ARTIFICIAL NEURAL NETWORK BASED APPROACH FOR CLASSIFICATION OF FAULTS IN PHOTOVOLTAIC ARRAYS

ARTIFICIAL INTELLIGENCE

A photovoltaic array (PV array) is a group of solar panels that are electrically linked together to generate more electricity. PV arrays can be used to generate electricity for the grid as well as to power residences. Since they have a minimal impact on the environment, they are becoming a more and more popular source of renewable energy. The accurate and timely detection of faults like Electric damages, Physical damages, and soiling in photovoltaic (PV) arrays is crucial for ensuring the efficient operation and maintenance of solar energy systems. Here, we propose a methodology for fault classification in PV arrays using Artificial Neural Networks (ANN), Voting classifier and Convolutional Neural Networks (CNN) to classify 6 distinct anomalies commonly found in PV arrays, including Dusty, electrical damage, physical damage, bird drop, snow covered and clean. An ANN architecture and a voting classifier is designed for binary classification, allowing for the detection of faults. A CNN model is built for the efficient classification of the faults. This result contributes to the development of automated fault detection systems in the solar energy industry, facilitating efficient maintenance planning and enhancing the reliability and efficiency of photovoltaic arrays.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.