Estimation of goodness of fit of synchrophasors during transient faults

Synchrophasor measurements are a useful tool, in particular for understanding dynamic or transient phenomena on large electrical systems. However, synchrophasors are only truly valid when the system is time invariant, that being when the frequency and magnitude of the waveforms is cyclical and continuous from one cycle to the next. This paper considers a metric for determining the quality of synchrophasors, Goodness of Fit (GoF). This metric returns a single scalar value which indicates how closely the estimated synchrophasors compare with the sampled values from which they are computed. The GoF metric is applied to empirical data obtained during a transient event on the power system of the island of Ireland. It is found that the metric returns a value indicating that the synchrophasors are of good quality before the transient event commences, and correctly indicates when the synchrophasors are no longer a good match to the sampled values. The authors propose that this technique should be added to PMUs and can be immediately incorporated with existing PMU communications. Additionally, as the GoF metric indicates when the sampled data is not a good sinusoidal fit, GoF can itself be useful for triggering captures of sampled values for post analysis.