GATE Electrical: Power Systems (lines and load flow), previous-year questions solved
Real Power Systems questions from the last five GATE Electrical papers, part 1: conductor GMR, the nominal-π line and its power limit, voltage drop and power-factor correction, line losses, Ybus row sums, bus types, DC and Newton-Raphson load flow and the Ybus of a phase-shifting transformer. Each one has a step-by-step solution that matches the official answer key, the traps that cost marks, and where the idea shows up in real grid and plant work. Faults, stability, dispatch and protection are on the second Power Systems page (linked on this page).
Questions from official GATE EE papers (IIT Kharagpur 2022, IIT Kanpur 2023, IISc Bengaluru 2024, IIT Roorkee 2025, IIT Guwahati 2026); solutions by Skill-Lync. · Updated 16 Nov 2026 · 13 questions
Four equal strands of radius r packed in a square: what is the conductor's geometric mean radius?
- (A)
- (B)
- (C)
- (D)
Answer: (D) Official key: D
- GMR is the -th root of the product of all distances between strands, where each strand's distance to itself is replaced by (this accounts for the flux inside the strand).
- Touching strands have centres apart; the two diagonal pairs are apart. Every strand sees the same set: itself (), two neighbours (, ) and one diagonal ().
- : option (D).
Trap: using instead of for the self term gives , which is not an option; picking (the centre spacing) ignores the averaging.
The same GMR idea sizes bundled conductors: extra-high-voltage lines (400 kV and above) hang 2 to 4 or more sub-conductors per phase on spacers. A bigger GMR lowers series inductance (more power over the same line) and the larger effective radius cuts surface gradient and corona loss.
400 km, 275 kV line on the nominal-π model: at what load angle is received power maximum?
Answer: 83.64° Official key: 83 to 84
- Receiving-end power of a two-port line with : , where and .
- Only the first term depends on , so is largest at .
- In the nominal- model , the total series impedance: .
- (key 83 to 84). A numerical sweep of over with the full and peaks at the same angle.
The capacitance changes (here ) but not the angle of maximum power. For a lossless line and the familiar comes back; resistance pulls it below .
Real lines are never run near this steady-state limit: planners keep the angle across a corridor well below it for transient-stability margin, and 400 km at 275 kV would normally be modelled as a long line with distributed parameters. Grid studies in PSS/E, PowerFactory or the open-source MATPOWER use the exact ABCD model for such lengths.
11 kV, 10 MVA generator feeding a 0.866 pf load through a 5 Ω line: what is the terminal power factor?
- (A)0.63 lagging.
- (B)0.87 lagging.
- (C)0.63 leading.
- (D)0.87 leading.
Answer: (A) Official key: A
- Rated current ; rated phase voltage at the terminals .
- Take as reference. The load voltage leads it by : . The line adds .
- Terminal voltage , and : solving the quadratic gives per phase.
- , at ahead of the current: power factor lagging, option (A).
The 30 Ω synchronous reactance is a distractor: it sits behind the terminals and changes the internal EMF, not the terminal power factor. The line's reactive drop is what makes the generator see a worse power factor than the load.
Every inductive line consumes reactive power (), so a generator feeding a remote load supplies more MVAr than the load asks for. That is why grid codes ask generators to hold a power-factor range at their terminals (typically 0.85 lagging to 0.95 leading) and why long lines get shunt compensation.
Feeder with R/X = 5: at what load power factor is the voltage drop largest?
Answer: 0.981 Official key: 0.975 to 0.985
- Approximate voltage drop: . For a fixed kVA (fixed current) only the bracket changes with power factor.
- Maximise : setting the derivative to zero gives .
- (lagging), inside the key 0.975 to 0.985.
The 350 kVA is not needed. The intuition: the drop is worst when the current lines up with the feeder impedance angle, which for a resistive feeder () is almost unity power factor.
Low-voltage and 11 kV distribution feeders are mostly resistive (small conductors, short spans), so in India's rural feeders it is real power, not reactive power, that drives the voltage drop. Capacitor banks help much less there than on high-voltage lines, and utilities use conductor upgrades, feeder bifurcation or on-load tap changers instead.
4 kV motor, 390 kVA at 0.77 pf corrected to 0.85: what line current does the capacitor bank draw?
Answer: 9.1 A Official key: 8.5 to 10.0
- Real power stays the same: .
- Reactive power before: . After: .
- Capacitor bank rating .
- Line current of the bank at 4 kV: .
Common slip: using (the old kVA instead of the kW) for . The key's wide range (8.5 to 10.0) allows for rounding of the trigonometric values.
Plant engineers size capacitor banks exactly like this, usually in automatic steps, because many utilities bill industrial consumers on kVAh or add a low power-factor penalty. Over-correcting to a leading power factor raises voltage at light load, which is why the target is usually 0.95 to 0.99 lagging, not unity.
Same end voltages, Z = 0.75∠0° at 0° vs Z = 0.75∠90° at 90°: which line losses are larger?
- (A)Real power loss in Case-1 is more than that in Case-2
- (B)Real power loss in Case-2 is more than that in Case-1
- (C)Reactive power loss in Case-1 is more than that in Case-2
- (D)Reactive power loss in Case-2 is more than that in Case-1
Answer: (A) and (D) Official key: A;D
- Case-1: p.u. The line is a pure resistance, so real loss p.u. and reactive loss .
- Case-2: , , so p.u. The line is a pure reactance: real loss , reactive loss p.u.
- Real loss is larger in Case-1 (): (A) true, (B) false. Reactive loss is larger in Case-2 (): (D) true, (C) false.
Answer: (A) and (D). A resistance can only absorb real power and a reactance only reactive power, whatever the current.
On transmission lines X is 5 to 15 times R, so heavily loaded lines absorb large amounts of reactive power and the receiving-end voltage sags. Grid operators counter this with switched shunt capacitors, SVCs and STATCOMs, and by asking generators to raise their reactive output during peak load.
Reading shunt capacitors off a 3-bus Ybus: which line-charging statements can NOT be true?
- (A)Line charging capacitor of finite value is present in all three lines
- (B)Line charging capacitor of finite value is present in line 2-3 only
- (C)Line charging capacitor of finite value is present in line 2-3 only and shunt capacitor of finite value is present in bus 1 only
- (D)Line charging capacitor of finite value is present in line 2-3 only and shunt capacitor of finite value is present in bus 3 only
Answer: (A) and (C) Official key: A, C
- The off-diagonal terms give the series line admittances: , , .
- The total shunt admittance at bus is the row sum: bus 1: ; bus 2: ; bus 3: (positive = capacitive).
- Bus 1 has zero net shunt admittance. With no shunt inductor anywhere, nothing can cancel a capacitor at bus 1, so there is no line charging on lines 1-2 and 1-3 and no capacitor at bus 1. (A) and (C) both put capacitance at bus 1: they can NOT be true.
- Line 2-3 charging puts equal halves at buses 2 and 3 ( each in the model), which explains bus 2 exactly; the remaining at bus 3 is a bus-3 shunt capacitor. That is (D), and (B) does not rule it out, so both are possible.
Answer: (A) and (C).
Load-flow programs build Ybus exactly this way from line data, so a wrong sign on a line-charging entry shows up as a phantom shunt at a bus. Checking row sums against the known shunt devices is a quick sanity test engineers run on network models imported from utility data.
A current source held at 90° to its bus voltage: how are the three buses classified for load flow?
- (A)Bus 1: Slack bus; Bus 2: bus; Bus 3: bus
- (B)Bus 1: bus; Bus 2: bus; Bus 3: Slack bus
- (C)Bus 1: bus; Bus 2: bus; Bus 3: Slack bus
- (D)Bus 1: bus; Bus 2: bus; Bus 3: Slack bus
Answer: (D) Official key: D
- Bus 3 is an infinite bus with fixed and angle: it is the slack (reference) bus.
- Bus 2: the alternator injects a known 200 MW and a known 40 MVAr, while its voltage is unknown: a - bus, even though a generator sits on it.
- Bus 1: the source current is from the bus voltage, so . Real power is known (zero), is held, and the reactive power is whatever the controller needs: a - bus.
- So Bus 1 PV, Bus 2 PQ, Bus 3 slack: option (D).
The trap is to label buses by the equipment on them (generator = PV). Load flow labels them by what is specified.
This is exactly how a STATCOM or synchronous condenser is modelled: it exchanges almost no real power and holds a bus voltage by injecting reactive current, so it enters a load flow as a PV bus with P = 0. A generator dispatched at fixed P and Q (for example a wind farm on power-factor control) is a PQ bus.
DC load flow on a 3-bus triangle: when is the flow from bus 1 to bus 2 exactly zero?
- (A)
- (B)
- (C)
- (D)
Answer: (A) Official key: A
- With 1 pu voltages and small angles, flow on a line is (the DC load-flow approximation).
- Zero flow from bus 1 to bus 2 means .
- Then all of leaves through line 1-3 and all of through line 2-3: , .
- Dividing, with : , so : option (A).
Note that drops out: when the two ends of a line sit at the same angle, its reactance does not matter. (Checked by solving the DC load flow for 200 random parameter sets per option: only (A) always gives zero flow.)
DC load flow is what planning and market studies fall back on when they need thousands of cases fast, for example screening contingencies or estimating how a trade loads each transmission corridor (power transfer distribution factors come straight from it). The same zero-flow logic is how phase-shifting transformers are set to steer power off an overloaded line.
Ring of four buses with branches 4z, z, 2z, 4z: which branch to open for minimum loss?
- (A)
- (B)
- (C)
- (D)
Answer: (C) Official key: C
- Opening one branch leaves a radial tree, and each branch then carries the sum of the load currents downstream of it. Loss , where is the branch's multiple of .
- Open : carries , , : loss .
- Open : , , : .
- Open : (left bus only), , : .
- Open : , , : .
- Minimum: open , option (C).
Intuition: keep the low-impedance branch doing the heavy lifting and avoid routing current through the long path twice.
Distribution networks are built as meshes but run radially through normally-open points, and choosing where to open them (network reconfiguration) is a daily loss-reduction tool for distribution utilities. Distribution management systems solve this same problem over hundreds of switches.
During a power cut, how is a home UPS represented in a Newton-Raphson load flow?
- (A)Slack bus
- (B)PV bus
- (C)PQ bus
- (D)PQV bus
Answer: (A) Official key: A
- During the outage the UPS is the only source in this small islanded network.
- It sets the voltage magnitude and the reference angle (its inverter is a voltage source), and it must supply whatever real and reactive power the lights and fans draw, plus losses. Nobody specifies its P or Q in advance.
- A bus with fixed and angle and free P and Q is the slack bus: option (A).
A PV bus needs a specified P, which the UPS does not have, and 'PQV bus' is not a standard bus type.
The same reasoning applies to islanded microgrids: the grid-forming inverter or diesel set that holds voltage and frequency is the slack, and other sources (solar inverters in grid-following mode) are PQ buses. Getting this wrong is a classic reason a microgrid load-flow study does not converge.
N buses with P generator buses: what is the size of the Newton-Raphson Jacobian?
- (A)
- (B)
- (C)
- (D)
Answer: (B) Official key: B
- One of the P generator buses is the slack: its and angle are known.
- The other buses each have an unknown angle: unknowns (and real-power equations).
- Only the Q load buses have an unknown voltage magnitude (generator buses hold ): more unknowns (and reactive-power equations).
- Jacobian size square: option (B).
Check with a 6-bus system with 2 generators: 5 angles + 4 magnitudes = 9 unknowns, and (option (C) would give 8). Choose the test case with care: with 5 buses and 2 generators, (B) and (C) both give 7. The 'no reactive limit violations' condition matters: a generator that hits its MVAr limit becomes a PQ bus and the Jacobian grows by one.
For a national grid model with tens of thousands of buses the Jacobian is huge but more than 99% zeros, which is why production load-flow solvers rely on sparse matrix factorisation and fast-decoupled variants. The same matrix also gives voltage-stability indicators near the nose of the PV curve.
Transformer with complex ratio VP/VS = 0.866 + j0.5: what is its Ybus?
- (A)
- (B)
- (C)
- (D)
Answer: (D) Official key: D
- Model: an ideal ratio at the primary, then to the secondary bus. Behind the ideal ratio the voltage is .
- Current through towards the secondary: , so .
- An ideal transformer passes complex power unchanged: , so .
- Here , so , , : option (D).
The key point: a phase-shifting transformer makes Ybus non-symmetric (), because the ratio is complex. Option (C), with both off-diagonals equal, is what you get if you forget the conjugate. Verified by computing , from the physical model for random , .
Phase-shifting transformers are installed on interconnectors to push or hold back power flow, and star-delta transformers carry a built-in 30° shift. Load-flow software stores the off-diagonal entries separately for exactly this reason, and a symmetric-Ybus shortcut gives wrong flows on any network with a phase shifter.