J. Cent. South Univ. (2020) 27: 876-890

DOI: https://doi.org/10.1007/s11771-020-4338-6

Performance of interface between TRC and existing concrete under a chloride dry-wet cycle environment

LI Yao(李耀)^{1, 2}, YIN Shi-ping(尹世平)^{1, 2}, LV Heng-lin(吕恒林)²

1. State Key Laboratory for Geomechanics & Deep Underground Engineering, China University of Mining and Technology, Xuzhou 221116, China;

2. Jiangsu Key Laboratory of Environmental Impact and Structural Safety in Engineering, School of Mechanics & Civil Engineering, China University of Mining and Technology, Xuzhou 221116, China

Central South University Press and Springer-Verlag GmbH Germany, part of Springer Nature 2020

Abstract: Textile-reinforced concrete (TRC) is suitable to repair and reinforce concrete structures in harsh environments. The performance of the interface between TRC and existing concrete is an important factor in determining the strengthening effect of TRC. In this paper, a double-sided shear test was performed to investigate the effects of the chloride dry-wet cycles on the average shear strength and slip at the interface between the TRC and existing concrete, also considering the existing concrete strength, bond length, textile layer and short-cut fiber arrangements. In addition, X-ray diffraction (XRD) technology was used to analyze the microscopic matter at the interface in the corrosive environment. The experimental results indicate that the interface performance between TRC and existing concrete would decrease with continued chloride dry-wet cycles. Compared with the specimen with a single layer of textile reinforcement, the specimens with two layers of textile with added PVA or AR-glass short-cut fibers could further improve the properties of the interface between the TRC layer and existing concrete. For the TRC with a single layer of textile, the average shear strength tended to decrease with increasing bond length. In addition, the strength grade of the existing concrete had a minor effect on the interface properties.

Key words: textile-reinforced concrete; chloride dry-wet cycles; double-sided shear; average shear strength; interface slip; X-ray diffraction technology

Cite this article as: LI Yao, YIN Shi-ping, LV Heng-lin. Performance of interface between TRC and existing concrete under a chloride dry-wet cycle environment [J]. Journal of Central South University, 2020, 27(3): 876-890. DOI: https://doi.org/10.1007/s11771-020-4338-6.

1 Introduction

Because of the temperature difference between the inside and outside of concrete and the effect of the marine environment, coastal concrete structures often contain cracks, and the effect of chloride erosion is considerable [1-3]. Additionally, buildings and civil infrastructures have increasingly been exposed to multihazard environments in recent years [4, 5]. Therefore, repairing and reinforcing concrete structures are important tasks for these constructions [6]. In recent decades, the use of fiber-reinforced polymer (FRP) wraps has become popular due to the favorable properties of this retrofitting method: extremely low weight-to- strength ratio, easy application, minimal change in the structure behavior, and protection against and prevention of corrosion [7, 8]. Despite these advantages, the FRP retrofitting technique has a few drawbacks; for instance, exposed fabric results in poor resistance to fire, high temperatures and radiation, and fabric inconsistencies with the existing concrete structure limit the strengthening effect [9, 10]. Therefore, scholars have introduced the concept of reinforcement fibers that combine textile fibers with inorganic substrates (such as cement-based mortar) [11]. Some scholars have proposed textile-reinforced concrete (TRC) [12, 13], textile-reinforced mortar (TRM) [10, 14, 15] and fabric-reinforced cementitious matrix (FRCM) [16, 17] systems as alternative methods of FRP reinforcement. These systems are merely variations of the same core idea and typically include only minor differences.

TRC is a new type of fiber-reinforced concrete. TRC uses carbon fiber, alkali-resistant (AR)-glass fiber, basalt fiber or a hybrid textile of these fibers as the main reinforcing material. These fibers are used for the restoration of existing buildings and for the strengthening of new buildings due to their tensile strength and corrosion resistance, and high-performance concrete is used as the matrix [15, 18]. Due to the combination of the textile and cementitious matrix, TRC has good corrosion resistance, tensile properties and anticrack characteristics. Compared with FRP, TRC can be better integrated with existing concrete structures [9, 10]. Recently, TRC has been applied to strengthen concrete structures [19].

The reinforcement effect of composites is largely determined by the interface properties between composites and existing concrete [13, 20]. Regarding FRP, many scholars have investigated the interface properties [20-22]. TUAKTA et al [20] studied the effects of cyclic moisture conditions on the fracture toughness of a CFRP-concrete bond system, considering the peel and shear fracture toughness. In addition, a predictive model has been developed for predicting the service-life of FRP-concrete bond systems.LI et al [21] found that the FRP-concrete bond strength was related to the number of dry-wet cycles and that the interface bond strength increased during the early stage of testing. LIANG et al [22] explored the effect of FRP type, dry-wet cycles and sustained loading level on the failure modes, stress transfer, and local bond-slip curves of the composite specimens. In addition, a reasonable model was proposed to predict the fracture energy that could be a reference for the design of bond durability factors. On the other hand, some scholars have been devoted to studying the interface properties between TRC/TRM/FRCM and existing concrete by performing pull-out tests [14, 16, 17, 23, 24] and push-in tests [13, 25]. These research results are valuable for investigating the interface performance. ORTLEPP et al [23] summarized the form of interface damage between TRC and existing concrete. Further, ORTLEPP et al [24] found that the strain distribution of the TRC sheet was not continuous in the cracking process of the interface. According to this theory, ORTLEPP et al [23] described the development of cracks during the destruction of the interface bond and proposed the basic bond-slip theory of TRC and existing concrete. ASKOUNI et al [14] noted that the effective bond length between glass TRM overlays and masonry substrates was approximately 130 mm. D’AMBRISI et al [17] also improved the bond-slip model between fiber and matrix based on the experiments. In addition, D’AMBRISI et al [16] verified the bond-slip curve between polyparaphenylene benzobisoxazole fiber reinforced cementitious matrix (PBO-FRCM) and existing concrete using the results from a double-side shear test and confirmed the validity of reinforcement from the perspective of fracture energy. XUN et al [25] investigated the interface properties between TRC and existing concrete in a conventional environment under different interface treatments by push-in testing. The results showed that the roughing treatment and roughness had significant impacts on the performance of the interface between TRC and existing concrete. YIN et al [13] studied the effect of TRC precracking, concrete strength, interface form, short-cut fibers, and freezing-and-thawing cycle number on the interfacial bond properties between TRC and existing concrete under chloride salt erosion and freezing-and-thawing cycles. In addition, POURASEE et al [26] explored the anti-infiltration performances of different forms of fabric-cement- based materials by conducting water permeability experiments. The results showed that the impermeability of concrete specimens with a TRC layer was significantly improved compared with that of concrete specimens without a TRC layer. That study also noted that when durability was considered for fabric-cement composites, it was not sufficient to consider only the crack width; instead, the fluid-transport properties and physical geometry of the yarns composing the fabric must also be considered. LIEBOLDT et al [6] further investigated the impermeability of TRC- strengthened concrete structures. The results showed a lower number of cracks in the concrete structure strengthened with TRC; thus, the impermeability of concrete was improved significantly.

To apply TRC to the construction reinforcement field, particularly to marine engineering, the properties of the interface between TRC and existing concrete in a chloride environment must be studied to improve its reinforcement effect. However, few studies have investigated the effects of different forms of reinforcement on the interface between TRC and existing concrete [10]; to the best of our knowledge, no studies have investigated the effect of chloride dry-wet cycles on the interface properties. This paper investigated the properties of the interface between TRC and existing concrete by analyzing the average shear strength and slip at the interface under chloride wet-dry cycles and different forms of TRC reinforcement through a double-side shear test. Furthermore, X-ray diffraction (XRD) technology was used to analyze the microscopic matter at the interface under the effects of the environment.

2 Materials and test methods

2.1 Experimental materials

2.1.1 Textiles

Textile was produced from carbon fiber and glass fiber using a warp knitting process. The carbon fiber was the reinforcing fiber (cross- sectional area was 0.45 mm²). The glass fiber was used only to fix the carbon fiber; its load-carrying contribution was not considered in this study. The mesh size of the textile was 10 mm×10 mm, and its thickness was approximately 2 mm (shown in Figure 1(a)). To increase the adhesive properties between the textile and fine-grained concrete, textile structures were impregnated with epoxy resin and sand. The particle size of the selected sand was 0.6-1.2 mm. Figure 1(b) shows the treated textile. In this experiment (push-in test), the properties of the interface between the TRC and existing concrete were the main research object, and the textile did not bear the direct bending or tensile load. Thus, in this experiment, the mechanical properties of the fibers were not exclusively studied; specific studies on this topic have been presented [27]. The details of the yarns used in textile, according to the manufacturer and the literature [27], are listed in Table 1.

2.1.2 Existing concrete

Ordinary Portland cement of 42.5R and 32.5R was chosen for the existing concrete. Gravel with a diameter of 5-10 mm was chosen as the coarse aggregate, and its apparent density was 2720 kg/m³. Medium sand was chosen as the fine aggregate, and its fineness modulus was 2.7. The superplasticizer was polycarboxylate. The laboratory fine-grained concrete mix proportion is shown in Table 2.

Figure 1 Hybrid textiles made of carbon and E-glass yarns:

Table 1 Mechanical properties and geometric parameters of yarns

2.1.3 Fine-grained concrete

The fine-grained concrete mainly contained fly ash, silica fume, cement and fine aggregate. Ordinary Portland cement of 52.5R was used. The particle size of the fine sand was 0-0.6 mm, and the particle size of the coarse sand was 0.6-1.2 mm. The laboratory mix proportion of the fine-grained concrete is shown in Table 2. Tests confirmed that the selected mix proportion met the performance requirements of the self-compacting concrete [28]. Furthermore, according to Ref. [29], tests were performed to study the slump flow (expansion), flow rate and permeability. The slump flow index of fine-grained concrete was 12.05, which was slightly higher than that of the self-compacting concrete; the time of fine-grained concrete outflow from the V-shaped container was 7 s. The outflow rate was 1.43, which surpassed the self-compacting concrete

requirements and was faster than the average flow of self-compacting concrete. The value of parameter L, which characterizes the permeability of fine-grained concrete, was 0.95, which was larger than 0.73 suggested by NAGAMOTO et al [30]. That is, the permeability of the fine-grained concrete satisfied the test requirements. In addition, the segregation phenomenon did not occur. The hardened properties of the fine-grained concrete are shown in Table 3.

2.1.4 Short-cut fibers

This paper investigated the effects of two types of short-cut fibers (AR-glass and polyvinyl alcohol (PVA)) on the properties of the interface with the reinforced layer. Previous research has explored a more reasonable volume fraction, and a value of 0.5% was found suitable for the matrix. Therefore, the volume fraction of PVA or AR-glass fiber was chosen as 0.5%. The fiber basic properties are shown in Table 4. Furthermore, to ensure the fibers were uniformly distributed and the results reported herein were not affected by the fiber distribution, before the reinforcement, the concrete was fully stirred.

Table 2 Laboratory mix proportion of concrete (kg/m³)

Table 3 Mechanism properties of fine-grained concrete

Table 4 Properties and geometric parameters of short-cut fibers

2.2 Specimen information

In this experiment, the properties of the interface between the TRC and existing concrete under different numbers of chloride dry-wet cycles and different forms of the interface were studied.

A total of 5 groups of specimens, each with 3 specimens, were designed to study the effects of different numbers of chloride dry-wet cycles on the properties of the interface between the TRC and existing concrete. A total of 7 groups of specimens, each with 3 specimens, were designed to study the influence of different interface forms (such as different bond lengths and concrete strengths) on the properties of the interface between the TRC and existing concrete under the chloride dry-wet cycles. The specimen details are shown in Table 5. L1 was the standard group, L2 was the group that did not undergo chloride dry-wet cycles, and L3 was the group treated with a chloride salt solution but not dry-wet cycling. The test period for groups L2 and L3 was 120 d.

2.3 Specimen production

As a strengthening material, TRC is typically applied by spraying or laminating. However, in some cases, such as when the strengthening side is the bottom surface of a beam, it is inconvenient to adopt these methods. Furthermore, adopting prefabricated TRC sheets could save time in a strengthening project. Thus, in this paper, a prefabricated TRC sheet was utilized to reinforce the existing concrete, and the interface properties were studied.

For the TRC sheet with a single layer of textile, the production was divided into four steps. First, each wooden mold with dimensions of 900 mm (length)×450 mm (width) was leveled, and then,5 mm thick wooden strips were nailed around the wooden mold to control the thickness of the fine-grained concrete. Second, a 5 mm layer of fine-grained concrete was poured into the wooden molds. Third, the textile under tension was fixed on the mold. Finally, the remaining fine-grained concrete was poured into the wooden molds to create another 5 mm layer after the 5 mm thick wooden strips were nailed around the wooden mold, as in the first step. For the TRC sheet with two layers of textile, the third and fourth steps were repeated, and the thickness of the specimens was approximately 19 mm. The molds were removed after 24 h at constant room temperature; then, the TRC sheets were placed in a curing room for 27 d. Using a concrete cutter, the large TRC sheets were cut into small TRC sheets with different lengths and cross-sectional dimensions of 100 mm×12 mm (a single layer) and 100 mm×19 mm (double layers). In this paper, the contribution of fine-grained concrete to the interface strength was not considered independently; the experimental results were attributed to the synergistic effect between textile and fine-grained concrete. In addition, the carbon fibers of the textile were always aligned in the longitudinal direction of the TRC sheet.

The concrete mix proportions of the existing concrete are shown in Table 2. Substrate specimens measuring 300 mm×100 mm×100 mm were produced. The molds were removed after 24 h at constant room temperature; then, the specimens were placed in a standard curing room for 27 d. The interface treatment was performed on the surfaces of the concrete specimens to enhance the bonding properties between the concrete and TRC. In this process, cement slurry was chiseled off to artificially expose the coarse aggregates. The chiseled depth △h of the existing concrete surface was controlled to between 2 and 4 mm to achieve a better interfacial property.

Table 5 Test scheme designed for specimens

The bonding agent adhered the prefabricated TRC sheet to the existing concrete. The mix proportion of the bonding agent was the same as that of the fine-grained concrete. The specimens used to investigate the interface performance were treated by chloride dry-wet cycle processing after production and standard curing for 28 d. Specimens were soaked in a 5% NaCl solution at room temperature for 12 h and then dried naturally for 12 h indoors. The duration of one cycle was 24 h [31-33]. A total of 120 cycles were completed. In this process, the specimens were soaked and were not sealed laterally to mimic actual construction conditions in a marine environment. To study the effects of different numbers of cycles on the properties of the interface between the TRC and existing concrete, specimens for 90 and 150 chloride wet-dry cycles were designed for the experimental groups. The environment simulated in this experiment was an ideal environment, i.e., an environment without tensile stress, with the goal of clearly observing the effects due to various factors. Thus, the effect of tensile stress was not considered. The properties of the interface between the existing concrete and TRC under the coupling action of tensile stress and various factors will be further investigated in the future.

2.4 Loading method

In the experiment, because the thickness of the TRC sheet used for strengthening was approximately 10 or 20 mm, a double-sided shear test was adopted using the push-in method by exerting pressure on top of the existing concrete (shown in Figure 2). The maximum pressure value was acquired when the TRC and/or existing concrete was damaged. The bonding effect of the fine-grained concrete adhered the TRC and existing concrete together, and the textile did not directly bear the tensile stress. The performance of the fine-grained concrete in the TRC was improved because of the effect of the textile, as the textile decreased the deformation caused by the nonuniform stress during the test and increased the plastic deformation of the concrete; otherwise, the failure phenomenon of fine-grained concrete would not be debonding from the interface.

Figure 2 Diagram of double-side shear loading:

Because the shear stress does not distribute uniformly along the bond interface between the TRC and existing concrete, it was not convenient to calculate the stress at each observation point. Thus, the average shear strength at the interface was calculated by Eq. (1) to reflect the shear resistance of the bond interface.

                                 (1)

where τ is the average shear strength, P_m is the vertical load exerted on the existing concrete, and A is the total bond area between the two sides of the TRC and existing concrete.

When the load was applied, the test specimen and concave steel base were centered; the existing bottom concrete surface was covered with this concave base, and the outer TRC layer was arranged on the concave base. A pressure sensor was placed on the loading plate to record the magnitude of the load applied on the specimen, and two displacement sensors were placed on the surface of the existing concrete to monitor the slip between the TRC and existing concrete (Figure 3). The specimen was monotonically loaded using the displacement-control method until the specimen was destroyed, and the loading speed was 0.5 mm/min. Before the test, specimens were loaded 2 to 3 times to a proper value in order to eliminate the gap between the specimens and device; then, the data of the test machine was reset to zero. A schematic diagram of the loading is shown in Figure 2, and a photograph of the test setup is provided in Figure 3. In the static test, the shear loads and slip were recorded by using a static strain acquisition system developed by the Jiangsu Donghua Testing Technology Co., Ltd., China.

2.5 XRD test

The XRD test was carried out after the double-sided shear test was completed. First, a small sample was obtained at the interface between the TRC and existing concrete from the outer midpoint of the bonding interface. Before grinding, the sample was cleaned with anhydrous alcohol. Then, the sample was ground. Next, the power was obtained by filtering through a 325-mesh sieve. Finally, XRD technology was used to analyze the microscopic matter.

Figure 3 Test setup

3 Experimental results and analysis

3.1 Experimental results

The failure modes, ultimate loads, ultimate shear strength and maximum slip of each specimen are summarized in Table 6. During the test, the shear strength-slip characteristics of the specimens from the same group were similar. Therefore, nearly all of the data in the shear strength-slip curve were from one sample of the three specimens, and the ultimate shear strength and corresponding slip in the curve represented the average values from the three samples.

Table 6 Specimen shear resistance and failure mode

3.2 Interface properties affected by different forms of chloride dry-wet cycles

3.2.1 Failure-mode analysis

The failure mode of L1-L5 was double-side shear damage, as shown in Figure 4. When the specimen was damaged, the interface between the side of the existing concrete and the TRC cracked thoroughly, and the TRC detached from the existing concrete. The destruction process occurred rapidly and with a loud sound. The typical shear failure form is shown in Figure 4(a). This damage phenomenon suggested that when double-side shear failure occurred, the weakest location of the specimens was the interface between the TRC layer and existing concrete. The poor performance of the interface was typically due to damage during production and the loose microstructure. When the damage occurred, the TRC layer and some fine-grained concrete peeled from the existing concrete, as shown in Figure 4(b).

3.2.2 Analysis of average shear strength and slip of interface between TRC and existing concrete

1) Effect of chloride-solution immersion on interface properties

As shown in Figure 5(a), the shear strength- slip curves of L3 and L2 largely coincided in the initial loading phase. During the loading process, the curve of L2 maintained a steady upward trend. For the curve of L3, the slip increased significantly, and the shear strength was nearly constant, indicating that slip occurred at the interface. In addition, the value of the slip was approximately 0.09 mm. Compared with L2, the ultimate shear strength of L3 was 9.0% lower, and the slip corresponding to the ultimate shear strength was 9.0% higher when the interface failure occurred, as shown in Table 5. The shear strength-slip curve of L3 indicated that the increase in the maximum slip at the interface was mainly due to the increase in slip during the loading process. This result suggests that continuous immersion in the chloride solution adversely affects the interfacial bonding property between the existing concrete and TRC. However, this deterioration was not extensive.

2) Effect of number of dry-wet cycles on interface properties

Figure 4 Failure modes:

The shear strength-slip characteristics between the existing concrete and TRC under 90, 120 and 150 chloride dry-wet cycles are shown in Figure 5(b). In the initial loading phase, the curves of L1, L4 and L5 exhibited similar rising trends, and a sliding phenomenon was observed in L5. With increasing load, the curve of L4 continued to rise, and the curve of L1 exhibited an increase in slip while the shear strength remained nearly unchanged, which indicated that an interface slippage occurred in L1. Damage occurred in L5, even though its slip was small. Table 6 shows that compared with L4, when the interface failure occurred, the ultimate shear strength of L1 decreased by 22.4%, and the corresponding maximum slip decreased by 10.4%; the ultimate shear strength of L5 decreased by 45.3%, and the corresponding maximum slip decreased by 73.8%. These test results illustrate that the ultimate shear strength and corresponding maximum slip between the TRC and existing concrete decreased with an increasing number of chloride dry-wet cycles.

Figure 5 Average shear strength-slip curves for specimens under different numbers of chloride dry-wet cycles:

3.3 Interface properties affected by different interface reinforcement types

3.3.1 Failure-mode analysis

1) Double-sided shear damage

In all specimens except L10, the failure mode of the specimen was double-sided shear failure; the detailed analysis is given in the discussion above.

2) Delamination at bottom of TRC layer

The failure mode for specimen L10 is shown in Figure 4(c), cracks extended upwardly along the TRC, and the delamination occurred between the textile and fine-grained concrete at the bottom of the TRC layer; however, the TRC layer did not peel. At the same time, cracks stretched into the existing concrete, but these cracks were minor. This phenomenon shows that when the bond length is increased, the interface performance between the TRC and existing concrete is improved and the weak position of the specimens is changed.

3.3.2 Analysis of average shear strength and slip of interface between TRC and existing concrete

1) Effect of existing concrete strength on interface properties

Figure 6(a) illustrates that during the initial loading phase, the specimens presented nearly the same shear strength for a given amount of slippage. When the specimens were damaged, the ultimate shear strengths of the three groups were nearly identical, whereas the corresponding maximum slip increased slightly with the concrete strength. The ultimate slips of L7 and L1 were 10% and 20% higher than that of L6, respectively. Therefore, the compressive strength of the existing concrete had a slight influence on the interface bond properties between the TRC and existing concrete. Consequently, the TRC possesses good applicability for a range of concrete strengths, ensuring the interfacial properties between the strengthening layers and existing concrete.

2) Effect of textile layers on interface properties

Figure 6(b) illustrates that the shear strength-slip curves of L1 and L8 were nearly identical during the loading process. When the shear strength approached the ultimate shear strength, the shear strength of L1 increased slowly, and its slip increased rapidly; however, the shear strength-slip curve of L8 continued to rise, and its slip did not increase significantly. As shown in Table 6, the ultimate shear strength of L8 was 17.1% higher than that of L1 possibly because the textiles were treated by epoxy resin and sand, which could play a bridging role and improve the integrity of the TRC and the interface between the TRC and existing concrete. Therefore, the ultimate shear strength of the specimen reinforced with two layers was slightly larger than that with one layer. On the other hand, in the double-sided shear test, the misalignment of the interfacial shear stress and bottom compressive stress led to an eccentric compression of the TRC. The cross-sectional area of the TRC with two layers was increased, and the contact area between the TRC panel and the concave steel socket was augmented, which improved the force transmission of the TRC and avoided the local compression of the TRC at the base. Thus, the ultimate shear strength of L8 was larger than that of L1, fluctuating within a narrow range. In addition, comparing the damage forms of L1 and L8, when the interface shear failure occurred in L8, more of the fine-grained concrete used as a binder peeled off from the existing concrete with the TRC layer. This observation indicates that the use of the two textile layers changes the interfacial failure mode but not inherently, as the double-sided shear failure still appeared in the specimen L8.

Figure 6 Average shear strength-slip curves for different interface types of specimens:

3) Effect of bond length on interface properties

Figure 7(a) reveals that the slip and shear load increased with the bond length. As shown in Table 6, compared with the results for L9, the maximum slip of L1 increased by 81.1%, and the maximum shear load increased by 55.8%; the maximum slip of L10 increased by 241.5%, and the maximum shear load increased by 152.8%. However, the changes in the shear strength failed to comply with this trend: the ultimate shear strengths of L9 and L10 both increased compared with that of L1; Table 6 illustrates that the ultimate shear strength of L10 increased by 8.6%, while that of L9 increased by 28.9%; however, the reasons for these increases were not consistent. The interface area of L9 was decreased because of the smaller bond length of its TRC layer. When damage occurred, the maximum shear load and maximum slip also decreased. In L1 and L9, a portion of the fine-grained concrete binder peeled off with the TRC layer from the existing concrete. This section of the fine-grained concrete was a core part of the interface and was directly connected to the existing concrete surface, which was roughly treated. Moreover, the stripped areas of fine-grained concrete in L1 and L9 were very similar; however, the total area of the TRC layer of L9 was small, which meant that the shear load of L9 decreased but its shear strength increased.

Figure 7 Influences of bond length on interface properties of specimens:

Furthermore, throughout the loading process, the slip of L10 under the same shear strength was reduced compared with that of L1 because of its longer bond length and larger interface area. When the interface cracked, the cracks extended into the existing concrete such that the interface between the existing concrete and TRC could withstand a greater load, and consequently, there was a greater slip at the interface. During the final destruction of specimen L10, delamination occurred at the bottom of the TRC layer, but double-sided shear failure did not occur.

4) Effect of short-cut fibers on interface properties

As shown in Figure 8, in the initial loading phase, the shear strength of specimens with fine-grained concrete modified by either PVA fibers or AR-glass fibers was reduced compared with that of L1 at the same slip; however, with continued loading, the shear strength-slip curves of both types of specimens rose rapidly. Unlike L1, the slip phenomenon was not observed in L11 and L12 during the loading process. Therefore, the use of PVA/AR-glass fibers improved the interfacial bonding properties. Table 6 reveals that when the interface failure occurred, the shear strengths of L11 and L12 increased by 33.5% and 29.6%, respectively, compared with that of L1, suggesting that the modification of fine-grained concrete with PVA/AR-glass fibers is effective for enhancing the shear strength between TRC and existing concrete. This result may have been due to the bridging effect of short fibers improving the microstructure between the TRC and existing concrete, and the fibers could restrict the shrinkage and creep of fine-grained concrete during hydration, reducing the damage and deterioration caused by tensile stress at the interface of the TRC and existing concrete. Therefore, the chloride ion and free water diffusion and permeability decreased at the interface between the TRC and existing concrete in a corrosive environment, reducing the shrinkage due to drying, and the expansion and the deterioration due to the presence of chloride ions. When failure occurred at the interface, the shear capacities of specimens L11 and L12 were nearly identical, but the maximum slips at the interface were slightly different. The maximum slip of L12 was higher than that of L11. Table 6 shows that compared with the results for L1, the maximum slip of L12, corresponding to the ultimate shear strength, was increased by 5%; therefore, this specimen expresses a more significant deformation capability when subjected to the interfacial shear load. However, compared with that of L1, the maximum slip of L11, corresponding to the ultimate shear strength, decreased by 19.7%, which means that this specimen exhibits a more significant slip-limiting capability when subjected to the interfacial shear load.

Figure 8 Average shear strength-slip curves of specimens with fine-grained concrete modified by short-cut fiber

3.4 Interfacial microstructure analysis under chloride dry-wet cycles

The concrete deterioration mechanisms under chloride dry-wet cycles have been previously studied [1], and the reasons for such deterioration can be summarized in terms of three aspects:

1) Concrete properties are affected by the shrinkage and imbibition in moisture circulation: a greater number of dry-wet cycles leads to greater shrinkage and imbibition;

2) During the drying process, the NaCl solution in the concrete becomes supersaturated and crystallizes, leading to increasing inner pressure;

3) Chloride solution and concrete hydration products react to cause chemical corrosion, resulting in internal damage to the concrete.

When concrete undergoes chloride dry-wet cycles, the combined effects of these three aspects lead to the deterioration of the concrete. XRD tests mainly determine whether the chloride chemical corrosion can still affect the performance of the interface between the TRC and existing concrete.

Samples were taken from the interface between the TRC and existing concrete in specimens L2-L5 to analyze the composition of the interface under the action of chloride dry-wet cycles. In addition, the composition of the interface material affected by chloride dry-wet cycles was deduced. In this test, the XRD patterns of the interfacial microstructure between the TRC and existing concrete were obtained under different conditions, which are shown in Figure 9. Under the influence of different factors, the main composition of the interface microstructure was consistent within each group of specimens: SiO₂ (DS: 0.334, 0.425, 0.246 and 0.228 nm. Note: “DS” is short for “d-spacing”, and it represents the interplanar distance at the characteristic peaks that identify the phase) was from the coarse sand, fine sand and silica fume of the fine-grained concrete. CaCO₃ (DS: 0.301, 0.334, and 0.228 nm) was mainly from the coarse aggregate of the existing concrete. In addition, the carbonization reaction between the concrete and CO₂ in the air might have also led to the generation of CaCO₃. The characteristic peaks of C₃A·CaCl₂·10H₂O (DS: 0.786 and 0.286 nm) were found in the XRD patterns of all the specimens except L2. When comparing the specimen XRD patterns under the influences of different factors, small amounts of C₃A·CaCl₂· 10H₂O were found in the patterns, and the characteristic peaks are described below:

Figure 9 Specimen XRD patterns:

1) The characteristic peak of C₃A·CaCl₂· 10H₂O was not found in the XRD patterns when the specimen remained in the natural environment;

2) Under the influence of continuous chloride immersion, the characteristic peak of C₃A·CaCl₂· 10H₂O was found in the XRD patterns but was not large;

3) Compared with the results from specimens undergoing 90 dry-wet cycles, the characteristic peak of C₃A·CaCl₂·10H₂O was larger for specimens undergoing 150 dry-wet cycles.

A comparison of the characteristic peaks of C₃A·CaCl₂·10H₂O with the shear strengths of the specimens affected by the same factor illustrates that for the same factors, a larger characteristic peak of C₃A·CaCl₂·10H₂O indicates a lower specimen shear strength.

This conclusion is consistent with the results of previous studies suggesting that C₃A·CaCl₂· 10H₂O is a corrosion product from the reaction of chloride ions in solution and concrete hydration products and that C₃A·CaCl₂·10H₂O expands the concrete internally, causing microscopic damage to the concrete structures. The results showed that C₃A·CaCl₂·10H₂O would be produced at the interface between the TRC and existing concrete, and chemical corrosion would lead to weaker interfacial bonding properties under chloride dry-wet cycles. However, C₃A·CaCl₂·10H₂O chemical corrosion is not the only reason for the degradation of the interfacial bonding performance between the TRC and existing concrete. The shrinkage and imbibition of concrete and salt crystallization expansion may also cause the degradation of the interfacial bond performance. The interface properties are considerably affected by different forms of interface reinforcement; whether C₃A·CaCl₂·10H₂O chemical corrosion remains the primary reason for the deterioration of the interfacial properties is unclear and must be further investigated.

4 Conclusions

In this paper, double-sided shear test was used to study the effects of chloride solution immersion, number of chloride dry-wet cycles, existing concrete strength, bond length, textile layer and short-cut fibers on the average shear strength and slip at the interface between the TRC and existing concrete. XRD technology was also used to analyze the microscopic matter at the interface under the effect of a chloride environment. The following conclusions are drawn from the test results.

1) Compared with the natural environment, continuous chloride immersion adversely affects the interfacial bonding properties between the existing concrete and TRC. Continuous chloride immersion reduces the ultimate shear strength by 9.0%.

2) The ultimate shear strength between the TRC and existing concrete is decreased with an increasing number of chloride dry-wet cycles. Compared with the specimen that does not undergo dry-wet cycles, the shear strength is reduced by 30%, 42.9% and 51.9% after 90, 120 and 150 cycles, respectively.

3) For specimens with a single layer of textile, the average shear strength tends to decrease with increasing bond length. The interface properties of the specimen are mainly determined by the nature of the TRC layer and the interface when the concrete strength is between C20 and C40. The strength of existing concrete presents a slight influence on the interfacial bond properties between the TRC and existing concrete.

4) Compared with using a single layer of textile, the use of two layers of textile further improves the interfacial properties between the existing concrete and TRC as well as the ultimate shear strength and maximum slip of the interface.

5) The average shear strength between the TRC and existing concrete is improved under the effect of fine-grained concrete modified by PVA or AR-glass fibers, but the effect of PVA fibers on limiting the slip at the interface is more apparent.

6) C₃A·CaCl₂·10H₂O is produced between the TRC and existing concrete under chloride dry-wet cycles, and the chemical corrosion led to the deterioration of the interfacial bonding properties.

Additional experiments are needed to generalize certain results and determine overarching rules. This research will be pursued in the future. Moreover, the determination of the depth-dependent distribution of chloride concentration is meaningful, and experiments toward this goal will be performed in future studies to establish an improved theory.

References

[1] SHI Xian-ming, XIE Ning, FORTUNE K, GONG Jing. Durability of steel reinforced concrete in chloride environments: An overview [J]. Construction and Building Materials, 2012, 30: 125-138. DOI: 10.1016/j.conbuildmat. 2011.12.038.

[2] FU Chuan-qing, YE Hai-long, JIN Xian-yu, YAN Dong-ming, JIN Nan-guo, PENG Zhao-xiong. Chloride penetration into concrete damaged by uniaxial tensile fatigue loading [J]. Construction & Building Materials, 2016, 125: 714-723. DOI: 10.1016/j.conbuildmat.2016.08.096.

[3] YE Hai-long, FU Chuan-qing, JIN Nan-guo, JIN Xian-yu. Influence of flexural loading on chloride ingress in concrete subjected to cyclic drying-wetting condition [J]. Computers and Concrete, 2015(15): 183-198. DOI: 10.12989/cac.2015.15.2.183.

[4] YAN Dong-ming, LIN Gao, CHEN Gen-da. Dynamic properties of plain concrete in triaxial stress state [J]. ACI Materials Journal, 2009, 106(1): 89-94.

[5] LU Yi-yan, HU Ling, LI Shan, WANG Kang-hao. Experimental study and analysis on fatigue stiffness of RC beams strengthened with CFRP and steel plate [J]. Journal of Central South University, 2016, 23(3): 701-707. DOI: 10.1007/s11771-016-3115-z.

[6] LIEBOLDT M, MECHTCHERINE V. Capillary transport of water through textile-reinforced concrete applied in repairing and/or strengthening cracked RC structures [J]. Cement and Concrete Research, 2013, 52: 53-62. DOI: 10.1016/ j.cemconres.2013.05.012.

[7] COLAJANNI P, DOMENICO F, RECUPERO A, SPINELLAA N. Concrete columns confined with fibre reinforced cementitious mortars: Experimentation and modelling [J]. Construction and Building Materials, 2014, 52: 375-384. DOI: 10.1016/j.conbuildmat.2013.11.048.

[8] TOUTANJI H, BALAGURU P. Durability characteristics of concrete columns wrapped with FRP tow sheets [J]. Journal of Materials in Civil Engineering, 1998, 10(1): 52-57. DOI: 10.1061/(ASCE)0899-1561(1998)10:1(52).

[9] OMBRES L. Debonding analysis of reinforced concrete beams strengthened with fibre reinforced cementitious mortar [J]. Engineering Fracture Mechanics, 2012, 81: 94-109. DOI: 10.1016/j.engfracmech.2011.06.012.

[10] BOURNAS D A, PAVESE A, TIZANI W. Tensile capacity of FRP anchors in connecting FRP and TRM sheets to concrete [J]. Engineering Structures, 2015, 82: 72-81. DOI: 10.1016/ j.engstruct.2014.10.031.

[11] KOUTAS L, PITYTZOGIA A, TRIANTAFILLOU T C, BOUSIAS S N. Strengthening of infilled reinforced concrete frames with TRM: Study on the development and testing of textile-based anchors [J]. Journal of Composites for Construction, 2014, 18(3): A4013015. DOI: 10.1061/ (ASCE)CC.1943-5614.0000390.

[12] MECHTCHERINE V, LIEBOLDT M. Permeation of water and gases through cracked textile reinforced concrete [J]. Cement and Concrete Composites, 2011, 33(7): 725-734. DOI: 10.1016/j.cemconcomp.2011.04.001.

[13] YIN Shi-ping, LI Yao, JIN Zhe-yu, LI Peng-hao. Interfacial properties of textile-reinforced concrete and concrete in chloride freezing-and-thawing cycle [J]. ACI Materials Journal, 2018, 115(2): 197-208. DOI: 10.14359/51701919.

[14] ASKOUNI P D, PAPANICOLAOU C G. Experimental investigation of bond between glass textile reinforced mortar overlays and masonry: The effect of bond length [J]. Materials and Structures, 2017, 50(2), 164: 1-15. DOI: 10.1617/s11527-017-1033-7.

[15] RAOOF S, KOUTAS L N, BOURNAS D A. Bond between textile-reinforced mortar (TRM) and concrete substrates: Experimental investigation [J]. Composites Part B-Engineering, 2016, 98: 350-361. DOI: 10.1016/ j.compositesb.2016.05.041.

[16] D’AMBRISI A, FEO L, FOCACCI F. Bond-slip relations for PBO-FRCM materials externally bonded to concrete [J]. Composites Part B-Engineering, 2012, 43(8): 2938-2949. DOI: 10.1016/j.compositesb.2012.06.002.

[17] D’AMBRISI A, FEO L, FOCACCI F. Experimental and analytical investigation on bond between carbon-FRCM materials and masonry [J]. Composites Part B-Engineering, 2013, 46(3): 15-20. DOI: 10.1016/j.compositesb.2012. 10.018.

[18] YIN Shi-ping, XU Shi-lang, LI He-dong. Improved mechanical properties of textile reinforced concrete thin plate [J]. Journal of Wuhan University of Technology- Materials Science Edition, 2013, 28(1): 92-98. DOI: 10.1007/s11595-013-0647-z.

[19] PELED A. Confinement of damaged and nondamaged structural concrete with FRP and TRC sleeves [J]. Journal of Composites for Construction, 2007, 11(5): 514-522. DOI: 10.1061/(ASCE)1090-0268(2007)11:5(514).

[20] TUAKTA C, BUYUKOZTURK O. Conceptual model for prediction of FRP-concrete bond strength under moisture cycles [J]. Journal of Composites for Construction, 2011, 15(5): 743-756. DOI: 10.1061/(ASCE)CC.1943-5614. 000021.

[21] LI Yue, LIU Xiong-fei, LI Jia-qi. Bond properties of FRP-concrete interface with nano-modified epoxy resin under wet-dry cycles [J]. KSCE Journal of Civil Engineering, 2017, 21(4): 1379-1385. DOI: 10.1007/s12205-016-0921-7.

[22] LIANG Hong-jun, LI Shan, LU Yi-yan, YANG Ting. The combined effects of wet–dry cycles and sustained load on the bond behavior of FRP–concrete interface [J]. Polymer Composites, 2019, 40(3): 1006-1017. DOI: 10.1002/pc. 24785.

[23] ORTLEPP R, CURBACH M. Bonding behavior of textile reinforced concrete strengthening [C]// NAAMAN A E, REINHARDT H W, eds. Fourth International Workshop on High Performance Fiber Reinforced Cement Composites (HPFRCC4). France: RILEM Publications, 2003: 517-527.

[24] ORTLEPP R, HAMPEL U, CURBACH M. A new approach for evaluating bond capacity of TRC strengthening [J]. Cement & Concrete Composites, 2006, 28(7): 589-597. DOI: 10.1016/j.cemeoncomp.2006.05.003.

[25] XUN Yong, XIAO Bao-hui, ZHANG Qin. Experimental study on interfaces bond behavior of textile reinforced concrete (TRC) with RC component [J]. Sichuan Build Sci, 2011, 37(4): 55-59. (in Chinese)

[26] POURASEE A, PELED A, WEISS J. Fluid transport in cracked fabric-reinforced-cement-based composites [J]. Journal of Materials in Civil Engineering, 2011, 23(8): 1227-1238. DOI: 10.1061/(ASCE)MT.1943-5533.0000289.

[27] YIN Shi-ping, XU Shi-lang, WANG Fei. Investigation on the flexural behavior of concrete members reinforced with epoxy resin-impregnated textiles [J]. Materials and Structures, 2015, 48(1, 2): 153-166. DOI: 10.1617/s11527-013-0174-6.

[28] XU Shi-lang, SHEN Ling-hua, WANG Ji-yang, FU Ye. High temperature mechanical performance and micro interfacial adhesive failure of textile reinforced concrete thin-plate [J]. Journal of Zhejiang University-Science A, 2014, 15(1): 31-38. DOI: 10.1631/jzus.A1300150.

[29] LI He, XU Shi-lang. Development and optimization of cementitious matrices for textile reinforced concrete [J]. Journal of Hydroelectric Engineering, 2006, 25(3): 72-76. (in Chinese)

[30] NAGAMOTO N, OZAWA K. Mixture proportions of self- compacting high-performance concrete [C]// High- performance Concrete: Design and Materials and Advances in Concrete Technology, ACI SP-172. 1997: 623-636.

[31] JULIO G R, JOSE M S, PEDRO F, JOSE X. Effect of dry-wet cycles on the bond behaviour of concrete elements strengthened with NSM CFRP laminate strips [J]. Composite Structures, 2015, 132(12): 331-340. DOI: 10.1016/ j.compstruct.2015.05.053.

[32] YU Yu-lin, YIN Shi-ping, NA Ming-wang. Bending performance of TRC-strengthened RC beams with secondary load under chloride erosion [J]. Journal of Central South University, 2019, 26(1): 196-206. DOI: 10.1007/s11771- 019-3993-y.

[33] LI Yao, YIN Shi-ping, CHEN Wen-jie. Seismic behavior of corrosion-damaged RC columns strengthened with TRC under a chloride environment [J]. Construction and Building Materials, 2019, 201: 736-745. DOI: 10.1016/j.conbuildmat. 2019.01.009.

(Edited by FANG Jing-hua)

中文导读

氯盐干湿循环下纤维编织网增强混凝土与既有混凝土的界面性能

摘要：纤维编织网增强混凝土(TRC)适用于恶劣环境下混凝土结构的修补和增强。TRC与既有混凝土的界面性能是决定TRC加固效果的重要因素。本文通过双面剪切试验研究了氯盐干湿循环对TRC与既有混凝土界面处平均抗剪强度和滑移的影响，同时考虑了既有混凝土强度、粘结长度、加固层数和掺加短切纤维等因素。此外，还利用X射线衍射(XRD)技术分析了腐蚀环境中界面处的微观物质。试验结果表明，随着氯盐干湿循环次数的增加，TRC与既有混凝土的界面性能降低。与加固单层试件相比，加固两层且基体中掺入PVA或AR-玻璃短切纤维的试件可以进一步改善TRC层与既有混凝土的界面性能。对于具有单层纤维编织网的TRC，试件的平均抗剪强度随粘结长度的增加有减小的趋势。此外，既有混凝土的强度等级对TRC与混凝土界面性能的影响较小。

关键词：纤维编织网增强混凝土(TRC)；氯盐干湿循环；双面剪切；平均剪切强度；界面滑移；X射线衍射技术

Foundation item: Project(2017XKZD09) supported by the Fundamental Research Funds for the Central Universities, China

Received date: 2019-02-22; Accepted date: 2019-09-29

Corresponding author: YIN Shi-ping, PhD, Professor; Tel: +86-15262013916; E-mail: yinshiping2821@163.com; ORCID: 0000-0001- 8304-5914