Transmission line tower failure warning based on FBG strain monitoring and prediction model

Abstract

• Early warning of transmission tower failure by strain on-line monitoring. • Complete fiber Bragg grating strain monitoring system including power supply and signal transmission. • The rolling short-term prediction model accurately predicts the strain changes at the weak location of the transmission tower. • The ARIMA-LSTM combined prediction model has higher accuracy and versatility than the single prediction model. Transmission tower failure will endanger the safe operation of power grid. This paper designed an online monitoring system of tower strain based on fiber Bragg grating strain monitoring, and proposed a medium-short term strain prediction model based on ARIMA and a ultra-short term prediction model based on ARIMA-LSTM combination, which provide a basis for the early warning of tower failure. A fixture that is convenient for sensor installation has been designed. The monitoring system has been installed on an actual 500 kV transmission tower, and the measured strain data of weak components has been obtained. The comparison between the prediction results of strain time series and the measured data show that the proposed prediction model has good accuracy. The prediction effect of the ARIMA-LSTM combined model is better than that of a single ARIMA, SVR and LSTM model. The research results in this paper are of great significance to the condition maintenance and disaster prevention of transmission lines.