Failure diagnosis and trend‐based performance losses routines for the detection and classification of incidents in large‐scale photovoltaic systems

Fault detection and classification in photovoltaic (PV) systems through real-time monitoring is a fundamental task that ensures quality of operation significantly improves the performance reliability operating systems. Different statistical comparative approaches have already been proposed literature for fault detection; however, accurate loss incidents based on PV time series remains key challenge. Failure diagnosis trend-based routines were developed this work detecting underperformance accurately identifying different types mechanisms. The focus mainly differentiation failures (e.g., inverter faults) from irreversible degradation) reversible snow soiling) factors analysis. benchmarked using historical data obtained 1.8 MWp power plant. results demonstrated effectiveness mechanisms capability pipeline distinguishing issues anomaly change-point (CP) models. Finally, CP model was used to extract significant changes data, detect soiling cleaning events estimate both degradation rates fielded