Full-range behavior of fiber reinforced cementitious matrix (FRCM)-concrete joints using a trilinear bond-slip relationship

• Published in: Composite structures Vol. 239; p. 112024
• Main Authors: Zou, Xingxing, Sneed, Lesley H., D'Antino, Tommaso
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.05.2020
• Subjects: Fiber reinforced cementitious matrix (FRCM) composite; Finite difference method (FDM); Neural network (NN); Trilinear bond-slip relationship; Finite difference method (FDM); Fiber reinforced cementitious matrix (FRCM) composite; Neural network (NN); Trilinear bond-slip relationship
• ISSN: 0263-8223; 1879-1085
• DOI: 10.1016/j.compstruct.2020.112024

Interfacial debonding of fiber reinforced cementitious matrix (FRCM)-concrete joints can be considered as a mainly mode-II fracture process, a problem that can be solved by accounting for one-dimensional interfacial shear stress-slip relationships. This paper presents an analytical approach to predict the load response of FRCM-concrete joints by adopting a trilinear bond-slip relationship consisting of a linear-elastic branch, a softening branch, and a friction branch. The applied load-global slip response of FRCM-concrete joints with (relatively) long bonded length includes five stages: elastic, elastic-softening, elastic-softening-debonding, softening-debonding, and debonding stages. Closed-form solutions of the interfacial slip, shear stress, and axial stress (or strain) distribution along the bonded length are provided. The response of FRCM-concrete joints with (relatively) short bonded length is examined. The effective bond length and a critical length for the existence of the snap-back phenomenon are derived. Experimental results reported in the literature are used to calibrate the parameters needed for the analytical approach. The analytical results are then compared with experimental results and with numerical results determined using a finite difference method (FDM). Finally, the capability of determining the parameters in the trilinear bond-slip relationship using a neural network (NN) with the experimental load response as the input is investigated.