Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Deep foundation pit deformation displacement prediction method based on VMD-FEDform network
According to the deep foundation pit deformation displacement prediction method based on the VMD-FEDform network, the modal decomposition VMD automatically extracts local features of signals, and the problem that a primary function needs to be manually selected in a traditional decomposition method is avoided. Meanwhile, the reconstructed signal is used as an input feature to replace a time-based feature in a traditional prediction model, so that the influence of noise on a prediction result is reduced, and the prediction precision is improved; the FEDformer model has high applicability and accuracy in foundation pit displacement prediction research, the problem that a combined model is complex in structure at the present stage is solved, the operation efficiency and the calculation performance are improved, and the model prediction precision is greatly improved.
本发明公开的基于VMD‑FEDformer网络的深基坑变形位移预测方法，模态分解VMD通过自动提取信号的局部特征，避免了传统分解方法中需要手动选择基函数的问题。同时，以重构信号作为输入特征，替代了传统预测模型中基于时间的特征，从而减少噪声对预测结果的影响，提高了预测精度；利用FEDformer模型在基坑位移预测研究中具有较强的适用性和准确性，克服了组合模型现阶段存在的结构复杂问题，提高了运行效率和计算性能，大大提高了模型预测的精度。
Deep foundation pit deformation displacement prediction method based on VMD-FEDform network
According to the deep foundation pit deformation displacement prediction method based on the VMD-FEDform network, the modal decomposition VMD automatically extracts local features of signals, and the problem that a primary function needs to be manually selected in a traditional decomposition method is avoided. Meanwhile, the reconstructed signal is used as an input feature to replace a time-based feature in a traditional prediction model, so that the influence of noise on a prediction result is reduced, and the prediction precision is improved; the FEDformer model has high applicability and accuracy in foundation pit displacement prediction research, the problem that a combined model is complex in structure at the present stage is solved, the operation efficiency and the calculation performance are improved, and the model prediction precision is greatly improved.
本发明公开的基于VMD‑FEDformer网络的深基坑变形位移预测方法，模态分解VMD通过自动提取信号的局部特征，避免了传统分解方法中需要手动选择基函数的问题。同时，以重构信号作为输入特征，替代了传统预测模型中基于时间的特征，从而减少噪声对预测结果的影响，提高了预测精度；利用FEDformer模型在基坑位移预测研究中具有较强的适用性和准确性，克服了组合模型现阶段存在的结构复杂问题，提高了运行效率和计算性能，大大提高了模型预测的精度。
Deep foundation pit deformation displacement prediction method based on VMD-FEDform network
基于VMD-FEDformer网络的深基坑变形位移预测方法
ZHANG JISONG (Autor:in) / XIANG YANCHUN (Autor:in) / ZHAI KEDONG (Autor:in) / WANG WENPING (Autor:in) / GONG XINGUO (Autor:in) / SONG LIYU (Autor:in) / HUANG XUEFENG (Autor:in) / ZHU JUNCHEN (Autor:in) / XU YONGCHAO (Autor:in)
26.04.2024
Patent
Elektronische Ressource
Chinesisch
Deformation prediction analysis of vertical displacement of deep foundation pit based on LIBSVM
| DOAJ | 2020
The Application of Neural Network to Deep Foundation Pit Retaining Structure Displacement Prediction
| British Library Conference Proceedings | 2008
Method for reducing displacement deformation of foundation pit
| Europäisches Patentamt | 2024