Fatigue strengthening of damaged metallic beams using prestressed unbonded and bonded CFRP plates

Abstract

Bonded fiber reinforced polymers (FRPs) reinforcement systems have traditionally been found to be an efficient method for improving the lifespan of fatigued metallic structures and have attracted much research attention. Nevertheless, the performance of a bonded FRP reinforcement system under fatigue loading is basically dependent on the FRP-to-metal bond behavior. In this paper, a prestressed unbonded reinforcement (PUR) system was developed. The proposed PUR system can be used as an alternative to bonded FRP reinforcement, particularly when there is concern about the effects of high ambient temperatures, moisture, water and fatigue loading on the FRP-to-metal bond behavior. The performance of cracked beams strengthened by the PUR system was compared with that of cracked beams strengthened by the prestressed bonded reinforcement (PBR) system. A theoretical method was developed to estimate the level of prestressing sufficient to arrest fatigue crack growth (FCG). Furthermore, the method was used to examine different passive, semi-active and active crack modes with a loaded, strengthened beam. The mechanism by which a prestressed FRP plate forms a compressive stress field at the vicinity of the crack tip was also examined. Finite element (FE) modeling was conducted and the results were compared with experimental results.

Introduction

A survey of European railway administrations has revealed the existence of a variety of metallic, concrete, arch and steel/composite bridges across Europe [1]. The survey has included different climates areas and has covered nearly 220,000 bridges in almost all of the main European railway administrations that geographically cover most of Europe. The results have shown that almost 22% of the bridges are of metallic construction. Among these metallic bridges, nearly 3% are cast iron, 25% are wrought iron, 53% are steel and 20% whose material cannot be identified. Furthermore, 28% of the metallic bridges are over 100 years old, and nearly 70% are more than 50 years old. Recent developments in the railway industry have been characterized by increasingly higher speeds and heavier trains. Consequently, bridge structures that support moving trains are subjected to high stresses due to intense vibrations and dynamic deflections that are far greater than ever before [2].

Carbon fiber reinforced polymers (CFRPs) have the potential to reinforce metallic railway and highway bridges that carry permanent loads (dead-weight). Strengthening using a non-prestressed CFRP plate has been traditionally used to reinforce flexural members; however, the permanent loads are not transferred into the strengthening element, and the CFRP only acts against live loads. By strengthening with prestressed CFRP plates, a portion of the permanent load on the metallic structure will be transferred to the CFRP plate. Although laboratory experiments by several researches have proven the effectiveness of using a CFRP-bonded reinforcement technique to improve the load carrying capacity and service life of metallic members, bonded reinforcement systems (with or without prestressing) suffer from several drawbacks generally associated with the long-term performance of the adhesive bond between the metallic substrate and CFRP plate.

• High temperature: compared to concrete, steel has a high thermal conductivity (about 50 W/mK) and has significant ability to transfer heat rapidly to the adhesive. Moreover, the rate of sunlight absorption by steel is much greater than the rate of steel electromagnetic radiation (black body radiation); therefore, steel members exposed to direct sunlight on a hot day will easily become much hotter than the ambient temperature. This effect makes the adhesive adjacent to a hot steel surface soften excessively when the service temperature of the steel substrate approaches the glass transition temperature of the adhesive.

• Galvanic corrosion: although CFRP is a non-corrosive material, when carbon fibers come into contact with steel they can form a galvanic cell. Tavakkolizadeh and Saadatmanesh [4] investigated the galvanic corrosion between steel and carbon with different thicknesses of epoxy coating in different electrolytes, such as de-icing salt solution and ocean water.

• Metallic riveted bridges: due to the flat configuration of FRP plates, they cannot be bonded to the surface of structures that are not sufficiently smooth. Because the cover plate is riveted to the steel girders in steel-riveted bridges, for example, there is a high rivet density and the bonded FRP reinforcement system cannot be used.

• Heritage structures: the components of strengthening systems for heritage structures need to be designed for easy removal when there is a need to restore the structure to its original unstrengthened construction design. In a bonded reinforcement system, FRP strengthening materials cannot be easily separated from the beam due to the applied glue [3].

The main difference between FRP-steel and FRP-concrete bonded joints is that in the former, failure will likely occur in the adhesive layer and in the latter failure is expected to occur in the concrete. Thus, by providing an adequate bond length, the optimal strength of a bond joint is dependent on the fracture energy of the adhesive for former and concrete for the latter. Results from the experimental researches have shown that in the FRP-strengthened steel structures, interfacial failure should happen within the adhesive layer in the form of cohesion failure to maximize the effectiveness of FRP strengthening and minimize variations of the interfacial bond capacity as a result of different surface preparations. Inappropriate surface preparation of the steel substrate prior to the bond application will result in an adhesion failure at the steel-to-adhesive interface. Therefore, assuming the adhesive as the weakest point of a FRP-steel bond joint, in this paper, a prestressed unbonded reinforcement (PUR) system was developed and used for fatigue strengthening of damaged steel beams. The PUR system uses only two mechanical anchorage clamps to fix the prestressed CFRP plate to the beam without using any adhesive layer. Consequently the performance of strengthened beam is no longer dependent on the fracture energy of the adhesive, however depends on the efficiency of the designed anchorage system. The unbonded reinforcement system does not suffer from the drawbacks of the bonded reinforcement system that are associated with the glue in heritage structures, riveted bridges and environments with high temperatures. The existence of a small gap between the CFRP and the metal in a PUR system prevents galvanic corrosion. Moreover, it is recommended to place two thin glass fiber reinforced polymer (GFRP) plates on the sides of CFRP plate in the mechanical anchorage system to avoid galvanic corrosion between the metallic clamps and the CFRP, which can also result in a better stress distribution along the anchorage length. The proposed PUR system can be used as an alternative to the prestressed bonded reinforcement (PBR) system when there are concerns about the effects of high ambient temperatures, moisture, water and fatigue loading on the glue between the CFRP plate and the metal.

Flexural strengthening of steel structures using CFRP plates has attracted recent research attention [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22]. While there has been some research into the fatigue strengthening of damaged metallic members with CFRP plates [23], [24], [25], [26], [27], [28], [29], [30], there are relatively few studies that have used a prestressed CFRP plate for reinforcement against cyclic loadings [31], [32], [33], [34], [35], [36]. Nevertheless, all of these studies have used a type of bonded FRP reinforcement system. In the current article, the effect of the two different reinforcement techniques (the PUR and PBR systems) on the fatigue and fracture performance of damaged steel beams was investigated. On the other hand, although the researches that used a prestressed CFRP plate have observed a significant decrease in the fatigue crack growth (FCG) rate and in some cases a crack arrest [33], nevertheless no theoretical study on the prestressing level required for a completely crack arrest has been performed. In this research an analytical method for estimating the prestressing level needed to arrest the fatigue crack propagation was proposed. A finite element (FE) analysis was performed, and the results were compared with those of the experiments.