A method and system for building structure health monitoring based on seismic background noise

Abstract

This invention relates to a method and system for monitoring the structural health of buildings based on seismic background noise. The method first uses a three-component seismograph to collect seismic background noise data for each floor, and then sequentially performs preprocessing, HVSR analysis, and polarization analysis on this data to extract the structural dynamic parameters of each floor. Next, a multi-input multi-output gradient boosting decision tree (MIMO-GBDT) model is constructed and trained. Finally, the feature weights of the input features to the output target in the trained MIMO-GBDT model are extracted, the difference trend of the feature weights is calculated, and the structural health status of the building is identified based on the difference trend. Compared with existing technologies, this invention can achieve continuous, real-time health monitoring of building structures without requiring historical damage data, and has advantages such as non-invasiveness, low cost, and accurate positioning.

Images