Research On Multi-peak MPPT Method Of Photovoltaic Array Based On Shadow Occlusion Type

The BP neural network can solve the incorrect tracking problem of the traditional Maximum Power Point Tracking(MPPT)algorithm,that is,in the case of shadow occlusion,the true maximum power point cannot be tracked because it is trapped in the local extreme point.However,there are still some problems when BP neural network performs multi-peak MPPT,including the problem that neural network requires a lot of training data,the problem that shadow occlusion type is difficult to express as neural network input vector,the problem of tracking inaccuracy due to temperature.In order to solve the above three problems,this paper proposes a multi-peak MPPT method based on the output characteristics of the PV array under the condition of shadow occlusion,which realizes multi-peak maximum power point tracking of the PV array under shadow occlusion.The method mainly includes: reducing the amount of training data required by the neural network by using the basic shadow occlusion type;using the image recognition technology to identify the shadow occlusion type to facilitate the quantization of the shadow type;using the Perturb and Observe method in the later stage of the tracking to eliminate the effect of temperature on tracking accuracyFirst,modeling from the photovoltaic cells output under uniform illumination,and analyzing the output characteristics of the photovoltaic cells under normal conditions.Then,the output characteristics of the photovoltaic array under partial shading conditions are simulated by Simulink software,and the first set of simulated data was obtained.At the same time,the actual photovoltaic array was also built,and the design experiment was carried out to obtain the data of the photovoltaic array under the actual partial shading conditions,and the second set of measured data was obtained.The first set of data is analyzed,and then the second set of data is used to verify the conclusions of the first set of data,finally the concept of basic shadow occlusion type is proposed.Simulation data and measured data are input to the BP neural network for training respectively in the Matlab software.The two training results indicate that BP neural network only training basic shadow occlusion type can accurately predict the voltage of maximum power point.It can be used as the basis for maximum power point tracking,that is,the basic shadow occlusion type can reduce the amount of training data acquired by the BP neural network to track the global maximum power point.Then,the shadow occlusion type is encoded,and the shadow occlusion type is identified by image recognition technology as a dimension of the BP neural network input vector.Mainly use image perspective,grayscale,image cutting,binarization.Finally,the image shadow ratio calculation results are in line with the actual situation,which proves the identifiability of the shadow occlusion type and ensures the integrity of the BP neural network training data input dimension.Finally,in order to eliminate the influence of temperature,the tracking method is used to further track after the tracking is stabilized.The photovoltaic MPPT system was built by Simulink platform to verify the effect of BP neural network after training on multi-peak MPPT.The multipeak MPPT effect of the contrast perturbation method,the fixed voltage method,and the BP neural network combined with the perturbation method in terms of shadow type,illumination intensity,and temperature.The simulation results show that the BP neural network combined with the perturbation method trained by the basic shadow occlusion type can effectively track the multipeak maximum power point.