Mechanical Specific Energy versus Depth of Cut in

Rock Cutting and Drilling

Yaneng Zhou1, Wu Zhang2, Isaac Gamwo2, Jeen-Shang Lin3,4 *
1
Department of Civil and Environmental Engineering, Louisiana State University, Baton Rouge, LA
70803, USA
2
U.S. Department of Energy, National Energy Technology Laboratory, Pittsburgh, PA 15236, USA
3
Department of Civil and Environmental Engineering, University of Pittsburgh, Pittsburgh, PA 15261,
USA; 4Department of Mechanical Engineering and Material Science, University of Pittsburgh, PA 15261,
USA
Abstract The relationship between Mechanical Specific Energy (MSE) and the Rate of
Penetration (ROP), or equivalently the depth of cut per revolution, provides an important measure
for strategizing a drilling operation. This study explores how MSE evolves with depth of cut, and
presents a concerted effort that encompasses analytical, computational and experimental
approaches. A simple model for the relationship between MSE and cutting depth is first derived
with consideration of the wear progression of a circular cutter. This is an extension of Detournay
and Defourny’s phenomenological cutting model. Wear is modeled as a flat contact area at the
bottom of a cutter referred to as a wear flat, and that wear flat in the past is often considered to be
fixed during cutting. But during a drilling operation by a full bit that consists of multiple circular
cutters, the wear flat length may increase because of various wear mechanisms involved. The wear
progression of cutters generally results in reduced efficiency with either increased MSE or
decreased ROP. Also, an accurate estimate of removed rock volume is found important for the
evaluation of MSE. The derived model is compared with experiment results from a single circular
cutter, for cutting soft rock under ambient pressure with actual depth measured through a
micrometer, and for cutting high strength rock under high pressure with actual cutting area
measured by a confocal microscope. Finally, the model is employed to interpret the evolution of
MSE with depth of cut for a full drilling bit under confining pressure. The general form of equation
of the developed model is found to describe well the experiment data and can be applied to interpret
the drilling data for a full bit.
Keywords: Mechanical specific energy, Depth of cut, Cutter wear, Rock cutting, Frictional contact
*Corresponding author: Jeen-Shang Lin, e-mail address: jslin@pitt.edu

1
 1. Introduction
Mechanical Specific Energy (MSE) and Rate of Penetration (ROP) are two key factors for
evaluating the efficiency of a drilling process. MSE is defined as the energy required to remove a
unit volume of rock [1]. ROP generally refers to the depth of cut per unit time and it is proportional
to the depth of cut per revolution which is equivalent to the depth of cut for a single cutter [2]. In
rock cutting, modes of failure would transit from ductile to brittle as the depth of cutting increases
[3, 4], and this study is focused on ductile failure mode of shallow cutting which is encountered in
most of the cutting and drilling operations. Most laboratory experiments have been focused on
evaluating MSE and ROP separately, to maximize ROP or to minimize MSE through investigating
the influence of rock properties and operation conditions [5-9]. Recent work shows that it is more
effective to strategize the drilling operation by combining these two parameters. For example, the
real-time surveillance of MSE is effective to optimize ROP [10, 11]. Also, extensive experiment
results show that the MSE generally decreases with ROP [12-15], indicating the possibility to
maximize ROP and minimize MSE simultaneously.
The purpose of this study is to establish a simple mathematical model between MSE and depth
of cut, to facilitate strategizing a drilling process. A phenomenological cutting model developed
by Detournay and Defourny explains well the physics of the decrease of MSE with depth of cut in
rock cutting with the incorporation of wear mechanics [2]. Specifically, the wear introduces a flat
contact area at the bottom of a cutter referred to as a wear flat, and that wear flat is considered to
be fixed during a cutting. But during a drilling operation by a full bit that consists of multiple
circular cutters, the wear flat length may increase because of various wear mechanisms [16-18].
The wear progression of cutters generally results in reduced efficiency with either increased MSE
[19, 20] or decreased ROP [6, 7, 21]. Thus, this study further takes into account the evolution of
wear during cutting. This study focuses only on drag bits such as Polycrystalline Diamond
Compact (PDC) bits.
For a single cutter, the relationship between MSE and depth of cut may be obtained in two
different ways. One approach uses a suite of cutting tests and each of the test is of a fixed cutting
depth when the cutter advances forward at a fixed speed. The other approach uses just a single
cutting test when the cutting depth steadily increases by advancing the cutter both horizontally and
vertically with their respective fixed speeds. The latter approach is similar to a drill-on or drill-off
test in which the depth of cut varies linearly in a single test [10, 22], and it is a more convenient

2
way to obtain the relationship between MSE and depth of cut. Both types of tests are analyzed in
this study.
This paper is organized as follows. A simple model for the relationship between MSE and
cutting depth is first derived by considering the wear progression of a circular cutter. The
importance of an accurate estimate of removed rock volume for the evaluation of MSE is then
illustrated through Finite Element Method (FEM) modeling. The model is further compared with
cutting experiment results from a single circular cutter, for cutting soft rock under ambient pressure
with actual depth measured through a micrometer [23], and for cutting high strength rock under
high pressure with actual cutting area measured by a confocal microscope [24, 25]. Finally, the
model is employed to interpret the evolution of MSE with depth of cut for a full drilling bit under
confining pressure [14].

2. A model of MSE versus depth of cut

2.1 Detournay and Defourny’s model
Detournay and Defourny’s phenomenological cutting model is derived from a consideration
that the bit-rock interaction is characterized by the coexistence of cutting and frictional contact [2].
The phenomenological model is simple and yet explains well the drilling action of drag bits. The
essence of the model can be explained by examining the forces acting on a cutter as illustrated in
Figure 1. For a sharp cutter, the force transmitted through a cutter is utilized in cutting rocks, and
is denoted as cutting force, Fc, while for a blunt cutter, part of the force exerted onto the cutter is
utilized in overcoming friction which is denoted as Ff. The mechanism with which the friction is
transmitted of a blunt cutter due to wear is effectively modeled by introducing a wear flat to the
cutter. For a sharp rectangular cutter, the MSE, E, for a specific depth of cut can be written as,
Fhc
E (1)
Ac

where Fhc is the horizontal component of the cutting force, and Ac is the cutting area which is the

cutting width, w, multiplied by the cutting depth, d. E so obtained is a reflection of rock properties,
and is referred to as intrinsic specific energy,  . For a blunt cutter, the additional force transmitted
F f acts on the wear flat that has an area of Af . Let the coefficient friction of the wear against the

3
rock be  , the normal force transmits through the wear flat is Fv f , which can be further written

as a contact stress  multiplied by Af , then MSE for a blunt cutter can be found as,

Fhc  Fh f Af
E     (2)
Ac Ac

where Fhf   Fv f   Af .

The contribution of the vertical force transmitted through the cutter, or the weight on bit, is
reflected in the term denoted as the drilling strength, S, as follows,
Fvc  Fv f
S (3)
Ac

For a rectangular cutter that is at least as wide as the rock sample, the contact friction at wear
flat can be obtained from the E-S diagram. In an E-S diagram, each fixed depth cut test gives a
single (S, E) data point. For a sharp cutter the data from a suite of tests plotted in an E-S diagram
will cluster around a single point, while for a blunt cutter it they would fall on a straight line. In
the case of rectilinear cut, the slope of the line is a good estimate of  [2]. Moreover, data from
the shallower depths of cut lie at the higher end of the line, while those from the deeper cuts the
lower end. In other words, the MSE decreases with depth of cut.
It is convenient to normalize the MSE, or E, of Eq. (2) with respect to some strength measure
of the rock. Under ambient pressure conditions, the unconfined compressive strength of rock is a
convenient measure. This is especially the case as extensive experiment data under ambient
pressure condition for rectangular slab cuts show that the intrinsic specific energy  is
approximately equal to the unconfined compressive strength of rock  c [26-28], and after

normalization Eq. (2) can be expressed as:

Af
   0   (4)
Ac
where  and  are the dimensionless MSE and dimensionless contact stress defined as
  E /  c and    /  c , respectively, and  0 represents the normalized energy consumed on
the cutting face with  0   /  c ; 1 for a rectilinear cutting under ambient pressure.

The MSE for a blunt cutter is strongly affected by the contact stress on the wear flat, and the
frictional contact process can be idealized as a rigid blunt tool sliding on a cohesive-frictional rock

4
[29]. The dimensionless contact stress  is predominantly governed by a dimensionless
parameter   E o tan  /  c , where Eo is the plane strain elastic modulus, and  denotes the

inclination angle of the wear flat. The frictional contact generally includes three regimes: elastic if
 is small, rigid-plastic if  is large, and elastoplastic if  is between these two limiting
behaviors [29, 31]. The dimensionless contact stress  reaches its limiting value  * in the rigid-

plastic regime, and  * is of order O(1) and it is around two when the interface friction angle is

identical to the internal friction angle under ambient pressure.  * increases with decreasing

interface friction angle for a fixed internal friction angle [30, 31].
At a depth of cut, d, for a rectangular cutter with a width of w, Ac  wd while Af  wl with l

being the length of the wear flat. The dimensionless MSE can be rewritten as follows:
l
   0   (5)
d
It can readily been seen that MSE decreases with depth for a blunt cutter with fixed wear flat
length, reflecting the fact that the wear induced friction is more significant at a shallower depth.
2.2 An extended phenomenological model
An extension to Detournay and Defourny’s phenomenological model to circular cutters is
presented herein. We consider a general scenario that a circular cutter advances with fixed but
different horizontal and vertical velocities and cuts a groove in the rock. Cuttings under ambient
and confining pressure conditions are generally different regarding the wear rate, contact stress,
and specific energy. The case under ambient pressure is discussed first. The geometrical measures
considered for a circular cutter include, radius r, rake angle  , wear flat length l , central angle
 , and curve length L along the cutter edge with contact, as shown in Figure 2. Three coordinate
systems are introduced, with two Cartesian coordinate systems, x’y’z’, xyz, and a polar coordinate
system. Based on geometrical relationships along the cutter edge, y  y ' cos  , y '  r 1  cos   ,

and the approximation for shallow cut, cos  ; 1   2 / 2 , the curve length along the cutter edge
can be expressed as:
2 ry
L  2r ; 2 (6)
cos 

5
 We further consider that the evolution of wear may increase steadily with cutting time. This is
introduced through the adoption of a wear rate defined as K  d l d , with l denoting the wear
flat length and  the contact time between the wear flat and rock. The wear flat length increases
from an initial value of l 0 before cutting to a final value of l f at the tip of the cutter when the

final depth of cut d f is reached.

For a cutter travelling both horizontally and vertically, the contact time is related to the vertical
velocity, v y , depth of cut, d , and the vertical distance between the bottom of the cutter and any

position below the rock surface, y . Specifically, the contact time  can be expressed as:
dy
 , 0 y d (7)
vy

As the wear length increases with contact time, at any location on the edge of the cutter below
the rock surface, the wear flat length can be expressed as:
K
l  l 0  K  l 0  d  y  , 0  y  d (8)
vy

The wear flat length is the largest at the bottom of the cutter after cutting, because of the longest
contact time there. With the curve length L and wear flat length l , then the friction area along
the cutter edge can be expressed as the function of depth d :
d dL 2r
Af   l dL   l dy ; 2 l 0 f d  (9)
0 dy cos 

where f  d   d 1/2  d 3/2  d r  ,   3v y l 0  2d r K  , K  0, l 0  0 . The parameter d r is a reference

depth introduced for convenience with the value of unity (i.e., d r  1 mm ). The auxiliary

dimensionless parameter  is inversely proportional to the wear rate, and thus a smaller value of
 represents a larger wear rate K . The parameter  can be conveniently determined from any
two states of the cutter, such as the initial and final states. In that way, the wear rate K can be
obtained as K   l f  l 0  v y d f , leading to the expression of the auxiliary dimensionless

parameter  :
3 df
 (10)
2 1    d r

6
where the parameter  represents the ratio of initial wear flat length over final wear flat length in

the bottom of cutter with   l 0 / l f , and it lies between 0 and 1.

The summation of two power functions can be further approximated with a simple power
function:
1
d 3/2
 1  d 
1  1
f  d   d 1/2  ; g d   d  2 (11)
 dr   dr 
Figure 3 shows that these two functions match well for different values of  , with  = 0.1, 1,
and 10. Thus the friction area can be approximated by a simple power function of depth as:
1
2r   1 12  d   1 (12)
Af ; 2 l0 d  
cos    dr 
By neglecting the influence of wear rate on the cutting area, then the cutting area can be
expressed as:
 1 
Ac  r 2    sin 2  cos  (13)
 2 

Based on the approximation of sin  ;    3! , the cutting area corresponding to y  d can

3

also be approximated as a power function of depth:

4 2 r 23
Ac ; d (14)
3 cos 
For shallow cuts under ambient pressure, the contact stress is assumed normal to the wear flat
and independent of depth [26], and both the cutting and frictional forces are accordingly obtained.
The above considerations lead to a simple model relating dimensionless MSE with depth as
follows:
1
 ; 0  (15)
d / dr 
2

with
 3l 0 
0  , 1 ; , 2 ; (16)
c 2 2 d r  1
Eqs. (15) and (16) are derived in a general case for an initially blunt circular cutter with wear
evolution (i.e., l 0  0, K  0 ). The values of 1 and  2 in some special cases are as follows:

7
 1   2  0 , for an initially sharp cutter without wear evolution  l 0  0, K  0  ;

 1 ;  l f d f ,  2 ; 0 , for an initially sharp cutter with wear evolution  l 0  0, K  0  ;

 1 ; 3l 0  2dr  ,  2 ; 1 , for an initially blunt cutter without wear evolution  l 0  0, K  0  .

For cutting in pressured environments, both the MSE and contact stress depend on confining
pressure, and thus the confined compressive strength instead of the unconfined confined strength
is used as a representative strength in the normalization. The dimensionless MSE,  ' , and
dimensionless contact stress,  ' , are defined as  '  E / CCS , and  '   / CCS , respectively,
where CCS denotes the confined compressive strength that can be estimated using the Mohr-
Coulomb failure criterion,
2 cos  1  sin 
CCS  c 3 (17)
1  sin  1  sin 

where  and c is the internal friction angle and cohesion of rock, respectively, and  3 is the

confining pressure.
The dimensionless MSE under confining pressure can thus be expressed as follows,
1
' ; 0  (18)
d / dr 
2

with
' 3  ' l 0 
0  , 1 ; , 2 ; (19)
CCS 2 2 d r  1
where  ' is the specific energy for a sharp cutter under confining pressure, and it is given as
 '  Fhc / Ac .
Eqs. (15) and (18) are the main equations of the extended model under ambient and confining
pressure conditions, respectively. The dimensionless MSE in these two equations includes two
components, with the first component  0 associated with the normalized energy dissipated

through cutting, and the second component  1 /  d / d r  2 associated with the normalized energy


dissipated through frictional contact. The dimensionless parameter  1 represents the normalized

energy associated with frictional contact at the reference depth. Figure 4 illustrates the
decomposition of the dimensionless MSE into cutting and frictional contact components in a

8
typical MSE versus depth plot, for a case with  0  1 ,   0.58 ,   1 ,   0.5 , l 0  0.2 mm ,

d r  1 mm , and d  0 ~ 1 mm . The values of  1 and  2 are computed to be 0.29 and 0.60,

respectively, based on Eq. (19). The dimensionless parameter  2 characterizes the ratio of the

initial wear flat length over the final wear flat length, as shown in Figure 5 for a typical case. The
parameter  2 decreases from 1 to 0 when  decreases from 1 to 0 with increasing wear rate, and

 2 is equal to 1 when there is no wear progression with   1. On the other hand, the parameter  1
increases with decreased  , as the normalized energy associated with frictional contact at the
reference depth is larger when the wear progression is severer with larger wear flat length.
The extended model provides a means to estimate the final wear flat length and the contact
stress, when applying the model to interpret experimental data. Based on Eqs. (10), (16) and (19),
the ratio of the final wear flat length over initial wear flat length can be estimated as:
l 3 1   2  d f
f
; 1 (20)
l0 2 2d r
The dimensionless contact stress can be estimated as by rewriting Eqs. (16) and (19) as:
2 1 2 d r
 (or  ') ; (21)
3 l 0

It is recognized that the physics of cutting under confined pressure is complex, and the specific
energy on the cutting face and contact stress on the wear flat are influenced by factors including
confining pressure and pore pressure [32]. But it still is possible to have some ideas as to what the
range of the values of the parameters of Eq. (18). Limited experiment data show that the
dimensionless MSE for a sharp or nominally sharp cutter is pressure dependent and that  0 could

increase from 1 under ambient condition to around 2.5~4.0 under the confining pressure of around
7 MPa for different rocks with different drilling fluids [33]. As for  1 , its value depends on the

contact stress which in turn depends on rock properties and pressure conditions [22]. By treating
the rock as a cohesive-frictional material, the limiting contact stress under confining pressure
condition can be related to that under ambient pressure condition based on a correspondence rule.
Specifically, the limiting contact stress under confining pressure is the product of the limiting
contact stress under ambient pressure and the ratio of differential stress over unconfined
compressive strength, when the interstitial fluid pressure between the wear flat and rock is identical
to the confining pressure and pore pressure is zero in an impermeable rock [34]. Thus the

9
dimensionless limiting contact stress, *' , can be scaled from that under ambient pressure as

follows,
d
 *'   * (22)
CCS
where  d is the differential stress that can be found as,

2 cos 
d   c   3 tan   (23)
1  sin 

The value of *' is also generally of order O(1) and this provides a basis for estimating the likely

value of  1 .

The dimensionless MSE,  ' , is the inverse of drilling efficiency that is often defined as the
ratio of confined compressive strength over MSE under confining pressure conditions [10].
Typical experiments show that the drilling efficiency generally converges to about 30~40% with
cutting depth [1, 9, 10, 15], and thus  ' would converge with cutting depth to 2.5~3.3 for these
experiments.
To test the derived model, finite element modeling is first conducted for a sharp circular cutter
without considering wear progression. Then, the model is used to interpret the experimental results
of cutting soft rock under ambient pressure and cutting hard rock under confining pressure.

3. FEM modeling of a sharp cutter

3.1 FEM model construct
FEM modeling of cutting by a sharp circular cutter under a circular motion is discussed here.
An FEM modeling of orthogonal cutting on a rectilinear path has previously been presented [35].
For the FEM model, the material model used was the “Continuous Surface Cap model” in LS-
DYNA[36], in which the fracture energy is part of the material characterization. The A cap model
is used to define yield surface, and the damage model kicks in when yielding is reached. When the
damage measure reaches about 0.999, the element is erased. Hexahedron element, which we used first,
tends to skew the fracture zone development because the regularity of mesh generally associates with such
element. The tetrahedron element provides a degree of inherent randomness that allows us to model cutting
effects well. Fracture energy is input as a material property that is independent of element size, and this
regularization reduces mesh dependence of the results. The material properties, including the elastic

10
stiffness, have been extensively calibrated based on experimental data. In the process, we do not observed
stiffer response from the element type.
The accurate measurement of depth of cut, and the actual volume of removal are critical for the
evaluation of MSE [28], both of which can be accurately determined in the FEM modeling. One-
eighth of a circular rock sample is modeled as shown in the shaded area in Relationships between
two model parameters γ1 and γ2 versus the ratio χ which is the initial wear flat length over the
final wear flat length
Figure 6(a). Sensitivity analysis has been carried out regarding the width of the sample to ensure
sufficient lateral distance[37]. A sample of the FEM mesh used is depicted in Relationships
between two model parameters γ1 and γ2 versus the ratio χ which is the initial wear flat length
over the final wear flat length
Figure 6(b), with a total number of 645,140 tetrahedron elements for rock. The width and
average length of rock sample are 11 mm and 23.4 mm, respectively; while the sample height is
3mm, where the top 2.5 mm a finely meshed zone formed with an average mesh size of 0.25 mm.
As for the PDC disc cutter, the diameter and thickness used are 13 mm and 2 mm, respectively,
and the rake angle employed is  = 15°. The tangential and vertical velocities are fixed at v = 4

m/s, and vv = 0.255 m/s, respectively. They are so selected so that the cutting depth would increase

linearly from zero on the left to 1.5 mm on the right. The maximum width of cut is 8.4 mm, when
the depth reaches the maximum value at 1.5 mm.
The present FEM cutting analysis is carried out based on Vosges sandstone with typical grain
size of 0.15 ~ 0.45 mm and an unconfined compressive strength of c = 39 MPa [38]. The adopted

material model uses an erosion algorithm that deletes the rock elements after failure [39, 40]. To
offset the influence of the erosion algorithm on the cutting force, special efforts are conducted by
tunning the dynamic strength of rock, to offer reliable cutting force observed in laboratory cutting
test [35, 41]. The dynamic yield strength added a strain-rate dependent term to the static strength.
The fracture energy, defined as the integration of the stress-displacement curve, also scales with
the dynamic yield strength. The srtain rate dependent term is calibrated via the numerical modeling
to make the specific energy close to that in a laboratory cutting test at a fixed depth.

11
3.2 Result analysis
Relationships between two model parameters γ1 and γ2 versus the ratio χ which is the initial
wear flat length over the final wear flat length
Figure 6(b) gives the configuration of a cutting in progress as well as the state of damages
incurred on the sample behind the cutter as it passes. Figure 7(a) shows the histories of three force
components, the horizontal cutting force, F , the vertical force or the thrust force, Fv , and the side

force, Fr . The cutting forces obtained are larger than the thrust forces, and the side forces are

around zero.
MSE computation requires an estimate of the volume of rock being removed. Two different
volume measures can be distinguished: the project removed volume and the actual removed
volume. The projected removed volume is the volume that is swept by a cutter through its path
over the sample; whereas the actual removed volume is literally the actual volume of rocks that is
being removed.
Given the histories of force components and projected cutting area during the cutting process,
the MSE using the projected volume, or , E p , may be calculated as:

dW F v  Fv vv  Fr vr F
Ep   ; (24)
dV p Ap v Ap

where dW is the total mechanical work per unit time, dV p = Ap v is the projected removed volume

per unit time, Ap the projected area that the cutter traces through its path in the rock orthogonal

to the cutter.
On the other hand, if the actual removed volume of rock is used, the MSE, or Ea , can be

calculated as:
F dV
Ea ; with Aa  a (25)
Aa v dt

where Aa is the actual cut area, dVa = Aa v is the actual volume cut per unit time.

In a FEM analysis both areas can be readily found. The projected area and actual cut area are
plotted in Figure 7(b). The actual removed or cut area, fluctuates as expected, whereas the
projected area is smooth. The actual removed area is generally larger, as one or two layers of rocks
beneath the cutter are damaged and thus deleted when the strength is lost. The dimensionless MSE

12
using the projected volume and actual volume, denoted as  p and a , respectively, are presented

in Figure 8, in which a moving average has been applied as a filter. To eliminate the possible effect
of boundary, the depth smaller than 0.3 mm and larger than 1.3 mm are taken out, thus only the
depth from 0.3 mm to 1.3 mm are plotted. The actual dimensionless MSE, a , computed using

the actual area is more or less independent of the depth of cut, which is consistent with the model
for a sharp cutter. However, the projected dimensionless MSE,  p , computed using the projected

area decreases with the depth and then levels off, and it is generally overestimated at small depth.

4. Interpret experiments with sharp cutters

4.1 Soft rock cutting under ambient pressure
The relationship between MSE and depth of cut for a circular cutter is analyzed using
experimental data obtained by Richard et al. from cutting soft rocks under ambient pressure
condition [23]. The tests are conducted on various low strength sedimentary rocks with different
shaped cutters, including circular, triangular and rectangular. The cutting depths are small,
typically less than 1 mm, and the rocks are deemed to have failed in a ductile mode. The cutting
depth is constant during each test, but the depth varies from test to test. For each test, the cutting
depth is measured with a micrometer at several points along the groove after cutting, and the
average depth is used as the actual cutting depth [23].
Two sets of the circular cutting data obtained using two different cutter radii are compared to
the present model. The radii of the cutter are 6.5 mm and 9.5 mm, respectively. The back rake
angle is fixed at  = 15°. The tested Lens limestone has an unconfined compressive strength, c

, of 30 MPa and an internal friction angle,  , of 38°. In addition to provide  , the E-S diagram
provides an efficient way to characterize a cutting response. Figure 9(a) clearly shows that the
cutters analyzed here are blunt and that the interface frictional between the wear flat and rock has
a coefficient around 1.2. The data with cutting depths smaller than 1 mm are studied. Eq. (15) is
found to match well these two sets of data as shown in Figure 9(b). In both cases as the cutting
was carried out under ambient pressure and the cut is rectilinear,  0 is fixed at 1.0. For the test

with the cutter radius of r  6.5 mm , the fit gives  1  0.18 ,  2  0.96 , R 2  0.93 ; while for the

test with the cutter radius of r  9.5 mm , the fit gives  1  0.10 ,  2  1.04 , R 2  0.81 .

13
 Based on the results of data fitting, the wear progression, the magnitude of the contact stress
and the influence of the wear on the MSE are analyzed. For these two sets of data, the regression
result of  2 ; 1 suggests that the wear flat length of the nominally sharp cutter is approximately

constant under ambient pressure based on Eq. (20). The contact stresses can be computed to be
approximately 44 MPa and 27 MPa respectively for the two cases, by using the estimated

interface friction coefficient of around 1.2 and an average wear flat length l 0 ; 0.065 mm in Eq.

(21). The estimated magnitude of contact stress is around one or two times the unconfined
compressive strength, and that is consistent with the solution in the elastoplastic regime and it also
agrees with other experimental data [30, 42]. The dimensionless MSE is shown to be larger than
unity at small depth (e.g., less than 0.4 mm) and then levels off at the value of around unity when
the depth is close to 1 mm, confirming that the friction of a nominally sharp cutter becomes
negligible at the deeper ends of shallow cuts.
4.2 Hard rock cutting under high pressure
To understand the complex interaction among the drilling bit, rock, and drilling fluid, an Ultra-
deep Drilling Simulator (UDS) was built at the National Energy and Technology Laboratory
(NETL) as shown in Figure 10(a) [24, 43]. The UDS was capable of simulating the cutting action
of a single PDC cutter under pressure as high as 205 MPa. Figure 10(b) gives a layout of the force
measurement system in the UDS, and one set of test results is analyzed here. The PDC disc cutter
has a diameter of 16 mm and a thickness of 1.8 mm, and the edge of the cutter is chamfered with
a length of 0.4 mm and an angle of 45°. The tests were run on a Carthage marble sample which
has a diameter of 0.2 m and a thickness of 0.2 m. The unconfined compressive strength of the rock
is around 110 MPa and the confined compressive strength is estimated to be around 317 MPa
under the applied confining pressure of 103 MPa [14]. The cutter has a rake angle of 15° and the
surface of the rock sample is tilted in a fashion so that the cutting depth will increase from 0 to 1.5
mm as the cutter is rotated without vertical translation. A clear mineral oil is used as the drilling
fluid, and a confining pressure of 103 MPa is applied through a hydraulic system. Figure 10(c)
presents a photo of the rock sample taken after cutting is completed.
Three force components are recorded during the cutting process, and two primary force
components obtained are depicted in Figure 11(a), as the side force is around zero. The E-S
diagram of Figure 11(b) shows that the cutting behavior is similar to that observed for a blunt

14
cutter [2], and friction between the wear flat and the rock surface is significant even though the
cutter is initially nominally sharp with a small chamfer. Both the MSE and drilling strength
decrease with the projected cutting depth.
To compute MSE, the projected area of cut is calculated based on the cutter geometry and
experiment setup, whereas the actual area of cut is measured from high resolution three-
dimensional images taken by a confocal laser scanning microscope. Confocal images are taken on
several radial sections of the rock sample as shown in Figure 10(c), and one example together with
the measure profile is presented in Figure 12.
Based on the projected and actual areas of cut, the corresponding dimensionless MSE are
obtained and plotted in Figure 13(a). It can be observed, just as from the FEM results, the projected
'p first decreases with depth and then levels off. Meanwhile, actual 'a also varies with depth,

albeit with a smaller change. 'p is much larger than 'a at small depth, but the difference is small

at larger depth due to small difference of the two areas of cut.

Under high confining pressure,  0 is pressure dependent and could be greater than 1. With few

data to base on, the 'a data are fitted by Eq. (18) using three different  0 values: 1, 1.5 and 2,

respectively, and the fitted curve with  0  1.0 is shown in Figure 13(b). It turns out that all lead

to reasonable good matches with the following parameter values:  0  1.0 ,  1  2.65 ,  2  0.57 ,

R 2  0.88 ;  0  1.5 , 1  2.19 ,  2  0.63 ; and R 2  0.87 ; and  0  2.0 ,  1  1.76 ,  2  0.71,

R 2  0.86 . The lower value of  0 is estimated to be 1.0 corresponding to a sharp cutter under

ambient pressure, and the upper bound value of  0 is estimated to be 2.4, based on an extended

Merchant model under confining pressure conditions [32], for a cohesive-frictional material that
follows Mohr-Coulomb failure criterion, with the unconfined compressive strength of 110 MPa,
an estimated internal friction angle of 26°, a typical interface frictional angle of 18° on the cutting
face [42], the confining pressure of 103 MPa, and zero pore pressure. All three fits clearly indicate
that the wear increase with cutting depth. For example, the final wear flat length is estimated to be
around 2.7 times the initial wear flat length based on Eq. (20) for the case with  0  1.0 . The shape

of cutter before and after cutting is shown in Figure 14. Even though the initial and final wear flat
length cannot be accurately determined, severe wear at the bottom of the cutter is observed. What

15
the results indicate is that the parameter  0 has little influence on the goodness of fit and on the

prediction of MSE for cutting hard rock under high pressure condition, since  0 contributes only

a small fraction of total MSE except when depth is large (e.g., 1 mm). The predicted values of 'a

at a depth of 0.2 mm are 7.63, 7.54, and 7.52, respectively, for the three cases with different values
 0 , and the difference is within 2%. The difference of 'a is within 5% at a depth of 1.2mm. This
implies that it suffices to study the relationship between MSE and depth of cut by simply setting
 0 to 1 in the model presented in Eq. (18), when data are not available to obtain a reliable  0
through tests with a sharp cutter under a confining pressure.

5. Application to a full bit

Considering the complex mechanism of cutter interaction and the frictional contact process, the
relationship between MSE and depth of cut for a full bit is in nature more complex than that for a
single cutter. However, experiment data show a generally similar trend of MSE to that of a single
cutter, in the sense that MSE decreases with depth and then almost levels off [12-14, 44]. Instead
of deriving a model considering all factors, Eq. (18) is applied here to see if it can be employed to
as a semi-empirical way of interpreting the drilling experiment data by a full bit, with three
empirical coefficients  0 ,  1,  2 . In a drilling test, the actual cut volume, even though unknown,

is not expected to vary significantly from the projected volume. This is mainly because the extra
cutting volume beyond the edge of bit is small compared to the volume that the bit traces. Thus
the volume of the rock removed is simply calculated by multiplying the cross area of the drilling
bit and the average depth per revolution.
In the literature there are extensive laboratory drilling test results that were carried out to
improve the drilling performance under high pressure and high temperature conditions [14]. The
tests consist of different rocks, drilling bits and drilling fluids. A set of data with similar drilling
bits and operation conditions is studied. The tested Crab Orchard sandstone has a confined
compressive strength of around 593 MPa under the confining pressure of 76 MPa [14]. The E-S
diagram in Figure 15(a) shows that the drilling response is similar to that of a blunt cutter, and the
dimensionless MSE obtained by normalizing with the confined compressive strength is presented
in Figure 15(b). Again, Eq. (18) provides a good match with the data, by assuming that the specific
energy on the cutting face to be one time, one and half and two times the confined compressive

16
strength with  0  1.0 ,  0  1.5 , and  0  2.0 , respectively, and the fitted curve with  0  1.0 is

shown in Figure 15(b). The regression results are:  1  2.08 ,  2  0.49 , R 2  0.89 ;  1  1.77 ,

 2  0.53 , R 2  0.88 ; and  1  1.47 ,  2  0.56 , R 2  0.88 , respectively. The value of  0

between 1 and 2 has little influence on the fitted curves, which have similar goodness of fit of the
lab data. It is worth noting that even though the mechanics of the drilling process by a full drilling
bit is much more complex than that of a single cutter, the relationship between MSE and depth of
cut is found to follow the same equation form.
Drilling efficiency,  , defined as the ratio of confined compressive strength over MSE can then
be expressed as follows [10]:
1 1
 ; (26)
 '  0   1  d / d r  2

The drilling efficiency increases with depth, and is around 30~50% when the depth is between
0.5~1.5 mm as shown in Figure 15(c). The drilling efficiency is less than 10% when the depth is
smaller than 0.1 mm, which may suggest that most of the energy is dissipated by frictional contact.
One possible way of employing the present model for drilling applications is the real-time
surveillance of drilling efficiency over ROP. This can be achieved by the real-time surveillance of
the evolutions of MSE and ROP during drilling [10, 11], in combination with the estimation of
unconfined and confined compressive strengths of rock based on petrophysical measurements
while drilling [45, 46].

6. Conclusions
A simple model between the MSE and depth of cut for a circular cutter is proposed by
considering the wear flat evolution with cutting depth. The derivation is an extension of Detournay
and Defourny’s phenomenological cutting model for a rectangular cutter with fixed wear flat
length. The evaluation of MSE depends on accurate measurement of the cutting volume, and using
the actual cut volume, instead of the volume the cutter swept through, is found to give a better
estimation of MSE through finite element modeling.
The proposed model could reasonably explain the evolution of MSE over depth, for cutting soft
rock under ambient pressure, and cutting hard rock under high pressure. The actual cuttings areas
are measured through a micrometer for the former and a confocal microscope, respectively. For
the former, the rate of wear progression is negligible with approximate constant wear flat length.

17
The small chamfer has large influence on MSE at small depth (e.g., less than 0.4 mm), making
cutting inefficient. Its impact diminishes at large depth (e.g., 1 mm), since the MSE levels off at
the value that is approximately equal to the unconfined compressive strength. For the latter, the
rate of wear progression is not negligible. Its MSE also generally decreases with depth and then
levels off, and is much larger in magnitude comparing with that of cutting soft rock under ambient
pressure condition, due to high confined compressive strength, high contact stress and increasing
wear flat length.
The proposed simple model also applies to the drilling by a full bit under confining pressure,
and it may serve as an approximation when a theoretical relationship between MSE and depth of
cut is not available.
Acknowledgements
This study was performed in support of the National Energy Technology Laboratory’s research
in drilling under extreme conditions under the RDS contract DE-AC26-04NT41817 and RES
contract DE-FE0004000. The authors would like to thank Professor Emmanuel Detournay at the
University of Minnesota for the benefits of stimulating discussions and comments.

References
[1] Teale R. The concept of specific energy in rock drilling. Int J Rock Mech Min Sci. 1965;2:57-
73.
[2] Detournay E, Defourny P. A phenomenological model for the drilling action of drag bits. Int J
Rock Mech Min Sci. 1992;29:13-23.
[3] Huang H, Detournay E. Intrinsic length scales in tool-rock interaction. Int J Geomech.
2008;8:39-44.
[4] Zhou Y, Lin J-S. On the critical failure mode transition depth for rock cutting. Int J Rock Mech
Min Sci. 2013;62:131-7.
[5] Kahraman S, Bilgin N, Feridunoglu C. Dominant rock properties affecting the penetration rate
of percussive drills. Int J Rock Mech Min Sci. 2003;40:711-23.
[6] Motahhari H, Hareland G, Nygaard R, Bond B. Method of optimizing motor and bit
performance for maximum ROP. J Can Petrol Technol. 2009;48:44-9.
[7] Fear M. How to improve rate of penetration in field operations. SPE Drill Completion.
1999;14:42-9.
[8] Altindag R. Correlation of specific energy with rock brittleness concepts on rock cutting. J
South Afr Inst Min Metall. 2003;103:163-71.
[9] Pessier R, Fear M. Quantifying common drilling problems with mechanical specific energy
and a bit-specific coefficient of sliding friction. In: Proceedings of the SPE Annual Technical
Conference and Exhibition. Washington, D.C.; 4-7 October 1992.

18
[10] Dupriest F, Koederitz W. Maximizing drill rates with real-time surveillance of mechanical
specific energy. In: Proceedings of the SPE/IADC Drilling Conference. Amsterdam, Netherlands;
23-25 February 2005.
[11] Amadi W, Iyalla I. Application of mechanical specific energy techniques in reducing drilling
cost in deepwater development. In: Proceedings of the SPE Deepwater Drilling and Completions
Conference. Galveston, Texas; 20-21 June 2012.
[12] Reddish D, Yasar E. A new portable rock strength index test based on specific energy of
drilling. Int J Rock Mech Min Sci. 1996;33:543-8.
[13] Huang S, Wang Z. The mechanics of diamond core drilling of rocks. Int J Rock Mech Min
Sci. 1997;34:134.e1-134.e14.
[14] TerraTek. Improving deep drilling performance. Salt Lake City: TerraTek company; 2008.
[15] Armenta M. Identifying inefficient drilling conditions using drilling-specific energy. In:
Proceedings of the SPE Annual Technical Conference and Exhibition. Denver, Colorado; 21-24
September 2008.
[16] Lin T-P, Michael H, George AC, Li X. Wear and failure mechanisms of polycrystalline
diamond compact bits. Wear. 1992;156:133-50.
[17] Appl F, Wilson CC, Lakshman I. Measurement of forces, temperatures and wear of PDC
cutters in rock cutting. Wear. 1993;169:9-24.
[18] Glowka D. Use of single-cutter data in the analysis of PDC bit designs: part 1-development
of a PDC cutting force model. J Petrol Technol. 1989;41:797-9.
[19] Rashidi B, Hareland G, Nygaard R. Real-time drill bit wear prediction by combining rock
energy and drilling strength concepts. In: Proceedings of the Abu Dhabi International Petroleum
Exhibition and Conference. Abu Dhabi, UAE; 3-6 November 2008.
[20] Waughman R, Kenner J, Moore R. Real-time specific energy monitoring reveals drilling
inefficiency and enhances the understanding of when to pull worn PDC bits. In: Proceedings of
the IADC/SPE Drilling Conference. Dallas, Texas; 26-28 February 2002.
[21] Warren T, Armagost W. Laboratory drilling performance of PDC bits. SPE Drill Eng.
1988;3:125-35.
[22] Detournay E, Richard T, Shepherd M. Drilling response of drag bits: theory and experiment.
Int J Rock Mech Min Sci. 2008;45:1347-60.
[23] Richard T, Coudyzer C, Desmette S. Influence of groove geometry and cutter inclination In
rock cutting. In: Proceedings of the 44th US Rock Mechanics Symposium and 5th US-Canada
Symposium. Salt Lake City, Utah; 27-30 June 2010.
[24] Zhang W, Thornton J, Whipple G, Lindner E. The initial single cutter test of the NETL ultra
deep drilling simulator. In: Proceedings of the 45th US Rock Mechanics / Geomechanics
Symposium. San Francisco, California; 26-29 June 2011.
[25] Zhou Y, Zhang W, Gamwo IK, Lin JS, Eastman H, Gill M, et al. Mechanical specific energy
versus depth of cut. In: Proceedings of the 46th US Rock Mechanics / Geomechanics Symposium.
Chicago, Illinois; 24-27 June 2012.
[26] Dagrain F. Influence of the cutter geometry in rock cutting with PDC cutters: an experimental
approach [M.S. Thesis]. University of Minnesota, Minneapolis, Minnesota2001.
[27] Dagrain F. Influence of the cutter geometry in rock cutting with PDC cutters: an experimental
approach [M.S. Thesis]: University of Minnesota; 2001.
[28] Richard T, Dagrain F, Poyol E, Detournay E. Rock strength determination from scratch tests.
Eng Geol. 2012;147:91-100.

19
[29] Adachi J. Frictional contact in rock cutting with blunt tools [M.S. Thesis]. University of
Minnesota, Minneapolis, Minnesota1996.
[30] Zhou Y, Detournay E. Analysis of the contact forces on a blunt PDC bit. In: Proceedings of
the 48th US Rock Mechanics/Geomechanics Symposium. Minneapolis, Minnesota; 1-4 June 2014.
[31] Zhou Y, Detournay E. Elastoplastic model of a blunt rigid tool sliding on rock. Manuscript in
preparation.
[32] Detournay E, Atkinson C. Influence of pore pressure on the drilling response in low-
permeability shear-dilatant rocks. Int J Rock Mech Min Sci. 2000;37:1091-101.
[33] Rafatian N, Miska SZ, Ledgerwood LW, Yu M, Ahmed R, Takach NE. Experimental study
of MSE of a single PDC cutter interacting with rock under simulated pressurized conditions. SPE
Drill Completion. 2010;25:10-8.
[34] Zhou Y. The applicability of correspondence rule with inclined load. Rock Mech Rock Eng.
2017;50:233-40.
[35] Jaime MC, Zhou Y, Lin J-S, Gamwo IK. Finite element modeling of rock cutting and its
fragmentation process. Int J Rock Mech Min Sci. 2015;80:137-46.
[36] Murray YD. User manual for LS-DYNA concrete material model 159. Mclean: U.S.
Department of transportation, Federal Highway Administration 2007.
[37] Zhou Y, Jaime MC, Gamwo IK, Zhang W, Lin JS. Modeling groove cutting in rocks using
finite elements. The 45th US Rock Mechanics / Geomechanics Symposium. San Francisco, CA,
USA: Curran Associates, Inc.; 2011.
[38] Bésuelle P, Desrues J, Raynaud S. Experimental characterisation of the localisation
phenomenon inside a Vosges sandstone in a triaxial cell. Int J Rock Mech Min Sci. 2000;37:1223-
37.
[39] Hallquist JO. LS-DYNA Keyword user's manual. Livermore, CA: Livermore Software
Technology Corporation; 2009.
[40] Murray YD. User manual for LS-DYNA concrete material model 159. Mclean, Virginia:
Federal Highway Administration, 2007.
[41] Richard T. Determination of rock strength from cutting tests [M.S. Thesis]. University of
Minnesota, Minneapolis, Minnesota1999.
[42] Almenara R, Detournay E. Cutting experiments in sandstones with blunt PDC cutters. In:
Proceedings of the International Society of Rock Mechanics Symposium EuRock. Chester, UK;
14–17 September 1992.
[43] Zhang W, Ayers WJ, Brown J, Bromhal G. Optimizing drilling parameters: a preliminary
model – drilling carthage mable with oil based mud and PDC bit. In: Proceedings of the 49th US
Rock Mechanics/Geomechanics Symposium. San Francisco, California; 28 June-1 July 2015.
[44] Hamrick TR. Optimization of operating parameters for minimum mechanical specific energy
in drilling [Ph.D. Thesis]. West Virginia University, Morgantown, West Virginia2011.
[45] Chang C, Zoback MD, Khaksar A. Empirical relations between rock strength and physical
properties in sedimentary rocks. J Petrol Sci Eng. 2006;51:223-37.
[46] Hareland G, Nygaard R. Calculating unconfined rock strength from drilling data. In:
Proceedings of the 1st Canada-US Rock Mechanics Symposium. Vancouver, Canada; 27-31 May
2007.

20
 List of Figs.

Figure 1 Forces considered in the phenomenological model for a shallow rectangular cut: (a) A
sharp cutter, (b) A blunt cutter [2]
Figure 2 (a)Cartesian coordinates and geometrical terms, (b) Cartesian coordinates, polar
coordinates and geometrical terms
Figure 3 Approximation of the summation of two power functions of f(d) into one signle power
function g(d)
Figure 4 Decomposition of dimensionless MSE into cutting and frictional contact components in
a typical MSE versus depth plot
Figure 5 Relationships between two model parameters γ1 and γ2 versus the ratio χ which is the
initial wear flat length over the final wear flat length
Figure 6 (a) Sketch of rock sample used in the circular cutting model, (b) A snapshot showing the
cutting in progress in FEM
Figure 7 (a) Histories of forces obtained from FEM, (b) History of actual cut area and projected
area from FEM
Figure 8 Evolution of dimensionless MSE with cutting depth from FEM
Figure 9 (a) E-S diagram under ambient pressure condition, (b) Evolution of dimensionless MSE
with cutting depth under ambient pressure condition
Figure 10 (a) The NETL UDS system, (b) UDS force measurement layout, (c) Carthage marble
sample [25]
Figure 11 (a) History of tangential force and vertical force with increasing projected cutting depth
in UDS, (b) E-S diagram in UDS
Figure 12 (a) Confocal image of a typical segment after cutting in UDS, (b) An extracted sample
cross-sectional profile
Figure 13 (a) Evolution of dimensionless MSE with cutting depth in UDS, (b) An interpretation
of this dimensionless MSE with depth
Figure 14 The wear of cutter in UDS: (a) Cutter before cutting,
(b) Cutter after cutting [24]
Figure 15 (a) E-S diagram of a full drill bit, (b) Evolution of dimensionless MSE with depth per
revolution, (c) Evolution of drilling efficiency with depth per revolution

21
 (a)

(b)

Figure 1 Forces considered in the phenomenological model for a

shallow rectangular cut: (a) A sharp cutter, (b) A blunt cutter [2]
(a)

(b)

Figure 2 (a) Cartesian coordinates and geometrical terms, (b)

Cartesian coordinates, polar coordinates and geometrical terms
 α = 0.1

α =1

α = 10

Figure 3 Approximation of the summation of two power

functions of f(d) into one single power function g(d)
 Frictional contact
Frictional contact

γγ 1
1

γγ 0 Cutting
0 Cutting

Figure 4 Decomposition of dimensionless MSE into

cutting and frictional contact components in a typical
MSE versus depth plot
Figure 5 Relationships between two model parameters γ1
and γ2 versus the ratio χ which is the initial wear flat length
over the final wear flat length
(a)

(b)

Figure 6 (a) Sketch of rock sample used in the circular cutting

model, (b) A snapshot showing the cutting in progress in FEM
 0.6
F
(a) 0.5
F
v
0.4
F
r

0.3
Force(kN)

0.2

0.1

-0.1
0 0.5 1 1.5
Cutting depth(mm)

(b) 12

10

8
)
2

6
A (mm

A
a
2
A
p
0
0 0.5 1 1.5

Cutting depth (mm)

Figure 7 (a) Histories of forces obtained from FEM, (b)

History of actual cut area and projected area from FEM
 3

p
2.5
a

1.5

0.5

0
0 0.5 1 1.5
Cutting depth (mm)

Figure 8 Evolution of dimensionless MSE with cutting

depth from FEM
(a) 80
r = 6.5 mm
r = 9.5 mm

60
E (MPa)

40

20

0
0 20 40 60 80

S (MPa)
(b)

Figure 9 (a) E-S diagram under ambient pressure condition,

(b) Evolution of dimensionless MSE with cutting depth
under ambient pressure condition
 (b)
(a)

(c)

Figure 10 (a) The NETL UDS system, (b) UDS force

measurement layout, (c) Carthage marble sample [25]
 16
F
(a) 14
F
v
12

10
Force (kN)

2
0 0.5 1 1.5

Cutting depth (mm)

(b) 30

20
E (GPa)

10

0
0 10 20 30

S (GPa)

Figure 11 (a) History of tangential force and vertical force with

increasing cutting depth in UDS, (b) E-S diagram in UDS
 (a)

(b)
2000

1500
m)

1000
z(

500

0
0 4000 6000 12000 16000
x( m)

Figure 12 (a) Confocal image of a typical segment after

cutting in UDS, (b) An extracted sample cross-sectional
profile
 30
(a) '

p
'
25
a

20

15
'

10

0
0 0.5 1 1.5

Cutting depth (mm)

(b)

Figure 13 (a) Evolution of dimensionless MSE with cutting depth

in UDS, (b) An interpretation of this dimensionless MSE with
depth
(a)

1mm

(b)

1mm

Figure 14 The wear of cutter in UDS: (a) Cutter before

cutting, (b) Cutter after cutting [24]
 20
(a)

E (GPa)

10

0
0 10 20

S (GPa)

20

(b) Lab data

2 ' 0.49
Fitted curve: R = 0.89 ( = 1.00 + 2.08/(d/d ) )
15 a r
a

10
'

0
0 0.5 1 1.5

Depth per revolution (mm)

Figure 15 (a) E-S diagram of a full drill bit, (b) Evolution of

dimensionless MSE with depth per revolution, (c) Evolution of
drilling efficiency with depth per revolution
 (c)50

40

30
(%)

20

10 Lab data
2 0.63
Fitted curve: R = 0.89 ( = 1/(1.00 + 1.57/d ))

0
0 0.5 1 1.5

Depth per revolution (mm)

Figure 15 Continued