基于改进PSO优化的RBF神经网络的特高压线损预测

Prediction of UHV Line Loss Based on RBFNN Optimized by Improved PSO

为了研究特高压线损与众多影响因素之间的关系,提出一种基于改进粒子群优化(particle swarm optimization,PSO)的径向基函数(radial basis function,RBF)神经网络模型,用于特高压线损的预测。首先,通过理论分析确定影响特高压线损的主要因素即电气和气象因素,从而确定特高压线损的特征指标体系;然后对特征指标体系建立RBF神经网络预测模型,采用惯性权重因子改进的PSO算法优化RBF神经网络的训练过程,保证RBF神经网络在训练时有足够全局寻优能力的同时具有较强的局部寻优能力;最后,基于一条特高压线路的历史运行数据进行仿真,并与其他预测方法的结果进行对比,验证本文所提方法的有效性和准确性。

In order to study the relationship between the UHV line loss and many influencing factors,this paper proposes a radial basis function neural network(RBFNN)model based on the improved particle swarm optimization(PSO)to predict UHV line loss.Firstly,it determines the main factors affecting UHV line loss including electrical factors and meteorological factors by means of theoretical analysis,and accordingly determines the characteristic index system of UHV line loss.Secondly,it builds the RBF neural network prediction model for the characteristic index system,and uses the PSO algorithm improved by the inertia weight factor to optimize the training process of the RBF neural network,which can ensure that the RBF neural network has enough global optimization ability and strong local optimization ability at the same time.Finally,it makes a simulation based on the historical operation data of one UHV line,and compares the result obtained with results of other prediction methods to verify the effectiveness and accuracy of the proposed method.