“The fault diagnosis of photovoltaic arrays faces complex working conditions challenges. A fault diagnosis method based on improved grey wolf algorithm optimized extreme learning machine is proposed. Researchers establish a 5-dimensional fault feature vector and improve the optimization algorithm, with a classification accuracy of 98.32%, providing an efficient solution for photovoltaic system state monitoring.”