Mean and RMS values

Plug an AC ammeter into a household circuit and it reads a steady number — yet the current behind that number is switching direction fifty times a second, tracing out a sine wave that swings between $+I_0$ and $-I_0$ . Average that wave over a full cycle and you get exactly zero, which is obviously useless for describing "how much current is flowing." So how does a single meaningful number survive? This article works through the two answers physics actually uses. The half-cycle mean rescues the idea of an average by only looking at one hump of the sine wave, giving $I_{mean} = 2I_0/\pi$ . The RMS (root-mean-square) value, $I_{rms} = I_0/\sqrt2$ , goes further: it is built specifically to match how heating and power actually accumulate, which is why every AC meter, every power rating, and the "230 V" on your wall socket are all RMS numbers, not peak or mean ones.

The core insight

For a sinusoidal AC quantity the useful "size" is not the peak $I_0$ but $I_{rms} = I_0/\sqrt2 \approx 0.707\,I_0$ , because RMS is defined so that $I_{rms}^2 R$ equals the true average power dissipated, unlike the half-cycle mean $I_{mean} = 2I_0/\pi \approx 0.637\,I_0$ , which is a different, less useful number.

The core idea

An alternating current driven by an AC generator varies sinusoidally in time:

i(t) = I_0\sin\omega t

where: $I_0$ is the peak (maximum) value of the current and $\omega = 2\pi f$ is the angular frequency, with $f$ the frequency in hertz. The same form applies to the EMF, $\varepsilon = \varepsilon_0\sin\omega t$ , and to voltage.

Full-cycle mean is zero, and that is a real problem. Over one complete cycle the current is positive for exactly half the time and negative, equal in size, for the other half, so the two halves cancel perfectly. A meter that reported this "mean" would always read zero for any AC supply, which tells you nothing about the current's actual strength.

The half-cycle mean fixes this by only averaging one hump. Averaging $I_0\sin\omega t$ from $t=0$ to $t=\pi/\omega$ (one half-period, where the sine stays positive) gives

I_{mean} = \frac{2I_0}{\pi} \approx 0.637\,I_0

This is a legitimate "typical size" of the current, but it is rarely the number engineers actually need, because electrical heating and power depend on the square of the current, not on the current itself.

RMS is built around that squaring. It is defined as the square root of the mean of the squares of the instantaneous current over a cycle, $I_{rms} = \sqrt{\langle i^2\rangle}$ . Since $i^2 = I_0^2\sin^2\omega t$ and the average of $\sin^2\omega t$ over a full cycle is exactly $\tfrac12$ (squaring removes the sign, so this average is never zero), $\langle i^2\rangle = I_0^2/2$ and

I_{rms} = \frac{I_0}{\sqrt2} \approx 0.707\,I_0

The identical relation holds for voltage, $V_{rms} = V_0/\sqrt2$ . Notice $I_{rms} > I_{mean}$ for a sinusoid: $0.707$ versus $0.637$ of the peak — squaring-then-rooting always weighs the larger instantaneous values more heavily than a plain linear average does. The figure below puts the three landmark values — peak, half-cycle mean, and RMS — on one sine wave so their relative heights are visible at a glance.

Figure

A sine wave for one cycle of alternating current with three horizontal reference lines marking the peak value, the RMS value at about seventy one percent of peak, and the half cycle mean value at about sixty four percent of peak

Why RMS, not mean, decides power. RMS is defined so that a steady DC current of value $I_{rms}$ dissipates exactly the same average heat in a resistor as the real AC does. So $I_{rms}$ is the AC current's "effective" value for power:

\bar P = I_{rms}^2R = \frac{I_0^2}{2}R

where: $\bar P$ is average power, $R$ is resistance. This is exactly why AC ammeters and voltmeters are built to read RMS, and why "230 V AC" is an RMS rating — the instantaneous voltage actually swings up to $230\times\sqrt2 \approx 325$ V.

Derivation ADeriving the RMS value of a sinusoidal current

Show that the root-mean-square value of a sinusoidal current $i = I_0\sin\omega t$ , taken over a full cycle, works out to $I_0/\sqrt2$ , and along the way see clearly why this squaring-averaging-rooting procedure is what correctly captures the current's power-delivering strength rather than its plain linear average.

Step 1. Square the instantaneous current. $i^2 = I_0^2\sin^2\omega t$ . Squaring is the crucial move: it makes every instant's contribution positive, so the negative half-cycle no longer cancels the positive half-cycle the way it does for the plain mean.

Step 2. Average $\sin^2\omega t$ over one full cycle. Using the identity $\sin^2\theta = \tfrac12(1-\cos2\theta)$ , the average of $\cos2\theta$ over a full cycle is zero (it is itself a complete oscillation), so the average of $\sin^2\omega t$ is $\tfrac12(1-0) = \tfrac12$ . This is a fixed number, independent of $I_0$ or $\omega$ — true for any sinusoid over an integer number of cycles.

Step 3. Take the mean of the square. $\langle i^2\rangle = I_0^2\times\langle\sin^2\omega t\rangle = I_0^2\times\tfrac12 = \dfrac{I_0^2}{2}$ .

Step 4. Take the square root to undo the squaring from Step 1. $I_{rms} = \sqrt{\langle i^2\rangle} = \sqrt{\dfrac{I_0^2}{2}} = \dfrac{I_0}{\sqrt2} \approx 0.707\,I_0$ . The same three-step "square, average, root" chain applies unchanged to voltage, giving $V_{rms} = V_0/\sqrt2$ .

Worked example

The peak value of the voltage in a household AC supply is 325 V, and this supply feeds a pure resistor of 20 Ω. Find (a) the RMS voltage, (b) the mean voltage over half a cycle, (c) the RMS current through the resistor, and (d) the average power dissipated.

Step 1. RMS voltage: $V_{rms} = \dfrac{V_0}{\sqrt2} = \dfrac{325}{1.414} = 229.9\ \text{V} \approx 230\ \text{V}$ . This confirms that "230 V mains" is exactly the RMS rating of a 325 V peak supply.

Step 2. Half-cycle mean voltage: $V_{mean} = \dfrac{2V_0}{\pi} = \dfrac{2\times325}{3.1416} = \dfrac{650}{3.1416} = 206.9\ \text{V}$ . Note this is a genuinely different, smaller number than the RMS value — about 11% lower — and it is not what appliances are rated for.

Step 3. RMS current through the resistor, using Ohm's law with RMS values (valid for a pure resistor): $I_{rms} = \dfrac{V_{rms}}{R} = \dfrac{229.9}{20} = 11.5\ \text{A}$ .

Step 4. Average power dissipated: $\bar P = I_{rms}^2R = (11.5)^2\times20 = 132.3\times20 = 2645\ \text{W}$ . Check using peak values directly: $\bar P = \tfrac12\dfrac{V_0^2}{R} = \tfrac12\times\dfrac{325^2}{20} = \tfrac12\times5281 = 2641\ \text{W}$ ✓ (small rounding difference from carrying 229.9 V forward). The RMS voltage is 230 V, the half-cycle mean is 207 V, the RMS current is 11.5 A, and the average power dissipated is about 2.64 kW.

Common traps

Quoting "the mean value of AC is zero" and stopping there — the full-cycle mean is zero, but the physically useful "mean value" quoted in problems is over a half-cycle and equals $2I_0/\pi$ , not zero.
Confusing RMS with a simple arithmetic average — RMS involves squaring, averaging, then square-rooting, and is always larger than the half-cycle mean for a sinusoid ( $0.707\,I_0$ versus $0.637\,I_0$ ).
Using the peak value directly in power formulas, e.g. writing $P = I_0^2R$ instead of $P = I_{rms}^2R = \tfrac12I_0^2R$ — this overestimates power by a factor of 2.
Forgetting that a rating like "230 V AC" is an RMS value — the instantaneous voltage actually swings between $+325$ V and $-325$ V, not staying at 230 V.
Applying $I_{rms} = I_0/\sqrt2$ to non-sinusoidal waveforms such as square or triangular AC — that specific ratio holds only for a sine wave; other shapes need the mean-of-squares calculation redone for that shape.

Ask the tutor: if you'd like the $\langle\sin^2\omega t\rangle = 1/2$ integral spelled out in full, or a worked comparison of RMS values for square and triangular waveforms, just ask.

Why do we bother with RMS at all if the half-cycle mean also gives a nonzero average

Because power dissipation depends on the square of the current, $P = i^2R$ , not on the current itself. Averaging $i$ first and then squaring it (which is what the half-cycle mean effectively represents) gives a different, smaller number than averaging $i^2$ first and then taking the square root. Only the second procedure — RMS — correctly reproduces the actual average heating effect of the AC, which is why it, not the mean, is used for power ratings, ammeters and voltmeters.

Is the factor 1/root 2 true for any AC waveform, not just sine waves

No. The result $I_{rms} = I_0/\sqrt2$ comes specifically from averaging $\sin^2\omega t$ over a cycle and getting exactly $1/2$ . A square wave that stays at $+I_0$ for half the cycle and $-I_0$ for the other half has $i^2 = I_0^2$ at every instant, so its RMS value is simply $I_0$ itself, with no $\sqrt2$ factor at all. A triangular wave gives yet another ratio. Whenever the waveform changes, you must redo the "square, average, root" calculation for that specific shape.

Why does an AC ammeter read a single steady number if the current keeps reversing direction

Because the meter is built (and calibrated) to respond to the heating or magnetic effect the current produces, both of which depend on $i^2$ rather than on $i$ directly, and $i^2$ never goes negative even though $i$ does. Averaging that squared response over many cycles and taking the square root at the display stage is exactly the RMS calculation, so the needle settles at $I_{rms}$ , not at zero and not at the peak.

Mean and RMS values

The core idea

Worked example

Common traps

Before this

Where this leads