Preprocessing synchronized phasor measurement data for spectral analysis of electromechanical oscillations in the Nordic Grid

Spectral analysis techniques have successfully been applied for near real-time monitoring of power system small-signal electromechanical oscillations using synchronized phasor data from PMUs. The methods used for this purpose commonly assume that random load variations adopt the distribution function of Gaussian noise. Hence, careful attention has to be paid so that the preprocessing of synchronized phasor measurements is capable of providing data with the characteristics expected for these methods to work properly. This can be viewed as a conditioning step in the data handling process that has an impact on the results from spectral analysis which consume this data. This article aims to revisit the crucial step of preprocessing of PMU data in a tutorial fashion. The goal is to share the authors' experience when dealing with real PMU data originating from the Nordic Grid. This article offers a systematic and detailed methodology developed by the authors which has been successfully used in studies on estimation of electromechanical modes in the Nordic Grid.
QC 20150303