Deformation Prediction of Transmission Pole Foundation by Using Improved BP Neural Network

Based on the field survey data of the goaf along the UHV path and by including the main geological and mining factors, the stability of UHV transmission pole foundation via Shanxi goaf have been analyzed in details. Using BP artificial neural network method, the paper set up the prediction model of subsidence deformation of pole foundation above the goaf through experiment and study of the data samples. Levenberg-Marquardt algorithm was applied in order to achieve better results. It is concluded that by using BP neural network model, predicting pole foundation stability of the goaf is convenient, reliable, and more applicable.