Performance Rate Analysis in Photovoltaic Solar Plants by Machine Learning

Photovoltaic (PV) solar plants require conditions monitoring systems and advanced analytics to enhance performance and reliability. Supervisory control and data acquisition systems are widely applied in PV solar plants to ensure suitable maintenance management. Novel data processing techniques are required for condition monitoring fault detection and diagnosis due to the volume and variety of data. This article presents a new approach for pattern recognition in the performance ratio (PR) associated with faults of the PV solar power plant. The approach developed in this work presents shapelet, K‐nearest neighbors, and support vector machine as classification algorithms for PR, being a novelty in the current state of the literature. It is analyzed as a real case study based on a PV solar plant located in Spain. The validation is studied with cross validation, and, finally, the results are compared using a nonparametric Mann–Whitney test.