PSI - Issue 45

Yuanpeng Zheng et al. / Procedia Structural Integrity 45 (2023) 96–103 Author name / Structural Integrity Procedia 00 (2019) 000–000

102

7

tests. The outer layers of carbon fiber sheets have limited contribution to the reinforcement because of the shear lag effect as the bonding area of the repaired CT specimen is insufficient for a complete stress redistribution by material stiffness under the concentrated tensile loading. Additionally, thicker patches might cause severer stress concentration close to the material edges but also poses a challenge for high-quality carbon-fiber-sheet wet lay-up on the damaged steel member. Both disadvantages may facilitate CFRP delamination and debonding from the steel substrate under fatigue loading which were observed in the test (Zheng et al. 2022). 5.2. Debonded region description The scalar damage variable SDEG provides a qualitative debonded region description in the numerical model. The calculated debonded region (Fig. 12) shares a similar shape (semi-ellipse) with the experimental post-mortem observation results (Fig. 13), though larger in size. The boundary of the debonded region does not reach the final crack tip of the specimen, which is different from the assumptions and inspections of some existing research (Colombi et al. 2003a, 2003b; Colombi et al. 2015). Instead of being determined by the crack dimension, it is likely that debonding depends on the relative deformation between the steel substrate and the FRP, which is more connected to the parameter of crack opening displacement (COD), especially for CT specimens.

CFRP sheet

Steel

Specimen R-D-1

Half width a and half height b < 5~6 mm

Debonded area

Fig. 12. Numerical debonded region description of Specimen R-D-1

Fig. 13. Semi-ellipse debonded region from post-test observations

5.3. Enlightenment for prospective design Besides manually predefining a debonded region, to conduct a more detailed simulation for metallic fatigue strengthening, it is also advisable to consider the actual behavior of the CFRP-steel interfaces in the analysis. In the current China national code of GB 50608-2020 (Ministry of Housing and Urban-Rural Construction of the People’s Republic of China, 2020), the stress range of steel members after FRP strengthening is simply determined by its stiffness ratio of the overall steel-FRP composite structure. But the concept does not sufficiently reflect the premise of reliable adhesion which in many cases is far from satisfying in engineering reality and limits the performance of FRP patching. This paper emphasizes the significance of the actual steel-FRP bonding behavior in prospective analysis and design via the numerical simulation explained above. 6. Conclusions This paper presents the numerical simulation of CT specimens repaired by CFRP, based on an experimental program in which carbon fiber sheets were utilized to elongate the fatigue life of CT specimens. The eXtended finite element method expedites SIF range calculation with good precision. A revised bond-slip model defined in cohesive elements details the response of the CFRP-steel adhesion under fatigue loading, which could be further extended for prospective simulations on other FRP-repaired cracked steel members considering the possible debonding. Good agreement is observed in terms of SIF range and fatigue life, and the error is discussed both from the perspective of the experiment and the numerical modeling. Debonded region estimation in the modeling instead of a manual predefinition corresponds with experimental observations, offering other options to include the debonding phenomenon during analysis. As considering the damage evolution of the adhesive is critical for safe predictions, caution is needed for prospective fatigue strengthening design on steel components with relatively limited bonding