Boosted Convolutional Neural Networks Based Hybrid Approach for Power Quality Disturbances Analysis

A hybrid approach using machine learning and deep learning was developed in this study to detect and classify power quality disturbances accurately with minimal error. Four different Boosting models of machine learning and deep learning were used to analyze power quality events, and a hybrid model called Boosting CNN SOS was proposed to enhance the accuracy of detection and classification by optimizing the deep learning and boosting algorithms’ parameters. The experimental setup generated real-time power quality events, which were pre-processed for normalization and cleaning. The proposed model was evaluated using various parameters, including RMSE, MSE, d2, MAE, and ME, and compared to other models using K-fold cross-validation. The study results indicated that the proposed model achieved an accuracy of over 99%. This hybrid approach could be applied in smart grids for real time applcations.