基于准最小二乘法的光伏阵列模型鲁棒参数估计方法

光伏阵列的模型参数估计在光伏发电系统的仿真、输出功率预测、最大功率点跟踪等方面有重要意义。当测量数据中只含随机误差时，以加权最小二乘（WLS）为优化函数的参数估计方法有较好的辩识效果。但是当测量数据中含有显著误差时，WLS参数辩识的效果较差。为解决此问题，本文提出了一种以准加权最小二乘法（QWLS）为优化函数的参数估计方法来减小显著误差的影响，采用了赤池信息量准则(AIC)设计QWLS最优参数，将该方法应用于光伏阵列中构造模型鲁棒参数估计问题。最后将WLS和QWLS分别结合序列二次规划（SQP）算法，进行光伏阵列模型的参数估计仿真与实验测试。测试结果显示QWLS参数估计结果更准确，验证了准最小二乘法的鲁棒性与有效性。

Quasi-least Squares Based Robust Parameter Identification of Photovoltaic Array Model

The parameter identification of photovoltaic array model is of great significance in the simulation of photovoltaic power generation system, the output power prediction and maximum power point tracking of the photovoltaic array. Weighted least squares (WLS), as a commonly used parameter identification optimization function, has a good recognition effect when the measurement data only contains random errors. When there are gross errors present in the measurement data, the effect of parameter identification with WLS is poor. To solve this issue, a quasi-weighted least squares (QWLS) method is proposed in this paper to reduce the influence of gross errors by using the QWLS as the optimization function, Akaike information criterion is used to design the optimal parameter of QWLS, and the method is applied in the model of photovoltaic arrays to construct robust parameter estimation problem. Finally, WLS and QWLS are combined with Sequential Quadratic Programming (SQP) algorithm to carry out the simulation and experimental testing of parameter identification of photovoltaic array model. The results show that the QWLS based robust parameter identification can achieve more accurate estimation, which further verifies the robustness and effectiveness of the quasi-least squares method.