一种基于机器学习的隧道跨断层变形预测方法
By constructing a fault plane coordinate system and a unified observation operator, and combining a dual-domain neural operator and differentiable projection correction, the problems of observation unification and physical constraints in the prediction of relative displacement of tunnels across faults were solved, achieving high-precision prediction and risk quantification.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- HOHAI UNIV
- Filing Date
- 2025-10-24
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies lack a unified observation operator and coordinate system for predicting relative displacement of tunnels across faults. Insufficient coupling of physical constraints makes it difficult to quantify uncertainties, resulting in inaccurate prediction results and inadequate risk assessment.
A unified operator for fault plane coordinate system and observation is constructed. A dual-domain neural operator is used for directional weighted modeling. Complementary constraints of non-penetration and Coulomb friction are applied. Through implicit correction of differentiable projection and assimilation update of uncertainty weight, multiquantile prediction and out-of-limit probability are output.
It achieves high-precision prediction of relative displacement, unifies observation data, ensures physical consistency and stability, quantifies uncertainty, and provides construction management and risk warning.
Smart Images
Figure CN121365593B_ABST