AC Analysis & Three-Phase Systems — Engineering Reference

AC Analysis & Three-Phase Systems

1. At a glance

AC analysis is the procedure for solving linear circuits driven by sinusoidal sources in the sinusoidal steady state. Instead of integrating differential equations in the time domain, every sinusoid of a single frequency ω is replaced by a complex number — its phasor — and every passive element is replaced by its complex impedance Z(jω). Kirchhoff’s laws still hold in the phasor domain, so all of the DC machinery (series/parallel reduction, voltage/current dividers, Thevenin/Norton, mesh, nodal, superposition) carries over verbatim, with complex arithmetic in place of real arithmetic. AC analysis is the foundation for everything that operates above DC: power distribution, transformers, motors and drives, filters, op-amp frequency response, transmission lines, antennas, and the bulk of utility-scale electrical engineering.

Three-phase is the dominant utility-scale power topology. Three sinusoidal sources phase-shifted 120° from one another deliver constant instantaneous total power (the time variation of one phase is exactly cancelled by the other two), produce uniformly rotating magnetic fields without needing brushes or commutators, and use roughly half the conductor copper of an equivalent set of single-phase circuits. North America and the western Pacific operate at 60 Hz; the rest of the world plus parts of western Japan operate at 50 Hz. The transition from a single-phase v(t) = V_m cos(ωt+φ) to a three-phase set {v_a, v_b, v_c} with v_b lagging v_a by 120° and v_c lagging by 240° is the bridge between desktop electronics and the grid.

A practising engineer reaches for this material every day: sizing a motor branch circuit, picking decoupling capacitors against a known noise spectrum, designing a notch filter, computing power-factor correction for an industrial load, sizing a transformer KVA rating, or sanity-checking the line currents on a 480/277 V panel feeding a mix of resistive and inductive load. The DC foundation lives at [[Engineering/circuit-analysis]]; this note picks up where that one ends.

2. First principles

Sinusoidal sources. A general AC voltage:

v(t) = V_m · cos(ωt + φ), units volts (V)

• V_m is the peak (amplitude), in V.
• ω = 2πf is the angular frequency, in rad/s. f is in hertz (Hz = 1/s).
• φ is the phase angle, in radians (or degrees by convention in power-engineering practice).
• Period T = 1/f = 2π/ω.

Sine and cosine differ only in phase reference (sin(x) = cos(x − π/2)); circuit-analysis convention is cosine. Power engineers default to cosine reference; signal-processing convention sometimes uses sine. Pick one and stay consistent.

Root-mean-square (RMS) values. RMS is the DC-equivalent value that delivers the same average power to a resistor:

V_rms = √(⟨v²⟩) = V_m/√2 (for a pure sinusoid)

I_rms = I_m/√2 similarly.

For non-sinusoidal periodic waveforms (square, triangle, distorted) the form factor V_m/V_rms differs. A “True-RMS” multimeter measures √(⟨v²⟩) directly; an average-responding meter assumes a sinusoid and reads correctly only for sinusoids.

All AC voltages and currents in power engineering are quoted as RMS unless explicitly labelled “peak.” “120 V residential” means 120 V_rms ≈ 170 V_peak ≈ 340 V_peak-to-peak.

Real average power into a resistor. P = V_rms · I_rms = V_rms² / R = I_rms² · R. Cosine convention drops out — the ½ from time-averaging cos² is already embedded in the √2 of RMS.

Phasors. A phasor is a complex number representing the magnitude and phase of a sinusoid at a known frequency. The Euler identity e^(jθ) = cos(θ) + j·sin(θ) lets us write:

v(t) = V_m · cos(ωt + φ) = Re{ V_m · e^(j(ωt+φ)) } = Re{ (V_m · e^(jφ)) · e^(jωt) }

The bracketed term V = V_m · e^(jφ) = V_m ∠φ is the phasor. By convention power engineers use RMS-magnitude phasors: V = V_rms ∠φ. The e^(jωt) carrier is implicit and dropped from all equations — every quantity in a phasor calculation rotates at the same ω, so the rotation factors cancel.

A phasor diagram plots V and I as arrows in the complex plane, all rotating counter-clockwise at ω. The relative angles between phasors are what matter — the reference (which phasor lies on the positive real axis) is chosen for convenience.

Complex impedance. Impedance is the phasor-domain generalization of resistance. For each passive element:

• Resistor: Z_R = R (real, frequency-independent). Current is in phase with voltage.
• Inductor: Z_L = jωL (positive imaginary, grows with frequency). Voltage leads current by 90° — equivalently, current lags voltage in an inductor.
• Capacitor: Z_C = 1/(jωC) = −j/(ωC) (negative imaginary, falls with frequency). Voltage lags current by 90° — equivalently, current leads voltage in a capacitor.

The general impedance Z = R + jX has a real part R (resistance, units Ω) and an imaginary part X (reactance, units Ω). X > 0 is inductive; X < 0 is capacitive. The magnitude |Z| = √(R² + X²) and angle ∠Z = arctan(X/R) determine how much the element scales current and shifts its phase relative to voltage.

The reciprocal Y = 1/Z = G + jB is admittance (siemens, S), with conductance G and susceptance B.

Memory aid: “ELI the ICE man.”

• Electromotive force leads current in an inductor (L): E-L-I.
• Current leads electromotive force in a capacitor (C): I-C-E.

Why phasors work. Differentiation in the time domain becomes multiplication by jω in the phasor domain: d/dt → jω. So v_L = L · di/dt becomes V_L = jωL · I_L, which is Ohm’s law with Z_L = jωL. Every linear differential equation in t with sinusoidal forcing collapses to a linear algebraic equation in the complex variables — analytically tractable, numerically cheap.

KCL and KVL in the phasor domain. Unchanged. ΣI_into_node = 0 and ΣV_around_loop = 0 still hold, where the I and V are now phasors and the sums are complex.

Three-phase sources. Three sinusoidal voltages of equal magnitude, displaced 120° (= 2π/3 rad):

• v_a(t) = V_m · cos(ωt)
• v_b(t) = V_m · cos(ωt − 120°)
• v_c(t) = V_m · cos(ωt − 240°) = V_m · cos(ωt + 120°)

In phasors with v_a as reference: V_a = V_rms ∠0°, V_b = V_rms ∠−120°, V_c = V_rms ∠+120°. This is the positive (abc) phase sequence. Reversing two phases gives negative (acb) sequence — important because three-phase motors reverse direction when you do this.

Y (wye / star) connection. Three sources share a common neutral point; each phase voltage v_a, v_b, v_c is measured from neutral to its line terminal. Line-to-line voltage V_LL is the difference between two phase voltages and is √3 times the line-to-neutral voltage V_LN, with a 30° phase shift:

V_LL = √3 · V_LN ∠V_ab = ∠V_a + 30°

A “208/120 V” or “480/277 V” or “400/230 V” label always means V_LL / V_LN of a wye system.

Δ (delta) connection. Three sources connected end-to-end in a triangle. No neutral. Phase voltage equals line-to-line voltage (V_phase = V_LL), but the line current is √3 times the phase current (the current circulating inside the delta):

I_line = √3 · I_phase (in delta)

Power generation is universally wye-connected with grounded neutral; loads can be either. Three-phase motors are commonly Y for starting (lower starting current) and switched to Δ for run (higher torque) in the “wye-delta starter” topology.

3. Practical math / design equations

Impedance combinations

Series and parallel rules from DC still work, with complex Z replacing real R:

• Series: Z = Z₁ + Z₂ + … + Z_n (add real and imaginary parts separately)
• Parallel: 1/Z = 1/Z₁ + 1/Z₂ + … + 1/Z_n
• Two in parallel: Z = Z₁·Z₂ / (Z₁ + Z₂) (complex multiplication and division)

Worked numerical example: a 10 Ω resistor in series with a 25 mH inductor at 60 Hz.

• ω = 2π·60 = 377 rad/s.
• Z_R = 10 Ω.
• Z_L = jωL = j·377·0.025 = j9.42 Ω.
• Total Z = 10 + j9.42 Ω = 13.73 Ω ∠43.3°.

If a 120 V_rms source drives this load:

• I = V/Z = 120 ∠0° / (13.73 ∠43.3°) = 8.74 ∠−43.3° A_rms.
• Current lags voltage by 43.3° — inductive load.

Voltage and current dividers

Same form as DC, with complex impedances:

V_out = V_in · Z₂ / (Z₁ + Z₂)

The magnitude and phase of V_out both follow from complex arithmetic. This is the workhorse of filter design — an RC pair becomes a first-order low-pass, an RLC becomes a band-pass, etc.

Real, reactive, and apparent power

In the phasor domain, complex power is defined for any two-terminal element:

S = V · I* = P + jQ

where I* is the complex conjugate of the current phasor. Units VA. The choice of conjugate is a sign convention that makes Q positive for inductive (lagging) loads, which is universal in power engineering.

• P = |V|·|I|·cos(θ) — real (active) power, in watts (W). Energy actually consumed by the load (heat, mechanical work, light).
• Q = |V|·|I|·sin(θ) — reactive power, in volt-amperes reactive (VAR). Energy oscillating between source and reactive elements (L and C); no net work done but real conductor current flows.
• |S| = |V|·|I| — apparent power, in volt-amperes (VA). Magnitude of S; sets conductor and transformer sizing.

with θ = ∠V − ∠I the phase angle by which voltage leads current.

Power triangle:

|S|² = P² + Q² tan(θ) = Q/P

Power factor

pf = cos(θ) = P / |S|

• pf = 1: purely resistive load. All current does useful work.
• pf < 1 lagging (inductive load, θ > 0, Q > 0): motors, fluorescent ballasts, induction heating. Standard situation; the load is consuming Q.
• pf < 1 leading (capacitive load, θ < 0, Q < 0): over-excited synchronous motors, large capacitor banks, some power-electronic loads. The load is supplying Q to the grid.

Low pf means the utility has to ship more current to deliver a given P, sizing conductors and transformers larger than the actual W consumption would suggest. Industrial utility tariffs typically penalise pf below 0.90–0.95.

Resonance

Series RLC resonance (R, L, C all in series). At the resonant angular frequency:

ω₀ = 1/√(L·C) f₀ = 1/(2π√(L·C))

the inductive and capacitive reactances cancel: jωL = −j/(ωC), so Z = R (purely real, minimum). Current is maximum and in phase with voltage. Quality factor:

Q = ω₀·L / R = 1/(ω₀·R·C) = (1/R)·√(L/C)

The half-power bandwidth is BW = ω₀/Q. High Q → sharp resonance peak → narrow filter; low Q → broad peak → wide filter. For radio receivers Q is typically 50–200; for power-system filters 5–30.

Parallel RLC resonance. Same ω₀, but at resonance the parallel combination has maximum impedance (looks like an open circuit). Q in the parallel case is Q = R/(ω₀·L) = ω₀·R·C — note the reciprocal form.

Bode plots and frequency response

The transfer function H(jω) = V_out/V_in is a complex function of ω. Bode plots show |H| (in dB, where dB = 20·log₁₀|H|) and ∠H (in degrees) versus log frequency. Asymptotic rules:

• Each pole at ω_p contributes a slope change of −20 dB/decade above ω_p, and a phase shift of −90° over a range from ω_p/10 to 10·ω_p centred at ω_p (where ∠H crosses −45°).
• Each zero contributes +20 dB/decade and +90°.
• At the pole/zero break frequency itself, the magnitude is −3 dB / +3 dB relative to the asymptote (3.01 dB exact; “−3 dB” = half power, the conventional bandwidth definition).

A first-order RC low-pass (R in series, C across output) has H(jω) = 1/(1 + jωRC), cutoff f_c = 1/(2πRC), −20 dB/decade roll-off above f_c, −90° asymptotic phase.

Three-phase power

For a balanced three-phase load, the total three-phase real power is:

P_3φ = √3 · V_LL · I_L · cos(θ)

with V_LL the line-to-line voltage (RMS), I_L the line current (RMS), and θ the per-phase load angle (same for all three phases in a balanced system). Works for both Y and Δ loads — the √3 absorbs the geometry. Likewise:

Q_3φ = √3 · V_LL · I_L · sin(θ) |S_3φ| = √3 · V_LL · I_L

Per-phase real power is P_phase = P_3φ / 3.

Per-phase analysis

For a balanced three-phase system, analyse one phase only (treat as a single-phase circuit between phase terminal and neutral), then copy the result to the other two with ±120° phase rotations. Δ-connected loads are converted to their Y equivalent for analysis using Z_Y = Z_Δ / 3. This reduction turns three-phase problems into single-phase problems and is the standard practice in power-system studies.

Worked example 1 — single-phase RL series load

A 240 V_rms, 60 Hz source feeds a 100 Ω resistor in series with a 0.5 H inductor. Find the current magnitude and phase, the real/reactive/apparent power, and the power factor.

• ω = 2π·60 = 377 rad/s.
• Z_L = jωL = j·377·0.5 = j188.5 Ω.
• Z_total = 100 + j188.5 Ω.
• |Z| = √(100² + 188.5²) = √(10 000 + 35 532) = √45 532 = 213.4 Ω.
• ∠Z = arctan(188.5/100) = 62.05°.
• I = V/Z = 240 ∠0° / (213.4 ∠62.05°) = 1.125 ∠−62.05° A_rms.

So |I| = 1.125 A_rms, lagging the voltage by 62.05° (inductive — current lags voltage in an inductor, as expected).

Power:

• pf = cos(62.05°) = 0.469 lagging.
• |S| = V·I = 240 · 1.125 = 270.0 VA.
• P = |S| · cos(θ) = 270.0 · 0.469 = 126.6 W.
• Q = |S| · sin(θ) = 270.0 · 0.883 = 238.5 VAR.
• Check: P² + Q² = 126.6² + 238.5² = 16 028 + 56 882 = 72 910; √72 910 = 270.0 VA. ✓
• Alternate check: P = I²·R = 1.125²·100 = 126.6 W. ✓

The 100 Ω resistor dissipates 126.6 W as heat. The inductor exchanges 238.5 VAR with the source twice per cycle, doing no net work but loading the 240 V source to 1.125 A. A purely resistive 126.6 W load at 240 V would draw only 0.528 A, so the inductor more than doubles the conductor current for the same useful power — the canonical motivation for power-factor correction.

Worked example 2 — three-phase load with power-factor correction

A three-phase, 4-wire, 208/120 V wye system (V_LL = 208 V, V_LN = 120 V, 60 Hz) feeds two parallel balanced loads:

• Load A: 5 kW resistive heater at pf = 1.0.
• Load B: 10 kW induction motor at pf = 0.8 lagging.

Find the total line current and design a capacitor bank to correct system pf to 0.95 lagging.

Step 1 — sum powers.

• P_A = 5 kW, Q_A = 0 VAR.
• P_B = 10 kW; |S_B| = P_B / pf = 10 / 0.8 = 12.5 kVA; Q_B = |S_B| · sin(arccos 0.8) = 12.5 · 0.6 = 7.5 kVAR.
• Total before correction: P = 15 kW, Q = 7.5 kVAR, |S| = √(15² + 7.5²) = √281.25 = 16.77 kVA; pf = 15/16.77 = 0.894 lagging.

Step 2 — line current before correction.

• I_L = |S| / (√3 · V_LL) = 16 770 VA / (√3 · 208 V) = 16 770 / 360.3 = 46.5 A_rms per line.

Step 3 — capacitor bank sizing.

• Target pf = 0.95 → target θ’ = arccos(0.95) = 18.19° → target Q’ = P · tan(θ’) = 15 kW · 0.3287 = 4.93 kVAR.
• Capacitor must supply Q_C = Q − Q’ = 7.5 − 4.93 = 2.57 kVAR (capacitors deliver Q to the system; equivalently they consume negative Q).
• For a balanced three-phase capacitor bank (Y-connected, one capacitor per phase, between phase and neutral), each phase capacitor supplies Q_C/3 = 0.857 kVAR.
• Per-phase capacitance: Q_C_phase = V_LN² · ω · C → C = Q_C_phase / (V_LN² · ω) = 857 / (120² · 377) = 857 / 5 428 800 = 1.58 × 10⁻⁴ F = 158 µF per phase.

For a Δ-connected capacitor bank the per-leg voltage is V_LL = 208 V, so C_Δ = Q_C_phase / (V_LL² · ω) = 857 / (208² · 377) = 857 / 16 314 752 = 5.25 × 10⁻⁵ F = 52.5 µF per leg (smaller capacitance, but rated for higher voltage — a typical power-engineering trade-off).

Step 4 — line current after correction.

• New |S’| = √(15² + 4.93²) = √(225 + 24.3) = √249.3 = 15.79 kVA.
• New I_L’ = 15 790 / 360.3 = 43.8 A_rms per line — a 5.8 % reduction.

Conductors, breakers, and the upstream transformer all see less current; the heater’s and motor’s behaviour is unchanged.

Worked example 3 — 60 Hz notch filter for instrumentation hum

Design a passive LC parallel-resonant trap to attenuate 60 Hz mains hum on a small-signal instrumentation line that has a 1 kΩ source impedance.

Topology: L in parallel with C placed in series with the signal path. At ω₀ the parallel LC presents very high impedance, blocking 60 Hz; at all other frequencies it presents low impedance and lets the signal through.

Design:

• f₀ = 60 Hz → ω₀ = 377 rad/s.
• ω₀² · L · C = 1, so L·C = 1/ω₀² = 1/142 129 = 7.036 × 10⁻⁶ s².
• Pick C = 10 µF (commonly available film cap): L = 7.036 × 10⁻⁶ / 10⁻⁵ = 0.7036 H ≈ 700 mH. (Practical: an audio-grade iron-core choke.)
• Or pick L = 100 mH: C = 7.036 × 10⁻⁶ / 0.1 = 70.4 µF (still feasible, larger cap).

Q and bandwidth: Q = ω₀·R_loss / (ω₀·L)? — for a parallel RLC the loss resistance is in parallel with the LC. If the inductor has DCR = 30 Ω and the capacitor ESR ≈ 0, the dominant loss is the inductor. Convert to the parallel equivalent at ω₀: R_p ≈ (ω₀·L)² / DCR = (377·0.7036)² / 30 = 265.4² / 30 = 70 437 / 30 = 2 348 Ω. Then Q = R_p / (ω₀·L) = 2 348 / 265.4 = 8.85. Bandwidth BW = f₀/Q = 60 / 8.85 = 6.8 Hz, so attenuation is significant from about 56.6 to 63.4 Hz.

Attenuation at 60 Hz: Z_trap at resonance ≈ R_p = 2.35 kΩ. With the 1 kΩ source impedance and (say) a 10 kΩ load, the voltage at the load is V_out = V_in · 10 kΩ / (1 kΩ + 2.35 kΩ + 10 kΩ) = V_in · 0.749. Without the trap V_out = V_in · 10/(10+1) = 0.909. The trap attenuates 60 Hz by 20·log(0.749/0.909) = −1.7 dB — modest. A higher-Q inductor (lower DCR) gives much sharper attenuation. For deep mains rejection (>40 dB) the practical answer is active notch filtering (twin-T with op-amp, or digital filter on the ADC side) — passive LC traps are limited by inductor Q at audio frequencies.

4. Reference data

Common AC voltage standards

Region	Residential	Commercial 3φ	Distribution MV	Frequency
US / Canada	120/240 V split-phase	208Y/120 V, 480Y/277 V	4.16 kV, 12.47 kV, 13.8 kV, 34.5 kV	60 Hz
Mexico, Brazil, Korea, Taiwan	120 V or 127 V	220 V, 480 V	similar to US	60 Hz
EU / UK / most of Asia	230 V single-phase (was 220 / 240)	400Y/230 V	11 kV, 22 kV, 33 kV	50 Hz
Japan	100 V	200 V (east 50 Hz, west 60 Hz)	6.6 kV, 22 kV, 66 kV	50 / 60 Hz
Industrial heavy / mining	—	4.16Y/2.4 kV, 6.9 kV, 13.8 kV	up to 34.5 kV	regional
Transmission	—	—	69, 115, 138, 161, 230, 345, 500, 765 kV (US)	regional

“208Y/120” syntax means a wye system with line-to-line 208 V and line-to-neutral 120 V. The 480Y/277 V system is universal in US commercial/light-industrial — motors run on 480 V_LL, fluorescent ballasts on 277 V_LN. IEC 60038 codifies preferred voltages internationally.

Standard transformer ratings (kVA)

Type	Common single-phase ratings (kVA)	Common three-phase ratings (kVA)
Pole-mounted distribution	5, 10, 15, 25, 37.5, 50, 75, 100, 167, 333	30, 45, 75, 112.5, 150, 225
Pad-mounted distribution	25, 50, 75, 100, 167	75, 112.5, 150, 225, 300, 500, 750, 1000, 1500, 2000, 2500
Dry-type indoor (low voltage)	—	15, 30, 45, 75, 112.5, 150, 225, 300, 500, 750, 1000
Substation power (oil)	—	5 MVA, 10, 15, 20, 25, 50, 75, 100 MVA

Standard ratings are codified in ANSI C57.12.00 (US) and IEC 60076. Picking a rating just above the calculated kVA load with some margin (typically 25 % growth, plus standard derating for ambient) is the design rule.

Reactive power compensation

Equipment	Typical pf	Typical sizing
Induction motor (no load)	0.20–0.30 lagging	Q ≈
Induction motor (full load)	0.80–0.92 lagging	depends on slip, frame
Synchronous motor (under-excited)	0.80–0.95 leading	adjustable via field current
Welder, arc furnace	0.30–0.70 lagging	highly variable
LED / electronic ballast (good driver)	0.90–0.99	mandatory pf > 0.9 by ENERGY STAR / EN 61000-3-2
Switching power supply (no PFC)	0.50–0.70	distorted, large harmonic Q
Capacitor bank (Y or Δ)	leading by definition	sized 10–40 % of transformer rating
Static VAR compensator (SVC)	continuously variable	utility scale
STATCOM	continuously variable	utility scale, fast

Skin depth in copper (high-frequency conductor sizing)

Frequency	Skin depth δ in Cu	Implication
60 Hz	8.5 mm	Solid copper bars are uniformly conducting; no concern below about 8 AWG.
1 kHz	2.1 mm	Litz wire starts to help in transformers.
100 kHz	0.21 mm	Switching converters use Litz wire or thin foil windings.
1 MHz	66 µm	RF circuit conductors must be plated, not just bulk.
1 GHz	2.1 µm	Surface roughness on PCB copper matters; “rolled” vs “electrodeposited” copper differs in loss.

δ = √(2ρ/(ωµ)) with ρ = 1.68×10⁻⁸ Ω·m, µ = 4π×10⁻⁷ H/m. AC resistance of a thin-walled tube of radius a is R_AC = ρ/(2πa·δ), so doubling frequency multiplies AC resistance by √2 once skin effect dominates.

5p. Theory

Fourier representation. Any periodic waveform v(t) with period T can be written as a sum of sinusoids at integer multiples of the fundamental f₀ = 1/T:

v(t) = a₀ + Σ_{n=1}^∞ [ a_n · cos(n·ω₀·t) + b_n · sin(n·ω₀·t) ]

For a linear circuit, by superposition, the response to v(t) is the sum of the responses to each Fourier component. So AC analysis at a single frequency is the building block for analysing any periodic signal in linear circuits: compute H(jω) at each harmonic, multiply by the harmonic’s amplitude, sum at the output.

Harmonics. Non-linear loads (switching power supplies without PFC, six-pulse rectifiers, variable-frequency drives, arc furnaces) draw non-sinusoidal current even from a sinusoidal voltage. The current spectrum contains the fundamental plus higher-order components — the harmonics. A typical six-pulse rectifier injects 5th, 7th, 11th, 13th, … (the 6k±1 harmonics) at predictable amplitudes (1/n of fundamental for an ideal rectifier).

Total Harmonic Distortion (THD). The standard scalar measure of waveform purity:

THD = √(Σ_{n=2}^∞ V_n²) / V_1 (often expressed as a percentage)

with V_n the RMS of the n-th harmonic and V_1 the fundamental. THD_V (voltage THD at a node) and THD_I (current THD in a branch) are distinct; both are regulated. IEEE 519-2022 limits voltage THD at the point of common coupling (PCC) to 5 % for general-purpose systems and current TDD to 5–20 % depending on the load’s short-circuit ratio. IEC 61000-3-2 limits per-harmonic current injection at the device terminals (rather than at PCC).

Laplace-domain extension. Phasor analysis is the special case of Laplace-transform analysis evaluated on the jω axis (s = jω). The general transfer function H(s) = V_out(s)/V_in(s) is a rational function of s with poles and zeros in the complex plane. Stability is the condition that all poles lie in the left half-plane (Re(s) < 0). For sinusoidal steady state we evaluate H(s) at s = jω; for transient response we use the full s-plane and inverse Laplace. The two views are unified in the standard control-theory framework — see [[Engineering/digital-control]].

Symmetrical components (Fortescue 1918). Any unbalanced set of three phasors {V_a, V_b, V_c} can be uniquely decomposed into three balanced sets:

• Positive sequence (V₁): equal magnitude, 120° apart, abc rotation. Normal balanced operation.
• Negative sequence (V₂): equal magnitude, 120° apart, acb rotation. Caused by unbalanced loads or single-line-to-line faults; produces a counter-rotating field in induction motors that causes heating without doing work.
• Zero sequence (V₀): three phasors in phase (no rotation). Flows only in systems with a neutral or ground return. Caused by ground faults.

Conversion uses a = e^(j120°) = −0.5 + j(√3/2):

[V₀; V₁; V₂] = (1/3) · [[1,1,1]; [1,a,a²]; [1,a²,a]] · [V_a; V_b; V_c]

Power-system fault analysis is entirely formulated in symmetrical components, because the sequence networks decouple for balanced impedances. Phase-domain analysis is a nightmare for any non-trivial unbalanced case.

Passive vs active filters. Passive filters use only R, L, C — no power gain, bounded amplitude response, no DC blocking issues but bulky inductors at low frequency. Active filters use op-amps to synthesize transfer functions without inductors, achieve sharp roll-off, can have gain, but need a power supply and have finite gain-bandwidth product. Sallen-Key, multiple-feedback, state-variable, and switched-capacitor topologies cover the practical design space. See [[Engineering/op-amps]].

Three-phase rotating field. A three-phase stator winding excited by balanced three-phase currents produces a magnetic field of constant magnitude that rotates at the synchronous speed n_s = 120·f/p (rpm), where f is the supply frequency and p is the number of poles. This is the physical basis of every AC motor — induction, synchronous, permanent-magnet — and is impossible with a single-phase supply without an auxiliary starting winding (and even then the rotating field is degraded). See [[Engineering/electric-motors]].

6p. Application

Power-factor correction (PFC). Two regimes:

• Passive PFC at the utility scale. Capacitor banks (fixed, switched-in by contactors, or thyristor-switched in SVCs) at industrial substations correct lagging pf back toward 0.95–0.99. Sized per the kVAR formula in §3. Excessive PFC overshoots into leading pf and can resonate with system inductance — utilities monitor capacitor switching transients carefully.
• Active PFC at the device scale. Switching power supplies in laptops, LED drivers, and computer power supplies shape their input current to be sinusoidal and in phase with voltage. A boost-converter front-end with a current-mode control loop is the standard topology, mandated by EN 61000-3-2 for devices >75 W in the EU. Pre-2000 SMPS without PFC presented apparent loads of pf ≈ 0.6 with massive 3rd-harmonic current; modern PFC supplies achieve pf > 0.95 across 90 to 90 % load.

Harmonic mitigation. Tools include passive trap filters tuned to specific harmonics (e.g. 5th and 7th trap branches at industrial PCC), broad-band passive filters, active filters that inject anti-phase harmonic currents, and drive isolation transformers (delta-wye transformers cancel triplen harmonics from VFD inputs). A 12-pulse rectifier (two six-pulse rectifiers phased 30° apart via a Δ-Y/Δ-Δ transformer) cancels 5th and 7th by construction; 18- and 24-pulse topologies cancel further.

Transformer sizing. Pick a kVA rating that exceeds the apparent power demand at design conditions including:

• Load |S| with diversity factor accounted (not every load runs at full nameplate simultaneously).
• 25 % growth margin (typical).
• Ambient temperature derating (sea-level ANSI ratings assume 30 °C ambient, dropping ~1 % per °C above).
• Altitude derating above 1000 m.
• Harmonic K-factor derating for non-linear loads (K-13 transformers for VFD-heavy loads).

For example: a panel computed at 380 kVA with 25 % growth → 475 kVA needed. Next standard size up is 500 kVA — pick that. If the load is 70 % electronic (VFDs, computers) specify a K-13 rated transformer.

Motor starting. A standard induction motor draws 6–8× its full-load current during direct-on-line (DOL) start because the rotor presents low impedance until it accelerates. Starting techniques:

• DOL: simplest, OK for small motors (<25 hp) or stiff utility supplies.
• Wye-delta starter: motor wound for Δ run-voltage; started in Y (gives V_LL/√3 across each winding, so 1/3 of starting current and 1/3 of starting torque), switched to Δ at ~75 % speed.
• Autotransformer starter: steps voltage down to 50–80 % of nominal for the starting period.
• Soft-starter: phase-controlled SCR pairs ramp voltage from ~30 % to 100 % over 5–30 seconds. Now the workhorse.
• Variable-frequency drive (VFD): ramps both voltage and frequency together to maintain V/f constant, holding torque at full value during the entire acceleration. Eliminates inrush entirely. See [[Engineering/power-electronics]].

Three-phase rectifiers. Three-phase six-pulse bridge (six diodes or thyristors) produces a DC output with 6f ripple frequency (360 Hz on 60 Hz mains, 300 Hz on 50 Hz). Average DC = 1.35 · V_LL, ripple amplitude ~4.2 % of average — much smoother than a single-phase rectifier’s 100 % ripple, which is why DC bus capacitors on three-phase drives can be much smaller for the same DC quality.

What changes when going from single-phase to three-phase design:

• Smoother instantaneous power. p_3φ(t) = constant for a balanced resistive load; single-phase pulses at 2f. No mechanical vibration from torque pulsation.
• Smaller conductors and transformers for given P. Three wires carry the same total real power as four single-phase wires (two return paths if you include neutral); about 75 % copper.
• Self-starting motors and rotating fields — single biggest reason three-phase exists.
• No neutral required for balanced loads. The three phase currents sum to zero at the wye point. Unbalanced loads need a fourth wire (the neutral).
• Equipment complexity rises — three sets of breakers, three contactors, balanced loading to consider.

7p. Edge cases & assumptions

Phasor analysis assumes pure sinusoidal steady state. Three things break it:

1. Transients. During motor starts, fault clearing, load steps, transformer energization, the system is not in steady state. Decaying DC offsets, sub-synchronous oscillations, and inrush all live in the time domain. Use SPICE, PSCAD, or EMTP for transient analysis.
2. Non-linear loads. Diode rectifiers, magnetic saturation (transformers driven near knee), corona discharge, arc loads. Current is non-sinusoidal even with sinusoidal voltage. Fourier decomposition + per-harmonic phasor analysis is the workaround at design time; full time-domain simulation is the workaround when accuracy matters.
3. Frequency-dependent parameters. Skin effect in conductors, hysteresis loss in iron cores, dielectric loss in cables. Modelled by frequency-dependent R(ω), L(ω); phasor analysis at one frequency at a time still works.

Unbalanced three-phase requires symmetrical components. A single-line-to-ground fault on phase A produces V_a → 0 but V_b and V_c unchanged. There is no “per-phase equivalent” for this; the three sequence networks (positive, negative, zero) must be solved together. Most utility relaying and protection engineering is formulated in sequence components.

Power-factor sign conventions. The IEC convention treats Q > 0 as absorbed by the load (inductive) and Q < 0 as delivered. Some North American utility documents reverse this in metering contexts (“kVARh delivered” vs “kVARh received”). Always check the metering convention before interpreting utility bills or SCADA values.

Capacitor banks resonate with system inductance. A 600 kVAR capacitor switched onto a stiff bus can form a parallel-resonant circuit with the upstream transformer’s leakage reactance. The resonant frequency is given by f_r = f · √(MVA_sc / Q_cap), with MVA_sc the system short-circuit MVA at the bus. If f_r happens to coincide with a load harmonic (commonly 5th = 300 Hz on 60 Hz mains), the capacitor draws huge harmonic current and fails. Always check the resonant frequency before sizing a capacitor bank. Detuning by adding a series reactor (typically 7 % or 14 % of the capacitor’s reactance) is the standard mitigation.

Tariff penalties on Q. Utilities bill not just for kWh but also for kVAR-hours or for peak kVA. A facility with pf = 0.7 will see its bill rise 30 % over an identical pf = 1.0 facility, motivating the PFC investment.

IEEE 519-2022 harmonic limits at PCC. Voltage THD must be ≤ 5 % at <69 kV bus, ≤ 2.5 % at >161 kV. Current TDD limits depend on the ratio I_sc / I_L (system short-circuit current to maximum load current) and harmonic order. A stiff bus (high I_sc) permits higher harmonic injection from a given customer. The current limits at PCC are the customer’s responsibility; the voltage limits are the utility’s.

Frequency tolerance. North American utilities maintain 60.000 Hz to within ±0.02 Hz under normal conditions; 50 Hz European systems to within ±0.05 Hz. Frequency deviation indicates active-power imbalance (deficit lowers frequency). Modern grids with high renewable penetration see wider excursions; this is one of the standard motivations for inverter-based resource (IBR) grid services and IEEE 1547-2018 ride-through requirements.

Transient over-voltages on capacitor switching. Energizing a discharged shunt capacitor against an inductive source produces a 1-cycle-decay-time over-voltage that can peak at 2.0 p.u. with no damping (resistor-less switching), or 1.4 p.u. with pre-insertion resistors. Repetitive switching can stress motor insulation downstream — particularly inverter-fed motors with their already-stressed cable terminations. Synchronized closing (zero-voltage point on the wave) eliminates the transient and is now the preferred technology for utility capacitor banks.

8p. Tools & software

Linear circuit simulation (sinusoidal AC analysis):

• LTspice / ngspice / PSpice — AC sweep mode computes the Bode plot of any node. Standard for analog circuits up to a few hundred MHz.
• QucsStudio — open-source, RF-flavoured, S-parameter support.
• Multisim — NI, education-flavoured, integrates with hardware.

Power-system analysis (three-phase, transformers, short-circuit, load flow):

• ETAP — commercial, dominant in industrial power engineering (refineries, data centres). Load flow, short-circuit, motor starting, arc-flash, harmonic analysis, protection coordination, dynamic stability — all in one.
• SKM PowerTools — commercial, US-focused, similar feature set to ETAP.
• EasyPower — commercial, North American.
• DIgSILENT PowerFactory — commercial, dominant in European utility transmission planning. Strong on RMS and EMT (electromagnetic transient) co-simulation.
• PSS/E — Siemens, US/Canadian utility transmission planning. Steady-state load flow and dynamic stability.
• PSCAD/EMTDC — Manitoba HVDC Research Centre, EMT-domain (instantaneous time-domain) for HVDC, FACTS, and IBR studies.
• EMTP-RV — Hydro-Québec / EDF, EMT for transmission and protection.
• OpenDSS — open-source from EPRI, distribution-system analysis, quasi-static time-series for high-PV-penetration studies.

Open-source Python ecosystem:

• pandapower — power flow and short-circuit on radial/meshed distribution. Cross-link [[Languages/Tier3/energy-power]].
• PyPSA — unit-commitment, optimal power flow, scenario analysis at transmission and energy-system scale.
• scikit-rf — RF and microwave network analysis (S-parameters, Smith chart, calibration).
• NumPy / SciPy with scipy.signal — generic transfer-function manipulation, Bode, root locus, IIR/FIR filter design.
• MATLAB Simulink with the Power Systems and Simscape Electrical toolboxes — graphical, very mature, expensive.

Bench instruments:

• Vector network analyzer (VNA): measures |H(jω)| and ∠H(jω) directly across a frequency sweep. Sub-$100 hobby-tier (NanoVNA) covers 1 kHz–6 GHz. Lab tier (Keysight E5071C, Rohde&Schwarz ZNB) covers 100 kHz–8.5 GHz with <0.1 dB accuracy.
• Power-quality analyser: measures V, I, P, Q, S, pf, THD, sag/swell, transient events on three-phase services. Fluke 435-II / Hioki PW3198 / Dranetz PowerXplorer. Mandatory for IEEE 519 compliance audits.
• LCR meter: measures component Z at one or a few frequencies. DE-5000 (hobby, <$200, 100 Hz–100 kHz); Keysight E4990A (lab, 20 Hz–120 MHz).
• Three-phase calibrator: Fluke 6105A — generates known V/I/φ for relay/meter calibration.

Modelling references:

• IEEE benchmark systems — the IEEE 9-bus, 14-bus, 30-bus, 39-bus, 118-bus systems are the canonical test cases for power-flow and stability algorithms. Data published in MATPOWER and pandapower.
• CIM (IEC 61970/61968) — the canonical XML data model for transmission and distribution networks; what utility EMS systems exchange. See [[Languages/Tier3/energy-power]].

11. Cross-references

• [[Engineering/circuit-analysis]] — the DC foundation. KCL, KVL, Thevenin, Norton, mesh, nodal — all carry into the phasor domain unchanged.
• [[Engineering/electric-motors]] — induction, synchronous, and PMSM motors as three-phase AC loads driven by the rotating field.
• [[Engineering/transformers-power-systems]] — phasor and per-unit analysis is the working language; turns ratios, impedance reflection, Y-Δ phase shifts.
• [[Engineering/electromagnetics-engineering]] — Maxwell’s equations origin of L (Faraday) and C (Gauss), and the basis for skin depth and transmission-line behaviour.
• [[Engineering/power-electronics]] — VFDs, inverters, rectifiers, PFC converters; the switching breaks the sinusoidal assumption, replaced by averaged-model state-space analysis.
• [[Engineering/op-amps]] — bandwidth, slew rate, and stability all evaluated via H(jω) and Bode plots.
• [[Engineering/digital-control]] — sampling and aliasing connect time-domain to frequency-domain; phasor steady state is the s = jω slice of Laplace.
• [[Engineering/rf-design]] — phasor analysis extended to distributed parameters; characteristic impedance, standing waves, Smith chart.
• [[Engineering/semiconductor-devices]] — small-signal models live in the phasor domain at the device’s operating point.
• [[Robotics/power-systems]] — battery → DC link → inverter → three-phase motor; the full AC/DC/AC path of a modern mobile robot.
• [[Languages/Tier3/energy-power]] — CIM/CGMES/IEEE 2030.5/OpenADR data models for the bulk power system and DER integration.
• [[Languages/Tier3/industrial-automation]] — IEC 61850 substation communications, IEC 61131 PLC programming for protection and control.

12. Citations

• Hayt, W. H., Kemmerly, J. E. & Durbin, S. M. (2018). Engineering Circuit Analysis (9th ed.). McGraw-Hill. Chapters on AC steady-state, phasors, complex power, three-phase circuits.
• Irwin, J. D. & Nelms, R. M. (2020). Basic Engineering Circuit Analysis (12th ed.). Wiley. Strong worked examples for phasors and three-phase.
• Sadiku, M. N. O. & Alexander, C. K. (2020). Fundamentals of Electric Circuits (7th ed.). McGraw-Hill.
• Glover, J. D., Overbye, T. J. & Sarma, M. S. (2016). Power System Analysis & Design (6th ed.). Cengage. Per-unit, symmetrical components, load flow, short-circuit, transient stability.
• Bergen, A. R. & Vittal, V. (2000). Power Systems Analysis (2nd ed.). Prentice-Hall. Classical rigorous treatment of three-phase and per-unit.
• Stevenson, W. D. (1982). Elements of Power System Analysis (4th ed.). McGraw-Hill. Historic textbook; still the cleanest exposition of symmetrical components.
• Chapman, S. J. (2020). Electric Machinery Fundamentals (5th ed.). McGraw-Hill. The three-phase chapter is one of the best practitioner introductions to balanced and unbalanced systems.
• Mohan, N., Undeland, T. M. & Robbins, W. P. (2003). Power Electronics: Converters, Applications, and Design (3rd ed.). Wiley. Rectifiers, inverters, PFC, drive electronics — the AC-DC-AC path in modern systems.
• Kundur, P. (1994). Power System Stability and Control. McGraw-Hill. Reference for dynamic and transient stability analysis; deep on synchronous-machine modelling.
• Fortescue, C. L. (1918). “Method of Symmetrical Co-ordinates Applied to the Solution of Polyphase Networks.” Transactions of the AIEE 37(2): 1027–1140. The original paper; readable today.
• IEEE Std 519-2022. IEEE Standard for Harmonic Control in Electric Power Systems. The reference for THD limits at the point of common coupling.
• IEEE Std 1547-2018. IEEE Standard for Interconnection and Interoperability of Distributed Energy Resources with Associated Electric Power Systems Interfaces. Inverter ride-through, voltage and frequency support, grid services.
• IEC 60038:2009+A1:2021. IEC standard voltages. Source of the harmonised LV/MV/HV nominal voltages.
• IEC 61000 series. Electromagnetic compatibility (EMC). Per-device harmonic current limits (61000-3-2), voltage flicker (61000-3-3), immunity test methods.
• NFPA 70-2023. National Electrical Code (NEC). The US installation code; Article 220 (load calculations), Article 310 (conductor ampacity), Article 430 (motors), Article 460 (capacitors).
• ANSI/IEEE C57.12.00-2021. Standard for General Requirements for Liquid-Immersed Distribution, Power, and Regulating Transformers. Standard kVA ratings, impedance ranges, temperature limits.
• Horowitz, P. & Hill, W. (2015). The Art of Electronics (3rd ed.). Cambridge University Press. Chapter 1 on impedance and frequency response remains the most readable introduction to phasor thinking for practising electronics engineers.