Data-Driven Dynamic Model Identification for Synchronous Generators

The objective of this paper is to find parameters of dynamic synchronous generator models through phasor measurement unit (PMU) data. Two estimation models are assumed. The first is an input/output model with electric power measurement as input and frequency as output. The second model assumes a single machine infinite bus system and the measurements are treated as outputs only. For these two models, various system identification methods are applied to achieve the goal, including autoregression exogenous (ARX)-based least squares estimation (LSE), subspace identification through n4sid, and linear grey-box model identification. Numerical examples are given to show the effectiveness of dynamic parameter estimation.