An Optimized Deep Learning Model Based PV Fault Classification for Reliable Power Generation

Abstract

Solar energy is considered important renewable energy due to its cleanliness and low-cost power generation. In hard-working conditions, various types of faults affect the performance of the Photovoltaic (PV) system. Detection and classification of PV faults are critical to assure a reliable power generation operation. Many fault detection schemes have been proposed based on visual and embedded approaches. However, the image processing-based fault detection technique achieved great attention by offering higher detection accuracy. This work proposes the metaheuristic optimized deep learning model for solar fault detection. The learning model of MobileNetV2 is developed with hyperparameter optimization. The performance of the proposed model is analyzed in terms of accuracy, precision, recall and F-Score rates.