8.6 Power System Stability: Swing Equation, Equal-Area Criterion & Critical Clearing Time

Key Takeaways

  • Steady-state stability is a static margin limited by the power-angle maximum at δ = 90°, while transient stability is a first-swing energy balance that a machine can fail even while sitting well inside its steady-state limit.
  • The swing equation (2H/ω_s)(d²δ/dt²) = P_m − P_e holds mechanical input constant through the first swing because governor and turbine response take seconds while the first swing completes in under one second.
  • The equal-area criterion declares a system stable when the decelerating area A₂ available after clearing is at least equal to the accelerating area A₁ absorbed during the fault, and it is valid only for two-machine or single-machine-infinite-bus systems.
  • Critical clearing time is t_cr = sqrt[4H(δ_cr − δ₀) / (ω_s · P_m)] with every angle expressed in radians; a typical heavily loaded machine yields roughly 0.1 to 0.25 seconds, which is why EHV lines use pilot protection rather than coordinated backup tripping.
  • Turbo-generator inertia constants run about 2 to 6 seconds, and replacing synchronous machines with grid-following inverter-based resources lowers aggregate inertia, raising RoCoF and shortening permissible clearing times.
Last updated: August 2026

8.6 Power System Stability: The Swing Equation, Equal-Area Criterion & Critical Clearing Time

Power system stability is a named sub-topic of Transmission and Distribution Analysis in the NCEES specification, and it is the analysis that decides how fast protection must clear a fault. Section 10.2 sets a coordination time interval from equipment constraints; stability sets the absolute ceiling from system constraints. When the two conflict, stability wins — which is why EHV transmission uses pilot protection to clear in three cycles rather than accepting a coordinated 0.3-second backup trip.

Stability is the ability of a system to return to synchronism after a disturbance. All synchronous machines on an interconnection must rotate at exactly the same electrical speed; a machine that permanently loses this lock has "gone out of step" and must be tripped.


1. The Three Classes of Stability

ClassDisturbanceTime frameGoverning analysis
Steady-state (small-signal)Slow, gradual load changeSeconds to minutesStatic power-angle limit; $\delta < 90^\circ$
TransientLarge disturbance — fault, line trip, sudden load rejectionFirst swing, ~1 secondSwing equation; equal-area criterion
DynamicSmall oscillations after the first swingSeconds to tens of secondsDamping, governor and exciter response, power system stabilizers

The exam concentrates on the first two. The distinction the test wants is that steady-state stability is a static margin question while transient stability is a first-swing energy question — a machine can be comfortably inside its steady-state limit and still go unstable transiently.


2. Steady-State Stability Limit

For a machine tied to an infinite bus through total reactance $X$:

Pe=EVXsinδP_e = \frac{|E||V|}{X}\sin\delta

The maximum transferable power occurs at $\delta = 90^\circ$, giving $P_{max} = \dfrac{|E||V|}{X}$. Beyond $90^\circ$ the slope $dP_e/d\delta$ turns negative: an increase in angle now reduces electrical output while mechanical input is unchanged, so the rotor accelerates without limit and synchronism is lost. The quantity $dP_e/d\delta = P_{max}\cos\delta$ is the synchronizing power coefficient, and it must remain positive for stable operation.

Utilities operate transmission corridors well below this theoretical limit — typically at $\delta$ of 30° to 45° — to preserve margin for contingencies. Reducing $X$ with series capacitors (Section 6.2) or a parallel circuit raises $P_{max}$ directly, which is the standard engineering fix for an angle-limited corridor.


3. The Swing Equation

Rotor dynamics follow directly from Newton's second law applied to a rotating mass:

2Hωsd2δdt2=PmPe=Pa\frac{2H}{\omega_s}\frac{d^2\delta}{dt^2} = P_m - P_e = P_a

where $H$ is the inertia constant in seconds (stored kinetic energy at rated speed divided by machine MVA rating, typically 2–6 s for turbo-generators and 2–4 s for hydro units), $\omega_s$ is synchronous speed in electrical rad/s, $P_m$ is mechanical input, $P_e$ is electrical output, and $P_a$ is the accelerating power.

Three exam-relevant consequences:

  1. Mechanical input is constant during the first swing. Governor and turbine response takes seconds; the first swing is over in well under a second. Every equal-area problem therefore holds $P_m$ fixed.
  2. Higher $H$ means slower angle excursion. Large thermal units ride through disturbances that would destabilize a small machine. This is precisely why displacing synchronous generation with inverter-based resources (which contribute no physical inertia unless explicitly programmed for synthetic inertia) raises system RoCoF and shortens permissible clearing times — the modern grid problem that makes synchronous condensers valuable.
  3. $H$ must be converted to the study base. $H_{new} = H_{old}\times\left(S_{old}/S_{new}\right)$, the inverse of the impedance base conversion.

For a three-phase bolted fault at the machine terminals, $P_e$ collapses to zero and the swing equation integrates directly:

δ(t)=δ0+ωsPm4Ht2\delta(t) = \delta_0 + \frac{\omega_s P_m}{4H}t^2


4. The Equal-Area Criterion

The equal-area criterion converts the differential equation into an energy balance that can be solved graphically or with a single integral. During the fault the rotor absorbs kinetic energy proportional to the accelerating area $A_1$ between $P_m$ and the faulted power-angle curve. After the fault clears, the rotor gives that energy back over the decelerating area $A_2$ between the post-fault curve and $P_m$.

The system is transiently stable if the available decelerating area is at least equal to the accelerating area: $A_2 \geq A_1$. If the maximum available $A_2$ is smaller than $A_1$, the rotor passes the point of no return at $\delta_{max} = 180^\circ - \delta_{post}$ and pulls out of step.

The criterion is valid only for a two-machine or single-machine-infinite-bus system, because it assumes a single relative angle. Multi-machine studies require step-by-step numerical integration. Exam items sometimes offer "apply the equal-area criterion to a 12-machine system" as a distractor — that is the wrong tool.

Critical Clearing Angle and Time

For the worst case — a bolted three-phase fault that drops $P_e$ to zero, with the pre-fault and post-fault curves identical (fault cleared without losing a line) — the critical clearing angle has a closed form:

cosδcr=(π2δ0)PmPmaxcosδ0\cos\delta_{cr} = (\pi - 2\delta_0)\frac{P_m}{P_{max}} - \cos\delta_0

and because $P_m/P_{max} = \sin\delta_0$ at the initial operating point, this reduces to the form worth memorizing:

δcr=cos1[(π2δ0)sinδ0cosδ0]\delta_{cr} = \cos^{-1}\left[(\pi - 2\delta_0)\sin\delta_0 - \cos\delta_0\right]

with $\delta_0$ expressed in radians. Substituting into the integrated swing equation gives the critical clearing time:

tcr=4H(δcrδ0)ωsPmt_{cr} = \sqrt{\frac{4H(\delta_{cr} - \delta_0)}{\omega_s P_m}}

with angles in radians, $\omega_s = 2\pi f$ in electrical rad/s, and $P_m$ in per unit.

Worked Example: Critical Clearing Time

A 60 Hz generator with $H = 4.0$ s delivers $P_m = 0.85$ pu into an infinite bus. A bolted three-phase fault at the machine terminals drops $P_e$ to zero, and the pre-fault and post-fault power-angle curves are identical with $P_{max} = 1.70$ pu.

Step 1 — Initial angle. $\sin\delta_0 = P_m/P_{max} = 0.85/1.70 = 0.500$, so $\delta_0 = 30^\circ = 0.5236$ rad.

Step 2 — Critical clearing angle. δcr=cos1[(π2(0.5236))(0.500)0.8660]\delta_{cr} = \cos^{-1}\left[(\pi - 2(0.5236))(0.500) - 0.8660\right] =cos1[(3.14161.0472)(0.500)0.8660]=cos1[1.04720.8660]= \cos^{-1}\left[(3.1416 - 1.0472)(0.500) - 0.8660\right] = \cos^{-1}\left[1.0472 - 0.8660\right] =cos1(0.1812)=79.56=1.3886 rad= \cos^{-1}(0.1812) = 79.56^\circ = 1.3886 \text{ rad}

Step 3 — Critical clearing time. tcr=4(4.0)(1.38860.5236)2π(60)(0.85)=13.840320.44=0.04319=0.208 st_{cr} = \sqrt{\frac{4(4.0)(1.3886 - 0.5236)}{2\pi(60)(0.85)}} = \sqrt{\frac{13.840}{320.44}} = \sqrt{0.04319} = 0.208 \text{ s}

Step 4 — Interpret. 0.208 s is about 12.5 cycles at 60 Hz. A relay-plus-breaker scheme clearing in 5 cycles (0.083 s) has comfortable margin; a 0.30-second coordinated backup trip does not, and this is exactly why the primary protection on such a machine is instantaneous.


5. Engineering Measures That Improve Stability

MeasureMechanism
Faster fault clearingReduces $A_1$ directly — the single most effective measure
High-speed reclosingRestores the post-fault curve to pre-fault height, enlarging $A_2$
Series capacitorsLower $X$, raising $P_{max}$ on both faulted and post-fault curves
Additional parallel circuitsLower $X$ and keep a path in service after a line trip
High-response excitationRaises $
Power system stabilizers (PSS)Damp the dynamic oscillations that follow the first swing
Single-pole trippingKeeps two healthy phases in service through a single line-to-ground fault (the great majority of faults), so $P_e$ never falls to zero

6. Common Exam Traps & Tactical Pitfalls

+-----------------------------------------------------------------------------+
|                    POWER SYSTEM STABILITY EXAM PITFALLS                     |
|                                                                             |
|   [!] Mixing degrees and radians in t_cr. The swing equation integral       |
|       requires RADIANS; only sin/cos arguments may be in degrees.           |
|   [!] Assuming P_m changes during the swing. Governor action is far too     |
|       slow for the first swing - P_m is CONSTANT.                           |
|   [!] Applying the equal-area criterion to a multi-machine system. It is    |
|       valid only for two-machine / single-machine-infinite-bus cases.       |
|   [!] Forgetting that clearing a fault by TRIPPING A LINE raises post-      |
|       fault X, lowering P_max and shrinking the available decelerating      |
|       area. The pre-fault and post-fault curves are then NOT identical.     |
|   [!] Confusing the steady-state limit (delta = 90 degrees) with the        |
|       transient limit. A machine can sit at 40 degrees steady-state and     |
|       still lose synchronism on the first swing.                            |
|   [!] Failing to convert H to the study MVA base: H_new = H_old*(S_old/     |
|       S_new), which is the INVERSE of the impedance base conversion.        |
+-----------------------------------------------------------------------------+
Loading diagram...
Transient Stability Assessment: Fault Inception Through Equal-Area Verdict
Test Your Knowledge

A 60 Hz turbo-generator with an inertia constant H = 3.0 s on its own base delivers a mechanical input of P_m = 0.90 pu into an infinite bus. A bolted three-phase fault at the machine terminals collapses electrical output to zero, and the pre-fault and post-fault power-angle curves are identical with P_max = 1.80 pu. Using the equal-area criterion, what is the critical clearing time?

A
B
C
D
Test Your Knowledge

A protection engineer must justify why a 500 kV transmission line uses a pilot (communication-assisted) scheme clearing in 3 cycles rather than accepting a coordinated stepped-distance backup that would clear the same fault in 0.35 seconds. Which statement gives the correct engineering basis?

A
B
C
D
Test Your Knowledge

A utility is retiring a 400 MVA coal unit with an inertia constant of H = 4.5 s and replacing its output with grid-following inverter-based photovoltaic generation of equal MW capacity. What is the principal transient-stability consequence, and which mitigation directly restores the lost physical property?

A
B
C
D