PV Array Fault Detection and Location Estimation Using Ensemble Learning Model

Currently, large-scale grid connected solar power plants are being installed in MW range for electricity generation. The PV arrays of these large-scale solar power plants are exposed to different electrical faults. The misdetection of faults can create problems in effective operation of solar power plant and can also lead to fire hazards. The magnitude of fault current depends on the irradiance level as well as number of modules getting affected. Consequently, the change in the magnitude of fault currents especially during high impedance ground faults and low mismatch faults is insignificant. The fixed fuse rating of conventional protection schemes like overcurrent protection devices and ground fault detection interrupters have limitations in detection of such crucial faults. In this paper, an ensemble learning model-based technique is used for detection of faults and faulted strings in PV arrays using string current and array voltage simulated through MATLAB/Simulink. The effectiveness of the proposed method is validated and demonstrated through simulated results.