This paper introduces a new way of system identification of dynamic based on artificial neural networks (ANN) and explains concretely how to predict seismic response of structures by ANN in a practical example. This ...This paper introduces a new way of system identification of dynamic based on artificial neural networks (ANN) and explains concretely how to predict seismic response of structures by ANN in a practical example. This paper identifies the structural model of a shear system by the feed forward network of the BP (back propagation) algorithm. First of all, the BP network described in this paper has been trained by practical seismic waves and the corresponding imitated seismic response. Then the seismic response of structures under other seismic excitation will be predicted by BP network of ANN that had been trained. The new ANN can identify the dynamical character and predict dynamical response of structures exactly. This paper also discusses the effects of network topology and input layer elements on the network learning and prediction, etc.展开更多
文摘This paper introduces a new way of system identification of dynamic based on artificial neural networks (ANN) and explains concretely how to predict seismic response of structures by ANN in a practical example. This paper identifies the structural model of a shear system by the feed forward network of the BP (back propagation) algorithm. First of all, the BP network described in this paper has been trained by practical seismic waves and the corresponding imitated seismic response. Then the seismic response of structures under other seismic excitation will be predicted by BP network of ANN that had been trained. The new ANN can identify the dynamical character and predict dynamical response of structures exactly. This paper also discusses the effects of network topology and input layer elements on the network learning and prediction, etc.
