Impact of PMU Data Errors on Modal Extraction Using Matrix Pencil Method

Phasor Measurement Units (PMUs) provide information on the dynamic response of power systems that can be used to extract modal information and determine the mode shape of electromagnetic oscillations in real-time. Modal analysis introduces a significant improvement in determining and assessing stability margins and provides insights into the dynamics of the power system for monitoring and preventive control. Performance of modal analysis depends primarily on the availability and quality of the time-domain signals. However, PMU data is often delivered with various data quality issues. This paper examines the impact of different data quality issues on the performance of the Matrix Pencil Method (MPM), which has been used for power system modal analysis. The paper also explores the ability of the MPM to accurately identify dominant system modes in presence of defective data, including noise, outliers, and un-updated data. The results are illustrated through six examples ranging in size from a single signal up to results from a 2000-bus system.