| 英文摘要 |
A neural network-based approach was developed to estimate linear and nonlinear parameters of a building from seismic response data. The proposed approach comprises three steps. In the first step, NARX(Nonlinear AutoRegressive with eXogenous) Neural Networks are used to identify the building. In the second step, HFRFs(Higher-order Frequency Response Functions) between input excitation and structural relative displacements are obtained from the weights of the NARX neural networks using a harmonic probing algorithm. In the third step, linear and nonlinear parameters of the building are estimated using the HFRFs obtained in the second step. Moreover, numerical and experimental examples were presented to demonstrate the feasibility of the proposed approach. |
本站僅提供期刊文獻檢索。
