Structural identification for condition assessment using modal non-destructive test data

This research explores the potential of structural parameter estimation for the purpose of condition assessment of existing structures. Structural parameter estimation is a mathematical approach for updating parameters of an a-priori model of the structure such as a finite element model to reproduce its measured response. It is the purpose of this research to use measured Non-Destructive Test (NDT) data to identify stiffness parameters of a laboratory-built grid-type steel structure representing a bridge deck. The grid was built and tested at the University of Cincinnati Infrastructure Institute (UCII). Measured modal NDT data sets were reported for use in parameter estimation. Prior to using the modal NDT data, several techniques are developed and a number of technical difficulties associated with parameter estimation are investigated.; A new error function, modal flexibility-based error function, is proposed for model updating. This error function represents the residual modal displacements measured at subset of degrees of freedom of the FEM. A study on the significance of modeling error in parameter estimation is also presented. It is concluded that modeling error locations and magnitudes, measurement locations, and type of error function most influence the degree of contamination in the parameter estimates.; Objective functions used in parameter estimation are usually highly nonlinear and require a robust algorithm to find their global minimum without converging to a local minima or diverging. This study examines the potential of Genetic Algorithms (GA) to find the global minimum associated with modal objective functions. It is concluded that GA's performance in locating the basin of the global minimum is superior to conventional optimization methods.; The proposed error function and GA are incorporated into Parameter Identification System (PARIS©), a Matlab®-based computer program. Utilizing PARIS, measured NDT data sets are used successfully to identify the stiffness parameters of the grid specimen. The identified parameters include bending and torsional moments of inertia of the grid components as well as vertical stiffness of its supporting neoprene pads.