Newfound Intelligent Solution for Grid Connected PV Systems Diagnosis Based on CANFIS Algorithm
Main Article Content
Abstract
The present article proposes a novel hybrid intelligent diagnosis solution for grid connected PV installations based on Co Active Neuro Inference System (CANFIS) algorithm. The considered faults are open circuit fault, short circuit fault, ground fault and by-pass diodes fault in PV array. This solution has been tested and validated on a 9.54 kWp grid connected PV installation. The performances of the proposed method has been tested by residual criteria citing: Mean square error (MSE), Root mean square error (RMSE), Mean absolute percentage error (MAPE), Mean absolute deviation (MAD) and coefficient of correlation (R2) which display 4.65 % and 0.99. The isolation process has been performed using the percentage linear scatter plot of both electrical data (Impp, Vmpp) in the goal to obtain the global diagnosis of PV system. The main novelty of this work is the fact that the proposed diagnosis solution takes into account a particular class of faults which has been until now discarded because of the difficulty of its isolation ; these are faults linked to the bypass diode.