Research On Automatic Identification Method Of Structural Modal Parameters Based On Joint Spectral Density

With the development of lightweight,large-scale and complex of civil structures,the influence of dynamic load on the safety status of such structures is more and more important,because the accident of dynamic failure happens from time to time,and is often accompanied by casualties or heavy economic losses.In order to effectively protect the safety of structures,the technology of structural health monitoring has been greatly developed and popularized in recent years.By installing sensors and other equipment on the structure to carry out real-time monitoring of the structure condition,health monitoring can realize the real-time assessment of the health of the structure,and then realize the vibration control of the monitoring structure,safety early warning and other work,and finally ensure the safety of the structure.In health monitoring,the study of dynamic characteristics of structures is an important content to ensure the dynamic safety of structures.Among them,the modal parameters are an important index to evaluate the dynamic characteristics,so the real-time identification of structural modal parameters under operating conditions has always been a hot topic in the field of structural health monitoring.But the existing modal parameter identification method in the practical application of overhand difficulty is high,false modal identification difficult,big issues such as man-made influence,and in order to meet the growing demand for structural health monitoring is blossoming,the movement modal of the structure of automatic identification,reduce the overhand research is a difficult subject has important practical value.In this paper,based on the theory of blind source separation,the weighted spectral density blind source separation method,combined with the tensor parallel factor decomposition,established a joint spectral density identification method,and realized the estimation of high-precision mode information.And on the basis of traditional sliding window to identify ideas,by combining frequency domain decomposition method(FDD)theory and the system of order modal confidence criterion(MAC),puts forward an adaptive frequency domain identification division and modal window order method,through the underdetermined system into multiple not owe the stator system,realize the identification of underdetermined system,finally,a high-precision operation mode autoidentification method based on weighted spectral density is proposed,which provides a method support for the automatic identification of structural modal parameters in the running state.The main research work and conclusions of this paper are as follows:(1)In view of the adaptability of blind source separation in modal application,the similarity relationship between the mathematical model of blind source separation and the mathematical model of mode superposition method is analyzed theoretically,and it is proved that the method of blind source analysis is suitable for modal recognition of linear structures based on mode superposition method.On this basis,according to the weighted spectral density blind source analysis method to identify the problem of large precision affected by the weighted function,by adopting different weighted spectral density function to build joint spectral density tensor,and combining the parallel factor analysis and minimum spectral density criteria respectively the corresponding to each order modal of the optimal selection of weighted function,so as to realize the high precision estimation modal information.At the same time,a numerical example of M-C-K consisting of a mass block and a spring shows that the joint spectral density method has the characteristics of higher identification accuracy,lower peak dependence and lower signal length requirement compared with the traditional FDD.(2)According to traditional sliding window method in identifying underdetermined problems on determination of the difficulties and the initial width of sliding window shall not adjust the sexual problems,this paper proposes a model based on FDD parameters and MAC criterion of adaptive identification division and window system set order method,has realized the frequency domain data is divided into many segments have rich not owe the stator modal information system,so as to realize the underdetermined system modal information identification.Numerical examples of mass block and spring are given to verify the feasibility of the proposed method.(3)In this paper,to analyzes the practicability of this method,the acceleration response of a continuous rigid frame bridge in Chongqing is monitored under the field environment excitation,and the modal information is extracted by using the new algorithm proposed in this paper.Finally,the theoretical results are compared with the ANSYS finite element model of the bridge.The results show that the first six modes identified by this method are in good agreement with the theoretical results,and the identification results are less affected by false peaks Finally,the practicability of the proposed method in practical engineering application is verified.