The equal-area criterion compares the energy a generator rotor gains while a fault lasts with the energy it can return after the fault is cleared. If the acceleration area does not exceed the available braking area, the generator keeps synchronism. The angle at which the two areas become equal gives the critical clearing time: the maximum time the fault can last. It is worked through with numbers for a machine against an infinite bus.
Short answer
When there is a fault near a generator, the electrical power it delivers drops while the turbine keeps pushing: the rotor accelerates and its angle relative to the grid grows. When the fault is cleared, the network can receive power again and the rotor slows down. The equal-area criterion compares the energy the rotor gained during the fault (acceleration area) with the energy it can return afterwards (braking area). If the acceleration area does not exceed the available braking area, the generator stays in synchronism. The angle at which the two areas are equal is the critical angle, and the time the rotor takes to reach it is the critical clearing time. If the fault lasts longer, the generator loses synchronism.
The model: one machine against an infinite bus
The method is explained with the simplest case: a generator connected by a line to a network so large that its voltage and frequency do not change (infinite bus).
- The electrical power the machine delivers is Pe = Pmax · sin δ, where δ is the angle between the machine’s internal voltage and the bus voltage, and Pmax depends on the voltages and the total reactance.
- The turbine mechanical power Pm is assumed constant during the event.
- The difference Pm − Pe accelerates or slows the rotor according to the swing equation: d²δ/dt² = (π·f / H) · (Pm − Pe), with δ in radians, f the grid frequency and H the machine inertia constant in seconds. Everything in per unit.
There are three network conditions, each with its own Pe(δ) curve:
- Before the fault: curve with Pmax1.
- During the fault: a lower curve, with Pmax2. In a three-phase fault next to the generator bus, the bus drops to zero voltage and Pmax2 is zero.
- After clearing: curve with Pmax3, normally lower than before if clearing removes a line.
The two areas and the critical angle
The initial angle δ0 comes from Pm = Pmax1 · sin δ0. During the fault, Pm is greater than Pe and the rotor accelerates; the area between Pm and the fault curve, from δ0 to the clearing angle δc, is the acceleration area. After clearing, Pe moves to the post-fault curve, which exceeds Pm: the rotor slows down. The area between the post-fault curve and Pm, from δc to the maximum angle δmax, is the braking area. The maximum angle satisfies Pm = Pmax3 · sin(180° − δmax).
The generator is stable if the rotor manages to slow down before δmax. The limit case is when the two areas are equal, and that gives the critical angle. For a fault in which Pe falls to zero during the event, the area balance has a closed-form solution:
cos δcr = [ Pm · (δmax − δ0) + Pmax3 · cos δmax ] / Pmax3
With Pe = 0 during the fault, the rotor accelerates at a constant rate, and the critical time is:
tcr = √[ 2H · (δcr − δ0) / (π · f · Pm) ]
If Pe does not fall to zero, the balance follows the same principle, but the critical time is obtained by integrating the swing equation step by step.
Worked example
Example data. Generator connected to an infinite bus; frequency 60 Hz; H = 5 s; Pm = 0.8 pu. Pmax before the fault: 1.8 pu. Three-phase fault on one of the two lines, next to the generator bus (Pe = 0 during the fault). Clearing that opens that line and leaves a single line in service, with Pmax3 = 1.2 pu. Total clearing time for the example: 0.10 s (relay plus breaker).
- Initial angle: sin δ0 = 0.8 / 1.8 = 0.4444, so δ0 = 26.39° = 0.4606 rad.
- Maximum post-fault angle: sin(180° − δmax) = 0.8 / 1.2 = 0.6667, so 180° − δmax = 41.81° and δmax = 138.19° = 2.4119 rad. Its cosine is −0.7454.
- Critical angle: cos δcr = [0.8 × (2.4119 − 0.4606) + 1.2 × (−0.7454)] / 1.2 = (1.5610 − 0.8944) / 1.2 = 0.5555, so δcr = 56.25° = 0.9818 rad.
- Critical time: tcr = √[2 × 5 × (0.9818 − 0.4606) / (π × 60 × 0.8)] = √(5.2126 / 150.80) = √0.03457 = 0.186 s, about 11 cycles at 60 Hz.
- Check with the 0.10 s clearing: δ = 0.4606 + (π × 60 × 0.8 / (2 × 5)) × 0.10² = 0.4606 + 15.08 × 0.01 = 0.6114 rad = 35.0°, below 56.25°. The generator is stable, with a margin of 0.186 − 0.10 = 0.086 s.
- If the breaker fails and the backup clears at 0.35 s: δ = 0.4606 + 15.08 × 0.1225 = 2.308 rad = 132.2°, far above 56.25°. The generator loses synchronism.
- How does the clearing time compare with the critical time?
- Clearing below critical, with margin: the generator keeps synchronism; the margin is the reserve against uncertain data
- Clearing almost equal to critical: no reserve; speed up the protection or reinforce the network
- Clearing above critical: loss of synchronism; the generator must be protected with a loss-of-synchronism protection (78) or a tripping scheme, or the fault must be cleared faster
Reading: the main protection clears with margin, but the breaker-failure backup (0.35 s) does not. If the plant must remain stable for a fault combined with a breaker failure, the 50BF time has to be shortened to below 0.186 s, if the equipment allows it, or generator tripping has to be accepted.
Scope and limits of the method
- It is valid for one machine against an infinite bus (or two equivalent machines). It does not replace simulation of the whole system.
- It assumes constant internal voltage, no voltage regulator, no speed governor and no damping. Real regulators change the result.
- A project’s stability study uses dynamic models of each machine and its controls, and compares the clearing time with the critical time for the faults the grid operator sets.
To see what decides stability in practice, see transient stability: what it is and what decides it.
Checklist
- The fault studied (type and location) is defined and justified.
- The three P-δ curves (before, during and after) have their source data.
- The inertia constant H comes from the machine manufacturer.
- The clearing time includes relay, breaker and, if applicable, the communication logic.
- The critical time was calculated for each fault and each network configuration.
- The breaker-failure case was reviewed with its backup time.
- The result was confirmed with time-domain simulation if the system has more than one machine.
Related guides
- Transient stability: what it is and what decides it
- Swing curves and damping in stability
- Stability of an industrial network with its own generation
- Stability in industrial plants: causes and solutions
- Studies the CNO requires to connect generation
- Short-circuit study: contributions, IEC 60909 and ANSI/IEEE
References
- Kundur, P., Power System Stability and Control, McGraw-Hill, 1994: chapter on transient stability (equal-area criterion).
- Anderson, P. M. and Fouad, A. A., Power System Control and Stability, 2nd ed., IEEE Press/Wiley, 2003.
Do you need this study?
We carry out transient stability studies with dynamic models of generators and controls, and calculate the critical clearing times your protections must meet. It is engineering design and studies work done remotely, valid for projects in any country. If you need it to connect to a grid, see the connection studies. Contact us.
Frequently asked questions
What does the equal-area criterion say?
That the generator is stable if the acceleration area (the energy its rotor gains while the fault reduces the electrical power it delivers) is equal to or smaller than the available braking area (the energy it can return afterwards, when the network is left with the maximum power of the post-fault condition). If the acceleration area is larger, the rotor runs too far ahead and loses synchronism.
What is the critical clearing time?
It is the maximum time the fault can last before the generator loses synchronism. It is obtained from the critical angle, at which the two areas become equal. The actual clearing time (relay plus breaker) must stay below it with a margin.
Does the equal-area criterion work for a system with many machines?
Directly, only for one machine against an infinite bus, or two equivalent machines. In a real system it is used as a way to understand the phenomenon and to check results, and the stability study is done with time-domain simulation of all the machines and their controls.