A cohesive model to predict the loading bond capacity of concrete structures repaired/reinforced with HPFRC/UHPFRC and stressed to mixed mode

https://doi.org/10.1016/j.cemconcomp.2020.103673Get rights and content

Abstract

The risk of cracking/debonding of a cement overlay used to repair or strengthen an existing structure is still a key issue. Current bond test methods are not designed to measure the combined effect of peeling (mode I) and shear (mode

II) on the interface. A few existing models propose theoretical approaches to predict that, but they were fitted on specific cases and lack in generality. In addition, controversial opinions about the influence of both the moisture level of the substrate surface prior to the application of the overlay and properties of the latter on the loading bond capacity call for further investigations. In this work, a cohesive model is developed to predict the loading bond capacity of an existing concrete structure overlaid by a layer of HPFRC/UHPFRC. Different bond tests were specifically designed for calibrating the cohesive parameters employed into the model, which also takes into account the type of the overlay used and the moisture conditioning level. An experimental campaign confirmed the reliability of the predictions of the proposed theoretical model.

Introduction

An increasing number of industrial applications of HPFRC (High Performance Fibre Reinforce Concrete) as repairing material on deteriorated structures has been observed in recent year [16,29]. This is not the case for UHPFRC materials (Ultra High Performance Fibre Reinforce Concrete), despite their higher performances in terms of strength, ductility and durability [13,[40], [41], [42]]. Among the reasons underlying the lower interest to use UHPFRC in the field of the civil engineering there are both the high manufacturing cost and the missing harmonization of existing codes. Nevertheless, recent applications have successfully involved the use of UHPFRC both as new materials (e.g. beam, panel manufacturing, etc.) [67] and as overlay material for rehabilitating and strengthening bridge decks [11,26] and hydraulic structures [25,39]. As far as the properties of a fiber reinforced overlay material could be excellent, if the substrate preparation and pouring operations of the overlay are not well designed, the risk of bond failure could be high. In the practice of retrofitted concrete structures, the bond failure is caused by both the different physical properties between substrate and overlay (thermal expansion coefficient, elastic modulus, etc.) and external loadings [22]. For both cases, the cracking/debonding along the interface is related by several aspects like the fracture energy magnitude and the shape of cohesive law governing the interface response. HPFRC/UHPFRC overlays might therefore reduce the risk of cracking/debonding, since they provided high bond strength and good adherence to the concrete substrate members, as observed in previous works [1,27,48]. This effect is also related to the presence of steel fibers within the HPCC/UHPC matrix that transmit the force through the cracks in the matrix1, thus the built-in peak stress2 at the interface decreases, which reduces the risk of premature cracking/debonding [23]. Further improvements of bond strength can be achieved by installing a series of dowel bars properly anchored both in the substrate and in the overlay, even though the reinforcement has to be deformed plastically prior to carry a relevant part of the load. Since they are more rigid than the overlay, they will carry the load only after the bond breaks [46]. This solution, even if increases the global capacity of the bond, does not prevent a premature deterioration of the interface. A premature cracking/debonding can be avoided also by both a correct testing of the bond loading capacity [52] and a subsequent accurate structural design [12,14,15].

The soundness and roughness of the substrate strongly influence the bond strength development. Both parameters seem to depend on the removal methods of the deteriorated concrete [8,46], like impact, high pressure water, or mixture of them. The impact methods are based on the use of breakers to fracture and spall the unsound concrete. Rougher surfaces of the interface are provided, which is beneficial to the bond strength [37]. But the heavy impact performs micro-cracks on the concrete surface [21,54]. Another removal method used in the practice is the hydro-jetting. Hydro-jetting disintegrates unsound or deteriorated concrete and ensures a substrate with a sound and rough surface profile. Hydro-jetting provides a less pronounced roughness profile than impact methods, but no micro-cracks are observed [37]. Nevertheless, Kauw and Dornbusch (1997) [28] and Silfwerbrand (2000) [43] concluded that a minimal compressive strength of the substrate is requested to avoid the rupture of sound concrete as well, as also confirmed by Bissonnette at al. (2008) [8]. Findings of Silfwerbrand (1990) [44] showed that a roughness surface profile provided by sandblasting leads to maximum gains of tensile bond strength. Also the moisture condition of the concrete substrate surface prior to overlay plays a key role on the development of the bond strength, even though such a phenomenon is still controversial. In fact, Beushausen (2010) [5] and Vaysburd et al. (2016) [50] stated that a “dry” substrate condition prior to overlay leads to better performances of the bond than “saturated-surface-dry” (SSD) conditions; in certain cases, SSD treatment was even detrimental. De la Varga et al. (2015) [49] and Lukovic and Ye (2016) [36] claimed that SSD condition provides the best bond strength. Bissonnette et al. (2014) [9] suggested that the optimal saturation level ranges from 55% to 90%.

Current specifications in the concrete repair technology suggest that bond strength is defined as the tensile strength measured at the interface (mode I) via “pull-off” tests [56,57]. However, test results can be affected by both eccentricity in the load application and damage during the coring. A solution can be found by using a “direct tension” test [35]. Both tests are limited to the fact that if the bond strength is higher than tensile strength of bonded materials, the failure will not be at the interface, and recorded data will be useless. In many practical cases the interface is subjected to pure tension only at the small zones close to edges. By contrast, shear stresses (mode II) occur along the entire interface, e.g. in composite slabs subjected to bending loads. For this reason, shear test methods have been developed as well [45], even though none of these has been accepted as standard. The “slant shear” test is the most used; the set-up is easy, the reliability of the results is good. Nevertheless, unrealistic loading conditions are applied to the interface. The failure of the interface depends on the angle of the plane with respect to the load3. In addition, the test is relatively insensitive to the surface preparation and roughness, since bond failure occurred only for smooth surfaces [2,17]. A more realistic loading condition is reproduced via “lateral shear” tests, but the presence of a bending moment at the interface, due to the shear force eccentricity, promotes the development of peeling stresses which affect the shear strength. In order to prevent such an inconvenient, Silfwerbrand (2003) [45] developed a “twist off” test, although, according to the theory of brittle material strength, the failure plane is not parallel to the torque plane, but it has an inclination around 45°. In the case of bonded materials sub-jected to the torsion torque, the plane of failure does not correspond with the plane of the interface. In fact, experimental results reported in previous works [7,45] confirmed such geometric incompatibility. A different test method, named “direct shear”, solved the problem of geometrical incompatibility observed in the twist-off test. In addition, the fact that the load shear was directly transmitted along the interface permitted to reduce the bending moments and tensile forces arising at the interface [6].

In the practice, cracking/debonding between substrate and overlay propagates in a mixed mode of stresses at the interface [24]. Such an aspect is not properly taken into account by current test methods, which could overestimate the bond capacity. Only one concerning investigation was found in Literature [2]. In such a work authors attempted to define an empirical bond failure envelope concept for normal concrete repairs, by supplementing pull-off and slant-shear data results. But, as also discussed above, slant shear method proved to have serious shortcomings, as few cases of bond failure were recorded. In the lack of data, few existing models propose theoretical approaches to predict the mixed mode behavior, but they were fitted on specific cases and lack in generality [24].

The aim of this work is to develop a cohesive model able to predict the loading bond capacity of retrofitted concrete structures. In particular, the model can predict the load-slip behavior of bonded materials subjected to mixed mode stresses, by taking into account both the moisture conditions of the substrate prior to the application of the overlay and the properties of the latter. The relationships for mode I, mode II and their coupling factor were calibrated according to bond tests specifically designed by authors. An independent experimental investigation permitted to validate both the proposed model and highlight the different cracking/debonding patterns observed in the system “overlay-interface-substrate” by changing the properties above discussed. In order to properly reproduce the rehabilitation in the practice, the hydro-jetting method was adopted for preparing the substrate prior to overlay. The roughness profile was carefully analyzed. A description of the experimental program is provided in Section 2; in Section 3 the experimental results are discussed; in Section 4 the cohesive model is presented and theo-retical results are compared with the experimental data; finally, conclusions are drawn in Section 5.

Section snippets

Materials and methods

In order to characterize the loading bond capacity of retrofitted composite concrete structures, several concrete slabs were cast, exposed to weather conditions for 90 days and then subjected to the surface treatment by hydro-jetting. The roughness profile of the surface was measured by photo-scanning. Then, the substrate was prepared to the application of the overlay. Two commercial fiber-reinforced-concretes were used as overlay, one HPFRC and one UHPFRC. After 28 days of curing, the

Experimental results

Specimens tested in this investigation were labeled according to the type of test carried out, type of overlay and moisture condition level of the substrate prior to overlay. In particular:

  • P, DT, DS, LS, TS: stand for pull-off, direct tensile, direct shear, lateral shear and tensile/shear test;

  • Dry, 75, SSD: stand for dry substrate, moisture surface level of 75% and saturated-surface-dry;

  • A, B: stand for overlay A (HPFRC) and overlay B (UHPFRC).

By varying the parameters above, thirty series were

Cohesive model

The cracking/debonding encountered in the practice near to discontinuities of the overlay was simulated by lateral shear bond tests on concrete slabs repaired by different overlays, see Section 3.5. In the experimental test, a concrete slab 200 mm thick was reinforced by 50 mm of HPFRC/UHPFRC. The edge side of the overlay was subjected to an incremental lateral shear load, until the cracking/debonding failure occurred along the interface zone. Experimental results confirmed that the majority of

Conclusions

In this work a cohesive model was developed to predict the loading bond capacity of retrofitted concrete structures, whose interface is subjected to mixed mode. The experimental investigation permitted to validate both the proposed model and highlight the different cracking/debonding patterns observed in the system “overlay-interface-substrate”, by varying both the moisture conditions of the substrate prior to the application of the overlay and the properties of the latter.

In such a model, mode

Declaration of competing interest

The authors declare that they have no conflict of interest.

Acknowledgments

Authors gratefully acknowledge the financial support provided by HEIG- VD (Haute Ecole d'ingénieurs et de gestion du canton de Vaud - Switzerland- ). Financial support from the Italian Ministry of Education, University and Research (MIUR) in the framework of the Project PRIN 2017 “Modelling of constitutive laws for traditional and innovative building materials (code 2017HFPKZY) is gratefully acknowledged.

References (67)

  • L. Lanzoni et al.

    Performance evaluation of a polypropylene-based draw-wired fibre for concrete structures

    Construct. Build. Mater.

    (2012)
  • L. Lanzoni et al.

    A loaded Timoshenko beam bonded to an elastic half plane

    Int. J. Solid Struct.

    (2016)
  • L. Lanzoni et al.

    Damaged hyperelastic membranes

    Int. J. NonLinear Mech.

    (2014)
  • L. Lanzoni et al.

    A simple nonlinear model to simulate the localized necking and neck propagation

    Int. J. NonLinear Mech.

    (2016)
  • A. Nobili et al.

    Experimental investigation and monitoring of a polypropylene-based fiber reinforced concrete road pave- ment

    Construct. Build. Mater.

    (2013)
  • V. Savino et al.

    Tensile constitutive be- havior of high and ultra-high performance fibre-reinforced concretes

    Construct. Build. Mater.

    (2018)
  • V. Savino et al.

    Simple and effective models to predict the compressive and tensile strength of HPFRC as the steel fiber content and type changes

    Composites Part B

    (2018)
  • V. Savino et al.

    An extended model to predict the compressive, tensile and flexural strengths of HPFRCs and UHPFRCs: definition and experimental validation

    Composites Part B

    (2019)
  • B.A. Tayeh et al.

    Mechanical and perme- ability properties of the interface between normal concrete substrate and ultra high performance fiber concrete overlay

    Construct. Build. Mater.

    (2012)
  • R. Walter et al.

    Debonding of FRC com- posite deck bridge. Brittle Matrix Composites 7

    Proceedings of an Inter- National Symposium

    (2003)
  • H. Yin et al.

    Numerical model for predicting the struc- tural response of composite UHPC–concrete members considering the bond strength at the interface

    Compos. Struct.

    (2019)
  • H.D. Beushausen

    Long-term Performance of Bonded Concrete Overlays Sub- Jected to Differential Shrinkage

    (2005)
  • H. Beushausen

    The influence of concrete substrate preparation on overlay bond strength

    Mag. Concr. Res.

    (2010)
  • H. Beushausen et al.

    Bond strength development between con- cretes of different ages

    Mag. Concr. Res.

    (2008)
  • B. Bissonnette et al.

    Development of specifications and performance criteria for surface preparation based on issues related to bond strength

    Final Report ST-2017-2886-1

    (2017)
  • Bissonnette B, Nuta A, Morency M, Marchand J, Vaysburd AM. Concrete repair and interfacial bond: influence of surface...
  • B. Bissonnette et al.

    Moisture Content Require- Ments for Repair, Part 1: Concrete Repair Testing. Report Number MERL-2013-63

    (2014)
  • E. Brühwiler et al.

    Rehabilitation of Concrete Structures Using Ultra-high Performance Fibre Reinforced Concrete. The Second Interna- Tional Symposium on Ultra High Performance Concrete

    (2008)
  • G. ChunPing et al.

    Ultrahigh performance concrete- properties, applications and perspectives

    Sci. China Technol. Sci.

    (2015)
  • N.J. Delatte et al.

    Laboratory and field testing of concrete bond development for expedited bonded concrete overlays

    ACI Mater. J.

    (2000)
  • N.J. Delatte et al.

    Bond strength development with maturity of high-early-strength bonded concrete overlays

    American Concrete Institute

    (2000)
  • D. Godaire et al.

    Bond Quality of Fiber Reinforced Polymer Concrete Strengthening Systems

    (2005)
  • J.-L. Granju

    Debonding of thin cement-based overlays

    J. Mater. Civ. Eng.

    (2001)
  • Cited by (10)

    • Interfacial shear performance of UHPC-strengthened RC structures using post-installed rebar connections: Experiments and analyses

      2023, Engineering Structures
      Citation Excerpt :

      Given the construction defects in this specimen, it is reasonable to conclude that the rebars were pulled out in specimen C-L-1. The previous studies [47–50] indicate that UHPC exhibits high adhesive strengths when it is used to attach the RC segment. In this study, specimens CN were used to explore the adhesive effect (without post-installed rebar).

    • Numerical insights on quasi-static behaviors of UHPC-NC composite members by a phase-field approach enhanced with a cohesive-frictional interface model

      2022, Composite Structures
      Citation Excerpt :

      As a compromise, combining UHPC and NC materials to form composite members attracts scholars and engineers in the fields of retrofitting existing structures [15,56], enhancing durability [4], accelerating construction speed [69], etc. These composite members, albeit cost-effective, inevitably face the risks of debonding failure at the UHPC-NC interface [58,57]. Interfacial debonding is very brittle and leads to the premature failure of composite members, as well as complex failure patterns.

    • A 2D generic multi-surface cohesive zone model for simulating FRP-to-concrete mixed-mode debonding failure

      2022, Composite Structures
      Citation Excerpt :

      The failure may also occur within the adhesive if it is weak, as shown in Fig. 1d; however, this phenomenon happens less frequently, except under some extreme conditions such as in a high-temperature environment where the bond strength will degrade with the increase of temperature [7]. Recent studies show a similar mechanical response of joints in cementitious members [8–10]. In addition to a larger number of analytical studies (e.g., [11–16]), numerical modelling has been widely used to comprehend the bond failure mechanism of FRP detached from concrete substrate.

    • Efficiency of high performance fiber reinforced cementitious composites as a retrofit material for fire-damaged concrete

      2022, Materials Today: Proceedings
      Citation Excerpt :

      Subsequent advancements in high strength concrete have seen the development and widespread usage of high performance fiber reinforced cementitious composites (HPFRCC) in civil engineering applications. The attraction of this material can be comfortably justified considering its remarkable advantages, such as its superior mechanical properties, greater strain capacity, and better durability performances compared to conventional concrete [25–35]. In addition, this material exhibits excellent interfacial bond strength with fire-damaged concrete [36–38].

    View all citing articles on Scopus
    View full text