Digital filtering of synchronous generator measurements for parameter estimation

This paper describes data processing of synchronous generator current and voltage measurements, which are used to estimate the machine parameters via Park's transformation. The data preprocessing employs infinite impulse response (IIR) digital filters. The challenge is to filter the measurement data while preserving the phase information within the ac voltage and current. This problem is addressed by converting the phase quantities into a time-invariant frame of reference using Park's transformation. The influence of Park's transformation on higher harmonics is analytically studied here to track the conversion of the harmonics into the transformed domain. Park's transformation is analyzed for positive, negative and zero sequence inputs of different frequencies under balanced and unbalanced conditions. Strategies using filters before and/or after Park's transformation are applied to simulation as well as operational generator data from an electric utility. Results demonstrate that the machine parameter estimation error is reduced from the application of the digital filtering.