GATE Electrical: Power Systems (faults, stability, dispatch, protection), previous-year questions solved
Real Power Systems questions from the last five GATE Electrical papers, part 2: sequence impedances and fault currents from sequence networks and Zbus, the swing equation and equal-area criterion, governor droop and AGC, economic load dispatch, and relay characteristics. 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. Line models and load flow are on the first 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 · 16 questions
Which set of positive, negative and zero-sequence impedances fits a transposed 220 kV line?
- (A)1.1, 0.15 and 0.08
- (B)0.15, 0.15 and 0.35
- (C)0.2, 0.2 and 0.2
- (D)0.1, 0.3 and 0.1
Answer: (B) Official key: B
- A line is a static, symmetrical element: positive- and negative-sequence currents see the same mutual coupling, just with phase order reversed, so . That rules out (A) and (D).
- Zero-sequence currents are in phase in all three conductors and return through earth and earth wires. The mutual flux adds instead of cancelling, so is much larger than , typically 2 to 3.5 times. That rules out (C).
- (B) 0.15, 0.15, 0.35 is the only valid set.
Contrast with a synchronous machine, where (the negative-sequence field rotates against the rotor) and is the smallest of the three.
Protection engineers need to set earth-fault distance relays: the zero-sequence compensation factor is entered into every distance relay on a line. Utilities measure line during commissioning because earth resistivity makes it hard to calculate accurately.
Bolted b-c fault at bus 1 of a generator-transformer-motor system: what is the fault current?
Answer: 7.22 p.u. Official key: 7.1 to 7.3
- A line-to-line fault connects the positive- and negative-sequence networks in parallel at the fault point; the zero-sequence network is not involved (no earth path), so figure (iv) is a distractor.
- Thevenin reactance at bus 1 in the positive network: the path back through the line and generator () in parallel with the motor (): . The negative network has the same values: .
- Pre-fault currents neglected, so the prefault voltage is p.u.: .
- Fault current p.u. (key 7.1 to 7.3).
Two traps: forgetting the motor's contribution (it also feeds the fault, giving ), and adding the zero-sequence network as you would for a single line-to-ground fault.
Short-circuit studies like this set the breaking capacity of every breaker and the settings of every relay in a substation. Motors and other rotating loads feed fault current for a few cycles, which is why IEC 60909 and IEEE short-circuit methods include motor contributions; leaving them out undersizes switchgear.
Fault through j0.007 at bus 2 of a known Zbus: what is the voltage at bus 1 during the fault?
Answer: 0.390 pu Official key: 0.380 to 0.400
- No pre-fault loading, so every bus starts at pu.
- Fault current at bus 2: pu.
- Voltage change at bus 1 is set by the transfer impedance : pu.
Zbus makes this a two-line calculation: the diagonal entry gives the fault current, and the column of the faulted bus gives the voltage at every other bus. Using instead of (a common slip) gives 0.41, outside the key.
This is how voltage-dip studies work: a fault on one bus pulls down voltages across the network, and sensitive loads (drives, data centres, process plants) a few buses away can trip. Fault-ride-through rules for wind and solar plants exist because of exactly these remote voltage dips.
Symmetrical components for a two-phase network: which matrices S are valid?
- (A)
- (B)
- (C)
- (D)
Answer: (A) and (D) Official key: A; D
- For phases, symmetrical components use the -th roots of unity. For two phases these are and , so each sequence is a balanced set: both phases equal (in phase) or equal and opposite.
- So every column of must have entries of equal magnitude, and the two columns must be independent so that any , can be decomposed ( invertible).
- (A): columns and : valid. (D): columns and : valid (same two sequences, ordered and signed differently).
- (B) is singular (both columns equal). (C) has a column , which is not a balanced set.
Answer: (A) and (D).
The same idea, generalised, is the Fortescue transform used for every unbalanced-fault calculation on three-phase grids, and its real-valued cousins (Clarke's alpha-beta transform) run inside every motor-drive and grid-inverter controller.
Unbalanced bus voltages all rotate by the same small angle: what happens to the zero-sequence component?
- (A)The zero sequence components in Case-1 and Case-2 have the same phase angle and magnitude
- (B)The magnitude of the zero sequence component in Case-1 is greater than that in Case-2
- (C)The magnitude of the zero sequence component in Case-2 is greater than that in Case-1
- (D)The zero sequence components in Case-1 and Case-2 have the same magnitude but different phase angles
Answer: (D) Official key: D
- Zero sequence: .
- In Case-2 every phasor is multiplied by the same : .
- Same magnitude, phase shifted by : option (D). (The same holds for the positive and negative sequence components.)
The set in the figure is unbalanced (different magnitudes), so ; that is what makes the phase angle meaningful. A common rotation of all phasors just changes the time reference.
This is why protection relays and phasor measurement units (PMUs) judge unbalance from sequence-component magnitudes, which stay put when the whole system's angle drifts. India's wide-area monitoring system streams such synchrophasors from PMUs at major substations to the grid control centres.
20 MVA, 4-pole alternator with H = 15 MJ/MVA, 15 MW in and 10 MW out: what is the rotor's acceleration?
Answer: 75 mechanical degree/s² Official key: 74 to 76
- Stored kinetic energy ; accelerating power .
- Swing equation in electrical degrees: with .
- electrical degree/s2.
- A 4-pole machine has 2 pole pairs, so mechanical angles are half the electrical ones: mechanical degree/s2.
Cross-check in SI units: with rad/s gives the same 75. The 11.2 kV rating is not needed. Forgetting the pole-pair conversion gives 150, outside the key.
The inertia constant H is what keeps grid frequency from collapsing in the first seconds after a big plant trips. As solar and wind (with no spinning mass) replace coal units, system inertia falls, which is why grid operators now track inertia in real time and buy fast frequency response from batteries.
Critical clearing angle π/2, post-fault curve (Pmax/√2) sin δ: what is the accelerating area?
Answer: 0.11 Pmax Official key: 0.10 to 0.12
- At the critical clearing angle the accelerating area exactly equals the largest decelerating area available, so compute the decelerating area (it needs no during-fault curve).
- Post-fault, the rotor can swing up to where on the far side: , (135°).
- Decelerating area:
- , so , and the accelerating area is (key 0.10 to 0.12).
The initial angle (, °) is not needed: the problem is built so you never need the during-fault curve.
Critical clearing time is why transmission protection must clear faults in about 100 ms (5 cycles at 50 Hz) on major lines: slower clearance lets the accelerating area grow past what the network can absorb, and the generator loses synchronism. Transient-stability programs integrate the swing equation for every machine to find these limits.
250 MW at 6% droop and 400 MW at 6.4% share 500 MW: how much does the 250 MW unit carry?
Answer: 200 MW Official key: 188 to 192 OR 198 to 202
- Droop in Hz per MW: unit 1 , unit 2 .
- In parallel both units run at the same frequency. If both start from the same no-load frequency, each one's frequency drop is the same : , so .
- Unit 1 carries (unit 2: 300 MW).
The official key also accepts 188 to 192 because the question does not say where the no-load frequencies sit. If instead each unit reaches 50 Hz at its own full load (no-load set-points Hz and Hz), equal frequency gives , so MW. Both readings are in the key.
Droop is how hundreds of generators share a load change without talking to each other: each governor responds in proportion to its rating and droop setting. India's grid code sets droop at 3 to 6% for most units, and the operator then restores 50 Hz with secondary control (AGC) by shifting set-points.
AGC and AVR on a synchronous generator: which statements about their effects are correct?
- (A)AGC response has a local effect on frequency while AVR response has a global effect on voltage.
- (B)AGC response has a global effect on frequency while AVR response has a local effect on voltage.
- (C)AGC regulates the field current of the synchronous generator while AVR regulates the generator’s mechanical power input.
- (D)AGC regulates the generator’s mechanical power input while AVR regulates the field current of the synchronous generator.
Answer: (B) and (D) Official key: B; D
- Frequency is one number for the whole synchronous grid: real power injected anywhere changes it everywhere. So AGC, which balances real power, has a global effect on frequency.
- Voltage is a local quantity, held up by reactive power, and reactive power does not travel far over inductive lines. So the AVR mainly affects voltage near its own bus. (B) is true, (A) is false.
- AGC acts through the governor set-point, i.e. the turbine's mechanical power input; the AVR acts on the exciter, i.e. the field current. (D) is true, (C) swaps them.
Answer: (B) and (D). This is the P-f / Q-V decoupling that also underlies the fast-decoupled load flow.
India's grid operator runs AGC from the national control centre, sending set-points to participating plants every few seconds to hold frequency within the 49.90 to 50.05 Hz band. Voltage control stays regional: AVRs, transformer taps and reactors managed close to each substation.
Two 600 MW plants, λ = 8 Rs/MWh, 550 MW demand: what is the optimal split?
- (A)200, 350
- (B)250, 300
- (C)325, 225
- (D)350, 200
Answer: (B) Official key: B
- Optimal dispatch (no losses, no limits hit): every plant runs at the same incremental cost .
- Plant 1: , so .
- Demand balance: : option (B).
- Check: plant 2 needs , so , a sensible value; both outputs are within 600 MW.
The unknown is a red herring: plant 1's curve and the demand fix the answer.
Grid-India runs security-constrained economic despatch (SCED) across inter-state thermal stations on this principle, re-dispatching every 15-minute block towards the cheapest units subject to network limits. Plant heat-rate curves supply the incremental costs.
Unit 1's quadratic cost coefficient rises 10% at the same 290 MW load: what happens to PG1 and total cost?
- (A) will decrease and F will increase
- (B)Both and F will increase
- (C) will increase and F will decrease
- (D)Both and F will decrease
Answer: (A) Official key: A
- Find : at the optimum the incremental costs are equal. , and , so .
- Raise by 10%: unit 1's incremental cost climbs faster, so the optimum shifts load to unit 2. Re-solving gives MW: it decreases.
- Total cost: was Rs/h and becomes Rs/h: it increases. It must: unit 1 is dearer at every output, so even the new best split costs more than the old optimum.
- Option (A).
This is what happens on a real dispatch desk when a plant's coal quality drops or its heat rate worsens: its incremental-cost curve steepens, the scheduler moves load to other units, and the system's total cost still rises. Merit-order dispatch in India's market (based on variable cost) follows the same logic.
Incremental cost Rs 7,400/MWh on these two curves: what is the common load?
Answer: 35 MW Official key: 35 to 35
- Read the lines: Gen A rises from 8000 Rs/MWh at 0 MW to 11200 at 160 MW, so ; Gen B rises from 6000 to 12400, so (both reach 10000 at 100 MW, where they cross).
- At : Gen A's cheapest megawatt already costs 8000, more than 7400, so Gen A produces nothing: .
- Gen B: , so MW.
- Common load MW.
Trap: solving gives MW and a load of 5 MW. A generator cannot run below zero, so the equal-incremental-cost rule only applies to units that are actually generating.
Real dispatch adds exactly this logic as minimum and maximum limits: below some load a unit is backed down to its technical minimum or taken off. Indian coal units have a technical minimum around 55% of rating, which is why low-demand hours (solar noon) create real scheduling problems.
Two units seen at λ = 10,000 and 10,600 Rs/MWh: what is the plant output at 11,800?
Answer: 216 MW Official key: 216 to 216
- Quadratic cost curves give straight-line incremental costs, so each unit's output is linear in . Unit 1: 30 MW more for Rs 600 more, i.e. 1 MW per Rs 20. Unit 2: 12 MW per Rs 600, i.e. 1 MW per Rs 50.
- At (Rs 1,800 above 10,000): unit 1 would want MW, above its 100 MW rating, so it is held at 100 MW.
- Unit 2: MW, within its 150 MW rating.
- Plant output MW.
Trap: ignoring the rating gives MW. That is why the question says 'without overloading any unit'.
Once a unit hits its maximum it drops out of the equal-incremental-cost set and the remaining units pick up the extra load at a higher λ. Dispatch software tracks these limits for every unit every 15-minute block, along with ramp-rate limits (MW per minute) that this question leaves out.
Which relay protects an alternator against loss of excitation?
- (A)offset Mho relay.
- (B)over current relay.
- (C)differential relay.
- (D)Buchholz relay.
Answer: (A) Official key: A
- When the field is lost, the machine keeps running as an induction generator, draws reactive power from the grid, and the impedance seen from its terminals moves into the fourth quadrant of the R-X plane, roughly between and .
- An offset mho relay (a circle offset downwards along the axis) covers exactly that region: option (A).
- Overcurrent relays do not see it (stator current may stay near rated), differential relays only see internal winding faults, and the Buchholz relay is for oil-filled transformers.
Loss-of-field protection (ANSI device 40) is fitted to every utility-scale generator because running without excitation overheats the rotor and can pull down system voltage. Modern numerical generator relays combine two offset mho zones with an undervoltage check.
Parallel lines between bus 1 and bus 2: which relays must be directional?
- (A)R3 and R4 are directional overcurrent relays blocking faults towards bus 2
- (B)R3 and R4 are directional overcurrent relays blocking faults towards bus 2 and R7 is directional overcurrent relay blocking faults towards bus 3
- (C)R3 and R4 are directional overcurrent relays blocking faults towards Line 1 and Line 2, respectively, R7 is directional overcurrent relay blocking faults towards Line 3 and R5 is directional overcurrent relay blocking faults towards bus 2
- (D)R3 and R4 are directional overcurrent relays blocking faults towards Line 1 and Line 2, respectively.
Answer: (A) Official key: A
- The only source is behind bus 1, so power normally flows bus 1 → bus 2 → bus 3. Everything beyond bus 2 is radial: plain overcurrent relays graded by time handle Line 3 and the loads. No directional relay is needed at B5, B6, B7 or B8.
- The parallel lines are the problem. For a fault on Line 1, current also flows from bus 1 down Line 2, into bus 2 and back up Line 1 through B3. The relay at B4 sees that fault current and, without direction, would trip healthy Line 2.
- Make R3 and R4 directional so they operate only for current flowing from bus 2 into their own line (towards bus 1) and block faults towards bus 2. Then a Line 1 fault trips B1 and B3 only.
- R1 and R2 at the source end can stay non-directional (fault current through them always flows away from bus 1). Option (A).
Directional overcurrent relays (ANSI 67) are standard at the receiving ends of parallel feeders and ring mains in sub-transmission networks. On higher-voltage lines, distance relays with communication schemes take over, but the same 'which way is the fault' logic is built in.
A reactance relay with X ≤ 1 Ω: what is its characteristic in the G-B plane?
- (A)
- (B)
- (C)
- (D)
Answer: (A) Official key: A
- , so , and .
- Therefore .
- becomes , i.e. , i.e. : option (A).
Geometrically: inversion maps the straight line to a circle through the origin, centre , radius ; the operating region is outside it. Checked by sampling 4,000 random impedances: only (A) agrees with every time.
Reactance relays are used for short lines, where arc resistance would push a mho relay's measured impedance out of its zone; measuring only X makes them immune to that resistance. Numerical distance relays implement this as the 'reactance line' of a quadrilateral characteristic.