基于级联PINN的盾构隧道施工期衬砌结构纵向变形预测方法
By using a cascaded physical information neural network model, the problem of poor interpretability in predicting the longitudinal deformation of the lining structure during shield tunnel construction was solved. This model achieved high-precision advance prediction, enhanced the credibility and usability of the model, provided real-time decision support, and reduced construction risks.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SHENZHEN UNIV
- Filing Date
- 2026-05-12
- Publication Date
- 2026-07-17
AI Technical Summary
During shield tunnel construction, inaccurate prediction of longitudinal deformation of the lining structure can lead to disasters such as segment misalignment, joint leakage, and structural damage. Furthermore, traditional data-driven neural network methods neglect the redistribution of ground stress caused by shield tunneling, resulting in poor model interpretability and difficulty in verifying prediction results.
The Cascaded Physical Information Neural Network (CPINN) model is adopted. Through the series load prediction model (PINN-q) and deformation prediction model (PINN-w), physical control equations are embedded. The model is trained using shield tunneling parameters and stratum parameters, and outputs equivalent load and tunnel longitudinal deformation, thereby enhancing the physical transparency and interpretability of the model.
It achieves high-precision and interpretable advanced prediction of longitudinal deformation of the lining structure during shield tunnel construction, improves the credibility and usability of the model, provides real-time decision support, dynamically optimizes tunneling parameters, and reduces construction risks.
Smart Images
Figure CN122197165B_ABST