Structural system tracking identification method based on BP neural network

tracking identification;  Structural system identification;  BP neural networks;  Seismic loads

Aiming at the problems of artificial neural networks in the identification of structural systems, a tracking recognition method based on BP neural networks is proposed. By dividing the actual structural model into a mechanism model and a real-time error model, the former is trained offline based on the conventional BP neural network, while the latter identifies the systematic error in real time through a small BP neural network, so that the improved system recognition network can have the ability of dynamic hierarchical recognition system. Computer simulation analysis shows that this method can effectively reduce the identification error of structural system caused by different external loads, and improve the accuracy and reliability of artificial neural network in system identification.