What does this research mean for the field?

The proposed CEEMDAN-SVR-PSO-LSTM hybrid framework provides highly accurate, millimeter-level predictions of retaining-pile displacement during deep-excavation construction, outperforming standard models like BP, EMD-LSTM, and VMD-GRU. Novelty: ClaimNovelty.METHODOLOGICAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

What question did this study set out to answer?

To develop a hybrid machine learning framework combining CEEMDAN, SVR, and PSO-LSTM for predicting nonstationary retaining-pile displacement during staged deep excavation construction.

September 5, 2026BuildingsOpen Access

A CEEMDAN–SVR–PSO-LSTM Hybrid Model for Construction-Induced Displacement Prediction of Metro Deep Excavations

View Full Paper

Ask AI

Bookmark

Authors

YLYanwei Liu TZTianjun Zhang WZWeiqin Zuo

Discussion

Member takes

Overview

Predictive modeling study demonstrates accurate retaining-pile displacement forecasting in metro excavations, indicating enhanced deformation control for deep underground construction.

Key Points

To develop a hybrid machine learning framework combining CEEMDAN, SVR, and PSO-LSTM for predicting nonstationary retaining-pile displacement during staged deep excavation construction.
Decomposed displacement monitoring series into a low-frequency residual trend and oscillatory intrinsic mode functions using CEEMDAN.
Predicted the trend component using support vector regression and fluctuation components using particle swarm optimization-tuned long short-term memory networks.
Validated performance using multi-depth monitoring data across six construction stages from two metro excavation projects.
Achieved average R2 of approximately 0.94, RMSE of 0.33 mm, and MAPE of 2.8% in Case 1, outperforming BP, EMD-LSTM, and VMD-GRU baselines.
Maintained millimeter-level prediction errors across 90 depth-wise monitoring points over six construction stages during external validation in Case 2.
Identified the late construction stages and upper pile segments as critical zones requiring the highest degree of deformation monitoring.

Cite This Study

Liu et al. (2026) studied this question.

synapsesocial.com/papers/6a9bd49d6b95aff0620ec6cf https://doi.org/10.3390/buildings16173510

View Full Paper

Ask AI

Bookmark

September 5, 2026BuildingsOpen Access

A CEEMDAN–SVR–PSO-LSTM Hybrid Model for Construction-Induced Displacement Prediction of Metro Deep Excavations

View Full Paper

Ask AI

Bookmark

Authors

YLYanwei Liu TZTianjun Zhang WZWeiqin Zuo

Discussion

Member takes

Overview

Predictive modeling study demonstrates accurate retaining-pile displacement forecasting in metro excavations, indicating enhanced deformation control for deep underground construction.

Key Points

To develop a hybrid machine learning framework combining CEEMDAN, SVR, and PSO-LSTM for predicting nonstationary retaining-pile displacement during staged deep excavation construction.
Decomposed displacement monitoring series into a low-frequency residual trend and oscillatory intrinsic mode functions using CEEMDAN.
Predicted the trend component using support vector regression and fluctuation components using particle swarm optimization-tuned long short-term memory networks.
Validated performance using multi-depth monitoring data across six construction stages from two metro excavation projects.
Achieved average R2 of approximately 0.94, RMSE of 0.33 mm, and MAPE of 2.8% in Case 1, outperforming BP, EMD-LSTM, and VMD-GRU baselines.
Maintained millimeter-level prediction errors across 90 depth-wise monitoring points over six construction stages during external validation in Case 2.
Identified the late construction stages and upper pile segments as critical zones requiring the highest degree of deformation monitoring.

Cite This Study

Liu et al. (2026) studied this question.

synapsesocial.com/papers/6a9bd49d6b95aff0620ec6cf https://doi.org/10.3390/buildings16173510

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

View Full Paper

Ask AI

Bookmark