Font Size: a A A

Structural Damage Identification Based On Extended Kalman Filter

Posted on:2015-10-21Degree:MasterType:Thesis
Country:ChinaCandidate:L DaiFull Text:PDF
GTID:2180330422977378Subject:Mechanics
Abstract/Summary:PDF Full Text Request
In the civil engineering structure health monitoring, damage assessment ofstructures in service is a very important work. At present, based on structuralvibration signal measurement system identification and damage identificationtechnology has become a focus of research both at home and abroad. In recent yearsbased on extended kalman filter (EKF) method in the structure of the systemidentification and damage detection has received the widespread attention in the study.It is a kind of real-time recursion method which can be used for structuralparameter identification for nonlinear system based on Least Mean Square ErrorCriterion. The basic idea is: Based on state space model of signal and noise,using theestimation of the previous value and the observation of the present value to update theestimate of the state variables,calculating the best estimate For the present value.This method can effectively identify the change of structural parameters of theposition and level, so that this method can identify structural damage online.In this paper, the extended kalman filter(EKF) is studied and applied to structuraldamage identification. The theoretical formula in single degree of freedom systemand multiple degrees of freedom system is derived under the free vibration dampingand small damping cases using extended kalman filter(EKF) in order to identify thedamage. Mass-spring model of numerical simulation results prove the validity andapplicability of this method in structural damage identification.Because of some problems like that the decline of identification accuracy in Incomplex structure by using the extended kalman filter(EKF). In this paper, making afurther research and improvement based on the extended kalman filtering(EKF).1,Build the extended kalman filter(EKF) state vector by replacing the initial value ofthe master degrees of freedom physical coordinates into the modal coordinates. Toguarantee the stability of the algorithm by transforming the modal coordinate whichcould reduce the freedom degrees to achieve the effect of reducing dimension;2,Dueto structural damage identification problem in essence belongs to the systemidentification problems in inverse problem,combining the Tikhonov regularizationand extended kalman filter(EKF) to improving the solution of ill-posed inver problems, the extended kalman filter(EKF) with Tikhonov regularization has goodrobust and identification accuracy. In the case of a simply supported beam,theimproved algorithm can effectively restrain the noise of the input signal andaccurately identify the single or multiple damages.
Keywords/Search Tags:Extended kalman filter(EKF), Tikhonov regularization, damageidentification
PDF Full Text Request
Related items