Impedance based damage detection under varying temperature and loading conditions

The impedance based damage detection technique utilizing piezoelectric materials has become a promising and attractive tool for structural health monitoring due to its high sensitivity to small local damage. However, impedance signals are also sensitive to time-varying environmental and operational conditions, and these ambient variations can often cause false-alarms. In this study, a data normalization technique using Kernel principal component analysis (KPCA) is developed to improve damage detectability under varying temperature and external loading conditions and to minimize false-alarms due to these variations. The proposed technique is used to detect bolt loosening within a metal fitting lug, which connects a composite aircraft wing to a fuselage. Model and full-scale tests are performed under realistic temperature and loading variations to validate the proposed technique. The uniqueness of this paper lies in that (1) a data normalization technique tailored for impedance based damage detection has been developed (2) multiple environmental parameters, such as temperature and static/dynamic loading are considered simultaneously for data normalization and (3) the effectiveness of the proposed technique is examined using data collected from a full-scale composite wing specimen with a complex geometry.

► A data normalization technique using Kernel principal component analysis (KPCA) for specifically impedance based damage detection is developed.
► Model and full-scale bolt loosening detection tests are performed under realistic environmental variations.
► The bolt loosening is successfully detected under varying environmental conditions.
► It shows better performance than other convention data normalization and impedance techniques.