Correntropy Based Robust Parameter Identification For Photovoltaic Array Model

With the global energy crisis and the growing environmental problems, solar energy has being widely valued around the word due to the advantage of zero pollution and inexhaustible .As an important part of photovoltaic power generation system,the accuracy of the PV array model is directly related to the utilization of solar energy. Parameter identification can improve the accuracy of the model parameters in the photovoltaic cell model, which can guarantee that the identifiedmodel is consistent with the actual plant. In general, the PV array model uses the weighted least squares parameter identification method. A robust parameter identification method based correntropy is proposed for the problem that the weighted least squares parameter identification method is significantly affected by significant errors. The output characteristics of the PV array under the uniform illumination condition differ greatly from those in the partical shadow condition.Therefore, this thesis analyzes the PV array model under ideal and partical shadow conditions, and studies the robust parameter identification method in two cases. The main work is as follows:(1)The working principle of the photovoltaic cell is analyzed in detail, and the theoretical mathematical model of the photovoltaic cell is constructed. The influence of illumination, the temperature of the panel and the diode quality factor on the output characteristics and the maximum power point of the photovoltaic cell are obtained by MATLAB simulation software. Finally, based on the principle of series-parallel connection of the circuit, the mathematical model of the PV array is established, and the correctness of the model is verified by the experimental data.(2) The robust estimator based on correntropy is studied. The influence of measurement error on the weighted least squares estimator and the correntropy estimator is analyzed by the influence function. It can be concluded that when the measurement data has only random error, the effect of the measurement error on the weighted least squares estimator and the correntropy estimator is almost the same. However,There is a significant error in the measurement data, the correntropy estimator is little affected.(3) Based the correntropy estimator, the problem of robust parameter identification is constructed for photovoltaic cells under ideal conditions.The mathematical model of photovoltaic cell is used as the nonlinear constraint of the parameter identification problem, and The weighted least squares parameter identification method (WLS-PI) and the correntroy parameter identification method (C-PI) are simulated and analyzed. It can be concluded that the C-PI is superior to the WLS-PI method, and the robustness and validity of the C-PI are further verified by the test data.(4)The causes and hazards of the local shadows of the PV array are analyst. The mechanism of the PV array mismatch and the hot spot effect are described in detail. Then, the PV array modeling under the shadow condition is studied and the correctness of the theoretical mode is verified. The C-PI and WLS-PI methods are used in the simulation examples, the results show that the C-PI method is more robust. The robustness and effectiveness of the C-PI method are verified by four sets of experimental data under partical shadow conditionsThis study can provide robust parameter identification technology for precise modeling of photovoltaic arrays. The PV array model obtained by robust parameter identification can be further applied to the simulation of photovoltaic power generation system and the maximum power point tracking control.