Electronics Guide

Steady State: Where the Power Flows

A power flow study, also called a load flow study, computes the voltage magnitude and phase angle at every bus for a given pattern of generation and load, and from those the real and reactive power in every branch and the losses in the network. A load bus, or PQ bus, is specified by the real and reactive power it consumes; a generator bus, or PV bus, by its real power output and the voltage magnitude its excitation holds. One slack bus sets the angle reference and supplies the balance of real power, including losses that are not known until the solution is found. The complex power at a bus is its voltage times the conjugate of a current that depends on every bus voltage, so the equations are nonlinear, and programs solve them iteratively, most often by the Newton–Raphson method.

Real power transfer between two buses shows why the phase angle matters. For buses with voltage magnitudes V₁ and V₂, joined by a series reactance X, with the voltage at bus 1 leading the voltage at bus 2 by an angle δ, the real power sent from bus 1 to bus 2 is

P = V₁ V₂ sin δ / X

The relation assumes a lossless connection in sinusoidal steady state, and in per-unit quantities it gives three-phase power directly. Real power follows the angle difference, while on a predominantly inductive network reactive power follows mainly the difference in voltage magnitude, which is why transmission voltage is controlled chiefly with reactive power. The transfer peaks at V₁ V₂ / X when δ = 90°, an early sign of the stability limits described below.

Planners solve the network for peak and light load, for future years, and for contingencies in which a line, transformer, or generator is out of service. Transmission planning commonly requires the system to stay within limits after the loss of any single element, the N−1 criterion. Distribution engineers run the same calculation on feeders, sometimes as a quasi-static time series across a year, to find how much rooftop solar or vehicle charging a circuit can accept, as Hosting Capacity and Distribution Impacts describes.

Faults: How Much Current Flows

A short-circuit study, or fault study, calculates the current that flows when a low-impedance path forms between phases or from a phase to ground. The three-phase fault is the simplest case because it leaves the network balanced. In a bolted three-phase fault, one with no impedance at the fault itself, the current is the prefault voltage at the fault location divided by the Thévenin impedance of the network seen from that point:

I_f = V_pre / Z_th

The relation assumes a linear network, and worked per phase in volts and ohms it takes the line-to-neutral voltage. In per-unit terms, a bolted three-phase fault at a bus with a Thévenin impedance of 0.1 per unit on a 100 MVA base and a prefault voltage of 1.0 per unit draws 10 per unit, a fault level of 1,000 MVA, or about 4.2 kA at 138 kV.

Most faults are unbalanced, and on overhead lines the single line-to-ground fault is the most common. Unbalanced faults are solved with symmetrical components, the method Charles L. Fortescue published in 1918, which resolves an unbalanced set of three-phase quantities into balanced positive-, negative-, and zero-sequence sets. Each set flows in its own sequence network, and each fault type joins those networks in a particular way at the fault point. With equal positive- and negative-sequence impedances, a line-to-ground fault draws more current than a three-phase fault wherever the zero-sequence impedance at the fault is lower than the positive-sequence impedance, as it often is near solidly grounded transformers and generators.

The first cycles of fault current also contain a decaying DC offset. Its size depends on the point in the voltage wave at which the fault begins, and it decays with the time constant X/(ωR), where ω is the angular frequency, so a higher X/R ratio means a slower decay and a higher first peak. Synchronous machines add a decay of their own in the alternating component as their internal flux changes, which study programs represent with subtransient, transient, and synchronous reactances. Inverter-based resources break this pattern, because their controls set the magnitude and phase angle of their fault current. The 2021 stability paper described below puts the contribution of a fully converter-interfaced resource often between zero, when the converter blocks for a close-in bolted fault, and about 1.5 per unit.

Stability: Whether the System Recovers

A stability study follows the system through time after a disturbance such as a fault, the loss of a generator, or the tripping of a line. The IEEE/CIGRE Joint Task Force on Stability Terms and Definitions, reporting in 2004, defined power system stability as “the ability of an electric power system, for a given initial operating condition, to regain a state of operating equilibrium after being subjected to a physical disturbance, with most system variables bounded so that practically the entire system remains intact.” Its classification separated rotor angle, voltage, and frequency stability. A 2021 paper in IEEE Transactions on Power Systems, based on an IEEE Power & Energy Society report by a task force chaired by Nikos Hatziargyriou, kept that definition and added two classes, converter-driven stability and resonance stability, because converter controls acting within microseconds to milliseconds extend the phenomena of interest down to electromagnetic transients.

Rotor angle stability asks whether synchronous machines stay in step. During a nearby fault a generator’s electrical output falls while its mechanical input does not, so its rotor accelerates. If the fault clears soon enough, the rotor swings back; if it clears after the critical clearing time, the machine loses synchronism. Small-disturbance rotor angle stability, the ability to keep synchronism under small disturbances, usually turns on whether electromechanical oscillations are adequately damped. The inertia constant H, the kinetic energy stored in the rotating mass at synchronous speed divided by the machine’s rated apparent power and expressed in seconds, sets how quickly rotor angles and system frequency move in the first seconds after an imbalance. Voltage stability asks whether bus voltages stay steady after a disturbance or as load and power transfers grow, and it is bounded by transfer capability and reserves of reactive power. Frequency stability asks whether frequency settles to an acceptable value after a severe upset, such as the loss of a large power plant, unbalances generation and load.

One-Line Diagrams

Nearly every study starts from a one-line diagram, also called a single-line diagram. It draws a three-phase network with one line per circuit and shows, with standard symbols, the generators, transformers, buses, lines and cables, circuit breakers, loads, and compensation equipment, together with their ratings. For a balanced system the simplification loses nothing, because the analysis of one phase describes all three, as Three-Phase Circuits and Power explains. Unbalanced studies rely on the transformer winding connections and grounding shown on the same diagram, which determine the zero-sequence network.

The same drawing documents the installed system for operators and gives study software the structure of its data, since each symbol becomes a model and each connection a node or branch. Keeping it current is continuing work, because an outdated conductor size, transformer tap, or relay setting yields precise answers about a system that no longer exists.

Per-Unit Quantities

Power engineers usually express voltages, currents, impedances, and powers as fractions of chosen base values. One base apparent power, S_base, applies to the whole network, often a round figure such as 100 MVA. Each voltage zone has a base voltage V_base, normally its nominal line-to-line voltage, and the base voltages on either side of a transformer stand in the ratio of its rated voltages. With S_base as three-phase power and V_base as line-to-line voltage, the base current and base impedance follow:

I_base = S_base / (√3 V_base)

Z_base = V_base² / S_base

At 138 kV and 100 MVA, the base current is about 418 A and the base impedance is 190.44 Ω, so a line with a series reactance of 19 Ω has a reactance of about 0.10 per unit.

The system earns its place for three reasons. When base voltages follow transformer ratings, the ideal transformer drops out, so a model spanning several voltage levels becomes one connected impedance network. Equipment impedances on their own ratings fall in narrow ranges for each kind of equipment, which makes bad data easy to spot. And with the line-to-neutral base set at the line-to-line base divided by √3, a per-unit voltage is the same whether it is read line to line or line to neutral. Nameplate impedances, given on each device’s own rating, are converted to the system base:

Z_pu,new = Z_pu,old × (S_base,new / S_base,old) × (V_base,old / V_base,new)²

A 50 MVA transformer with an impedance of 10 percent on its own rating, for example, has an impedance of 0.20 per unit on a 100 MVA base at the same voltage.

Component Models

Each piece of equipment enters a study as a small equivalent circuit whose detail depends on the question. A short line needs only its series resistance and reactance, a medium-length line adds its shunt capacitance as a nominal-π circuit with half at each end, and a long line uses the distributed-parameter equations, usually as an equivalent-π circuit, because a lumped circuit loses accuracy as length grows. Transformers appear as a series leakage impedance, sometimes with a shunt magnetizing branch, and an off-nominal tap ratio where the tap setting departs from the base voltage ratio. Their winding connections matter twice: a delta–wye transformer shifts positive-sequence voltages and currents by 30 degrees, and the connections and grounding decide whether zero-sequence current can pass.

Synchronous generators appear as a voltage behind a reactance. A power flow represents them by real power and terminal voltage, a fault study by the subtransient or transient reactance, and a stability study adds the rotor’s mechanical dynamics, the excitation system, and the turbine governor. Loads are aggregated at substations as constant power, constant current, constant impedance, or a mixture, with large induction motors modeled explicitly when their dynamics matter. Inverter-based resources are the least settled part of the model. A plant behaves as its control software dictates, so its generic or vendor-specific model must be validated against tests, and a model that serves a power flow says little about the first cycles of a fault.

Protection and Equipment Ratings

Fault calculations are the foundation of protection engineering. Relay engineers use the maximum and minimum fault currents at each location to choose current transformer ratios, to set overcurrent elements so that the device nearest a fault operates first, and to confirm that every relay detects the faults it must clear. Power System Protection covers those schemes and their coordination. The same results confirm that circuit breakers can interrupt the available current and that buses and cables can withstand its heating and mechanical forces. The available fault current is also the starting point for the incident energy calculations described in Arc Flash Analysis Equipment.

Interconnection of Inverter-Based Resources

A large new generator, storage plant, or load goes through interconnection studies before it connects: power flow studies for thermal and voltage violations, short-circuit studies for breaker duties and protection, and dynamic simulations for ride-through and stability. As inverter-based resources replace synchronous machines, the grid loses the short-circuit current and inertia that rotating plant supplied as a by-product; the 2021 stability paper notes that converter-interfaced generation inherently provides neither. Studies must then check whether protection still detects faults and whether frequency holds after the largest credible loss of generation. Grid-forming control, which establishes voltage and frequency rather than following them, is one response, described in Grid-Forming Inverters.

Reactive compensation draws on the same studies. Where the network impedance seen from a bus is mainly reactive, a small reactive injection ΔQ at a bus of short-circuit capacity S_sc changes the voltage by roughly ΔV / V ≈ ΔQ / S_sc. A 100 Mvar device at a bus with a 2,000 MVA fault level therefore moves the voltage by about 5 percent, which is why fault levels, together with the reactive power the network needs, set the rating of a static VAR compensator.

HVDC and Weak Grids

Transmission-scale converters depend on the strength of the AC network at their terminals, usually measured first by the short-circuit ratio, the fault level at the connection point divided by the rating of the converter or plant. A line-commutated HVDC converter needs a strong network to commutate reliably, while a voltage-source converter tolerates a much weaker one, as HVDC Transmission Systems describes. Inverter plants that synchronize through a phase-locked loop have limits too: the 2021 stability paper reports that their controls can become oscillatory below a short-circuit ratio typically between 1.5 and 2, depending on the vendor and the network. Power system analysis supplies the fault levels, the reactive power balance, and the dynamic studies behind the choice of converter and of supporting equipment such as synchronous condensers or harmonic filters.

Operations and Substation Automation

In a control center, analysis runs continuously. A state estimator combines SCADA measurements, and in some systems synchronized phasor measurements, with the network model to estimate the present operating state. That estimate feeds online power flow and contingency analysis, which shows operators which outages would overload equipment or depress voltages. Substation Automation and IEC 61850 describes how intelligent electronic devices, merging units, and station networks gather those measurements and carry out protection and control with settings that trace back to fault and coordination studies.

Steady-State and Short-Circuit Programs

Hand calculation teaches the methods, but practical studies run on software that stores the network model, solves it, and checks the results against equipment ratings. Siemens describes its PSS/E as transmission planning and analysis software, citing power flow for up to 200,000 buses, contingency and fault analysis, and dynamic simulation. DIgSILENT lists load flow, short-circuit, contingency, protection, harmonic, RMS stability, and electromagnetic-transient functions for PowerFactory. Open-source tools serve research, teaching, and distribution planning. MATPOWER is a package of M-files that solves power flow, continuation power flow, and optimal power flow problems in MATLAB or GNU Octave, and EPRI’s OpenDSS is a distribution system simulator intended to support distributed resource integration and grid modernization. Both are released under BSD licenses.

Every program is only as accurate as its impedances, transformer taps, load data, and generator and inverter models, so utilities validate models against recorded disturbances and field tests.

RMS and Electromagnetic-Transient Simulation

Dynamic simulation takes two broad forms. Root-mean-square (RMS) simulation, also called phasor-domain simulation, represents the network by fundamental-frequency phasors and integrates the slower dynamics of machines, excitation systems, governors, and controls, which makes it efficient enough for studies of whole interconnections. Electromagnetic-transient (EMT) simulation solves the differential equations of the network’s inductances and capacitances directly in the time domain. It reproduces instantaneous waveforms, harmonics, switching events, and fast converter control actions, at the cost of much shorter time steps and far more computation. PSCAD, from Manitoba Hydro International, is an EMT program, and PowerFactory offers both forms.

The phenomenon decides the tool. The 2021 stability paper notes that the phasor approximation properly models the classical stability classes but usually not converter-driven or resonance stability, the possible exception being slow-interaction converter-driven stability, which involves oscillations typically below 10 Hz. Interconnection reviews of inverter-based plants increasingly ask for validated EMT models for that reason. Laboratories also connect real controllers to real-time EMT simulators for hardware-in-the-loop tests, a technique covered in Real-Time Simulation Hardware.