Research On Bridge Health Assessment Based On Multi-Stage Modal Parameter Characteristicsc

Damage can have a significant impact on the safety and reliability of bridges.Therefore,accurately assessing the service condition of bridges through structural performance parameters is crucial in guiding subsequent maintenance and strengthening efforts.However,traditional bridge health assessments mainly rely on section stress as an index,which poses issues with data processing such as confusion and large quantities.This does not provide a quick and effective comprehensive overall picture of the safe use of the bridge.To address this issue,we propose a bridge health assessment method based on multi-stage modal parameter characteristics through structural dynamics theory.Furthermore,we develop an XGBoost-based bridge health assessment model that incorporates machine learning algorithms in order to achieve a rapid and refined assessment of the bridge’s usage status:(1)In this paper,a structural health assessment method based on multi-stage modal parameter characteristics is proposed with self-oscillation frequency and modal vibration pattern as indicators.In this study,the sensitivity analysis of the modal parameters under the damage state of the structure was firstly carried out.By taking the super-stationary beam as the research object,its self-oscillation frequency and modal vibration pattern under different damage states were calculated by using ABAQUS software.The correspondence exists between the degree of structural damage and the modal parameters.When the damage reaches a certain degree,the self-oscillation frequency of the structure will change.Based on this,the structure will have the phenomenon of vibration jumping and vibration localization..Therefore,this paper constructs a structural health assessment method based on multi-stage modal parameter characteristics,which can realize the refinement and grading of structural safety assessment based on self-oscillation frequency and vibration pattern changes.(2)Based on the structural health assessment method with multi-stage modal parameter characteristics,a spatial finite element model of the bridge was established by using ABAQUS finite element software for an inclined leaning arch bridge to analyze the impact of damage on the safety of the structure in different parts.The results show that when the change of self-oscillation frequency is greater than 10%,the damage degree of arch rib is greater than 70% or the damage degree of boom is greater than 80%.When the change of self-oscillation frequency is less than 10%,further refinement of assessment can be achieved based on the damage degree of different parts corresponding to the vibration jump and local vibration pattern.Moreover,according to the characteristics of modal parameters,the importance sequence of damage effects on different parts can be determined.The arch ribs,spreader bars and tied beams damage to the overall vertical stiffness and torsional stiffness of the bridge are reduced more,and the crossbeams have relatively small effects on the damage to the structural dynamic performance.(3)A multi-stage XGBoost-based bridge health assessment model with modal parameter characteristics was developed based on the modal analysis data of inclined arch bridges.The model is capable of determining the corresponding structural health status categories based on input structural modal parameter features.In order to optimize the model parameters and improve the classification accuracy,the particle swarm algorithm is optimally tuned,trained and tested using finite element theory data.By comparing the analysis with the random forest classification model,the XGBoost model is proved to perform better in terms of accuracy,recall,precision and F1 value.Meanwhile,the arithmetic analysis verifies the effectiveness of the evaluated model.With the use of a multi-stage modal parameter analysis method for structural health assessment,as well as the incorporation of structural dynamic test parameters and damage judgment processing as samples,a bridge health assessment model was constructed using the XGBoost algorithm.The model not only provides a more accurate and comprehensive bridge health assessment,but also provides ideas for fully automated damage diagnosis of bridge structures.