A novel combinatorial hybrid SFL-PS algorithm based neural network with perturb and observe for the MPPT controller of a hybrid PV-storage system

摘要

In recent years, various control methods have been proposed for maximum power point tracking (MPPT) of photovoltaic (PV) power plants. Different MPPT methods for PV systems in the literature have been evaluated in terms of energy efficiency, energy conversion, dynamic performance and reliability in different environmental conditions. Among the various MPPT methods, the Artificial Neural Network (ANN) MPPT is one of the best methods due to its ability in noise rejection and no need for prior information of physical parameters. For implementing the ANN-based MPPT two input variables including temperature and irradiance and an output variable containing voltage of MPP are taken into account. In this paper, a hybrid shuffled frog leaping and pattern search (HSFL-PS) algorithm is used for optimizing ANN-based MPPT in a grid-tied PV system. The P&O approach is used for the tracking cycle procedure and starts a precise tracking scheme after training the ANN and specification of neuron weights. MATLAB/Simulink is utilized for simulation tests to confirm the performance of the offered MPPT method. The outcomes from simulation tests validate the improved performance of the recommended MPPT in comparison with the conventional methods with a fast response of 011 sec.