Cable tension prediction of Hongfeng Lake cable-stayed bridge using BP neural network

Due to the funds problem, it is impossible to have acceleration sensors installed on bridgepsilas every cable. In order to get each cablepsilas tension, a three-layer BP neural network model is presented and the neural network with a number of measured data is trained. Using this neural network model, the cable tension can be predicted without any acceleration sensor installed. Then, the neural network model is certificated by using the measured data. By comparison with measured data and the predicted data, the predicted cable tensions by the neural network are credible.