Abstract
In this work we present the ongoing PV4.0 project, which aims to develop a cost-effective PV plant monitoring and intervention system. By linking PV plant measurement data to a digital ticketing and asset management system, and additionally including the recently developed Cost-Priority-Number method [1], a tool will be developed that applies the concepts of Industry 4.0 to the PV sector. The developed tool aims to improve O&M by prioritizing maintenance activities, and streamlining and further digitalizing the workflow of technicians. Fault detection algorithms are developed based on machine learning approaches like clustering and outlier detection to automatically detect and classify anomalies in PV system output and consequently create maintenance tickets that specify a further course of action. By applying automatic failure detection, the aim is to decrease the time required for failure detection and intervention, in order to minimize revenue losses. By analysing historical and real-time PV plant data and historical failures in operation, the potential for predictive monitoring will be investigated. The digital tools will furthermore produce valuable statistics on PV plant failure modes within each analysed PV plant, and for PV plants in general.