Detection and Classification of Power Quality Disturbances Using Deep Learning Algorithms

Ensuring uninterrupted electrical energy transmission and distribution, free from voltage and current disturbances, is imperative in response to the growing emphasis on distributed energy systems and micro-power grids. Power quality disturbances (PQDs) serve as the primary cause of electricity quality deterioration that can pose efficiency and safety concerns. To achieve the desired power quality, it is critical to identify and classify various types of power quality disturbances. This chapter aims to give a thorough overview of power quality disturbances, their types, causes, and consequences. It delves into the detection as well as the classification of PQDs, and to tackle this challenge in real time, a deep learning-based method is suggested. Furthermore, a case study is conducted to classify a power-quality voltage disturbance signal using synthetic data generated based on the IEEE-1159 standard. The PQDs dataset comprises 14 types of signals, both single and multiple. The classification is performed using the convolutional neural network-long short-term memory algorithm, known for its high speed and accuracy. The convolutional neural network-long short-term memory algorithm analyses the dataset, and the model’s performance is assessed utilizing four metrics. Moreover, results are compared with convolutional neural networks, k-nearest neighbors, and decision trees to ensure the effectiveness of the convolutional neural network-long short-term memory algorithm. This algorithm performs effectively in PQD detection and classification with a high accuracy of 99.26%.