Modular Subspace-based System Identification and Damage Detection on Large Structures

The aim of the present study is to propose a procedure for seismic cracking identification of concrete dams using a coupling of the extended finite element method (XFEM) based on cohesive crack segments (XFEM-COH) and continuous wavelet transform (CWT). First, the dam is numerically modeled using the traditional finite element method (FEM). Then, cracking capability is added to the dam structure by applying the XFEM-COH for concrete material. The results of both the methods under the seismic excitation have been compared and identified to damage detection purposes. In spite of predefined damage in some of the structural health monitoring (SHM) techniques, there is an advantage in the XFEM model where the whole dam structure is potentially under damage risk without initial crack, and may not crack at all. Finally, in order to evaluate any change in the system, that is, specification of any probable crack effects and nonlinear behavior, the structural modal parameters and their variation have been investigated using system identification based on the CWT. The results show that the extended finite element–wavelet transform procedure has high ability for the online SHM of concrete dams that by analysis of its results, the history of physical changes, cracking initiation time, and exact damage localization have been performed from comparing the intact (FEM) and damaged (XFEM) modal parameters of the structural response. In addition, any small change in the system is observable while the final crack profile and performance simulation of the dam body under strong seismic excitations have obtained.

Download Full-text

Experimental System Identification of the Vibro-Impact Dynamics Towards Structural Health Monitoring and Damage Detection

Volume 8: 22nd Reliability, Stress Analysis, and Failure Prevention Conference; 25th Conference on Mechanical Vibration and Noise ◽

10.1115/detc2013-13534 ◽

2013 ◽

Author(s):

Heng Chen ◽

Young S. Lee ◽

Mehmet Kurt ◽

D. Michael McFarland ◽

Lawrence A. Bergman ◽

...

Keyword(s):

System Identification ◽

Structural Health Monitoring ◽

Damage Detection ◽

Mode Shape ◽

Health Monitoring ◽

Correlation Spectroscopy ◽

Impact Dynamics ◽

Shape Functions ◽

Modal Assurance Criterion ◽

Structural Health

We perform nonlinear system identification (NSI) on the acceleration signals that were experimentally measured at ten, almost evenly spaced positions along a cantilever beam undergoing vibro-impacts between two rigid stops with clearances. The NSI methodology is based on the correspondence between analytical and empirical slow-flow dynamics, with the first step requiring empirical mode decomposition (EMD) analysis of the measured time series leading to sets of intrinsic modal oscillators (IMOs) governing the vibro-impact dynamics at different time scales. By comparing the spatiotemporal variations of the nonlinear modal interactions (and hence the IMOs), we examine how vibro-impacts influence the low- and high-frequency modes in global and local senses. In applications of the NSI results to structural health monitoring and damage detection (SHM / DD), we calculate typical measures such as the modal assurance criterion (MAC) and the coordinate modal assurance criterion (COMAC) by extracting information about the mode shape functions from the spatiotemporal IMO solutions. Whereas the MAC provides a global aspect of damage occurrence (i.e., which modes are more affected by induced defects), the COMAC can narrow down the damage locations (i.e., where in the structure defects exist that yield low correlation values in specific modes). Finally, we discuss the use of the 2-dimensional correlation spectroscopy technique to SHM / DD, which is frequently used in optical chemistry areas. With the spatiotemporal IMOs the 2-D correlation intensity for the linear beam is proportional to the product of the two mode shape functions at the respective positions; hence any deviations from that may indicate the occurrence and locations of damage in the structure.

Download Full-text