This study reports the development of an innovative fault detection and diagnosis scheme to monitor the direct current (DC) side of photovoltaic (PV) systems. Towards this end, we propose a statistical approach that exploits the advantages of one-diode model and those of the univariate and multivariate exponentially weighted moving average (EWMA) charts to better detect faults. Specifically, we generate array's residuals of current, voltage and power using measured temperature and irradiance. These residuals capture the difference between the measurements and the predictions MPP for the current, voltage and power from the one-diode model, and use them as fault indicators. Then, we apply the multivariate EWMA (MEWMA) monitoring chart to the residuals to detect faults. However, a MEWMA scheme cannot identify the type of fault. Once a fault is detected in MEWMA chart, the univariate EWMA chart based on current and voltage indicators is used to identify the type of fault (e.g., short-circuit, open-circuit and shading faults). We applied this strategy to real data from the grid-connected PV system installed at the Renewable Energy Development Center, Algeria. Results show the capacity of the proposed strategy to monitors the DC side of PV systems and detects partial shading

This paper presents a model-based anomaly detection method for supervising the direct current (dc) side of photovotaic (PV) systems. Toward this end, a framework combining the benefits of $k$ -nearest neighbors ( $k$ NN) with univariate monitoring approaches has been proposed. Specifically, $k$ NN-based Shewhart and exponentially weighted moving average (EWMA) schemes with parametric and nonparametric thresholds have been introduced to suitably detect faults in PV systems. The choice of $k$ NN method to separate normal and abnormal features is motivated by its capacity to handle nonlinear features and do not make assumptions on the underlying data distribution. In addition, because the EWMA approach is sensitive in detecting small changes. First, a simulation model for the inspected PV array is constructed. Then, residuals generated from this model are employed as the input for $k$ NN-based schemes for anomaly detection. Parametric and nonparametric thresholds using kernel density estimation have been used to detect faults. The effectiveness of the $k$ NN-based procedures is verified using actual measurements from a 9.54-kWp grid-connected system in Algeria. Results proclaim the efficiency of the proposed strategy to supervise the dc side of PV systems.