Detection of Multiple Power Quality Disturbances using Stockwell Transform and Convolutional Neural Network in Electrical Power System

In modern times, most of sophisticated electrical and electronic equipments must be connected with disturbance free power supply. Therefore, to sustain reliability in electrical power system networks, detection of single and multiple power quality disturbances is essential. For that reason, an automatic detection method is developed using Stockwell transform and convolutional neural network in this paper. The simulated power quality disturbance signals are generated by maintaining international standards. Therefore, disturbances are extracted from the generated power quality disturbances signals using a fundamental frequency signal tracker. Extracted disturbances are transformed into time-frequency grey scale images. Thereafter, those images are applied to the input of customized convolutional neural network architecture for classification purpose. The suggested method is efficiently detected the power quality disturbances from the images with 99.83% accuracy. Therefore, power quality monitoring can be settled using this method.