doi:10.1016/j.compstruct.2020.112916

Article start

Bond-slip response of steelfibers after exposure to elevated temperatures:Experimental program and design-oriented constitutive equationRamoel Serafinia,b,⇑, Ronney Rodrigues Agraa,b, Luís A.G. Bitencourt Jr.c, Albert de la Fuented,Antonio D. de FigueiredoaaDepartment of Civil Construction Engineering, Polytechnic School at the University of São Paulo Avenida Professor Almeida Prado, Travessa do Biênio, 83, 05424-970 São Paulo, BrazilbInstitute of Technological Research, Avenida Professor Almeida Prado, Travessa do Biênio, 83, 05424-970 São Paulo, BrazilcDepartment of Structural and Geotechnical Engineering, Polytechnic School at the University of São Paulo, Avenida Professor Almeida Prado, Travessa do Biênio, 83,05424-970 São Paulo, BrazildDepartment of Construction Engineering, Polytechnic University of Catalonia, Jordi Girona, 1-3, 08034 Barcelona, SpainARTICLE INFOKeywords:PulloutBond‐slip responseHooked‐end steelfibersElevated temperaturesNumerical simulationABSTRACTThis study aimed to evaluate the effect of elevated temperatures on the bond‐slip behavior of hooked‐end steelfibers. A total of 180 pullout specimens were tested in post‐cooling conditions using a double‐sided pullout testwith multiple embeddedfibers for target temperatures between 25 and 750 °C. Results proved that the bondstrength significantly increases for temperatures up to 450 °C, and drastically decreases for temperatures of 600and 750 °C. The contribution of hooks reduced with temperature and is negligible for temperatures higher than600 °C, while thefiber‐matrix frictional interaction seems to improve for all temperatures evaluated. Atemperature‐sensitive constitutive equation that allows simulating the bond‐slip behavior of hooked‐end steelfibers is proposed and its suitability confirmed through a numerical model.1. IntroductionThe fundamental principle of steelfiber reinforced concrete (SFRC)is based on the interaction betweenfibers and the cementitious matrixto provide plain concrete a pseudo‐ductile behavior and enhance post‐crack mechanical properties. The steelfibers bridge the cracks throughthe matrix and enhance the post‐crack tensile strength of the compos-ite, which is a key feature for structural applications[1,2]. Three maincomponents that must be characterized to analyze these compositesafter temperature exposure are the cementitious matrix, thefibers,and thefiber–matrix interface. In this regard, the bond‐slip responsebetween thefiber and the cementitious matrix is a key parameter tomodel the SFRC mechanical behavior[3]and is known to stronglyinfluence the post‐crack behavior of the composite[4,5]. The bond‐slip response of hooked‐end steelfibers in the cementitious matrixcan be parameterized by pullout test results and can be divided intofive main stages[6], which are schematically shown inFig. 1.Stage 1 and 2 are defined as the partial and full debonding betweenfiber and matrix, respectively. Once the full debonding takes place, thedynamic frictional interaction betweenfiber and matrix begins. At thisstage, thefibers must undergo a considerable plastic deformation dueto the straightening of the hooks before the dynamic frictional interac-tion occurs[6], which defines Stages 3 and 4. The mechanical inter-locking caused by the plastic deformation of the hooks at M1 andM2 increases the maximum pullout load value at Stage 3.As soon as the tip of thefiber goes beyond M1, the pullout load sub-stantially decreases and the mechanical interlocking is caused only bythe plastic deformation of the hook at M2. It is important to remindthat the matrix cracks as the pullout occurs[5]and that the energyrequired to yield the hooks is intrinsically bound to the bending stiff-ness and the hook geometry of the steelfiber, as well as the interfacialbond properties[7,8]. After the hook is completely straightened, thedynamic frictional interaction betweenfiber and matrix takes placein Stage 5. This last phase results in a rapid drop in the pullout loadvalues and prevails until the wholefiber is removed from the matrixpathway[6].Pullout tests are conducted in a wide variety of setups in the liter-ature. The difference between methodologies adopted is usuallyrelated to the number offibers embedded (single or multiplefibers)and the method of applying the tensile force (single‐or double‐https://doi.org/10.1016/j.compstruct.2020.112916Received 18 February 2020; Revised 10 August 2020; Accepted 1 September 2020Available online 5 September 20200263-8223/© 2020 Elsevier Ltd. All rights reserved.⇑Corresponding author at: Department of Civil Construction Engineering, Polytechnic School at the University of São Paulo Avenida Professor Almeida Prado, Travessa do Biênio,83, 05424-970 São Paulo, Brazil.E-mail address:rserafini@usp.br(R. Serafini).Composite Structures 255 (2021) 112916Contents lists available atScienceDirectComposite Structuresjournal homepage: www.elsevier.com/locate/compstruct

Ramoel Sera

Ronney Rodrigues Agra

Luís A.G. Bitencourt

Albert de la Fuente

Antonio D. de Figueiredo

sided). The pullout tests are commonly performed employing single‐sided tests on a singlefiber, owing to the simplicity of the methodol-ogy during the preparation of specimens and during the test[6]. How-ever, a major drawback of this methodology is that the test setup needsto be capable of precision measurements due to the intrinsically lowpullout forces. More than that, difficulties encountered in single‐sided tests are mainly associated with the interaction between the gripand thefiber. Also, high variability is common in single‐sided tests ona singlefiber, which requires a considerable number of specimens toguarantee the reliability of the results[9].The drawbacks of using single‐sided tests with a singlefiber are fur-ther aggravated for pullout specimens that are exposed to elevatedtemperatures. This occurs because the pullout forces are expected toreduce and the steelfiber mechanical properties to be negativelyaffected after high‐temperature exposure. Several studies have investi-gated the bond‐slip properties of hooked‐end steelfibers at room tem-perature conditions in the last decades[9–12]. However, the studiesregarding the bond‐slip behavior of steelfibers after elevated temper-atures are very scarce in the literature. In this sense, there is a need foridentifying constitutive equations and analytical formulations to thedesigners’community so that the effect offire on SFRC structurescan be properly assessed.Results obtained by recent studies show that the pullout load valueswere comparable up to~400 °C and significantly reduced for highertemperatures[13–16]. However, the non‐significant effect of temper-ature up to 400 °C may be a side effect of the intrinsic dispersion ofsingle‐sided tests with a singlefiber, since the standard deviation val-ues were omitted. Moreover, numerical models focusing on the explicitand discrete representation of the steelfibers in SFRC have been devel-oped recently and require an accurate description of the steelfiberbond‐slip behavior as input[17,18]. So far, the results published inthe literature have not provided a microstructural based explanationfor the changes in the bond‐slip behavior or proposed a constitutiveequation for design purposes, which denotes that the topic still needsto be deeply investigated.The present study aims to evaluate the bond‐slip behavior ofhooked‐end steelfibers after exposure to elevated temperaturesemploying a double‐sided pullout test using multiplefibers. This testmethodology aims at increasing the stability of the test and avoidingthe drawbacks associated with single‐sided and singlefiber pullouttests. The interfacial transition zone of the steelfibers was character-ized to assess the effect of temperature on the vicinity of thefiberand relate the microstructural results with the mechanical behavior.Additionally, an analytical model was proposed and the pullout testswere validated through numerical simulations using a discrete andexplicit representation of steelfibers inside the pullout specimens.2. Materials and methodsFig. 2shows a schematic drawing of the experimental program con-ducted in this study. The investigation herein conducted took placewithin the framework of a Ph.D. research project regarding the studyof the effect of temperature on the properties offiber‐reinforced com-posites. In this sense, all the characterizations were conducted using amortar that followed the SFRC mix design based on the work of Ser-afini et al.[19]and a detailed description and characterization ofthe materials employed can be found in the referenced study. Evenwith those considerations, a brief description is presented in thissection.The bond‐slip response of the hooked‐end steelfibers was evalu-ated after exposure to high temperatures. Based on the experimentalresults, an analytical equation that computes the effect of temperatureon the bond‐slip behavior of steelfibers is proposed. This equation isused as input for a refined numerical model that is capable of repre-senting the steelfibers discretely and explicitly inside the plain con-crete[18]. As supplementary investigations, the characterization ofthe interfacial transition zone (ITZ) is performed and aims to verifythe effect of temperature on the vicinity of thefiber, which can be cor-related with the bond‐slip response. Additionally, the effect of temper-ature on the compressive strength and dynamic elastic modulus of themortar were evaluated and serve as input for the numerical simulationconducted in this study.2.1. MaterialsThe cementitious materials used in this study were a Type I Port-land cement (CEM I 52.5R) and silica fume type Elkem 920‐U. The par-ticle packing was increased by using river and artificial sand asfineaggregates and two coarse granite aggregates. A polycarboxylate‐based superplasticizer, GCP ADVA Cast 525, was used to provide con-sistency to the mix. A cold‐drawn, hooked‐end steelfiber, Dramix 3D80/60‐BG, was employed. This steelfiber was chosen since it is com-Fig. 1.Schematic diagram of the pullout mechanism for hooked-end steelfibers.R. Serafini et al.Composite Structures 255 (2021) 1129162