MS18.1 - System Identification and Damage Detection

A direct scheme is presented for output feedback assignment and uncertainty quantification of perturbation-invariant eigenvalues in controllable and observable systems. The feedback is realized through offline signal processing of open-loop input-output data, and the eigenvalue invariance is imposed in a deterministic eigenstructure assignment formulation with left eigenvectors. The uncertainties of the assigned eigenvalues are, under the assumption that subspace identification is used to infer the mathematical model of the given system, quantified using the statistical delta method. The paper outlines and validates the proposed scheme and subsequently presents an application of it for damage localization in mechanical systems that comply with the controllability and observability assumptions.

Damage to bridges is related to road safety and traffic disruption, therefore to substantial direct and indirect economic losses. Hence, there is an urgent need for development and application of systems that can assess the structural condition of bridges. In this study, a methodology for damage identification utilizing measured vibrational data of the initial and the damaged response of a monitored R/C bridge system is proposed. In this context, a finite element model is developed, and calibration of its parameters is performed by applying optimization algorithms to match the numerical to the measured data . The same procedure is applied for the estimation of the model parameters of the damaged structure, in order to identify, localize and quantify damage. Capacity curves are developed for both the calibrated and the damaged bridge and a dimensionless damage indicator is introduced to assess the effect of the identified damage on the global response of the bridge. This approach is applied to a case study monolithic R/C bridge and simulated experiments (e.g. damage scenarios) are conducted to demonstrate the proposed methodology.

Vibration-based Structural Health Monitoring (SHM), exploits a variety of approaches for novelty detection. In particular, some data-based methods try to recognise patterns by exploiting analogies with the human body’s natural defences at a cellular level. These algorithms often require the use of a large variety of data correlated to different environmental and operation conditions to ensure performance robustness and to aid interpretation of results. Regardless, the scarcity of such data often limits the extent of their applicability. In the framework presented here, a possible solution is provided by a novel approach based on modelling and sharing already acquired knowledge between sufficiently similar structures, i.e., population-based structural health monitoring (PBSHM). This study investigates the process of damage detection in a group of three different structures, obtained by applying structural modifications to a small-scale glider model, which follows the characteristic geometry of the GARTEUR benchmark project. Damage identification is performed by exploiting the Negative Selection Algorithm, (NSA), already applied by some of the Authors on numerically simulated case studies, and chosen for its capability of self/non-self discrimination under varying operational or environmental conditions. The research is expanded by using sparse autoencoders for feature dimensionality reduction. The method is applied to three experimental datasets acquired by laser vibrometer measurements, to identify consistent damage-sensitive features from the frequency response functions, and to obtain a reliable fault-detection performance.

Vibration based damage detection is a particularly challenging problem for reinforced concrete beams and frames for which evolution of natural frequencies with damage often do not properly reflect local distributions of stiffness losses, even when they are large. With the recent development of MEMS rotation rate sensors, it is now possible to measure angle variations along the bar axis during vibrations of the structures. This way, for flexural elements, it is possible to directly obtain translational natural modes of the structure and their spatial derivatives (so called rotational natural modes). Having a rotational natural mode, one can obtain more easily a curvature natural mode which is directly related to flexural stiffness of the element. Early numerical simulations and experiments demonstrated many potential advantages for these new, angular measurements of axes of beams and rods of frames. The purpose of the planned presentation during the EURODYN2022 Conference is to report results of the experiment carried out on 6 m reinforced concrete (r/c) and ultrahigh performance concrete (UHPC) beams. The beams have been damaged using static actuator in a number of stages: from the moment when only 1-2 cracks where barely visible to the yielding of reinforcing steel bars. After each stage the beams were hanged in free-free conditions and their translational and rotational vibrations induced by a modal hammer have been measured with translational accelerometers and rotation rate sensors respectively. Measurements were later used to find average stiffness losses of the beams with model-updating technique. To help finding optimal solution, genetic algorithms were used. Our results show that by using rotation rate sensors one can improve determination of flexural stiffness losses in ordinary r/c or UHPC beams. Selected early results of the modal analysis of the UHPC beam specimens were recently published: https://www.mdpi.com/1424-8220/20/17/4711/htm