基于GA-SVM的高压输电线路弧垂预测模型

High Voltage Transmission Line Sag Prediction Model Based on GA-SVM

导线弧垂是反映输电线路运行状态的重要参数之一,为了预知和预警高压输电线路弧垂的变化,提出了一种基于遗传算法(genetic algorithm, GA)特征自适应赋权的支持向量机(support vector machine, SVM),预测输电线路弧垂的方法(GA-SVM).该方法主要分为两个阶段,首先使用GA对实验数据自适应赋权,以突出重要属性,抑制冗余或次要属性,然后使用SVM预测输电线路弧垂.实验结果表明,该方法在预测输电线路弧垂方面是可行有效的,并且优于贝叶斯(Bayes)算法、K-最近邻算法(KNN)、决策树算法和BPNN神经网络算法.

The wire sag was one of the important parameters of transmission lines. In order to predict the line sag variation, a support vector machine model based on genetic algorithm（GA-SVM） was proposed. The method was divided into two stages. Firstly, GA algorithm was applied to adaptively weight the features to highlight the important attributes and suppress the redundant or secondary attributes. Then, SVM algorithm was used to predict the line sag. The empirical analysis showed that the proposed method was feasible and effective. It was superior to Bayes algorithm, K-Nearest Neighbor algorithm（KNN）, Decision Tree algorithm, and BPNN algorithm.