Method and system for predicting ground surface settlement limit of subway tunnel underpassing high-speed railway
By integrating multi-source monitoring data and using regional prediction methods, combined with improved models and dynamic threshold adjustments, the accuracy and timeliness of settlement prediction for subway tunnels passing under high-speed railways have been solved, improving construction safety and operational controllability.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CCCC RAILWAY DESIGN & RES INST CO LTD
- Filing Date
- 2026-04-14
- Publication Date
- 2026-06-16
AI Technical Summary
Existing technologies for subway tunnels passing under high-speed railways lack the accuracy and timeliness of settlement prediction, making it difficult to meet the high safety and smooth operation requirements of high-speed railways. In particular, the risk assessment is not targeted due to incomplete monitoring data, failure to consider regional differences within the section, and fixed thresholds.
By acquiring multi-source monitoring data, collaborative analysis sub-segments are divided based on feature similarity and spatial continuity. An improved LSTM model and an improved Peck formula model are used in parallel for independent prediction and weighted fusion. Combined with the dynamic adjustment of the settlement risk threshold, a settlement risk threshold matching the real-time status is generated to achieve zonal risk early warning.
It improves the accuracy and reliability of settlement prediction, dynamically adjusts risk thresholds to adapt to the real-time status of different sub-sections, provides targeted guidance, reduces construction risks, and ensures track smoothness and operational safety.
Smart Images

Figure CN122220992A_ABST