Calculation of transient cable conductor temperature based on support vector machine optimized by particle swarm optimization algorithm

Abstract

It is difficult to directly measure the transient conductor temperature of operating cable, however, which plays a significant role in reflecting its operation state and controlling its load. So it is essential to calculate and monitor the transient cable conductor temperature. A model of calculating transient cable conductor temperature based on support vector machine (SVM) is proposed in this dissertation. In the model, particle swarm optimization (PSO) algorithm is introduced to improve the calculation accuracy and avoid blindly selecting training parameters, taking the skin temperature and load current as input and conductor temperature as output. The comparison of calculated and measured values of temperature rise experiments shows that PSO-SVM model can calculate cable conductor temperature accurately, and the calculation result of conductor temperature is independent of the cable' physical parameters and the laying condition. Moreover, the accuracy is higher than that of result of BP neural network (BPNN) and thermal circuit model, and the model has a good ability of generalization.