Fibre-reinforced-polymer (FRP) composites have been widely used for the flexural strengthening of reinforced concrete (RC) structures. Flexural strengthening methods with FRP include external bonding of FRP composites (EB system) and insertion of FRP strips or bars into grooves cut into the concrete (near-surface-mounted or NSM system). Recently, a prestressed FRP strengthening system has been developed and investigated, whereby the FRP reinforcement is pretensioned prior to attachment to the concrete to maximize the use of the high tensile strength of the FRP reinforcement. Existing studies have shown that the ultimate load carrying capacity and serviceability were greatly improved in FRP flexural strengthened beams. However, the only disadvantage of the FRP strengthening system is the reduction of deformability compared to that of unstrengthened structures due to the limited strain capacity of the FRP reinforcement and premature debonding failure. Structures with low deformability may fail suddenly without warning to evacuate, resulting in catastrophic failure. Therefore, a study on the improvement of deformability is critical for the effective use of FRP strengthening systems.
In this study, a partially bonded concept is introduced and applied to various FRP strengthening methods, with the specific objective of increasing deformability in FRP strengthened beams. The FRP reinforcement is usually completely bonded to the concrete tensile surface, while a portion of the FRP length is intentionally unbonded in the partially bonded system in order to improve deformability while sustaining high load carrying capacity. To investigate the general behaviour of the partially bonded system, a new analytical model has been developed because conventional section analysis used for analysis of the fully bonded system is not applicable due to strain incompatibility at the FRP reinforcement level within the unbonded length. The analysis shows that a partially bonded system has a high potential to improve deformability without the loss of strength capacity.
An extensive experimental program was conducted to verify the analytical model and to investigate the actual behaviour of the partially bonded beams. A total of seventeen, 3.5m long, RC T-beams were constructed and tested. One of them is an unstrengthened control beam, while the other 16 beams consist of four test groups that were strengthened by different strengthening methods: non-prestressed EB, non-prestressed NSM, 40% prestressed NSM, and 60% prestressed NSM. To allow investigation of the effect of partially unbonding, each group has different unbonded lengths and includes a fully bonded beam.
For the non-prestressed EB strengthened beams, the failure mode of all beams was premature FRP debonding failure without regard to the bond condition. The ultimate strength and the ultimate deformability of the partially bonded beams were improved compared to the fully bonded beam. This was because the typical intermediate debonding failure that occurred in the fully bonded beam was avoided due to intentional unbonding in the partially bonded beams. The analytical model predicted the general behaviour of the EB strengthened beams well except at the ultimate response due to the premature debonding failure. A three-dimensional nonlinear finite element (FE) analysis was performed utilizing interfacial elements and contact modeling to investigate the debonding failure of this system. The FE analysis represented the behaviour of the debonding failure and bond stress distributions at FRP-concrete interface of both the fully bonded and partially bonded beams well.
For the non-prestressed NSM strengthened beams, the premature debonding failure that occurred in the EB strengthened beams was not observed, and almost the full capacity of FRP was exhibited. Prominent stiffness reduction was observed in terms of load-deflection diagrams at the post-yielding stage with the increase of the unbonded length. This stiffness reduction increased the deformability of the partially bonded beams for a given applied load after steel yielding in comparison to the fully bonded beam. The FRP started to slip at high load levels and the concrete crushed gradually with a gradual loss of the beam’s cross-section, inducing nonlinear behaviour near the ultimate state of the beams. To address this behaviour, an advanced analytical model utilizing idealized section model and slip model is proposed to consider the FRP slip and concrete gradual failure.
Prestressed NSM strengthened beams were very effective to improve the cracking load, to decrease the deflection at service load, and to increase the ultimate load compared to non-prestressed NSM strengthened beams. This improvement was greater as the prestressing level increased. The partially bonded prestressed beams showed an improvement in deformability compared to the fully bonded prestressed beams while minimizing the reduction of the ultimate load carrying capacity and serviceability. The partially bonded system was more effective to improve the deformability at higher levels of prestressing force.
Based on the model developed, a parametric study was performed varying the main parameters. This showed that the FRP strengthened beam that has an FRP area (Af) less than the balanced FRP area (Af,b) of the beam has a high potential to improve the deformability as the unbonded length increases. The balanced FRP area is increased as the concrete strength and the FRP prestressing force are increased, or as the area of the steel reinforcement decreases. Finally, design recommendations and procedures are proposed for the effective use of the partially bonded system to improve the deformability of FRP strengthened concrete beams.
Identifer | oai:union.ndltd.org:LACETR/oai:collectionscanada.gc.ca:OWTU.10012/4048 |
Date | 19 September 2008 |
Creators | Choi, Han Tae |
Source Sets | Library and Archives Canada ETDs Repository / Centre d'archives des thèses électroniques de Bibliothèque et Archives Canada |
Language | English |
Detected Language | English |
Type | Thesis or Dissertation |
Page generated in 0.0024 seconds