Reference : Damage detection in prestressed concrete bridges based on static load testing, saggin...
Dissertations and theses : Doctoral thesis
Engineering, computing & technology : Civil engineering
Damage detection in prestressed concrete bridges based on static load testing, sagging and modal parameters, using measurements and model updating
Schommer, Sebastian mailto [University of Luxembourg > Faculty of Science, Technology and Communication (FSTC) > Engineering Research Unit >]
University of Luxembourg, ​​Luxembourg
Docteur en Sciences de l'Ingénieur
xviii, 176 + 11
Maas, Stefan mailto
Zürbes, Arno mailto
Schäfer, Markus mailto
Strauss, Alfred mailto
Link, Michael mailto
[en] structural health monitoring ; damage detection ; prestressed concrete ; bridges ; damage indicators ; modal parameters ; static load tests ; model updating
[en] Bridges are an essential part of nowadays infrastructure to cross natural and artificial obstacles like rivers, valleys or other roads and railways. Many concrete bridges were built in the last 70 years. The traffic density has increased immensely over the last decades and the bridges are suffering from corrosion and wear. Nevertheless, the safety of the infrastructure has to be guaranteed and therefore it is very important to find efficient methods for structural health monitoring.
For this purpose, visual inspections are the most widely adopted in reality today. Considering the size of most bridge structures, it is understandable that these tests are generally very time-consuming and many personnel are needed, so they are cost-intensive. However, it is not always guaranteed that all damage can be found as only the surface is accessible. For instance, internal damage, such as corrosion of passive reinforcements or prestressed tendons, is difficult to detect. In addition, small cracks can remain undetected when covered by paints or dirt.
Therefore, it is important to complement the standard methods with advanced alternatives. The aim is therefore not necessarily to replace visual inspections, but rather to find efficient methods for amendment.
An idea being vigorously discussed in the scientific community is based on vibration measurements of a structure to assess its dynamic behaviour. The occurrence of damage will change the system properties, as it changes above all the stiffness distribution. So the system identification process in principle allows detection of changes of eigenfrequencies and hence stiffness.
The main problem in practice on real bridges is that the robustness of a method is often insufficient, as the measured parameters are often also influenced by temperature changes. It will be shown that the impact of temperature change, e.g. between night and day, on the system properties is much higher than the influence of small damage. Furthermore, changes in soil and bearing conditions between different seasons can play a role. These environmental effects have to be taken into account while performing measurements for damage assessment. For this purpose, strategies are proposed to compensate environmental effects.
Therefore, this thesis focuses on measurements under real environmental conditions, outside a laboratory. Different methods for damage assessment or stiffness tracking based on measured static and on dynamic properties of structures are deployed. Finally, the measured and analysed physical properties of the bridges in this thesis are: eigenfrequencies, mode shapes, sagging under own weight and the deflection line under a static test-loading. These quantities are tracked and artificial damage is introduced stepwise to a test-beam of a real bridge. Damage assessment and localisation is tried directly with the measured quantities but also by model-updating of a finite element model. This solid model is divided in a special way in different slices. It is possible to change the stiffness distribution along the axis of the simulated beam by varying the Young’s moduli of these slices. Furthermore, to reduce the number of free parameters for the subsequent up-dating process, an exponential damage function is introduced that describes the stiffness distribution. At first, the model was designed to fit a healthy reference state. Now measurement data from the artificially damaged test-beam are introduced and the model is updated by changing the Young’s moduli of the slices to minimise a special objective function containing the measured and simulated physical quantities. The comparison of initial and updated model allows a quantification and localisation of damage. Finally, the slice width is reduced around the identified damage region to improve the process.
Researchers ; Professionals ; Students

File(s) associated to this reference

Fulltext file(s):

Open access
PhD-Thesis_Sebastian_Schommer.pdfAuthor postprint13.13 MBView/Open

Bookmark and Share SFX Query

All documents in ORBilu are protected by a user license.