WAMS measurements pre-processing for detecting low-frequency oscillations in power systems

2nd International Conference on Measurement Instrumentation and Electronics

物理学;无线电电子学

Kovalenko, P.Y.^1

Automated Electrical Systems Department, Ural Federal University Named after the First President of Russia B.N. Yeltsin, 19 Mira st., Yekaterinburg, Russia^1

关键词: Distributed measurements;    Low frequency oscillations;    Measurement system;    Median filtering;    Outliers detection;    Pre-processing;    Sample collection;    Simulated signals;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/870/1/012011/pdf
DOI  :  10.1088/1742-6596/870/1/012011

来源: IOP

【 摘 要 】

Processing the data received from measurement systems implies the situation when one or more registered values stand apart from the sample collection. These values are referred to as "outliers". The processing results may be influenced significantly by the presence of those in the data sample under consideration. In order to ensure the accuracy of low-frequency oscillations detection in power systems the corresponding algorithm has been developed for the outliers detection and elimination. The algorithm is based on the concept of the irregular component of measurement signal. This component comprises measurement errors and is assumed to be Gauss-distributed random. The median filtering is employed to detect the values lying outside the range of the normally distributed measurement error on the basis of a 3σ criterion. The algorithm has been validated involving simulated signals and WAMS data as well.