Cross-bridge displacement monitoring data reconstruction method, device and equipment

By using a cross-bridge displacement monitoring data reconstruction method, and combining graph neural networks with structural dynamics and energy constraints, the problem of long-term missing bridge monitoring data was solved, achieving high-precision data reconstruction and scientific bridge health assessment.

CN122451282AActive Publication Date: 2026-07-24JSTI GRP CO LTD +1
2 Cites 0 Cited by

Patent Information

Application Number
CN202610947849.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-06-29
Publication Date
2026-07-24
Estimated Expiration
2046-06-29

AI Technical Summary

Technical Problem

When existing bridge monitoring data is missing for long periods, traditional methods struggle to handle the strong nonlinear dynamic response caused by vehicle loads. Furthermore, deep learning models are prone to overfitting under underdetermined conditions, leading to distorted reconstructed data and failing to meet the requirements for high-precision structural condition assessment.

Method used

By acquiring monitoring data between bridges of the same origin, and using a graph neural network model combined with structural dynamics, macroscopic energy constraints, and historical deformation patterns, a spatiotemporal mapping mechanism for bridges is established to reconstruct the monitoring data.

Benefits of technology

It achieves high-precision data reconstruction under underdetermined conditions, ensuring that the reconstructed data strictly satisfies the boundary conditions and energy conservation in physical terms, thereby improving the confidence and scientific rigor of bridge health assessment.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention discloses a cross-bridge displacement monitoring data reconstruction method, device and equipment, and belongs to the technical field of bridge structure health monitoring and data processing. The method comprises the following steps: acquiring monitoring data of a first bridge; determining time-varying delay of mapping the monitoring data to a second bridge time axis based on time-space transfer characteristics of homologous excitation; projecting the monitoring data to a modal feature space to obtain a first bridge modal feature; obtaining a graph neural network model based on structural dynamics constraint, macroscopic energy constraint and historical deformation mode constraint update parameters to learn a transverse coupling relationship between the second bridge space measurement points; and using the model to output predicted modal features based on the modal features and the time-varying delay and reconstruct second bridge monitoring data. According to the method, the fusion of a physical mechanism and data driving is realized, the problem of single-bridge data missing is effectively solved, and the physical consistency and confidence of reconstructed data are improved.
Need to check novelty before this filing date? Find Prior Art