These notes synthesize the standard graduate-level treatment of power systems analysis as covered in ECE 360 at the University of Waterloo. The primary reference is Glover, Sarma & Overbye, Power System Analysis and Design (5th/6th ed., Cengage). Supplementary material draws from Bergen & Vittal, Power Systems Analysis (2nd ed., Prentice Hall) and Kundur, Power System Stability and Control (McGraw-Hill). Open courseware from MIT OCW 6.061 is also referenced throughout.

---

Sources and References

• J. D. Glover, M. S. Sarma, and T. J. Overbye, Power System Analysis and Design, 5th ed. (SI version), Cengage Learning, 2012.
• A. R. Bergen and V. Vittal, Power Systems Analysis, 2nd ed., Prentice Hall, 2000.
• P. Kundur, Power System Stability and Control, McGraw-Hill, 1994.
• MIT OpenCourseWare 6.061, Introduction to Electric Power Systems, Massachusetts Institute of Technology (open access).
• IEEE Standard C37.118.1-2011, IEEE Standard for Synchrophasor Measurements for Power Systems.
• IEC 61850, Communication Networks and Systems for Power Utility Automation.

---

Chapter 1: Power System Structure and Fundamentals

1.1 The Interconnected Power System

Modern electric power systems are among the most complex engineering artifacts ever constructed. At their core, they must solve a deceptively simple problem: at every instant, the electrical energy produced by generators must precisely equal the energy consumed by loads plus the losses in the network. Because electricity cannot be economically stored at grid scale (at least not yet), this balance must be maintained in real time across networks spanning thousands of kilometres.

The power system is conventionally divided into three functional domains: generation, transmission, and distribution. Generators — whether thermal, hydro, nuclear, or renewable — convert primary energy into three-phase alternating current at voltages typically in the range of 11–30 kV. Step-up transformers raise these voltages to transmission levels (115 kV, 230 kV, 345 kV, 500 kV, or 765 kV in North American practice) for bulk transport over long distances. At the receiving end, step-down transformers reduce voltage for subtransmission (26–69 kV), then for distribution primaries (4–35 kV), and finally to service entrance levels (120/240 V single-phase or 208/120 V three-phase in residential and light commercial contexts).

The reason for high-voltage transmission is straightforward: for a given power \(P\) transmitted at voltage \(V\) and current \(I\), the resistive loss in the line scales as \(I^2 R\). Since \(P = VI\), doubling the voltage halves the current and reduces losses by a factor of four. The economics of conductor material and insulation cost dictate an optimal voltage for each power level and distance combination.

1.2 Three-Phase Power — Review and Motivation

Power is transmitted as three-phase AC because it offers higher power density per conductor than single-phase and because three-phase induction and synchronous machines have simpler, more efficient construction. The three-phase instantaneous voltages form a balanced set:

\[ v_a(t) = V_m \cos(\omega t + \phi), \quad v_b(t) = V_m \cos\!\left(\omega t + \phi - \tfrac{2\pi}{3}\right), \quad v_c(t) = V_m \cos\!\left(\omega t + \phi - \tfrac{4\pi}{3}\right) \]

where \(\omega = 2\pi f\) with \(f = 60\) Hz in North America. This is called positive (abc) sequence. The sum \(v_a + v_b + v_c = 0\) at every instant, which is why the neutral conductor carries no current in a balanced system.

Using phasor notation (RMS amplitudes):

\[ \mathbf{V}_a = V\angle\phi, \quad \mathbf{V}_b = V\angle(\phi - 120°), \quad \mathbf{V}_c = V\angle(\phi - 240°) \]

with \(V = V_m/\sqrt{2}\).

1.2.1 Wye and Delta Connections

In a wye (Y) connection, each impedance is connected between one phase terminal and the neutral point. The line voltage \(V_L\) relates to the phase voltage \(V_\phi\) by:

\[ V_L = \sqrt{3}\, V_\phi \]

This follows directly from phasor subtraction: \(\mathbf{V}_{AB} = \mathbf{V}_A - \mathbf{V}_B = \sqrt{3}\,V_\phi\angle(\phi + 30°)\). The line current equals the phase current: \(I_L = I_\phi\).

In a delta (\(\Delta\)) connection, each impedance is connected between two line terminals. The line current relates to the phase (branch) current by:

\[ I_L = \sqrt{3}\, I_\phi \]

and the phase voltage equals the line voltage: \(V_\phi = V_L\).

1.2.2 Balanced Three-Phase Power

For a balanced three-phase load with power factor angle \(\theta\) (the angle between voltage and current in each phase):

\[ P_{3\phi} = 3\, V_\phi I_\phi \cos\theta = \sqrt{3}\, V_L I_L \cos\theta \]\[ Q_{3\phi} = 3\, V_\phi I_\phi \sin\theta = \sqrt{3}\, V_L I_L \sin\theta \]\[ S_{3\phi} = P_{3\phi} + jQ_{3\phi} = \sqrt{3}\, V_L I_L \angle\theta \]

where \(P\) is real power (watts), \(Q\) is reactive power (VAr), and \(S\) is complex apparent power (VA). Inductive loads absorb reactive power (\(Q > 0\)); capacitive loads supply it (\(Q < 0\)).

A crucial insight: in a balanced system, the total instantaneous three-phase power is constant (no pulsation), unlike single-phase power which oscillates at twice the supply frequency. This is one reason large machines are three-phase: the torque on the shaft is smooth.

1.3 Load Profiles and Characteristics

Power systems must be designed for the maximum foreseeable demand — the peak load — but most of the time they operate at significantly lower levels. The ratio of average load to peak load over a period is the load factor, typically 0.5–0.7 for a utility system. A high load factor is economically desirable because it means capital-intensive generation and transmission assets are utilized more uniformly.

Load is modelled in several ways depending on the application:

• Constant power (PQ): load draws fixed real and reactive power regardless of voltage. Used in most power flow studies.
• Constant current: load current magnitude is fixed; power varies linearly with voltage.
• Constant impedance (ZIP): load behaves as a fixed impedance; power varies as voltage squared.
• Composite ZIP model: \(P = P_0[a_P(V/V_0)^2 + b_P(V/V_0) + c_P]\) with \(a_P + b_P + c_P = 1\).

1.4 Deregulation and Market Structure

Historically, the power sector operated as a vertically integrated monopoly: one entity owned generation, transmission, and distribution within a service territory and was regulated by a government authority. Beginning in the 1990s, many jurisdictions restructured (“deregulated”) the generation sector to allow competition. Under restructuring:

• Transmission remains a regulated natural monopoly (high infrastructure cost, natural barrier to competition).
• An Independent System Operator (ISO) or Regional Transmission Organization (RTO) operates the grid neutrally.
• Generators bid into energy markets; market prices clear based on merit order (cheapest generators dispatch first).
• Distribution typically remains a regulated local monopoly.

Understanding deregulation is essential because it fundamentally changes the objectives of power system operation: the ISO must minimize cost (economic dispatch) while satisfying physical constraints (thermal limits, voltage limits, stability margins).

---

Chapter 2: Power System Components

2.1 Synchronous Generators

The synchronous generator is the dominant source of electrical energy in the power grid. A rotor carrying a DC field winding (excited by an exciter) rotates at synchronous speed \(n_s = 120f/p\) RPM, where \(p\) is the number of poles and \(f\) is the system frequency. The rotating magnetic field induces sinusoidal EMFs in the stator (armature) windings displaced 120° in space, producing the three-phase output.

2.1.1 Equivalent Circuit and Phasor Diagram

The steady-state per-phase equivalent circuit of a round-rotor synchronous generator is a voltage source \(\mathbf{E}_f\) (the internal or excitation voltage) in series with the synchronous impedance \(Z_s = R_a + jX_s\), where \(R_a\) is the armature resistance (usually small) and \(X_s\) is the synchronous reactance. With the terminal voltage \(\mathbf{V}_t\) and armature current \(\mathbf{I}_a\) (positive out of generator), the KVL equation is:

\[ \mathbf{E}_f = \mathbf{V}_t + \mathbf{I}_a Z_s \]

Neglecting \(R_a\) (valid for large machines):

\[ \mathbf{E}_f = \mathbf{V}_t + j X_s \mathbf{I}_a \]

The internal EMF \(E_f = k\Phi \omega_s\), where \(\Phi\) is the rotor flux and \(k\) is a machine constant. Increasing the field current increases \(E_f\) and hence the reactive power output (overexcited operation supplies \(Q\) to the system; underexcited absorbs \(Q\)).

The real power output of a generator connected to an infinite bus with bus voltage \(V\) through reactance \(X\) (assumed purely inductive) is:

\[ P = \frac{E_f V}{X} \sin\delta \]

where \(\delta\) is the power angle — the angle by which \(\mathbf{E}_f\) leads \(\mathbf{V}\). This \(P\)–\(\delta\) relationship is central to power system stability analysis.

2.1.2 Synchronization

Before a generator is connected to the grid, it must be synchronized: its terminal voltage must match the grid voltage in magnitude, frequency, phase sequence, and phase angle. The process uses a synchroscope or automatic synchronizer. Failure to synchronize properly results in large current transients (synchronizing current) that can damage the machine.

The infinite bus is an idealized bus whose voltage magnitude and frequency remain constant regardless of the generator’s real and reactive power output — it represents the aggregate of all other generators in the system. Real grids approach this ideal for a single small generator connected to a large system.

2.2 Transformers

Transformers are the enabling technology of high-voltage transmission. The ideal transformer with turns ratio \(a = N_1/N_2\) satisfies:

\[ \frac{V_1}{V_2} = a, \qquad \frac{I_1}{I_2} = \frac{1}{a}, \qquad Z_1 = a^2 Z_2 \]

The last relation is the impedance transformation property: a load impedance \(Z_2\) on the secondary appears as \(a^2 Z_2\) at the primary. This is fundamental for per-unit analysis.

2.2.1 Exact and Approximate Equivalent Circuits

The exact equivalent circuit of a practical transformer includes:

• Primary leakage reactance \(X_1\) and resistance \(R_1\)
• Magnetizing branch: shunt susceptance \(B_m\) (core loss) and shunt conductance \(G_c\) (iron loss)
• Secondary leakage reactance \(X_2\) and resistance \(R_2\) (referred to primary: \(a^2 X_2\), \(a^2 R_2\))

For power systems analysis, the magnetizing branch is often neglected (it draws only 1–3% of rated current) and the equivalent circuit reduces to a series impedance:

\[ Z_{eq} = (R_1 + a^2 R_2) + j(X_1 + a^2 X_2) = R_{eq} + jX_{eq} \]

The short-circuit test yields \(Z_{eq}\) directly; the open-circuit test yields the magnetizing parameters.

2.2.2 Three-Phase Transformer Connections

Three-phase transformers can be connected in four configurations: Y–Y, Y–\(\Delta\), \(\Delta\)–Y, and \(\Delta\)–\(\Delta\). The choice has important consequences:

The Y–\(\Delta\) (or \(\Delta\)–Y) connection is preferred for most power transformers because:

1. The Y-connected winding provides a neutral point that can be grounded, controlling zero-sequence currents.
2. The \(\Delta\)-connected winding provides a path for circulating third-harmonic currents, suppressing third-harmonic voltage distortion.
3. The 30° phase shift blocks zero-sequence current from propagating through the transformer.

In per-unit analysis with properly chosen bases, the transformer equivalent circuit is simply the series leakage impedance — the ideal turns ratio disappears.

2.2.3 Voltage Regulation and Efficiency

The voltage regulation of a transformer at full load, expressed as a percentage, is:

\[ VR\% = \frac{|V_{2,\text{no load}}| - |V_{2,\text{full load}}|}{|V_{2,\text{full load}}|} \times 100 \]

For lagging (inductive) loads, voltage regulation is positive (voltage drops under load). For leading (capacitive) loads, regulation can be negative (terminal voltage actually rises above no-load value).

Efficiency is:

\[ \eta = \frac{P_{out}}{P_{out} + P_{core} + P_{Cu}} = \frac{S\cos\theta}{S\cos\theta + P_c + I^2 R_{eq}} \]

Maximum efficiency occurs when copper losses equal core losses: \(I^2 R_{eq} = P_c\).

2.3 Transmission Lines

Transmission lines are distributed-parameter devices. However, for the purpose of power flow and stability analysis, lumped-parameter models suffice for most line lengths encountered in practice.

2.3.1 Line Parameters

Four distributed parameters characterize a transmission line per unit length:

• Series resistance \(r\) [\(\Omega\)/km]: due to conductor resistivity; increases with temperature; skin effect increases effective \(r\) at high frequency.
• Series inductance \(l\) [H/km]: due to magnetic flux linkage both internal to the conductor and external in the surrounding space.
• Shunt capacitance \(c\) [F/km]: due to the electric field between conductors.
• Shunt conductance \(g\) [S/km]: due to leakage current through insulation and corona; usually negligible.

For a solid cylindrical conductor with Geometric Mean Radius (GMR) \(r'\) and Geometric Mean Distance (GMD) \(D\) between conductors:

\[ l = \frac{\mu_0}{2\pi}\ln\!\frac{D}{r'} \quad \text{[H/m]}, \qquad c = \frac{2\pi\epsilon_0}{\ln(D/r)} \quad \text{[F/m]} \]

where \(r\) here is the conductor’s actual radius (not GMR). Notice that \(lc = \mu_0\epsilon_0 = 1/v_{prop}^2\), where \(v_{prop} \approx c_{light}\) for lines in air.

2.3.2 Short, Medium, and Long Line Models

Phasor circuit equations:

\[ \mathbf{V}_S = \mathbf{V}_R + Z\mathbf{I}_R, \qquad \mathbf{I}_S = \mathbf{I}_R \]

For the nominal \(\pi\) model:

\[ \mathbf{V}_S = \left(1 + \frac{ZY}{2}\right)\mathbf{V}_R + Z\mathbf{I}_R \]\[ \mathbf{I}_S = Y\left(1 + \frac{ZY}{4}\right)\mathbf{V}_R + \left(1 + \frac{ZY}{2}\right)\mathbf{I}_R \]

In matrix form, this is the ABCD (two-port) representation:

\[ \begin{pmatrix}\mathbf{V}_S \\ \mathbf{I}_S\end{pmatrix} = \begin{pmatrix}A & B \\ C & D\end{pmatrix}\begin{pmatrix}\mathbf{V}_R \\ \mathbf{I}_R\end{pmatrix} \]

with \(A = D = 1 + ZY/2\), \(B = Z\), \(C = Y(1 + ZY/4)\), and \(AD - BC = 1\).

The telegraphers’ equations for voltage and current as functions of position \(x\) along the line (positive toward the receiving end) and time lead to the phasor equations:

\[ \frac{d^2\mathbf{V}}{dx^2} = \gamma^2 \mathbf{V}, \qquad \frac{d^2\mathbf{I}}{dx^2} = \gamma^2 \mathbf{I} \]

where \(\gamma = \sqrt{zy} = \alpha + j\beta\) is the propagation constant, \(\alpha\) is the attenuation constant (Np/m), and \(\beta = \omega/v_{phase}\) is the phase constant (rad/m).

The general solution gives the hyperbolic form:

\[ \mathbf{V}(x) = \mathbf{V}_R \cosh(\gamma x) + Z_c \mathbf{I}_R \sinh(\gamma x) \]\[ \mathbf{I}(x) = \frac{\mathbf{V}_R}{Z_c}\sinh(\gamma x) + \mathbf{I}_R \cosh(\gamma x) \]

where \(Z_c = \sqrt{z/y}\) is the characteristic impedance (also called surge impedance), a complex number for lossy lines but approximately real for lossless lines.

2.3.3 Surge Impedance Loading

When the line operates at SIL, the voltage profile is flat (no voltage variation along the line) and the reactive power requirement from external sources is zero. This is the natural operating condition for which the line is “self-compensating.”

For loading above SIL, the line consumes net reactive power (voltage drops toward the receiving end). For loading below SIL, the line generates net reactive power (Ferranti effect — the receiving-end voltage can actually exceed the sending-end voltage on a lightly loaded long line).

---

Chapter 3: The Per-Unit System

3.1 Motivation and Concept

Power systems contain transformers that introduce voltage-level changes at many points. A detailed analysis that tracks actual voltages and currents in each voltage zone is cumbersome and error-prone. The per-unit (pu) system normalizes all quantities relative to chosen base values, so that the ideal transformer disappears from the circuit and voltages cluster near unity, making it easy to spot errors.

3.2 Base Quantities and Their Relations

Two independent base quantities must be chosen: a three-phase apparent power base \(S_{base}\) (in MVA) and a line-to-line voltage base \(V_{base}\) (in kV). The remaining bases are then determined:

\[ I_{base} = \frac{S_{base}}{\sqrt{3}\, V_{base}} \quad \text{[kA]} \]\[ Z_{base} = \frac{V_{base}^2}{S_{base}} \quad \text{[}\Omega\text{, with } V_{base}\text{ in kV, }S_{base}\text{ in MVA]} \]\[ Y_{base} = \frac{1}{Z_{base}} \quad \text{[S]} \]

Note that the same \(S_{base}\) applies throughout the entire system. The voltage base, however, changes with each transformer, tracking the nominal turns ratio. If the high-voltage side has base \(V_{base,H}\) and the turns ratio is \(a\), then the low-voltage base is \(V_{base,L} = V_{base,H}/a\), and the impedance base on the low side is \(Z_{base,L} = V_{base,L}^2/S_{base}\).

3.3 Change of Base

Equipment impedances are given on the manufacturer’s nameplate as a percentage or per-unit value on the device’s own rating base \((S_{base,old}, V_{base,old})\). When the system base differs, the impedance must be converted:

\[ Z_{pu,new} = Z_{pu,old} \times \frac{S_{base,new}}{S_{base,old}} \times \left(\frac{V_{base,old}}{V_{base,new}}\right)^2 \]

3.4 Advantages of the Per-Unit System

1. Transformer equations simplify: The ideal transformer with turns ratio \(a:1\) becomes a short circuit (0 pu voltage drop) when both sides use their own voltage bases derived consistently from one chosen base.
2. Values cluster near unity: All voltages are near 1.0 pu, all impedances are in a narrow range (typically 0.05–0.20 pu for transformers, 0.01–0.30 pu for lines), making errors immediately obvious.
3. Machine equations simplify: Generator per-unit reactances (\(X_d''\), \(X_d'\), \(X_d\)) are directly comparable across machines of different ratings.
4. Symmetrical component analysis: Sequence networks use per-unit quantities consistently, eliminating factors of \(\sqrt{3}\) that would otherwise appear.

---

Chapter 4: Power Flow Analysis

4.1 The Power Flow Problem

The power flow (or load flow) problem is the computation of steady-state voltages and power flows throughout the network given specified generation and load conditions. It is the most fundamental calculation in power systems engineering, forming the basis for planning, operations, stability studies, and optimal dispatch.

At each bus (node) in the network, four variables are associated:

• \(|V_i|\): voltage magnitude
• \(\theta_i\): voltage angle
• \(P_i\): net real power injection (generation minus load)
• \(Q_i\): net reactive power injection

Two of these four quantities are known (specified), and two are unknown (to be solved). The bus type classification determines which two are specified:

4.2 The Bus Admittance Matrix \(Y_{bus}\)

The network topology and impedances are encoded in the bus admittance matrix \(\mathbf{Y}_{bus}\). For an \(n\)-bus system, \(\mathbf{Y}_{bus}\) is an \(n \times n\) complex symmetric matrix (assuming no phase-shifting transformers) defined by:

\[ Y_{ii} = \sum_{k \neq i} y_{ik} + y_{i,shunt} \quad \text{(diagonal elements)} \]\[ Y_{ij} = -y_{ij} \quad (i \neq j) \quad \text{(off-diagonal elements)} \]

where \(y_{ij} = 1/Z_{ij}\) is the admittance of the branch between buses \(i\) and \(j\), and \(y_{i,shunt}\) includes transformer magnetizing branches and line charging capacitances connected to bus \(i\).

The nodal current injection equation is:

\[ \mathbf{I}_{bus} = \mathbf{Y}_{bus}\, \mathbf{V}_{bus} \]

For a simple example: consider a three-bus system with line admittances \(y_{12}\), \(y_{13}\), \(y_{23}\) and shunt admittances \(y_{1s}\), \(y_{2s}\), \(y_{3s}\):

\[ \mathbf{Y}_{bus} = \begin{pmatrix} y_{12} + y_{13} + y_{1s} & -y_{12} & -y_{13} \\ -y_{12} & y_{12} + y_{23} + y_{2s} & -y_{23} \\ -y_{13} & -y_{23} & y_{13} + y_{23} + y_{3s} \end{pmatrix} \]

4.2.1 Building \(Y_{bus}\) — Systematic Rules

The construction of \(\mathbf{Y}_{bus}\) follows a simple algorithm: (1) set all elements to zero; (2) for each branch between buses \(i\) and \(j\) with series admittance \(y_{series}\) and half-shunt admittances \(y_{shunt}/2\) at each end, add \(y_{series}\) to \(Y_{ii}\) and \(Y_{jj}\), subtract \(y_{series}\) from \(Y_{ij}\) and \(Y_{ji}\), and add \(y_{shunt}/2\) to \(Y_{ii}\) and \(Y_{jj}\); (3) for each shunt element at bus \(i\), add its admittance to \(Y_{ii}\).

The resulting matrix is sparse for realistic systems (each bus is connected to only a few others) — a 1000-bus system might have \(\mathbf{Y}_{bus}\) with fewer than 5000 nonzero entries out of \(10^6\) possible. Sparse matrix techniques are essential for efficient computation.

4.3 Power Flow Equations

The complex power injection at bus \(i\) is:

\[ S_i = P_i + jQ_i = \mathbf{V}_i \mathbf{I}_i^* \]

where the current injection \(\mathbf{I}_i = \sum_{k=1}^n Y_{ik} \mathbf{V}_k\). Writing \(\mathbf{V}_i = |V_i|e^{j\theta_i}\) and \(Y_{ik} = |Y_{ik}|e^{j\gamma_{ik}} = G_{ik} + jB_{ik}\):

\[ P_i = \sum_{k=1}^n |V_i||V_k|\bigl(G_{ik}\cos(\theta_i - \theta_k) + B_{ik}\sin(\theta_i - \theta_k)\bigr) \]\[ Q_i = \sum_{k=1}^n |V_i||V_k|\bigl(G_{ik}\sin(\theta_i - \theta_k) - B_{ik}\cos(\theta_i - \theta_k)\bigr) \]

These are the power flow equations. For an \(n\)-bus system with one slack bus and \(n_{PV}\) PV buses and \(n_{PQ} = n - 1 - n_{PV}\) PQ buses, there are \(2n_{PQ} + n_{PV}\) nonlinear equations in the same number of unknowns.

4.4 Newton-Raphson Load Flow

The Newton-Raphson (NR) method is the workhorse of modern power flow computation. Starting from an initial guess (typically a “flat start”: \(|V_i| = 1.0\) pu, \(\theta_i = 0\) for all buses), it linearizes the power flow equations around the current estimate and solves for a correction:

\[ \begin{pmatrix}\Delta \mathbf{P} \\ \Delta \mathbf{Q}\end{pmatrix} = \mathbf{J} \begin{pmatrix}\Delta \boldsymbol{\theta} \\ \Delta |\mathbf{V}|\end{pmatrix} \]

where \(\mathbf{J}\) is the Jacobian matrix of partial derivatives:

\[ \mathbf{J} = \begin{pmatrix} \partial \mathbf{P}/\partial \boldsymbol{\theta} & \partial \mathbf{P}/\partial |\mathbf{V}| \\ \partial \mathbf{Q}/\partial \boldsymbol{\theta} & \partial \mathbf{Q}/\partial |\mathbf{V}| \end{pmatrix} = \begin{pmatrix} \mathbf{J}_1 & \mathbf{J}_2 \\ \mathbf{J}_3 & \mathbf{J}_4 \end{pmatrix} \]

The NR iteration proceeds:

1. Compute mismatches \(\Delta P_i = P_i^{spec} - P_i^{calc}\) and \(\Delta Q_i = Q_i^{spec} - Q_i^{calc}\).
2. Solve the linear system \(\mathbf{J} \cdot [\Delta\boldsymbol{\theta}; \Delta|\mathbf{V}|/|\mathbf{V}|] = [\Delta\mathbf{P}; \Delta\mathbf{Q}]\) (usually with LU factorization).
3. Update: \(\theta_i^{new} = \theta_i + \Delta\theta_i\), \(|V_i|^{new} = |V_i|(1 + \Delta|V_i|/|V_i|)\).
4. Check convergence: if \(\max|\Delta P_i|, \max|\Delta Q_i| < \epsilon\) (typically \(\epsilon = 10^{-4}\) pu), stop.

The NR method converges quadratically near the solution — the error roughly squares each iteration. Typically 3–7 iterations suffice from a flat start, regardless of system size. The principal cost is factorizing the Jacobian, which is \(\mathcal{O}(n^{1.5})\) to \(\mathcal{O}(n^2)\) for sparse systems.

4.5 Gauss-Seidel Method

The Gauss-Seidel (GS) method, historically used before NR became practical, iterates bus voltages one at a time. For PQ bus \(i\):

\[ \mathbf{V}_i^{(k+1)} = \frac{1}{Y_{ii}}\left[\frac{(P_i - jQ_i)^*}{(\mathbf{V}_i^{(k)})^*} - \sum_{j \neq i} Y_{ij}\mathbf{V}_j^{(\text{latest})}\right] \]

where “latest” means using \(\mathbf{V}_j^{(k+1)}\) for \(j < i\) (already updated in this sweep) and \(\mathbf{V}_j^{(k)}\) for \(j > i\).

For a PV bus, after computing \(\mathbf{V}_i^{(k+1)}\), the reactive power \(Q_i\) is adjusted to enforce \(|V_i| = |V_i|^{spec}\), and then the voltage is corrected in magnitude while keeping the computed angle.

GS converges linearly, requiring 50–200 iterations for typical systems — far slower than NR. Its main advantage is simplicity and low memory per iteration. Acceleration factors (typically 1.4–1.6) applied to the voltage correction improve convergence somewhat.

4.6 Fast Decoupled Load Flow

In high-voltage transmission systems, there is approximate decoupling between \(P\)-\(\theta\) and \(Q\)-\(|V|\) relationships: real power flow is primarily determined by angle differences, and reactive power is primarily determined by voltage magnitudes. The fast decoupled load flow (FDLF) exploits this decoupling by using constant, simplified Jacobian submatrices.

Under the assumptions that \(|B_{ik}| \gg G_{ik}\), \(\cos(\theta_{ik}) \approx 1\), \(G_{ik}\sin(\theta_{ik}) \ll B_{ik}\), and shunt susceptances are neglected, the Jacobian simplifies to:

\[ \frac{\Delta\mathbf{P}}{|\mathbf{V}|} \approx \mathbf{B}' \Delta\boldsymbol{\theta} \]\[ \frac{\Delta\mathbf{Q}}{|\mathbf{V}|} \approx \mathbf{B}'' \Delta|\mathbf{V}| \]

where \(\mathbf{B}'\) and \(\mathbf{B}''\) are constant real matrices derived from the susceptance part of \(\mathbf{Y}_{bus}\). Because these matrices are constant (factorized once), each iteration requires only two forward-backward substitutions rather than a full Jacobian assembly and factorization. The method converges in 2–5 times more iterations than NR, but each iteration is much cheaper, making FDLF competitive for large systems.

4.7 Control of Active and Reactive Power

Power flow control is accomplished through:

• Generator real power dispatch: Changes the real power balance across the network (changes angles).
• Generator excitation (AVR): Controls terminal voltage at PV buses (changes reactive dispatch).
• Transformer tap changers: Voltage regulating transformers with variable turns ratio control voltage magnitude at a bus or reactive power flow on a branch.
• Phase-shifting transformers: Control real power flow on a specific branch by introducing a phase shift.
• Switched capacitor/reactor banks: Provide discrete reactive power compensation.
• FACTS devices (SVC, STATCOM, TCSC): Provide fast, continuously variable reactive compensation and, for series devices, control of impedance.

---

Chapter 5: Fault Analysis and Protection

5.1 Symmetrical Three-Phase Faults

A three-phase fault (3LG — three-line-to-ground) is the most severe type of short circuit in power systems. Although less common than single-line-to-ground faults, it is often used as the design-basis fault for circuit breaker rating and relay setting because it produces the largest fault current in most circumstances.

5.1.1 Thevenin Equivalent and Fault Current

The computation of fault current uses the pre-fault Thevenin equivalent seen at the fault bus \(k\). Before the fault, let the bus voltage be \(\mathbf{V}_{f}\) (the pre-fault voltage, often taken as 1.0 pu for a “flat start” pre-fault condition). The Thevenin impedance \(Z_{kk}\) is the \(k,k\) diagonal element of the bus impedance matrix \(\mathbf{Z}_{bus} = \mathbf{Y}_{bus}^{-1}\).

The voltage at every other bus \(i\) during the fault is:

\[ \mathbf{V}_i^{fault} = \mathbf{V}_i^{prefault} - Z_{ik}\mathbf{I}_f \]

where \(Z_{ik}\) is the \(i,k\) element of \(\mathbf{Z}_{bus}\). The current flowing from bus \(i\) to bus \(j\) through a branch with impedance \(Z_{ij}\) during the fault is:

\[ \mathbf{I}_{ij}^{fault} = \frac{\mathbf{V}_i^{fault} - \mathbf{V}_j^{fault}}{Z_{ij}} \]

5.1.2 Subtransient, Transient, and Steady-State Reactances

For fault analysis, the synchronous generator is not modelled with its steady-state synchronous reactance \(X_d\); instead, the subtransient or transient reactance is used depending on the time frame of interest.

For circuit breaker interrupting duty calculations, the subtransient reactances of all generators are used to form \(\mathbf{Z}_{bus}\).

5.2 Symmetrical Components

Unsymmetrical faults (single-line-to-ground, line-to-line, double-line-to-ground) are far more common than three-phase faults. Analysing them with the full three-phase network is complex. The method of symmetrical components, introduced by Fortescue (1918), decomposes any unbalanced three-phase set into three balanced sequence components.

The transformation matrix is denoted \(\mathbf{A}\); its inverse is:

\[ \mathbf{A}^{-1} = \frac{1}{3}\begin{pmatrix}1 & a & a^2 \\ 1 & a^2 & a \\ 1 & 1 & 1\end{pmatrix} \]

5.2.1 Sequence Networks

The power of symmetrical components is that, for balanced networks (symmetric impedance matrices), the sequence networks are decoupled: positive-sequence currents flow only through the positive-sequence network, negative-sequence through the negative-sequence network, and zero-sequence through the zero-sequence network.

Each sequence network is a standard single-phase circuit described by its Thevenin equivalent:

• Positive-sequence: All EMF sources are present (generators produce positive sequence). Impedances are the normal positive-sequence values.
• Negative-sequence: No EMF sources (generator negative-sequence impedance: \(Z_2 \approx X_d''\) for most machines). Impedances are slightly different from positive-sequence (equal for balanced lines and transformers).
• Zero-sequence: No EMF sources. Zero-sequence impedances depend critically on transformer winding connections and system grounding. Delta windings block zero-sequence; Y-grounded windings pass it.

5.2.2 Single-Line-to-Ground Fault

For a single-line-to-ground (SLG) fault on phase \(a\) at bus \(k\) with fault impedance \(Z_f\):

Boundary conditions: \(\mathbf{I}_{fb} = \mathbf{I}_{fc} = 0\), \(\mathbf{V}_{fka} = Z_f \mathbf{I}_{fa}\).

Applying the symmetrical component transformation to the boundary conditions shows that the sequence currents are equal:

\[ \mathbf{I}_{a0} = \mathbf{I}_{a1} = \mathbf{I}_{a2} = \frac{\mathbf{V}_f}{Z_{kk}^{(1)} + Z_{kk}^{(2)} + Z_{kk}^{(0)} + 3Z_f} \]

where \(Z_{kk}^{(1)}\), \(Z_{kk}^{(2)}\), \(Z_{kk}^{(0)}\) are the Thevenin impedances of the positive-, negative-, and zero-sequence networks at bus \(k\). The total fault current is:

\[ \mathbf{I}_{fa} = 3\mathbf{I}_{a0} = \frac{3\mathbf{V}_f}{Z_{kk}^{(1)} + Z_{kk}^{(2)} + Z_{kk}^{(0)} + 3Z_f} \]

This is represented by connecting the three sequence networks in series at the fault terminals.

5.2.3 Line-to-Line and Double-Line-to-Ground Faults

For a line-to-line fault (phases \(b\) and \(c\), with \(Z_f\)): The boundary conditions give \(\mathbf{I}_{a0} = 0\) and \(\mathbf{I}_{a1} = -\mathbf{I}_{a2}\). The sequence networks connect positive in parallel with negative (no zero-sequence).

\[ \mathbf{I}_{a1} = \frac{\mathbf{V}_f}{Z_{kk}^{(1)} + Z_{kk}^{(2)} + Z_f} \]

For a double-line-to-ground fault (phases \(b\) and \(c\) to ground): all three sequence networks are involved. The positive-sequence current is:

\[ \mathbf{I}_{a1} = \frac{\mathbf{V}_f}{Z_{kk}^{(1)} + Z_{kk}^{(2)}||(Z_{kk}^{(0)} + 3Z_f)} \]

where \(||\) denotes parallel combination.

5.3 Protection Systems

5.3.1 Objectives and Basic Concepts

The purpose of a protection system is to detect abnormal conditions and quickly isolate the faulted element to limit damage and maintain system integrity. The key performance metrics are:

• Reliability: The relay must operate when required (security against failure-to-trip) but not operate when not required (security against mal-trip). Reliability = (number of correct operations)/(number of operations required).
• Speed: Faster fault clearing reduces equipment damage and improves stability. Modern numerical relays can operate in 1–2 cycles (16–33 ms at 60 Hz).
• Selectivity (discrimination): Only the faulted element is isolated; as much of the healthy system as possible remains in service.
• Sensitivity: The relay must detect the minimum fault current.

5.3.2 Protection Zones

Protection systems are organized in overlapping zones of protection. Each zone is bounded by current transformers (CTs) and protected by a relay or relay set. The overlap between zones ensures that no point in the system is unprotected.

Primary protection provides the first line of defence for a given zone. Backup protection (either local or remote) operates if the primary fails:

• Local backup: A second relay at the same location operates the same breaker(s) with a time delay if the primary relay does not operate.
• Remote backup: Relays at adjacent substations detect faults in a neighbouring zone and clear them with a time delay after the primary relay fails.

5.3.3 Types of Protective Relays

---

Chapter 6: Distribution Systems

6.1 Distribution System Configurations

The distribution system delivers electrical energy from bulk-power substations (typically 69/26 kV or 115/13.8 kV) to end consumers. Unlike transmission systems (which form meshed networks for reliability), distribution systems are typically operated radially — the power flows in one direction from the substation to the loads. Radial operation simplifies protection coordination (overcurrent relays can be easily coordinated) but means that a single fault isolates all downstream customers.

Common distribution configurations:

• Simple radial: One feeder, one path to each customer. Low cost, poor reliability.
• Loop (open ring): Two feeders connected to a normally open switch. When a section faults, the switch closes, restoring service to the unfaulted portion. Provides sectionalizing capability.
• Network: Multiple transformers feed a common bus through network protectors. Used in dense urban areas for high reliability (automatic restoration).

6.2 Voltage Drop and Loss Calculations

For a radial feeder segment with impedance \(R + jX\) carrying current \(I\) at power factor angle \(\phi\), the approximate voltage drop is:

\[ \Delta V \approx I(R\cos\phi + X\sin\phi) \]

This is the approximate formula (ignoring the quadrature term) widely used in distribution planning. The exact formula, using the Pythagoras theorem on the receiving-end voltage phasor:

\[ |V_S|^2 = (|V_R| + IR\cos\phi + IX\sin\phi)^2 + (IX\cos\phi - IR\sin\phi)^2 \]

yields the same approximate result when the second term (quadrature component) is small relative to \(|V_R|\).

For a feeder with multiple uniformly distributed loads, the voltage drop along the feeder can be computed by summing the contribution of each load-current segment. With load density \(\rho\) (A/km) over a feeder length \(L\):

\[ \Delta V_{total} = \frac{1}{2}\rho L \cdot z \cdot L = \frac{I_{total}L\,z}{2} \]

where \(z\) is the impedance per unit length. The factor of 1/2 arises because, for a uniformly distributed load, the effective electrical centre of the load is at the midpoint.

The total real power loss in a feeder with uniformly distributed load:

\[ P_{loss} = \frac{1}{3} I_{total}^2 R_{total} \]

(The factor \(1/3\) reflects the same centre-of-load principle.)

6.3 Shunt Capacitors in Distribution Systems

Distribution feeders supply predominantly inductive loads (motors, fluorescent lighting). The lagging power factor increases line current for a given real power delivery, causing excess voltage drop and resistive losses. Shunt capacitors compensate by supplying reactive power locally.

With capacitor reactive compensation \(Q_c\) (kVAr) at the load:

\[ I_{new} = \frac{\sqrt{P^2 + (Q - Q_c)^2}}{V} \]

The power factor improves from \(\cos\phi_1\) to \(\cos\phi_2\):

\[ Q_c = P(\tan\phi_1 - \tan\phi_2) \]

6.4 Distribution Automation

Distribution Automation (DA) uses remote-controllable switches, sensors, and communication infrastructure to achieve:

• Automatic fault isolation and service restoration (FLISR): Upon detection of a sustained fault, automated switches isolate the faulted section and close normally-open ties to restore service to unfaulted sections — all within seconds, without requiring a field crew.
• Volt/VAR optimization (VVO): Automated control of tap changers, capacitor banks, and distributed generators to minimize losses and maintain voltage within limits throughout the day as load changes.
• Outage management: Integration with customer outage calls and SCADA data to rapidly identify fault locations.

---

Chapter 7: Economic Dispatch

7.1 The Economic Dispatch Problem

Given a set of generating units committed to serve a load, the economic dispatch problem finds the output of each unit that minimizes total generation cost while meeting the load demand. Unlike the unit commitment problem (which decides which units to turn on), economic dispatch assumes the set of online units is fixed.

Each generating unit \(i\) has a cost function \(C_i(P_i)\) [$/h] expressing the fuel cost as a function of real power output \(P_i\). For thermal units, this is commonly approximated by a quadratic:

\[ C_i(P_i) = \alpha_i + \beta_i P_i + \gamma_i P_i^2 \quad \text{[\$/h]} \]

where \(\alpha_i\), \(\beta_i\), \(\gamma_i\) are cost coefficients determined from heat rate curves and fuel prices. The incremental cost (marginal cost) is:

\[ \lambda_i = \frac{dC_i}{dP_i} = \beta_i + 2\gamma_i P_i \quad \text{[\$/MWh]} \]

7.2 Equal Incremental Cost Criterion

The optimality condition for economic dispatch (without transmission losses) follows from applying the KKT conditions to the constrained minimization:

\[ \min \sum_{i=1}^N C_i(P_i) \quad \text{subject to} \quad \sum_{i=1}^N P_i = P_D, \quad P_{i,min} \leq P_i \leq P_{i,max} \]

7.2.1 Lambda Iteration

For quadratic cost functions, the optimal dispatch for a given \(\lambda\) is:

\[ P_i^*(\lambda) = \frac{\lambda - \beta_i}{2\gamma_i} \]

clipped to \([P_{i,min}, P_{i,max}]\). The lambda iteration algorithm:

1. Guess an initial \(\lambda\).
2. Compute \(P_i^*(\lambda)\) for each unit (with limit enforcement).
3. Compute total generation \(\sum P_i^*(\lambda)\).
4. If \(\sum P_i^* > P_D\), decrease \(\lambda\); if \(\sum P_i^* < P_D\), increase \(\lambda\).
5. Use bisection or Newton’s method on \(\lambda\) to satisfy the balance equation:

\[ \sum_{i=1}^N P_i^*(\lambda) = P_D \]

7.3 Economic Dispatch with Transmission Losses

When transmission losses are significant, the power balance becomes:

\[ \sum_{i=1}^N P_i = P_D + P_L \]

where \(P_L\) is the total system loss. Using the B-coefficient (Kron’s) loss formula:

\[ P_L = \sum_{i=1}^N \sum_{j=1}^N P_i B_{ij} P_j + \sum_{i=1}^N B_{0i}P_i + B_{00} \]

The optimality conditions become:

\[ \frac{dC_i}{dP_i} \cdot \frac{1}{1 - \partial P_L/\partial P_i} = \lambda \quad \Longrightarrow \quad \frac{dC_i}{dP_i} \cdot PF_i = \lambda \]

where \(PF_i = 1/(1 - \partial P_L/\partial P_i)\) is the penalty factor for unit \(i\). A unit in a heavily loaded area (where an additional MW causes more system losses) has a higher penalty factor and therefore operates at a lower output than it would in the lossless case.

---

Chapter 8: Power System Stability

8.1 Introduction to Stability

Power system stability refers to the ability of the system to regain a state of operating equilibrium after being subjected to a physical disturbance, with most system variables bounded so that the entire system remains intact. The IEEE/CIGRE classification distinguishes:

• Rotor angle stability: Ability of synchronous machines to remain in synchronism. Subdivided into small-signal stability (response to small perturbations) and transient stability (response to large disturbances like faults).
• Voltage stability: Ability to maintain acceptable voltages at all buses after a disturbance. Voltage collapse occurs when the system cannot supply sufficient reactive power.
• Frequency stability: Ability to maintain frequency within acceptable limits after a large power imbalance. Involves governor action and load shedding if necessary.

8.2 The Swing Equation

The mechanical equation of motion of a synchronous generator rotor is derived from Newton’s second law for rotation:

\[ J\frac{d^2\theta_m}{dt^2} = T_{mech} - T_{elec} \]

where \(J\) is the moment of inertia [kg·m²], \(\theta_m\) is the mechanical rotor angle, \(T_{mech}\) is the mechanical torque input, and \(T_{elec}\) is the electromagnetic torque. Converting to electrical radians and per-unit quantities, defining the power angle \(\delta\) as the angle of the rotor’s internal EMF relative to a reference rotating at synchronous speed \(\omega_s\):

Including a damping term proportional to the speed deviation \(\Delta\omega = d\delta/dt\):

\[ \frac{2H}{\omega_s}\frac{d^2\delta}{dt^2} + D\frac{d\delta}{dt} = P_{mech} - P_{elec} \]

For a generator connected to an infinite bus through total reactance \(X\), \(P_{elec} = P_{max}\sin\delta\) where \(P_{max} = E'V/X\).

8.3 Small-Signal Stability

For small perturbations about an equilibrium point \((\delta_0, \omega_0 = \omega_s)\), let \(\delta = \delta_0 + \Delta\delta\). Linearizing the swing equation:

\[ \frac{2H}{\omega_s}\frac{d^2\Delta\delta}{dt^2} + D\frac{d\Delta\delta}{dt} + P_{max}\cos\delta_0 \cdot \Delta\delta = 0 \]

This is a second-order linear ODE with the characteristic equation:

\[ \frac{2H}{\omega_s}s^2 + Ds + K_s = 0 \]

where \(K_s = P_{max}\cos\delta_0\) is the synchronizing power coefficient. The system is stable if \(K_s > 0\) (which requires \(\delta_0 < 90°\) for the classical model) and \(D > 0\). The natural frequency of oscillation (undamped) is:

\[ \omega_n = \sqrt{\frac{K_s \omega_s}{2H}} \quad \text{[rad/s]} \]

This typically yields oscillation frequencies of 0.5–2 Hz for a machine oscillating against an infinite bus, and 0.1–0.7 Hz for inter-area modes involving groups of machines.

8.4 Transient Stability and the Equal Area Criterion

Transient stability concerns the ability of the system to maintain synchronism after a large, sudden disturbance — typically a three-phase fault followed by circuit breaker operation.

The equal area criterion provides geometric intuition but is limited to single-machine-infinite-bus (SMIB) systems and the classical model. Multi-machine transient stability requires numerical integration of the coupled swing equations.

---

Chapter 9: Smart Grids

9.1 Philosophy and Motivation

The traditional power grid was designed for unidirectional power flow — from large central generators to passive loads — and was controlled by a relatively small number of dispatchable generators. The smart grid concept emerged in the 2000s to describe the integration of advanced sensing, communication, and control technologies into the power grid to:

1. Enable two-way power flow (to accommodate distributed generation and storage).
2. Enable two-way communication (between utilities and customers).
3. Improve reliability through self-healing capabilities.
4. Improve efficiency through demand response and real-time optimization.
5. Integrate renewable energy and electric vehicles.
6. Empower customers with real-time information and pricing signals.

The US Energy Independence and Security Act of 2007 and subsequent policies worldwide catalysed significant investment in smart grid infrastructure. The European Smart Grid Technology Platform has articulated similar objectives under the concept of “active distribution networks.”

9.2 Advanced Metering Infrastructure (AMI)

Advanced Metering Infrastructure consists of smart meters at customer premises, a communication network, and data management systems at the utility. Smart meters differ from traditional meters in several key ways:

• Two-way communication: Can receive commands (e.g., demand response signals, pricing information) as well as transmit consumption data.
• Granular measurement: Record energy consumption at short intervals (15 minutes or hourly), enabling time-differentiated billing and detailed load profiling.
• Remote connect/disconnect: Eliminates the need for a service truck for routine service changes.
• Power quality monitoring: Some advanced meters record voltage sags, momentary interruptions, and harmonics.

Communication technologies used in AMI include Power Line Communication (PLC), RF mesh networks (900 MHz, IEEE 802.15.4g), cellular (4G/5G), and fibre. The choice involves trade-offs between cost, latency, bandwidth, and reliability.

9.3 Demand Response

DR programmes take several forms:

• Price-based: Time-of-Use (TOU) rates, Critical Peak Pricing (CPP), Real-Time Pricing (RTP). Customers voluntarily shift load in response to price signals. Effective for elastic loads (EV charging, water heaters, pool pumps, smart thermostats).
• Incentive-based: Direct Load Control (DLC) — utility can directly curtail or switch off enrolled devices (typically central A/C compressors). Customers receive a bill credit. Fast-responding, but requires customer consent and appropriate devices.
• Emergency DR: Interruptible service contracts with large industrial customers, who receive rate discounts in exchange for curtailment upon request.

The aggregate demand response resource can be modelled as a virtual generator providing negative load (load reduction) at a bid price determined by the customer’s value of electricity at that time.

9.4 Distributed Energy Resources and Microgrids

The integration of DER creates bidirectional power flow in distribution feeders, violating the assumptions of traditional protection and voltage regulation schemes:

• Reverse power flow: During high PV output and low load, feeders may export power back toward the substation, confusing directional overcurrent relays.
• Voltage rise: DER injection can cause voltage at the end of a feeder to rise above limits (the Ferranti-like effect in distribution).
• Protection blinding: DER can reduce the fault current seen by the substation relay, causing it to fail to detect a fault.
• Islanding: DER may energize a section of the feeder that has been disconnected from the grid — dangerous for utility workers. Anti-islanding protection is mandatory.

Key microgrid components:

• Point of Common Coupling (PCC): The electrical boundary between the microgrid and the main grid; a controllable switch allows islanding.
• Energy Management System (EMS): Optimizes dispatch of DER within the microgrid.
• Droop control: Each inverter or generator in islanded mode uses frequency droop (\(\Delta f = -D\Delta P\)) and voltage droop to share load proportionally without requiring a communication link.

9.5 SCADA and Energy Management Systems

A modern Energy Management System (EMS) at an ISO/RTO contains:

• State Estimation (SE): The foundational application. Using redundant real-time measurements (\(P\), \(Q\), \(|V|\) from meters, and line flows from SCADA), SE solves a weighted least-squares problem to find the best estimate of the system state (all bus voltage magnitudes and angles). Bad data detection algorithms identify and remove telemetry errors or cyber attacks.
• Automatic Generation Control (AGC): Continuously adjusts generator setpoints to maintain system frequency at 60 Hz and to control inter-area power flows to their scheduled values. Uses the Area Control Error (ACE):

\[ ACE = \Delta P_{tie} + B\Delta f \]

where \(\Delta P_{tie}\) is the deviation in tie-line flow from schedule and \(B\) is the frequency bias coefficient [MW/0.1 Hz]. AGC acts through the secondary (load frequency) control loop on governor setpoints.

• Economic Dispatch and Optimal Power Flow: Optimal Power Flow (OPF) extends economic dispatch to include the full nonlinear power flow equations and security constraints (line flow limits, voltage limits). The AC-OPF is a nonlinear non-convex optimization problem; DC-OPF (linearized power flow) is a linear program used in real-time markets.
• Contingency Analysis: Automatically evaluates the impact of the loss of each generator or transmission line (N-1 contingency) on line flows and voltages, flagging violations for operator attention.
• Short-term Load Forecasting: Predicts system load 1 hour to 24 hours ahead using weather data, historical patterns, and economic indicators. Essential for unit commitment and scheduling.

9.6 Phasor Measurement Units (PMUs)

Traditional SCADA systems collect measurements asynchronously with scan rates of 2–4 seconds. For dynamic monitoring and wide-area control, this is far too slow. Phasor Measurement Units (PMUs, also called synchrophasors) provide a quantum leap in observability.

The GPS timestamp allows PMU measurements from geographically distant substations to be meaningfully compared (unlike unsynchronized SCADA measurements). This enables:

• Wide-Area Situational Awareness (WASA): Real-time monitoring of power angles across the entire interconnection, detecting stress before it becomes a stability problem.
• Linear State Estimation: Because PMU measurements are direct phasor values (voltage and current magnitude and angle), the state estimation problem becomes linear (no iterative solution needed) if sufficient PMU coverage is available.
• Post-disturbance Analysis: PMU recordings of major events (inter-area oscillations, voltage collapse) provide high-resolution data for root-cause analysis.
• Out-of-Step Protection: PMU-based relays can detect loss of synchronism (pole slipping) across a large area and initiate controlled separation to prevent a cascading blackout.

9.7 Self-Healing Grids and Restoration Paths

The concept of a self-healing grid refers to the ability to anticipate, detect, and respond to disturbances automatically, with minimal operator intervention. Key capabilities:

• Automated fault location, isolation, and service restoration (FLISR): Within seconds of a sustained fault, the system identifies the faulted section using sensors and communication, opens sectionalizing switches to isolate it, and closes normally-open switches to restore the unfaulted sections from alternate paths.
• Predictive analytics: Analysis of equipment sensor data (transformer oil temperature, dissolved gas analysis, cable partial discharge monitoring) to predict failures before they occur and schedule maintenance proactively.
• Adaptive protection: As the network topology changes (DER connecting/disconnecting, microgrid islanding), protection relay settings are updated in real time to reflect the new fault current levels and network topology.

The prerequisite for self-healing is a communication infrastructure with sufficient bandwidth, low latency, high reliability, and cybersecurity. The IEC 61968/61970 Common Information Model (CIM) and IEC 61850 substation communication standard provide the data models and protocols for interoperable smart grid equipment.

---

Chapter 10: Renewable Integration and Future Directions

10.1 Characteristics of Renewable Sources

Renewable energy sources — primarily wind and solar PV — are inherently variable and non-dispatchable. Their integration raises fundamentally different operational challenges than those posed by conventional thermal or hydro generation.

Wind power output follows a roughly cubic relationship with wind speed within the operating range:

\[ P_{wind} = \frac{1}{2}\rho A v^3 C_p(\lambda, \beta) \]

where \(\rho\) is air density [kg/m³], \(A = \pi R^2\) is the rotor swept area [m²], \(v\) is wind speed [m/s], and \(C_p\) is the power coefficient (maximum theoretical value: Betz limit = 16/27 ≈ 0.593). Modern variable-speed turbines use doubly-fed induction generators (DFIG) or full-power converters to extract power at the Betz-optimal tip-speed ratio across a wide wind speed range.

Solar PV output depends on solar irradiance \(G\) [W/m²], cell temperature \(T_c\), and module characteristics:

\[ P_{PV} = P_{rated} \cdot \frac{G}{G_{STC}} \cdot [1 + \alpha(T_c - T_{STC})] \]

where \(G_{STC} = 1000\) W/m² and \(T_{STC} = 25°C\) are standard test conditions and \(\alpha \approx -0.004/°C\) is the temperature coefficient.

10.2 Grid Integration Challenges

The variability and uncertainty of renewable output creates several technical challenges:

• Frequency regulation: Conventional synchronous generators contribute rotational kinetic energy (inertia \(H\)) that resists sudden frequency changes. Inverter-based renewables provide no inherent inertia. As renewables displace synchronous generators, the system’s Rate of Change of Frequency (ROCOF) after a generation loss event increases: \(ROCOF = \Delta P/(2H_{sys})\). Low-inertia systems require faster frequency response (synthetic inertia from inverter controls, or faster governor response).
• Reactive power and voltage control: Large solar farms and wind plants connected through long transmission lines must provide reactive power support. Grid codes increasingly require DER to ride through voltage dips and provide reactive current injection during faults (similar to conventional generators).
• Flexibility requirements: The “duck curve” phenomenon (high solar output at midday depresses net load, then rapid ramp-up in the evening as solar drops and load peaks) requires flexible generation — fast-ramping gas peakers, pumped hydro, or battery storage.

10.3 Battery Energy Storage Systems

Battery Energy Storage Systems (BESS) are increasingly deployed at both utility scale and distributed levels. Their operational characteristics differ fundamentally from generators:

• Reversibility: BESS can charge (absorb power) or discharge (inject power) within milliseconds, making them ideal for fast frequency response.
• State of Charge (SOC): Constrains operation. A battery at 100% SOC cannot absorb more energy; at 0% it cannot inject.
• Efficiency: Round-trip efficiency is typically 85–95% for lithium-ion batteries.
• Degradation: Cycling causes capacity fade; optimal scheduling must account for degradation costs.

In power flow and dispatch models, a BESS is typically modelled as a PQ bus with adjustable \(P_{charge}/P_{discharge}\) bounds constrained by SOC dynamics:

\[ SOC(t+\Delta t) = SOC(t) + \frac{\eta_{ch}P_{ch}\Delta t - P_{dis}\Delta t/\eta_{dis}}{E_{capacity}} \]

---

Chapter 11: Worked Examples and Problem Solutions

11.1 Per-Unit System — Full Three-Bus Example

11.2 Newton-Raphson Power Flow — Two-Bus Example

11.3 Symmetrical Fault — Complete Calculation

11.4 Economic Dispatch — Three-Unit Example

11.5 Equal Area Criterion — Numerical Example

---

Chapter 12: Power System One-Line Diagrams and Modelling

12.1 One-Line Diagram Conventions

A one-line (single-line) diagram is a simplified representation of a three-phase power system in which three-phase components are shown as single symbols, and the three-phase nature of the system is implied. Key symbols:

• Generator: circle with a sine wave or \(G\) label, connected to a bus through a transformer.
• Transformer: two touching circles (for iron-core magnetic coupling).
• Transmission line: a series of symbols depending on the level of detail (just a line segment for load flow, or the \(\pi\) equivalent for detailed analysis).
• Bus: a horizontal or vertical bar to which generators, loads, and lines connect.
• Circuit breaker: a small rectangle or specific symbol on each branch.
• Load: an arrow pointing away from the bus, or an impedance symbol.

One-line diagrams abstract the three-phase symmetry — which is valid as long as the system is balanced. For fault analysis involving asymmetrical faults, the full three-phase (or sequence-network) representation must be used.

12.2 Per-Unit Equivalent Circuit of the Complete System

After converting all impedances to the system base and constructing per-unit models for each component, the power system equivalent circuit is assembled as follows:

1. Each generator is modelled as an internal voltage source \(\mathbf{E}''\) in series with its subtransient reactance \(X''\).
2. Each transformer is modelled as its per-unit leakage reactance (ideal turns ratio disappears).
3. Each transmission line is modelled as its nominal \(\pi\) equivalent: series impedance \(Z_{line}\) and half-shunt admittances at each end.
4. Loads are modelled as constant power (for power flow) or constant impedance (for fault studies, where they are typically neglected because load impedance \(\gg\) fault path impedance).
5. The resulting network is described by \(\mathbf{Y}_{bus}\) or \(\mathbf{Z}_{bus}\).

12.3 Off-Nominal Tap Transformers

When a transformer operates at a tap setting that differs from the nominal ratio used to define the voltage bases, the per-unit model must include an ideal transformer element. If the nominal turns ratio is \(a_{nom}\) but the actual ratio is \(a = a_{nom}(1 + \epsilon)\) where \(\epsilon\) is the fractional tap change:

The off-nominal tap transformer model in per-unit has an admittance \(y_T\) (the leakage admittance) in series with an ideal transformer of turns ratio \(t = 1 + \epsilon \approx 1\) for small taps. In matrix form, this contributes to \(\mathbf{Y}_{bus}\) as:

\[ Y_{ii} \mathrel{+}= \frac{y_T}{t^2}, \quad Y_{jj} \mathrel{+}= y_T, \quad Y_{ij} = Y_{ji} \mathrel{+}= -\frac{y_T}{t} \]

where bus \(i\) is on the tap side and bus \(j\) is on the non-tap side. Note the asymmetry: \(Y_{ii} \neq Y_{jj}\), making \(\mathbf{Y}_{bus}\) non-symmetric for systems with off-nominal tap transformers.

---

Chapter 13: Protection System Coordination

13.1 Overcurrent Relay Coordination

For a radial feeder protected by a cascade of overcurrent relays (at the feeder source, at each sectionalizing point), coordination requires that the upstream relay always waits for the downstream relay to clear the fault first. The coordination time interval (CTI) between adjacent relays is typically 0.2–0.4 s.

The inverse-time characteristic of a relay is described by:

\[ t = \frac{TD \cdot A}{(I/I_{pickup})^B - 1} \]

where \(TD\) is the time-dial setting, \(I_{pickup}\) is the pickup current, and \(A\), \(B\) are curve constants defined by the relay standard (IEEE/ANSI or IEC). For the ANSI Moderately Inverse curve: \(A = 0.0515\), \(B = 0.02\). For the Very Inverse: \(A = 19.61\), \(B = 2.0\).

13.2 Distance Relay Characteristics

Distance relay characteristics are represented in the complex \(R\)–\(X\) impedance plane. The measured impedance is:

\[ Z_{meas} = \frac{\mathbf{V}_{relay}}{\mathbf{I}_{relay}} \]

Under load flow conditions, \(Z_{meas}\) lies far from the origin (large resistance, moderate reactance). Under a fault on the protected line, \(Z_{meas}\) moves toward the origin along the line’s impedance angle (typically 60°–80° from the real axis).

Common characteristic shapes:

• Mho (circular): Passes through the origin; the boundary is a circle in the \(R\)–\(X\) plane. Directional by nature. Zone 1 circle has diameter \(Z_1 = 0.85 Z_{line}\), Zone 2 has diameter \(Z_2 = 1.2 Z_{line}\).
• Quadrilateral: Separate reach settings for reactance (\(X_{reach}\)) and resistance (\(R_{reach}\)). More flexible for lines with high fault resistance (e.g., ground faults through trees).
• Lens: Combination, provides improved security against load encroachment.

13.3 Differential Protection of Transformers

Transformer differential protection compares the currents on both sides of the transformer (referred to a common base using CT ratios and the transformer turns ratio). For a healthy transformer:

\[ \mathbf{I}_{diff} = \mathbf{I}_{primary,referred} - \mathbf{I}_{secondary,referred} \approx 0 \]

(In practice, the magnetizing current creates a small differential current even under no-fault conditions.) A fault inside the transformer produces a large differential current.

Complications unique to transformer differential protection:

1. CT saturation: During external faults with large current, CTs can saturate, producing a false differential current. Stabilizing (restraint) windings add a fraction of the through-current to the pickup threshold: the relay operates only if \(I_{diff} > k \cdot I_{restraint}\) where \(k\) is the slope setting (15–30% typically).

2. Inrush current: When a transformer is energized, the magnetizing core can saturate asymmetrically, drawing an inrush current with a large second-harmonic component (up to 80% of fundamental). Since inrush is only in the primary, it appears as a large differential current. Second-harmonic restraint blocks tripping during inrush.

3. Over-excitation (overfluxing): At low frequency or high voltage, the core saturates, drawing fifth-harmonic-rich magnetizing current. Fifth-harmonic restraint prevents inadvertent tripping.

4. Y–Δ phase shift: The 30° phase shift between the Y and Δ winding currents must be compensated. Modern numerical relays perform this mathematically; older electromechanical schemes used CTs connected in Delta on the Y side and in Wye on the Delta side to compensate.

---

Appendix A: Key Mathematical Tools

A.1 Complex Power and Power Factor

For a voltage phasor \(\mathbf{V} = |V|\angle\theta_v\) and current phasor \(\mathbf{I} = |I|\angle\theta_i\) (using load sign convention, positive into load):

\[ S = P + jQ = \mathbf{V}\mathbf{I}^* = |V||I|\angle(\theta_v - \theta_i) = |V||I|\angle\phi \]

where \(\phi = \theta_v - \theta_i\) is the power factor angle. The power factor is \(\cos\phi\); lagging if \(\phi > 0\) (inductive load); leading if \(\phi < 0\) (capacitive load).

A.2 Thevenin and Norton Equivalents

Any linear network seen from a terminal pair can be replaced by a Thevenin equivalent:

• \(\mathbf{V}_{th}\): open-circuit voltage at the terminals.
• \(Z_{th}\): impedance seen looking in with all independent sources zeroed (voltage sources → short circuit, current sources → open circuit).

For power systems fault analysis, \(Z_{th} = Z_{kk}\), the diagonal element of \(\mathbf{Z}_{bus}\).

A.3 Matrix Inversion Lemma for \(Z_{bus}\) Building

The bus impedance matrix can be built incrementally using the following rules without explicit matrix inversion, starting from \(\mathbf{Z}_{bus}\) for a reduced network and adding one element at a time (the Kron reduction algorithm), or built directly by inverting \(\mathbf{Y}_{bus}\) using LU factorization.

When a branch with impedance \(Z_b\) is added from the reference to a new bus \(k\):

\[ \mathbf{Z}_{bus,new} = \begin{pmatrix}\mathbf{Z}_{bus} & \mathbf{0} \\ \mathbf{0}^T & Z_b\end{pmatrix} \]

When a branch \(Z_b\) is added from existing bus \(p\) to existing bus \(q\) (both already in \(\mathbf{Z}_{bus}\)):

\[ [Z_{bus,new}]_{ij} = [Z_{bus}]_{ij} - \frac{([Z_{bus}]_{ip} - [Z_{bus}]_{iq})([Z_{bus}]_{pj} - [Z_{bus}]_{qj})}{Z_b + Z_{pp} + Z_{qq} - 2Z_{pq}} \]