Tension Prediction Model of Ice-Coated Transmission Line Based on YULE-WALK Auto-Regressive Model and Support Vector Machine

Abstract

Large tension of ice-coated transmission line will cause line overload and conductor galloping, accidents such as break line and tower collapse will be caused, it bring great threat to safety and stability of power systems. Therefore, there is an important physical meaning for preventing above accidents to in-depth study tension prediction model of ice-coated transmission line.In this paper,we establishes a tension prediction model of ice-coated transmission line based on the Yule-Wake auto-regressive model and support vector machine, the model contains the micrometeorological and tension historical data, etc. Through studying the tension prediction of Gan-Zhen 155# transmission line in Zhaotong area of Yunnan province,it shows the prediction obtained by this model in the next eight hours is in accord with the actual monitoring data pretty well, the absolute maximum error is less than 5.86%, and the maximum absolute mean error is less than 2.74%.So, the feasibility and accuracy of this model is verified.