An Overall Improved Ant Colony Optimization algorithm trained BPNN for PV MPPT

Abstract

A novel PV MPPT algorithm based on the overall improved ant colony optimization algorithm-trained BP neural network (OIACO-BPNN) has been proposed in this paper to overcome the poor prediction accuracy and slow convergence rate of the BP Neural Network (BPNN). Firstly, the pheromone updating model of the Ant Colony Optimization (ACO) algorithm is improved, and the weight coefficient is added to improve the convergence rate of the ACO algorithm. Secondly, the optimal weight threshold of BPNN is updated by Overall Improved Ant Colony Optimization (OIACO) algorithm. Thirdly, the optimized BPNN is employed to predict the Maximum Power Point (MPP) voltage of the photovoltaic (PV) array. Finally, the deviation value between the voltage of the PV array and the predicted voltage is employed as the input of PID controller. In addition, the duty cycle of the Boost circuit is adjusted by PID controller to achieve MPPT. Matlab/Simulink is employed to verify the feasibility and effectiveness of the proposed MPPT algorithm. Simulation results illustrate that the OIACO-BPNN algorithm is superior to the ACO and the BPNN in prediction accuracy and tracking performance, moreover has a good robustness and response speed.