A high-pile wharf damage identification method based on xgboost-RF combined algorithm
By using a joint model of wavelet packet decomposition and xgboost-RF, the problems of insufficient reflection of local damage and inaccurate localization in the structural damage identification of high-pile wharves are solved, achieving efficient and accurate damage detection and localization, which is suitable for online health monitoring of high-pile wharves.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- TIANJIN RES INST FOR WATER TRANSPORT ENG M O T
- Filing Date
- 2026-04-28
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies for identifying damage in high-pile wharf structures suffer from insufficient ability to reflect local damage, low accuracy of mode shape testing, and poor accuracy of damage location. In particular, they are difficult to accurately identify when multiple damage locations and degrees are being identified together.
Wavelet packet decomposition is used to extract the component energy of transient acceleration response signals, construct damage degree and location indicators, and use the XGBoost-RF joint model for damage identification. Damage feature indicators are constructed by wavelet packet decomposition, and feature discretization is performed by combining the XGBoost algorithm. The random forest model is used to identify the damage location and degree.
It achieves efficient and full-coverage damage detection, avoids secondary structural damage, improves detection efficiency and accuracy, and can accurately identify damage at multiple locations and to varying degrees, with a positioning accuracy rate of nearly 100%, providing reliable support for online health monitoring of high-pile wharves.
Smart Images
Figure CN122412937A_ABST