Research Of MPPT Control Algorithm For Photovoltaic System Based On DSP

As coal,oil,natural gas and other energy sources continue to decrease,solar energy has been gradually developed and utilized as a kind of clean,resource-rich and high-efficiency energy source.As an effective measure to improve power generation efficiency,the maximum powerpoint tracking(MPPT)technology of photovoltaic system has attracted increasing attentions.Due to the nonlinearity of the output of the photovoltaic cell and the existence of multiple local maximum power points in the case of partial shading,the existing algorithms have shortcomings such as low tracking efficiency,attachment oscillation at the maximum power point,inability to track the maximum power point under local shadow,and long tracking time.Aiming at the above defects,an MPPT algorithm based on radial basis function(RBF)neural network is proposed.The algorithm can track the maximum power point according to the light intensity and temperature,and it has faster tracking speed and higher prediction accuracy.The thesis first introduces the working principle and mathematical model of photovoltaic cells,and analyzes the simulation characteristics of monolithic photovoltaic cells and photovoltaic arrays with engineering mathematical models.Finally clarifies the output characteristics of photovoltaic cells.On this basis,the single-peak classical MPPT algorithm constant voltage method,disturbance observation method,increment conductance method,multi-peak MPPT algorithm particle swarm algorithm tracking principle are introduced in detail,and simulation analysis is carried out on Matlab/Simulink platform.In response to the problems of the above algorithms,the MPPT algorithm based on RBF neural network uses illumination intensity and temperature as input neurons.The gradient descent method is used as the learning algorithm to train the RBF neural network.The weight,center and extended constants of the RBF neural network are updated to improve the learning speed and prediction accuracy.The simulation results show that the algorithm can achieve the expected error of 0.0001 after training 76 times,and the absolute value of the predicted voltage error is within 0.01 V.Finally,the experimental circuit is designed and built to verify the effectiveness of the algorithm.The hardware circuit takes Boost as the main control circuit,and the DSP chip is the controller.After the sampling circuit collecting external light intensity and temperature,the maximum power point is tracked by the MPPT algorithm.The experimental results show that the algorithm can track the maximum power point within 1.6s when the external environment changes,and the system voltage amplitude is less than 1V after stabilization.The dynamic performance and stability have good effects.