Soil seismic response modeling of KiK-net downhole array sites with CNN and LSTM networks

Li, Lin; Feng, Jiashi; Huang, Duruo; Wang, Gang

doi:10.1016/j.engappai.2023.105990

Cited by 16 publications

(3 citation statements)

References 46 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…52 The J 2 bounding surface plasticity constitutive model proposed by Wang and Sitar 53 is applied to simulate the dynamic behavior of the soil. More details about the FEM model can be found in Li et al 40 . At each station, the shear wave velocity profile is utilized to determine the maximum shear modulus of soil layers.…”

Section: The Fem Model For the Seismic Response Of Soil Columnmentioning

confidence: 99%

“…The J 2 bounding surface plasticity constitutive model proposed by Wang and Sitar 53 is applied to simulate the dynamic behavior of the soil. More details about the FEM model can be found in Li et al 40 …”

Section: Analysis Of the Prediction Performance Of The Proposed Modelmentioning

confidence: 99%

“…Subsequently, the physics constraints were embedded in CNN and LSTM for structural response modeling by the same team, 38,39 outperforming non-physics-guided neural network. Li et al 40 developed a deep learning framework for predicting soil seismic response using the recorded acceleration histories from KiK-net stations. The performance of CNN and LSTM based models is compared with that of the state-of-art FEM model in both the time and frequency domain.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Multi‐input integrative neural network for soil seismic response modeling at KiK‐net downhole array sites

Li,

Jin,

Huang

et al. 2024

Earthq Engng Struct Dyn

Self Cite

View full text Add to dashboard Cite

Prediction of the soil seismic response is of primary importance for geotechnical earthquake engineering. Conventional physics‐based models such as the finite element method (FEM) often face challenges due to inherent model assumptions and uncertainties of model parameters. Furthermore, these physics‐based models require significant computational resources, particularly when simulating seismic responses across numerous soil sites. In this study, a multi‐ input integrative neural network is developed for predicting soil seismic response based on the recorded data from a large number of downhole array sites. Ground motions, seismic event information, and wave velocity structures of the sites are utilized as input data in the proposed neural network, enabling the model to adapt to various site conditions. Comparative assessments against state‐of‐the‐art FEM models demonstrate that the proposed models exhibit superior prediction performance with increased efficiency. Furthermore, the pre‐training technique, a transfer learning method, is employed to predict the seismic response at new stations. By fine‐tuning the pre‐trained model derived from the extensive dataset with limited recorded data from new stations, high‐precision seismic response predictions can be realized, illustrating the adaptability and efficacy of the proposed approach in data‐scarce conditions.

show abstract

Section: The Fem Model For the Seismic Response Of Soil Columnmentioning

confidence: 99%

Section: Analysis Of the Prediction Performance Of The Proposed Modelmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Multi‐input integrative neural network for soil seismic response modeling at KiK‐net downhole array sites

Li,

Jin,

Huang

et al. 2024

Earthq Engng Struct Dyn

Self Cite

View full text Add to dashboard Cite

show abstract

Data‐Physical Fusion Deep Learning for Site Seismic Response Using KiK‐Net Records

Chen,

Hu,

Jiang

et al. 2024

Earthq Engng Struct Dyn

View full text Add to dashboard Cite

In the realm of earthquake engineering, response spectra play a crucial role in characterizing the effects of site dynamic characteristics under seismic activity. Consequently, accurately predicting seismic response spectra is of paramount importance. We have developed a physics‐guided bidirectional long short‐term memory neural network model (Phy‐BiLSTM) that is proficient in predicting site seismic response based on bedrock records. The core principle of the Phy‐BiLSTM is to improve the alignment between the solution space and the ground truth by integrating physics knowledge obtained from the physical model. The model introduced in this study utilized the 5%‐damped response spectra, which were derived from strong ground motion records collected at the KiK‐net downhole array. The results substantiate the performance enhancement of Phy‐BiLSTM in comparison to the data‐driven BiLSTM model. Furthermore, we conduct a comparative analysis of the Phy‐BiLSTM model against traditional methods (EQ, SBSR) as well as other neural network architectures (CNN and LSTM). The result highlights the advantages of Phy‐BiLSTM in accurately predicting the site seismic response.

show abstract

Nonlinear Intelligent Inversion Method and Practice for In-situ Stress in Stratified Rock Masses with Deep Valley

Song,

Jiang,

Chen

et al. 2024

Rock Mech Rock Eng

View full text Add to dashboard Cite

Soil seismic response modeling of KiK-net downhole array sites with CNN and LSTM networks

Cited by 16 publications

References 46 publications

Multi‐input integrative neural network for soil seismic response modeling at KiK‐net downhole array sites

Multi‐input integrative neural network for soil seismic response modeling at KiK‐net downhole array sites

Data‐Physical Fusion Deep Learning for Site Seismic Response Using KiK‐Net Records

Nonlinear Intelligent Inversion Method and Practice for In-situ Stress in Stratified Rock Masses with Deep Valley

Contact Info

Product

Resources

About