Stability, Power Quality, and Reliability

Introduction

Modern power systems must remain secure while supplying voltage and frequency within acceptable limits, even when generators, loads, and network topology change. In the laboratory, this requirement is examined through four related questions:

1. Will synchronous machines remain in synchronism after a disturbance?
2. Will bus voltages recover to acceptable values?
3. Is the supplied waveform sufficiently sinusoidal, balanced, and continuous?
4. How frequently do interruptions and equipment failures occur, and how effectively are they mitigated?

These questions define the interconnected subjects of stability, power quality, and reliability. Stability concerns the response of a power system to disturbances. Power quality concerns the characteristics of the delivered voltage and current. Reliability concerns the ability to provide acceptable service over time.

Laboratory investigation combines simulation, measurement, signal processing, protection, and engineering judgment. A sound experiment must distinguish a physical disturbance from an instrument artifact, use appropriate reference quantities, record operating conditions, and interpret results against applicable standards and utility requirements.

Learning Objectives

After completing this chapter, students should be able to:

• distinguish rotor-angle, voltage, and frequency stability;
• use the swing equation and equal-area concept for transient-stability analysis;
• evaluate frequency response and primary frequency control;
• identify harmonics, flicker, sags, swells, interruptions, and unbalance;
• calculate distortion, unbalance, and reliability indices;
• design and assess grounding and filtering arrangements;
• diagnose disturbances from voltage, current, frequency, and event records;
• propose technically sound and safe mitigation strategies.

1. Stability of Power Systems

1.1 Classification of stability

Rotor-angle stability is the ability of synchronous machines to remain in synchronism. It includes small-disturbance stability, associated with gradual changes and oscillations, and transient stability, associated with severe faults, generator trips, or sudden line switching.

Voltage stability is the ability of the system to maintain acceptable bus voltages after a disturbance or change in load. It is strongly influenced by reactive power balance, transmission impedance, transformer taps, and load characteristics. Voltage collapse is a progressive loss of voltage caused by insufficient reactive support and unfavorable system conditions.

Frequency stability is the ability to maintain frequency following a major imbalance between generation and demand. It depends on rotating inertia, governor response, turbine control, load damping, and under-frequency protection.

These categories interact. A fault may cause rotor-angle oscillations, a voltage depression, and a frequency deviation simultaneously; therefore, measurements should be time-synchronized.

1.2 The swing equation

For a simplified synchronous generator represented on a common power base, the rotor motion is described by

where [[M:M]] is the inertia constant in dynamic form (MW·s² per electrical radian on the selected base), [[M:δ]] is rotor electrical angle in radians, [[M:P_m]] is mechanical input power, and [[M:P_e]] is electrical air-gap output power. A commonly used per-unit form is

where [[M:H]] is the inertia constant in seconds, [[M:ω_s = 2πf_0]] is synchronous electrical angular speed in rad/s, and [[M:P_a]] is accelerating power in per unit. Mechanical damping may be included as an additional term proportional to speed deviation. The equation assumes a coherent machine model, constant frequency over the short transient interval, and consistent per-unit quantities.

For a cylindrical-rotor generator connected to an infinite bus through an equivalent reactance, the classical power-angle relation is

where [[M:E′]] is the internal transient emf, [[M:V]] is infinite-bus voltage, and [[M:X]] is the transfer reactance. The maximum transferable power is [[M:P_max = E′V/X]]. The model neglects resistance, saliency, excitation transients, and detailed controls; it is useful for introductory transient studies but not for all operating conditions.

1.3 Equal-area criterion

For a single machine against an infinite bus, a disturbance is transiently stable if the accelerating area equals or is less than the available decelerating area before the rotor reaches an unstable equilibrium. In energy form,

Here [[M:δ₀]] is the pre-disturbance angle, [[M:δ_c]] the clearing angle, and [[M:δ_max]] the maximum post-disturbance angle. The inequality applies to the classical one-machine infinite-bus model and a specified fault, clearing action, and post-fault network. In a multimachine system, numerical time-domain simulation is normally required.

1.4 Worked example: initial rotor angle

A generator transfers 0.8 per unit through a reactance with [[M:P_max = 1.6]] per unit. Before a disturbance, mechanical input equals electrical output. Therefore,

The result is an initial operating angle, not a stability margin. A transient assessment additionally requires the fault-network and post-fault power-angle curves, inertia, fault duration, and clearing angle.

1.5 Laboratory rotor-angle experiment

A laboratory study may use a synchronous-machine model, programmable fault switch, adjustable mechanical input, and voltage/current acquisition. The procedure is:

1. Establish a steady operating point and record [[M:P_m]], voltage, current, and rotor position.
2. Apply a controlled disturbance, such as a short line fault or sudden line opening.
3. Clear the disturbance at several times and identify the longest stable clearing time.
4. Compare measured rotor-angle trajectories with simulation.
5. Repeat with excitation control, fast fault clearing, or altered transfer reactance.

The main errors arise from inaccurate machine parameters, sensor delay, saturation, shaft torsional dynamics, and assuming an infinite bus when the laboratory source is not stiff.

2. Voltage and Frequency Response

2.1 Voltage stability and reactive power

A simple two-bus system illustrates voltage sensitivity. For a lossless line, increasing transmitted real power increases the angle difference, while reactive-power demand and line reactance influence voltage magnitude. Practical voltage-stability studies use power-flow continuation, reactive-power margins, sensitivity analysis, or time-domain simulation.

Useful operational indicators include:

• low bus voltage and declining voltage after corrective action;
• high reactive-power flow and generator field-current limits;
• transformer tap changers reaching limits;
• increasing load voltage sensitivity;
• negative or very small [[M:dV/dQ]] sensitivity at a monitored bus.

Capacitor banks, static var compensators, STATCOMs, synchronous condensers, transformer tap control, and load shedding can improve voltage security. Capacitors may, however, amplify harmonics or produce overvoltage at light load.

2.2 Frequency response

Immediately after a generation loss, kinetic energy in rotating machines supplies part of the deficit. The simplified frequency-balance equation is

where [[M:Δω]] is speed deviation, [[M:D]] is load-frequency damping in consistent per-unit torque or power units, and the other symbols retain their previous meanings. The equation is a short-term aggregated model; it excludes detailed governor dead bands, turbine delays, network effects, and nonlinear protection.

The frequency response characteristic is often represented by

where [[M:R]] is governor speed regulation in per-unit frequency change per per-unit power change, and [[M:β]] is the combined frequency-response characteristic. For a small generation deficit [[M:ΔP_L]], the quasi-steady frequency change is approximately

provided the same base and sign convention are used and secondary control has not restored scheduled frequency.

2.3 Frequency-response experiment

Measure frequency, active power, and generator speed before and after a controlled load step. Estimate the initial rate of change of frequency from a fitted line rather than a single noisy sample. Compare responses with and without governor control. Record dead band, ramp rate, settling time, nadir, and steady-state error. A frequency relay or under-frequency load-shedding scheme must be tested with safe, limited disturbances and an emergency stop.

3. Power Quality Phenomena

Power quality is commonly assessed at the point of common coupling. Definitions and limits should be taken from the applicable edition of standards such as IEEE 1159, IEEE 519, IEC 61000 series, and local grid codes; numerical limits are not universal and depend on voltage level, equipment class, and measurement method.

3.1 RMS voltage and events

For a sinusoidal voltage, [[M:V_rms = V_peak/√2]]. For sampled data with [[M:N]] samples,

The formula applies to a defined observation window. For nonstationary waveforms, specify the window length and whether the value is refreshed each cycle or over a longer interval.

A sag is a short-duration reduction in RMS voltage, while a swell is an increase. An interruption is a severe reduction, often near zero, for a defined duration. Causes include faults, motor starting, transformer energization, single-line conductor problems, and incorrect switching. Sag mitigation includes fault clearing, feeder reconfiguration, dynamic voltage restorers, energy-storage systems, and improved equipment ride-through.

3.2 Harmonics and distortion

A periodic nonsinusoidal waveform can be represented by a fundamental and harmonic components. Voltage total harmonic distortion is

where [[M:V_1]] is the RMS fundamental voltage and [[M:V_h]] is the RMS component at harmonic order [[M:h]]. A current version replaces voltage quantities with [[M:I_1]] and [[M:I_h]]. The fundamental is excluded from the numerator. Harmonics originate from rectifiers, variable-speed drives, switched-mode supplies, arc furnaces, and magnetic saturation.

Harmonic consequences include transformer heating, neutral-current amplification from triplen harmonics, capacitor resonance, motor torque pulsation, measurement error, and nuisance tripping. Mitigation includes passive tuned filters, broadband passive filters, active power filters, phase-shifting transformers, multipulse converters, and suitable converter controls. A passive filter must be checked for parallel resonance with the supply impedance; reducing one harmonic can unintentionally increase another.

3.3 Flicker

Voltage flicker is the visible or perceptible fluctuation of lamp light caused by repeated voltage changes. Arc furnaces, welders, compressors, and rapidly varying drives are common sources. Flicker assessment uses standardized measurement algorithms, commonly reporting short-term severity [[M:P_st]] and long-term severity [[M:P_lt]]. These indices are not interchangeable with simple percentage voltage variation because repetition rate, waveform, and human visual sensitivity are included.

3.4 Unbalance

For a three-phase system, symmetrical components decompose phase quantities into positive-, negative-, and zero-sequence sets. Voltage unbalance may be expressed as

where [[M:V_1]] and [[M:V_2]] are positive- and negative-sequence voltage magnitudes. This sequence definition is preferred for technical analysis. A phase-magnitude deviation method may produce a different result and must not be substituted without disclosure. Negative-sequence voltage produces double-frequency currents in machine rotors and can cause severe heating. Causes include unequal single-phase loads, blown fuses, asymmetrical impedances, and open conductors.

3.5 Grounding and common-mode disturbances

Grounding provides a reference and a low-impedance path for fault current so protective devices can operate. It is not identical to bonding: bonding connects exposed conductive parts, whereas grounding connects a system or equipment to earth or a reference conductor. Protective grounding must be designed according to local electrical codes and prospective fault current.

An idealized ground-fault loop gives

where [[M:I_f]] is fault current, [[M:V_phase]] is phase-to-ground voltage, and [[M:Z_loop]] includes source, transformer, conductor, and grounding impedances. This approximation assumes a bolted fault and linear impedances; arc faults, grounding electrodes, and protective-device characteristics require more detailed analysis.

Poor grounding can increase touch voltage, electromagnetic interference, common-mode noise, and shock risk. Grounding resistance alone does not prove safety: bonding continuity, fault-loop impedance, protective disconnection time, and touch voltage must also be evaluated.

4. Measurement, Filtering, and Disturbance Diagnosis

4.1 Instrumentation practice

A power-quality analyzer should have isolated voltage inputs, appropriately rated current probes, adequate bandwidth, synchronized sampling, and a sampling rate high enough for the highest frequency of interest. Anti-alias filtering is essential. If the sampling frequency is [[M:f_s]], frequencies above the Nyquist frequency [[M:f_s/2]] can alias into the measured band.

Before testing, verify:

• probe ratings and CAT category;
• phase sequence and polarity;
• calibration status;
• grounding and isolation;
• trigger threshold and pre-trigger memory;
• time synchronization;
• channel scaling and uncertainty.

Never connect an oscilloscope ground clip to a live conductor unless the instrument and connection method are specifically rated for that arrangement.

4.2 Digital filtering

A moving-average filter of length [[M:M]] is

It reduces random noise but also attenuates rapid events and can obscure sag inception. An RMS event detector should use a defined window and threshold with hysteresis to prevent repeated triggering near a boundary. Filtering must not be used to conceal an actual disturbance; retain the raw waveform and document all processing.

4.3 Diagnostic signatures

Diagnosis should correlate electrical records with switching logs, protection targets, weather, process data, and maintenance history. A single waveform rarely proves causation.

5. Reliability Assessment

Reliability has adequacy and security dimensions. Adequacy asks whether sufficient facilities exist to supply demand under expected conditions. Security asks whether the system withstands disturbances while remaining within operating limits.

For interruption statistics, common distribution indices include:

Here [[M:N_i]] is the number of customers interrupted in event [[M:i]], [[M:r_i]] is interruption duration in hours, and [[M:N_T]] is total customers served. SAIFI is interruptions per customer, SAIDI is customer-hours interrupted per customer (hours/customer), and CAIDI is average interruption duration per interrupted customer (hours/interruption). Momentary interruptions may be excluded or reported separately according to the adopted standard.

For a two-state component with failure rate [[M:λ]] and repair rate [[M:μ]], the steady-state availability is

and unavailability is [[M:U = 1 − A = λ/(λ+μ)]]. This assumes constant independent failure and repair rates, exponential state durations, and a two-state model. Real systems require common-cause failures, weather, maintenance, spare parts, and dependent outages.

Worked example: reliability indices

A feeder serves 1,000 customers. During a year, event 1 interrupts 200 customers for 2 hours, and event 2 interrupts 100 customers for 1 hour. Assuming no overlap and counting each event once,

The results describe this recorded population and period; they are not predictions of future performance without a suitable statistical model.

6. Integrated Case Study: Industrial Feeder Disturbances

Background

The following is an illustrative case. A 415 V industrial feeder supplies variable-speed drives, induction motors, welding equipment, and sensitive control computers. Operators report nuisance trips, visible lamp fluctuation, and overheating of the neutral conductor.

Problem and analysis

A power-quality analyzer is installed at the point of common coupling for one week. The record shows repeated voltage sags coincident with motor starts, elevated fifth and seventh current harmonics during drive operation, and high neutral current despite moderate phase-current values. A sequence calculation shows increased negative-sequence voltage during a period when one single-phase production line is heavily loaded.

The diagnosis is not simply “poor power quality.” Three mechanisms coexist:

• motor starting causes short voltage sags through feeder impedance;
• six-pulse drives inject characteristic low-order harmonics;
• unequal single-phase nonlinear loads produce neutral and zero-sequence current.

Solution

The engineering team first tightens motor-starting procedures and evaluates soft starters or variable-speed ramp control. Drives are grouped by phase and supplied through line reactors or an active front end where justified. A harmonic study checks transformer impedance and capacitor-resonance risk before any capacitor bank is installed. Single-phase loads are redistributed, neutral conductors are inspected for heating, and sensitive controllers receive ride-through support. Protective settings are reviewed so that temporary, documented events do not cause unnecessary trips, without disabling genuine fault protection.

Outcome

Post-installation measurements show fewer deep sags, lower harmonic current, reduced neutral temperature, and fewer control resets. The result is accepted only after measurements cover representative production cycles and protective coordination is verified.

Learning points

Power-quality mitigation should follow measurement and system modeling. Filtering without resonance analysis is unsafe; phase balancing without checking actual load diversity may be ineffective; and improving one metric can worsen another. Reliability improves when root causes, protection, maintenance, and equipment immunity are addressed together.

7. Mitigation, Governance, and Contemporary Practice

A mitigation plan should follow the hierarchy: define the event, measure at the correct location, identify the source, model alternatives, implement the least-risk effective solution, and verify performance. Typical actions include faster fault clearing, network reinforcement, reactive support, energy storage, harmonic filtering, load balancing, grounding and bonding correction, sectionalizing, automation, and preventive maintenance.

Engineering decisions must respect safety legislation, utility interconnection rules, electromagnetic-compatibility standards, and data-governance requirements. Disturbance records may contain production information and should be protected. Automated controls and AI-based classifiers can identify waveform patterns, forecast equipment failure, and prioritize maintenance, but training-data bias, false alarms, cyberattack, explainability, and human override must be considered. AI should support—not replace—protection principles, safety procedures, and accountable engineering review.

Contemporary systems increasingly use phasor measurement units, synchronized edge sensors, inverter-based resources, grid-forming controls, battery energy storage, and high-resolution event analytics. These technologies improve observability but also introduce converter control interactions, reduced mechanical inertia, communication dependence, and new cyber-physical failure modes.

Best-Practice Laboratory Workflow

1. State the hypothesis and measurable acceptance criteria.
2. Draw the single-line diagram and identify the point of common coupling.
3. Calculate expected magnitudes before energizing the apparatus.
4. Verify ratings, isolation, grounding, and emergency shutdown.
5. Capture steady-state data before applying a disturbance.
6. Apply one controlled change at a time.
7. Save raw synchronized voltage, current, frequency, and status data.
8. Process data with documented windows, filters, and uncertainty.
9. Compare measurements with simulation and explain discrepancies.
10. Restore the system safely and report limitations.

Chapter Summary

Power-system stability, power quality, and reliability are complementary dimensions of supply performance. Rotor-angle stability is governed by the balance between mechanical input and electrical output and can be studied with the swing equation, power-angle relation, and equal-area criterion. Voltage stability depends strongly on reactive-power availability and network impedance, while frequency stability depends on inertia, damping, governor response, and load shedding.

Power-quality analysis requires defined RMS windows and careful distinction among sags, swells, interruptions, harmonics, flicker, and unbalance. THD, symmetrical components, and event records provide useful quantitative tools. Grounding and bonding must be assessed as safety and interference-control systems, not merely by measuring electrode resistance. Reliability indices such as SAIFI, SAIDI, and CAIDI summarize interruption performance, while availability models provide component-level insight.

Successful laboratory work combines safe instrumentation, synchronized measurement, transparent signal processing, validated models, and standards-aware interpretation. Mitigation must address the physical source and verify that corrective action does not create resonance, protection, safety, or cyber-physical problems.

Key Terms

Stability: Ability of a power system to maintain an acceptable operating condition after a disturbance.

Rotor-angle stability: Ability of synchronous machines to remain in synchronism.

Transient stability: Stability following a severe, sudden disturbance.

Voltage stability: Ability to maintain acceptable bus voltages.

Frequency stability: Ability to maintain frequency after a generation-load imbalance.

Swing equation: Dynamic equation relating rotor acceleration to accelerating power.

Power-angle curve: Relationship between transferred power and rotor angle.

Equal-area criterion: Energy-based test for transient stability in a simplified system.

Sag: Short-duration RMS-voltage reduction.

Swell: Short-duration RMS-voltage increase.

Interruption: Severe loss of supply voltage for a defined duration.

Harmonic: Sinusoidal component whose frequency is an integer multiple of the fundamental.

THD: RMS magnitude of harmonic components relative to the fundamental, usually expressed as a percentage.

Flicker: Perceptible light fluctuation caused by voltage variation.

Unbalance: Inequality among three-phase magnitudes or phase angles.

Negative sequence: Symmetrical-component set rotating opposite to the normal phase sequence.

Grounding: Connection of a system or equipment to earth or a reference point.

SAIFI: Average number of sustained interruptions per customer.

SAIDI: Average interruption duration per customer.

CAIDI: Average duration per interrupted customer.

Review Questions

Short Answer Questions

1. Distinguish rotor-angle, voltage, and frequency stability.
2. What physical meaning does the accelerating power [[M:P_m − P_e]] have?
3. Why can capacitor banks worsen harmonic problems?
4. Define voltage sag and voltage swell.
5. What information does SAIDI provide?

Long Answer Questions

1. Explain the swing equation and discuss the assumptions of the classical generator model.
2. Describe the equal-area criterion and its limitations in multimachine systems.
3. Explain the causes, effects, and mitigation of harmonics in an industrial feeder.
4. Discuss the role of grounding, bonding, and fault-loop impedance in safety.
5. Design a laboratory procedure for measuring frequency response after a controlled generation loss.

Analytical / Application Questions

1. A generator has [[M:P_max = 2.0]] per unit and operates at [[M:P_m = 1.0]] per unit. Calculate its initial power angle and state why this result alone does not establish transient stability.
2. A feeder serves 2,000 customers. Two events interrupt 400 customers for 1.5 hours and 200 customers for 3 hours. Calculate SAIFI, SAIDI, and CAIDI, stating any assumptions.
3. An analyzer reports a high fifth-harmonic current, neutral overheating, and repeated motor-starting sags. Propose a measurement-based diagnosis and a prioritized mitigation plan.

Multiple Choice Questions with Answers

Question 1

For a classical generator connected to an infinite bus, which relation describes electrical power transfer?

A. [[M:P_e = E′V/X sin δ]]

B. [[M:P_e = E′V/X cos δ]]

C. [[M:P_e = X/(E′V) sin δ]]

D. [[M:P_e = P_m + DΔω]]

Correct Answer: A. [[M:P_e = E′V/X sin δ]]

Question 2

A generator has [[M:P_m = 0.8]] per unit and [[M:P_max = 1.6]] per unit. Its initial rotor angle is approximately:

A. 15°

B. 30°

C. 45°

D. 60°

Correct Answer: B. 30°

Question 3

In the swing equation, positive accelerating power means that:

A. Electrical output exceeds mechanical input

B. Mechanical input equals electrical output

C. Mechanical input exceeds electrical output

D. The generator has zero inertia

Correct Answer: C. Mechanical input exceeds electrical output

Question 4

Which quantity is excluded from the numerator of voltage THD?

A. Second-harmonic RMS voltage

B. Fifth-harmonic RMS voltage

C. Fundamental RMS voltage

D. Seventh-harmonic RMS voltage

Correct Answer: C. Fundamental RMS voltage

Question 5

Negative-sequence voltage is particularly harmful to rotating machines because it produces:

A. A counter-rotating field and rotor heating

B. Only a higher positive-sequence voltage

C. Perfect phase balance

D. Zero-frequency current in every winding

Correct Answer: A. A counter-rotating field and rotor heating

Question 6

A feeder serves 1,000 customers and records interruptions affecting 200 and 100 customers. The SAIFI is:

A. 0.10 interruptions/customer

B. 0.20 interruptions/customer

C. 0.30 interruptions/customer

D. 3.00 interruptions/customer

Correct Answer: C. 0.30 interruptions/customer

Question 7

Which device is primarily intended to reduce rapid voltage fluctuations that cause flicker?

A. Static var compensator

B. Fuse link only

C. Energy meter

D. Insulation tester

Correct Answer: A. Static var compensator

Question 8

If a system has failure rate [[M:λ]] and repair rate [[M:μ]] under a two-state constant-rate model, availability is:

A. [[M:λ/(λ+μ)]]

B. [[M:μ/(λ+μ)]]

C. [[M:λμ]]

D. [[M:λ+μ]]

Correct Answer: B. [[M:μ/(λ+μ)]]

Question 9

The Nyquist frequency for a sampling frequency [[M:f_s]] is:

A. [[M:f_s/4]]

B. [[M:f_s/2]]

C. [[M:2f_s]]

D. [[M:f_s²]]

Correct Answer: B. [[M:f_s/2]]

Question 10

Which practice is most appropriate before installing a passive harmonic filter?

A. Disconnect all protective devices

B. Perform a resonance and impedance study

C. Increase neutral impedance deliberately

D. Ignore the supply short-circuit level

Correct Answer: B. Perform a resonance and impedance study