Prediction of Deep Foundation Pit Deformation Based on ESMD-FE-AJSO-LSTM Algorithm

WeChat ID: nsbdqk

Long press to recognize the QR code to follow us

Recommended Articles

South-to-North Water Diversion and Water Conservancy Technology (Chinese and English)

2024 02

Prediction of Deep Foundation Pit Deformation Based on ESMD-FE-AJSO-LSTM Algorithm

Editor’s Note

The deformation of deep foundation pits excavated for sluices is influenced not only by objective factors such as engineering conditions in the sluice area, hydrogeological conditions, pit dimensions, support structure types, and excavation stages but also by random environmental factors such as vibration from construction machinery on-site, loads from surrounding public transport, rainfall, and temperature. Therefore, the deformation of deep foundation pits for sluices exhibits significant nonlinear, unstable, and uncertain characteristics.

Volume 2024, Issue 2 of South-to-North Water Diversion and Water Conservancy Technology (Chinese and English) published the article titled “Prediction of Deep Foundation Pit Deformation Based on ESMD-FE-AJSO-LSTM Algorithm.” The deformation of deep foundation pits for sluices has significant nonlinear and unstable characteristics. Based on this, the Extreme Point Symmetric Mode Decomposition (ESMD) algorithm is introduced to perform multimodal decomposition on the prototype monitoring sequence of sluice deep foundation pit deformation. Furthermore, based on the theory of fuzzy entropy (FE), fuzzy multimodal phase space reconstruction is conducted for each decomposed component, effectively identifying the physical characteristics of sluice pit deformation across different time scales. An artificial jellyfish search optimizer (AJSO) optimized Long Short-Term Memory (LSTM) artificial neural network model is constructed, which is trained based on the reconstructed subsequences, and the trained predictive modal components are combined to achieve dynamic prediction and analysis of sluice pit excavation deformation. Taking the deep foundation pit deformation monitoring of the Jiangsu Zhangjiagang City Eleven Xu River Hub Reconstruction Project as an example, the aforementioned method is used to predict and analyze the deformation process during the excavation of this hub project. The results indicate that the prediction method based on the ESMD-FE-AJSO LSTM algorithm can effectively predict the nonlinear characteristics of foundation pit excavation deformation, achieving higher prediction accuracy and stability compared to traditional LSTM, Recurrent Neural Network (RNN), and Support Vector Machine (SVM) algorithms, providing a technical reference for real-time scientific diagnosis and analysis of excavation safety.

Article Information:

Zhang Wei, Deng Binbin, Qiu Jianchun, et al. Prediction of Deep Foundation Pit Deformation Based on ESMD-FE-AJSO-LSTM Algorithm [J]. South-to-North Water Diversion and Water Conservancy Technology (Chinese and English), 2024, 22(2): 378-387, 408.

Author Information

Zhang Wei1,3, Deng Binbin2, Qiu Jianchun3,4, Xia Guochun1,3, Yao Zhaoren1, Liu Zhanwu1, Zhu Xinyu1, Wang Yujin1

(1. Jiangsu Provincial Water Conservancy Construction Engineering Co., Ltd., Yangzhou, Jiangsu 225002; 2. Zhangjiagang City Yangtze River Flood Control Engineering Management Office, Suzhou, Jiangsu 215600; 3. Yangzhou University, College of Water Conservancy Science and Engineering, Yangzhou, Jiangsu 225100; 4. Nanjing Hydraulic Research Institute, National Engineering Research Center for Efficient Utilization of Water Resources and Engineering Safety, Nanjing 210098)

Corresponding Author:

Qiu Jianchun (1989—), male, from Yangzhou, Jiangsu, PhD, mainly engaged in research in the field of hydraulic structures.

Main Text

If you think it’s good, please share it with your friends!

Citation format:

This report is published with the authorization of the authors. For reproduction, please contact the authors for authorization.

Disclaimer: The articles/images shared by this public account are intended to disseminate beneficial information to the public. The content reflects the personal views of the authors and is for reference only. It does not imply the endorsement of the public account’s views or the verification of the content’s authenticity. Other media, networks, or individuals using content from this public account must bear copyright and legal responsibilities.