Solar-ASP Tool For Early Fault Detection and Remote Diagnosis in Photovoltaic Generators

This work presents results of SOLAR-ASP, a novel early Fault Detection ad Diagnosis (FDD) tool developed by TECNALIA and validated on a MAGTEL-operated photovoltaic (PV) plant. It combines data-driven analysis from SCADA measurements with domain-specific physical knowledge to assess PV asset health. This hybrid approach reduces reliance on labelled historical datasets (a common limitation observed in purely data-centric methods) while maintaining interpretability for PV experts. Field trials at a MAGTEL plant demonstrated its ability to estimate main characterization parameters, called Array State Parameters (ASPs), such as Maximum Power Point (MPP) current and voltage normalized to Standard Test Conditions (STC). SOLAR-ASP shows comparable performance to I-V curve measurements in the field, with errors lower than 3%. while enabling the earlier identification of performance degradation. This detection capability can inform proactive maintenance scheduling strategies, allowing for the reduction of in-field inspection campaigns and unnecessary preventative maintenance. This directly lowers OPEX costs while increasing performance by addressing issues before they increase.