Voltage-based prediction technique for efficient MPPT method for SPV systems under non-uniform insolation and partial shading conditions

• Adopting K-futures samples of the PV power in the tracking process to avoid trapping in the local maximum power point by sensing in advance the future values of the PV power using the previous values of the PV voltage. • Utilizing normalized ZALMS algorithm (NZALMS) to adjust the coefficients of the prediction system by adaptively tuning the stepsize using the PV voltage as input vector to the algorithm. • Combining the P&O method with the proposed approaches in (1 and 2) to form an MPPT technique demands less iteration to track the MPP. • MATLAB simulations have been carried out to compare the proposed technique’s performance against the P&O algorithm in order to assess the efficacy and robustness of the technique under fast changes in solar radiation and PSCs. • The technique can efficiently track the maximum power point under PSCs and minimize power loss. • Results of different cases demonstrate the superiority of the proposed technique in terms of achieving fast-tracking, accelerated dynamic response, and quick settling to the MPP under PSCs and rapid atmospheric environments. The combination of different modules of photovoltaic (PV) cells produces many peaks under partial shading conditions. Therefore, extracting the global maximum power is crucial to attain better efficiency. However, conventional methods, such as Perturb and Observe (P&O), usually struggle to track the global maximum power point under such circumstances, as they tend to converge to local maxima. Due to this attribute, this paper proposes a robust voltage-based adaptation technique for predicting k-future samples of the photovoltaic power. The motivation of this prediction is to accurately extract the maximum power of the PV arrays under different partial shading conditions by adjusting the stepsize of the well-known Zero-Attracting LMS (ZALMS) algorithm in its normalized form using the priorly estimated PV voltages. The PV system has been developed in MATLAB/Simulink, and various patterns of non-uniform insolation and partial shading conditions were used to investigate the performance of the proposed technique. In comparison with the conventional approach, our technique improves dynamic response to environmental variations and the ability to avoid local maximum power point while still maintaining lower power oscillations. Furthermore, in adverse conditions, the proposed method improves power generation and efficiency more than the conventional P&O method, by 97.812% and 98.923%, respectively, as opposed to 80.236% and 88.686%. The computational complexity of the proposed work is relatively comparable to the traditional P&O algorithm, making it more suitable for real-time implementation in smart energy systems.