After a fault clears, the generators on a system do not always settle straight back to where they were. The rotor angle of a machine swings against the rest of the grid, power transfer swings with it, and for a few seconds the relays watch an impedance that wanders instead of jumping. That wandering is a power swing, and telling it apart from a fault is one of the quiet hard jobs in protection.
The one equation underneath it
In the classical model, a machine connected to a strong grid through a transfer reactance transfers power as:
Pe = Pmax sin δ, where Pmax = EV/X
E is the machine's internal voltage, V the grid voltage, X the transfer reactance between them, and δ the rotor angle. Mechanical input Pm pushes the rotor forward; electrical output Pe pulls it back. When the two are equal, the angle sits still. Disturb one of them and the angle oscillates around that balance point while damping slowly bleeds the swing away.
A worked example you can check
Take E = 1.1 per unit, V = 1.0 per unit and X = 0.6 per unit. Then:
Pmax = 1.1 × 1.0 / 0.6 = 1.8333 per unit
At a rotor angle of 40 degrees, Pe = 1.8333 × sin(40°) = 1.1784 per unit. If the mechanical input is Pm = 0.8 per unit, the machine is putting out more electrical power than it takes in mechanically at that angle, so it decelerates — the initial acceleration is negative. The balance point, where Pe equals Pm, sits at arcsin(0.8 / 1.8333) = 25.87 degrees. A swing that starts near 40 degrees oscillates around that 25.87-degree equilibrium and, with any damping at all, settles there.
Push the starting angle far enough past the crest of the sine curve, though, and there is no coming back: Pe falls as the angle grows, the machine accelerates away, and the swing becomes a pole slip. That boundary is why the same disturbance can be a shrug on a strong interconnection and an out-of-step event on a weak one.
Why the relay cares
A distance relay does not see angles. It sees impedance, and during a swing the measured impedance travels slowly across the impedance plane — slowly, that is, compared with the near-instant jump of a fault. A stable swing can carry that travelling impedance straight through a zone characteristic and out the other side. Entry into the characteristic alone is not proof of instability, and tripping a healthy line mid-swing can make the disturbance worse.
That is the whole reason power swing blocking and out-of-step tripping exist as separate functions: block the distance elements while a swing is passing, and trip deliberately, at a chosen angle, only when the swing is genuinely unstable. The SEL tutorial linked from the calculator page below walks through both schemes properly.
Try the swing yourself
If you want to watch the angle move, the free Power Swing & Rotor Angle Calculator on GridCircuits runs this classical single-machine model in your browser — change the transfer reactance or the mechanical input and see how the equilibrium and the swing change. Its worked example is the one above, and it is honest about its limits: it is an educational model with a fixed network, not a settings tool, and real out-of-step work needs a proper transient-stability study.
Next time a swing study crosses your desk, run the sine curve by hand first. The shape of Pe against δ explains most of what the relay is about to do.
Top comments (0)