基于PSO-RBF的输电线路覆冰预测研究

Prediction Research on Transmission Line Icing Based on Particle Swarm Optimization-Radial Basis Function

覆冰后的架空输电线路在风载荷的作用下,容易产生导线舞动现象,严重危害输电线路安全。提出一种基于PSO-RBF的神经网络模型对输电线路的覆冰情况进行预测,对微气象参数影响因子进行排序,选取合适的微气象因素作为模型的输入,降低建模输入的维度,并通过粒子群算法对RBF神经网络参数进行优化,与单一的RBF神经网络相比提高了预测精度,能及时了解导线覆冰的趋势并给出预警,有效防止严重覆冰事故的发生。

The icing overhead transmission lines under the action of wind load are easy to generate the galloping phenomenon, which is seriously harmful to the safe operation of transmission lines. This paper proposed a kind of neural network model based on particle swarm optimization-radial basis function（PSO-RBF） to carry out prediction for the icing condition of transmission lines, to sort the micro meteorological parameters, to select the suitable micro meteorological parameters as the input of model and to reduce the dimensions of modeling input. The algorithm of PSO was used to optimize the RBF neural network and compared with the single RBF neural network, its prediction accuracy was improved, which makes the icing trend of transmission lines known in time with warning to effectively prevent serious icing accidents.