基于粒子群优化支持向量机的电缆温度计算

Calculation of Cable Temperature Based on Support Vector Machine Optimized by Particle Swarm Algorithm

导体温度是影响运行电缆使用寿命和材料利用率的最主要因素,也是反映电缆运行状态的参数.由于技术上尚难以实现对运行电缆导体温度的直接测量,因此有必要进行导体温度计算.文中以电流和外皮温度作为模型输入,以导体温度作为模型输出,构建基于支持向量机的电缆暂态导体温度的数学模型;为提高该模型计算的精度,避免盲目选取训练参数,引入粒子群算法对其惩罚因子C和核参数γ进行寻优.仿真与试验对比结果表明:基于粒子群优化的支持向量机模型(PSO-SVM模型)可以用于电缆暂态导体温度计算,且计算误差小于热路模型和BP神经网络;模型具有良好的泛化能力.

Cable conductor temperature is a main factor affectingthe life and material utilization of the cable,and is an important parameter reflecting cable's operation state. However,it is difficult to directly measure the conductor temperature of in-use cables,so that a temperature calculation is necessary. In this paper,a model to calculate the transient temperature of cable conductor based on the support vector machine( SVM) is proposed. In this model,both the load current and the skin temperature are used as the inputs and the conductor temperature is taken as the output. Moreover,in order to improve the calculation accuracy and avoid blind selection of training parameters,the particle swarm optimization( PSO) algorithmis introduced in the model for optimizing the punishment index C and the core parameter γ. In addition,a comparison between the simulated and the experimental results is made,finding that the proposed PSO-SVM model is superior to the thermal circuit model and the BP neural network because it helps to obtain more accurate transient temperature and possesses good generalization ability.