Electrical Engineering › Section L · Power (ac) and resonance › Topic L.3

Power-factor correction

Motors and other inductive loads draw more current than the real power they use would need. A capacitor placed across them supplies their reactive power on the spot, so the supply delivers less current for exactly the same work.

Why the supply wants less current

A distribution system works at a fixed voltage, so the current it must deliver is set by the apparent power: \(I = S/V\). Current is what costs money. Line losses grow with its square, \(I^2R\); bigger currents need thicker conductors and larger transformers and generators; and they cause larger voltage drops. For a given real power \(P\), the current is smallest when \(Q_T = 0\), so that \(S = P\) and the power factor is 1. Power-factor correction means adding reactive elements to push the power factor towards 1. Most loads (motors, transformers, magnetic ballasts) are inductive, so the correcting element is almost always a capacitor.

The idea: supply the VAR locally

Connect a capacitor in parallel with an inductive load. The load still sees the same voltage, so it draws the same current and uses the same real and reactive power; it cannot tell the capacitor is there. But the supply now delivers the sum of the load current and the capacitor current. The capacitor's negative reactive power cancels part or all of the load's positive reactive power, so \(Q_T\) falls, \(S_T\) shrinks towards \(P\), and the supply current falls. In energy terms, the stored energy now swaps between the motor's magnetic field and the capacitor's electric field next to it, instead of travelling all the way back to the power station.

The phasor picture

Take the supply voltage as the reference. An inductive load current has a real part in phase with the voltage and a negative imaginary part: \(\mathbf{I}_L = I_{\text{re}} - jI_{\text{im}}\). An ideal capacitor's current leads the voltage by \(90^\circ\), so it is purely positive imaginary, \(\mathbf{I}_C = +jI_C\). The supply current is their sum:

\mathbf{I}_s = I_{\text{re}} + j(I_C - I_{\text{im}})

Choose the capacitor so that \(I_C = I_{\text{im}}\) and the supply current is just \(I_{\text{re}}\), in phase with the voltage. The supply then sees a purely resistive load, and its current is the smallest possible for that real power.

Sizing the capacitor

To move from power-factor angle \(\theta_1\) to a target \(\theta_2\) at constant real power \(P\), the capacitor must supply the difference in reactive power. Its \(Q_C\) is negative, so its size is

|Q_C| = P(\tan\theta_1 - \tan\theta_2)

For full correction \(\theta_2 = 0\), so \(|Q_C|\) equals the whole \(Q\) of the load. A capacitor across voltage \(V\) has \(|Q_C| = V^2/X_C\), so

X_C = \frac{V^2}{|Q_C|},\qquad C = \frac{|Q_C|}{2\pi f V^2}

Power-factor capacitors are normally sold by their kVAR rating at a stated voltage rather than by capacitance, so in practice you choose the nearest standard kVAR rating and check that its voltage rating suits the supply.

Worked example: correcting a motor to unity

A 7.5 hp motor with an efficiency of 88% and a power factor of 0.72 lagging runs on 240 V, 60 Hz. 1. Output: \(7.5 \times 746 = 5595\) W, so input \(P = 5595/0.88 = 6358\) W. 2. \(\theta_1 = \cos^{-1}0.72 = 43.9^\circ\) and \(Q = P\tan\theta_1 = 6128\) VAR. 3. \(S = P/0.72 = 8830\) VA, so \(I = 8830/240 = 36.8\) A. 4. For unity, \(Q_C = -6128\) VAR, so \(X_C = 240^2/6128 = 9.40\,\Omega\). 5. \(C = 1/(2\pi \cdot 60 \cdot 9.40) = 282\,\mu\text{F}\). 6. Now \(S = P\), so \(I = 6358/240 = 26.5\) A, a 28% cut. 7. Check: the capacitor current, \(240/9.40 = 25.5\) A, equals the imaginary part of the motor current, \(6128/240 = 25.5\) A.

Worked example: a workshop raised to 0.95

A workshop on 240 V, 60 Hz has 4.0 kW of heaters (\(F_p = 1\)) and motors totalling 10 kVA at 0.70 lagging. Raise its power factor to 0.95 lagging. 1. Motors: \(P = 7.0\) kW and \(Q = 10\sin(\cos^{-1}0.70) = 7.14\) kVAR. 2. Totals: \(P = 11.0\) kW, \(Q = 7.14\) kVAR, \(S = 13.1\) kVA, \(F_p = 0.839\) lagging, \(I = 54.6\) A. 3. Target: \(\theta_2 = \cos^{-1}0.95 = 18.2^\circ\), so \(Q' = 11.0\tan 18.2^\circ = 3.62\) kVAR. 4. \(|Q_C| = 7.14 - 3.62 = 3.53\) kVAR. 5. \(C = 3530/(2\pi \cdot 60 \cdot 240^2) = 162\,\mu\text{F}\). 6. New \(S = 11.0/0.95 = 11.6\) kVA, so \(I = 48.2\) A.

Why stop at 0.95 lagging?

In the workshop, the first 3.53 kVAR of capacitors cut the current from 54.6 A to 48.2 A. Going on to unity would need another 3.62 kVAR, about as much again, yet would only lower the current to \(11.0\text{ kW}/240\text{ V} = 45.8\) A. Near unity the triangle's long side is almost the same as its base, so each extra kVAR buys less. Loads also vary through the day, and a capacitor sized for full load can overcorrect at light load, leaving a leading power factor and a current that rises again. That is why targets of about 0.95 lagging are common, and why large sites often use switched capacitor banks.

What correction does and does not do

Correction leaves the load itself untouched: its voltage, current, real power and reactive power are unchanged, so the motor runs exactly as before. It reduces the current, and therefore the losses, only between the supply and the point where the capacitor is connected, which is why capacitors are placed close to large loads. It does not reduce the energy the load converts, so a home's energy bill in kWh barely changes. Utilities, though, often add a charge for industrial customers whose power factor falls below a set level, commonly about 0.9 lagging. Many electronic power supplies use active power-factor correction, a circuit that shapes the input current to follow the voltage.

Common mistakes to avoid

First, sizing the capacitor as \(P\tan\theta_2\): that is the reactive power left over, not the amount to remove. Second, putting the capacitor in series with the load, which changes the load's voltage; correction capacitors go in parallel. Third, forgetting \(2\pi\) or using peak voltage in \(C = |Q_C|/(2\pi fV^2)\). Fourth, assuming correction lowers the load's own current or real power: only the supply current changes. Fifth, adding so much capacitance that the power factor becomes leading. As a safety note, large capacitors can hold a dangerous charge after disconnection, which is why banks include discharge resistors.

Key takeaways

For fixed voltage and real power, supply current is lowest at unity power factor, when \(Q_T = 0\). Most loads are inductive, so a capacitor in parallel supplies their reactive power locally and cancels the imaginary part of the load current. The capacitor needed has

|Q_C| = P(\tan\theta_1 - \tan\theta_2)

with \(X_C = V^2/|Q_C|\) and \(C = |Q_C|/(2\pi fV^2)\). The load is unaffected; only the current and losses upstream of the capacitor fall. Gains shrink near unity, and overcorrection gives a leading power factor, so targets of about 0.95 lagging are usual.

Flashcards

Flashcard

Power-factor correction

Adding reactive elements to bring the power factor towards 1

Usually capacitors across inductive loads.

Test your understanding

Question 1 of 5

A load takes 12 kW at 0.6 lagging. What capacitor reactive power raises its power factor to 1?

Question 2 of 5

A load draws 10 kW at 0.75 lagging. How much capacitor reactive power raises the power factor to 0.95 lagging?

Question 3 of 5

A correction capacitor must supply 4.0 kVAR at 240 V, 60 Hz. What capacitance is needed?

Question 4 of 5

A motor's power factor is raised from 0.70 lagging to 0.95 lagging with a parallel capacitor. What happens to the motor's own current?

Question 5 of 5

Why is power-factor correction usually aimed at about 0.95 lagging rather than exactly 1?

← Previous topicNext topic →