Structural Dynamic Model Identification through Neural Network

Due to its attributes, such as parallelism, adaptability, robustness and inherent ability to handle non-linearality, the artificial neural network has shown great promise in the function mapping, pattern recognition, imagine processing, etc. The dynamic function mapping, dynamic pattern recognition and dynamic imagine processing are still challenging topics in neural network applications. A neural network approach for dynamic function mapping is presented in this paper. The network is trained, tested and verified by using the responses recorded in a real apartment building during earthquakes. The results show that the dynamic behaviors of the building can be very well modeled by the trained neural network. The results also explore the great potential of using neural network in dynamic function mapping/structural dynamic model identification.