Stochastic Subspace Method For Automotive Identification Of Structural Modal Parameters

Real-time online monitoring of the structure puts forward the demand forautomotive identification of structural modal parameters. The current structuralmodal parameter identification methods require manual intervention. Professionalsneed to make decision on the intermediate results to complete the modal parameteridentification. In order to achieve real-time monitoring of structures, we are trying todevelop an automotive structural modal parameters identification method.Stochastic subspace identification method is based on the state spacerepresentation of systems, and the only parameters that need to be determined by thepractitioners is the "system order". In SSI algorithm, QR decomposition and/orSingular Value Decomposition are used, and the iterative operation is avoided, whichensures convergence. The system order is significantly reduced through thecomputation of covariance or matrix orthogonal projection, which allows highefficient computation. Based on these characteristics of SSI, this paper aims todevelop a method which can identify the modal parameters automatically based onSSI algorithm, and identify modal parameters without human involvement.The key for automotive modal parameter identification using SSI is todetermine the system order automatically. Singular value jump method andstabilization diagram method are the two conventional algorithms for system orderdetermination, which require human intervention.. This paper presents a way toautomatically determine the system order, which enable automotive identification ofmodal parameters. The main contents and conclusions of this paper are as follows:(1) The two forms of SSI method for modal parameter identification, i.e.,data-driven and covariance based forms, are derived respectively, and the differencebetween these two forms is identified. With higher computation efficiency,data-driven SSI is a better choice for automotive identification.(2) The principles of singular value jumps method and stabilization diagrammethod are introduced, and their advantages and disadvantages are analyzed. Bothmethods require human intervention, and the results are somewhat subjective. Anautomotive order determination method is proposed. The automotive algorithm istheoretically developed, and its procedure is presented. A comparison is made among the two conventional methods and the automotive approach, and the results of thelater are validated.(3) In order to verify the reliability and applicability of the automotive SSImethod, three examples are used, i.e., a three-span continuous girder bridge model,Binzhou Yellow River Bridge scaled model (cable-stayed bridge), Pingsheng Bridge(real suspension bridge). The outputs of the automotive approach are compared withthe ANSYS finite element results, as well as with the FFT, FDD and HHT resultsrespectively. It shows that the automotive SSI approach consistently performs well inthe identification of frequencies and modal shapes of the three structures.