Skip to Main content Skip to Navigation

Structural Health Monitoring Optimization

Abstract : Following the growing expansion of civil engineering infrastructure throughout the twentieth century, the problem of Inspection, Maintenance and Rehabilitation (IM&R) of structures is currently given a particular attention. The importance of the problem manifests itself in industrialized countries because of the importance of their ever-growing heritage in terms of aging civil engineering structures. Thus, the reallocation of budgetary expenditure towards the inspection, maintenance and rehabilitation of the structures was made. As for the developing countries, a different pattern is observed. Despite the accumulation of maintenance needs due to the lack of budgets, available budgets are devoted to new constructions.In this context, managing the lifespan of existing structures is becoming a major challenge for the society. The evaluation of structures’ health state customarily relied on intermittent surveillance of structures at specific points in time by visual inspections and / or non-destructive detection techniques. However, these intermittent surveillance techniques make it difficult to detect any defect in the structure during an inspection visit. In some cases, for instance, a critical defect could appear between two successive inspections and not be detected in time. Monitoring structures using permanent sensors (known as Structural Health Monitoring "SHM") overcome this shortcoming and makes it possible to continuously identify and monitor the state of deterioration. The obtained results would be used in order to draw indicators on the structure’s health and to assess its residual life. Unavoidable budget and resource limitations lead to the need for an optimal configuration of sensors. The aim of the thesis is therefore to develop a framework consisting of several algorithms for the detection, localization and characterization of damage as well as the optimization of the sensors configuration.First, a state-of-the-art review considering works done on structural monitoring, detection methods and optimization methods is presented. Four methodologies are then developed:The first methodology, based on a hierarchical Approximate Bayesian inference, concerns the detection of structural damage without having to solve the inverse problem which is generally ill-posed. The main advantage of this method lies in its ability to take into account, systematically and transparently, all uncertainties affecting the structural system as well as the measurement system.This methodology is further developed to amplify the information about less monitored elements and/or structures (whose condition states are defined with high uncertainty) using information collected from well monitored structures and/or elements (whose condition states are defined with low uncertainty). An approach is then proposed to optimize the planning for the monitoring and maintenance of structures using data fusion of SHM results and conventional inspections outcomes. Finally, a new predator-prey approach is proposed for optimizing the configuration (i.e. type, number and location) of sensors in a structure. All these methods have shown their effectiveness through numerical applications on different types of structures.
Document type :
Complete list of metadata
Contributor : ABES STAR :  Contact
Submitted on : Tuesday, May 10, 2022 - 3:40:48 PM
Last modification on : Wednesday, May 11, 2022 - 3:48:53 AM


Version validated by the jury (STAR)


  • HAL Id : tel-03663965, version 1


Christelle Geara. Structural Health Monitoring Optimization. Civil Engineering. Université Clermont Auvergne; Université Saint-Joseph (Beyrouth), 2021. English. ⟨NNT : 2021UCFAC087⟩. ⟨tel-03663965⟩



Record views


Files downloads