Identification Of Moving Loads On The Bridge Based On BP Neural Network

Because of the influence of the moving vehicles and over-load transport, the bridges tend to be damaged, even destructed and collapse, and this threatens the safe service of the bridges. So, the identification of the moving force to determine the magnitude and kinds of the moving force is of important theory signification and is valuable in application in health monitoring, daily maintenance, safe assessment and traffic planning and so on.In this paper, the BP neural network is used in the identification of the moving force, based on the dynamic model of the moving force. Firstly, based on the Euler beam theory and the modal superposition method, the vibration equations of the simple beam and continuous beam under the multiple moving forces are established. The four-degree-of-freedom vehicle model is adopted and the road roughness is obtained by the discrete inverse Fourier transform. The exact frequency equation and mode function are adopted in continuous beams. The dynamic response and the dynamic interaction forces of a simple beam and a continuous beam under the moving force are simulated.Secondly, the identification of moving force on bridge based on BP neural network is studied. The strain responses of the bridge act as the input signals and the dynamic interaction forces as the output signals. The storehouse of samples is established according to different conditions (different velocity, different load and different location). Also, the sensibility of the input parameters is analyzed, and the conventional processing and feature extraction on the samples are carried out. Then the structure and training parameters of the net are confirmed. A multi-step identification procedure, which can be used to effectively identify multiple parameters of the moving vehicle load is proposed. The genetic algorithm is used to optimize the net weights.Finally, the simulations of the force identification show that the proposed BP neural network method can identify the moving vehicle load with acceptable accuracies. The force identification of the constant force model and dynamic force model on the simple bridge, continuous bridge and other...