Tensile Deformability of Shotcrete in Tunnel Primary Support: A Case Study

Abstract

Shotcrete strain in the primary support of a tunnel will produce non-loading strain at an early age due to the influence of its own temperature change, hardening shrinkage, spraying force, and other factors, which means that current strain-monitoring results fail to reflect the real strain, and the strain value after stabilization is high. In addition, tensile strain may be evident in the final result, even exceeding the tensile warning value, but, in actuality, the on-site lining is very stable, with no cracks or damage. Therefore, it is necessary to understand the strain characteristics of shotcrete in the primary support of a tunnel. Based on the long-span tunnel project at Shishan Road Station on the Qingdao Metro Line 6, in situ and indoor pull tests of concrete strain were designed while only considering temperature change, hardening shrinkage, and spraying force. This study shows the following: (1) The strain in shotcrete is greatly affected by temperature changes, hardening shrinkage, and shotcrete force in the first three days, reaching its peak value in the second to third days, while tending to be stable at about the seventh day. (2) The real strain of the shotcrete was tested, and the warning value was adjusted from 90 με to 120 με. (3) The strain value at the third day was taken as the initial value, and the previous monitoring results were revised. The revised results align with the trends shown during real tests performed on-site, providing guidance for tunnel engineering support monitoring.

1. Introduction

With the rapid development of underground engineering, geological disasters often occur in such projects, such as large deformation, rock burst, water inrush, and surface subsidence, causing casualties, equipment damage, project delays, engineering failure, etc. [1,2,3,4,5]. Using shotcrete support has been considered an effective measure to prevent such disasters. In recent decades, shotcrete has seen wide application in tunnel primary support [6,7,8,9]. However, shotcrete shows some detriments, such as low tensile strength, high brittleness, and susceptibility to tensile failure. Therefore, shotcrete is prone to cracking, spalling, the separation of layers, and insufficient thickness during construction, raising quality problems that greatly affect construction safety [10].

The tensile failure of shotcrete is a common instability phenomenon occurring in tunnel primary supports; thus, evaluating the tensile deformability of shotcrete is key to ensuring its quality. Common methods used for testing the tensile deformability of shotcrete materials include splitting tests, bending tests, and direct tensile tests [11,12]. A number of studies have obtained results on the tensile deformability of shotcrete through testing. For example, some scholars obtained full tensile stress–strain/deformation curves by improving the stiffness of the material testing machine [13,14,15]. Guo et al. [16] attached a rigid frame composed of a beam and a tie rod to an ordinary hydraulic material testing machine and derived the full tensile stress–deformation curve of the sample.

According to the above, shotcrete also plays a key role in primary supports, so the quality of the shotcrete used in the project needs to be strictly controlled. Therefore, the on-site monitoring of shotcrete is an important step to ensuring the safety and stability of the primary support. At present, there are many types of strain gauges used to monitor the quality of shotcrete in situ, such as resistance strain gauges, differential strain gauges, embedded vibrating string strain gauges, and optical fiber strain gauges [17,18,19]. The volumetric parameters of shotcrete change constantly with changes in cement hydration, water loss, temperature, and load conditions [20,21,22,23,24]; in effect, the strain that shotcrete is subjected to constantly changes after its pouring. According to our understanding of the causes of strain and the relationship between strain and structural stress, concrete strain can be divided into the following categories [25]: strain caused by temperature changes in the concrete; strain caused by concrete shrinkage (expansion); strain caused by concrete creep; and concrete strain caused by external load. Do et al. [26] proposed that in the process of the early hydration of mass concrete, the degree of temperature deformation in different parts of the interior is relatively large, resulting in strong mutual restraint, which is the main factor causing early internal stress in mass concrete. Witasse et al. [27] used Diana 10.5 analysis software to simulate an inverted T-shaped wall composed of mass concrete and analyzed the hydration heat therein, as well as the development of cracks. Yu et al. [28] used an optical fiber light grid sensor to monitor the temperature and strain within the inner part of concrete and thereby derived the thermal characteristics. All of these studies show that the strain that develops in concrete at an early age is greatly affected by changes in temperature and humidity, such that the strain gauge reading cannot be used to reflect the concrete strain caused by the external load. Taking the project building a tunnel at a station of the Qingdao Metro as an example, we see that this initial value makes the strain value of the concrete more stable, and no cracks appear in the shotcrete on-site. Obviously, the results obtained via this test method are not accurate. Therefore, it is of great necessity to study the tensile deformability of shotcrete to ensure tunnel safety.

In order to solve the aforementioned problem, the non-loading deformability of the shotcrete used in the tunnel’s primary support is studied herein. Specifically, through direct tensile tests of shotcrete samples, the ultimate failure strain values are determined in Section 2, and they are utilized together with the non-loading strain results to modify the monitored deformation values (Section 3). An innovative method is proposed to implement tests on the non-loading deformation of shotcrete in Section 3. A stress relief device of our own design is adopted; this can not only ensure that the concrete is subjected to local stress adjustments in its steel bars and other materials but also eliminate the influence of structural stress in a large arch frame. Meanwhile, the regular approaches used for deformation monitoring in shotcrete are presented in Section 3, where their shortcomings are also presented. This study is intended to provide guidance in the prediction of the stability of shotcrete used in tunnel primary support.

2. Determination of Tensile Failure Strain of Shotcrete through Laboratory Tests

2.1. Engineering Background

Shishan Road Subway Station on Qingdao Metro Line 6 is an underground station that was excavated with a length of 224 m. The station is situated 33~39 m deep, and the tunnel’s body is primarily set in lightly weathered granite, with a compressive strength of 55~90 MPa. In the vicinity, zones with joints and fractures have been developed. The tunnel’s primary support system consists of anchoring bolts, shotcrete, and a steel arch, as shown in Figure 1. The parameters of the monitoring sensors are shown in

Table 1. Monitoring sensor parameters.

Monitoring Items	Physical Quantity	Instrument Name	Specification
Bolt axial force	$F/\mathrm{KN}$	Bolt axial force gauge	MSJ
Concrete strain	${\epsilon }_c$	Concrete strain gauge	EM-30
Steel rib stress	$\sigma /\mathrm{MPa}$	Rebar stress gauge	STG-25

. The C30 shotcrete utilized had a mix ratio of cement, water, sand, and stone of 0.38:1:1.11:2.72. The thickness of the shotcrete was 300 mm, and the spraying pressure was strictly controlled within a range of 0.3–0.4 MPa.

In order to forecast the stability of the primary support system, the shotcrete’s deformation was monitored. Vibrating wire extensometers were utilized to monitor the values of deformation in the shotcrete, and their setup is displayed in Figure 2. Once the monitoring data had been obtained, the engineers were tasked with making a judgment as to the safety of the situation according to a warning value and taking appropriate measures. Therefore, a warning value of tensile strain in the shotcrete used in Shishan Road Subway Station was thus first determined through laboratory tests.

2.2. Test Scheme Design

To set a warning value for tensile strain, one must know the ultimate tensile failure strength of the shotcrete material used in the engineering field. Thus, a direct tensile test scheme was developed, and the results are presented in this section. Although a direct tensile test can be performed to easily and quickly determine the ultimate tensile failure strength of shotcrete materials, factors such as secondary bending, slip, and a high stress concentration at the end of the sample make such tests very difficult to perform accurately [29,30,31,32,33]. Thus, we propose an innovative sampling method to solve the above problem.

Firstly, test samples were cast with the C30 shotcrete that is used in the primary support in the Qingdao Metro station. In order to ensure that we derived an accurate and realistic tensile failure effect, and for the sake of convenience in mold-making, site casting, and the test operation, the diameter of the end section of the block used in this tensile test was set to 150 mm, the size of the middle section was 80 mm, the total length of the sample was 250 mm, and the length of the middle section was 90 mm. The vertical steel bar used in the steel claw was 80 mm, and four evenly distributed lateral steel bars were welded to the bottoms of the vertical bars. The length of each lateral steel bar was 60 mm, and their angles with the vertical steel bars were 45°. All of the bars were made of Φ18 HPB335 steel. The shapes and sizes of the samples are shown in Figure 3, and the preparation and curing process is shown in Figure 4. The test schedule and loading details are shown in

Table 2. Test schedule and loading details.

Item	Specification	Loading Rate	Quantity
Shotcrete	C30	0.002 mm/min	2
Strain gauge	120-88AA	-	3

.

2.3. Test Process

Before running the test, the test samples were polished with sandpaper and three resistance strain gauges (model 120–88AA) were attached within the stretching area, distributed uniformly in the vertical direction, to measure strain changes in the test samples. Each sample was mounted on a loading test machine. Then, the terminal was connected to the DH5902 dynamic data acquisition instrument (Jiangsu Donghua Testing Technology Co., Ltd., Taizhou, China), and the steel claws at both ends of the test block were clamped and fixed onto the electronic universal testing machine with an anti-rotation fixture. The normal load of the sample was given by the high-precision load sensor via the TEDS self-identification function of the electronic universal testing machine. The loading rate was always maintained at 0.002 mm/min until the sample became damaged (as shown in Figure 5a). The failure modes of the samples are shown in Figure 5b, where tensile cracks can be seen developing through the cross-section in the middle of the samples. Secondary bending, slip, and high stress concentrations did not occur in the tests, but the failure feature of the samples was obvious.

2.4. Analysis of Test Results

Through laboratory direct tensile testing, two types of stress–strain curves were seen to develop after testing without grouping the samples in advance, as shown in Figure 6. According to the curve type, the samples were divided into two groups, named sample group A and sample group B.

In the initial tensile stage, the strain changed less and the stress changed more, but the strain increased proportionally with the increase in stress. When the stress of curve B reached about 0.2 MPa, the strain increased gradually, and the curve began to show a rough convex shape. For curve A, when the stress reached 0.17 MPa, the curve began to show a convex shape. At the same stress level, the strain of curve B was slightly lower than that of curve A, and the strain growth rate of curve A was greater than that of curve B. When approaching peak stress, the development of strain in the sample was fast, the growth of stress was slow, and the curve remained almost level until the point of final failure. The peak stress values of the two curves were 1.0 MPa and 0.8 MPa, respectively. The peak strain values for the two sample groups were 141.4 με and 137.6 με, respectively, and the average value of the two was calculated at 139.5 με. The ultimate tensile strain at failure for the samples should thus be 139.5 με. The results show that the mean error of the peak strain in the two groups of samples was 1.9 με, which meets the accuracy requirements of the test. The differences in the stress–strain curves may be due to errors during sample fabrication and human errors.

According to the relevant specifications and our results, the safety warning value of strain should be set at 121.8 με. The warning value for C30 concrete as regards tensile strain is outlined as 90 με in the specifications, which is much lower than that given by our test results. Direct tensile testing can be used to reflect the ultimate tensile failure strength of concrete, and the effects of strain caused by temperature differences and hardening were excluded in this test. As such, the warning value as regards tensile strain for C30 concrete can be increased to 120 με, according to the results of our direct tensile testing.

3. Field Monitoring Results and Analysis

3.1. Regular Monitoring

The monitoring method used here involved setting vibrating string strain gauges at the positions of the foot, the waist, and the top of the grid arch. The strain gauges were tied to the steel ribs of the grid arch, distributed in the ring direction, and the lead wire was fed along the arch into the constructed concrete’s surface so that the strain in the shotcrete could be continuously monitored at a later stage. Finally, C30 concrete was sprayed onto the grid arch. Monitoring commenced immediately after concrete spraying. The natural vibration frequency of the strain gauge, set at the factory, was taken as the initial frequency f₀, and the natural vibration frequency measured here was taken as f. After that, monitoring was performed once a day. The strain value can be calculated using Equation (1) [34]:

$$\epsilon ={\displaystyle \frac{K}{EA}}\left(f^2-f_0^2\right)$$

(1)

where K is related to the length of the string, with the mass set per unit length, and EA is the axial stiffness of the steel string.

The variation in the strain of the shotcrete with time is shown in Figure 7. Note that positive strain is compressive strain, while negative strain is tensile strain. Accordingly, it can be seen that the shotcrete was mainly subjected to compression, but tension developed in some positions. The strain variation with time in the shotcrete shows a basically varying trend of “rise–stabilization”. Most notably, within a few days after concrete spraying, the strain value changed dramatically, which may have been caused by the large impact force produced on the strain gauge during the spraying process, the change in temperature of the concrete itself, and the hardening and shrinkage process, which would affect its real force. These factors resulted in the measured strain being much greater than the concrete’s actual strain. Finally, the value eventually tended towards stability at a high level. In addition, tension developed in some positions that necessitate early warning, according to relevant specifications. However, the strain at some of these significant positions exceeded the warning value, but no crack or damage could be observed in the concrete on-site. Obviously, this warning value is inconsistent with the actual conditions of the project and needs to be adjusted.

Table 3. Shotcrete tensile positions.

Position	Number of Measuring Points	Number of Tensile Measuring Points	Proportion of Tensile Measuring Points
Arch foot	124	34	27.42%
Arch waist	71	13	18.31%
Arch top	31	6	19.35%

shows that, during in situ monitoring, the arch foot, top, and waist all featured tension zones, which are ranked as follows in order of the rate of their occurrence, from high to low: arch foot > arch waist > arch top. Moreover, there were more arch foot measurement points in this study, and the number of tension zones at this position was the greatest, which has a significant impact on the safety of the tunnel. Therefore, the arch foot tension zone is the focus of this study.

3.2. Non-Loading Deformation Test of Shotcrete

3.2.1. Test Setup

Due to the large difference between the laboratory environment and an engineering site, the shotcrete deformation values measured in the laboratory will be quite different from those measured in the field. In order to study the deformation of shotcrete as induced by non-loading factors, such as temperature changes, hardening shrinkage, and spraying pressure, an innovative test scheme is proposed in the present study. Figure 8 shows a schematic diagram of the field layout of the strain test.

In the field test, a stress relief device of our own design was employed. The device is composed of four rubber plates. Figure 9a shows the dimensions of the stress relief device. The dimensions of the top plate and the bottom plate are 500 × 400 × 40 mm, and the dimensions of the side plate are 700 × 400 × 40 mm. The rubber plate is connected by way of a bolt anchor. The rubber material produces a strong buffer effect and permits the absorption of external stress. In addition, the device has a certain degree of internal headroom, which can eliminate the influence of external stress on the development of strain in the shotcrete itself. Figure 9b shows a schematic diagram of the stress relief device. Before installing this stress relief device, it is necessary to use an electric welding machine at the corresponding position to heat the steel until it leaks out of the surrounding rock mass, so as to avoid the influence of the steel on the strain of the shotcrete itself. In order to prevent the device from loosening and falling off, the device was welded and fixed with steel bars on the air-exposed side. After the installation of the device, it was necessary to lay strain transducers. Two strain transducers were laid on each stress relief device, set in the circumferential direction. In order to avoid contact between the strain transducers and the stress relief device, we used overhanging steel bars to fix the transducers and then proceeded with continuous monitoring for 10 days. The field installation process followed for the stress relief device is shown in Figure 10.

3.2.2. Test Results

Figure 11 shows the variations in the strain in the shotcrete with time, obtained via continuous monitoring for 10 days after the completion of the concrete spraying process. Figure 11 shows that the shotcrete was significantly affected by the spraying pressure, hardening shrinkage and temperature changes. Specifically, most of the shotcrete deformation was induced by compression, and the compressive strain range was 200~300 με. On the other hand, a small part of the shotcrete suffered from tension, and the tensile strain was about 200 με. The trend of variation in the shotcrete’s strain with time can be divided into two stages:

• The strain growth stage—Within 3 days after concrete spraying was completed, the internal temperature of the concrete began to drop sharply, resulting in temperature strain developing around the strain gauge. At the same time, the shotcrete began to harden and shrink, causing the strain gauge to record shrinkage strain. In addition, during the spraying of concrete, the spraying pressure generated by the spray gun was high and acted directly on the strain gauge such that it became stressed, thus producing additional strain.
• The strain stabilization stage—On the second to third days after the completion of concrete spraying, the internal temperature of the concrete tended to become consistent with the ambient temperature, and the self-hardening and shrinkage processes effectively ended. The strain rose very slowly under the action of external load and tended to be stable by the seventh day.

Above all, the strain in the shotcrete increased rapidly in the first 3 days, which was not caused by the real load but rather by the influence of changes in its own temperature, hardening and shrinkage, and spraying pressure. Therefore, the initial strain value of the shotcrete within the primary support of a tunnel can be altered based on the strain value measured on the third day.

3.3. Modification and Safety Assessment of Field Monitoring Results

Taking the data from some measurement points given in Table 1 as an example, according to the conclusions obtained via the in situ testing of shotcrete strain, the factory value of the strain gauge was adjusted to the strain value measured on the third day to give the initial value. According to Figure 12, after excluding non-loading strain, the curve showed a general change law of “rising to stability”, and the final stable value remained low. Only the left arch foot showed a higher stable strain value. At the same time, the tensile strain also decreased, and the warning value stabilized at 120 με. The modified test data are consistent with the actual values for the supports at the project site and are within the range of relative safety. The uncorrected data include the non-loading strain produced by the spraying force, temperature changes, and hardening and shrinkage; as a result, the test data are generally too high and deviate from the actual working conditions of the project. The modified monitoring results are within the range of the warning value, strongly suggesting that there is no cracking or damage in the lining. At the same time, this proves that the modifications performed on the warning value are accurate.

4. Conclusions

Focusing on the problems identified in the process of strain testing for shotcrete, specifically in the tunnels of the Qingdao Metro, a new method for the non-loading strain testing and indoor direct tensile testing of shotcrete is proposed in this paper. Based on the test requirements, the field test and casting molds of the tensile samples were designed to ensure that the concrete would be subjected to local stress adjustments in steel bars and other materials, and to exclude the influence of the stress within the large arch structure. Based on the above two tests, the following conclusions were obtained:

(1)

In the process of concrete pouring, the hydration of concrete releases a lot of heat. Concrete expands under heat. Due to the constraining effect of the upper and surrounding rock, the concrete was compressed. At the same time, the arch frame in the initial support restricted the concrete’s expansion. At the initial stage of concrete pouring, the humidity in the near-surface area of the concrete decreased rapidly, forming a large disparity between the moisture level near the surface and that inside the concrete, resulting in tensile strain in the near-surface area. As a result, the strain suffered by the concrete was mainly compressive, with some cases of tensile strain.

(2)

In view of the large tensile strain measured in the field test, an indoor direct tensile test was carried out to test the real strain when the primary shotcrete was damaged by tensile failure, and the warning value was adjusted from 90 με to 120 με.

(3)

This in situ strain testing of shotcrete excludes the influence of other factors such as arch structure, and the data measured concern the strain suffered by the concrete under the influence of its own hardening and shrinkage, as well as temperature changes and the spraying force. The strain of the concrete reached its peak at 2–3 days, after which there were small fluctuations, and it tended to become stable at around 7 days. The main reason for this is that the early strain of shotcrete is greatly determined by internal temperature changes and hardening and shrinkage, and the shotcrete spraying force greatly impacts the vibrating string strain gauge. Young shotcrete features a large degree of non-loading deformation resulting from shotcrete spraying.

(4)

The test results obtained on the third day were finally used as the initial value to calculate the monitoring results for the primary shotcrete, which remedied the inappropriate practice of taking the strain value at the end of the shotcrete spraying process as the initial value. The modified results are in line with the trends derived during real testing in the field, and they can provide guidance and support for tunnel engineering.