Photovoltaic industryPhotovoltaic Solar Power Plant isStatistical analysis requiring novel monitoringMonitoring systemsMaintenance management and data analysis techniques to determine the actual state of photovoltaic systems. Current analysis methodologies based on machine learning and artificial intelligent require complex trainings and high computational costs. This paper presents a novel approach to detect abnormal decreases caused by failures through the application of statistical techniques. This methodology selects possible periods where the correlationCorrelation between energyEnergy generation data and solar irradiationSolar irradiation data provides abnormal values to detect possible failures with low computational costs. The study uses Supervisory Control and Data Acquisition data from a working solar photovoltaic plant to test the effectiveness of the approach. The main objective is the identification of faults in periods that are not connected to environmental parameters that may justify the reduction in production. The results demonstrated that the application of statistical tools increases the reliability of fault detection.