PSI - Issue 64

Quentin Sourisseau et al. / Procedia Structural Integrity 64 (2024) 893–900 Quentin SOURISSEAU/ Structural Integrity Procedia 00 (2019) 000 – 000

894

2

1. Introduction

Maintenance of steel structures (Floating Production Storage and Offloading units , naval ships, …) installed in harsh offshore environment is gaining more and more interest. For such applications, adhesively bonded FRP (Fiber Reinforced Polymer) present several advantages (short down-time, non-intrusive process, high performance materials, non-corrosive materials, cold work, complex 3D geometry, … ), Zhao et al. (2007), Kamruzzaman et al. (2014), Ghafoori et al. (2019). However, the design of such reinforcement patch for fully stressed area present important challenges as high stresses in the bond line may occur. Thus, a good understanding of adhesion mechanics and strength is critical for designing reliable patch repairs. A better understanding of the failure of bonded joint and the development of reliable design methodologies are necessary to increase confidence and to lead to more efficient application of composite bonded repairs in offshore applications. Such methodology could be further developed to study damage tolerance and ageing. When designing adhesively bonded joints, two steps are generally required. The first step aims at assessing the adhesive or the joints capacities through standardized experimental investigations and allows obtaining failure criteria data and sometimes mechanical behavior information. It may also give information related to surface preparation and failure modes. The second step requires having access to mechanical analysis tools that may be analytical or numerical. Their use provides information related to design predictions. Comparison of both results (failure criteria and design predictions) is then carried out, and, if needed, design is modified. Several design approaches have been developed so far for adhesively bonded joints (relying on continuous mechanics, fracture mechanics or both), Da Silva et al. (2008). It was decided in this work to investigate fracture mechanics approach, using cohesive zone modeling tools, Sourisseau et al. (2022a). When using such approaches, it is necessary to determine critical energy release rates for the different modes of loading present in the structure and for each existing interface through standardized methods such as Double Cantilever Beam (DCB) for mode I or End Notched Flexure (ENF) Test for mode II. Those critical energy release rates can then be used to determine adapted cohesive zone models used during the design of the real patches. For the specific case of FRP patches that may include different types of internal plies, the need to investigate all the existing interfaces is a clear drawback of the methodology as it drastically increases the required number of investigations. Thus, investigations were carried out to develop modified fracture energy samples able to characterize several interfaces at once. In the first part, the equivalent interface samples will be presented and the led investigations briefly described. Those aimed at determining the fracture energy characteristics in mode I, mode II and in mixed mode (combinaison of modes I and II). The second part will detail the methodology used to determine the cohesive behavior laws. Finally, in order to assess the efficiency of the proposed design methodology, large scales samples were tested up to failure, and experimental ultimate loads are compared to design expectations using the developed methodology. This is developed in the last part before the conclusion. 2. Description of the investigations conducted on equivalent interface samples 2.1. Geometry of specimens The studied adhesively bonded FRP patch consists of glass fiber plies and carbon fiber plies. Specifically, glass plies are located between the carbon plies and the steel surface to soften the interfacial stresses and prevent galvanic corrosion risks. The FRP patch is fabricated by infusion, Sourisseau et al. (2023). When using the classical approach, it would then be needed to investigate the glass ply to steel interface, the glass ply to glass ply interface, the carbon ply to glass ply interface and the carbon ply to carbon ply interface for the different loading modes (here we only investigate mode I, mode II and mixed mode). To limit the number of investigations, a sample was designed in accordance with existing standards used for steel to-steel bonded joints but that includes the different interfaces (ASTM standards D5528, D7905 and D6671 respectively for the DCB in mode I, ENF in mode II and Mixed Mode Bending (MMB) for mixed mode I and II). The final sample geometry is given in Figure 1. The same sample is used for all the investigations. It is important to note