Voltage Stability Part 5: Voltage Collapse, QV Curves, OLTC Feedback and Stability Indices
Voltage Collapse: QV Curves, OLTC Feedback, Load Restoration and Stability Indices
- Part 1 of 5 — What is Voltage Stability? Definition, Classification and Physical Intuition
- Part 2 of 5 — The Two-Bus Model: Setting Up the Voltage Stability Problem
- Part 3 of 5 — Deriving the PV Curve: The Nose Curve from First Principles
- Part 4 of 5 — The Nose Point: Maximum Loadability, Critical Voltage and the Jacobian
- Part 5 of 5 — Voltage Collapse: QV Curves, OLTC Feedback and Stability Indices
1. QV curves and reactive power margin
While the PV curve gives the loadability limit, the QV curve is the primary tool for assessing reactive power margin at individual buses. It plots the reactive injection Q needed at a bus to hold voltage at V, with active power held constant.
The reactive power margin is the vertical distance from the bottom of the QV curve to the Q = 0 axis. A negative value means the bus is already in reactive deficit — collapse is imminent unless reactive support is injected. In practice, the QV curve is obtained by attaching a fictitious reactive source to the bus under study and running a parametric sweep of power flows.
2. Voltage collapse mechanism
Knowing the static nose point is necessary but not sufficient. Real collapses develop dynamically through a sequence of device actions that each erode the system’s reactive margin.
3. The OLTC feedback loop
The most common collapse scenario begins with on-load tap changer (OLTC) action. OLTCs are designed to restore distribution-level voltages after disturbances. But near the nose point, their corrective action can accelerate collapse.
1. A disturbance (line outage, generator trip) reduces the system’s reactive reserve.
2. Load-bus voltages fall. OLTCs detect low secondary voltage and tap up — increasing the turns ratio.
3. The effective load seen from the HV bus increases (more current drawn). Reactive demand rises.
4. HV-bus voltages fall further. OLTCs tap up again.
5. Each OLTC step depletes more reactive reserve, pushing the operating point toward the nose.
6. Generators hit field current limits — OXL activates, capping reactive output.
7. Voltages collapse across a wide area within 2–10 minutes of the triggering event.
This mechanism caused the 1987 Tokyo blackout (27 million people, air-conditioner load surge on a summer day) and contributed to the 2003 Northeast US/Canada event.
4. Generator reactive limits and OXL activation
A synchronous generator’s reactive output is bounded by its capability curve. The overexcitation limiter (OXL) enforces a maximum field current limit, typically activated 60–180 seconds after overexcitation begins. Once the OXL activates, the generator switches from voltage control (PV bus) to reactive power limit control (PQ bus) and can no longer provide additional reactive support.
In planning studies, the proximity of generator operating points to their OXL limits is critical. A network that appears stable when generators are in voltage-control mode may collapse rapidly once several machines hit their limits simultaneously.
5. Voltage stability indices
For large multi-bus systems, scalar indices efficiently rank buses by voltage stability risk without plotting a full QV curve at every bus.
5.1 L-index (Kessel and Glavitsch, 1986)
where αG is the set of generator buses and F̄ⱼᵢ is a transfer coefficient derived from the Y-bus. Lⱼ lies between 0 (no load) and 1 (voltage collapse). Values above 0.5–0.6 signal high risk. The global index is L = max(Lⱼ).
5.2 Voltage collapse proximity index (VCPI)
where P_j_max is the maximum deliverable power to bus j. VCPI → 0 signals imminent collapse at that bus.
5.3 Reactive power margin
The most physically intuitive index. Negative Q_margin means reactive deficiency. Many TSOs use this as a real-time control room signal.
| Index | Range | Collapse condition | Computational cost |
|---|---|---|---|
| L-index | 0 → 1 | L → 1 | Low (Y-bus matrix algebra) |
| VCPI | 0 → 1 | VCPI → 0 | Medium (needs P_max per bus) |
| Q_margin | +∞ → −∞ | Q_margin = 0 | Low (from power flow output) |
| Δ (discriminant) | +∞ → 0 | Δ = 0 | Low (analytical, two-bus) |
6. Series summary and what comes next
This five-part series has built the complete theoretical foundation for voltage stability analysis:
| Part | Topic | Key result |
|---|---|---|
| 1 | Definition and classification | IEEE/CIGRÉ definition; large-disturbance long-term stability is the planning focus |
| 2 | Two-bus model | Power flow equations (1)–(4); quadratic structure |
| 3 | PV curve derivation | V₂² = ½{[V₁² − 2(PR+QX)] ± √Δ}; two branches; effect of pf |
| 4 | Nose point | P_max = V₁²/(2X); V₂_cr ≈ 0.707 pu; det(J) = 0 at collapse |
| 5 | Collapse and indices | QV curve; OLTC feedback; OXL; L-index; VCPI; Q-margin |
The next series moves from theory to tool. Part 6 builds the QV curve and runs a continuation power flow in DIgSILENT PowerFactory — configuring the voltage stability calculation module, setting up the study cases, and interpreting the nose curve and Q-margin results for a realistic test network.
References
- P. Kundur, Power System Stability and Control, McGraw-Hill, 1994 — Chapters 14–15.
- T. Van Cutsem and C. Vournas, Voltage Stability of Electric Power Systems, Springer, 1998.
- IEEE Std 1110-2002, Guide for Synchronous Generator Modelling Practices.
- CIGRÉ Technical Brochure 325, Working Group C4.601, 2007.
- P. Kessel and H. Glavitsch, “Estimating the voltage stability of a power system,” IEEE Trans. Power Delivery, vol. 1, no. 3, 1986 — original L-index paper.
- B. Gao, G. K. Morison, P. Kundur, “Voltage stability evaluation using modal analysis,” IEEE Trans. Power Systems, vol. 7, no. 4, 1992.
- NERC, Voltage Stability Criteria and Reactive Power Reserve Practices, Sep-2010.