Case of study: Photovoltaic faults recognition method based on data mining techniques


Data Mining techniques have been applied to data collected from a 222 kWp CdTe (Cadmium Telluride) photovoltaic (PV) generator to predict faults or special conditions that occurs due to shadows, bad weather, soiling, and technical faults. Five types of errors have been distinguished and its impact on the PV system performance has been evaluated. Up to date, this computing approach has needed the simultaneous measurement of environmental attributes that an array of sensors collected. This study presents a model to assess the state of the PV (photovoltaic) generator and an algorithm that classifies its state without measuring ambient conditions. The result of a 222 kWp CdTe PV case study shows how the application of computing learning algorithms can be used to improve the management and performance of the photovoltaic generators and underlines the environmental parameters as clue attributes to find faults during the PV performance. Although the application of this method requires computational effort, the result deals with an easy-implementing decision tree, which can be installed in small device

Journal of Renewable and Sustainable Energy