A Novel Approach to the Automatic Analysis of Disturbance Data: Technology and Implementation

Abstract

The analysis of faults and disturbances is a fundamental foundation for a secure and reliable electrical power supply. Huge amount of disturbance data from the digital fault recorders also brings the challenge of automatically converting data to knowledge, which frees the complex and time-consuming manual analysis. This paper presents a novel approach to the automatic analysis of disturbance data. First, the disturbance signal is segmented based on abrupt change detection in the signal model parameter. This is followed by feature vector construction for specific segments and pattern matching using those feature vectors. Recorded disturbance signals from the power network in South Africa have been used for practical testing.