Priority queue-based switching matrix algorithm for adaptive neuro-fuzzy inference system assisted MPPT controlled PV system

Abstract

The photovoltaic array output is substantially mitigated by regularly occurring, inevitable shadowing conditions. Subsequently, the array's characteristics exhibit several peaks, which causes the traditional maximum power point tracking (MPPT) controllers to inevitably get trapped at the local optimum. Therefore, an adaptive neuro-fuzzy inference system (ANFIS) approach has been proposed for predicting the optimal duty ratio to track the global maximum power among numerous peaks. To dispense the shading impact for improving the GMP and minimization of multiple peaks, a novel priority queue-based reconfiguration algorithm is proposed. The efficacy of the proposed algorithm has been validated for 46 distinct uniform, non-uniform, and dynamic shading conditions for symmetrical 10 × 10, 9 × 9 and unsymmetrical 6 × 4, 6 × 20, 6 × 21 arrays and its performance is compared with the 20 existing algorithms. Further, the efficacy of the combined ANFIS-MPPT with the proposed algorithm has also been validated. The performance of the proposed ANFIS has been compared to the conventional perturb and observe algorithm with and without reconfiguration. Additionally, the ease of global power tracking using a conventional MPPT due to the alleviation of peaks after reconfiguring the array has been presented in detail. Upon reconfiguration, the output is improved by 47.39%, 31.41%, 31.08%, and 9.48% employing the MPPT controller for the various cases. The combined reconfiguration and ANFIS-based methodology effectively tracks the global power time within 0.06 sec with minimal steady-state oscillations yielding a higher tracking efficiency of 99.49%. The proposed methodology is further validated experimentally using a laboratory prototype model.