Experimental investigation and analytical modelling of shear strength of thin walled textile-reinforced UHPC beams

.


Introduction
The building construction industry is constantly on the lookout for sustainable, innovative materials and solutions in order to primarily decrease building costs and energy consumption in comparison to classical reinforced concrete. A material that has the potential to meet this requirement in certain areas of applications is textile-reinforced concrete (TRC). TRC is a combination of fine-grained concrete (generally high-performance concrete) and multi-axial textile fabrics (longitudinal and transversal rovings) typically made of aramid, basalt, alkaliresistant glass or carbon fibres. Over the last decade, German research institutes have been the main pioneers in the investigation of this highperformance composite material [1,2]. Recently, the research interest has also sparked in other countries, e.g. [3][4][5]. In various research projects the material (compression, tension, bond, etc.) as well as structural behaviour (bending, shear, etc.) of TRC using different types of textile fabrics was investigated [6][7][8][9][10][11][12] with the results are currently being summarized in a RILEM Technical Committee.
The material is characterized by a high durability, especially for carbon textiles [13], which allows for a reduction of the concrete cover and thus, the construction of slender, lightweight structures with a long estimated service life, particularly suitable for pedestrian bridges [14], shells [15][16][17] and façade panels [18]. In addition, the resulting weight reduction is of particular interest for the production of precast elements for building constructions [19,20]. Furthermore, it was discovered, that TRC is a suitable material for strengthening of existing structures like shells [21], columns [22], webs of shear critical beams [23,24] and others. Nowadays, mainly impregnated textiles (preferably carbon and ARglass) are used as reinforcement. Due to the impregnation, the tensile strength and bond behaviour of the fabrics is essentially enhanced by a better activation of the core filaments [25]. Thus, tensile stresses up to 3500 MPa can be achieved e.g. for carbon textiles. In order to further improve the bond strength of TRC the surface of the rovings can additionally be coated with sand [26]. Beyond that, it is possible to pre-shape the textile elements three-dimensionally preserving the form using an ensuing impregnation. Currently, acrylate and styrene based polymers (easily bent to limited bending radii) and epoxy resins (high stiffness) are the most commonly used impregnation materials [27]. A recent study [28] showed, that the possible reduction of concrete in TRC considerably decreases the cumulative energy demand and environmental impact in some rating categories. In certain areas of application, the ecological footprint lies below that of conventional reinforced concrete. However, this can only be ensured if structures are optimized accordingly and the high-performance materials used are fully utilized [10,20].
In this paper, the structural performance of thin-walled T-beams made of carbon textile reinforcement and ultra-high performance concrete (UHPC) is investigated by means of a test series of 18 shear tests on 9 T-beams. The construction elements tested and presented below could be used in practice as precast ribbed slabs for building constructions. While there are already several studies on the bending behaviour of TRC [10,29], the presented research focused on the little investigated shear behaviour of filigree, exclusively carbon reinforced T-beams. First preliminary results of this study are published in [30]. To complement the described test results this manuscript will focus on the detailed evaluation of the shear carrying capacity of the used textile reinforcement.

Specimens
To gain a better understanding of the shear behaviour of thin-walled T-beams made of carbon reinforcement (rods and textiles) and UHPC a comprehensive experimental programme was performed. The main investigating parameters consisted of the type of the textile reinforcement grid (impregnation: acrylonitrile-styrene-acrylate copolymers (ASA); type 1 vs. epoxy resin; type 2); different shear reinforcement ratios, the shear span to effective depth ratio a/d, and the thickness of the walls (web and flanges). Furthermore, the effect of prestressing of the carbon fibre reinforced polymer (CFRP) rods on the shear strength of carbon reinforced concrete was analysed. The main parameters are summarized in Table 1.
Three different types of test specimens were produced. Fig. 1 shows the dimensions as well as the reinforcement layouts of the different cross sections used in this study. The height and the width of the upper flange was the same for all specimens and was set to 260 mm and 300 mm respectively. To investigate the effect of different wall thicknesses, the web width was 30 mm for cross section type 1 (see Fig. 1a) and increased to 40 mm for cross sections types 2 and 3 (see Fig. 1b and c). Furthermore, the height of the compression flange (20 vs. 30 mm) as well as the width of the tension flange (100 vs. 120 mm) were varied. Another major difference between cross section type 1 and cross sections types 2 and 3 was the diameter of the used CFRP rods, which was changed from a diameter of 8 mm to a diameter of 10 mm. The only difference between specimens type 2 and specimens of type 3 was that a prestressing force was applied on specimens of type 3. In addition to the CFRP bars a flat carbon textile grid of the same type as the respective shear reinforcement was used as longitudinal reinforcement in the compression and tension flanges of each specimen.
The effective depth d of the textile reinforcement acting in tension is listed in Table 2. The effective depth d of the CFRP rods is shown in Fig. 1. The shear reinforcement consisted of two preformed L-shaped carbon textile grids. The L-shape grids were installed in such a manner that resulted in a straight anchorage in the bottom flange and a longer almost horizontal anchorage in the top flange. Two different types of textiles were used in this study (see Table 4) as shear reinforcement with varying cross sections of the rovings A rov , varying distances between the rovings s rov as well as different impregnation materials. Furthermore, single-sided textile pull-out the surface of the textile reinforcement with ASA-impregnation was additionally sand coated. Details of the transverse reinforcement are given in Table 2. The CFRP rods in the type 3 specimens (PTx.xS40) were prestressed in a stressing bed using a conventional prestressing jack by coupling the rod with a steel strand. To improve the bond, the rods in the coupler were notched (for detailed information on the prestressing system see [31]). An additional, prestressed carbon rod was added at the compression flange to avoid cracking due to the arising bending moment caused by the bottom focused pre-tensioning. The prestressing forces measured in the specimens on the day of testing are listed in Table 1

Test setup
A simple 3-point bending test setup (see Fig. 2), with an applied single load, was chosen for testing. The load was introduced at specific distances which would ensure clearly defined testing regions, where shear failure could be expected. To investigate the effect of shear slenderness on the shear strength, the position of the load introduction was varied. The tested a/d ratios are listed in Table 1. It was possible to test each specimens twice, since the testing region of the second test was situated in the cantilever section of the first test (see Fig. 2a). For the second test setup, attention was paid to the placement of the entire damaged area from the first test in the cantilever (see Fig. 2b). The ratio between the shear span and the total span was kept constant at 0.3.

Concrete
All beams were made of UHPC with a maximum aggregate size of 1.6 mm. The material tests were performed immediately after a set of two to four beam loading tests. The mean values of the cube compressive strength (on 100 mm × 100 mm cubes) and the flexural tensile strength varied for all beams between 159.7 and 173.0 MPa and 9.2 and 10.5 MPa, respectively. According to [32] the uniaxial tensile strength was calculated between 4.7 and 5.4 MPa. The material properties of the concrete are summarised in Table 3. One cubic metre of UHPC consisted of approximately 980 kg of quartz sand (one half 0.1-0.5 mm grain size and one half 1.0-1.6 mm grain size), 10 kg of mixing water, 750 kg of cement (CEM I 52.5 N C3A free), 60 kg of superplasticiser, 40 kg of shrinkage compensator and admixture (200 kg stone powder and 310 kg microsilica slurry) and 1 kg of defoamer.

Textile reinforcement
The material properties and dimensions of the various utilised carbon textile reinforcement are listed in Table 4. The test setup for the determination of the material parameters was chosen according to [34]. The tensile behaviour was linear elastic until reaching the tensile strength [20]. As will be described in detail later, the shape of the textiles plays an important role in the structural behaviour. The cross section of a single fibre strand of the used textiles has an elliptical shape. The connection of the warp and weft strands in the textiles type 2 by warp knitting results in a compression of the fibre strands at the intersection creating a certain regularly repeating asperity on the surface. In the textiles type 1, the connection between the two directions is facilitated by an intersection of the weft roving through the warp roving resulting in no regularly repeating variation in cross-sectional dimensions. Fig. 3 shows a comparison of the two textile fabric types by means of digital 3d-models of the fabrics generated by laser scans.
The bond behaviour of the used textiles was investigated after the shear tests showed pronounced spalling of the concrete cover due to high bond action (see section 3.1ff). A detailed description of the experiments can be found in [35,36]. In contrast to conventional metallic reinforcement bars, carbon textile reinforcement has no standardized test method, as the pull-out test according to RILEM [37], for determining the bond strength. Due to this fact the authors chose the singlesided textile pull-out (SPO)-test based on the work of Lorenz [38], in which a single roving is pulled out of a thin-walled concrete element. Details of the test setup are shown in Fig. 4a.
The thickness of the SPO-test specimens was chosen based on the beam web width of 30 mm. The width of the elements was set to 100 mm and reduced to 20 mm or 38 mm in the anchorage zone, according to   [33]. adaptions made by Bielak et.al [39], to account for a spalling of the concrete cover due to splitting tensile stresses. The bond length of the roving and the reduced width was chosen to be the mesh size of the grid (20 and 38 mm). In contrast to Lorenz [38], a crack was initiated by placing a PE sheet at the beginning of the bond length in the specimens before casting. This allowed for a more precise testing of the bond length without exceeding the tensile capacity of concrete beforehand. Fig. 4b shows the bond flow T (=F/l b ) in relation to the crack opening w for the two tested textile types (5 tests per textile type). The fibre strands with the ASA impregnation and sand-coated surface (textile type 1) reached a maximum mean bond flow of 49.6 N/mm. After reaching the maximum value, the bond flow dropped to 15.7 N/mm remaining at the value while the crack opening increased. The remaining bond strength can be traced back to the friction capacity of the sand-coated surface. The ultimate failure resulted in a total pull-out of the fibre strands in all test specimens without cracking of the concrete matrix. The fibre strands with 6400 tex and an impregnation made of epoxy resin (textile type 2) showed a completely different bond behaviour. The maximal bond flow T of 115.1 N/mm (mean value of 5 tests) surpassed that of the ASAimpregnated textiles. The main bond mechanism acting in this phase can be assigned to a mechanical interlock, which is enabled by the regularly repeating variation in the cross-sectional dimensions [40].
After the maximum bond flow T was reached, a longitudinal crack initiated, leading to a sudden drop to almost zero of the bond flow transmitted in the interface. Minimal slowly decreasing friction remained while the crack opening increased. The cause of the longitudinal cracking can, on the one hand, be traced back to the increased bond flow capacity due to the higher amount of filaments compared to textile type 1, on the other hand to the specific geometric properties of textile type 2, e.g. the elliptical shape of the fibre strand and a wedge effect on account of the regularly repeating variation in cross-sectional dimensions (for further details compare [41]). Furthermore, the epoxy resin impregnation is much stiffer than the ASA impregnation, resulting in a much more pronounced wedge effect. As it will be discussed later, this phenomenon has a significant influence on the shear behaviour of thin-walled beams with textile fabrics as shear reinforcement. In order to describe the bond behaviour, a simple multilinear bond law can be used. Fig. 4c illustrates the details regarding the bond strength and corresponding slip for both textile types. The derivation of the bond-slip relationship from the load-crack opening-diagram can be done either by semi-analytical [42] or numerical methods [43,48]. A difference in the concrete cover of the bond performance test specimens described in this section and the overall shear performance test specimens has to be noted. While the bond stiffness in the uncracked state is not affected significantly by the thickness of the concrete cover, which is supported by the results of the recalculations of the stress state in section 4.1, the bond strength for the epoxy-impregnated textiles may be overestimated due to the bigger concrete cover.

CFRP rods
Sand coated CRFP rods were used as bar reinforcement. The impregnation was made of epoxy resin. The sand coating was achieved by roughing the plain bars, coating them with epoxy resin and subsequently sprinkling them with quartz sand of a grain size between 0.3 and 0.7. This method provides a high bond strength (mean value of approx. 15.0 MPa) and a high bond stiffness (for detailed information on the bond performance see [44]). The material properties for the CFRP rod reinforcement were taken from material tests provided by the producer. The tensile strength of the bars was stated as 2,048 MPa, the modulus of elasticity as 161 GPa and the ultimate strain as 12.72 mm/m.

Measurements
During testing, continuous measurements of the load, deflections as well as strains on the concrete surface were recorded. In addition, the kinematics (opening and sliding of the crack surfaces) of critical cracks during the experiments could be analysed with the help of digital image correlation (DIC). The photogrammetric measurement system ARAMIS 4 M from the company GOM with two 4 megapixel cameras (resolution 2,352 × 1,728 pixels) was used. The post-processing of the images was performed with the DIC software GOM Correlate Professional 2016. A random high-contrast speckle pattern was applied to the concrete surface in order for the photogrammetric system to identify an overlapping grid of facets (15 × 15 pixels for the evaluation of crack kinematics). The image acquisition rate of the cameras was changed at specific load steps, and the frequency was increased to 5 Hz just before failure of the specimen occurred.

Description of the failure modes
The shear capacity V exp of all tests is summarized in Table 5. A shear induced failure was observed in all conducted experiments. Fig. 5 shows the crack pattern after failure for the tests with an a/d ratio of 4.0.
Different failure modes were observed throughout the varying series. Thereby, a clear tendency regarding the type of textile reinforcement (used as shear reinforcement) was visible. All tests with an epoxy resin impregnation and a high yarn count (textile type 2; see Table 4) exhibited a distinct spalling of the concrete cover due to longitudinal cracking in the web without a rupture of the textile fabrics (see Fig. 5b). As the bond performance was reduced significantly after initiation of a longitudinal cracking, a sudden loss of stiffness of the system was noticed; see Fig. 6. Subsequent failure occurred due to crushing of the concrete compression strut preceded by a distinct spalling of the concrete cover. The distinct spalling of the concrete cover in the case of epoxy-impregnated textiles is primarily caused by high splitting tensile stresses due to bond action (see section 4.3). A separation of the concrete cover due to the big surface area of textile fabrics is not supported by the data, as the ASA-impregnated textiles exhibited a bigger surface area, while simultaneously no spalling of the concrete was observed in these tests.
The ASA-impregnated TRC-beams (textile type 1) showed a typical flexural-shear crack failure (the critical shear crack formed from a flexural crack) with rupture of the textile fabrics and crushing or splitting of the compression zone near the load introduction. A third failure mode, a shear-tension failure, was observed in the specimen PT3.0S40 (prestressed beam) with an a/d ratio of 3.0. The critical shear crack was formed in the web next to the support leading to a fracture of the textile grid (ASA-impregnation) without spalling of the concrete cover.  Table 5) depending on several influencing parameters. In all the tests a clear trend regarding the influence of the shear span to effective depth ratio a/d was visible.

Effect of the influencing parameters
With an increasing a/d ratio, a decreasing shear resistance could be observed. Furthermore, it was noted that a thicker web and flange resulted in an improvement of the shear carrying capacity of around 20% (see Fig. 7a and b). Although the shear performance index ρ w ⋅f t,tex was more than four times higher for the beams with the textile reinforcement type 2, the difference of the shear carrying capacity only accumulated to 23% (compare Fig. 7a and b). The results showed that the most effective way (especially for the chosen configurations) to improve the shear capacity of carbon reinforced beams is by prestressing the CFRP rods. Although the prestressing force was relatively small (20% of the tensile capacity of the CRFP bars), an improvement of the maximum experimental shear force of 44% was achieved (see Fig. 7c). The various effects and mechanisms leading to the observed shear carrying capacity of the different specimens are discussed in the following sections.

Measured stress state in textile reinforcement at failure
Specimens reinforced with epoxy impregnated fabrics (type 2) failed at comparably low shear forces in contrast to their high shear performance index ρ w ⋅f t,tex (more than four times higher than specimens reinforced with ASA-impregnated fabrics; see Table 5). This low efficiency can be traced back to the initiation of a longitudinal cracking in the web at an early stage, causing a spalling of the concrete cover and subsequently leading to a premature failure due to a significant degradation of the bond performance. The initiation of the longitudinal cracking is dependent upon the geometric and mechanic properties of the textile. The stress state in the textile, which allows conclusions on the effectiveness of the reinforcement at initiation of the spalling, can be determined in terms of measured crack-opening at the location of the textile yarns. The development of the crack-opening of the critical shear crack during testing at the place of a representative fibre strand crossing the crack is depicted in Fig. 8 for all tests specimens.
In reference to the bond performance tests described in section 2.3.2, longitudinal cracking is induced at a low crack-opening of the primary Table 5 Model parameters and results of the testes specimens.
For the calculation of the stress state with a given (measured) crack opening, the bond slip at the crack is needed. This value is obtained by numerically integrating the differential equation of the relocatable slip (see Eq. (2), with τ(s) being the bond-slip relationship derived from the bond performance tests described in section 2.3.2 (see also Fig. 4, respectively Fig. 10).
The obtained stress-crack opening diagrams for both types of textile reinforcements are depicted in Fig. 8c.

Evaluation of shear strength due to textile reinforcement
By means of the stress-crack-opening diagrams given in section 4.1, the stress state in the fabric at failure σ tex,eff was evaluated for all test specimens. The calculated data is given in Table 5. It can be seen that the stress state at failure is significantly lower for the epoxy impregnated textiles when looking at it in relation to its higher tensile strength. In addition, the inclination of the critical shear crack was also evaluated and is listed for all specimens in Table 5. The method of determination of the critical shear crack inclination β cr is depicted in Fig. 9 (a). The typical curved form of the shear crack was idealized with a straight line which connected the junction of the crack with top and lower flange. The shear carrying capacity of the textile reinforcement was then calculated according to Eq (3).
The calculated values of the shear resistance V R,tex are given in Table 5. In Fig. 9 (b) and (c) the shear carrying capacity of the textile reinforcement is graphically illustrated as part of the shear resistance of the specimens -V exp in the case of the ASA-impregnated textiles and V spal in the case of the epoxy-impregnated textiles. It can be seen, that the shear resistance of the textile reinforcement V R,tex was accountable for almost the full shear resistance, while the shear carrying capacity of the concrete V R,c was comparably low. The contribution of the textile reinforcement varied between 54% and 98% of the shear strength of the beams. Furthermore, the relation of the attribution of the fabrics to the shear carrying capacity decreased with an increasing width of the web and the depth of the top flange. A possible physical explanation could be the effect of the compression flange (acting as Timoshenko-beam) as the Fig. 5. Crack pattern at failure for TRC specimen: (a) reinforced with ASA-impregnated textile; (b) reinforced with epoxy impregnated textile; (c) reinforced with ASA-impregnated textile and prestressed CFRP rods [30].  (c) normal stresses due to prestressing σ cp = P test /A c [30].
shear-transfer action takes over. Generally, it must be stated, that the shear-transfer actions for beams with moderate shear slenderness (a/ d ratio > 3.0 for tests with point loads) contributed to concrete (e.g. [45]) are rather negligible. Moreover, the small aggregate size in combination with the relatively large crack openings, normally results in little contributions of aggregate interlock as well as shear strength in the fracture process zone. In addition, the dowelling effects of the longitudinal CFRP rods are not pronounced [9], with no dowelling cracks observed along the rebars during testing. On the other hand the shear strength due to arching action plays a very important role in beams with a low shear slenderness. A possible way to take this effect into account, is to use strut-and-tie models [46] or to increase the shear resistance by the factor β = 3/(a/d) if a/d is smaller than 3 [47].

Efficiency of used textile reinforcement
In Fig. 10b the bond stress τ b acting at the interface between the textile and the surrounding cementitious matrix is depicted for a shear crack crossing the fabric with an angle of α cr . Additionally the force progression in the textile F tex (Fig. 10c) as well as the concrete F c (Fig. 10d) can be seen. The used textiles in the experimental programme exhibit significantly different bond behaviours. While the ASAimpregnated textiles (type 1) are able to transfer bond stresses due to friction once the maximum bond stress τ b is reached, the bond performance of the epoxy impregnated textiles (type 2) is achieved by mechanical interlock. Due to the mechanical interlock, high splitting tensile stresses occur, which lead to a concrete failure once F c,spal is surpassed.  Splitting tensile stresses are significantly less distinct for the ASAimpregnated textiles resulting in failure once the textile tensile strength F t,tex is reached. The reduction in the tensile strength in comparison to the uniaxial tensile tests can be mainly traced back to the inclined direction of the crack and the resulting non-uniaxial loading of the textile. This phenomenon leads, in the case of the prestressed specimens, to a higher efficiency factor due to the less steep inclination of the critical shear crack compared to that of non-prestressed specimens.
With the method described in section 4.1 a stress state in the textile reinforcement at failure for the epoxy impregnated textiles of approx. 900 MPa for a web thickness of 30 mm (crack width at failure of 0.20 mm; specimen T2.5E30 excluded as no excessive spalling was observed in this case) and 1000 MPa for a web thickness of 40 mm (crack width at failure of 0.24 mm) was calculated. This corresponds to an efficiency factor when related to the tensile strength of the fabrics of 0.27 and 0.30 respectively. The results are also in accordance with the bond performance tests described in section 2.3.2, as they exhibited an equivalent stress state in the fabric at failure. The same method is chosen for the ASA-impregnated textiles. The stress state at failure of approx. 1700 MPa is calculated. This corresponds to an efficiency factor of 0.82 for both web thicknesses of the non-prestressed specimens as well as the prestressed specimens (crack width at failure of 0.56 mm). As it can be seen here, the effectiveness of ASA impregnated textiles in this case is much higher than of the epoxy impregnated fabrics with a high yarn count. This correlates with the shear carrying capacity, as the enhancement of the shear carrying capacity between ASA impregnated textiles to those impregnated with epoxy resin was on average only 23% although the shear performance factor was more than four times higher.
The lower effectiveness of both textile fabrics will be considered in the following by specifying the textile stress σ tex,eff that can be transmitted via the shear crack. As the lower level of effectiveness in the case of the two impregnation materials are resting upon two different mechanisms, the effectiveness is considered with two different reduction factors, see Eq. (4).
In the case of epoxy impregnated textiles a bond reduction factor k b is introduced, which considers a premature shear failure of the specimen induced by a longitudinal cracking in the web in the layer of the textile fabric. For ASA-impregnated textiles the reduced effectiveness of the reinforcement is considered by a reduction factor k T implemented on the textile strength, which is primary influenced by the non-uniaxial loading of the fabrics as in the past it was shown that the tensile strength of the fabrics decreases with increasing deviation from a uniaxial load condition. Nonetheless also other effects that may decrease the tensile strength of the fabric can be considered within factor k T as for example an unfavourable folding radius of the fabric. In Table 6 the suggested parameters k b and k T on the base of measured crack openings in this study are summarized.
As the initiation of a spalling is dependent upon the resistance of the concrete, bond reduction factor k b is influenced by the web thickness and has to be currently determined experimentally for every different combination of concrete and textile characteristics, e.g. by the bond performance test described in section 2.3.2. Amongst other parameters influencing the initiation of spalling (e.g. the compression force in the concrete cover, tension force in the out-of-plane direction), it was shown that the geometry of the fibre strand is a decisive parameter influencing the size and magnitude of the resulting splitting forces due to bond action [40], which lead the aforementioned longitudinal cracking. Alternatively the efficiency factor can be calculated by means of numerical solution of the differential equation of the relocatable slip and comparison of the resulting splitting tensile stresses with a concrete resistance against spalling, as it was described in [48]. As no spalling of the concrete cover was observed in the case of ASA-impregnated textiles, factor k b was therefore set to be one in size. In contrast, factor k T is only dependent upon the textile characteristics, therefore it can be determined for different fabrics individually and subsequently used for the calculation of other building components. As for epoxy impregnated textiles no rupture of the fabrics was observed in this study, parameter k T was only determined for the ASA-impregnated textiles in this case.

Conclusions
In the course of a research project at TU Wien, ultra-high performance concrete (UHPC) and carbon reinforcement (textiles and rods) were combined to create lightweight, sustainable thin-walled structural elements. In this paper a test series investigating the shear performance of such elements, specifically T-beams with and without prestressed CFRP rods, is presented. Based on the influencing parameters investigated and the results obtained, the following conclusions can be drawn: • The shear carrying capacity differs widely depending on the used textile reinforcement and the web and flange thickness of the Tbeams. A thicker web and flange brings an improvement in the shear carrying capacity of around 20%. The shear carrying capacity of the epoxy-impregnated textiles was noted to be higher by only 23% in average compared to that of the ASA-impregnated textiles with the shear performance factor surpassing it by the factor four. It is noticeable that prestressing of the CFRP rods is an efficient way to improve the structural behaviour of thin-walled carbon textile reinforced T-beams. This is especially evident in the shear carrying capacity. • Based on the DIC-measurements the crack width of the critical shear crack was evaluated and the stress state within the textile fabrics was calculated under consideration of bond models derived from bond performance tests. It was seen, that in both cases the uniaxial tensile strength of the fabrics could not be fully exploited. While there is a reduction of the tensile strength of the fabrics due to a non-axial loading for the ASA-impregnated textiles, the specimens reinforced with epoxy-resin impregnated textile shear reinforcement tend to fail at an early load stage due to a spalling of the concrete cover in the web. As a consequence, the high tensile strength of the reinforcement cannot be exploited efficiently. As shown in various studies this failure mechanism can be tracked down to high splitting tensile stresses [40].  • In contrast to steel-reinforced concrete, where longitudinal cracking is avoided by minimum values of concrete cover or a minimum amount of shear reinforcement overlapping the longitudinal reinforcement, such detailing would often oppose the design philosophy of TRC of thin-walled filigree structural components. The focus therefore should lie on the consideration of the reduced effectiveness of the textile reinforcement by means of appropriate design models. However, if the utilization rate is too low, it is possible that, in view of the high price of the material, economic efficiency may no longer be guaranteed. This design philosophy will therefore lead to a balancing act between reduced efficiency of the textile reinforcement on the one hand, and a reduction of concrete consumption on the other. • The determination whether a TRC-component tends to concrete spalling can either be done experimentally using a SPO-test with adaptions according to Bielak et al. [39], or by design models (see for example [49]). The lower effectiveness of textile fabrics can then be considered by specifying the textile stress σ tex,eff that can be transmitted via the shear crack. As the lower level of effectiveness in the case of ASA-impregnated textiles and epoxy-impregnated textiles rest upon different mechanisms, the two efficiency factors k b and k T are introduced. While factor k b considers a premature shear failure of the specimen induced by a longitudinal cracking in the web, factor k T accounts for different mechanisms which reduce the breaking strength of the textile reinforcement (for example an unfavourable folding radius of the fabric).

Declaration of Competing Interest
The authors declared that there is no conflict of interest.