基于遗传算法和非线性规划求解信息交互的光伏阵列模型鲁棒参数辨识方法

对光伏阵列进行建模不仅可以研究温度、光照等因素对V-I特性曲线的影响,还可以用模型代替实际光伏阵列进行各种光伏实验,降低实验成本,节省实验时间;参数辨识可以使光伏阵列模型的参数值设置更精确,使其与实际值相一致;针对基于非线性规划的光伏阵列模型鲁棒参数辨识方法容易陷入局部搜索的问题,提出了遗传算法与非线性规划求解信息交互的鲁棒参数辨识方法;将遗传算法与非线性规划求解信息交互,既可以进行全局搜索,又可以进行局部搜索,以得到问题的全局最优解;通过仿真测试,使用该方法得到的结果均方误差降低了8倍,均方误差量级达到了1.0E-3,表明了该方法在光伏阵列模型参数辨识方面具有较高的精确度。

Robust parameter identification of photovoltaic arrays model based on genetic algorithm and nonlinear programming

The Modeling of photovoltaic (PV) arrays can analyze the influence of temperature, light and other factors on the V-I characteristic curve and use the models of PV arrays for various PV experiments to reduce the experimental cost and save the experimental time. Parameter identification can make the parameter values of the PV arrays model set more accurately and make it consistent with the actual outputs of the PV device. In the view of the fact that the robust parameter identification method for PV arrays model based on nonlinear programming is easy to fall into local search, a robust parameter identification method combining genetic algorithm with nonlinear programming is proposed. By combining genetic algorithm with nonlinear programming and using their solving information for interaction, the global search and local search can be carried out to obtain the global optimal solution. Through simulation, it is proved that this method has high accuracy in parameter identification of photovoltaic arrays model.