Investigation Of Nonlinear Damage Detection Of Frame Structure Based On AR Model And Principal Component Analysis

Affected by various working loads and environmental conditions,in-service engineering structures are prone to damage in the form of fatigue,fractures,looseness,and delamination,which in turn threatens the structural operation safety.In order to ensure the safe operation of the structure and evaluate the health status of the structure,structural health monitoring technology emerged and developed rapidly.The structure is vulnerable to structural damage along with nonlinear features in the vibration environment.The Auto Regressive Model(AR Model)coefficients extracted from the structural acceleration response data contain information that represents the structural state.The vibration-based structural damage identification method mainly realizes the damage identification according to the change of the characteristic parameters before and after the structural damage.However,generally,the change of the structural characteristic parameters caused by the change of the environmental factors will often cover the change of the characteristic parameters caused by the structural damage.In view of the above problems,this paper develops a Principal Component Analysis(PCA)based damage identification method and realizes the identification of nonlinear damage of the frame structure.Kernel Principal Component Analysis(KPCA)method is used to discuss the effectiveness of damage identification under different environmental conditions.The specific work is summarized in the following aspects:Firstly,the research background of this paper and the research status of structural health monitoring based on statistical analysis are introduced,and the impact of environmental effects on the structural damage detection is outlined.The methods of damage detection under environmental effects are emphatically presented and the main work of this article is detailed.Secondly,the principle of AR model is introduced and the mechanical significance of AR model coefficients is expounded.Then the damage index based on AR model coefficients is constructed and as a further discussion,a numerical simulation of a 2D cantilever beam through ANSYS is used to validate the effectiveness and limitation of the damage index.Thirdly,the principle of statistical analysis is introduced and the process of PCA and KPCA analysis is deducted.The theory of multivariate charts is introduced and the rationality and the construction of the damage index based on statistical analysis method is discussed.Fourthly,the vibration experiment study of a frame structure is introduced.The nonlinear damage of the structure is simulated by a nonlinear collision model and the AR model using the acceleration data is constructed.The experimental data is compressed and characterized by the method of PCA.And the identification of the frame structure damage is validated via the damage index.The environmental effects are simulated by applying different mass blocks on the third layer of the frame structure.The KPCA method is used to project the standardized AR model coefficients into a high-dimensional feature space.The Square Prediction Error(SPE)statistic is constructed as the damage index and the structural damage detection under different environmental effects is validated and discussed.Finally,the full paper is summarized and future research is prospected.