Distributed fibre optic sensing (DFOS) offers millimetre-scale, continuous strain measurements that can reveal the longitudinal behaviour and cracking of concrete members far beyond what conventional point sensors can provide. For existing concrete structures, however, its effective use is still limited by three issues: the lack of an evidence-based installation strategy for surface-bonded fibres, limited quantification of how strain is transferred from concrete to the fibre, and incomplete validation of crack widths derived from DFOS under realistic data conditions.

This thesis addresses these gaps through a combination of literature review and laboratory experiments on reinforced-concrete members with surface-bonded DFOS, complemented by a conceptual application to an existing prestressed concrete box-girder bridge. As a qualitative pilot, an inverted T-girder tested in three-point bending is instrumented with DFOS and digital image correlation (DIC). The distributed strain profiles clearly reveal the formation and growth of flexural and shear cracks, but they also expose weaknesses of generic installation guidelines, such as non-uniform adhesive layers, local debonding and data gaps near steep strain gradients. These observations are used to formulate a refined, evidence-based installation strategy for surface-bonded DFOS on concrete.

In a second phase, four reinforced-concrete beams are tested in four-point bending with DFOS, strain gauges and digital image correlation (DIC). Comparisons between DFOS and strain-gauge measurements in both tension and compression show that the fibre systematically underestimates the true concrete surface strain, but with an almost constant ratio for a given installation. This allows a strain-transfer efficiency factor to be identified so that DFOS strains can be converted into realistic concrete strains in the uncracked range. DFOS-based crack widths, obtained by integrating the corrected strain peaks around cracks, are then validated against DIC. For cracks above a practical resolution limit, good agreement is achieved as long as the DFOS signal around each crack is largely intact. When substantial parts of the peak are missing, the error in DFOS crack widths increases and the results become unreliable.

Overall, the thesis demonstrates that surface-bonded DFOS can be used quantitatively for strain and crack-width monitoring in existing concrete structures, provided that installation is treated as a carefully designed process, strain-transfer efficiency is calibrated, and simple data-quality checks are incorporated into the interpretation of crack measurements.