Regulation and efficiency

Buy a transformer and its nameplate tells you a rated kVA, a rated voltage, a frequency — but it does not tell you, directly, two things every practising engineer needs before switching it into service: how much the secondary voltage will sag once real load current starts flowing, and what fraction of the input power actually reaches the load rather than turning into waste heat. These two numbers, voltage regulation and efficiency, both fall straight out of the equivalent-circuit parameters $R_{01}, X_{01}$ built up in the previous topic. Regulation is a load-dependent voltage story; efficiency is a loss-accounting story with a surprisingly clean optimum. This article builds both formulas from the equivalent circuit, derives the elegant condition for maximum efficiency, and works two IIT-KGP style examples end to end.

The core insight

Voltage regulation is the load-dependent volt-drop across the equivalent series impedance $R_{01}, X_{01}$ , sign-flipping between lagging and leading power factor, while efficiency trades a load-independent iron loss $P_i$ against a load-squared copper loss $P_{cu}=x^2P_{cu,FL}$ , reaching its maximum exactly when the two are equal.

The core idea

Two different questions, one equivalent circuit. Regulation asks: as I load the transformer from no-load to full-load, how far does the secondary terminal voltage fall short of its no-load value? Efficiency asks: of the real power fed in, how much comes out at the secondary terminals? Both questions are answered using the same series impedance $R_{01}, X_{01}$ (referred to the primary; an entirely parallel treatment holds referred to the secondary using $R_{02}, X_{02}$ ) that was derived from the short-circuit test in the equivalent-circuit topic.

Voltage regulation. Percentage regulation is the drop in secondary voltage from no-load to full load, expressed as a percentage of the rated secondary voltage $V_2$ . The standard approximate formula used throughout this course is

\%\text{Reg} \approx \frac{I(R_{01}\cos\theta \pm X_{01}\sin\theta)}{V_2}\times100

where: $I$ is the referred load current, $\theta$ is the load power-factor angle, $R_{01}$ and $X_{01}$ are the equivalent series resistance and reactance, and the $+$ sign applies for a lagging power factor while the $-$ sign applies for a leading power factor. Physically, a lagging load draws current that lags the voltage, so both the resistive drop $IR_{01}\cos\theta$ and the reactive drop $IX_{01}\sin\theta$ subtract from the secondary voltage in the same sense, and regulation is positive. A leading load reverses the sense of the reactive term, so it can partially cancel the resistive drop — and for a sufficiently leading power factor, regulation can even go negative, meaning the loaded secondary voltage slightly exceeds the no-load value.

Efficiency and the two loss types. Transformer losses split cleanly into two kinds with opposite dependence on load. Core (iron) loss $P_i$ is set by the peak flux density $B_{max}$ and supply frequency $f$ (from the B-H curve topic) and is essentially constant regardless of load current, since flux is fixed by the applied voltage. Copper loss $P_{cu}=I^2R_{01}$ grows with the square of load current. Writing the load as a fraction $x$ of the rated (full-load) current, copper loss scales as $P_{cu}=x^2P_{cu,FL}$ while $P_i$ stays fixed, giving

\eta = \frac{xS\cos\theta}{xS\cos\theta + P_i + x^2P_{cu,FL}}

where: $\eta$ is efficiency, $S$ is the rated kVA, $\cos\theta$ is the operating power factor, $x$ is the load fraction (1 at full load), $P_i$ is the constant iron loss, and $P_{cu,FL}$ is the copper loss at full load ( $x=1$ ).

Figure

Efficiency rises then falls as load fraction increases with a peak where copper loss equals iron loss

The condition for maximum efficiency

Derivation AMaximum-efficiency condition and the optimal load fraction

The goal of this derivation is to find, starting only from the efficiency formula $\eta = xS\cos\theta / (xS\cos\theta + P_i + x^2P_{cu,FL})$ , the exact load fraction $x$ at which efficiency peaks, and to show that this peak occurs precisely when copper loss equals iron loss — a condition compact enough to be used directly in problems without redoing the calculus each time.

Step 1. Efficiency is easier to maximise if rewritten as one minus the loss fraction, since maximising output-over-input is equivalent to minimising total-loss-over-output-plus-loss for a fixed numerator shape. Instead, differentiate $\eta$ directly with respect to $x$ at constant power factor and rated values, treating $\eta$ as a ratio $N/D$ with $N=xS\cos\theta$ and $D=xS\cos\theta+P_i+x^2P_{cu,FL}$ ; the quotient rule gives $d\eta/dx = (N'D - ND')/D^2$ , and setting the numerator to zero (since $D^2\neq0$ ) is the condition for an extremum.

Step 2. Compute the two derivatives: $N' = S\cos\theta$ and $D' = S\cos\theta + 2xP_{cu,FL}$ . Substituting into $N'D=ND'$ and expanding, every term containing $S\cos\theta \cdot xS\cos\theta$ cancels identically on both sides (they are the same product), leaving $S\cos\theta\,(P_i + x^2P_{cu,FL}) = xS\cos\theta\,(2xP_{cu,FL})$ .

Step 3. Divide through by the common factor $S\cos\theta$ (nonzero for any real operating point) to get $P_i + x^2P_{cu,FL} = 2x^2P_{cu,FL}$ , which rearranges immediately to $P_i = x^2P_{cu,FL}$ . Since $x^2P_{cu,FL}$ is precisely the copper loss $P_{cu}$ at load fraction $x$ , this is exactly the condition $P_{cu}=P_i$ : maximum efficiency occurs when copper loss equals iron loss.

Step 4. Solve for the load fraction directly: $x_{max\,\eta} = \sqrt{P_i / P_{cu,FL}}$ . This is a single-line, plug-in formula — given only the constant iron loss and the full-load copper loss (both readable off a nameplate or test data), the load fraction for peak efficiency follows with no further calculus needed. Verified: setting $x=1$ recovers the full-load condition $P_i \overset{?}{=} P_{cu,FL}$ , correctly showing that full load is the efficiency peak only in the special case where iron and full-load copper loss happen to be equal — otherwise the peak sits at some other $x$ , consistent with Step 4.

Worked example

A transformer has $R_{01}=1.5\,\Omega$ (referred to primary), $X_{01}=4\,\Omega$ , rated secondary voltage $V_2=400\,\text{V}$ , and delivers a full-load referred current $I_1=25\,\text{A}$ at a lagging power factor of $0.8$ ( $\theta=36.87^\circ$ , $\sin\theta=0.6$ ). Separately, at rated 50 kVA this same class of transformer has full-load copper loss $P_{cu,FL}=640\,\text{W}$ and constant core loss $P_i=250\,\text{W}$ , and operates at $0.8$ lagging. Find (a) the percentage voltage regulation, (b) the full-load efficiency, and (c) the load fraction and value of maximum efficiency.

Step 1 — regulation drop components. $IR_{01}\cos\theta = 25\times1.5\times0.8=30\ \text{V}$ and $IX_{01}\sin\theta=25\times4\times0.6=60\ \text{V}$ .

Step 2 — sum with the lagging sign and find regulation. $\text{Drop}=30+60=90\ \text{V}$ , so $\%\text{Reg}=(90/400)\times100=22.5\%$ . (These numbers are chosen for arithmetic clarity; a well-designed power transformer typically regulates under 10%.)

Step 3 — full-load efficiency. Output power $P_{out}=S\cos\theta=50{,}000\times0.8=40{,}000\ \text{W}$ , so $\eta_{FL}=40{,}000/(40{,}000+250+640)=40{,}000/40{,}890\approx0.9782=97.82\%$ .

Step 4 — load fraction for maximum efficiency. $x_{max\,\eta}=\sqrt{P_i/P_{cu,FL}}=\sqrt{250/640}=\sqrt{0.3906}\approx0.625$ .

Step 5 — maximum efficiency itself. At $x=0.625$ : $P_{out}=0.625\times50{,}000\times0.8=25{,}000\ \text{W}$ , and $P_{cu}=0.625^2\times640=250\ \text{W}$ (equal to $P_i$ , as expected). So $\eta_{max}=25{,}000/(25{,}000+250+250)=25{,}000/25{,}500\approx0.9804=98.04\%$ .

Percentage regulation $=22.5\%$ ; full-load efficiency $\approx97.82\%$ ; maximum efficiency $\approx98.04\%$ occurs at $62.5\%$ of full load — slightly higher than the full-load figure, illustrating the general pattern that transformers rarely run exactly at their maximum-efficiency point.

Competitive stretch

Suppose the same 50 kVA transformer is instead operated at only 40% of full load ( $x=0.4$ ), still at $0.8$ lagging. Then $P_{out}=0.4\times50{,}000\times0.8=16{,}000\ \text{W}$ and $P_{cu}=0.4^2\times640=0.16\times640=102.4\ \text{W}$ , so $\eta=16{,}000/(16{,}000+250+102.4)=16{,}000/16{,}352.4\approx0.9785=97.85\%$ — noticeably below the $98.04\%$ peak, confirming that efficiency falls away on both sides of $x\approx0.625$ , not just above it.

Common traps

Using the wrong sign in the regulation formula — the $+$ (lagging) and $-$ (leading) distinction is not optional; using $+$ for a leading load gives a wrong, wrongly-signed answer.
Forgetting that copper loss scales as $x^2$ while iron loss stays constant — a common error treats both losses as fixed, or scales copper loss linearly with load.
Confusing maximum efficiency with full-load operation — the peak occurs where $P_{cu}=P_i$ , which is generally below rated current, not at $x=1$ .
Mixing up $x$ and $S$ in the output-power formula — output real power is $xS\cos\theta$ , not $S\cos\theta$ scaled some other way; always check whether $x$ has already been folded in.

Ask the tutor: ask the AI tutor to rework the maximum-efficiency derivation using loss-fraction minimisation instead of the quotient rule, or to solve a fresh regulation problem with a leading power factor where the sign flips the answer.

Why does the regulation formula use an approximate plus-or-minus form instead of an exact vector calculation?

The exact secondary voltage under load is found by vector (phasor) addition of the no-load voltage and the impedance drop, which involves a small correction term that is usually negligible for typical transformer impedances. The approximate formula keeps only the dominant in-phase component of that vector sum, which is accurate enough for the regulation values (a few percent) transformers normally exhibit. The syllabus formula given here is this standard approximation, and it is what you should use unless a problem explicitly asks for the exact phasor treatment.

Does maximum efficiency depend on the power factor of the load?

The load fraction for maximum efficiency, $x_{max\,\eta}=\sqrt{P_i/P_{cu,FL}}$ , does not depend on power factor at all — it comes purely from balancing $P_i$ against $x^2P_{cu,FL}$ , and $\cos\theta$ cancelled out of that condition in the derivation. However, the value of maximum efficiency itself does depend on power factor, since output power $xS\cos\theta$ appears in the numerator: a lower power factor gives lower output for the same current, so maximum efficiency is smaller at a poorer power factor even though it still occurs at the same $x$ .

Why is iron loss treated as exactly constant when the load current is changing?

Because iron loss depends only on peak flux density $B_{max}$ and frequency $f$ , both of which are fixed once the transformer is connected to its rated supply voltage — the transformer's EMF equation ties flux to the applied voltage, not to the secondary load current. As long as the transformer stays connected to a constant-voltage, constant-frequency supply (the normal operating assumption in this course), the core flux and hence the iron loss do not change no matter how much or how little current the load draws, which is exactly the assumption that makes the $P_i$ term in the efficiency formula a fixed number rather than a function of $x$ .