This paper presents detailed procedure for automatic fault detection and diagnosis of possible faults occurring in a grid-connected photovoltaic (GCPV) plant using statistical methods. The approach has been validated using an experimental data of climate and electrical parameters based on a 1.98 kWp plant installed at the University of Huddersfield, United Kingdom. There are few instances of statistical tools being deployed in the analysis of PV measured data. The main focus of this paper is, therefore, to create a system capable of simulating the theoretical performances of PV systems and to enable statistical analysis of PV measured data. The fault detection algorithm compares the measured and theoretical output power using statistical t-test. In order to determine the location of the fault, the ratio between the measured and theoretical DC power and voltage is monitored. The obtained results indicate that the fault detection algorithm can detect and locate accurately different types of faults. Some of the typical faults are fault in a photovoltaic module, photovoltaic string and faulty maximum power point tracker (MPPT) unit. A virtual instrumentation (VI) LabVIEW software was used in the system development and implementation. This system was used successfully for fault detection on the GCPV plant.
- Department of Engineering and Technology - Senior Lecturer
- School of Computing and Engineering
- Centre for Planning, Autonomy and Representation of Knowledge - Member