Dropout and Pruned Neural Networks for Fault Classification in Photovoltaic Arrays

Automatic detection of solar array faults reduces maintenance costs and increases efficiency. In this paper, we address the problem fault detection, localization, classification in utility-scale photovoltaic (PV) arrays using machine learning methods. More specifically, develop a series customized neural networks for faults. We evaluate metrics such as accuracy, confusion matrices, Risk Priority Number (RPN). examine assess use with dropout regularizers. network pruning strategies illustrate trade-off between model accuracy algorithm complexity. Our approach promises to elevate performance robustness PV compares favorably against existing