Distributed fibre optical sensing in structural concrete – exploring its potential to improve the mechanical modelling of bond
Author: Tena Galkovski
Language: English
external pageDOI: 10.3929/ethz-b-000642506call_made
Abstract
The efficiency of structural concrete is closely linked to the effectiveness of the force transfer between concrete and reinforcing steel, commonly referred to as bond. Thanks to this interaction between both materials, the steel's high tensile strength and the concrete's good performance under compression can be fully exploited in the composite material. Bond governs not only the anchorage of reinforcement and the ductility of the structure (structural safety), but also crack spacing, crack widths and deflections (serviceability). With this significant influence on the Ultimate Limit State and the Service Limit State, considering bond is essential to accurately assess the structural behaviour of reinforced concrete members. The importance of bond for a safe and efficient design is reflected in how intensely it has been studied. However, despite the wide range of theoretical, numerical and experimental investigations, numerous aspects of bond are not satisfactorily resolved or understood, which underlines the complexity of the reinforcing steel-concrete interaction.
The difficulty in accessing and instrumenting the interface between concrete and reinforcing steel has been a critical obstacle in the experimental investigation of bond. Until recently, measuring the interaction over representative lengths without biasing the response required disproportionate efforts, such as glueing strain gauges between reinforcing bar halves or utilising a multitude of expensive discrete fibre optical sensors. These obstacles to instrument reinforcing bars were bypassed by conducting simplified tests in which only global parameters are measured. This is the case for pull-out tests, the most established method to study bond, where the applied load and the bar slip with respect to the concrete surface at the passive end are measured. However, the average bond shear stress-slip relationships obtained from pull-out tests are not necessarily representative of the bond behaviour under actual boundary conditions in structural elements (tension ties, uniaxial and biaxial bending, shear) because of the different stress states activated in the concrete. Nonetheless, most established models, as well as current design standards rely on bond shear stress-slip relationships calibrated on from pull-out tests, and merely account for the numerous influencing factors, including the structural configuration of the element, by semi-empirical corrections.
This thesis addresses the above-mentioned issues by experimentally investigating bond in reinforced concrete members leveraging the use of distributed fibre optical strain sensing (DFOS). Thanks to technical advances in Rayleigh backscatter-based coherent frequency domain reflectometry, inexpensive commercial telecommunication glass fibres can serve as sensors. The strains of instrumented reinforcing bars can be measured quasi-continuously with minimal impact on the reinforcing bar-concrete interface. Local steel stresses can then be inferred from the strains, and bond stresses obtained from the local steel stress gradient. This method has the potential to enable a better understanding and modelling of the bond behaviour of structural concrete under general loading and member geometry.
To this end, the first part of the thesis is dedicated to establishing a reliable basis for applying DFOS to structural concrete. Its suitability and drawbacks are explored, and a guideline is proposed for the appropriate instrumentation and data post-processing. The measurement technique is compared to established methods, and suitable measures for dealing with the high resolution are elaborated. Pilot test applications accompany this part with the aim of validating the measurement technique.
The second part of the thesis builds on the insights gained from the refined measurements and explores ways to capitalise on them. Established bond models are scrutinised for the fundamental loading cases of uniaxial tension and uniaxial bending, with the aim of revisiting common modelling assumptions and identifying potential flaws. To this end, in-depth analyses of reinforced concrete members tested in uniaxial tension and compression are performed. Moreover, the gained insights on the local bond distribution and the observed effects of factors such as the bar diameter, concrete cover, loading, and material properties inspired the development of a stress field model for the load transfer from the reinforcing bar to the concrete. The proposed stress field does not rely on empirical expressions and treats bond as a result of the bearing capacity of the surrounding concrete instead of reducing it to an interface characteristic. By assigning stiffnesses to the steel and the concrete and formulating a simplified compatibility condition requiring contact between concrete and steel at the ribs, the local and average bond stresses can be determined as a function of the steel strains, respectively, the applied load.