Monotonic and cyclic axial compressive responses of concrete specimens externally confined with different types of FRP composites: experimental and analytical investigations

The main objective of this study was to investigate the monotonic and cyclic axial stress versus strain response of the concrete confined with different types of fiber reinforced polymer (FRP) composites such as cotton fiber rope reinforced polymer (CFRRP), glass fiber reinforced polymer (GFRP) composites and carbon fiber reinforced polymer (CFRP) composites. Another objective was to investigate the applicability of existing ultimate strength models to predict the ultimate strength of concrete confined with CFRRP, GFRP and CFRP composites. For this, a total number of 28 concrete cylinders were cast, strengthened and tested under pure axial compression. The concrete cylinders were strengthened with two and four layers of FRPs. The experimental results indicate that all kinds of FRPs are very useful to alter the ultimate strength and strain of the confined concrete. The highest increase in ultimate strength was recorded for CFRP composite confined specimens and the highest increase in ultimate strain was observed for CFRRP composite confined specimens. Further, both ultimate strength and ultimate strains are found higher in the case of cyclic loads compared with the monotonic load. In general, the stress versus strain curves of cotton fiber rope and glass FRPs are found trilinear. Whereas, stress versus strain curves of CFRP confined specimens are observed to be bilinear. Existing ultimate strength models developed for CFRRP and CFRP composites are found well accurate to predict the ultimate strengths of CFRRP and CFRP confined concrete specimens tested in this study.


Introduction
Natural hazards (such as earthquakes, floods and land sliding) are occurring frequently around the world [1]. These natural hazards cause vast damage to the roads, bridges, dams and buildings. Especially, earthquake leads to higher amount of loads on reinforced concrete structures. During earthquake, concrete structures with insufficient and poor detailing of reinforcement totally collapsed [2]. A recent survey of existing reinforced concrete buildings in Thailand indicate that most columns in existing buildings are poorly detailed i.e., widely spaced stirrups [3]. An earthquake of intensity Mw 6.1 occurred in northern Thailand on 5th May 2014. The epicenter of this earthquake was located in Chiang Rai. A total number of 594 buildings found damaged during earthquake [4]. In buildings, mainly columns were highly damaged [5]. A typical damage to RC columns is shown in figure 1. Therefore, strengthening and rehabilitation of the existing structures with poor reinforcement details is very vital. Traditionally, the strengthening of concrete structures is performed by using concrete and steel jacketing methods [5][6][7]. In concrete jacketing method, an additional concrete shell is usually constructed at the perimeter of the existing concrete members to enhance the strength and ductility of the strengthened member. The concrete jacketing is usually performed in three steps. In first step, steel bars are placed around the existing concrete member, whereas in the second step, plywood or steel formwork is provided and in the last step, concrete or cement grout is placed in the form work.
Although concrete jacketing method is effective and useful for strengthening purposes [4][5][6][7][8][9][10][11], however, there are few disadvantages of concrete jacketing method. For example, increase in weight and volume of the strengthened concrete column. Also, this method is time consuming and requires skilled labor to install. In contrast to the concrete jacketing, steel jacketing method is relatively easy to install and construct. In steel jacketing method, steel plates are welded around the perimeter of the concrete columns and space between steel plates and concrete surface is usually filled with cement grout [12], thus eliminating the use of formwork. Although steel jacketing method solve some issues related with the concrete jacketing method [13,14], however, steel jacketing method itself involves some problems such as increase in dead mass and corrosion of the steel plates. Thus, there are further needs to investigate, low weight and easily installed strengthening technique for the retrofitting and rehabilitation of concrete structures.
During last two decades, use of fiber reinforced polymer (FRP) composite have gained very popularity in the field structural strengthening and rehabilitation due to high tensile strength and light weight [15][16][17][18]. In this method, a thin fabric sheet of carbon, glass, or polyethylene terephthalate fabric is wrapped around the concrete columns following by application of epoxy resin using a brush or roller. Ozbakkaloglu [19] investigated the axial compressive response of square and circular RC columns and concluded that CFRP composites can be effectively adopted to alter the strength and stiffness of strengthened members. Despite of successful application of CFRP composites, the practical use of CFRP composite is still very limited because of very high cost of carbon fiber and epoxy resin. Glass fiber which is normally made from melted glass is much cheaper as compared to the carbon fiber. The glass fiber is similar to carbon fiber in weight but tensile strength of glass fiber is lower than the carbon fiber. Also, the glass fiber is slightly more flexible than carbon fiber. Benzaid et al [20] investigated the axial compressive response of square concrete columns externally confined with Glass fiber reinforced polymer (GFRP) composites. Different parameters such as thickness of GFRP composites and the corner radius were considered. Research results indicate that GFRP can produce good lateral confining pressure and it can be used for strengthening and or rehabilitation of concrete structures. Cao et al [21] studied axial compressive behavior of concrete confined with GFRP composite tubes. The authors concluded that confinement due to GFRP tubes is effective to enhance ultimate strength and strain of the confined concrete. Lam and Teng (2001) presented a study on the compressive strength of FRP-confined concrete. Existing strength models for FRP-confined concrete have been reviewed and a new strength model for concrete confined by different types of FRP materials has been proposed [22]. Dai et al (2011) performed a study on the behaviour of concrete confined by PET and PEN fibres. They concluded that the stress-strain model proposed by Jiang and Teng (2007) [23] could be applied to predict the compressive strength of concrete confined by PET and PEN fibres but might overestimate the ultimate axial strain. A modified model was developed based on the stress-strain model proposed by Jiang and Teng (2007) [23], which provided a better prediction of the ultimate strength, strain and stress-strain behaviour of PET-and PEN-confined concrete [24]. Further, design guidelines for FRP confinement and strengthening of concrete members has been published in international standards such as American Concrete Institute  Although, CFRP and GFRP composites are well suitable for strengthening purposes, however, their availability in some developing countries is still an issue. Recently, the authors of this article developed a low cost, easily available and more environment friendly strengthening technique i.e., fiber rope reinforced polymer (FRRP) composites [29]. In FRRP composites, locally available fiber ropes are wrapped around the concrete members following by application of suitable resin by using a brush or roller. The salient features of the ropes are easy availability all around the world, low cost, light weight and simple application. A previous study indicates that FRRP composite are very effective to enhance the ultimate load carrying capacity and ductility of the confined circular concrete columns [29]. All above mentioned FRP composites had unique mechanical properties and stress versus strain behavior. For example, the ultimate tensile strength of CFRP composite is approximately 511% and 725% higher than the GFRP and FRRP composites, respectively. Whereas, the ultimate strain of FRRP composite is much higher than the CFRP and GFRP composites. For accurate and reliable FRP design, it is important to know the effect of mechanical properties of each FRP composite on the stress versus strain behavior of concrete. To date, limited studies have been published on direct comparison of stress versus strain behavior of concrete confined with different types of FRP composites. Therefore, the main objective of this research work is to investigate the axial compressive response of concrete confined with CFRRP, GFRP and CFRP composites. Concrete cylinders strengthened with different types of FRP composites are tested under uniaxial compression. Also, the experimental results are used to evaluate the applicability of existing ultimate strength models previously developed for CFRP and GFRP composites to the tested FRPs in this study.

Experimental program
In this study, standard concrete cylinders with dimensions of 300 × 150 mm (height × diameter) were cast by using low strength concrete. A total number of 28 concrete cylinders were cast and tested under pure axial compression. Among 28 specimens, two specimens were tested as control for each type of loading i.e., cyclic and monotonic. The remaining specimens were confined with CFRRP, GFRP and CFRP composites. Each FRP composites was applied by using two and four layers. The details of test specimens are provided in table 1. Test specimens were designated to refer the research parameters. For example, for CFRRP-2L, the first letters before the hyphen represent type of FRP composite i.e., cotton fiber rope reinforced polymer composite and the letters behind the hyphen indicate the thickness of FRP composites i.e., two layers. Two identical specimens were prepared for each type and thickness to access the average values.

Material properties
A single batch of concrete was used to cast all the test specimens. The mix proportion of the concrete is given in table 2. Ordinary Portland cement (Type-1) was used with natural granite coarse aggregates with a maximum size of 13 mm and locally available river sand to prepare the concrete. In Thailand, river sand is generally used for residential and commercial projects. The target unconfined compressive strength at 28 days was 15 MPa. However, the actual compressive strength of the concrete at the test age was slightly higher than the desired target strength. Locally available cotton fiber ropes (figure 2), glass fiber and carbon fiber were impregnated in a two-part (epoxy resin and hardener) epoxy resin to develop FRP composites. The mechanical properties of FRP composites were determined by conducting tensile tests in accordance with ASTM standards [30]. For tensile tests, a computer controlled machine M500-50AT manufactured by Testometric was used as shown in figure 3.  figure 4 and typical failure of FRP tensile strips are shown in figure 5.

Preparation of concrete specimens
All concrete specimens were prepared using steel molds. A mechanical mixer was used to mix the concrete ingredients. The concrete was filled in the steel molds following standard guidelines. After 24 h, the steel molds were removed and the concrete specimens were placed in water for seven days for curing process. After 28 days of curing, FRP composites were applied on the on the concrete specimens by using hand layup process. In the first step, first end of cotton fiber rope and or glass/carbon fibers were attached to the concrete by using a super glue and in the next step fiber rope and or glass/carbon fibers were wrapped on the side surface of each concrete specimen ( figure 6) and in the final step, epoxy resin was applied using a brush as shown in figure 7. The CFRRP strengthened specimens are shown in figure 8. Further details on the cotton fiber rope wrapping process can be found in Hussain et al [29]. For both CFRP and GFRP, an overlap of 150 mm were provided to avoid the debonding of FRP composites. Since, there is a continuous wrapping of cotton fiber rope, therefore, there is no need to provide overlap for the last end, but the last end of cotton fiber rope was attached again by using a super glue.

Instrumentation and loading setup
All specimens were tested under pure axial compression until ultimate failure by a universal testing machine with 2000 kN capacity. A typical load setup is shown in figure 9. Three transducers 120 degree apart were attached to the side surface of the specimen to record axial deformation during monotonic loading. A calibrated load cell was placed at the top of the specimen to record applied load.

Results and discussions
Axial stress versus strain responses tested specimens are shown in figure 10. Experimental results in terms of average first peak strength, ultimate strength and ultimate strain are also summarized in table 4. It can be seen     that all types of FRP composites are very effective to improve the ultimate load carrying capacity and axial strain of confined concrete. Further discussion on experimental results is provided in the following sections.

Overall response
From figure 10, it can be seen that all types of FRP composites i.e., cotton FRRP, GFRP and CFRP result in higher ultimate strength and strain values of the FRP confined concrete specimens as compared to the control specimen. The highest increase in ultimate strength was recorded for CFRP confined specimens and the highest increase in ultimate strain was observed for cotton FRRP composite. In the case of cyclic loads, the increase in ultimate strength was observed 111% and 224% for 2 layers and 4 layers CFRRP confined specimens as compared to the control specimens, respectively. The increase in ultimate strength was observed 187% and 334% for 2 layers and 4 layers GFRP confined specimens as compared to the control specimens, respectively. Whereas the increase in ultimate strength was observed 239% and 342% for 2 layers and 4 layers CFRP confined specimens as compared to the control specimens, respectively. In the case of monotonic loads the increase in ultimate strain was observed 109% and 243% for 2 layers and 4 layers CFRRP confined specimens as compared to the control specimens, respectively. The increase in ultimate strength was observed 181% and 325% for 2 layers and 4 layers GFRP confined specimens as compared to the control specimens, respectively. Whereas the increase in ultimate strength was observed 233% and 300% for 2 layers and 4 layers CFRP confined specimens as compared to the control specimens, respectively. Strengthening material with low tensile strength and higher tensile strain can be effectively utilized for ductility enhancement purposes. Further, the percentage increase in ultimate strength and strain is graphically shown in figure 11 to assess the effects of the both loading types i.e., monotonic and cyclic loads on the ultimate load carrying capacity and strain of the FRP confined specimens. In general, both ultimate strength and ultimate strains are found higher in the case of cyclic loads as compared with the monotonic load. For example, in the case of cyclic loads, the increase in ultimate strength was observed 239% and 342% for 2 layers and 4 layers CFRP confined specimens as compared to the control specimens, respectively. Whereas in the case of monotonic loads, the increase in ultimate strength was observed 233% and 300% for 2 layers and 4 layers CFRP confined specimens as compared to the control specimens, respectively. More studies are needed to further elaborate this phenomenon. There was an exception in the case of GFRP confined  specimens in which ultimate strength is found lower in cyclic loads as compared with the monotonic loads as shown in figure 11(d). Rousakis et al 2017 also reported that the efficiency of CFRP confinement was higher than the GFRP composites [31].

Stress versus strain responses
From figures 10 and 11, it can be seen that mechanical properties of FRP composites had significant effect on the stress versus strain response of FRP confined concrete. FRP composite with low tensile strength but high ultimate strain could result in high ductility enhancement. Particularly, two types of stress versus strain responses are observed for FRP confined concrete specimens. In case of CFRRP confined specimens, the stress versus strain response is essentially trilinear. In which the first part is almost similar to the unconfined concrete until first peak strength. The second part is descending downward indicating large lateral dilation of the concrete and third part is ascending upward until ultimate state. Ultimate state is defined as the rupture of the FRRP composite. Similar responses were also observed in the previous studies [32,33]. The third part indicates that confinement due to the FRRP composite is initiated and stresses are mainly taken by the FRRP composite. Whereas, the pattern of stress versus strain responses of both GFRP and CFRP confined specimens are similar, showing bi-linear response. In bilinear response, the first portion is once again similar to the unconfined concrete following a transition zone. The second part is ascending linearly until the rupture of the FRP composite. Comparatively, the stiffness of the second part of CFRP confined specimens is large than the GFRP Figure 10. Test results.
confined specimens. This phenomenon is mainly due to the larger deformability of the glass fiber as compared to the carbon fiber. In the past, similar bi-linear responses were also observed for FRP confined circular specimens [34,35].

Failure modes
The ultimate failure modes of all FRP confined specimens are shown in figure 12. The control specimens mainly failed due to concrete crushing ( figure 12(a)). Overall, the failure of the FRP confined specimens occurred due to the tensile rupture of the FRP composite as shown in figure 12. The failure characteristic of all FRP confined specimens was very explosive with large sound especially for CFRP confined specimens both for monotonic and cyclic loads. Previous studies [36,37] also reported tensile fracture of the FRP composites due to lateral outward expansion of the concrete in the middle portion.

Analytical analysis
In the past, several models have been proposed by researchers to predict the ultimate tensile strength of FRP confined concrete specimens [36,37]. These models can be generally written in the following form; (

Assessment of existing strength models
In this section, the applicability of existing strength models (table 5) is accessed for the CFRPR, GFRP and CFRP confined specimens. For this purpose, the existing strength models are used to predict the ultimate strength values of cotton FRRP, GFRP and CFRP confined specimens. Comparison of predicted values and experimental results is shown in figure 13. The theoretical predictions of Hussain et al [29] model are found accurate for CFRRP confined specimens ( figure 13(a)), low for GFRP confined specimens ( figure 13(b)) and high for CFRP confined specimens ( figure 13(c)). On the other hand, among existing ultimate strength developed for GFRP  [40] models are found accurate for GFRP confined specimens tested in this study ( figure 13(e)). Whereas, in the case of ultimate strength models developed for CFRP composite, the theoretical predictions of almost all models are found accurate for CFRP confined specimens tested in this study (figure 13(i)).

Conclusions
This study presents a detail experimental investigation on axial compressive response of concrete confined with different types of FRP composites such as cotton FRRP, GFRP and CFRP composites. Circular concrete specimens confined with two and four layers of FRP composites were tested under pure axial compression to evaluate the effect of mechanical properties of FRP composites on stress versus strain response of confined concrete. Also, applicability of existing strength models was assessed for cotton FRRP, GFRP and CFRP confined specimens. Following conclusions are drawn based on experimental and analytical investigation.
1. All types of FRP composites are very effective to alter the ultimate load carrying capacity and ductility of confined concrete. Overall, the highest increase in ultimate strength was recorded for CFRP composite confined specimens and the highest increase in ultimate strain was observed for CFRRP composite confined specimens.
2. There is found increase in ultimate load and strain of FRP confined specimens with an increase in the FRP thickness. For example, In the case of cyclic loads, the increase in ultimate strength was observed 111% and 224% for 2 layers and 4 layers CFRRP confined specimens as compared to the control specimens, respectively. In the case of cyclic loads the increase in ultimate strain was observed 133% and 2214% for 2 layers and 4 layers CFRRP confined specimens as compared to the control specimens, respectively.
3. The stress versus strain curves of cotton fiber rope and glass FRPs are found as tri-linear. Whereas, stress versus strain curves of CFRP confined specimens are observed as bi-linear. 4. In general, both ultimate strength and ultimate strains are found higher in the case of cyclic loads compared with the monotonic load. For example, in the case of cyclic loads, the increase in ultimate strength was observed 239% and 342% for 2 layers and 4 layers CFRP confined specimens as compared to the control specimens, respectively. Whereas in the case of monotonic loads, the increase in ultimate strength was observed 233% and 300% for 2 layers and 4 layers CFRP confined specimens as compared to the control specimens, respectively.
5. Existing ultimate strength models developed for CFRRP and CFRP composites are found well accurate to predict the ultimate strengths of CFRRP and CFRP confined concrete specimens tested in this study, respectively. Among selected ultimate strength models developed for GFRP composite, only the theoretical predictions of Vin and Di Lud models are found accurate for GFRP confined specimens tested in this study.

Recommendations for future research
The model assessment that has been presented herein clearly indicates that there is still need to develop a uniform models which can be used to predict the ultimate strength and strain of CFRRP, CFRP and GFRP confined concrete specimens. Therefore, it is recommended that a carefully chosen selection criteria for thickness of FRP is applied in the future database development efforts.