A comparison of damage detection methodologies using analytic models and identified models

Sensitivity enhancing control (SEC) uses the intrinsic relationship between feedback control gains and structure dynamics to enhance the sensitivity of vibration-based damage metrics to structural damage. Previous research has shown that the SEC concept can enhance the sensitivity of the lowest few structural modal frequencies to local stiffness damage significantly, making such metrics more attractive in damage detection. In this paper, SEC is explored in the context of damage localization using inverse approaches. Using full state feedback control laws designed from an analytic model, inverse or model updating approaches to damage localization can be developed such that only the lowest few modal frequencies and mode shapes are required to localize stiffness damage. Using modal data from both closed-loop systems and open-loop systems for updating a global stiffness matrix in a finite-element model avoids the need to measure higher order modal properties. However, recognizing that practical means of realizing SEC are necessary, the paper also considers the use of identified models in designing sensitivity enhancing control laws. Issues regarding the use of state estimate feedback in SEC are described, and an example demonstrates that a controller-observer system design from an identified model has similar properties in enhancing modal frequency shifts due to stiffness damage as a full state feedback controller designed from an analytic (finite-element) model.