Predictive Duty Cycle of Maximum Power Point Tracking Based on Artificial Neural Network and Bootstrap Method for Hybrid Photovoltaic/ Wind Turbine System Considering Limitation Voltage of Grid

Abstract

In this paper, we propose a new control-based the neural network and bootstrap method to get the predictive duty cycle for the maximum power point of hybrid Photovoltaic (PV) and Wind Turbine generator system (WTG) connected to 380 V grid. The neural network is designed to be controller by learning the data control of multi-input DC/ DC converter. The artificial neural network (ANN) needs many data for training then the ANN can give the predictive duty cycle to multi input DC/ DC converter. To get much data, we can use the bootstrap method to generate data from the real data. From Photovoltaic characteristic, we can get 344 real data after the data are made by bootstrap method we can get 8000 data. The 8000 data of PV can be used for training artificial neural network (ANN) of PV system. From wind turbine characteristic we can get 348 real data after the data are made by bootstrap method we can get 6000 data. The 6000 data of WT can be used for training artificial neural network of WT system. This new control has two responsibilities, are to shift the voltage of PV and WTG to optimum condition and to maintain the stability of grid system. From the simulation results those can be seen that the power of hybrid PV / WTG system using MPPT controller is in maximum power and has constant voltage and constant frequency of grid system.