**1. Introduction**

Shear failure may occur near the geometric discontinuity or joint interfaces of concrete structures, where the cracks are usually perpendicular to the axis of the member, without a bending moment. This shear behavior is known as "direct shear" [1]. Direct shear failure is a sudden and catastrophic failure mode in traditional concrete structures [2]. Although this behavior has been studied on ordinary concrete structures for more than 40 years, it was still not clear if the empirical models can accurately predict the actual shear behaviors [1]. Ultra-high performance concrete (UHPC) is a novel fiber reinforced concrete (FRC) with high strength, excellent service durability and low permeability [3–6]. It has been extensively used in buildings, bridges and other structural projects with a thin-walled structure [7–9]. Thin-walled reinforced concrete structures subjected to a distributed load of short duration (such as explosive loading and seismic load) may not behave plastically at the mid-span and fail there. Some of the beams might fail at positions very close to the support owing to direct shear failure [10]. Hence, it is very important to study the direct shear properties of UHPC structures. Through the numerous experimental studies of FRC [11–13], it had turned out that steel fibers have a great effect on the improvement of the shear properties of concrete. Due to the poor interface

properties between the coarse aggregates and cements and less fiber volume fraction used in FRC, the shear failures mostly show brittle modes. However, unlike ordinary concrete, FRC and high-strength concrete, UHPC has no coarse aggregates and possesses high compactness, as well as a high fiber volume fraction. Thus, the shear behavior of UHPC may be different from those of traditional concrete, and its shear failure process is worth studying.

Owing to the limitation of the mixing of UHPC, transportation and maintenance ability, structures will inevitably show joint connection problems [14–16]. Even in precast UHPC structures, there still remains some components or joints of segments to be cast in situ [17]. Thus, the study of shear properties at joint interfaces is very important for all composite structures [17–20]. At present, the ultimate limit and serviceability limit state that calculation for traditional concrete structures, because of its low tensile strength, do not need to take into account concrete tensile strength [21,22]. However, the tensile strength of UHPC is very well owing to the bridging effect of continuous steel fibers, and the utilization of the tensile capacity of UHPC has a significant impact on its economic rationality. For segmental precast UHPC structures, current practice is to use multiple key joints that are generally unreinforced and may be dry or epoxied [23–25]. For this reason, it could not utilize the beneficial effect of the continuity of steel fiber distribution at the interface and the natural occlusion between the UHPC aggregate matrix well. Hyun-O Jang [26] studied the shear properties of UHPC specimens with Z-shaped specimens. Their results show that the shear strength of the waterjet treated specimens (WJTSs) can reach 32.2% of the monolithic placement specimens (MPSs). However, it is insufficient that each group of specimens contain only one specimen without considering different types of steel fibers. Direct shear performance with different interface treatments for segmental cast-in-place UHPC is worthy of further research.

The objective of this paper is to obtain the failure modes, shear strength and shear slip properties of UHPC in situ wet joints. The variables include fiber types (13 mm straight-type steel fibers (13SSF), 13 mm hooked-type steel fibers (13HSF) and 16 mm hooked-type steel fibers (16HSF)), fiber volume fractions (2.0%, 2.5% and 3.0%) and interface treatment. The direct shear tests were performed on 15 MPSs and 12 WJTSs (flat joints). Besides, in order to evaluate the load-carrying capacity of cast-in-place UHPC structures, an interaction formula with respect to shear and compressive strength and the relative reduction of the shear strength ratio for the WJTSs are offered on the basis of the experimental results.

### **2. Experimental Program**

### *2.1. Experimental Specimens*

Herein, to make the shear transfer of the specimens more consistent with that of the segmental concrete structure, the direct shear test of Z-shaped specimens was used. Dimensions of each specimen are given by 200 mm × 400 mm × 100 mm, in which the shear plane size is 100 mm × 200 mm. In order to avoid the failure of other parts prior to the shear plane, the reinforced bar with a diameter of 8 mm was arranged in these specimens for strengthening. The dimensions for the specimens are illustrated in Figure 1.

The UHPC mixture used in the tests is composed of cementitious material (mix of Portland cement, silica fume and mineral powder), quartz sand and solid polycarboxylate superplasticizer (water reducing efficiency of 30%), and the mix proportion of UHPC is shown in Table 1.

To study the effect of steel fiber types on the shear properties of UHPC structures, three kinds of steel fibers were selected (see Figure 2), namely 13 mm straight-type steel fibers (13SSF), 13 mm hooked-type steel fibers (13HSF) and 16 mm hooked-type steel fibers (16HSF), respectively. The physical and mechanical properties of the steel fibers are shown in Table 2. Besides, three kinds of fiber volume fraction were selected to study the effect of steel fiber volume fraction on the UHPC shear properties, which is 2.0%, 2.5% and 3.0%, respectively (Table 3). In this way, 15 MPSs (Figure 3a) and 12 WJTSs (Figure 3b) were fabricated.

**Types Density** 

**(g·cm<sup>−</sup>3)** 

stirring process lasted about 3 min.

MPSs

WJTSs

and "NN" stands for non-doped fibers.

**(mm)** 

**Figure 1.** Direct shear test specimen. **Figure 1.** Direct shear test specimen.

The UHPC mixture used in the tests is composed of cementitious material (mix of Portland **Table 1.** Mix proportion of the ultra-high performance concrete (UHPC; mass).


**Table 1.** Mix proportion of the ultra-high performance concrete (UHPC; mass).

**Length Diameter Ratios of Length– of Tensile Strength Figure 2.** Steel fiber types: (**a**) 13SSF, (**b**) 13HSF and (**c**) 16HSF. **Figure 2.** Steel fiber types: (**a**) 13SSF, (**b**) 13HSF and (**c**) 16HSF.

**Modulus** 

**Elasticity** 

MP25S13 2.5% 13SSF

MP20H16 2.0% 16HSF MP25H16 2.5% 16HSF MP30H16 3.0% 16HSF

WJ25S13 2.5% 13SSF WJ25H13 2.5% 13HSF WJ25H16 2.5% 16HSF WJ-NN 0 /

**Volume Fraction Types of Steel Fibers** 

**External Features** 


**Diameter (GPa) (MPa)**  The following steps were conducted to mix the UHPC ingredients: **Table 2.** Physical and mechanical properties of the steel fibers.

**(mm)** 

**Specimen Type Specimen Number Steel Fiber** 


balling and to produce a concrete with uniform material consistency and good workability. Stirring occurred until the steel fibers were well encapsulated and evenly distributed in the slurry. The

**Table 3.** Shear specimen parameters.

Number Description: "MP" stands for the condition of monolithic placement, "WJ" stands for the condition of waterjet treatment; "20", "25" and "30" stands for the volume fraction of steel fibers; "H" stands for the hooked-type fibers, "S" stands for the straight-type fibers; "13" and "16" stands for the length of the fibers;

The MPSs were poured at one time and cured by high-temperature steam above 95 ± 3 °C for 48 h after conventional standard curing (temperature 20 ± 2 °C, relative humidity 95%) for two days. In contrast, the preparation of WJTSs appears more complex and follows the procedures below. First, the first portion was poured and cured in the standard environment for two days. Secondly, the shear bond interface of the first portion was treated by the high-pressure waterjet, and they were cured in high-temperature steam condition for 36 h. Then, the second portion was poured and cured in the


**Table 3.** Shear specimen parameters.

Number Description: "MP" stands for the condition of monolithic placement, "WJ" stands for the condition of waterjet treatment; "20", "25" and "30" stands for the volume fraction of steel fibers; "H" stands for the hooked-type fibers, "S" stands for the straight-type fibers; "13" and "16" stands for the length of the fibers; and "NN" stands for non-doped fibers. *Materials* **2017**, *10*, x FOR PEER REVIEW 5 of 17 standard environment for 48 h. Finally, whole specimens, including the first and second portions,

were cured in high-temperature steam conditions for 48 h.

(**c**) (**d**)

**Figure 3.** Manufacturing of Z-shaped specimens: (**a**) monolithic placement specimens (MPSs), (**b**) waterjet treated specimens (WJTSs), (**c**) finished specimens after maintenance and (**d**) all completed specimens. **Figure 3.** Manufacturing of Z-shaped specimens: (**a**) monolithic placement specimens (MPSs), (**b**) waterjet treated specimens (WJTSs), (**c**) finished specimens after maintenance and (**d**) all completed specimens.

The following steps were conducted to mix the UHPC ingredients:

In order to utilize the beneficial effect of the steel fibers continuity distribution and the natural occlusion between the UHPC aggregate matrix at the shear plane, the treatment of the shear bond (1) In a dry mixer (pre-wetting), dry components (cement, silica fume and mineral powder) were added and mixed for 2 min.

interface needs to ensure the retention of steel fibers and the length of fiber exposure as far as possible. (2) Then the mixer was suspended, and fine quartz sand added and stirred for 1 min.

fibers at the UHPC shear plane, which has been widely accepted by engineers [26,27].

As shown in the Figure 4, the roughing on the concrete surface with a high-pressure waterjet is converging water flow at a point through a high-pressure device and the energy will be greatly weakened once the water flow impacts the specimen surface. Therefore, the process will not cause (3) The required solid superplasticizer was poured into the total water outside of the mixer and the solution was added to the mix gradually and stirred for 4 min.

damage to the inside of the specimen. It is theoretically possible to retain a certain amount of steel

(4) Finally, steel fibers were added manually by slowly sprinkling them into the mixer, to avoid balling and to produce a concrete with uniform material consistency and good workability. Stirring occurred until the steel fibers were well encapsulated and evenly distributed in the slurry. The stirring process lasted about 3 min.

The MPSs were poured at one time and cured by high-temperature steam above 95 ± 3 ◦C for 48 h after conventional standard curing (temperature 20 ± 2 ◦C, relative humidity 95%) for two days. In contrast, the preparation of WJTSs appears more complex and follows the procedures below. First, the first portion was poured and cured in the standard environment for two days. Secondly, the shear bond interface of the first portion was treated by the high-pressure waterjet, and they were cured in high-temperature steam condition for 36 h. Then, the second portion was poured and cured in the standard environment for 48 h. Finally, whole specimens, including the first and second portions, were cured in high-temperature steam conditions for 48 h.

In order to utilize the beneficial effect of the steel fibers continuity distribution and the natural occlusion between the UHPC aggregate matrix at the shear plane, the treatment of the shear bond interface needs to ensure the retention of steel fibers and the length of fiber exposure as far as possible. As shown in the Figure 4, the roughing on the concrete surface with a high-pressure waterjet is converging water flow at a point through a high-pressure device and the energy will be greatly weakened once the water flow impacts the specimen surface. Therefore, the process will not cause damage to the inside of the specimen. It is theoretically possible to retain a certain amount of steel fibers at the UHPC shear plane, which has been widely accepted by engineers [ *Materials* **2017** 26,27]. , *10*, x FOR PEER REVIEW 6 of 17

(**a**) (**b**)

temperature.

*2.2. Material Properties* 

(*fcu*) and flexural strength (*fcf*).

**Figure 4.** Interface treatment with a high-pressure waterjet. (**a**) Waterjet; (**b**) surface treatment. **Figure 4.** Interface treatment with a high-pressure waterjet. (**a**) Waterjet; (**b**) surface treatment.

Figure 5 illustrates the variation of cubic compressive strength of UHPC with curing days under normal temperature. UHPC strength develops rapidly after initial solidification. After 4 days of maintenance, the cubic compressive strength of UHPC is close to 90 MPa. Therefore, in order to ensure the shear bond interface possesses excellent chiseling effect, the specimens should be Figure 5 illustrates the variation of cubic compressive strength of UHPC with curing days under normal temperature. UHPC strength develops rapidly after initial solidification. After 4 days of maintenance, the cubic compressive strength of UHPC is close to 90 MPa. Therefore, in order to ensure the shear bond interface possesses excellent chiseling effect, the specimens should be controlled to chisel after 1.5~4 days of maintenance in practical engineering.

controlled to chisel after 1.5~4 days of maintenance in practical engineering. Owing to the existence of a large number of steel fibers at the interface, it is difficult to accurately measure the interface roughness by conventional measurement methods. Thus, the effect of the roughness of the interface is not considered in this experiment. Considering the complexity of steel fiber dispersion, the distribution quantity of steel fiber types is also not considered.

fiber dispersion, the distribution quantity of steel fiber types is also not considered.

Owing to the existence of a large number of steel fibers at the interface, it is difficult to accurately measure the interface roughness by conventional measurement methods. Thus, the effect of the roughness of the interface is not considered in this experiment. Considering the complexity of steel

Three cubic compressive specimens (100 mm × 100 mm × 100 mm) and three flexural specimens (100 mm × 100 mm × 400 mm) were prepared to obtain the UHPC material properties. All specimens were cured under the same environment to determine the actual strength of UHPC materials during the test. As illustrated in Figure 6, some material properties experiments of UHPC have been carried out. Table 4 summarizes the material properties of UHPC, including the cubic compressive strength

controlled to chisel after 1.5~4 days of maintenance in practical engineering.

(**a**) (**b**)

**Figure 4.** Interface treatment with a high-pressure waterjet. (**a**) Waterjet; (**b**) surface treatment.

Figure 5 illustrates the variation of cubic compressive strength of UHPC with curing days under normal temperature. UHPC strength develops rapidly after initial solidification. After 4 days of maintenance, the cubic compressive strength of UHPC is close to 90 MPa. Therefore, in order to

**Figure 5.** The change in cubic compressive strength of UHPC with curing days under normal temperature. **Figure 5.** The change in cubic compressive strength of UHPC with curing days under normal temperature.

## *2.2. Material Properties*

test.

MN25H13

MN20H16

MN25H16

MN30H16

7.

Owing to the existence of a large number of steel fibers at the interface, it is difficult to accurately measure the interface roughness by conventional measurement methods. Thus, the effect of the roughness of the interface is not considered in this experiment. Considering the complexity of steel fiber dispersion, the distribution quantity of steel fiber types is also not considered. *2.2. Material Properties*  Three cubic compressive specimens (100 mm × 100 mm × 100 mm) and three flexural specimens Three cubic compressive specimens (100 mm × 100 mm × 100 mm) and three flexural specimens (100 mm × 100 mm × 400 mm) were prepared to obtain the UHPC material properties. All specimens were cured under the same environment to determine the actual strength of UHPC materials during the test. As illustrated in Figure 6, some material properties experiments of UHPC have been carried out. Table 4 summarizes the material properties of UHPC, including the cubic compressive strength (*fcu*) and flexural strength (*fcf Materials* **2017**, *10*, x FOR PEER REVIEW ). 7 of 17

152.21

158.24

184.38

*2.3. Loading Process and Measuring Arrangement* 

(**a**) (**b**)

**Figure 6.** Performance test of UHPC material: (**a**) cubic compression test and (**b**) four-point bending **Figure 6.** Performance test of UHPC material: (**a**) cubic compression test and (**b**) four-point bending test.

**Table 4.** Test results of the UHPC materials (unit: MPa).


The shear test was conducted on a 2000 kN universal testing machine. In order to examine deformation characteristics, a set of two linear variable differential transducers (LVDTs) was installed on the vertical direction of the specimen to measure the relative deformation under direct shear load at the construction joint. Besides, a set of two LVDTs was arranged at the center of the horizontal shear plane of the specimen for the purpose of measuring the variation of crack width along with the increase of load. To ensure that the LVDTs and strain gauges were fixed firmly as well as the test device was connected reliably, preloading was carried out before formal loading. Besides, when the load-carrying capacity of the specimen drops sharply, the test machine will automatically stop loading. In the process of formal loading, the condition of the crack initiation and extension was observed directly by a high-power magnifier (zoom in 30 times). The test set-up is shown in Figure

160.77 35.68 36.43 0.033

150.38 32.24 32.81 0.035

173.12 46.89 42.66 0.077

31.78

44.52

42.17

161.43 0.012

159.42 38.11

150.25 0.011

148.15 34.42

164.21 38.51

178.50 0.026

178.01 38.92

where τ

the shear plane area of specimens (mm2).

**3. Experimental Results and Discussion** 

*3.1. Test Results and Analysis of the MPSs* 

3.1.1. Load-Carrying Capacity and Failure Modes


**Table 4.** *Cont.*

### *2.3. Loading Process and Measuring Arrangement*

The shear test was conducted on a 2000 kN universal testing machine. In order to examine deformation characteristics, a set of two linear variable differential transducers (LVDTs) was installed on the vertical direction of the specimen to measure the relative deformation under direct shear load at the construction joint. Besides, a set of two LVDTs was arranged at the center of the horizontal shear plane of the specimen for the purpose of measuring the variation of crack width along with the increase of load. To ensure that the LVDTs and strain gauges were fixed firmly as well as the test device was connected reliably, preloading was carried out before formal loading. Besides, when the load-carrying capacity of the specimen drops sharply, the test machine will automatically stop loading. In the process of formal loading, the condition of the crack initiation and extension was observed directly by a high-power magnifier (zoom in 30 times). The test set-up is shown in Figure *Materials* **2017**, *10*, x FOR PEER REVIEW 7. 8 of 17

**Figure 7.** Test set-up. (**a**) Photo of test set-up and (**b**) arrangement of the transducers. **Figure 7.** Test set-up. (**a**) Photo of test set-up and (**b**) arrangement of the transducers.

After loading, the shear strength of the UHPC specimens under ultimate load can be obtained from Equation (1): After loading, the shear strength of the UHPC specimens under ultimate load can be obtained from Equation (1):

$$
\tau = \frac{F\_{cr}}{A} \tag{1}
$$

*A Fcr* (1) where τ represents the shear strength (MPa), *Fcr* represents the ultimate load (kN) and *A* represents the shear plane area of specimens (mm<sup>2</sup> ).

represents the shear strength (MPa), *Fcr* represents the ultimate load (kN) and *A* represents

the MP25H16 specimen as an example. Firstly, there were no changes in the surface of these specimens at the initial loading stage. As the loading continued, the fragments at the shear plane began to exfoliate. After that, the initial cracks appeared on the shear plane and several small cracks appeared instantaneously. Herein, it should be noted that the crack width of 0.05 mm is adopted as the criterion of visible initial cracking [28]. With the increase in loading, fine cracks further spread, connected and penetrated to form a crack zone along the shear failure surface, and the fibers between the crack zones were gradually pulled out or pulled off. Finally, with the further increase in the load, along with a huge sound, specimens were sheared and damaged. The condition of the crack development is shown in Figure 8. In order to verify whether the UHPC specimens still possess the bearing capacity after the main crack occurs, the test machine was restarted to continue loading. It turns out the load could still reach 1/2~2/3 of the ultimate load. On the basis of the testing results, it can be seen that there are two main crack modes in the failure modes for these specimens, namely

the single main crack (Figure 9a) and multiple diagonal cracks (Figure 9b), respectively.
