In-service health assessment of real structures at the element level with unknown input and limited global responses

Abstract

A novel nondestructive defect detection procedure is proposed for the health assessment of existing structures. The procedure can also be used to evaluate the health of structures just after natural events like strong earthquakes or high winds, or after a man-made event like explosion. The procedure is essentially a time domain system identification technique. The two most important features of the method are that it does not need excitation information to identify a structural system and it can identify defects at the local element level using noise-contaminated limited response information, i.e., responses need not be available at all dynamic degrees of freedoms. Any structure that can be represented by finite elements can be identified by this method. The method identifies defects by tracking the changes in the stiffness of the structural elements in the finite element representation. Once a defective element is identified, it can locate the defect spot more accurately within the defective element. The method is denoted as the Generalized Iterative Least-Square Extended Kalman Filter with Unknown Input (GILS-EKF-UI). Since the Kalman filter-based algorithm requires information on excitation force and initial state vector, the GILS-EKF-UI technique consists of two stages. In the first stage, based on the response information, a sub-structure model is developed that satisfying all the requirements for the generalized ILS-UI method. The unknown excitation force f(t) and all the elements in the substructure are identified. The identified stiffness and damping coefficients provide information on the initial values of the state vectors. In the second stage, the EKF-WGI method is used to identify the whole structure since all the information required to implement the EKF-WGI method is now available from Stage 1. With the help of numerous examples, the GILS-EKF-UI method was verified for relatively small and large structures in the presence of noise in the responses. It can detect relatively small defects. The study also helps to define the minimum response information required to implement the Kalman filter-based algorithm. The method is very accurate, robust, and economical and has a potential to be a non-destructive defect evaluation technique.