Structural Nonlinear Damage Detection Method Using AR/ARCH Model

Along with the social modernization process,structural damage identification has become a very active research field.In the field,a lot of methods for damage detection are appearing,such as damaged and undamaged methods,local and global methods.As one of the global damage detection methods,the method that based on structural vibration information is more intelligent,involving many modern technologies,such as sensor technology,data acquisition,data mining,data communication,signal processing,data management and so on,which makes the method become the research hotspot in the structural damage detection field during the recent decades.Among them,the method based on time series analysis has been applied in the structural damage detection and also obtained many research results.The principle of this time series analysis method is to apply a appropriate time series model to model structural vibration information(such as acceleration response),then extract the damage sensitive indicator form the built model,and subsequently use the damage indicator to detect structural damage.As we all known,the damage of practical engineering structures usually possess the different degree nonlinear feature,however,most time series methods are utilized to research structural linear damage,so the research of structural nonlinear damage detection is more close to practical engineering structures.The major works and innovations of this paper are described as below: Firstly,this paper respectively generally summarized the research results of structural damage detection based on relevant linear time series models(such as AR,ARMA and ARX model)and nonlinear time series models(such as GARCH and ARCH model)in recent decades.Then this paper introduced the basic theories and features of linear AR model and nonlinear ARCH model,which are employed to detect structural nonlinear damage,subsequently further presented the principle of nonlinear damage detection using the AR/ARCH model and the proposed nonlinear damage detection indicator: second order variance indicator(SOVI).The data from an experimental three-storey structure and a simulated eight-storey shear building structure model have been used to verify the effectiveness of the algorithm and the indicator: SOVI,the results show that the method and damage indicator have the advantages of accurately positioning nonlinear damage,the ability of reflecting the change of nonlinear damage degree,sensitivity to small damage and anti-environment disturbance ability etc.In addition,in order to further show the superiority of this method,as a contrast,a traditional linear damage detection method or indicator: cepstral metric indicator(CMI)has also been introduced to diagnose the same nonlinear damages.The main innovations of this paper is that applying the economic time series ARCH model and the proposed second order variance indicator(SOVI)to detect structural nonlinear damage,and obtained relevant results.