Covariance of dynamic strain responses for structural damage detection. (October 2017)

• Record Type: Journal Article
• Authors: Li, X.Y.; Wang, L.X.; Law, S.S.; Nie, Z.H.
• Is Part Of: Mechanical systems and signal processing. Volume 95(2017)
• Page Start: 90
• Page End: 105
• Publication Date: 2017-10
• Subjects: Damage detection -- Inverse problem -- Strain response -- Covariance -- Long-term health monitoring; Structural dynamics -- Periodicals; Vibration -- Periodicals; Constructions -- Dynamique -- Périodiques; Vibration -- Périodiques; Structural dynamics; Vibration; Periodicals; 621
• DOI: 10.1016/j.ymssp.2017.03.020
• Languages: English
• ISSNs: 0888-3270

Abstract

Highlights: The analytical formula of CoS is derived. The CoS sensitivity matrix is nearly diagonal and negative. The stiffness reduction causes the increase of the element's CoS monotonously. The CoS sensitivity is approximated as a constant. The damage identification equation based on CoS is uncoupled. Abstract: A new approach to address the practical problems with condition evaluation/damage detection of structures is proposed based on the distinct features of a new damage index. The covariance of strain response function (CoS) is a function of modal parameters of the structure. A local stiffness reduction in structure would cause monotonous increase in the CoS. Its sensitivity matrix with respect to local damages of structure is negative and narrow-banded. The damage extent can be estimated with an approximation to the sensitivity matrix to decouple the identification equations. The CoS sensitivity can be calibrated in practice from two previous states of measurements to estimate approximately the damage extent of a structure. A seven-storey plane frame structure is numerically studied to illustrate the features of the CoS index and the proposed method. A steel circular arch in the laboratory is tested. Natural frequencies changed due to damage in the arch and the damage occurrence can be judged. However, the proposed CoS method can identify not only damage happening but also location, even damage extent without need of an analytical model. It is promising for structural condition evaluation of selected components.