基于GA-SVR模型的配电网线路参数辨识

Line Parameter Identification of Distribution Network Based on GA-SVR Model

准确的线路参数对于配电网运行与控制具有重要意义,而配电网线路参数因受到环境、工况与温度等因素的影响而改变,同时,由于配电网结构复杂度、随机性与波动性日益加强,难以对配电网建立精确的参数辨识模型。提出一种基于GA-SVR模型的配电网线路参数辨识方法,实现配电网线路参数的准确辨识。通过遗传算法(GA)对支持向量回归(SVR)机的惩罚因子与核函数参数进行优化,解决了传统支持向量回归机采用默认参数导致模型预测效果不佳的问题。采用不同阻抗参数下的配电网潮流值对回归网络进行训练,构建改进型SVR参数辨识模型,实现对配电网线路参数的辨识。通过33节点配电网算例的验证,表明该改进型SVR参数辨识模型与传统SVR参数辨识模型相比,能够实现更高精度的配电网线路参数辨识。

Accurate line parameters are of great significance to the operation and control of the distribution network. However,the line parameters of distribution network are affected by environment,working condition and temperature,etc. At the same time,due to the increasing complexity,randomness and volatility of distribution network structure,it is difficult to establish an accurate parameter identification model for the distribution network.A method for identifying the line parameters of the distribution network based on the GA-SVR model was proposed to realize the accurate identification of the line parameters of the distribution network. The genetic algorithm(GA)was used to optimize the penalty factor and kernel function parameters of the support vector regression(SVR)machine,the problem that the default parameters of traditional support vector regression lead to poor model prediction effect was solved. The regression network was trained by using power flow values of distribution network under different impedance parameters,and an improved SVR parameter identification model was constructed to realize the identification of the line parameters of the distribution network. The verification of a 33-node distribution network example shows that the improved SVR parameter identification model can achieve higher-precision distribution network line parameter identification compared with the traditional SVR parameter identification model.