Classification of short-term power quality disturbances with wavelet analysis and random forest method

The concept of quality in electrical power system is of increasing importance. Conditions that adversely affect the quality of power are of great economic importance. In recent years, intensive studies have been carried out to detect power quality problems and to develop solution methods. In this study, it is aimed to determine short term power quality problems. Discrete Wavelet Transform is used to determine the characteristics of power quality problems. Energy, skewness and kurtosis values and attributes were obtained by this method. Features obtained from power quality problems were classified using Random Forest method. Power quality problems with and without noise were classified with 99.6% accuracy. Power quality with higher noise level were classified with accuracy of 97.8%. In the studies conducted by using Random Forest method, high accuracy results can be obtained in the classification of noisy power quality problems.