Neural Network Simulating the Tensility of Transmission Line Ice Shedding

Abstract

Aiming at the galloping of power transmission line ice shedding, using the methods of polynomial fitting and BP neural network simulating to approach the tensile curves was presented. Due to the tensile test data of power transmission line are discrete numerical values which not only can be fitted by combination polynomials, but also the nonlinear mapped characteristics of BP neural network can be used to simulate the tensile test data. The results indicated that this method not only accurately simulated the dynamic tensility, but also effectively solve the disruption problems of galloping, furthermore, compare polynomial fitting with neural network simulating that the fitting error reduce five times more, the method of BP neural network simulating can improve both the approach precision and the simulating efficiency.