power (kVA) into PVSyst where it expects real power (kW), you’ll overstate the real-power demand and misread the load curve. Always confirm which column your logger software is writing before importing.

The Power Factor Formula, Explained Simply

Power factor (PF) is the ratio of real power to apparent power:

PF = P (kW) ÷ S (kVA)

That’s the core of it. But to use it properly on-site, you need to understand what’s behind each of those terms.

Real power (P, in kW) is the power actually doing work — spinning a motor, heating an element, running a compressor. This is what the watt-hour meter records and what your client pays for in kWh.

Apparent power (S, in kVA) is the total power the supply must deliver, including the reactive component. It’s the vector sum of real and reactive power. This is what stresses the cables and the transformer.

Reactive power (Q, in kVAR) is the power that oscillates back and forth between source and load — it does no useful work, but it occupies current capacity. Inductive loads (motors, transformers, some lighting ballasts) consume it; capacitive loads supply it.

The three are related by the power triangle:

S² = P² + Q²

So if you know any two, you can calculate the third. For sinusoidal waveforms, power factor also equals the cosine of the phase angle (θ) between voltage and current — hence the term cos φ you’ll see on some instruments. For a pure resistive load, θ = 0°, cos φ = 1.0, and P = S. For a heavily inductive load, θ increases, cos φ drops, and you’re drawing a lot of current for the actual work being done.

One important note: in systems with significant harmonic distortion — VFDs, switch-mode power supplies, LED drivers with poor design — cos φ doesn’t tell the whole story. Total power factor also includes the distortion power factor, driven by THD in the current waveform. You can have cos φ of 0.98 and still have a poor true power factor if THD is high. More on that below.

A Worked Example: Three-Phase Motor Load

Here’s a typical scenario: a 75 kW motor driving a pump, three-phase 400 V supply, running at roughly 80% load, so around 60 kW of actual shaft demand.

You measure with a power analyser and get:

• Real power (P): 60 kW
• Apparent power (S): 80 kVA

PF = 60 ÷ 80 = 0.75

That’s a poor result. Now let’s calculate the reactive power:

Q = √(S² – P²) = √(80² – 60²) = √(6400 – 3600) = √2800 ≈ 52.9 kVAR

So you’ve got nearly 53 kVAR of reactive demand that’s drawing current, heating cables, and potentially attracting a penalty tariff — all without doing any useful work. A correctly sized power factor correction capacitor bank (in this case, around 50–55 kVAR) would bring PF up toward 0.98–0.99, reducing apparent power to close to 61 kVA and cutting current draw proportionally.

On a three-phase system, remember you’re calculating total three-phase power. If you’re measuring per-phase, add the phases:

P_total = P_L1 + P_L2 + P_L3

And check for phase imbalance while you’re at it — an unbalanced load will give you misleading totals if you only measure one phase and multiply by three.

For how these three-phase power numbers feed into inverter and battery sizing for a real solar or backup system, see Using logged energy consumption data to plan solar and backup power systems.

Where the Right Tool Makes the Difference

Calculating power factor from first principles is straightforward. Actually measuring the underlying values accurately on a live three-phase system — while it’s running, on conductors you can’t disconnect, over a representative logging period — is where most jobs go wrong.

This is where I reach for the Meatrol Mi550. It’s a portable handheld three-phase energy logger and power quality analyser that installs on a live circuit via Rogowski coils — no load disconnection, no shutdown required. The coils open and close around the conductor, so you’re up and logging in minutes without touching a breaker. On a busy manufacturing site or a commercial building that can’t afford downtime, that alone makes it the right tool.

The Mi550 measures real power, apparent power, reactive power, and power factor simultaneously across all three phases — and it distinguishes between displacement power factor (cos φ) and true power factor, which matters enormously on sites with VFDs or switch-mode power supplies. It logs to a 32 GB internal memory card, handles voltage up to 690 V L-L, and covers three current ranges (600 A, 3 kA, 6 kA) to suit everything from a small commercial board to a heavy industrial supply. Data exports via RJ45 over Modbus TCP, or you pull the card and run it through AmpX’s Excel analysis template, which plots your kW, kVA, and PF trends into ready-to-use graphs without needing a data analyst. For the site audit work I do, that combination of live-circuit portability, harmonic-aware measurement, and practical data output covers the vast majority of jobs.

For lighter-duty work — a single-board commercial site, a short verification audit, or a job where the Mi550’s full feature set is more than you need — the rest of AmpX’s portable energy logger range covers it at lower price points. The Meatrol ME440 is a sensible step-down for general PF and energy measurement when you don’t need the Mi550’s harmonic-grade analysis.

Common Mistakes and Field-Tested Tips

Don’t try to derive power factor from current alone. A current-only clamp meter gives you amps. Without simultaneous voltage and phase angle measurement, you cannot calculate power factor. You need a proper power analyser.

Log for at least 24–48 hours. A point-in-time reading is almost meaningless. Power factor varies with load — it’s often worst at light load, when motors are running below their rated point. Capture a full operating cycle before drawing conclusions.

Watch for leading vs. lagging power factor. Inductive loads (motors) give lagging PF. Over-compensated sites — where someone has installed too much capacitance — can run at leading PF, which causes its own problems with voltage rise and equipment stress. Your analyser should tell you which way you’re sitting.

High THD distorts displacement power factor readings. If you’re on a site with lots of VFDs, UPS systems, or switch-mode power supplies and your instrument only reports cos φ, you may be missing significant distortion power factor. Use a harmonic-capable analyser and report true power factor.

Check phase imbalance before you report. Power factor per phase can vary significantly if loads are unevenly distributed. Always report per-phase values alongside the three-phase total, especially on a site with single-phase loads across a three-phase distribution board.

Frequently Asked Questions

What is a good power factor for industrial and commercial sites?

For industrial and commercial installations, aim for 0.95 or better. Most network operators apply penalties below a defined threshold — typically 0.9 in Germany (depending on the DSO and tariff), 0.95 on many UK industrial tariffs, and around 0.85–0.95 in South Africa under NRS 048 and municipal supply agreements. Below 0.9 you’re paying for it somewhere — explicit kVA demand charges, a reactive energy line on the bill, or implicit cost through oversized cabling and transformer headroom. A site at 0.85 draws roughly 18% more current than its real load requires; at 0.7 it draws 43% more. Target 0.97 after correction; pushing to a strict 1.0 risks over-compensation and leading PF problems.

How do I measure power factor on a live three-phase circuit without shutting down?

Use a portable power analyser with Rogowski coils. The coils are flexible loops that open, wrap around each phase conductor, and close — no breaker work, no shutdown, no risk to the running plant. The analyser captures voltage and current simultaneously across all three phases, calculates real, apparent, and reactive power, and logs over time. Log for at least 24–48 hours to capture a full operating cycle, because power factor varies with load and a point-in-time reading can be misleading. The Meatrol Mi550 is what I reach for here — the Rogowski coils handle conductors that physically cannot be disconnected.

Why does displacement power factor (cos φ) differ from true power factor?

Displacement power factor (cos φ) only accounts for the phase shift between the fundamental voltage and current waveforms — it assumes both are pure sine waves. True power factor also accounts for harmonic distortion (3rd, 5th, 7th harmonics and above), which draws current without doing useful work. On a site with significant VFDs, UPS systems, switch-mode power supplies, or LED drivers with poor input design, you can see cos φ around 0.98 while true PF sits at 0.80 — the distortion power factor drags the total down. Instruments that only report cos φ miss this entirely. On non-linear-load sites, always use a harmonic-capable analyser that reports true PF.

How do I size a power factor correction capacitor bank?

Take the worked example above: P = 60 kW, S = 80 kVA, Q = 52.9 kVAR, existing PF = 0.75. To correct to a target PF of 0.98, calculate the reactive power you want to keep: Q_new = 60 × tan(arccos(0.98)) ≈ 12.2 kVAR. The capacitor bank supplies the difference: 52.9 − 12.2 ≈ 41 kVAR. Round up to the nearest standard size — a 45 or 50 kVAR bank. Two practical caveats: target 0.97 rather than 1.0 to avoid over-compensation, and if the site has significant harmonics, use a detuned (reactor-filtered) capacitor bank to prevent resonance. Always verify with a before-and-after logged measurement.

Measure It, Then Fix It

Power factor isn’t complicated in theory — PF = kW ÷ kVA, the power triangle, cos φ. What takes experience is knowing where the numbers hide, how long you need to log to get representative data, and when THD is distorting the picture. Get that measurement right, and you’ve handed your client a clear, defensible case for correction — or confirmed that their system is already running clean.

If you’re looking for a portable analyser that handles the full measurement picture on a live three-phase system, take a look at the Meatrol Mi550 on AmpX. Or if you’re not sure whether it’s the right fit for a specific job, contact the AmpX team — we’re happy to talk through the application before you commit.

Already have logger data from another instrument? AmpX’s Excel analysis template turns CSV exports into ready-to-use kW, kVA, and PF graphs — no advanced Excel skills required.