A highway geological disaster early warning method and system based on deep learning

By using a deep learning-based approach that combines multi-source monitoring data and physical information neural networks, the abnormal overflow replenishment of the drainage system is quantified and the dynamic pore water pressure field is predicted. This solves the problems of sparse monitoring data and incomplete assessment in existing technologies, and enables more accurate geological disaster early warning.

CN122245026APending Publication Date: 2026-06-19INST OF COMM SCI YUNNAN PROV
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
INST OF COMM SCI YUNNAN PROV
Filing Date
2026-02-07
Publication Date
2026-06-19

AI Technical Summary

Technical Problem

Existing highway slope geological disaster monitoring systems suffer from high sensor network deployment and maintenance costs, sparse data leading to poor prediction reliability, and a lack of multi-dimensional quantification in drainage system assessment, making it impossible to accurately determine the dynamic impact of poor drainage on the slope.

Method used

A deep learning-based approach is adopted to acquire multi-source monitoring data, calculate the functional integrity index of the drainage system, quantify the abnormal overflow supply, iteratively predict the dynamic pore water pressure field, generate early warning signals by combining geological technology instability risk index, and integrate physical information neural network to constrain physical laws.

Benefits of technology

It enables a quantitative reflection of the geological disaster process caused by drainage system failure, improves the authenticity of risk assessment and the accuracy of early warning, and can identify potential functional failure risks earlier and more accurately.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure FT_1
    Figure FT_1
  • Figure FT_2
    Figure FT_2
  • Figure FT_3
    Figure FT_3
Patent Text Reader

Abstract

This invention provides a deep learning-based method and system for early warning of geological disasters on highways, belonging to the field of geological disaster early warning technology. The invention includes: S1. Acquiring models and data; S2. Calculating the functional integrity index of the drainage system, which integrates visual congestion degree and seasonal deposition rate, to quantify abnormal overflow replenishment; S3. Using this replenishment as a boundary condition, iteratively solving it through a Physical Information Neural Network (PINN) to generate a dynamic pore water pressure field that conforms to physical laws; S4. Calculating the instability risk index and generating an early warning based on the area, gradient non-uniformity, and geological parameters of this pressure field. This invention solves the problem of decoupling hydrological and geotechnical models, accurately simulating the disaster chain caused by drainage failure through a coupled prediction mechanism, and providing a reliable early warning scheme. This invention establishes a coupled transmission mechanism from external hydrological data to internal physical response, improving the accuracy of early warning.
Need to check novelty before this filing date? Find Prior Art