Search

In this paper a methodology for studying the mechanical behaviour of complex structures made of GFRP structural members joined by means of adhesive connections and subject to variable intensity loads is presented. The fundamental equations are derived by an asymptotic approach applied on composite structures made of two elastic solids bonded together by a third thin one represented by a thin layer with a nonlinear behaviour . The adhesive layers are considered micro cracked according to the Kachanov’s assumptions. Within this framework, to calibrate the parameters of the imperfect interface model, the mechanical properties and damage evolution of an epoxy adhesive have been experimentally evaluated under cyclic loadings. The experimental tests have been performed on aluminium cylinders considering different thicknesses of adhesive. The experimental results evidence how the adhesive thickness influences the strength, stiffness and consequently the initial damage parameter (initial cracks) of the bonded connections. Finally, the robustness and accuracy of the imperfect interface model is demonstrated by the excellent comparison with experimental results of a GFRP hollow column to built-up beam adhesive connection, under static, cyclic and fatigue loads using a description of damage evolution.

In the field of external strengthening and repairing of existing reinforced concrete structures, steel reinforced polymers (SRP) systems have emerged as a competitive alternative to the use of the more common carbon and glass fiber reinforced polymer (FRP) composites. Experimental investigations have frequently shown the potentials of these innovative composite systems in improving the performance of deficient structural members. At the same time, additional studies are needed to expand the existing knowledge and either to provide design recommendations or to develop specific guidelines. The paper fills some of the foregoing knowledge gaps by presenting an analytical investigation on the bond behaviour between SRP and concrete in which closed-form solutions are derived to predict the entire debonding propagation process. In particular, accurate and simplified local shear stress-slip (τ-s) laws are employed in the proposed modelling from which different expressions for the interfacial shear stress distribution, the axial stress profile and the concrete-SRP relative displacement (slip) are developed and commented in the paper; analytical estimates of the SRP effective bonded length are provided as well as relationships for calculating the maximum axial stress (or peak force) at SRP laminate debonding. The analytical procedure was, firstly, applied to simulate some single-lap shear tests performed in a previous experimental program with the purpose to investigate the influence on the debonding propagation process of the following main parameters: a) concrete strength, b) concrete surface finish and c) steel tape density. Then, the comparisons between the numerical simulations and the experimental results available for some bond tests have allowed for verifying the accuracy of the proposed modelling.

A novel semi-analytical formulation of the governing equations of an FRP retrofitted reinforced concrete beam under bending, based on the concept of moment-curvature, is presented and it is applied to compute the complete nonlinear load-deflection curve, FRP-concrete interfacial shear stress, the FRP laminate strain and the debonding load of statically determinate RC beams retrofitted by an adhesively bonded FRP laminate. The beam may be subjected to any loading, including uniformly distributed or discrete point load(s). The governing equations satisfy the equilibrium and strain compatibility requirements of the beam but slip at the FRP-concrete interface is permitted. A novel procedure for constructing the full nonlinear moment-curvature relationship of beams undergoing interfacial slip is presented in detail, but it is shown that practically a trilinear relationship is satisfactory for obtaining an accurate estimate of the beam deflection and stresses. The robustness and accuracy of the method are demonstrated by analyzing several beams tested by other researchers under uniformly distributed load or four-point bending, with remarkably close agreement between the experimental and corresponding computed values.

Steel Reinforced Polymer (SRP) systems have recently emerged as an attractive solution for the external strengthening of reinforced concrete structures. They entail unidirectional fabrics made out of high tensile strength steel cords that can be externally bonded to the substrate via wet lay-up, using either epoxy or polyester resin. Currently, research about the behavior of SRP strengthening systems for concrete structures is evolving but further systematic and comprehensive studies are still needed to ensure the consistency and reliability of the studies performed to date. The present paper contributes to expanding the existing knowledge by presenting the results of an extensive experimental program devoted to investigate the bond behaviour between SRP and concrete substrate. To this purpose, a number of SRP strips were bonded to concrete blocks by a thixotropic organic matrix and the lap joint was subsequently subjected to direct shear tests performed in displacement control. Besides the concrete strength, the following study parameters were considered: (a) the concrete surface roughness in the bonded region, (b) the density of the steel fabric, (c) the ratio of the epoxy covered concrete surface width to the SRP strip width, and (d) the bonded interface length.

Adhesive durability and joint reliability, strictly related to the bonding agents, are key parameters still under evaluation in civil as well as in other engineering fields. Moisture, different environmental agents and temperature (in particular) can strongly affect the performance of the adhesive joints over the time limiting their applicability. The environmental temperature may exceed the glass transition temperature (T g ) of the adhesive formulation entailing relevant changes in its properties, determining, for instance, a transition from a hard to a rubbery behaviour, thus compromising its specific application. Furthermore, due to changes of the temperature values, the structural adhesive can be naturally subjected to a delay or increase in the curing degree. Hence adverse or positive changes in strength and stiffness can be manifested. Within this framework, the topic of the present paper is the study of the hygro-thermal durability of two commercial epoxy resins, suitable for civil engineering applications , respect to the immersion in tap water and sea water for a period of fifteen months at the temperature of 30 °C. To this scope a wide experimental program was developed comprising both End Notch Failure (ENF) tests on the adhesive samples (adherent in glass fiber reinforced polymer , GFRP) for evaluating the pure fracture energy in Mode II of the resins and the water absorption tests for resins and GFRP materials. In general, the results, in terms of fracture energy, show an initial increment (first three-four months) followed by a decrement up to the reaching of a plateau (in the ninth-twelfth month of conditioning). For what concerns the water absorption, the results show that the equilibrium value of both resins is reached in about one month, while that of GFRP samples depend on the type of liquid: three months for tap water and about five months for sea water. Considering the lower activity of the seawater, the longer time to reach the equilibrium value was an expected result.

In most full-composite structures nowadays, pultruded profiles are joined together to form more complex cross-sections that are not produced by the pultrusion process. These unconventional cross-sections represent a critical point in terms of strength and deformability of the whole structure, especially due to the presence of adhesive. This fundamental unknown aspect recently inspired the authors of the present paper to study the mechanical behavior of a glass fiber reinforced polymer (GFRP) profile with a complex cross-section shape obtained by bonding simple pultruded plates (bonded beams) by epoxy resin. The first step was to compare the behaviors of bonded beams with those of similar pultruded ones. With this aim, in a previous paper the flexural global behavior of an I-bonded beam was investigated, highlighting its better performance with respect to the pultruded I-profile in terms of failure load and stiffness. Now the authors would like to extend the study, both experimentally and numerically, to the mechanical behavior of the adhesive web–flange connections in terms of axial stiffness. Seventeen full specimens were tested by performing pull-out tests under force control. A uniformly distributed load was applied to the lower flange in order to favor the detachment of the flange from the web of the profile by using a steel device designed and realized ad hoc . The overall result of the research is that the bonded profile shows a good mechanical behavior justifying its use in technical practice. In addition, the bonded beams represents a valid alternative to the pultruded profile especially when the latter is not available on the market. In more detail, in bonded beams, the presence of the resin (instead of resin and fibers as in pultruded beams) at the web-to-flange connection does not affect the flexural global behavior but does penalize the axial behavior of their connections, as here demonstrated.

Relatively recently a steel fabric/laminate has been proposed for externally strengthening concrete structures, using a polymeric resin. The steel fiber-polymer composite system is termed Steel Reinforced Polymer (SRP). To determine the ultimate load capacity of an SRP retrofitted concrete structure, one must accurately predict the SRP-concrete interface debonding load, which requires a robust local bond-slip model. Many design guidelines recommend mode II interfacial fracture energy limit as the failure criterion for FRP-concrete interface. For SRP strengthened members, a suitable constitutive law and failure criterion have not been established yet. Consequently, in this study the applicability of five existing bond-slip interface models for FRP-concrete interface to SRP-concrete interface is examined. The models’ parameters are calibrated for SRP-concrete interface using an experimental database by the present authors and compared with the values suggested by the original authors for FRP-concrete interface. The database involves results of tests on concrete prisms bonded to SRP strips. The experimental interfacial bond-slip relationship for the former interface is observed to have a more precipitous descent after the peak stress than predicted by the existing models; consequently, a new model is proposed here to capture this phenomenon. All the models are calibrated using a classical technique which minimizes the difference between the measured and computed interfacial shear stress values at different slip levels. The results indicate that all the models predict relatively well the slope of the ascending branch of the shear stress-slip curve, but they give substantially different values for the maximum shear stress attainable and noticeably different descending branch profiles. Among these, overall, the proposed model is in relatively better agreement with the experimental results.

The results of geometrically nonlinear analyses on 43 built-up Pultruded Fibre-Reinforced Polymer (PFRP) columns with closely spaced chords and intermittent interconnections are presented. A comparison between columns with the end sections entirely loaded and columns loaded at the end battens only is reported, showing no appreciable difference in the P-δ response. The effects due to variations of column length and battens spacing are then investigated. It is found that stocky columns with small battens spacing attain pre-buckling failure at the web-flange junctions of the chords for loads approximately equal to 70% of the crushing load. Slender columns fail by global buckling, whereas intermediate-slenderness columns may experience interaction between local and global buckling. A design method is finally proposed.

International audience; This paper presents some experimental results dealing with the mechanical performance of composite beams obtained by bonding Glass Fibre Reinforced Polymer (GFRP) rectangular pultruded panels by means of an epoxy structural adhesive. The flexural response of these bonded beams was compared with those obtained by the pultrusion process with the same geometrical and material properties. As a matter of fact, no significant loss of performance emerged in terms of failure load; moreover, an increase of pre-failure stiffness was observed. This result may allow us to consider bonded GFRP beams as a viable simplification within the field of composite structures.

Steel Reinforced Polymer (SRP) systems have recently emerged as attractive solutions for external strengthening and repairing of existing structures. Experimental studies have frequently shown the potentials of these innovative composite systems in improving the performance of concrete and masonry structures. However, additional studies are needed to expand the existing knowledge and either to provide design recommendations or to develop specific guidelines. The paper fills some of the foregoing knowledge gaps by discussing the experimental results of 130 direct single-lap shear tests performed to investigate the bond behavior between the SRP reinforcement and the concrete substrate; different concrete surface finishes were investigated, i.e., sandblasted, bush hammered and grinded which are among the surface treatments mostly used in practice. Comparisons between the maximum load resisted by the SRP-concrete joints and the theoretical value calculated using some relationships available for similar FRP-concrete assemblages are discussed. Then, the experimental results relative to SRP-concrete joints with sandblasted finish were used to: a) calibrate the parameters of some bond-slip interface models available in the literature for FRP-concrete interface, and b) compare the obtained relationships with the interface models proposed in a previous study for SRP-concrete joints with bush-hammered and grinded concrete surface finish only. The bond-slip models calibrated for the three considered concrete surfaces were finally used to develop theoretical relationships between the maximum bond stress in the steel strip at debonding and the concrete strength. Comparisons with the experimental data were also performed.

International audience; In this paper a methodology to model a GFRP adhesive connections by using an imperfect soft interface model is presented. The model based on Kachanov's theory considered a cracked thin adhesive. Within this framework, the mechanical properties and the initial damage (diffuse initial cracks) of the adhesive layer has been experimentally evaluated. With a modified Arcan system, static tests were performed on adhesively bonded assemblies in tensile and shear solicitation mode considering three different adhesive thicknesses. The experimental results highlighted how the thickness of adhesive influences the mechanical strength and stiffness of the bonded connection. CT-scans were performed to measure the porosity rate in the adhesive layer. Furthermore, the excellent comparison of numerical and experimental data of an adhesive GFRP bonded connections allow us to consider the imperfect soft interface model proposed as highly competitive to evaluate complex structure performance in civil engineering context. A parametric analysis has been proposed to provide a formula able to describe the full response of the structure at varying adhesive property.

This study presents the results of an experimental investigation on the durability of a commercial High Performance Fiber Reinforced Concrete (HPFRC) subjected to rapid freeze/thaw cycles. High-performance fiber reinforced concrete (HPFRC) is currently being used in the construction industry both for strengthening existing members and for controlling opening cracks in new ones. In literature there are many studies focused on the evaluation of the HPFRC mechanical properties varying the matrix composition and/or the type and quantity of the fibers but a lack of knowledge exists for what concerns its durability. Durability of concrete may be defined as the ability of concrete to resist weathering conditions, chemical attack and abrasion while maintaining its desired engineering properties. In particular, durability can be defined by the percentage ratio of the dynamic modulus of elasticity after a number of freeze and thaw cycles to the corresponding value before the freeze and thaw cycles. In this paper, three different concrete mixtures with various fiber volume fraction , i.e. 0%, 1.25% and 2.50%, were examined. Seventy-five freeze-thaw cycles were performed according to UNI 7087-2017 on five prismatic specimens per each concrete mixture while other three specimens were selected as reference samples. In order to evaluate the durability factor of the three mixture, the resonant frequencies (transversal, longitudinal and torsional) and the dynamic moduli of elasticity were measured using the Impact Resonance Method (IRM) provided by the ASTM C215.

A semi-analytical model is developed to determine in FRP retrofitted reinforced concrete (RC) beams the interfacial shear and peeling stresses, the FRP laminate and the RC section strain and stresses at all loading stages up to failure. The FRP is assumed to be externally bonded to the beam but can undergo slip and relative vertical displacement at its interface with the concrete. The model is developed by satisfying the requirements of equilibrium and strain compatibility while concurrently allowing for interfacial deformations. FRP is treated as a linear elastic, steel as elasto-plastic strain hardening and concrete as fully nonlinear material in compression and tension, including tension stiffening. The governing equations are formulated as two second order differential equations with their dependent variables being the strain in the FRP and the relative normal displacement of the interface. The equations are solved for discrete states (uncracked, cracked, yielded) experienced by the RC section and their associated level of interfacial slip. The model results are compared with available experimental results for several beams retrofitted with carbon FRP or steel reinforced polymer (SRP) laminates subjected to either four point bending or simulated uniform load, with satisfactory agreement between them.

Composite materials consisting of high tensile strength steel cords embedded in a cementitious matrix (Steel Reinforced Grout, SRG) are becoming an attractive solution for strengthening existing structures thanks to the lower cost than “traditional” FRP systems, the isotropy of the steel and the ability of the mortar to resist fire. With respect to Steel Reinforced polymer (SRP) systems – whose mechanical behavior is predominantly dependent on the strength of the substrate to which the steel fibers are epoxy bonded – the mechanical properties of SRG systems are mainly dependent on the cement-based matrix which is sensitive to curing the installation conditions and methods. The experimental results presented and discussed in this paper contribute to expanding the existing knowledge on the bond behavior between the SRG and concrete substrate. To this purpose, a number of SRG strips were bonded to concrete blocks by an inorganic matrix and the lap joint was subsequently subjected to direct shear tests performed in displacement control. Since mortars are very sensitive to environmental conditions and humidity as previously mentioned, two different curing conditions were used. In the first case – termed “curing a” – specimens were cured at room temperature for 28 days and wet cloths were placed on top of the composite surface each day for two weeks. In the second case – termed “curing b” – specimens were just cured at room temperature for 28 days. In addition to the curing condition, the following study parameters were considered: (a) the concrete surface roughness in the bonded region, (b) the density of the dry steel fabric, (c) the bonded interface length, and (d) the concrete strength. Relatively to the latter parameter, concrete prisms were broadly divided into two strength groups: Normal Strength Concrete (NSC) having an average concrete cylindrical strength between 13 and 25 MPa and High Strength Concrete (HSC) having an average strength equal to 40 MPa. Experimental evidence underlines that SRG systems mainly exhibited a debonding failure at steel fibers-matrix interface (due to the sliding phenomena) irrespective of the concrete strength and surface finish. Furthermore, like the SRP systems, the low tape density showed better performances with respect to the higher densities considered in this paper, both in terms of the efficiency factor (i.e., the ratio between the tensile stress experienced by SRG in each test and the average tensile ultimate strength of dry steel fabric) and of the maximum force transferred by the system.

Glass Fiber Reinforced Polymer (GFRP) I-beam-column adhesively bonded connections are tested under combined bending and shear. The special feature of the novel connection is the wrapping of the seat angles at the connection by a carbon fiber reinforced polymer (CFRP) fabric wrap. The wrap is primarily intended to alter the connection failure mode from brittle to pseudo-ductile, thus providing adequate warning of impending failure. Four moment resisting connection configurations are tested, including the reference configuration without the wrap. It is observed that the connection failure is initiated by the fracture of the adhesive, but the provision of the wrap, together with a steel seat angle, alters the failure mode from brittle to pseudo-ductile. The post-peak load deformation is achieved without a large drop in the resistance of the connection. On other hand, the connection with the wrapping and a GFRP seat angle can also change the failure mode to pseudo-ductile, but it could not be done without a large reduction in the connection resistance after the peak load.

For the first time, the feasibility of adhesively bonded connections in FRP frame structures is explored as an alternative to bolted connections. Eight full-scale GFRP beam-column connections are tested and their failure mode, strength and rotational stiffness are investigated. A single pultruded GFRP I-profile is used for the two members. In four of the specimens the beam and the column are connected by epoxy adhesive and GFRP seat angles, similar to the so-called “standard bolted connection”. In the remaining four specimens, the seat angles are supplemented by additional GFRP angles and stiffeners to strengthen the column flange and web. The beam-column assembly forms an inverted L-shape frame, with the column being fixed at the bottom and attached to the beam near the top. The beam, acting as a cantilever, is loaded by a point load near its free end, which subjects the connection to bending and shear. The current standard connection failed by debonding within the column flange while the improved/strengthened connection failed within the adhesive or at the adhesive-column flange interface. The test results reveal that both the standard and improved connection can have at least the same strength as the corresponding bolted connection, irrespective of whether GFRP or steel bolts are used to make the connection. Hence, the current restrictions against the use of adhesive beam-column connections in GFRP frame structures may be unjustified. In making this comparison, the observed failure load of each connection is normalized by the ultimate moment capacity of the GFRP profile in the beam-column assembly.

A new adhesive beam-column connection is tested which possess the highest strength and stiffness compared to any other similar adhesive or bolted connection tested in the past. A square GFRP hollow section, acting as a column, was connected to a built-up beam made of two GFRP U-profiles by means of either epoxy or steel bolts. The beam-column assembly formed an L-shaped frame which was tested by applying a point load at the beam free end while the column was fixed at its base. Five bolted and five adhesive replicate connections were subjected to quasi-static loading up to failure. Another three adhesive connections were subjected to 400, 800 or 1200 cycles of loading and unloading with the maximum load being equal to 0.50 Pu,avg, where Pu,avg is the average static strength of the replicate adhesive specimens. At the end of the cyclic loading, the latter specimens were loaded quasi-statically to failure. Finally, another two adhesive connections were subjected to fatigue type loading. They were successively subjected to at least 196 cycles of loading and unloading with the load amplitude being 0.50 Pu,avg in the first 60 cycles, 0.75 Pu,avg in the next 60 cycles, 0.85 Pu,avg in the following 60 cycles and 0.95 Pu,avg after the 180th cycle. The test results show that the proposed adhesive connection can achieve on average 82% higher strength and 380% higher rotational stiffness than the companion bolted connection. Furthermore, the above cyclic loading has negligible effect on either the strength or the stiffness of the connection. Finally, the connection can sustain the foregoing fatigue load up to almost 180 cycles without significant damage but it will not be able to withstand the full 60 cycles of the load with 0.95 P u,avg amplitude. The current results demonstrate the superior strength and stiffness of the new adhesive connection compared to a similar bolted connection.

When dealing with composite laminate bonded to pre-existing members (masonry, concrete, metal, wood), the performance at collapse of the adhesion is substantially given by two features: the failure load and optimal bonding length. The last one can be defined as the length required for ensuring the maximum value of the transferred load between the laminate and the substrate. Within the context of structural strengthening, many technical design formulae are available for predicting both the failure load and optimal bonding length, but they have been generally derived for special cases only. There is not a unique formula with a general validity regardless of the nature of the substrate (concrete, masonry, metal, wood). In this paper, a mechanical model previously formulated by the authors is used to derive closed form predictive equations for the failure load and optimal bonding length which have, instead, a general validity due to the circumstance that they account for the cohesive parameters of the interface as well as the mechanical properties of the composite laminate. Comparisons with results obtained by other authors are also given with reference to the composite–concrete interface.

International audience; This paper presents a wide numerical investigation about the influence of an external reinforcement of the web/flange connection on the performance of beams obtained by bonding simple panels to each other by means of epoxy resin.Within this framework, two different models were developed: a 2D-model and a 3D-model. The 2D-model was used to understand the role played by the bonding joints in the experimental tests previously performed and to understand the influence of different types of reinforcements on the stress distribution in the adhesive layer. The 3D-model was implemented in order to study the influence of different web/flange reinforcements on the mechanical response of bonded beams and to predict their failure load and flexural stiffness.The numerical results have shown how it is possible to obtain, for these new bonded beams, levels of performance higher than those of the current pultruded beams, allowing us to consider the bonding system proposed as highly competitive in the field of construction of pultruded profiles.

The first results of a numerical and experimental investigation on the shear forces distribution in a bolted joint made entirely from FRP materials are presented. It is also proposed an experimental equipment for investigating the strains and stresses distributions around the holes of the connection as well as the bearing stresses at the interface between plate and steel bolt. A good agreement between numerical and experimental results allows to use the proposed testing set-up for analysing the bearing failure of several joint configuration with different lamination scheme, geometry and type of load.

In this paper a finite element analysis of the behavior of adhesive joints between FRP (fiber-reinforced polymers) adherents is presented. In particular, double-lap joints, in the case of both normal and shear/flexure stresses, are considered. The problem is nonlinear due to the mathematical formulation of the cohesive laws introduced to model the interfacial interactions. On the contrary, the adherents are supposed to be linear elastic up to failure. The common approach of analyzing shear/flexure behavior separately from extensional behavior, as well as disregarding the mutual effects between the normal and tangential stresses acting at the joint interfaces, has been updated. The paper takes into account the coupling effects between shear/flexure and extensional equilibrium problems. As highlighted in literature, only a coupled analysis can be considered correct when the mechanical characteristics of the adherents are quite different, as usually occurs in practice. Furthermore it also has been assumed th...

In this paper the debonding of adhesive lap joints between FRP adherents is analyzed with regard to the influence of the chosen interface failure criterion. In particular, the results obtained by using the well-known criteria of Hutchinson and Suo, Xu and Needleman, and Camacho and Ortiz, are compared. Due to the mathematical formulation of such cohesive interface models, the examined mechanical problem is non linear, although the adherents are supposed to be linear elastic up to failure. Furthermore, shear deformability of the adherents is taken into account. The numerical results of a finite element analysis are presented and discussed. Suggestions for the design of adhesive lap joints are also given.

In this paper the results of a comprehensive numerical investigation regarding the axial–flexural–torsional response of pultruded slender beams is presented. The goal is to propose GFRP standard cross-sections of such proportions and shapes that would possess improved strength, stability and deformational characteristics compared to the corresponding existing sections whose proportions are generally based on standard steel sections. As GFRP sections are thin-walled but are significantly less stiff than similar steel sections, the study focuses on enhancing their appropriate stiffness and buckling strength. The novel and efficient numerical model used in this investigation was developed by the writers and can be used to trace the complete pre-buckling geometrically nonlinear response of any GFRP or steel thin-walled member with open or closed cross-section. The bucking load is computed by the asymptotic value of the load–displacement curve. Members with I-, L-, T- and box sections are analyzed, considering different loading and boundary conditions. It is demonstrated that due to their unsuitable proportions, available standard GFRP sections do not have adequate stiffness and buckling strength. Consequently, recommendations are made for new sectional proportions and modified shapes, and some graphical results are presented to demonstrate how the results of the proposed method could be utilized in practical design situations. The superiority of the proposed sections is quantified by an efficiency factor, defined in terms of ratio of strength gain to material volume increase.

In this paper a geometrically nonlinear model for studying the lateral global buckling problem of a generic thin-walled composite beam is presented. The model is based on full second-order deformable beam theory and accounts for axial, flexural, shear, torsional and warping displacements. Moreover, the web/flange junctions are supposed to be deformable. Governing nonlinear equations are derived from the principle of virtual displacements. Initial numerical results, obtained by means of a finite element approximation, deal with open cross-section composite profiles under flexural/torsional loads, making it possible to identify the relevance of the connection stiffness over the pre-buckling range of the mechanical response. Comparisons with numerical and experimental results available in the literature are also discussed.

A micro-scale numerical investigation on the response of adhesive lap-joints between FRP adherents has been performed. Such an analysis allows to compare the prediction of classical theories with those ones obtained by the proposed approach, showing relevant differences in terms of ultimate behavior.