Experimental and Numerical Investigation of Shear Behavior of RC Beams Strengthened by Ultra-High Performance Concrete

Concrete structures need repairing or strengthening when they have some deficiencies in their structural performance and/or durability properties. Such deficiencies could be due to many reasons such as errors in deign calculations or construction practices; unexpected increasing in loads; change in service conditions; deteriorations resulting from corrosion of steel rebars or other chemical attacks, etc. As far as structural performance is concerned, both flexural and shear strengths requirements should be satisfied. RC members are mainly designed to develop their full strength (Altin et al. 2005). However, in some cases the failure of beams can take place due to deficiency in the design for shear. Since the shear failure occurs suddenly and may lead to catastrophic consequences, the necessity of the adequate shear capacity of RC beams should be given due importance. Accordingly, researches pertaining to strengthening the RC beams deficient in shear are reported.

Ultra-high performance concrete (UHPC), which is a hybrid of the cementitious materials and high-tensile strength steel fibers, can be used to strengthen the RC members (Al-osta 2018). UHPC is reported to have outstanding properties such as ultra-high strength, good flowability, excellent ductility, high serviceability, high strength-to-weight ratio, aesthetically appearance through self-levelling property, and overall superior durability properties such as low permeability and highly resistant against reinforcement corrosion (Ahmad et al. 2015). Moreover, the UHPC can effectively be bonded with the sandblasted surfaces of existing old reinforced structures and thus making it suitable for rehabilitation and strengthening of RC structures (Martinola et al. 2010; Al-Osta et al. 2017).

UHPC strengthening system is an alternative approach to rehabilitate or restore the deteriorated concrete members or to retrofit or strengthen the sound concrete members. It has exceptional advantages over traditional methods such as steel plate-bonding (Altin et al. 2005), fiber reinforced polymer (FRP) strengthening (Chen and Teng 2003), section enlargement, etc. For example, FRP possesses desired properties such as high strength, corrosion resistance, ease to apply, and without much change in the size of the structural member. However, FRP system has some shortcomings, which are mainly related to bonding, compatibility and fire-resistance problems. On the other hand, UHPC can be used as a strengthening material for existing structures having either sound or deteriorated concrete surfaces. Therefore, for repairing or rehabilitating of concrete structures, UHPC can be considered as a good option which can enhance the structural performance and durability of substrate concrete (Li 2004).

Throughout the last decade, an attempt was made by various researchers to use the high strength concretes, generally the steel fiber reinforced concrete, for purposes of structural strengthening. The flexural and shear behavior of the RC beams retrofitted using high-performance fiber reinforced concrete (HPFRC) was studied by Alaee et al. (2003). The results of this study indicated the feasibility of using HPFRC for upgrading the flexural and shear capacities of member as well as enhancing the durability properties through the dense mixture of such concrete. Farhat et al. (2007) investigated the behavior of damaged beams strengthened using the high-performance fiber reinforced cementitious composite (HPFRCC). The results showed that if the strengthening is done by applying HPFRCC on the tension face as well as on the side faces, the failure load would increase up to 86%. The strengthening technique using a 40 mm layer of high performance fiber reinforced concrete (HPFRC) was experimentally and numerically studied by Martinola et al. (2010). The results showed that the use of HPFRC jacketing for strengthening has a significant effect in increasing the load carrying capacity by a factor of 2.15. Furthermore, a good enhancement in the durability of the beams was observed due to use of HPFRC jacketing. Noshiravani and Brühwiler (2013) experimentally investigated the composite section made of RC and UHPC. This study concluded that a layer of UHPC applied at the tensile face could be used as an effective shear strengthening. Bastiien-Masse and Brühwiler (2014) investigated the structural behavior of the beams and slabs retrofitted using UHPC. Composite beams and slabs, which included 50 mm thick layer of UHPC were prepared and tested under different types of loading. The results clearly demonstrated that the use of UHPC layer over RC section had an effective enhancement on the load bearing capacity. Ruano et al. (2015) reported the shear behavior of RC beams retrofitted using steel fiber reinforced concrete (SFRC). The results indicated that the presence of fiber prevents debonding and generally enhances overall integrity of the beams. Chalioris et al. (2014) investigated the use of thin reinforced self-compacting concrete for strengthening of conventional RC beams. The results showed an increase in the strength with improvement in the ductility and favorable failure behavior. Hussein and Amleh (2015) evaluated the flexural and shear capacities of beams made with UHPC-normal strength concrete composite without stirrups. The results showed that such composite technique improved the performance of members strengthened in flexure and shear. The benefits of using UHPC for strengthening of conventional RC beams was demonstrated by Lampropoulos et al. (2016). Different configurations of UHPC layers used for strengthening consisted of jacketing of tensile face alone, compressive face alone, and three-side jacketing. A significant increase in the moment capacity was observed in case of UHPC jacketing from three sides. The flexural behavior of strengthened conventional RC beams using UHPC was experimentally studied by Al-Osta et al. (2017). The results showed that the proposed strengthening technique was enhanced the structural performance of retrofitted beams through increasing flexural capacity and overall stiffness.

Various studies are reported about use of finite element modelling for studying the shear behavior of strengthened beams including prediction of failure loads and cracking patterns. However, the modelling of concrete cracking is the most challenging task. Lampropoulos et al. (2016) used the smeared crack approach in the analysis of strengthened beams. Concrete damage plasticity model is most commonly used for simulation of the cracking in concrete (Al-Osta et al. 2017). Al-Osta et al. (2017) developed a finite element model of strengthened beams in flexural using the concrete damage plasticity theory and they found that their proposed model predicted the load–deflection response and the crack patterns in good agreement with the experimental results.

Despite the above mentioned research works pertaining to the use of UHPC in strengthening of RC beams, it can be observed that limited works had considered the jacketing by applying the UHPC along the vertical sides. In addition, there is a lack of information about the effect of UHPC jacketing on the shear behavior of strengthened RC beams. In addition, the effect of the shear span-to-depth ratio on the behavior of strengthened RC beams was not investigated. Consequently, in this research work, the behavior of two different configurations for strengthening of RC beams using UHPC is investigated experimentally. Additionally, a numerical model is developed to validate the experimental results. Moreover, the failure behavior of the non-strengthened and strengthened RC beams was studied considering the experimental variables.

In addition to the experimental investigation, a numerical modelling was carried out using finite element method to study the shear behavior of the strengthened beams. The main purpose of the numerical modelling in the present study was to confirm the sufficiency of the experimental data for highlighting the shear behavior that included depiction of load versus deflection plots, failure loads and cracking patterns.

The numerical modeling was developed using the Abaqus finite element analysis software. The finite element model consisted of modelling the geometry of elements with their materials and related constraints, such as boundary conditions, applying loads and the contacts between the different surfaces. The normal concrete, UHPC, and steel-plates were modeled using the three-dimensional eight-noded brick elements. Whereas, the reinforcement steel (longitudinal and transverse) were modeled with two nodes 3D truss elements. The bond between concrete and reinforcement steel was modeled as an embedded region, whereas the concrete is considered as the host element. The bond between normal concrete and UHPC was considered as perfect-bond because during all experimental tests there was no debonding observed. The steel plates were bonded to the concrete surfaces with tie-bond. The concrete damage plasticity model was used, which is reported to give reliable results (Sümer and Aktaş 2015). Accordingly, by using such model, the complete behavior of full-scale strengthened beams can be achieved without conducting any experimental testing of beam.

3.1 Concrete Damage Plasticity Model

The plasticity theory is commonly used in modelling the quasi-brittle materials such as a concrete. However, the use of plasticity theory is suitable only in compression zones. Several models based on fracture mechanics such as: smeared crack model, fictitious crack model, and crack-band theory are used in tension zones (Lee and Fenves 1998). Therefore, an approach is needed that could consider the non-linear behavior of concrete in a single constitutive model. Lubliner and Oliver (Lubliner et al. 1989) formulated a plastic damage model for concrete based on the plasticity theory.

Concrete damage plasticity (CDP) approach develops the constitutive behavior of concrete by presenting the scalar damage variables for both compressive and tensile response as illustrated in Fig. 8. The damage variables in tension and compression are denoted by d_t and d_c, respectively. The values of d_t and d_c range from zero to one. Abaqus user manual assumed zero for undamaged material and one for completely damaged (i.e., loss of stiffness) (Online Documentation Simulia 2016).

Open image in new window

The CDP approach describes mainly two failure mechanisms, tensile cracking and compressive crushing of concrete. The yield surface is governed by two hardening variables ɛ _t ^PL and ɛ _c ^PL , which are associated to the failure mechanisms under tension and compression loading, respectively. The compressive and tensile damage parameters are calculated based on the equations provided by Birtel and Mark (Birtel 2007):

• Compressive damage parameter (d_c):

  $$d_{c} = 1 - \frac{{\sigma_{c} E_{c}^{ - 1} }}{{\varepsilon_{c}^{pl} \left( {1/b_{c} - 1} \right) + \sigma_{c} E_{c}^{ - 1} }}.$$

  (1)

• Tensile damage parameter (d_t):

  $$d_{t} = 1 - \frac{{\sigma_{t} E_{c}^{ - 1} }}{{\varepsilon_{t}^{pl} \left( {1/b_{t} - 1} \right) + \sigma_{t} E_{c}^{ - 1} }}.$$

  (2)

where d_c and d_t are compressive and tensile damage parameters, σ_c and σ_t are compressive and tensile stresses of concrete, E_c is the modulus of elasticity of concrete, ɛ _c ^pl and ɛ _t ^pl are plastic strains corresponding to compressive and tensile strengths of concrete. b_c and b_t are constant parameters, 0 < b_c,t ≤ 1.

3.2 FE Modeling Considerations

The materials (normal concrete, UHPC and steel reinforcement) were modeled using the data generated through the experimental program. The nonlinear behavior in tension as well as in compression of both the normal concrete and UHPC, as shown in Fig. 9, were used. The bonding between different surfaces was modelled using the available options in Abaqus library. The bond between concrete and reinforcement steel was taken as embedded region, where the concrete is the host element. The adhesion between normal concrete and UHPC was considered as perfect-bond because during all experimental tests, debonding was not observed.

Open image in new window

In Abaqus, the most dependable approach of applying the load is the explicit dynamic method. This method is reported to be successful for two main reasons: first, it gives reliable results with less problems of convergence, second, it is the most suitable for materials like concrete to capture the concrete cracks and overall failure behavior (Mercan 2011). Furthermore, in explicit dynamic analysis, the inertial effects can be minimized by either reducing the loading rate or increasing the mass density of concrete in order to approach the static solution. Thus, in Abaqus, the time increments are automatically calculated and the loading rate is set as one second.

All experimental results, including failure load, crack pattern, failure mode and load–deflection curves were compared with those obtained through the FE modelling. This comparison showed that the FE model is able to capture most of the failure modes with good accuracy.