Detection of Grid Voltage Anomalies via Broadband Subspace Decompositon

Due to the increase of sensitive loads on the mains power grid, measurement and monitoring of the power quality (PQ) have become an important factor for both consumers and operators. As is well-known, PQ problems occur in a very short time period with specific characteristics. In transmission or distribution systems, power quality data are collected from monitoring devices such as digital fault recorders, power quality and dynamic system monitors, etc. The recorded data has to be analysed in order to understand system anomalies. These anomalies may be due to sources of broadband noise. In this study, we employ broadband subspace decomposition, using polynomial eigenvalue decomposition, to detect these anomalies. Results demonstrate that this method may be considered as a new and effective tool for measurement and monitoring of PQ problems.