A long transmission line is best modelled using

distributed parameters, using the exact ABCD constants from the propagation constant

nominal-pi with lumped shunt admittance at each end

distributed parameters, using the exact ABCD constants from the propagation constant

Classical economic load dispatch, ignoring transmission loss, allocates load among generating units such that

every operating unit's incremental fuel cost is equal

every unit generates the same power

every operating unit's incremental fuel cost is equal

the cheapest unit alone supplies the entire load

every unit runs at its rated capacity

A lightly loaded, long transmission line experiences the Ferranti effect, meaning its receiving-end voltage tends to

rise above the sending-end voltage

fall well below the sending-end voltage

rise above the sending-end voltage

stay exactly equal to the sending-end voltage

In a bus admittance matrix Ybus, the off-diagonal element Yij (i not equal to j) for two buses with no direct line between them is

the sum of all line admittances

equal to the diagonal element Yii

Which of the following are correctly true when comparing Gauss-Seidel and Newton-Raphson load flow?

Newton-Raphson converges quadratically, needing far fewer iterations for large systems.

Gauss-Seidel's per-iteration computational effort is generally lower than Newton-Raphson's.

Newton-Raphson requires building and solving a Jacobian matrix at every iteration.

Newton-Raphson converges quadratically, needing far fewer iterations for large systems.

Gauss-Seidel's per-iteration computational effort is generally lower than Newton-Raphson's.

Newton-Raphson requires building and solving a Jacobian matrix at every iteration.

Gauss-Seidel is the preferred method for very large modern transmission networks.

A symmetrical three-phase fault at a bus produces fault current in

the positive-sequence network only

all three sequence networks equally

the negative-sequence network only

For a line-to-line (LL) fault, the sequence networks are connected

positive and negative in parallel, zero absent

positive, negative and zero all in series

positive and negative in parallel, zero absent

positive-sequence network alone

Differential protection is best described as a

unit scheme, fully selective within its own defined zone

non-unit scheme relying on graded time delays

unit scheme, fully selective within its own defined zone

backup protection only, never primary

a scheme that measures apparent impedance to the fault

By the equal area criterion, a power system's transient stability after a fault clearance improves when

the fault is cleared faster, shrinking the accelerating area

the fault is cleared more slowly

the fault is cleared faster, shrinking the accelerating area

the generator's inertia is reduced

the mechanical input power is increased during the fault

Transmission, Load Flow, Fault Analysis and Protection

Section 6 of the GATE Electrical Engineering (EE) paper, the largest single section of this syllabus and the one that scales everything else in the paper up to a nationwide grid. Start with generation, transmission and distribution basics and the per-unit system that makes a multi-voltage-level network tractable at all; then how a load flow study finds the steady-state operating point of a network too large to solve by mesh or node analysis directly; then symmetrical components, the tool that turns an unbalanced three-phase fault into three independent balanced problems; and finally the protection apparatus — relays and circuit breakers — that clears a fault before it can do lasting damage, and the stability question of whether the system as a whole stays in synchronism afterward.

1. Generation, transmission, economic dispatch and distribution

A transmission line is modelled, depending on its length, as a short line (series impedance only), a medium line (nominal-pi or nominal-T, adding half the shunt admittance at each end or the middle), or a long line (distributed parameters, needing the exact ABCD parameters derived from hyperbolic functions of the propagation constant) — the choice of model is a length-based approximation trade-off, not a different physical phenomenon. HVDC transmission is preferred over AC for very long distances or asynchronous interconnections because it has no reactive power / stability limit tied to line length and no need for the two ends to be in synchronism, at the cost of expensive converter stations at each end.

Economic load dispatch allocates total demand among several generating units to minimise total fuel cost, and the classical result (ignoring transmission loss) is equal incremental cost: every operating unit's incremental fuel cost (the derivative of its cost curve) must be equal, because if one unit's next MW is cheaper than another's, shifting load toward the cheaper unit reduces total cost. Including transmission loss adds a penalty factor to each unit's incremental cost, since a unit far from the load centre costs more per delivered MW once line losses are counted — the two versions of the same optimisation, with and without loss, are a standard pairing of questions.

ℹ️ Compensation and per-unit quantities

Series compensation (a capacitor bank in series with the line) cancels part of the line's own series inductive reactance, raising power transfer capability and improving stability margin; shunt compensation (a reactor or capacitor bank in parallel, at the receiving end or mid-line) manages voltage — a shunt reactor absorbs excess reactive power on a lightly loaded line (which otherwise rises in voltage from the Ferranti effect), while a shunt capacitor supplies reactive power to support voltage on a heavily loaded one. The per-unit system expresses every quantity as a fraction of a chosen base, chosen so that transformers of different voltage ratios vanish from a per-unit one-line diagram (their turns ratio becomes 1:1 in per-unit, provided the base voltages on the two sides are chosen in that same ratio) — this is the entire reason per-unit exists rather than working in actual amperes and volts.

2. Bus admittance matrix, Gauss-Seidel and Newton-Raphson load flow

The bus admittance matrix Ybus relates bus current injections to bus voltages, I = Ybus * V, and is built directly from the network's line and shunt admittances without needing to invert anything: the diagonal entry Yii is the sum of all admittances connected to bus i, and the off-diagonal Yij is the negative of the admittance directly between buses i and j (zero if no direct line exists) — a sparse matrix for any realistic network, since most bus pairs have no direct connection. A load flow (power flow) study finds the bus voltage magnitudes and angles that satisfy the specified power injections at every bus, classified as slack (voltage and angle specified, absorbs the system's unknown loss), PV (voltage magnitude and real power specified, typically a generator bus) or PQ (real and reactive power specified, typically a load bus).

Gauss-Seidel versus Newton-Raphson load flow
Property	Gauss-Seidel	Newton-Raphson
Convergence	linear, slow, more iterations	quadratic, fast, few iterations
Per-iteration effort	small, simple update	large — builds and inverts a Jacobian each time
Sensitivity to initial guess	less sensitive	more sensitive, but rarely a problem in practice
Scaling to large systems	poor — iteration count grows badly with bus count	the industry standard for large networks

🧠 Voltage and frequency control, and power factor correction

In an interconnected power system, frequency is controlled by matching total generation to total demand — a governor's speed-droop characteristic on each generator shares any sudden load change among all running units automatically, and load frequency control (LFC) restores frequency to exactly nominal afterward via a slower secondary loop. Voltage is a more local quantity, controlled bus-by-bus with generator excitation (AVR), shunt reactive compensation and tap-changing transformers — reactive power does not travel far efficiently, which is why voltage control is local while frequency control is system-wide. Power factor correction (adding shunt capacitors near an inductive load) reduces the reactive current a source and its lines must carry for the same real power, cutting losses and the apparent-power (kVA) rating needed upstream.

3. Symmetrical components and fault analysis

Fortescue's theorem says any unbalanced set of three phasors can be resolved into three balanced sets: positive sequence (same phase order as the original, balanced), negative sequence (reverse phase order, balanced) and zero sequence (all three in phase, equal magnitude). A symmetrical fault (three-phase, all phases equally involved) has only positive-sequence current, since the system stays balanced throughout — this is why a three-phase fault is analysed with an ordinary single-phase equivalent and no symmetrical-component machinery is even needed for it, despite the topic living in the same section.

Sequence networks in an unsymmetrical fault
Fault type	Sequence network connection
Single line-to-ground (LG)	positive, negative and zero networks in SERIES
Line-to-line (LL)	positive and negative networks in PARALLEL (zero absent)
Double line-to-ground (LLG)	positive, negative and zero networks all in PARALLEL
Three-phase (LLL / LLLG)	positive network alone; negative and zero carry no current

⚠️ Which fault is worst is not a fixed answer

A single line-to-ground fault usually draws the LARGEST fault current at a solidly grounded generator (because the zero-sequence network there is often low-impedance), while a three-phase fault is conventionally used to size circuit breakers because it is the worst-case SYMMETRICAL condition and historically the easiest to compute by hand — treating 'three-phase fault always gives the highest current' as a universal rule, rather than a system-dependent outcome that must be checked via the sequence networks, is the standard error.

4. Protection, circuit breakers and stability

Protection principles
Principle	How it decides to trip
Overcurrent	current magnitude exceeds a set threshold, often with an inverse time-current curve
Differential	compares current entering and leaving a protected zone; a mismatch means an internal fault
Directional	trips only for fault current flowing in one specified direction
Distance	measures apparent impedance (V/I) to the fault; trips if it falls within a set reach, in zones of increasing reach and time delay

A circuit breaker's job is to interrupt fault current — the hard part is not opening its contacts but extinguishing the arc that forms between them as they separate, which different breaker types do by cooling and de-ionising the arc in oil, air-blast, SF6 gas or vacuum; its rating includes both a continuous current rating and a much larger breaking (interrupting) capacity for fault conditions. Differential protection is the classic example of UNIT protection — it defines an exact zone (a transformer, a generator, a short bus section) and is completely selective, tripping only for a fault genuinely inside that zone, whereas overcurrent and distance protection are NON-UNIT schemes that rely on graded time delays or reach settings between successive zones to achieve selectivity (coordination) instead.

🎯 The equal area criterion

System stability asks whether all synchronous machines remain in synchronism after a disturbance (a fault, a sudden load change); the equal area criterion is a graphical shortcut for transient stability of a single machine against an infinite bus, comparing the accelerating area (while the fault is on, rotor angle increasing, mechanical power exceeding electrical) against the decelerating area available (after clearing, electrical power exceeding mechanical, up to the maximum angle the system can tolerate) — the system remains stable if and only if the available decelerating area is at least as large as the accelerating area already accumulated, which is exactly why FASTER fault clearing (a smaller accelerating area) directly improves stability margin.

Key takeaways

The per-unit system exists to make transformer turns ratios vanish from a one-line diagram — chosen bases on the two sides in the transformer's own ratio, so its per-unit ratio becomes 1:1.
Newton-Raphson converges quadratically (fast, few iterations, heavy per-iteration work) and is the industry standard for large networks; Gauss-Seidel converges linearly (slow but each iteration is cheap).
A symmetrical (three-phase) fault has only positive-sequence current; an LG fault puts the three sequence networks in series, an LL fault puts positive and negative in parallel, and an LLG fault puts all three in parallel.
Differential protection is unit (fully selective within its own zone); overcurrent and distance protection are non-unit, achieving selectivity through graded time or reach settings between zones instead.
The equal area criterion says a system stays stable exactly when the available decelerating area is at least as large as the accelerating area already built up — faster fault clearing shrinks the accelerating area and directly buys stability margin.

Practice questions (10)

Attempt each one before opening the answer. Every explanation names the tempting wrong option as well as the right one, because that is where marks are lost.

A long transmission line is best modelled using
1. series impedance only
2. nominal-pi with lumped shunt admittance at each end
3. distributed parameters, using the exact ABCD constants from the propagation constant
4. a single lumped resistor
Show answer
Answer: C — distributed parameters, using the exact ABCD constants from the propagation constant
A long line's parameters cannot be lumped at a few points without introducing significant error, so it is modelled with distributed parameters and exact hyperbolic-function ABCD constants; short and medium lines use progressively simpler lumped approximations because the error from lumping is small enough over their length to ignore.
Classical economic load dispatch, ignoring transmission loss, allocates load among generating units such that
1. every unit generates the same power
2. every operating unit's incremental fuel cost is equal
3. the cheapest unit alone supplies the entire load
4. every unit runs at its rated capacity
Show answer
Answer: B — every operating unit's incremental fuel cost is equal
Equal incremental cost across all running units is the condition for minimum total cost: if one unit's next megawatt is cheaper, shifting load to it and away from a costlier unit reduces total cost, so the optimum is reached only when no such profitable shift remains — which is exactly when every unit's incremental cost is equal.
A lightly loaded, long transmission line experiences the Ferranti effect, meaning its receiving-end voltage tends to
1. fall well below the sending-end voltage
2. rise above the sending-end voltage
3. stay exactly equal to the sending-end voltage
4. drop to zero
Show answer
Answer: B — rise above the sending-end voltage
A lightly loaded long line's distributed shunt capacitance dominates over its series inductance's loading effect, so the line behaves like a capacitive load on the source and its receiving end can end up at a HIGHER voltage than the sending end — exactly the condition a shunt reactor is installed to correct.
In a bus admittance matrix Ybus, the off-diagonal element Yij (i not equal to j) for two buses with no direct line between them is
1. the sum of all line admittances
2. zero
3. always negative and non-zero
4. equal to the diagonal element Yii
Show answer
Answer: B — zero
Yij is minus the admittance of the direct line between buses i and j; with no direct line, that admittance is zero, so Yij is zero — this is exactly why Ybus is sparse for any real network, where most bus pairs are not directly connected.
Which of the following are correctly true when comparing Gauss-Seidel and Newton-Raphson load flow?
1. Newton-Raphson converges quadratically, needing far fewer iterations for large systems.
2. Gauss-Seidel's per-iteration computational effort is generally lower than Newton-Raphson's.
3. Newton-Raphson requires building and solving a Jacobian matrix at every iteration.
4. Gauss-Seidel is the preferred method for very large modern transmission networks.
Show answer
Answer: A — Newton-Raphson converges quadratically, needing far fewer iterations for large systems.; B — Gauss-Seidel's per-iteration computational effort is generally lower than Newton-Raphson's.; C — Newton-Raphson requires building and solving a Jacobian matrix at every iteration.
The first three are all correctly true — quadratic convergence, cheaper Gauss-Seidel iterations, and Newton-Raphson's Jacobian requirement are the standard comparison. Newton-Raphson, not Gauss-Seidel, is the preferred method for large modern networks, since Gauss-Seidel's iteration count grows too quickly with system size to be practical there.
A symmetrical three-phase fault at a bus produces fault current in
1. the positive-sequence network only
2. the zero-sequence network only
3. all three sequence networks equally
4. the negative-sequence network only
Show answer
Answer: A — the positive-sequence network only
A three-phase fault keeps the system exactly balanced throughout, so by definition there is no negative- or zero-sequence current at all — only the positive-sequence network carries current, which is exactly why this fault type can be solved with an ordinary per-phase equivalent circuit.
For a line-to-line (LL) fault, the sequence networks are connected
1. positive, negative and zero all in series
2. positive and negative in parallel, zero absent
3. zero-sequence network alone
4. positive-sequence network alone
Show answer
Answer: B — positive and negative in parallel, zero absent
A line-to-line fault does not involve ground, so no zero-sequence current path exists at all; positive- and negative-sequence networks are connected in parallel to represent the fault between the two faulted phases — this is the standard sequence-network diagram for an LL fault.
Differential protection is best described as a
1. non-unit scheme relying on graded time delays
2. unit scheme, fully selective within its own defined zone
3. backup protection only, never primary
4. a scheme that measures apparent impedance to the fault
Show answer
Answer: B — unit scheme, fully selective within its own defined zone
Differential protection compares current at the two ends of a precisely defined zone (a transformer, a generator winding, a bus section) and trips only for a genuine internal fault — it needs no coordination with neighbouring zones and is the classic example of a unit protection scheme, unlike overcurrent (graded time) or distance (impedance-measuring, non-unit) protection.
By the equal area criterion, a power system's transient stability after a fault clearance improves when
1. the fault is cleared more slowly
2. the fault is cleared faster, shrinking the accelerating area
3. the generator's inertia is reduced
4. the mechanical input power is increased during the fault
Show answer
Answer: B — the fault is cleared faster, shrinking the accelerating area
Faster clearing directly shrinks the accelerating area accumulated while the fault is on (rotor angle advances for less time before the fault is removed), leaving more of the available decelerating area as margin — the whole point of the equal area criterion's stability condition, decelerating area at least equal to accelerating area.
A generator is rated 100 MVA at 11 kV. On a system base of 100 MVA, its per-unit reactance equals its reactance in ohms multiplied by 100 divided by (base kV)^2. If the base voltage is chosen equal to the generator's own rated voltage of 11 kV, a reactance quoted as 0.15 per unit on the generator's own rating is, on this same 100 MVA base, equal to ___ per unit
Numerical answer — type the value.
Show answer
Answer: 0.15
When the chosen base MVA and base kV exactly match the machine's own rated MVA and kV, the per-unit value does not change at all when converting bases — it stays 0.15 per unit, since the conversion factor (own rating / new base) is exactly 1 in both voltage and MVA.