1 Introduction
When used on high-strength and highly abrasive rock formations, tunnel boring machines (TBMs) exhibit undesirable characteristics, such as low permeability, severe disc cutter wear, overload breakage in the disc cutter ring or bearing, and frequent cutter replacement and maintenance [
1–
7]. These problems typically result in low tunnel excavation efficiency and increased costs, which have caused widespread concern in the engineering industry. Therefore, new rock-breaking methods for industrial applications are highly demanded.
Currently, methods to assist mechanical rock breaking primarily include using high-pressure waterjets [
8], high-energy laser irradiation [
9], high-pressure liquid nitrogen jets [
10], and microwave irradiation [
11,
12]. The high-pressure waterjet method can generate pre-cracks on rocks, which improves the rock-breaking ability of disc cutters. The high-energy laser irradiation method can cause the local area of rocks to heat up rapidly, thus improving the rock breakability by weakening the mechanical properties of the rock via thermal stress. The high-pressure liquid nitrogen jet method exploits the advantage of the large temperature gradient generated rapidly to damage rocks via the intense thermal shock effect. The microwave irradiation method uses the different thermal responses of various mineral components inside a rock to weaken the ability of hard the rock to resist external loads. However, the abovementioned rock-breaking methods, in particular the high-energy laser irradiation, high-pressure liquid nitrogen jet, supercritical CO
2 jet, and microwave irradiation methods, are associated with small rock-breaking volumes and high-energy consumption, are subject to the operating conditions and primarily applied in projects such as oil and gas drilling. For hard-rock-breaking projects involving a large excavation face, the high-pressure waterjet method provides a good application foundation and provides the best application value [
13]. Existing studies show that, among various new rock-breaking methods, the high-pressure waterjet method offers high rock-breaking efficiency, low energy consumption, and easy implementation; moreover, it is based on extensive research and presents significant research and application potential.
To date, several types of waterjet-assisted rock-breaking techniques have been developed. For example, the erosion rock-breaking method [
14], where water is used to scour and destroy rock for a significant duration as well as to remove fine silt on the rock surface to form erosion pits. Additionally, the stress intersection rock-breaking method [
15–
17] has been developed. In this method, the rock mass is subjected to both the waterjet (arranged in an appropriate position) and cutter simultaneously to satisfy two types of failure stress fields; additionally, the waterjet is used to cut grooves or drill holes on the rock surface to form free surface pressure relief, which reduces the difficulty in rock breaking by the mechanical cutter. Additionally, researchers have proposed the wedge penetration rock-breaking method [
18,
19]; in this method, the action points of a waterjet and the mechanical cutter on a rock overlap to the maximum extent such that the waterjet, which exhibits a certain pressure, penetrates into cracks generated by the mechanical cutter, thus causing the rock to fracture under the action of water pressure. In waterjet application as an assisted rock-breaking method for hard-rock formations, two rock-breaking methods are primarily used: the stress intersection and water wedge penetration rock-breaking methods. From 1965 to 1968, the American scholar Summers was the first to investigate high-pressure waterjet-assisted (HPWA) drilling technology for his doctoral thesis [
20]. In terms of the assisted rock-breaking method for TBMs, the Japanese Railway Technical Research Institute first used high-pressure waterjet technology to facilitate rock breaking by disc cutters; subsequently, the results obtained were published in 1972, which garnered the attention of experts in related fields. Henceforth, waterjet-assisted rock-breaking technology was applied to a TBM with a diameter of 3.8 m, and on-site tunneling tests were performed in a hard-rock area. The results showed that the advance rate can be increased by 2.5 times [
21]. Knickmeyer and Baumann [
22] investigated the cutting of highly abrasive sandstone using disc cutters assisted by a high-pressure waterjet. The results showed that the normal force required to break the same volume of rock under waterjet-assisted rock breaking was approximately 50% lower than that required by only mechanical cutters. Fenn et al. [
17] investigated a typical waterjet-assisted cutting system to improve the performance of a TBM on extremely hard and abrasive rocks (norite with a uniaxial compressive strength (UCS) of 250 MPa). They discovered that using a sharp disc cutter significantly reduced the normal and rolling forces (by approximately 40%) and thus the specific energy. In the abovementioned studies pertaining to rock cutting using a TBM assisted by a high-pressure waterjet, only Japanese scholars conducted industrial tests that were implementable without requiring specific monitoring data, whereas other scholars performed only laboratory tests.
Focusing on hard-rock tunnel projects, hard-rock cutting tests were performed in this study using waterjets under different water pressures, nozzle diameters, and nozzle speeds, in addition to on-site industrial penetration tests using the first HPWA TBM developed in China. The results of this study may benefit the application of high-pressure waterjets in hard-rock TBMs.
2 Laboratory rock-cutting tests
2.1 Test design and equipment
A high-pressure waterjet rock-cutting test was conducted on a multifunctional mechanical rock-breaking test platform. The equipment used is shown in Fig.1. Rock samples were obtained from a construction site in Longyan, Fujian Province. The rock samples were granite, which exhibited the following mechanical properties: UCS (172 MPa), internal friction angle (70.3°), cohesion (22.5 MPa), and natural density (2.61 g/cm3). Results of X-ray diffraction petrophysical analysis showed that the granite samples were primarily composed of 26% quartz, 18.1% feldspar, 27.4% plagioclase, 13.2% biotite, 5.8% calcite, and 9.5% clay minerals.
2.2 Rock breaking with high-pressure waterjet
The rock samples were placed under high-pressure waterjets with different cutting parameters. The test results were quantified, and the index parameters were obtained. The relationship between the rock-cutting performance and waterjet parameters was analyzed, of which the result supported the selection of waterjet parameters for the HPWA TBM applied in the Wan’anxi Water Diversion Project in Longyan, Fujian Province.
To investigate the rock-cutting performance under different waterjet parameters, a series of cutting tests was conducted by changing the water pressure, nozzle diameter, and nozzle moving speed while fixing the target distance at 40 mm. Considering the actual situation on site and the maximum capacity of the test equipment, the water pressures were set to 140 and 280 MPa. The test scheme is presented in Tab.1. The groove depth and width were the main indices used to evaluate the cutting performance. During the test, the flow and power of a single nozzle were measured at different water pressures and nozzle diameters.
The relationship between the width of the high-pressure waterjet cutting crack and nozzle diameter is shown in Fig.2. As shown, the slit width increased linearly with the nozzle diameter under different water pressures. Comparing Fig.2(a) and Fig.2(b), under the same nozzle diameter and nozzle speed, the notch width under 280 MPa water pressure was smaller than that under 140 MPa. This is because a higher water pressure results in less scattering of the waterjet and a narrower slit. However, no significant correlation was indicated between x and y.
Meanwhile, the kerf depth decreased logarithmically as the nozzle speed increased, and the kerf depth increased with the nozzle diameter (Fig.3). This is because when a nozzle with a larger diameter is used, more water is jetted per unit time and the work and momentum exerting on the rock are greater, thus resulting in a deeper kerf. Under the same conditions, the kerf depth under 280 MPa water pressure was two to three times that under 140 MPa.
In terms of cutting performance, when the water pressure was 140 MPa, the kerf was wide but shallow, and small rock fragments on the rock surface collapsed near the kerf. When the water pressure was 280 MPa, the kerf was narrow but deep, and almost no collapse occurred on the rock surface near the kerf, as shown in Fig.4.
Meanwhile, the most important index for facilitating the HPWA TBM in rock breaking was the kerf depth. When the water pressure was 280 MPa and the nozzle diameter was 0.74–0.97 mm, the maximum kerf depth achieved was 5–14 mm. This indicates that the waterjet can be used to assist the TBM in rock breaking and can be integrated into the TBM cutterhead for industrial testing. However, the moving speed of the waterjet remained lower than the normal linear speed of the cutterhead; hence, the rotational speed of the cutterhead must be reduced when a waterjet is applied.
2.3 Rock breaking by disc cutter under different kerf conditions
This section presents the rock-breaking experiment using the disc cutter under different slit conditions. First, rock samples with different slits were prepared using a high-pressure waterjet in three modes: no slit cutting, identical cutting, and staggered cutting. Slotless cutting implies that the cutter cuts the complete rock sample without slotting, as shown in Fig.5(a). Meanwhile, identical cutting implies that the slit is located directly below the penetration position of the cutting teeth, as shown in Fig.5(b). Finally, staggered cutting implies that the cut is located next to the insertion position of the cutter, as shown in Fig.5(c).
The applied disc cutter was a commercial cutter with a constant cross-section cutter, a ring diameter of 423 mm, and a tip width of 17 mm. The penetration depth and cutting speed were set to 4 mm and 20 mm/s, respectively. To perform staggered cutting, the spacing between the kerf and cutting groove was set to 40 mm. During the test, the three directional forces (normal, rolling, and side forces) of the disc cutter were measured at a frequency of 100 Hz. After each test, the rock surface was photographed, and rock debris was collected.
The normal and rolling forces under the different kerf conditions are shown in Fig.6. As shown in Fig.6(a) and Fig.6(b) for the identical and staggered cutting, respectively, the normal force is significantly lower than that of the no-kerf cutting test. This proves that the kerf prepared using the high-pressure waterjet can effectively reduce the indentation load of the coupled mechanical cutting. For both the identical and staggered cutting, the average normal force decreased as the kerf depth increased; however, the decreasing trends were different for the two cases. When the kerf depth was 2 mm, the average normal force of the staggered cutting exceeded that of the identical cutting. This is attributable to the following reason: in identical cutting, the kerf is located directly below the disc cutter. Thus, the dense core beneath the cutter tip breaks more easily and the stress in the dense core can be released early. By contrast, in staggered cutting, the effect of the kerf in reducing the indentation load cannot be exploited as the kerf depth is extremely small (only 2 mm); hence, staggered cutting resembles no-kerf cutting. When the kerf depth exceeded 4 mm, the average normal force of staggered cutting was significantly lower than that of identical cutting. However, the fluctuation amplitude of the normal force in staggered cutting was much larger than that in identical cutting. This indicates that the rock-breaking mechanisms of identical cutting and staggered cutting might differ significantly.
The average rolling force can be calculated to predict the cutterhead torque during TBM tunneling. The lowest average rolling force was indicated by no-kerf cutting. For both identical cutting and staggered cutting, the average rolling force generally increased with the kerf depth. This is because a larger kerf depth allows the kerf to penetrate the rock more easily [
23,
24]. Thus, the center angle of the contact arc between the cutting ring and rock increases, which results in an increase in the rolling force.
The weights of the rock fragments under different experimental conditions are shown in Fig.7. Under no-kerf cutting and staggered cutting, the weight of the rock fragment was extremely low when the kerf depth was 2 mm. Under identical cutting, the weight of the rock fragment was the highest a kerf depth of 4 mm. Under staggered cutting, the weight of the rock fragment was the highest at a kerf depth of 6 mm. To further evaluate the rock-cutting efficiency of the different cutting modes, the cutting specific energy (SE, i.e., the energy required to cut through a unit volume of rock) of each test was calculated as follows:
where SE is expressed in MJ/m3; Fr is the rolling force (kN), l is the cutting length of the rock cut by the hob; V (cm3) is the volume of the rock fragment, which can be calculated using the data provided in Fig.7(a).
Because the weight of the rock fragment was extremely low under no-kerf cutting and staggered cutting when the kerf depth was 2 mm (Fig.7(b)), the SE for the two tests were not calculated. In fact, the SE of the staggered cutting tests was lower than that of the identical cutting tests, indicating that the rock-cutting efficiency afforded by staggered cutting is higher than by identical cutting.
The results above indicate that the rock-cutting performance yielded by staggered cutting is better than that by identical cutting and no-kerf cutting. The kerf depth should be at least 4 mm to fully exploit the effect of the kerf in improving rock-cutting performance. However, as only seven groups of cutting tests were performed, the findings obtained were insufficient to explain the rock-breaking mechanism under different cutting modes; thus, an optimal kerf structure prepared via the high-pressure waterjet could not be proposed. Consequently, numerical simulations were conducted to investigate the rock-breaking performance and mechanism under different cutting modes and kerf structures.
3 Numerical analyses of rock breaking by disc cutters under different kerf conditions prepared using high-pressure waterjet
3.1 Numerical modeling and calibration of the rock material
The particle discrete element method can be used to effectively investigate the damage and fracture mechanisms of materials from the perspective of micromechanics [
25]. As the rock sample in this study presents a granular texture under a high-power microscope, it can be regarded as a discrete particle. The PFC software uses the method of particle aggregation to characterize the rock material; in particular, by applying a linear parallel bond between adjacent round particles, the particle contact area exhibits a certain elasticity and resistance to tension, shear, and torsion, thus presenting macro-mechanical properties similar to those of actual rocks [
26,
27]. Previous studies [
28] show that the ratio of UCS to tensile strength of rocks directly aggregated by linear parallel bonds is much smaller than that of actual rocks. This is because the contact boundary between round particles neither exhibits a mosaic structure nor the self-torsion resistance of actual crystalline rock particles; consequently, the compressive strength of the rock is underestimated.
The modeling process of the granular cluster rock material using a grain-based discrete element method (GB-DEM) based on the optimized Tyson polygon algorithm is shown in Fig.8.
1) The rock area is meshed using Delaunay triangles.
2) The center of the circumscribed circle of each Delaunay triangle is connected to the triangle vertices.
3) The Delaunay triangle mesh is deleted, and the Tyson polygon meshes remain, as shown in Fig.8(a).
4) The rock model area is filled with round particles.
5) The particles are categorize into groups based on their Tyson polygon mesh. The particles in the same group are represented by the same color, as shown in Fig.8(b).
6) A linear model is applied to the contact between the particles and walls, as shown by the short red line in Fig.8(c), and a linear parallel bonding model is applied to the contact between all particles. The parameters of the intragranular (short blue line in Fig.8(c)) and intergranular (short green line in Fig.8(c)) contacts are calibrated and assigned, respectively.
To avoid errors caused by the random shape and size of the grains, a function embedded in the MATLAB software was optimized, and a more uniform Tyson polygon was generated by controlling the vertex distribution of the Delaunay triangles.
This method offers the following advantages: first, it can simulate the fracture of grains caused by compression and cracks caused by the dislocation or separation of grain boundaries under the action of shear, tension, and expansion, which resembles the actual situation; second, irregular grain boundaries cannot only promote the propagation of tensile cracks, but also improve the torsion resistance between grains [
29]. Therefore, the macro-mechanical properties of the aggregated rock will be similar to those of an actual rock.
Meanwhile, the microscopic parameters were calibrated as follows. For the granite investigated in this study, uniaxial compression and Brazilian splitting tests were performed based on the test method recommended by the International Society for Rock Mechanics; each test was performed five times to obtain the macro-physical parameters of the rock specimen and the average values of those parameters. A model for the rock was established by selecting the appropriate microscopic mechanical parameters; subsequently, the uniaxial compression and Brazilian splitting were simulated, and the result shows that the rock model and specimen showed similar macro-physical parameters. Based on previous studies, the loading speed for the uniaxial compression and Brazilian splitting simulation was set to 0.1 m/s, and the time step was set to 5 × 10
−8 s to ensure that a quasi-static equilibrium simulation. Li et al. [
28] discovered that particle size did not significantly affect the simulation results, and discovered that when the particle diameter was less than 0.73 mm, the macro-physical parameters of a rock remained unchanged, although the total number of particles increased rapidly as the particle diameter decreased, thus prolonging the calculation time. Therefore, after performing multiple sets of trial calculations and considering the effect of the particle diameter on the simulation accuracy and calculation time, the average particle diameter was set as 0.5 mm. The microscopic mechanical parameters determined are listed in Tab.2, and the obtained rock macro-physical parameters are listed in Tab.3. The simulation and test results for the uniaxial compression and Brazilian splitting are shown in Fig.9.
The results show that the macro-physical parameters of the rock model and specimen were highly similar and that the errors of the elastic modulus, Poisson’s ratio, compressive strength, and tensile strength were less than 4%. The macro-failure modes of the rock under uniaxial compression and Brazilian splitting obtained via simulation and test were similar, where oblique shear and central splitting were observed under uniaxial compression and Brazilian splitting, respectively. The stress–strain curves of the uniaxial compression obtained via simulation and test were similar; however, the simulated axial strain was slightly less than the test value because the original microcracks in the rock gradually approached the initial stage of loading, which rendered the rock compact and caused it to deform in a concave nonlinear manner in the early stage. However, the aggregated rock model failed to simulate the process of microcrack compaction. The results above show that the GB-DEM method for modeling rock materials is reliable. In the uniaxial compression simulation, tensile cracks constituted 84.0% (of all cracks), of which 60.7% were intergranular tensile cracks. Meanwhile, in the Brazilian splitting simulation, tensile cracks constituted 85.4%, of which 56.3% were intergranular tensile cracks. This implies that the main failure mode of the rock was intergranular tensile failure.
3.2 Setup of numerical model for rock indentation under different kerf conditions
The rock indentation models under different kerf conditions, as shown in Fig.10, includes two modes, i.e., staggered cutting (Fig.10(a)) and identical cutting (Fig.10(b)), where the initial position of the cutter is 0.3 mm above the upper surface of the rock, the maximum vertical displacement of the cutter is 2.3 mm, and the width of the rock model is 500 mm.
To select the appropriate model height based on the simulation results and speed, for staggered cutting, a rock indentation simulation was performed based on model heights of 400 and 200 mm, as shown in Fig.11 and Fig.12. Results show that the rock-cutting performance, crack distribution, and rock-cutting forces of the two models were similar and that their simulation times were 3.1 and 1.4 h, respectively. Thus, the height of the model was set to 200 mm. The rock model comprised 31000 grains and 113000 particles. A cutter with a constant cross-section, a diameter of 432 mm, and a tip width of 20 mm was selected. The cutter indentation speed was 0.1 m/s and the time step was 5 × s. Therefore, the cutter indentation speed was converted to 5 × m/step to ensure a quasi-static simulation.
3.3 Effect of kerf depth on rock breaking
3.3.1 Effect of presence or absence of slits on rock breakage
To compare the rock-breaking performance with and without pre-cut kerfs, two rock-breaking models were established. For the pre-cut kerf model, the cutter spacing was set as 80 mm; the cutter spacing was set as 80 mm; the cutter was placed in the middle of adjacent kerfs; and the kerf spacing, width, and depth were set as 80, 2, and 30 mm, respectively. The rock-breaking performances with and without the per-cut kerfs are shown in Fig.13. Based on previous studies, the rock-breaking mechanism of the non-kerf model, i.e., the conventional rock breaking, involves lateral cracks between adjacent cutters being connected to each other. However, the UCS of the hard rock investigated in this study was approximately 180 MPa. Thus, the lateral cracks between the cutters could not connect even when the main crack was fully developed, and the dense core was broken. In the pre-cut kerf model, the lateral cracks propagated actively toward the pre-cut kerfs. Although the interaction between the adjacent cutters weakened, the lateral constraint of the rock weakened as well; thus, the rock can be fully broken into triangular chips. The peak value of the normal force of cutter 1 in the non-kerf model was 1.8 times higher than that of the model with pre-cut kerfs (Fig.13), and the penetration specific energy (PSE) of the former and latter were 43.0 and 2.2 N/mm2, respectively. The PSE of the former is 19.5 times higher than of the latter, which indicates that the pre-cut kerf can effectively reduce the cutter force and significantly improve the efficiency of hard-rock breaking.
3.3.2 Effect of cutting mode on rock breaking
For the staggered cutting model, the cutter was placed in the center of adjacent kerfs, the cutter and kerf spacings were both 80 mm, and the kerf width and depth were 2 and 30 mm, respectively. The cutter and kerf spacings as well as the kerf width and depth of the identical cutting model were the same as those of the staggered cutting model. The rock-breaking performances of the abovementioned models with different cutting modes are shown in Fig.14. Results show that the rock was completely broken and that the rock-breaking efficiency of the staggered cutting model was higher. For the identical cutting mode, the cutter normal force was lower than that of the staggered cutting mode, and the normal force decreased as the kerf width increased. This is because a kerf was present immediately underneath the cutter tip; hence, a dense core could not be established.
The analyses above show that staggered cutting is an ideal mode for kerf-assisted rock breaking, which is characterized by the weakening of the surrounding rock by the kerf to remove the lateral restraint and crack growth driven by the dense core. In this mode, the cutter and kerf spacings were the same. In the identical cutting mode, the rock-breaking effect was unsatisfactory because a dense core established underneath the cutter. In the mixed cutting mode, the rock-breaking effect of each cutter differed significantly but not ideal. Therefore, the subsequent analyses are based on the staggered cutting mode.
3.3.3 Effect of kerf depth on rock breaking
Based on the staggered and middle-arranged cutting modes, where the cutters were placed in the middle of adjacent pre-cut kerfs, 10 models were established based on the following parameters: 1) the kerf width was set as 2 mm; 2) the cutter/kerf spacing was set as 80 mm; and 3) the kerf depth was set as 2, 4, 6, 8, 10, 14, 18, 22, 26, and 30 mm. The effect of the kerf depth on rock breaking is shown in Fig.15. When the kerf depth was less than 14 mm, the chips appeared primarily flat; when the kerf depth exceeded 18 mm, the chips appeared primarily triangular. As the kerf depth increased, the length of the median crack decreased, whereas the lateral cracks propagated sufficiently. This is because the lateral constraint of the rock decreased as the kerf depth increased, thus causing the lateral cracks to propagate more easily to the end of the kerfs. Thus, once the propagation of the lateral cracks had completed, the dense core was broken and could no longer support and transfer the normal force.
The PSEs of the 10 models are shown in Fig.15. The PSE vs. kerf depth was fitted using a power function, and the result shows that the PSE decreased as the kerf depth increased owing to the decreasing and increasing normal force and chip amount, respectively. This shows that the PSE decreased significantly when the kerf depth increased from 14 to 18 mm and then remained generally stable as the kerf depth continued to increase. This is primarily because the chip shape changed from flat to triangular, and the chip amount increased significantly when the kerf depth increased from 14 to 18 mm. This implies that the critical threshold for the kerf depth is 18 mm, considering its effect on the PSE.
The analyses above show that the rock-breaking force decreased while the rock-breaking efficiency increased with as the kerf depth increased. Two critical thresholds were indicated for the kerf depths, i.e., approximately 8 and 18 mm, considering the effect of the kerf depth on the rock-breaking force and rock-breaking efficiency, respectively. However, based on the waterjet experiment described in Subsection 1.2, generating kerf depths that exceed 4 mm at high moving speeds is difficult, and the kerf depth is generally between 2 and 4 mm. The analysis presented in this section shows that at a crack depth of 2–4 mm, the PSE can still be reduced by a factor of more than two, thus improving the rock-breaking efficiency.
4 Field penetration tests on HPWA TBM
Severe problems may arise when during hard-rock excavation using TBMs, such as low rock-breaking efficiency, low advancing speed, and high cutter consumption. Hence, China’s first HPWA TBM, which was developed by the China Railway Engineering Equipment Group Co., Ltd., was applied in the Wan’anxi Water Diversion Project in Longyan, China, as shown in Fig.16.
4.1 Geological conditions on site
The Wan’anxi Water Diversion Project aims to satisfy the medium- and long-term water supply requirements of Longyan, the main urban area in Fujian Province. The surrounding rock of the tunnel is primarily composed of Yanshanian biotite granite with a UCS of 80.6–197 MPa, which is categorized as hard rock. The surrounding rock is fresh rock, and the rock mass is relatively complete to complete. The thickness of the overburden rock along the alignment is generally 100–800 m, and the thickness of the overburden rock of the gully section is relatively small; meanwhile, weak to slightly weathered rock mass exhibits weak water permeability. Except for tunnel portals, the surrounding rock of the tunnel is generally relatively stable to stable and can be classified into the following grades (total length of 27.68 km): grade II, 77.3%; grade III, 13.2%; grade IV, 9.4%; and grade V, 0.1%.
The corresponding stratum in the tests were Devonian quartz conglomerate with quartz sandstone and grade II surrounding rock, with a UCS of 162–197 MPa and quartz content of 60%.
4.2 Field test scheme
In the rock-breaking experiment assisted by high-pressure waterjet coupling, the cutterhead was fist delivered to the tunnel face, the high-pressure waterjet (150 MPa) was started, and 400 mm of driving distance was excavated. Subsequently, 280 MPa of high-pressure waterjet was started and 700 mm of driving distance was excavated. The cutterhead was operated under no-load conditions and the high-pressure waterjet (280 MPa) was operated for 3 min. Subsequently, the high-pressure waterjet and main engine were stopped. Finally, a construction personnel entered the manhole to observe the cutting performance of the heading face. During the entire test, the cutterhead thrust was maintained at 6000–6500 kN, and the cutterhead speed was 6 r/min. The test scheme of the high-pressure waterjet coupling rock-breaking test is presented in Tab.4.
Within the test area marked with chainage, the stratum lithology was quartz conglomerate and the rock mass was complete. The UCS of the rock was estimated to be approximately 170 MPa, based on measurement obtained using a high-strength rebound instrument, and the surrounding rock was classified as grade II, as shown in Fig.17. In the tests, the high-pressure waterjet was 280 MPa, and no abnormal phenomenon occurred. During the tests, the data pertaining to the TBM tunneling parameters, high-pressure waterjet pressure parameters, ambient temperature at the #1 and #2 belt conveyor interface area, and ambient temperature at the high-pressure pump unit area and rock were obtained. A cutterhead hose protection clip and a rubber hose retainer were adopted to protect the high-pressure hose suspended on the back of the cutterhead or U-steel, which was difficult to set, as shown in Fig.18.
4.3 Field test data analysis
1) Tunnel boring machine driving data analysis
Based on the test data obtained, the thrust force, penetration, torque, and other parameters during the TBM tunneling by disc cutters only or coupled with a waterjet were compared and analyzed. Fig.19–Fig.21 show the data distribution of the cutterhead thrust force, penetration, and torque during pure disc cutter tunneling or when coupled with a high-pressure waterjet at different water pressures. Fig.22 shows the change curves of the FPI (Field Penetration Index) and TPI (Torque Penetration Index) under different water pressures.
The FPI represents the thrust required for each disc cutter to penetrate a rock by 1 mm. A lower FPI indicates a higher rock-breaking efficiency. Meanwhile, the TPI represents the torque required for each disc cutter to penetrate a rock by 1 mm. A lower TPI indicates a higher rock-breaking efficiency. Based on Fig.19–Fig.22 and the table presenting a comparison of the TBM tunneling parameters, the penetration of the TBM increased significantly, the cutterhead torque increased slightly, and the FPI and TPI decreased when the rock was broken by the disc cutter coupled with high-pressure waterjet. The average FPI was 84.6 kN/mm for pure disc cutter rock breaking, whereas it was 52.4 kN/mm for 280 MPa high-pressure waterjet-coupled tunneling, which indicates that the thrust of each disc cutter penetrating a rock by 1 mm decreases by 38.1%. When tunneling coupled with a water pressure of 280 MPa was performed, the TPI decreased from 6.33 kN·m/mm during pure disc cutter tunneling to 5.26 kN·m/mm, which is a decrease of approximately 16.9%, thus further confirming the desirable effect of high-pressure waterjet-coupled rock breaking.
After the high-pressure waterjet-coupled rock-breaking tests were completed, the cutterhead was rotated for 3 min only with the waterjet but without a load, and the tunnel face cutting effect is shown in Fig.23. Under a pressure of 280 MPa, a 5–8 mm deep annular kerf remained on the rock face after cutting was performed with the waterjet.
2) Analysis of tunnel environment temperature
In the high-pressure waterjet tests, the ambient temperature in the tunnel, the rock chip temperature, and the high-pressure water temperature at the nozzles were obtained, and the variation of the temperature field in the tunnel under different water pressures was obtained, as shown in Tab.5. The result shows that as the water pressure increased, the three abovementioned temperatures increased significantly.
The ambient temperature at the interface of the #1 and #2 belt conveyors (approximately 25 and 10 m away from the cutterhead and high-pressure pump unit, respectively) was 28.6 °C when the high-pressure waterjet was not turned on, and the maximum temperature increased to 33.3 °C after approximately 20 min of operation with a 280 MPa high-pressure waterjet. The highest temperature was recorded in the high-pressure pump unit area of the tunnel. When the high-pressure waterjet system was operated at 280 MPa, the highest temperature recorded in the pump unit area was 46 °C. In addition, the humidity in the tunnel increased significantly when the high-pressure waterjet was operated. The humidity at the #1 and #2 interface areas exceeded 80%, and the humidity in the pump unit area exceeded 95%.
As the pressure of the high-pressure waterjet system increased, the outlet water temperature of the cutter-head nozzle increased significantly. The outlet water temperature of the nozzle was approximately 45 °C during normal tunneling (where the high-pressure system is operated at 30–40 MPa); however, when the pressure increased to 280 MPa, the highest outlet water temperature recorded at the pump unit was 90 °C. The increase in the outlet water temperature in the high-pressure system increased the rock chip temperature. According to the test results, the highest temperature of the rock chip at the interface of the #1 and #2 belt conveyors was 57.3 °C.
During the tests, the continuous operation time of the high-pressure waterjet system was controlled for approximately 1 h. If the system operates under high pressure for a significant amount of time, then the construction ambient temperature in the tunnel and rock chip temperature will increase further. In addition, the high-pressure water temperature at the nozzle of the cutterhead will become extremely high, thereby diminishing the cooling effect of the cutterhead. Generally, 1 h after the excavation is halted, the overall temperature around the cutterhead will decrease gradually, thus allowing the construction personnel to enter the cutterhead to conduct inspections; however, this deteriorates the TBM construction efficiency. Therefore, temperature control must be prioritized during the operation of high-pressure waterjets.
3) Improvement afforded by new TBM
The industrial test results of the Longyan Water Diversion Tunnel in China show that HPWA rock breaking can significantly improve the efficiency of TBM tunneling as well as improve the adaptability of the TBM in rock strata with a UCS of 162–197 MPa. Based on comparing the tunneling parameters under different water pressures and observing the cutting depth yielded only by the high-pressure waterjet when the cutterhead operates without a load, the cutting depth of the high-pressure waterjet can match the penetration depth of the disc cutter tunneling.
Additionally, industrial test results show that the TBM coupled with a high-pressure waterjet require further improvements.
First, the stability of the high-pressure waterjet system must be improved. Because the maximum water pressure of the high-pressure waterjet system tested in this study is 280 MPa, the vibration during TBM tunneling is significant, which causes the key components to loosen and possibly pipe burst. Thus, the waterjet pipeline must be designed appropriately to ensure its safety.
Second, the high-pressure waterjet system causes the temperature field in the tunnel to increase. In particular, the maximum temperature around the cutterhead is 85–90 °C, which affects the cutter change process. The ventilation pipeline of a TBM must be designed to allow air to circulate effectively in the vicinity of the cutterhead such that the temperature around the cutterhead can be reduced.
5 Conclusions
In this study, rock-breaking tests were performed using a TBM cutterhead assisted by a high-pressure waterjet in the Wan’anxi water diversion tunnel in Longyan, China. The following conclusions were obtained.
1) Under a water pressure of 280 MPa and a nozzle diameter of 0.74–0.97 mm, the maximum cutting depth achieved was 5–14 mm, which indicates that waterjets can be used to assist the TBM cutterhead in rock breaking and can be integrated with the TBM for industrial tests.
2) When the depth of the cutting crack exceeded 4 mm, the rock-breaking volume in the double-waterjet cutting crack tests was greater than that in the single waterjet cutting crack tests, and the effect of the double-waterjet cutting cracks on rock breaking was much greater than that of the single waterjet cutting crack.
3) The critical mechanisms of kerf-assisted rock breaking by the TBM cutter were as follows: a) the pre-cut kerf weakens the surrounding rock, which renders it easier for the lateral cracks to propagate to the kerf end; b) the dense core is maintained, which facilitates the transfer of the rock-breaking force and the propagation of the lateral cracks. These two mechanisms should be fully utilized as they enable excellent rock-breaking performance to be achieved.
4) The test section was composed of Devonian quartz conglomerate interbedded with quartz sandstone, with a UCS of 162–197 MPa and a quartz content of 60%. The FPI and TPI decreased by 38.1% and 16.9%, respectively, which further proves the desirable effect of high-pressure waterjet-coupled rock breaking.
5) During the operation of the high-pressure waterjet, the ambient temperature in the tunnel and the water spray temperature of the cutterhead both increased significantly, thereby diminishing the cooling effect of the cutterhead and deteriorating the construction efficiency of the TBM. Therefore, temperature control during the long-term operation of a high-pressure waterjet must be further investigated, and the corresponding cooling measures should be implemented.
PDF
(9105KB)
