Maximum Power Point Tracking Of Photovoltaic System With Support Vector Regression

Maximum power point tracking(MPPT)is one of the important ways to improve efficiency.The existing MPPT technology cannot either balance the tracking speed and accuracy,or the implementation cost is too high due to the complexity of the calculation.In this paper,Support Vector Regression(SVR)is introduced in the traditional perturb and observe method.Compared with fuzzy control and neural network technology,SVR has shorter learning time,lower data dependence and easier implementation.In this paper,a photovoltaic array model is established and simulated to obtain the training and testing data sets of SVR,which not only overcomes the difficulty of measuring actual data under a variety of weather changes in a short time,but also eliminates the need for irradiance and temperature sensors.In the training process of the model,the parameter optimization process combining grid search and cross-validation is added.Multiple sets of different disciplinary factors and gamma parameters are pre-trained,then a set of parameters with the lowest mean square error will be selected for formal training,which avoids poor model fitting effects caused by determining directly the disciplinary factor c and the kernel function parameter gamma according to experience.SVR model is written into the S-Function predict module in M language.The current maximum power point voltage is predicted by the model and compared with the current operating voltage to predict a reasonable step size that can be adjusted automatically in real time,balancing the tracking speed and tracking accuracy.The simulation results show that the perturb and observe method with SVR has been significantly improved in tracking accuracy and tracking speed.The convergence time is reduced by up to 96.5%,and the tracking accuracy is increased by 2.60 percentage points.