Rock Mechanics and Rock Engineering (2019) 52:4763–4783

https://doi.org/10.1007/s00603-019-01865-x

ORIGINAL PAPER

Full‑Scale Linear Cutting Tests to Propose Some Empirical Formulas

for TBM Disc Cutter Performance Prediction
Yucong Pan1 · Quansheng Liu1 · Xingxin Peng2 · Qi Liu3 · Jianping Liu4 · Xing Huang5 · Xianze Cui6 · Tao Cai7

Received: 15 October 2018 / Accepted: 21 May 2019 / Published online: 13 June 2019
© Springer-Verlag GmbH Austria, part of Springer Nature 2019

Abstract
Determination of the TBM disc cutter performance at the optimum rock cutting condition is very important, but the existing
theoretical, laboratory, numerical or empirical methods still have some shortcomings. Full-scale rock cutting test is regarded
as the most accurate and reliable method in the laboratory, but it still requires large rock blocks and specific testing equip-
ment, which make the use of this method very limited and costly. Thus, by collecting, analyzing and formulizing the results
of many full-scale linear cutting tests, this study partly overcomes the shortcoming of the full-scale rock cutting test and
proposes some new empirical formulas to predict TBM disc cutter performance in a much easier and less costly way. First,
this study reveals the general rules to characterize the change of rock cutting results with the increase of rock uniaxial com-
pressive strength and cutter penetration depth by conducting full-scale linear cutting tests on five different rock types. Then,
by analyzing a large amount of full-scale linear cutting test results, this study correlates the TBM disc cutter performance at
the optimum rock cutting condition with rock uniaxial compressive strength and a widely used semi-theoretical prediction
model. The new proposed empirical formulas in this study only use rock properties, cutter geometries and cutting geometries
to design the machine specifications and select the TBM operation parameters, which is a much easier and less costly method.

Keywords Tunnel boring machine (TBM) · Linear cutting machine (LCM) · Disc cutter · Rock cutting · Performance
prediction · CSM Model
List of Symbols
* Quansheng Liu β Acting angle of the disc cutter resultant force
[email protected] referred to the normal direction (rad)
1
θ Angle of the studied point on the rock-cutter
Key Laboratory of Safety for Geotechnical and Structural
Engineering of Hubei Province, School of Civil
contact arc referred to the normal direction
Engineering, Wuhan University, Wuhan 430072, Hubei, (rad)
People’s Republic of China φ Contact angle between the rock surface and
2
China Railway 11th Bureau Group, 4th Engineering Co., Ltd, disc cutter (rad)
Wuhan 430074, Hubei, People’s Republic of China ν Poisson’s ratio of the rock
3
Changjiang Institute of Survey, Planning, ρ Natural density of the rock (g/cm3)
Design and Research, Wuhan 430010, Hubei, σc Uniaxial compressive strength of the rock
People’s Republic of China (MPa)
4
College of Water Conservancy and Hydropower Engineering, σt Brazilian tensile strength of the rock (MPa)
Sichuan Agricultural University, Ya’an 625014, Sichuan, ψ Contact pressure distribution constant in the
People’s Republic of China
semi-theoretical CSM model, typically − 0.2
5
State Key Laboratory of Geomechanics and Geotechnical to 0.2
Engineering, Institute of Rock and Soil Mechanics,
BI Rock boreability index obtained in the full-
Chinese Academy of Sciences, Wuhan 430071, Hubei,
People’s Republic of China scale linear cutting tests (kN/mm)
6 BINor Normalized rock boreability index obtained
College of Hydraulic and Environmental Engineering,
China Three Gorges University, Yichang 443002, Hubei, in the full-scale linear cutting tests
People’s Republic of China C Constant in the semi-theoretical CSM model,
7
China Construction Third Engineering Bureau Co., Ltd, usually taken as 2.12
Wuhan 430064, Hubei, People’s Republic of China

13
Vol.:(0123456789)

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

4764 Y. Pan et al.

CC Disc cutter cutting coefficient obtained in the s/p Cutter spacing to penetration depth ratio
full-scale linear cutting tests (%) obtained in the full-scale linear cutting tests
CCNor Normalized disc cutter cutting coefficient (s/p)opt Optimum cutter spacing to penetration depth
obtained in the full-scale linear cutting tests ratio obtained in the full-scale linear cutting
D Disc cutter diameter (mm) tests
E Static elasticity modulus of the rock (GPa) SE Specific energy obtained in the full-scale
FN Mean disc cutter normal force obtained in linear cutting tests (MJ/m3)
the full-scale linear cutting tests (kN) SEopt Optimum specific energy obtained in the
FNp Peak disc cutter normal force obtained in the full-scale linear cutting tests (MJ/m3)
full-scale linear cutting tests (kN) (SEopt)CSM Optimum specific energy predicted by the
FNCSM Disc cutter normal force predicted by the semi-theoretical CSM model (MJ/m3)
semi-theoretical CSM model (kN) SRMBI Specific rock mass boreability index (kN/
FNNor Normalized disc cutter normal force cutter/mm/rev)
obtained in the full-scale linear cutting tests T Disc cutter tip width (mm)
FNopt Normal force at the optimum condition V Rock cutting volume per cutting distance
obtained in the full-scale linear cutting tests ­(mm3/mm)
(kN)
Abbreviations
(FNopt)CSM Normal force at the optimum condition pre-
CCS Constant cross section
dicted by the semi-theoretical CSM model
CSM Colorado School of Mines
(kN)
CSU-LCM The linear cutting machine in Central South
FPI Field penetration index (kN/cutter/mm/rev)
University, China
FR Disc cutter rolling force obtained in the full-
ISRM International Society for Rock Mechanics
scale linear cutting tests (kN)
and Rock Engineering
FRp Peak disc cutter rolling force obtained in the
LCM Linear cutting machine
full-scale linear cutting tests (kN)
TBM Tunnel boring machine
FRCSM Disc cutter rolling force predicted by the
semi-theoretical CSM model (kN)
FRNor Normalized disc cutter rolling force obtained
1 Introduction
in the full-scale linear cutting tests
FRopt Rolling force at the optimum condition
When using TBMs (tunnel boring machines) to excavate
obtained in the full-scale linear cutting tests
underground tunnels, it is very important to identify the
(kN)
optimum rock cutting condition and determine the corre-
(FRopt)CSM Rolling force at the optimum condition pre-
sponding rock cutting results. This is done by analyzing the
dicted by the semi-theoretical CSM model
specific energy curve during the rock cutting process. Here,
(kN)
specific energy is defined as the energy input required to
FTCSM Disc cutter resultant force predicted by the
cut a unit volume of rock material (Teale 1965). As seen in
semi-theoretical CSM model (kN)
Fig. 1, when cutter penetration depth is too small (for a fixed
KN Normal force modification factor considering
cutter spacing) or cutter spacing to penetration depth ratio
rock uniaxial compressive strength
is too large, very few rock mucks will be cut off and specific
KR Rolling force modification factor considering
energy will be very high (i.e., insufficient rock breakage
rock uniaxial compressive strength
occurs). When cutter penetration depth is too large (for a
KS Specific energy modification factor consider-
fixed cutter spacing) or cutter spacing to penetration depth
ing rock uniaxial compressive strength
ratio is too small, rock material under TBM disc cutter will
NRF Normalized resultant force of the disc cutter
be excessively crushed/grinded and specific energy will be
obtained in the full-scale linear cutting tests
also very high (i.e., excessive rock crushing occurs). When
p Cutter penetration depth (mm)
cutter penetration depth matches well with cutter spacing,
P0 Base contact pressure immediately under-
the interaction between two adjacent disc cutters will be
neath the disc cutter (MPa)
maximized to produce large, thick and elongated rock chips
Pθ Contact pressure distribution function within
and specific energy will be very low (in this case, it will be
the rock-cutter contact area (MPa)
called the optimized specific energy). Thus, this minimum
R Disc cutter radius (mm)
point of the SE–p or SE–s/p curve is marked as the optimum
s Disc cutter spacing (mm)

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Full-Scale Linear Cutting Tests to Propose Some Empirical Formulas for TBM Disc Cutter… 4765

Fig. 1  Specific energy with cutter penetration depth and the ratio of cutter spacing to penetration depth

rock cutting condition (Snowdon et al. 1982; Roxborough used in the field TBM projects, so the disc cutter cut-
and Phillips 1975). ting forces and rock fragmentation behaviours obtained
Rock cutting results under the optimum rock cutting in full-scale linear cutting tests can be well correlated
condition mainly include the optimum cutter spacing, with the field TBM excavation results (e.g., Balci 2009;
optimum cutter penetration depth, optimum specific Balci and Tumac 2012; Bilgin et al. 2008; Cardu et al.
energy and the corresponding disc cutter normal and roll- 2017; Copur et al. 2014; Pan et al. 2018e). However, full-
ing forces. During the project planning and TBM design scale linear cutting tests also have some shortcomings.
stage, these parameters will be used as the input param- For example, the large rock blocks are usually difficult
eters to select the proper cutter shape/size, design the or even impossible to be obtained and the required test-
TBM cutterhead layout, calculate the installed machine ing equipment (i.e., full-scale linear cutting machine) are
specifications, and predict the field TBM excavation per- not available to many researchers. Thus, some alternative
formance. Detailed examples about this procedure can be methods have been proposed.
found in Balci (2009), Balci and Tumac (2012), Bilgin One alternative method is to correlate the full-scale
et al. (2008), Cardu et al. (2017), Copur et al. (2014) linear cutting test results with some rock properties,
and Pan et al. (2018e). Thus, many methods have been e.g., rock compressive and tensile strengths, static and
proposed to predict the TBM disc cutter performance, dynamic elasticity modulus, P- and S-wave velocities.
especially at the optimum rock cutting condition, includ- For examples, Bilgin et al. (2006) investigated the domi-
ing the theoretical, numerical, laboratory or empirical nant rock properties affecting the performance of conical
methods (e.g., Bruland 1998; Chen and Labuz 2006; Cho picks and correlated the cutter force and specific energy
et al. 2010; Delisio et al. 2013; Farrokh et al. 2012; Geng with rock properties and the results obtained from some
et al. 2016a, b; Gong et al. 2005, 2006; Labra et al. 2016; widely used theoretical approaches. Wang et al. (2018)
Liu et al. 2002, 2016; Macias 2016; Ozdemir and Wang also investigated the dominant cutting parameters affect-
1979; Pan et al. 2018b; Roxborough and Phillips 1975; ing the specific energy of the selected sandstones when
Sanio 1985). These methods all have their unique advan- using conical picks and developed some empirical pre-
tages and disadvantages, and among them, full-scale diction models based on rock strength parameters. Rox-
linear cutting test is regarded as the most accurate and borough and Phillips (1975) and Moon and Oh (2012)
reliable laboratory testing method (Balci 2009; Balci and derived the theoretical formulas to predict the optimum
Tumac 2012; Cho et al. 2013; Copur et al. 2014; Gertsch cutter spacing to penetration depth ratio using some rock
1993, 2000; Gertsch et al. 2007; Gong et al. 2015, 2016; strength and brittleness properties. Many more investiga-
Pan et al. 2018a, c, d, e, f; Rostami 1991, 2013; Roxbor- tions have been conducted to correlate the specific energy
ough and Phillips 1975; Snowdon et al. 1982; Tumac and of rock cutting to some basic rock properties (Atici and
Balci 2015). This kind of rock cutting tests uses large rock Ersoy 2009; Balci et al. 2004; Dogruoz et al. 2018; Ersoy
blocks and full-size cutting tools, so the scaling effect, and Atici 2007; Tiryaki 2009; Tiryaki and Dikmen 2006;
boundary effect and unnatural rock cutting behaviours Tumac et al. 2007; Yurdakul et al. 2014). However, the
can be avoided. This kind of rock cutting tests also uses above studies are usually about conical picks, sharp cut-
cutter spacings and penetration depths same as the ones ting picks, drilling bits or diamond circular saws, not

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

4766 Y. Pan et al.

specifically concerning the TBM disc cutters. Thus, the semi-theoretical CSM model and then modifies it to pre-
empirical prediction formulas proposed in these studies dict the TBM disc cutter performance at the optimum rock
cannot be applied to TBM disc cutters. Following the cutting condition more reliably. Generally, the new empiri-
similar methodology, this study conducts full-scale lin- cal formulas proposed in this study will help us design the
ear cutting tests on five different rock types using a CCS machine specifications and select the TBM operation param-
(constant cross section) disc cutter. The general rules to eters in a much easier and less costly way.
characterize the change of rock cutting results with the
increase of rock compressive strength and cutter penetra-
tion depth are deeply studied. Combined with the previ- 2 Analysis of the Full‑Scale Linear Cutting
ously published literatures, some empirical formulas are Test Results Obtained Using CSU‑LCM
proposed to use the rock compressive strength to predict
the full-scale linear cutting test results using TBM disc 2.1 Full‑Scale Linear Cutting Tests Conducted Using
cutters. CSU‑LCM
The other alternative method is to formulize the full-
scale linear cutting test results. For examples, Rostami and The five rock types used for the full-scale linear cutting
Ozdemir (1993) proposed a widely used semi-theoretical tests in this study are Cement mortar, Longchang sandstone,
CCS disc cutter cutting force prediction model (also known Hezhou marble, Miluo granite and Yueyang granite. Table 1
as the semi-theoretical CSM model) based on the full-scale summarizes the physical and mechanical properties of these
linear cutting tests conducted at Colorado School of Mines. five rock types. The uniaxial compressive tests are conducted
Roxborough and Phillips (1975) offered and examined a using the rock samples with a diameter of 50 mm and a
force prediction model for V-type disc cutters based on the length of 100 mm. The Brazilian tensile tests are conducted
full-scale linear cutting tests conducted using different disc using the rock samples with a diameter of 50 mm and a
cutter diameters, penetration depths, edge angles, cutting length of 50 mm. These tests are conducted based on the
speeds and spacings. Sanio (1985) derived the equations for ISRM-suggested testing procedure (Ulusay 2015). As seen
the prediction of TBM disc cutter performance in bedded in Table 1, the uniaxial compressive strength of these five
and schistose rocks by conducting full-scale linear cutting rock types varies from 21.25 to 176.88 MPa, and the Bra-
tests on rocks with distinctive strength anisotropy. These zilian tensile strength of these five rock types varies from
models make the combined use of the full-scale linear cut- 1.59 to 7.66 MPa; therefore, they can generally represent the
ting test results and some useful rock cutting mechanisms, commonly encountered rock types during TBM tunneling.
so they can usually offer prediction results close to the The full-scale linear cutting tests in this study are con-
experimental results (e.g., Geng et al. 2016a; Tumac and ducted at Central South University, China, as shown in
Balci 2015). However, in some cases, these models are not Fig. 2a. During these rock cutting tests, a CCS disc cutter
so accurate (e.g., Cardu et al. 2017; Entacher et al. 2014; with a diameter of 432 mm and a cutter tip width of 19 mm
Pan et al. 2018b). Pan et al. (2018c, f) and Xia et al. (2018) is fixed tightly to the cutter saddle using high-strength bolts.
have shown that the prediction results of the semi-theoreti- The cutter saddle is also fixed tightly to the head part of the
cal CSM model are always deviated from the experimental machine frame using a transition device. The disc cutter,
results obtained in the full-scale linear cutting tests and the cutter saddle and machine head part can be moved together
deviation degree will surely change when rock compressive vertically to make the disc cutter penetrate into the rock
strength changes. To predict the TBM disc cutter perfor- sample surface at the pre-set cutter penetration depth. The
mance more accurately, some modification factors should rock sample with the size of 980 mm × 380 mm × 280 mm
be proposed. Thus, combined with the previously published in length, width and height, respectively, is cast into the rock
literatures, this study checks the prediction accuracy of the sample box immediately beneath the disc cutter. The gaps

Table 1  Physical and Parameter Cement mortar Longchang Hezhou marble Miluo granite Yueyang granite
mechanical properties of the sandstone
five rock types
σc (MPa) 21.25 37.02 79.07 127.24 176.88
σt (MPa) 1.59 1.75 3.26 5.14 7.66
E (GPa) 8.80 9.98 51.65 56.74 62.19
ν (ratio) 0.28 0.33 0.24 0.26 0.29
ρ (g/cm3) 2.08 2.33 2.68 2.73 2.78

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Full-Scale Linear Cutting Tests to Propose Some Empirical Formulas for TBM Disc Cutter… 4767

Fig. 2  Full-scale linear cutting

tests conducted at Central South
University, China

between the rock sample and the rock sample box are filled the first cutting pass are also excluded because they repre-
by concrete grout (Grade: C60) to offer perfect contact and sent the rock cutting behaviours when cutter penetration
avoid sample vibration. The rock sample box can be moved depth is suddenly enlarged. The data obtained from last
laterally by the horizontal hydraulic cylinders to adjust the three cutting passes after excluding the two outer cutting
cutter trajectory to the pre-set cutter spacing. It can also be grooves are used for analysis; so for one cutter penetration
moved by the thrust hydraulic cylinders to make the disc depth, the analyzed data are obtained from nine cutting
cutter roll ahead on the rock sample surface to fulfil the grooves, which is sufficient to offer reliable and repre-
continuous rock cutting. During the rock cutting process, sentative results. After one cutting pass is finished, the
totally three groups of strain gauges are attached to the four rock mucks on the rock sample surface are collected and
columns of the cutter saddle. Their strain information will be weighted to calculate the specific energy, and then the
used as the input parameters to obtain the three-dimensional following cutting pass will be started. Low-strength rocks
cutting force acting on the disc cutter. Detailed information are easy to be cut, so the employed ranges of cutter pen-
about this full-scale linear cutting machine can be referenced etration depth are large, e.g., from 4 to 12 mm for Cement
in Xia et al. (2018). mortar and Longchang sandstone (Table 2). High-strength
The cutter spacings for these five rock types are all rocks are difficult to be cut, so the employed ranges of
70 mm because hard rock TBMs tend to operate at cut- cutter penetration depth are small, e.g., from 1 to 4 mm
ter spacings around 76 mm with the maximum penetra- for Yueyang granite (Table 2).
tion depth per revolution allowed by the machine power Table 2 summarizes the full-scale linear cutting test
(Gertsch et al. 2007). The rock sample width is 380 mm, results obtained from these five rock types. In this table,
so there are totally five cutting grooves created on the disc cutter normal force is the force component acting
rock sample surface. The distances from the outer cut- vertically to the rock surface, and disc cutter rolling force
ting grooves to the rock sample boundary are 50 mm, so is the force component acting along the cutter advancing
disc cutter cutting forces obtained from these two cut- direction, as shown in Fig. 2b. FN and F ­ Np are the mean
ting grooves are excluded because of the boundary effect. and peak values of the disc cutter normal force while
When disc cutter cuts the five cutting grooves sequentially FR and F ­ R p are the mean and peak values of the disc
once, a cutting pass is finished. Before the formal rock cutter rolling force. Disc cutter cutting coefficient (CC)
cutting tests, the top surface of the rock sample should be is calculated as the ratio of disc cutter normal force to
fully pre-conducted by conducting several cutting passes disc cutter rolling force, i.e., CC = FR/FN. Normalized
using the smallest pre-set cutter penetration depth. When resultant force of the disc cutter (NRF) is calculated as
disc cutter cutting forces obtained from different cutting the ratio of resultant force angle of the disc cutter (β) to
grooves are stable without any unnatural curve shapes, the rock-cutter contact angle (φ), i.e., NRF= β/φ. Here,
this procedure is finished. This procedure can make the resultant force angle of the disc cutter (β) is a func-
top surface of the rock sample have a damaged degree tion of the disc cutter normal and rolling forces, i.e.,
similar to that of the real tunnel face. During the formal β = arctan(FR/FN), and rock-cutter contact angle (φ) is a
rock cutting tests, for one cutter penetration depth, totally function of the disc cutter radius and penetration depth,
four cutting passes are conducted. The data obtained from i.e., φ = arccos((R − p)/R). Rock boreability index (BI) is

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

4768 Y. Pan et al.

Table 2  Full-scale linear cutting test results in the five rock types
Rock type p FN FR FNp FRp CC NRF BI SE FNCSM FRCSM SECSM
Unit mm kN kN kN kN % Ratio kN/mm MJ/m3 kN kN MJ/m3

Cement mortar 4 55.37 7.33 117.83 14.05 13.24 0.6828 13.84 53.47 28.44 2.75 9.82
6 67.68 10.07 136.28 18.50 14.88 0.6252 11.28 28.01 32.47 3.85 9.18
8 74.36 11.71 146.84 21.76 15.75 0.5721 9.30 20.92 35.66 4.90 8.75
10 79.95 13.90 157.67 26.12 17.39 0.5635 8.00 16.68 38.32 5.90 8.43
12 84.81 16.21 158.35 29.72 19.11 0.5639 7.07 17.64 40.63 6.87 8.17
Longchang sandstone 4 64.32 7.97 139.98 16.32 12.39 0.6396 16.08 65.09 42.51 4.11 14.68
6 78.31 10.50 146.08 22.36 13.41 0.5642 13.05 32.15 48.54 5.76 13.72
8 86.26 13.22 166.64 28.77 15.33 0.5570 10.78 24.05 53.30 7.32 13.07
10 92.23 14.89 169.07 30.09 16.14 0.5240 9.22 23.74 57.28 8.82 12.60
12 100.57 17.36 195.25 36.01 17.26 0.5104 8.38 24.22 60.73 10.27 12.22
Hezhou marble 2 72.12 6.06 159.30 13.52 8.40 0.6155 36.06 188.80 69.01 4.71 33.62
4 99.74 11.44 198.94 24.23 11.47 0.5925 24.94 83.15 86.74 8.39 29.95
6 102.84 13.25 207.84 26.83 12.88 0.5424 17.14 50.75 99.06 11.76 27.99
7 118.43 15.87 237.47 36.04 13.40 0.5218 16.92 55.54 104.16 13.37 27.28
8 125.49 18.62 236.62 39.51 14.84 0.5395 15.69 56.21 108.78 14.94 26.68
9 130.38 20.60 244.35 42.45 15.80 0.5409 14.49 58.94 113.00 16.48 26.16
Miluo granite 1 76.24 4.27 169.81 11.96 5.60 0.5812 76.24 342.81 87.64 4.22 60.31
2 103.65 8.09 209.12 18.56 7.81 0.5720 51.83 144.60 110.30 7.52 53.73
3 118.55 10.73 234.66 24.81 9.05 0.5410 39.52 135.66 126.11 10.55 50.22
4 130.36 12.51 257.32 29.35 9.60 0.4964 32.59 141.30 138.64 13.40 47.87
5 140.28 15.18 272.99 33.67 10.82 0.5000 28.06 146.57 149.17 16.14 46.12
6 148.96 16.69 281.71 36.67 11.20 0.4723 24.83 150.06 158.33 18.79 44.74
Yueyang granite 1 85.03 5.13 188.70 16.67 6.03 0.6260 85.03 320.05 124.69 6.01 85.80
2 112.07 8.32 233.42 23.02 7.42 0.5441 56.04 149.48 156.92 10.70 76.44
3 128.17 11.53 254.86 30.40 9.00 0.5377 42.72 155.69 179.42 15.00 71.45
4 140.90 13.68 276.97 34.68 9.71 0.5021 35.23 165.18 197.25 19.07 68.10

calculated as the ratio of disc cutter normal force to cut- efficiently. Figure 3 shows the relation between disc cutter
ter penetration depth, i.e., BI = FN/p. Specific energy of normal force and cutter penetration depth for the five rock
the rock cutting process (SE) is calculated as the ratio of types. As seen, these rock types vary from artificial rock-like
the energy input to the rock cutting volume, and it can material (Cement mortar), sedimentary rock (Longchang
also be simplified to SE = FR/V. Here, V is the rock cut- sandstone), metamorphic rock (Hezhou marble) to magmatic
ting volume per cutting distance and it can be calculated rock (Miluo and Yueyang granites). Also, these rock types
using the rock muck mass, rock density and the whole have the uniaxial compressive strength varying from 21.25
cutting distance. The last three parameters in Table 2, i.e., to 178.66 MPa and Brazilian tensile strength varying from
semi-theoretical normal force ­(FNCSM), semi-theoretical 1.59 to 7.66 MPa. However, as seen from Fig. 3a and b, the
rolling force (­ FRCSM) and semi-theoretical specific energy ­FNp–p and FN–p curves for these five rock types all can
­(SECSM), are calculated using the semi-theoretical CSM be described accurately using power functions. The powers
model, which will be introduced in details later. of these functions are within two very narrow ranges. As
shown in Fig. 3a, for peak disc cutter normal force, it ranges
2.2 Analysis of the Full‑Scale Linear Cutting Test from 0.2710 to 0.3029 and has an average value of 0.2827.
Results Considering Rock Uniaxial Compressive As seen in Fig. 3b, for mean disc cutter normal force, it
Strength ranges from 0.3587 to 0.3883 and has an average value of
0.3723. However, the coefficients of these functions increase
TBM disc cutter must have the ability to offer normal force monotonically when rock uniaxial compressive strength
high enough to penetrate into the rock surface at the pre-set increases from 21.25 to 176.88 MPa. As shown in Fig. 3a,
cutter penetration depth and crack the inside rock material it increases from 82.6900 to 190.4328. As shown in Fig. 3b,

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Full-Scale Linear Cutting Tests to Propose Some Empirical Formulas for TBM Disc Cutter… 4769

Fig. 3  Disc cutter normal force for the full-scale linear cutting tests conducted in the five rock types

it ranges from 33.9144 to 86.1401. Therefore, rock type and disc cutter normal force, marked as FN, is studied in details
rock uniaxial compressive strength do not affect the function in the following sections.
type of the ­FNp–p and FN–p curves, and they also hardly TBM disc cutter should also have the ability to offer roll-
affect the powers of these curves; however, they can surely ing force high enough to overcome the pre-torque of the cut-
affect the coefficients of these ­FNp–p and FN–p curves. If ter hub and drive the disc cutter rolling ahead to fulfil contin-
the mean disc cutter normal force is normalized using cutter uous rock chipping. Figure 4 shows the relation between disc
penetration depth with the power of 0.3723, then a very good cutter rolling force and cutter penetration depth for the five
increasing linear relation can be made between the normal- rock types. As seen in Fig. 4a and b, these F ­ Rp–p and FR–p
ized normal force and rock uniaxial compressive strength, as curves can also be described accurately using power func-
shown in Fig. 3d. That is, when rock uniaxial compressive tions. As seen in Fig. 4a, for peak disc cutter rolling force,
strength increases, disc cutter normal force per cutter pen- the powers of these functions vary from 0.5474 to 0.7735
etration depth increases continuously. Moreover, as shown and have an average value of 0.6618. As seen in Fig. 4b, for
in Fig. 3c, the peak and mean disc cutter normal forces can mean disc cutter rolling force, the powers of these functions
be well correlated using a linear fitting function with a very vary from 0.7043 to 0.8042 and have an average value of
high determination coefficient (adj. R2 = 0.9678), consistent 0.7272. When rock uniaxial compressive strength increases
with the previous studies (Cho et al. 2013; Snowdon et al. from 21.25 to 176.88 MPa, the coefficients of the F ­ Rp–p
1982). In these three studies, the intercepts of the ­FNp–FN functions increase from 5.3413 to 16.3119 (Fig. 4a) while
curves are 17.7306, 69.8 and 18.31, and the slope of the the coefficients of the FR–p curves increase from 2.7409 to
­FNp–FN curves is 1.8101, 1.59 and 1.48. The peak and mean 5.1545 (Fig. 4b). As seen in Fig. 4d, the relation between
disc cutter normal forces are well correlated, so only mean the normalized rolling force and rock uniaxial compressive

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

4770 Y. Pan et al.

Fig. 4  Disc cutter rolling force for the full-scale linear cutting tests conducted in the five rock types

strength can also be well described using an increasing Disc cutter cutting coefficient (CC) indicates the relation
linear function. Thus, similar to disc cutter normal force, between disc cutter normal and rolling forces. As seen in
the change of rock type and rock uniaxial compressive Fig. 5a, the five CC–p curves can also be described accu-
strength cannot affect the type and powers of the ­FRp–p and rately using power functions with the powers varying from
FR–p curves, but the increase of rock uniaxial compres- 0.3142 to 0.4043 and averaged at 0.3552. Unlike disc cutter
sive strength will surely increase the coefficients of these normal and rolling forces, the coefficients of these CC–p
functions. Thus, when rock uniaxial compressive strength curves decrease from 8.1856 for Cement mortar to 5.8696
increases, disc cutter rolling force per cutter penetration for Miluo granite. Thus, as seen in Fig. 5b, the normalized
depth also increases continuously. Moreover, as shown in cutting coefficient decreases quickly with the increase of
Fig. 4c, similar to disc cutter normal force, the peak and rock uniaxial compressive strength. Therefore, for same cut-
mean disc cutter rolling forces can also be well correlated ter penetration depth (e.g., 5 mm), disc cutter cutting coef-
using a linear fitting function with a very high determination ficient for low-strength rocks is higher than that of high-
coefficient (adj. R2 = 0.8910), consistent with the previous strength rocks, as shown in Fig. 5a. Normalized resultant
studies (Cho et al. 2013; Snowdon et al. 1982). In these three force of the disc cutter (NRF) indicates the acting point of
studies, the intercepts of the ­FRp–FR curves are 3.3500, 7.57 the disc cutter resultant force. As seen in Fig. 2b, if NRF
and 2.08, and the slope of the F ­ Rp–FR curves is 1.9007, equals to 0.5000, β will be half of φ because β = NRF·φ. In
1.76 and 1.65. The peak and mean disc cutter rolling forces this case, disc cutter resultant force acts at the mid-point of
are well correlated, so only mean disc cutter rolling force, the rock-cutter contact arc, bisects the rock-cutter contact
marked as FR, is studied in details in the following sections. angle and crosses the centre of the cutter hub. If NRF is
smaller than 0.5000, β will be smaller than 0.5000·φ. In

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Full-Scale Linear Cutting Tests to Propose Some Empirical Formulas for TBM Disc Cutter… 4771

Fig. 5  Disc cutter cutting coefficient for the full-scale linear cutting tests conducted in the five rock types

this case, disc cutter resultant force acts below the mid- contact arc. Also, at same cutter penetration depth (e.g.,
point of the rock-cutter contact arc and stronger indenta- 5 mm), the NRF value for low-strength rocks is generally
tion effect occurs. If NRF is larger than 0.5000, β will be higher than that for high-strength rocks, as shown in Fig. 6a.
larger than 0.5000·φ. In this case, disc cutter resultant force Therefore, when cutter penetration depths are same, disc cut-
acts above the mid-point of the rock-cutter contact arc and ter resultant force for low-strength rocks acts higher while it
weaker indentation effect occurs. As seen in Fig. 6, most of acts lower for high-strength rocks. In other words, indenta-
the NRF values are higher than 0.5000, averaged at 0.5588. tion effect of the disc cutter is more obvious when cutting
Thus, when cutting rocks using the linear cutting machine high-strength rocks while it is less obvious when cutting
at Central South University, disc cutter resultant force low-strength rocks.
generally acts higher than the assumed mid-point of the Rock boreability index (BI) is defined in the laboratory.
rock-cutter contact arc. As seen in Fig. 6a, for the five rock In the field, it is usually named field penetration index (FPI).
types, the NRF–p curves all decrease when cutter penetra- The former one is calculated as the ratio of disc cutter nor-
tion depth increases, and then they tend to remain at around mal force to cutter penetration depth and the latter one is cal-
0.5000 when cutter penetration depth is large. Thus, with culated as the ratio of thrust per cutter to penetration depth
the increase of cutter penetration depth, the acting point of per revolution. A higher rock boreability index means that
the disc cutter resultant force acts higher at first, but then it disc cutter normal force required for each unit of cutter pen-
acts lower and lower till to the mid-point of the rock-cutter etration depth is higher. As seen in Fig. 7a, decreasing power

Fig. 6  Disc cutter normalized resultant force for the full-scale linear cutting tests conducted in the five rock types

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

4772 Y. Pan et al.

functions can be used to accurately describe the five BI–p cutting grooves, or in other words, the combination of cut-
curves. The powers of these functions vary from − 0.5962 ter spacing and penetration depth. As depicted in Fig. 1, the
to − 0.6270, averaged at − 0.6107. Pan et al. (2018e) stated SE–p and SE–s/p curves can be characterized by a convex
that, for intact rock block or formation, the power of the function having a minimum point, and this minimum point
BI–p curve is about − 0.57. The value of − 0.6107 in this is marked as the optimum rock cutting condition. This state-
study is slightly higher than this value but the absolute ment can be verified by Cho et al. (2010, 2013), Pan et al.
error between them is very small (i.e., − 0.0407). As seen in (2018d, e), Roxborough and Phillips (1975) and Snowdon
Fig. 7a, when cutter penetration depth is small, rock bore- et al. (1982), and it can also be strongly proven in this study.
ability index is very high, so rock surface remains rather As seen in Fig. 8a, for the five rock types, the SE–p curves
intact and it is very difficult to be penetrated and cut off. all can be roughly fitted to quadratic functions with mini-
When cutter penetration depth is large, rock boreability mum points. For same cutter penetration depth range, the
index is very low, so rock material is seriously broken and change of specific energy for high-strength rocks is signifi-
it is very easy to be penetrated and cut off. Also, for same cant while the change of specific energy for low-strength
cutter penetration depth range, the decrease of rock bore- rocks is unobvious. The specific energy to cut low-strength
ability index for high-strength rocks is very significant while rocks remains less changed whenever cutter penetration
it is very unobvious for low-strength rocks. Therefore, the depth is small or large, and thus no obvious minimum points
increase of cutter penetration depth will surely decrease the can be marked (e.g., the SE–p curves for Cement mortar and
difficulty to cut high-strength rocks at a very high degree. To Longchang sandstone). Thus, the adjustment of cutter pen-
eliminate the influence of machine design/operation param- etration depth to obtain the optimum rock cutting condition
eters and find out a parameter to assess the rock mass bore- is not meaningful for low-strength rocks, and the cutterhead
ability, Gong (2005) defined the specific rock mass boreabil- penetration depth per revolution should be maximized to the
ity index (SRMBI) as the boreability index at the penetration allowable one to obtain the maximum excavation produc-
rate equal to 1.0 mm per revolution. Obviously, the SRMBI tion. However, for high-strength rocks, the specific energy
values are the coefficients of the BI–p curves in Fig. 7a. As is very high when cutter penetration depth is too small or
seen, they will increase when rock uniaxial compressive too large, and it is very low when cutter penetration depth
strength increases, which is consistent with the increase of is medium (e.g., the SE–p curves for Miluo and Yueyang
normalized rock boreability index in Fig. 7b. granites). Thus, the adjustment of cutter penetration depth to
Specific energy (SE) is defined as the energy input obtain the optimum rock cutting condition is very useful for
required to cut a unit volume of rock material (Teale 1965). high-strength rocks, and the cutterhead penetration depth per
Low specific energy indicates that the main part of energy revolution should be adjusted to the optimum cutter penetra-
input is used to produce elongated thick large rock chips and tion depth to obtain the highest energy utilization condition.
only a small part of energy input is used for the useless rock The horizontal coordinate of the minimum point of the
crushing and grinding, and in this case, TBM works at the SE–p curve is defined as the optimum cutter penetration
highest energy utilization condition. The value of specific depth (popt). In the field, TBM is suggested to be operated
energy depends on the interaction between two adjacent with the cutterhead penetration depth per revolution equal

Fig. 7  Rock boreability index for the full-scale linear cutting tests conducted in the five rock types

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Full-Scale Linear Cutting Tests to Propose Some Empirical Formulas for TBM Disc Cutter… 4773

to the optimum cutter penetration depth. There should also penetration depth. The vertical coordinate of the minimum
be a parameter called optimum cutter spacing used for cutter point of the SE–p curve is defined as the optimum specific
layout design. However, this parameter is not investigated in energy (or the minimum specific energy). This parameter is
this study because cutter spacings for the five rock types are used to calculate the cutterhead power and net cutting rate
all 70 mm. Hence, to consider the influence of cutter spac- (Balci 2009; Copur et al. 2014; Pan et al. 2018e). Figure 8c
ing, the ratio of cutter spacing to penetration depth (s/p) is shows the relation between the optimum specific energy
used. Figure 8b shows the relation between the optimum and rock uniaxial compressive strength. As seen, S ­ Eopt also
ratio of cutter spacing to penetration depth and rock uni- increases almost linearly with σc, consistent with the previ-
axial compressive strength. As seen, (s/p)opt increases almost ous studies (e.g., Bilgin et al. 2006; Wang et al. 2018). The
linearly with σc, consistent with the previous studies (e.g., formula in Fig. 8c also seems to suggest that the optimum
Snowdon et al. 1982). The formula in Fig. 8b seems to sug- specific energy is only determined by some rock properties
gest that the optimum ratio of cutter spacing to penetration (e.g., σc) and independent from cutter or cutting geometries
depth is only determined by some rock properties (e.g., σc) (i.e., D, T, s and p). This statement should be further checked
and independent from cutter or cutting geometries (i.e., D, because some previous studies show that the optimum spe-
T, s and p). This statement can be verified by many previ- cific energy is higher when the corresponding cutter spacing
ous studies (Cho et al. 2013; Gong et al. 2015; Roxborough and/or penetration depth is larger (Cho et al. 2013; Gong
and Phillips 1975; Snowdon et al. 1982; Tumac and Balci et al. 2015; Snowdon et al. 1982).
2015). In these studies, the optimum ratios of cutter spac- From Figs. 3, 4, 5, 6, 7, and 8, some empirical formulas
ing to penetration depth for a certain rock type are same or can be proposed to predict the TBM disc cutter performance
only slightly different whenever using large/small disc cut- simply using rock uniaxial compressive strength. These
ters, wide/narrow cutter spacings, and deep/shallow cutter empirical formulas can be summarized as follows:

Fig. 8  Optimum rock cutting condition for the full-scale linear cutting tests conducted in the five rock types

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

4774 Y. Pan et al.

( ) Brazilian tensile strength (σt), rock-cutter contact angle (φ)

FN = 27.7266 + 0.3550 ⋅ 𝜎c ⋅ p0.3723 , (1) and an equation constant (C = 2.12).
( ) �
FR = 2.3912 + 0.0166 ⋅ 𝜎c ⋅ p0.7272 , (2) �
� s ⋅ 𝜎c2 ⋅ 𝜎t
( ) P = C�
0 3
√ . (9)
CC = 7.8578 − 0.0125 ⋅ 𝜎c ⋅ p0.3552 , (3) 𝜑 R⋅T

NRF = 0.5588, (4) Then, the disc cutter resultant force ­(FTCSM) is deter-
( ) mined by integrating the contact pressure distribution func-
BI = 26.3846 + 0.3550 ⋅ 𝜎c ⋅ p−0.6107 , (5)
tion (Pθ) along the whole rock-cutter contact angle (φ), as
(s∕ p)opt = 3.2932 + 0.1075 ⋅ 𝜎c , (6) shown in Eq. (10). The contact pressure (Pθ) is assumed
to be a power function of the base contact pressure (P0),
SEopt = −7.5772 + 0.8118 ⋅ 𝜎c . (7) the rock-cutter contact angle (φ) and the contact pressure
These empirical formulas are only applicable for some distribution constant (ψ): Pθ = P0·(θ/φ)ψ. The contact pres-
specific case, i.e., the disc cutter should be 432 mm and the sure distribution constant (ψ) varies from − 0.2 to 0.2 and
cutter spacing should be 70 mm. Meanwhile, some other decreases with the increased tip width of the CCS disc cut-
important parameters are not considered in these empiri- ter (Tumac and Balci 2015). For calculation simplification,
cal formulas, including disc cutter tip width, rock Brazilian ψ can be calculated as ψ = − 0.0229·T + 0.3714 (Pan et al.
tensile strength, etc. Thus, for a better prediction, the widely 2018c) where T is the disc cutter tip width.
used semi-theoretical CSM model is checked and modified. 𝜑 𝜑 ( )𝜓
T ⋅ R ⋅ P0 ⋅ 𝜑
∫ ∫
𝜃
FTCSM = 𝜃
T ⋅ P ⋅ Rd𝜃 = T ⋅ R⋅P0 d𝜃 = .
2.3 Analysis of the Full‑Scale Linear Cutting Test 0 0
𝜑 1+𝜓
Results Considering the Semi‑Theoretical CSM (10)
Model Finally, the disc cutter normal and rolling forces are deter-
mined by Eqs. (11) and (12). During this process, the acting
Conducting full-scale linear cutting tests usually requires angle (β) of the disc cutter resultant force is assumed as half
large rock block and specific testing equipment, very time/ of the rock-cutter contact angle (φ), i.e., NRF = 0.5000 and
expense costly and sometimes even impossible. Thus, to β = 0.5000·φ.
make the better use of these tests, Rostami and Ozdemir �𝜑� 1
(1993) proposed a widely used semi-theoretical CCS disc FNCSM = FTCSM ⋅ cos(𝛽) = FTCSM ⋅ cos = 2− 6
2
cutter cutting force prediction model (also known as the 2 1 √ (11)
C 1 5 1
semi-theoretical CSM model) based on the full-scale linear ⋅ 3 3
⋅ s 3 ⋅ 𝜎c ⋅ 𝜎t ⋅ T 6 ⋅ p 3 ⋅ 2R − p,
𝜓 +1
cutting tests conducted at Colorado School of Mines. The
main advantage of this model is that it considers the influ- (𝜑) 1
FRCSM = FTCSM ⋅ sin(𝛽) = FTCSM ⋅ sin = 2− 6
ences of all important parameters (i.e., D, T, s, p, σc, σt) 2
at a quantitative level, so it can easily and accurately pre- C 1 2 1 5 5 (12)
⋅ s 3 ⋅ 𝜎c3 ⋅ 𝜎t3 ⋅ T 6 ⋅ p 6 .
dict the rock cutting results under new conditions in which
⋅
𝜓 +1
these parameters are different. This model assumes that the The meanings of these parameters can also be referenced
contact pressure is continuously distributed along the whole from List of Symbols. Using Eqs. (8)–(12), the semi-theo-
rock-cutter contact area, as shown in Fig. 2b. The basic pro- retical CSM prediction results for the full-scale linear cutting
cedure and formulas of this model are as follows. tests conducted in this study can be obtained, and they are
First, the rock-cutter contact angle (φ) is determined using also summarized in Table 2. The semi-theoretical specific
disc cutter radius (R) and cutter penetration depth (p) by energy in Table 2 is calculated as ­SECSM = FRCSM/(s·p). In
( ) Figs. 9, 10, and 11, the experimental and semi-theoretical
R−p
𝜑 = arccos . (8) disc cutter normal forces, rolling forces, and optimum spe-
R
cific energies are compared.
Second, the base contact pressure (P 0) immediately As seen in Figs. 9a and 10a, the FN–FNCSM and
underneath the disc cutter is determined by the regression FR–FRCSM curves all can be well described using linear
analysis on the large database of full-scale linear cutting test functions. Meanwhile, the y-intercepts of these linear func-
results conducted using different rock types, cutter and cut- tions all tend to be 0, but some of their slopes are larger than
ting geometries, as shown in Eq. (9). The input parameters 1.00 while some of them are smaller than 1.00. For low-
are disc cutter radius (R), disc cutter tip width (T), cutter strength rocks (Cement mortar, Longchang sandstone and
spacing (s), rock uniaxial compressive strength (σc), rock Hezhou marble), the experimental disc cutter normal and

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Full-Scale Linear Cutting Tests to Propose Some Empirical Formulas for TBM Disc Cutter… 4775

Fig. 9  Semi-theoretical and laboratory disc cutter normal forces for the five rock types

Fig. 10  Semi-theoretical and laboratory disc cutter rolling forces for the five rock types

Fig. 11  Semi-theoretical and laboratory optimum specific energies for the five rock types

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

4776 Y. Pan et al.

rolling forces all tend to be higher than the semi-theoretical penetration depth is too high, rock material under TBM disc
CSM prediction results. Thus, the semi-theoretical CSM cutter will be excessively crushed and grinded, so the spe-
model underestimates the experimental results. For high- cific energy will increase, as shown in Fig. 1a. Thus, the
strength rocks (Miluo and Yueyang granites), the experimen- right part of the S
­ ECSM–p curve for very high cutter penetra-
tal disc cutter normal and rolling forces all tend to be lower tion depth is indeed wrong. Therefore, in this study, only the
than the semi-theoretical CSM prediction results. Thus, the semi-theoretical and experimental specific energies obtained
semi-theoretical CSM model overestimates the experimen- at the optimum rock cutting condition are analyzed. The
tal results. Therefore, the semi-theoretical CSM model will experimental optimum specific energy can be obtained from
underestimate or overestimate the experimental results at the minimum point of the SE–p curve, as shown in Fig. 8a.
different degrees when rock uniaxial compressive strength In this section, the optimum cutter penetration depth for the
changes. This statement has also been investigated and veri- semi-theoretical CSM model is also determined by Fig. 8a,
fied by some previous studies (Pan et al. 2018c, f; Xia et al. and then the optimum cutter penetration depth will be incor-
2018), and thus it can be regarded as the general rule. To porated into the semi-theoretical CSM model to calculate the
fully address this issue, the FN/FNCSM and FR/FRCSM ratios semi-theoretical optimum specific energy. The semi-theoret-
are calculated and they are named as normal force modifica- ical and experimental optimum specific energies for the five
tion factor (KN) and rolling force modification factor (KR). rock types are compared in Fig. 11. As seen in Fig. 11a, the
As seen in Figs. 9b and 10b, the KN–σc and KR–σc curves semi-theoretical and experimental optimum specific ener-
both can be fitted accurately using a decreasing power func- gies can be correlated accurately using an increasing linear
tion, and they can also be written as function, as shown in Eq. (15). As seen in Fig. 11b, for dif-
ferent rock uniaxial compressive strengths, the experimental
FN = KN ⋅ FNCSM = 8.6740 ⋅ 𝜎c−0.4694 ⋅ FNCSM , (13) optimum specific energies are always nearly 1.8171 times
higher than the semi-theoretical optimum specific energies.
FR = KR ⋅ FRCSM = 12.3528 ⋅ 𝜎c−0.5282 ⋅ FRCSM . (14)
( ) ( )
As seen in Fig. 9b, when rock uniaxial compressive SEopt = −2.4634 + 1.9908 ⋅ SEopt CSM = 1.8171 ⋅ SEopt CSM .
strength is 100 MPa, normal force modification factor (KN) (15)
equals to 1.0000. As seen in Fig. 10b, when rock uniaxial Compared to Eqs. (1)–(7), Eqs. (13)–(15) can offer more
compressive strength is 117 MPa, rolling force modification accurate and reliable prediction results because they consider
factor (KR) equals to 1.0000. Thus, the experimental disc cut- more influencing parameters (i.e., D, T, s, p, σt). However,
ter normal and rolling forces are higher than the semi-theoret- all these equations have the following shortcoming. They are
ical CSM prediction results when rock uniaxial compressive derived based on the full-scale linear cutting tests conducted
strength is lower than 100 MPa while they are lower than the using same linear cutting machine (i.e., CSU-LCM), and only a
semi-theoretical CSM prediction results when rock uniaxial few cutting conditions are conducted (e.g., disc cutter diameter
compressive strength is higher than 117 MPa. These two val- is fixed at 432 mm and cutter spacing is fixed at 70 mm in this
ues are quite close to the results in Xia et al. (2018) in which study). Previous studies also have this shortcoming. For exam-
this critical value of rock uniaxial compressive strength is ples, the study of Bilgin et al. (2006) was conducted using the
marked as 120 MPa. The use of Eqs. (13) and (14) can surely linear cutting machine at Istanbul Technical University (Tur-
increase the prediction accuracy of the semi-theoretical CSM key) and an S-35/80 H conical pick manufactured by Sandvik,
model when rock uniaxial compressive strength changes. and the study of Wang et al. (2018) was conducted using the
The SE–p curves in the laboratory are in quadratic linear cutting machine at Chongqing University (China) and
function shapes, as shown in Fig. 8a. However, the SE–p several conical picks. As the improvement of previous studies
curves obtained from the semi-theoretical CSM model are and the novelty of this study, more full-scale linear cutting test
in decreasing power function shapes, as shown in Table 2. results under different cutting conditions are collected from
They have completely different curve shapes, so they cannot the previously published literatures (Abu Bakar et al. 2014;
be correlated directly. This statement can also be supported Balci 2009; Cho et al. 2010; Copur et al. 2014; Gertsch and
by Cardu et al. (2017) in which the authors stated that the Ozdemir 1992; Pan et al. 2018e; Rostami 1991; Thyagarajan
semi-theoretical CSM model does not allow the effective 2018; Tumac and Balci 2015), and some generalized empirical
estimation of the (s/p)opt and the optimal rock excavation formulas are proposed in the next section.
condition. When cutter penetration depth increases, the
continuous decrease of the semi-theoretical specific energy
shown in Table 2 does not happen realistically. When cutter

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 Table 3  Full-scale linear cutting test results at the optimum rock cutting condition for different rock types
Rock type σc σt D T s popt FNopt FRopt SEopt (FNopt)CSM (FRopt)CSM (SEopt)CSM Data resources
Unit MPa MPa mm mm mm mm kN kN MJ/m3 kN kN MJ/m3

Cement mortar 21.25 1.59 432 19 70 10.1 80.30 14.17 14.61 38.44 5.95 8.41 In this study
Shale 27.60 5.50 330 12 80 10.0 45.50 6.90 10.80 36.57 6.46 8.08 Tumac and Balci (2015)
Shale 27.60 5.50 432 18 80 9.0 69.50 8.80 8.64 65.06 9.49 13.18 Tumac and Balci (2015)
Shale 31.30 5.50 330 12 80 8.0 45.50 6.90 10.80 37.03 5.84 9.12 Balci (2009)
Fossilized limestone 31.90 3.90 330 12 75 7.0 61.10 7.80 10.44 31.35 4.62 8.79 Tumac and Balci (2015)
Fossilized limestone 31.90 3.90 432 18 75 7.0 84.60 7.60 7.56 57.65 7.40 14.09 Tumac and Balci (2015)
Arcose 34.10 4.20 330 12 80 7.0 88.30 14.00 20.16 34.33 5.05 9.02 Tumac and Balci (2015)
Arcose 34.10 4.20 432 18 80 7.0 89.90 8.30 12.60 63.12 8.10 14.47 Tumac and Balci (2015)
Busan granite 36.00 4.70 432 14 64 4.0 – 3.90 15.60 38.89 3.76 14.68 Cho et al. (2010)
Longchang sandstone 37.02 1.75 432 19 70 9.7 92.38 14.85 19.13 56.73 8.60 12.66 In this study
Saturated sandstone 43.00 1.00 292 11 76.2 9.5 23.20 3.50 3.24 23.05 4.23 5.84 Abu Bakar et al. (2014)
Dry sandstone 51.00 1.02 292 11 76.2 9.5 34.00 5.50 7.09 26.00 4.77 6.59 Abu Bakar et al. (2014)
Chongqing sandstone 60.76 1.81 432 15 80 6.1 92.56 11.80 9.92 53.71 6.43 13.17 Pan et al. (2018e)
Yeongwol limestone 64.00 8.90 432 14 60 4.0 – 3.90 30.70 69.10 6.68 27.83 Cho et al. (2010)
Limestone 65.60 7.40 432 18 80 8.0 91.00 10.30 14.76 123.14 16.91 26.43 Balci (2009)
Kartal limestone 65.60 7.40 330 12 80 7.0 73.30 9.10 15.60 64.14 9.44 16.86 Copur et al. (2014)
Paintbrush tuff 68.00 12.80 432 13.8 76.2 7.62 127.00 20.59 35.46 106.80 14.31 24.65 Gertsch and Ozdemir (1992)
Full-Scale Linear Cutting Tests to Propose Some Empirical Formulas for TBM Disc Cutter…

Hezhou marble 79.07 3.26 432 19 70 7.0 117.52 16.41 45.51 104.16 13.37 27.28 In this study
Hudong granite 91.00 12.30 432 14 48 4.0 96.20 7.30 48.60 90.34 8.73 45.49 Cho et al. (2010)
Sungnam gneiss 92.00 15.20 432 14 52 4.0 116.61 10.29 48.40 100.30 9.70 46.62 Cho et al. (2010)
Lesotho basalt-NAB 97.20 12.90 432 12.7 69.9 7.37 123.90 13.58 26.39 119.03 15.68 30.44 Rostami (1991)
Macheon granite 108.00 10.40 432 14 52 4.0 113.25 9.32 39.70 98.35 9.51 45.71 Cho et al. (2010)
Foliated rock 110.00 6.41 432 19 75 5.1 162.00 13.30 34.99 150.07 16.40 42.88 Thyagarajan (2018)
Busan tuff 115.00 25.20 432 14 68 4.0 – 7.80 26.30 150.64 14.56 53.54 Cho et al. (2010)
Paldang gneiss 124.00 13.80 432 14 56 4.0 75.50 6.53 29.00 121.47 11.74 52.42 Cho et al. (2010)
Miluo granite 127.24 5.14 432 19 70 4.2 132.18 13.18 111.56 140.88 13.96 47.48 In this study
Colorado spring granite 143.90 7.80 432 12.7 69.9 5.08 164.90 15.51 43.70 115.80 12.63 35.57 Rostami (1991)

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Yueyang granite 176.88 7.66 432 19 70 3.0 127.83 11.26 129.70 179.42 15.00 71.45 In this study
Colorado red granite 179.00 8.90 432 19 75 3.2 153.00 8.00 33.91 198.73 17.17 71.53 Thyagarajan (2018)
Hwangdeung granite 183.00 9.80 432 14 44 4.0 108.65 8.55 52.80 129.62 12.53 71.20 Cho et al. (2010)
Lyon sandstone 194.00 8.61 432 19 75 5.1 187.00 14.60 38.52 241.71 26.42 69.07 Thyagarajan (2018)

13
4777
4778 Y. Pan et al.

Fig. 12  Optimum ratio of cutter spacing to penetration depth and rock Fig. 14  Disc cutter normal force and rock uniaxial compressive
uniaxial compressive strength strength at the optimum rock cutting condition

Fig. 13  Optimum specific energy and rock uniaxial compressive Fig. 15  Disc cutter rolling force and rock uniaxial compressive
strength strength at the optimum rock cutting condition

3 Development of New Generalized test results obtained using different linear cutting machines,
Empirical Formulas for TBM Disc Cutter different disc cutters and different cutting parameters. There-
Performance Prediction fore, the results obtained by analyzing Table 3 can surely offer
us some generalized empirical formulas for TBM disc cutter
Table 3 summarizes the full-scale linear cutting test results at performance prediction. Figures 12, 13, 14, and 15 show the
the optimum rock cutting condition obtained from the previ- generalized empirical formulas to predict the optimum cut-
ously published literatures. In Table 3, rock uniaxial compres- ter spacing to penetration depth ratio, the optimum specific
sive strength varies from 21.25 to 194.00 MPa, rock Brazilian energy and the corresponding disc cutter normal and rolling
tensile strength varies from 1.00 to 25.00 MPa, disc cutter forces based on rock uniaxial compressive strength. The nor-
diameter varies from 292 to 432 mm, disc cutter tip width malized normal and rolling forces are calculated concerning
varies from 11 to 19 mm, cutter spacing varies from 44 to disc cutter diameter, disc cutter tip width and cutter spacing,
80 mm, and optimum cutter penetration depth varies from 3.0 as shown in Eqs. (18)–(19). As seen from these four figures,
to 10.1 mm. Thus, Table 3 involves the full-scale linear cutting the data points are rather scattered, which is apparent due to

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Full-Scale Linear Cutting Tests to Propose Some Empirical Formulas for TBM Disc Cutter… 4779

Fig. 16  Ratio of the semi-theoretical to laboratory optimum specific energies

Fig. 17  Ratio of the semi-theoretical to laboratory normal forces at the optimum rock cutting condition

the fact that many different cutting conditions are involved.

However, some general trends can still be marked because ⎧ 14.1261 + 0.0521 ⋅ 𝜎c (upper boundary)
these data points show a narrow band-shape distribution, ⎪
(s∕ p)opt = ⎨ 8.0088 + 0.0528 ⋅ 𝜎c (linear fitting line) , (16)
namely the optimum ratio of cutter spacing to penetration ⎪ 6.2570 + 0.0259 ⋅ 𝜎 (lower boundary)
depth, the optimum specific energy, the normalized normal ⎩ c

and rolling forces all increase when rock uniaxial compressive

strength increases. To fully address the narrow band-shape
⎧ 25.0300 + 0.2590 ⋅ 𝜎c (upper boundary)
distribution of the data points, the upper and lower boundaries ⎪
of the data points are also calculated along with the general SEopt = ⎨ 7.4914 + 0.2215 ⋅ 𝜎c (linear fitting line) , (17)
fitting trend line. The generalized empirical formulas consid- ⎪ −6.4571 + 0.2255 ⋅ 𝜎 (lower boundary)
⎩ c
ering rock uniaxial compressive strength are as follows:

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

4780 Y. Pan et al.

Fig. 18  Ratio of the semi-theoretical to laboratory rolling forces at the optimum rock cutting condition

� �
⎧ D0.5000 ⋅ s0.3333 ⋅ p0.3723 ⋅ T 0.8333 ⋅ 51.0654 + 0.5243 ⋅ 𝜎c (upper boundary)
⎪ � �
FNopt = ⎨ D0.5000 ⋅ s0.3333 ⋅ p0.3723 ⋅ T 0.8333 ⋅ 28.6873 + 0.4287 ⋅ 𝜎c (linear fitting line) ,
(18)
⎪ 0.5000 0.3333 0.3723 0.8333 � �
⎩D ⋅s ⋅p ⋅T ⋅ −2.5669 + 0.4965 ⋅ 𝜎c (lower boundary)

� �
⎧ s0.3333 ⋅ p0.7272 ⋅ T 0.8333 ⋅ 87.3101 + 0.3586 ⋅ 𝜎c (upper boundary)
⎪ � �
FRopt = ⎨ s0.3333 ⋅ p0.7272 ⋅ T 0.8333 ⋅ 42.3693 + 0.3557 ⋅ 𝜎c (linear fitting line) . (19)
⎪ 0.3333 0.7272 0.8333 � �
⎩ s ⋅p ⋅T ⋅ 6.5262 + 0.3547 ⋅ 𝜎c (lower boundary)

Figure 16, 17, and 18 show the generalized empirical for- From these formulas, not only the average values of
mulas to predict the optimum specific energy and the cor- (s/p)opt, ­SEopt, ­FNopt and F
­ Ropt are obtained, but also their
responding disc cutter normal and rolling forces based on possible maximum and minimum values are obtained. Thus,
the widely used semi-theoretical CSM model. In this case, the use of Eqs. (16)–(22) can help us to fully address the sto-
the optimum ratio of cutter spacing to penetration depth is chastic nature of the rock excavation process and to gain bet-
predicted using Eq. (16), and then it is taken into the semi- ter TBM performance prediction results. Similar procedure
theoretical CSM model to calculate the other semi-theoreti- has also been taken by Copur et al. (2014). In their study, the
cal prediction results. Though the data points are scattered, performance of two earth pressure balance (EPB) TBMs was
some generalized empirical formulas considering the semi- predicted using a stochastic model implemented into a deter-
theoretical CSM model can also be derived concerning the ministic model, where the probabilistic distributions of each
upper and lower boundaries of the data points: model parameter (not only their mean values) are considered.
This study mainly focuses on the influence of rock uni-
( )
SEopt = 0.9713 ⋅ SEopt CSM , (20) axial compressive strength on TBM disc cutter force and

⎧ 10.9561 ⋅ 𝜎 −0.4106 ⋅ �FN � (upper boundary)

⎪ c � opt� CSM � �
FNopt = ⎨ 5.2663 ⋅ 𝜎c−0.3513 ⋅ �FNopt �CSM = 0.9499 ⋅ FNopt CSM (power fitting line) , (21)
⎪ 2.3038 ⋅ 𝜎 −0.2718 ⋅ FNopt (lower boundary)
⎩ c CSM

⎧ 20.3531 ⋅ 𝜎 −0.5651 ⋅ �FR � (upper boundary)

⎪ c � opt� CSM � �
FRopt = ⎨ 6.5795 ⋅ 𝜎c−0.4372 ⋅ �FRopt �CSM = 0.8566 ⋅ FRopt CSM (power fitting line) . (22)
⎪ 1.4531 ⋅ 𝜎 −0.2192 ⋅ FRopt (lower boundary)
⎩ c CSM

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Full-Scale Linear Cutting Tests to Propose Some Empirical Formulas for TBM Disc Cutter… 4781

performance. However, TBM-excavated tunnels are usu- the TBM disc cutter performance. However, these empirical
ally constructed in hard, abrasive and fractured rock masses formulas are more suitable to the full-scale linear cutting
under overburden or/and groundwater. Therefore, geological tests conducted using the linear cutting machine at Central
discontinuous planes, in situ stress condition, rock saturation South University, China. Thus, more full-scale linear cutting
state, rock abrasivity properties and other important factors test results under different cutting conditions are obtained
may also greatly influence the TBM disc cutter force and from the previously published literatures and some general-
performance. The detailed investigation of these factors is ized empirical formulas are proposed. These new proposed
beyond the extent of this study; however, by conducting empirical formulas only use rock properties, cutter geom-
specially designed full-scale linear cutting tests, previous etries and cutting geometries to design the machine speci-
studies have already obtained some useful results, shown as fications and select the TBM operation parameters. This
follows. Yin et al. (2016) investigated the influence of joint study partly overcomes the shortcoming of the full-scale
spacing on the rock fragmentation process by TBM cutter linear cutting test and offers us a much easier and less costly
and found two crack initiation and propagation modes to method for TBM disc cutter performance prediction.
fragment the rock mass due to the variation of joint spacing.
Entacher and Schuller (2018) investigated the influence of Acknowledgements This work was financially supported by National
Natural Science Foundation of China under Grant Nos. 41807250,
foliation angle on the rolling force of a mini disc cutter and 41602326 and 41702254, National Key Basic Research Program of
found that the scaled rolling force value increases by a fac- China under Grant No. 2015CB058102, and China Postdoctoral Sci-
tor of about 2 at an alpha angle of 0° compared to 90°. Pan ence Foundation Program under Grant No. 2017M622515. The authors
et al. (2018a, d) and Ma et al. (2016a, b) investigated the are grateful for their continuous support, and also very much grateful
to the authors’ colleagues for their valuable help in organizing this
influence of confining stress on the cutting force and effi- article. Prof. Yimin Xia’s postgraduates at Central South University
ciency of a TBM disc cutter and found that, with the increase are sincerely acknowledged for helping the authors prepare the rock
of confining stress, its influence on rock fragmentation can samples and conduct the full-scale linear cutting tests. The anonymous
be divided into three stages, i.e., strengthening effect stage, reviewers are also deeply acknowledged for reviewing this article and
giving their valuable comments.
damaging effect stage and rupturing effect stage. Abu Bakar
et al. (2014) investigated the influence of moisture content
on rock behaviour and rock excavation by disc cutter and
quantified the differences between cutting dry and saturated
References
sandstone samples. Lin et al. (2017) investigated the influ- Abu Bakar MZ, Gertsch L, Rostami J (2014) Evaluation of frag-
ence of rock abrasivity properties on the wear mechanism of ments from disc cutting of dry and saturated sandstone. Rock
TBM disc cutters and obtained some matching rules of the Mech Rock Eng 47(5):1891–1903. https​://doi.org/10.1007/s0060​
TBM cutter ring properties and rock. Many more inspiring 3-013-0482-8
Atici U, Ersoy A (2009) Correlation of specific energy of cutting saws
results on full-scale rock cutting tests under different con- and drilling bits with rock brittleness and destruction energy. J
ditions can be found from the previous studies, e.g., using Mater Process Technol 209:2602–2612. https:​ //doi.org/10.1016/j.
different linear cutting machines, different disc cutters, dif- jmatp​rotec​.2008.06.004
ferent rock types, different cutting geometries and different Balci C (2009) Correlation of rock cutting tests with field performance
of a TBM in a highly fractured rock formation: a case study in
cutter control modes. In the future, their findings should be Kozyatagi-Kadikoy metro tunnel, Turkey. Tunn Undergr Space
incorporated into the empirical formulas proposed in this Technol 24(4):423–435. https:​ //doi.org/10.1016/j.tust.2008.12.001
study to establish a generalized prediction method. Balci C, Tumac D (2012) Investigation into the effects of different
rocks on rock cuttability by a V–type disc cutter. Tunn Undergr
Space Technol 30(4):183–193. https ​ : //doi.org/10.1016/j.
tust.2012.02.018
4 Conclusions Balci C, Demircin MA, Copur H, Tuncdemir H (2004) Estima-
tion of optimum specific energy based on rock properties for
This study is to correlate the TBM disc cutter performance assessment of roadheader performance. J S Afr Inst Min Metall
104(11):633–642
at the optimum rock cutting condition to rock uniaxial com- Bilgin N, Demircin MA, Copur H, Balci C, Tuncdemir H, Akcin N
pressive strength and a widely used semi-theoretical predic- (2006) Dominant rock properties affecting the performance of
tion model. First, full-scale linear cutting tests on five rock conical picks and the comparison of some experimental and theo-
types are conducted to obtain some general rules to charac- retical results. Int J Rock Mech Min Sci 43:139–156. https​://doi.
org/10.1016/j.ijrmm​s.2005.04.009
terize the change of rock cutting results with the increase of Bilgin N, Copur H, Balci C, Tumac D, Akgul M, Yuksel A (2008) The
rock uniaxial compressive strength and cutter penetration selection of a TBM using full scale laboratory tests and com-
depth. Then, the experimental results at the optimum rock parison of measured and predicted performance values in Istanbul
cutting condition are correlated with the prediction results Kozyatagi–Kadikoy metro tunnels. In: World Tunnel Congress
2008–Underground Facilities for Better Environment and Safety,
of the widely used semi-theoretical CSM model. During this India
process, some empirical formulas are proposed to predict

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

4782 Y. Pan et al.

Bruland A (1998) Hard rock tunnel boring. PhD Thesis. Norwegian Gong QM (2005) Development of a rock mass characteristics model for
University of Science and Technology, Trondheim, Norway TBM penetration rate prediction. PhD Thesis, School of Civil and
Cardu M, Iabichino G, Oreste P, Rispoli A (2017) Experimental and Environmental Engineering, Nanyang Technological University,
analytical studies of the parameters influencing the action of TBM Singapore
disc tools in tunnelling. Acta Geotech 12(2):293–304. https​://doi. Gong QM, Zhao J, Jiao YY (2005) Numerical modeling of the effects
org/10.1007/s1144​0-016-0453-9 of joint orientation on rock fragmentation by TBM cutters. Tunn
Chen LH, Labuz JF (2006) Indentation of rock by wedge–shaped Undergr Space Technol 20:183–191. https​://doi.org/10.1016/j.
tools. Int J Rock Mech Min Sci 43(7):1023–1033. https​://doi. tust.2004.08.006
org/10.1016/j.ijrmm​s.2006.03.005 Gong QM, Jiao YY, Zhao J (2006) Numerical modelling of the effects
Cho JW, Jeon S, Yu SH, Chang SH (2010) Optimum spacing of TBM of joint spacing on rock fragmentation by TBM cutters. Tunn
disc cutters: a numerical simulation using the three–dimen- Undergr Space Technol 21:46–55. https​://doi.org/10.1016/j.
sional dynamic fracturing method. Tunn Undergr Space Technol tust.2005.06.004
25(3):230–244. https​://doi.org/10.1016/j.tust.2009.11.007 Gong QM, He GW, Zhao XB, Ma HS, Li XZ (2015) Influence of dif-
Cho JW, Jeon S, Jeong HY, Chang SH (2013) Evaluation of cutting ferent cutter spacings on rock fragmentation efficiency of Beishan
efficiency during TBM disc cutter excavation within a Korean gra- granite by TBM. Chin J Geotech Eng 37(1):54–60. https​://doi.
nitic rock using linear–cutting–machine testing and photogram- org/10.11779​/CJGE2​01501​005 (in Chinese)
metric measurement. Tunn Undergr Space Technol 35(4):37–54. Gong QM, Du XL, Li Z, Wang QX (2016) Development of a mechani-
https​://doi.org/10.1016/j.tust.2012.08.006 cal rock breakage experimental platform. Tunn Undergr Space
Copur H, Aydin H, Bilgin N, Balci C, Tumac D, Dayanc C (2014) Technol 57:129–136. https​://doi.org/10.1016/j.tust.2016.02.019
Predicting performance of EPB TBMs by using a stochastic model Labra C, Rojek J, Oñate E (2016) Discrete/finite element modelling
implemented into a deterministic model. Tunn Undergr Space of rock cutting with a TBM disc cutter. Rock Mech Rock Eng
Technol 42(3):1–14. https​://doi.org/10.1016/j.tust.2014.01.006 50(3):621–638. https​://doi.org/10.1007/s0060​3-016-1133-7
Delisio A, Zhao J, Einstein HH (2013) Analysis and prediction of TBM Lin LK, Mao QS, Xia YM, Zhu ZM, Yang D, Guo B, Lan H (2017)
performance in blocky rock conditions at the Lötschberg Base Experimental study of specific matching characteristics of tunnel
Tunnel. Tunn Undergr Space Technol 33:131–142. https​://doi. boring machine cutter ring properties and rock. Wear 378–379:1–
org/10.1016/j.tust.2012.06.015 10. https​://doi.org/10.1016/j.wear.2017.01.072
Dogruoz C, Rostami J, Keles S (2018) Study of correlation between Liu HY, Kou SQ, Lindqvist P, Tang CA (2002) Numerical simula-
specific energy of cutting and physical properties of rock and tion of the rock fragmentation process induced by indenters. Int J
prediction of excavation rate for lignite mines in Çayırhan area, Rock Mech Min Sci 39:491–505. https​://doi.org/10.1016/S1365​
Turkey. B Eng Geol Environ 77:533–539. https:​ //doi.org/10.1007/ -1609(02)00043​-6
s1006​4-017-1124-2 Liu QS, Pan YC, Liu JP, Kong XX, Shi K (2016) Comparison and dis-
Entacher M, Schuller E (2018) Angular dependence of rock cutting cussion on fragmentation behavior of soft rock in multi–indenta-
forces due to foliation. Tunn Undergr Space Technol 71:215–222. tion tests by a single TBM disc cutter. Tunn Under Space Technol
https​://doi.org/10.1016/j.tust.2017.08.009 57:151–161. https​://doi.org/10.1016/j.tust.2016.02.021
Entacher M, Lorenz S, Galler R (2014) Tunnel boring machine per- Ma HS, Gong QM, Wang J, Yin LJ, Zhao XB (2016a) Study on the
formance prediction with scaled rock cutting tests. Int J Rock influence of confining stress on TBM performance in granite rock
Mech Min Sci 70(9):450–459. https​://doi.org/10.1016/j.ijrmm​ by linear cutting test. Tunn Undergr Space Technol 57:145–150.
s.2014.04.021 https​://doi.org/10.1016/j.tust.2016.02.020
Ersoy A, Atici U (2007) Correlation of P and S–waves with cutting Ma HS, Gong QM, Wang J, Zhao XB, Yin LJ, Miao CT, He GW
specific energy and dominant properties of volcanic and car- (2016b) Linear cutting tests on effect of confining stress on rock
bonate rocks. Rock Mech Rock Eng 40(5):491–504. https​://doi. fragmentation by TBM cutter. Chin J Rock Mech Eng 35(2):346–
org/10.1007/s0060​3-006-0111-x 355. https​://doi.org/10.13722​/j.cnki.jrme.2014.0926
Farrokh E, Rostami J, Laughton C (2012) Study of various models Macias FJ (2016) Hard rock tunnel boring: Performance predictions
for estimation of penetration rate of hard rock TBMs. Tunn and cutter life assessments. PhD Thesis, Norwegian University of
Undergr Space Technol 30(4):110–123. https​://doi.org/10.1016/j. Science and Technology, Norway
tust.2012.02.012 Moon T, Oh J (2012) A study of optimal rock–cutting conditions
Geng Q, Wei ZY, Meng H (2016a) An experimental research on the for hard rock TBM using the discrete element method. Rock
rock cutting process of the gage cutters for rock tunnel boring Mech Rock Eng 45(5):837–849. https​://doi.org/10.1007/s0060​
machine (TBM). Tunn Undergr Space Technol 52:182–191. https​ 3-011-0180-3
://doi.org/10.1016/j.tust.2015.12.008 Ozdemir L, Wang FD (1979) Mechanical tunnel boring prediction and
Geng Q, Wei ZY, Meng H, Macias FJ, Bruland A (2016b) Free–face– machine design. Nasa Sti/recon Technical Report N80
assisted rock breaking method based on the multi–stage tun- Pan YC, Liu QS, Liu JP, Peng XX, Kong XX (2018a) Full–scale linear
nel boring machine (TBM) cutterhead. Rock Mech Rock Eng cutting tests in Chongqing Sandstone to study the influence of
49(11):4459–4472. https​://doi.org/10.1007/s0060​3-016-1053-6 confining stress on rock cutting forces by TBM disc cutter. Rock
Gertsch R (1993) Tunnel boring machine disk cutter vibrations. MS Mech Rock Eng 51(6):1697–1713. https​://doi.org/10.1007/s0060​
thesis, Colorado School of Mines, Golden, Colorado, USA 3-018-1412-6
Gertsch R (2000) Rock toughness and disc cutting. PhD thesis, Uni- Pan YC, Liu QS, Peng XX, Kong XX, Liu JP, Zhang XB (2018b)
versity of Missouri, Rola, Missouri, USA Full–scale rotary cutting test to study the influence of disc cutter
Gertsch R, Ozdemir L (1992) Performance prediction of mechanical installment radius on rock cutting forces. Rock Mech Rock Eng
excavators from linear cutter tests on Yucca Mountain welded 51(7):223. https​://doi.org/10.1007/s0060​3-018-1460-y
tuffs: Yucca Mountain Site Characterization Project. Off Sci Tech Pan YC, Liu QS, Liu JP, Huang X, Liu Q, Peng XX (2018c) Com-
Inform Tech Rep. https​://doi.org/10.2172/13847​0 parison between experimental and semi–theoretical disc cut-
Gertsch R, Gertsch L, Rostami J (2007) Disc cutting tests in Colorado ter cutting forces: Implications for frame stiffness of the linear
Red Granite: implications for TBM performance prediction. Int cutting machine. Arab J Geosci. https​://doi.org/10.1007/s1251​
J Rock Mech Min Sci 44(2):238–246. https​://doi.org/10.1016/j. 7-018-3593-4
ijrmm​s.2006.07.007

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Full-Scale Linear Cutting Tests to Propose Some Empirical Formulas for TBM Disc Cutter… 4783

Pan YC, Liu QS, Liu JP, Liu Q, Kong XX (2018d) Full–scale linear cut- Tiryaki B (2009) Estimating rock cuttability using regression trees and
ting tests in Chongqing Sandstone to study the influence of con- artifcial neural networks. Rock Mech Rock Eng 42(6):939–946.
fining stress on rock cutting efficiency by TBM disc cutter. Tunn https​://doi.org/10.1007/s0060​3-008-0012-2
Undergr Space Technol 80:197–210. https​://doi.org/10.1016/j. Tiryaki B, Dikmen AC (2006) Effects of rock properties on specific
tust.2018.06.013 cutting energy in linear cutting of sandstones by picks. Rock
Pan YC, Liu QS, Kong XX, Liu JP, Peng XX, Liu Q (2018e) Full–scale Mech Rock Eng 39(2):89–120. https​://doi.org/10.1007/s0060​
linear cutting test in Chongqing Sandstone and the comparison 3-005-0062-7
with field TBM excavation performance. Acta Geotechnica. https​ Tumac D, Balci C (2015) Investigations into the cutting character-
://doi.org/10.1007/s1144​0-018-0702-1 istics of CCS type disc cutters and the comparison between
Pan YC, Liu QS, Liu JP, Kong XX, Peng XX, Liu Q (2018f) Investiga- experimental, theoretical and empirical force estimations. Tunn
tion on disc cutter behaviors in cutting rocks of different strengths Undergr Space Technol 45:84–98. https​://doi.org/10.1016/j.
and reverse estimation of rock strengths from experimental cutting tust.2014.09.009
forces. Euro J Environ Civil Eng. https​://doi.org/10.1080/19648​ Tumac D, Bilgin N, Feridunoglu C, Ergin H (2007) Estimation of
189.2018.15129​04 rock cuttability from shore hardness and compressive strength
Rostami J (1991) Design optimization, performance predictions, and properties. Rock Mech Rock Eng 40(5):477–490. https​://doi.
economic analysis of TBM application in the proposed Yucca org/10.1007/s0060​3-006-0108-5
mountain nuclear waste repository. MS thesis 4139, Colorado Ulusay R (2015) The complete ISRM suggested methods for rock
School of Mines, Golden, Colorado, USA characterization, testing and monitoring: 2007–2014. Springer
Rostami J (2013) Study of pressure distribution within the crushed International Publishing, Cham
zone in the contact area between rock and disc cutters. Int J Rock Wang X, Wang QF, Liang YP, Su O, Yang L (2018) Dominant cutting
Mech Min Sci 57(1):172–186. https​://doi.org/10.1016/j.ijrmm​ parameters affecting the specific energy of selected sandstones
s.2012.07.031 when using conical picks and the development of empirical pre-
Rostami J, Ozdemir L (1993) A new model for performance prediction diction models. Rock Mech Rock Eng. https​://doi.org/10.1007/
of hard rock TBMs. In: Proceedings of rapid excavation and tun- s0060​3-018-1522-1
neling conference, USA, pp 794–809 Xia YM, Guo B, Tan Q, Zhang XH, Lan H, Ji ZY (2018) Compari-
Roxborough FF, Phillips HR (1975) Rock excavation by disc cutter. son between experimental and semi–theoretical cutting forces of
Int J Rock Mech Min Sci Geomech Abstr 12(75):361–366. https​ CCS disc cutters. Rock Mech Rock Eng 51:1583–1597. https​://
://doi.org/10.1016/0148-9062(75)90547​-1 doi.org/10.1007/s0060​3-018-1400-x
Sanio HP (1985) Prediction of the performance of disc cutters in aniso- Yin LJ, Miao CT, He GW, Dai FC, Gong QM (2016) Study on the
tropic rock. Int J Rock Mech Min Sci Geomech Abstr 22(3):153– influence of joint spacing on rock fragmentation under TBM cut-
161. https​://doi.org/10.1016/0148-9062(85)93229​-2 ter by linear cutting test. Tunn Undergr Space Technol 57:137–
Snowdon RA, Ryley MD, Temporal J (1982) A study of disc cutting in 144. https​://doi.org/10.1016/j.tust.2016.02.018
selected British rocks. Int J Rock Mech Min Sci Geomech Abstr Yurdakul M, Gopalakrishnan K, Akdas H (2014) Prediction of specific
19(3):107–121. https​://doi.org/10.1016/0148-9062(82)91151​-2 cutting energy in natural stone cutting processes using the neuro-
Teale R (1965) The concept of specific energy in rock drilling. Int fuzzy methodology. Int J Rock Mech Min Sci 67:127–135. https​
J Rock Mech Min Sci Geomech Abstr 2(1):57–73. https​://doi. ://doi.org/10.1016/j.ijrmm​s.2014.01.015
org/10.1016/0148-9062(65)90022​-7
Thyagarajan MV (2018) The comparison of cutting forces on disc cut- Publisher’s Note Springer Nature remains neutral with regard to
ters in constant vs variable penetration modes. MS thesis, Colo- jurisdictional claims in published maps and institutional affiliations.
rado School of Mines, Golden, Colorado, USA

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:

1. use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
2. use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
3. falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
4. use bots or other automated methods to access the content or redirect messages
5. override any security feature or exclusionary protocol; or
6. share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at

[email protected]