DeepLSTM

We introduce the Deep LSTM for highly nonlinear system modeling and prediction. Two schemes of the long short-term memory (LSTM) network are proposed for data-driven structural seismic response modeling, including (1) LSTM-f: full sequence to sequence; and (2) LSTM-s: stacked sequence to sequence. The proposed deep learning model is capable of accurately predicting both elastic and inelastic response of building structures in a data-driven fashion as opposed to the classical physics-based nonlinear time history analysis using numerical methods. The performance of the proposed approach is successfully demonstrated through three proof-of-concept studies that include a nonlinear hysteretic system (Bouc-Wen model), a real-world building with field sensing data, and a steel moment resisting frame. Our model is also scalable to model the dynamics of other types of materials and structural systems, offering significant potential in seismic fragility analysis of buildings for reliability assessment.

For more information, please refer to the following:

• Zhang, R., Chen, Z., Chen, S., Zheng, J., Büyüköztürk, O., & Sun, H. (2019). Deep long short-term memory networks for nonlinear structural seismic response prediction. Computers & Structures, 220, 55-68.

Citation

@article{zhang2019deep,  
         title={Deep long short-term memory networks for nonlinear structural seismic response prediction},  
         author={Zhang, Ruiyang and Chen, Zhao and Chen, Su and Zheng, Jingwei and Büyüköztürk, Oral and Sun, Hao},  
         journal={Computers & Structures},  
         volume={220},  
         pages={55--68},  
         year={2019},  
         publisher={Elsevier}  
         }