Over the last decades steel reinforced concrete was the most widely used structural material in construction. Nevertheless, it is well known that, under certain environments, the corrosion of steel reinforcement can lead to considerable deterioration or even collapse of structural elements, requesting expensive repair and strengthening or reconstruction activities. This disadvantageous attribution of the steel reinforcement has contributed to the researches started to focus on alternative reinforcements.
FRP (Fibre Reinforced Polymer) materials offer a promising solution since for many years they are successfully used in other industries (such as the automobile and sports manufacturing industries) and more recently in construction. There are structures reinforced with FRP rebars that have been in service in aggressive environments in various parts of the world, for more than 15 years, without considerable structural problems.
The widespread adoption of any new type of reinforcement, such as FRP, requires the development of product specification, testing standards and codes of design practice, which is usually a process that can take many years to be completed. One of the fundamental aspects of structural behaviour is the bond development, since the bond stress transfer between reinforcement and the surrounding concrete is the basis of the theory and applicability of reinforced concrete whenever end anchorages are not used.
Development of new codes requires understanding of bond between FRP rebars and concrete and the relationship between the bond strength and various material parameters as well as test conditions. Despite the large amount of research done in order to understand the bond behaviour of FRP rebars, owing to the high number of parameters which are affecting the bond properties, there is still debate among researchers about the effect of particular parameters. Due to various constituent materials, manufacturing processes and surface treatments of FRP reinforcement, both bond performance and failure of bond can occur in different ways than in the case of conventional steel reinforcement.
There is still no generally accepted model developed for bond behaviour of internal FRP reinforcement in concrete. Consequently, one major expected outcome of this project is to validate existing models and possibly develop improved model. To achieve this aim an extensive literature review and database developing is needed. However, further experimental data might be also necessary for bond models in case of those parameters which still present lack of experiments.
Development of a generally accepted bond model and updating the experimental database with necessary data are considered as the major expected benefits of the project, which would result in increased confidence in use of FRP materials in construction in order to build durable structures. Owing to the benefits of this research project the annual cost of repair and maintenance of the infrastructure could be reduced, which in Europe alone is estimated to be about 50% of the construction budget.
The main aim of present research project is to investigate the bond behaviour of FRP (Fibre Reinforced Polymers) rebars in RC (Reinforced Concrete) elements in order to validate existing bond models between FRP rebars and concrete and if necessary to develop improved design model for the bond of FRP internal reinforcement.
Consequently, the planned main objectives of this research project are:
The bond performance of FRP rebars in concrete depends on a number of parameters like: strength and type of concrete, surface configuration of the rebar, type of fibres, loading as well as service temperature and many others. Although several models have been developed to predict bond behaviour, none of them is generally accepted. Extensive further research is still needed. At the first phase of the experimental work pull-out bond tests are carried out to examine local bond–slip behaviour and develop models that can capture material-related properties as well as loading rate, loading conditions and service temperature.
In Fig.1 experimental details are presented: pull-out test setup (left) for testing of bond behaviour of FRP rebars in concrete; test specimens (right); FRP bars ready for testing (lower middle) and moulds (upper middle) prepared for casting of concrete. Following a preliminary experimental series done both on cylindrical and cubic concrete specimens, the cubic shape was chosen, owing to the difficulty of unmoulding of cylindrical shaped specimens. A soft synthetic material was applied on the surface of FRP bars as bond breaker along the unbonded length.
Three LVDTs (Linear Variable Differential Transformer) are used to measure the slip at loaded end and one LVDT is used at the unloaded end of specimen.
Fig.1 - Test setup for pull-out testing (left), test specimens (right) and FRP bars (lower middle) in moulds (upper middle)
In the first series of experimental work the effect of rebar surface on bond behaviour of FRP rebars was studied in case of two different ribbed and one sand coated FRP rebar surfaces. Furthermore, two different embedment lengths (3db and 5db, where db is the nominal diameter of rebar) were considered. These experiments are considered beneficial, since in available design guidelines the effect of rebar surface characteristic is not taken properly into consideration.
The diameters used were similar (between 14.7 and 16.0 mm), they were all GFRP (Glass FRP) bars with similar modulus of elasticity (55-60 GPa) so the differences in bond strength, failure mode of bond and stiffness of bond can be attributed to the surface characteristics. In addition to the well-known fact, that rebars with larger diameters give lower bond strength values, in Fig.2 and Fig.3 that the highest and stiffest bond strength was reached when rebars with sand coated surfaces were used.
Fig.2 - Effect of surface characteristics on bond strength of FRP bars. Embedment lengh 3*db (left) and 5*db (right) where db is the bar diameter
Fig.3 - Bond stress - loaded end slip diagram. C2 concrete mixture, different surface characteristics
In the second experimental series, the proposed aim was to investigate the effect of different short fibres (e.g.: steel and synthetic) on bond behaviour of FRP rebars in fibre reinforced concrete (FRC), however, owing to numerous concrete mixtures and different FRP bars the effect of concrete strength and the effect of FRP bar fibre type (modulus of elasticity) can be studied as well. In this experimental stage BFRP (Basalt FRP), CFRP (Carbon FRP) and GFRP (Glass FRP) bars were used, along with synthetic micro and macro as well as steel short fibres.
When BFRP bars were used with helical wrapped surface, the also well-known fact that, higher concrete compressive strengths results in higher bond strengths (Fig.4) could be observed, however this was not the case when sand coated CFRP and GFRP where considered. This can be explain with the different failure modes (in case of sand coated bars the failure always occurred by shearing off the surface).
Fig.4 - Bond strength as a function of concrete compressive strength for BFRP bars
There were some small scatters between different concrete mixtures, considering the effect of short fibres, however, generally the steel fibres gave the best results for bond strength. In Fig.5 the effect of different short fibres is visible in case of C1 concrete mixture. The average concrete compressive strength measured on 150 mm cubic specimens at the age of 28 days was 40.3 MPa. In this case synthetic macro and steel fibres practically have the same effect on bond stress of BFRP bars in concrete.
The effect of short fibres is represented in Figure 6, as well. In this case the results of concrete mixture C3 are shown, with an average concrete comressive strength of plain concrete (C31 – no short fibres added) of 50.9 MPa. The highest average bond strength was reached by applying steel fibres.
Fig.5 - Effect of short fibres on bond strength. C1 concrete mixture. Error bars represent the maximum and minimum values
Fig.6 - Effect of short fibres on bond strength (C3 concrete mixture)
The effect of bar type is shown in Fig.7. For all concrete mixtures, CFRP bars had higher average bond strength than GFRP bars. Also, the ascending branch of bond stress - loaded end slip diagram is stiffer in case of CFRP bars. This effect can be explained with the different moduli of elasticity, since they both had similar sand coated surfaces. CFRP bars have higher modulus of elasticity, which results in lower relative displacement (slip) between concrete and FRP bar, which in turn results in better bond stress distribution along the embedded FRP bar and higher bond strength.
Fig.7 - Effect of fibre type (carbon and glass) on bond strength of FRP bars
In Fig.8, different FRP bars are shown after pull-out test. In case of ribbed bar type 1 the failure occurred in concrete (upper left corner) while in case of ribbed bar type 2 the outer surface of bar was peeled off (upper right corner), in case of CFRP and GFRP bars the whole sand coated surface was sheared off (lower left corner), finally in case of BFRP bars (lower right corner) the helical wrapped ribs were partially or totally sheared off, depending on concrete mixture.
Fig.8 - Specimens after failure. GFRP ribbed type 1 and 2 (upper left and right) CFRP/GFRP sand coated (lower left) and BFRP (lower right) bars
Specimens with further concrete mixtures (two different designed concrete mixtures, one with lower concrete compressive strengths than the actual lowest and the other one with a higher designed concrete compressive strength than the actual highest) were prepared recently, they will be tested during next weeks (at the age of 28 days) and later the results will be updated. The same short fibres were used, a total of 8 concrete mixtures (each mixture consist of 15 concrete specimens) were prepared in addition to the 16 mixtures already investigated.
After the above mention experimental work, as next stage of this project, analytical and numerical studies will take place and larger scale experimental work will follow as well.
Additional experimental work has been done since previous update/report. Two series of pull-out test were performed.
The first series indents to complete the experimental work previously presented. Namely, two additional concrete compositions were used to prepare a total of 2*4=8 concrete mixes, with or without fibres, just as in previously reported cases. These concrete mixes were meant to widen the studied concrete strength range, one with lower compressive strength (~28 MPa) than the lowest (~40 MPa) prepared in the previous stage and one with higher (~67 MPa) compressive strength than the highest prepared (~92 MPa) in the previous stage.
In the second series the effect of concrete cover was investigated by eccentric pull-out tests. Three different concrete covers and three different rebar types were considered.
Analysis of the result are still in progress and will be presented in next report.
Since last report, analysis of the result of the two last experimental series took place.
In the first series the effect of short fibers on bond behaviour of the above mentioned different types of FRP bars were further studied, by preparing 8 additional concrete mixes. These extra mixes also give the possibility to investigate the effect of concrete strength on bond between FRP bars and concrete.
In the second experimental series the bond strength of a new FRP bar was investigated when different concrete covers were used and comparing to other similar (by means of diameter, modulus of elasticity, but not surface characteristics) FRP bars. The new rebar has performed well in terms of bond strength, however the failure mode was different than in case of other rebars. Bond strength increased with increase of the concrete cover for all three rebar types.
When studying the bond of FRP rebars in concrete, not only the bond strength should be investigated since that involves high slip values, which are only valid in case of ULS, but bond stresses in SLS are also important.
Effect of short fibres on bond stresses was studied and the following conclusions were drew. The trend in bond stress is not obvious, however it gives the impression, that neither the concrete compressive strength, neither the addition of short fibres have effect on the bond stress at 0.15 mm loaded end slip.
0.15 mm slip was chosen, since there are slips at both sides of a crack, consequently the crack width is 2*0.15=0.3 mm, which is often the limit for design.
Furthermore, the preparation for the next experimental part is undergoing, along with analytical model validation and numerical modelling. The results will be disclosed in next reports.
Conferences and meetings
Membership in international organisations