Tunnelling and Underground Space Technology 72 (2018) 162–173

Contents lists available at ScienceDirect

Tunnelling and Underground Space Technology

journal homepage: www.elsevier.com/locate/tust

Horizontal directional drilling: State-of-the-art review of theory and T

applications
⁎
Xufeng Yana, Samuel T. Ariaratnamb, Shun Donga, Cong Zenga,
a
Faculty of Engineering, China University of Geosciences, Wuhan 430074, China
b
Ira A. Fulton Schools of Engineering, Arizona St. University, P.O. Box 873005, Tempe, AZ 85287-3005, United States

A R T I C L E I N F O A B S T R A C T

Keywords: Historically, new underground utility pipelines have typically been installed by traditional open-cut methods,
Horizontal directional drilling which sometimes results in environmental impacts or damage to existing infrastructure such as roadways and
State-of-the-art other surface structures. Furthermore, open-cut construction possesses challenges for installing pipelines be-
Models neath water bodies such as rivers and lakes. Horizontal Directional Drilling (HDD) provides a method for in-
Technology
stalling underground utility pipelines in conditions where open-cut methods are unsuitable. The adoption of
HDD has increased over the past decade, as new pipelines are being installed in crowded urban areas.
Subsequently, researchers have sought to develop basic engineering theoretical models and technological in-
novations to further increase its adoption. This paper provides a state-of-the-art review and evaluation on trends
in the theoretical development of pullback loads, borehole stability and borehole mud pressure estimation
models. Innovative techniques and new pipe materials are discussed that have served to expand the technolo-
gical envelope of HDD.

1. Introduction
Horizontal Directional Drilling (HDD), with origins in the oil and
gas industry, is a trenchless technology employed to install under-
ground pipelines with minimal impacts on the environment or damage
to existing infrastructure such as roadways and other surface structures.
The process starts with the surface launched drilling of a pilot hole
along the designed design path at an entry angle of 8–16° by a drill rig.
The initial pilot bore is subsequently enlarged with a series of different
diameter reamers before the product pipe is installed as shown in Fig. 1
(Ariaratnam and Lueke, 2002). During the drilling process, the con-
tractor uses high performance drilling fluids to transport the drill cut-
tings to the surface, maintain borehole stability, and cool the drill bit
(or reamers). The first installation using HDD was in 1971 for a crossing
of the Pajaro River in Watsonville, California to install a 187.5 m steel
natural gas pipe. Today, with the development of HDD innovations, it
has become an important and effective method for pipeline installation
for different uses including product oil, natural gas, water, sewer,
electrical and telecommunications (Ma and Najafi, 2008).
In HDD practice, there are three main engineering concerns: (1)
Pullback Load; (2) Borehole Instability; and (3) Mud Pressure
Prediction. Since the pullback load balances the resistance forces during
pipe installation, it becomes a key parameter in selecting the
appropriate drill rig and evaluating the pipe stress level during pullback
(Chehab, 2008). Borehole instability is the main reason that caused the
mud loss, and it is related to the borehole mud pressure. When the
borehole pressure exceeds the maximum allowable pressure of the soil,
failure of the soil initiated, and the mud in the annular space region will
erode into the overburden accordingly. With the development of soil
failure, a considerable amount of mud will be lost, which is called a
“hydro-fracture” or “blow-out” (Xia and Moore, 2006).
This paper provides a state-of-the-art review and evaluation on
trends in the theoretical development of pullback loads, borehole sta-
bility and borehole mud pressure estimation models. It then discusses
new techniques that improve the efficiency of HDD.
2. Development of HDD theory
2.1. Pullback load estimation
The pullback load balances the resistance forces during pipe in-
stallation using HDD. It is a key parameter for selecting a drill rig with
appropriate pullback capacity and evaluating the stresses during pro-
duct pipe installation. Therefore, it is necessary to accurately estimate
the anticipated pullback load when designing HDD projects. The key
resistance forces during pullback include: (1) resistance force between

⁎
Corresponding author.
E-mail addresses: ariaratnam@asu.edu (S.T. Ariaratnam), zengcong@126.com (C. Zeng).

https://doi.org/10.1016/j.tust.2017.10.005
Received 7 June 2017; Received in revised form 6 September 2017; Accepted 16 October 2017
Available online 22 November 2017
0886-7798/ © 2017 Elsevier Ltd. All rights reserved.
X. Yan et al.
Fig. 1. Horizontal directional drilling process.
pipe and ground surface and borehole wall (caused by gravity and
buoyancy of pipe); (2) resistance force at the curves (caused by Capstan
effect and bending/stiffness effect); and (3) drag force of drilling fluid
(caused by its viscosity).
2.1.1. Components of pullback load
2.1.1.1. Resistance force between pipe and ground surface. The product
pipe is on the ground surface prior to being pulled into the borehole.
When the pullback process starts, the product pipe begins to transcend
into the borehole and incurs a reverse friction force as a result of the
interaction between the pipe and ground surface as illustrated in Fig.2.
The ground friction is maximum at first, but then decreases as more
pipes are pulled into the borehole. This ground friction is based on the
Coulomb Friction law and can be expressed as:
Tg = (wp gμg cosβ0 + wp g sinβ0) Lg
where Tg is the ground friction force, wp is the pipe weight per unit
length, β0 is the angle of ground surface, Lg is the total length of pipe
lying on the ground surface, μg is the friction factor between pipe and
ground surface (ranges from 0.1 to 0.5, depend on the pipe material and
roughness, ground surface condition, Baumert and Allouche, 2002;
Baumert et al., 2004). In practice, constructors usually use rollers,
slings or even water ditches to minimize friction (Fig. 3), and thus μg
could be as small as 0.1 (ASTM F 1962 – 99).
2.1.1.2. Resistance force between pipe and borehole wall. When the
Fig. 2. Resistance force between product pipe and ground surface.
Tunnelling and Underground Space Technology 72 (2018) 162–173
product pipe is pulled into the borehole, the interaction between the
pipe and borehole wall produces a resistance force that counters
movement of the pipe. The drill path is not a perfect horizontal line
or a skew line, and therefore can be assumed to be a polyline comprised
by several skew lines (Lb1, Lb2…Lbi…). The resistance force for one skew
segment includes: (1) friction force between pipe and borehole wall;
and (2) the net buoyancy force component along the pipe axial
direction as illustrated in Fig.4.
The total borehole resistance force is the sum of all skew segments
resistance forces as shown in Eq. (2):
k
Tb = ∑ (|wb gμb cosβbi Lbi | + wb g sinβbi Lbi )
i=1 (2)
where Tb is the borehole resistance force, wb is the net buoyancy force
per unit length, βbi is the pipe inclination to horizontal of the pipe
segment Lbi, μb is the friction factor between pipe and borehole wall
(the suggested values for this variable range from 0.21 to 0.3 (Maidla
and Wojtanowicz, 1987; Driscopipe, 1993)). The suggested values for
HDPE pipe and the borehole range from 0.2 to 0.5 (El Chazli, 2005;
Rabiei et al., 2016). In practice, product pipes may be filled with water
for additional weight to reduce friction and lower the contact pressure
caused by buoyancy within the borehole, especially when crossing
water bodies.
2.1.1.3. Resistance force at the curves. Resistance force at the curves is
produced by capstan and bending (stiffness) effects. When the product
pipe is pulled crossing a curve section, the direction of the pull force
changes, resulting in an increase in the contact pressure between the
pipe and borehole wall. The increase in frictional force at the curve
section is referred to as the Capstan effect (Fig. 5) (Lasheen and Polak,
2001). In addition, because of the flexural rigidity of the product pipe,
the increasing bending stress will also result in an increase of normal
force between the pipe and borehole wall during crossing of curve
sections. The frictional force increases and is referred to as the bending/
stiffness effect (Fig. 6) (Huey et al., 1996). The capstan effect usually
occurs in flexible pipe, while the bending effect is more prevalent in
large diameter steel pipes (ASTM F 1962-05).
Assuming the capstan effect and bending effect are independent,
and the borepath does not change during pullback, the resistance force
caused by the capstan effect can be shown in Eq. (3) (Chehab, 2008):
(Tb)i = (Tb)i − 1e μb·Δβ (3)
(1) where (Tb)i and (Tb)i-1 are the tensile forces before and after the curve,
respectively, and Δβ is the curve angle.
Based on the mechanics of materials, the additional force caused by
the bending effect was investigated by various researchers (Dareing and
Ahlers, 1991; Polak and Lasheen, 2001; Polak and Chu, 2005). As-
suming the borehole is rigid, and the contact pressure between pipe and
borehole wall is concentrated, the resistance force related to the
bending effect can be calculated using Eq. (4). In reality, the contact
force between the pipe and borehole wall is a non-uniform pressure
acting on a contact area around the middle span of the curved section
(Hair, 2002).
Tbc =
2
(
3μb Ep Ip (Δβ ) · 1 + 4cos
Δβ
2 )
Lbc2
(
sin (Δβ )· 1 + cos
Δβ
2 ) (4)
where Tbc is the additional force caused by bending effect, Lbc is the
curved length, Ep is the elastic modulus of pipe material, Ip is the mo-
4
ment of inertia of the pipe (I = π (Dop −Dip4 )/64 ), Dop is the outer dia-
meter of pipe and Dip is the inner diameter of pipe.
2.1.1.4. Drag force of drilling fluid. When the product pipe is pulled
through the borehole, the relative movement between pipe and viscous
drilling fluid produces a drag force as illustrate in Fig. 7. Huey et al.
163
X. Yan et al. Tunnelling and Underground Space Technology 72 (2018) 162–173

Fig. 3. Methods to reduce friction force between pipe and

ground surface.

Fig. 7. Fluid drag force.

Fig. 4. Resistance force between pipe and borehole wall. borehole is Power Law fluid; and (2) the pipe and borehole are con-
centric. Subsequently, the drag force can be expressed as:
n i=1
dν (r )
(Td )i = K ⎛ |r = Rp ⎞ ·πDp ∑ Lbi
⎝ dr ⎠ k=1 (5)

r dp 1 R1 21
1
⎧
⎪ νp + ∫Rp [ 2K (− dz )] n ( r −r ) n drRp ≦ r ≦ Rl
ν (r ) =
⎨ Rb 1 dp 1 R2 1
⎪ ∫r [ 2K (− dz )] n (r − rl ) n drRl ≦ r ≦ Rb. (6)
⎩
Rb
Q= ∫R p
2πrν (r ) dr
(7)
where (Td)i is the drag force of mud, n is the flow index, and K is the
consistency index, Dp and Rp are the diameter and radius of pipe re-
spectively, v(r) is the velocity distribution of mud, Rl is the radius where
maximum mud velocity is, Rb is the radius of borehole, vp is the pull-
back velocity, dp/dz is the pressure grade along pipe axial, and Q is the
flow rate.
Fig. 5. Capstan effect. Then, dp/dz and v(r) can be obtained from Eqs. (6) and (7), and the
drag force can be calculated from Eq. (5). Since this method is im-
practical to calculate on the job site, a simplified equation to estimate
the drag force was developed as Eq. (8) (Chevron, 1999; ASTM F
1962–99).
π
THK = q (Db2−Dp2)
8 (8)
where q is the hydrokinetic pressure (estimated to be 70 kPa), Db is the
Fig. 6. Bending effect. borehole diameter.
In reality, the pipe and borehole should always be eccentric, thereby
leading to lower drag forces that can be calculated by multipling a
(1996) and Puckett (2003) suggest using pipe size or HDD installation
reduction eccentricity factor (Haciislamoglu and Langlinais, 1990;
load measurements to estimate the drag force for steel pipe, while
Chin, 2001; Baumert et al., 2005).
Driscopipe (1993) and Chin (2001) believe that the drag force is too
small for HDPE pipe and can be neglected.
Theoretically, the drag force can be determined by the mechanics of 2.1.2. Pullback load estimation models
viscous fluids with the following assumptions: (1) the mud in the Polak and Lasheen (2001), Polak and Chu (2005), and Cheng and
Polak (2007) proposed a model considering the bending effect and drag

164
X. Yan et al.
Fig. 8. Drill path.
force in estimating pullback load. However, the assumption of New-
tonian fluid affects the accuracy of the results. Currently, the ASTM F
1962 model and PRCI model are the most widely used estimation
models, while in China, the GB50424-2015 model and GB50268-2008
model are used primarily for oil (or natural gas) pipe and municipal
pipeline installations, respectively.
2.1.2.1. ASTM F 1962 model. This model simplifies the drill path as five
segments as shown in Fig.8 with the three main assumptions: (1) the
entry and exit points are in the same elevation; (2) the bending effect is
neglected; and (3) the middle segment is horizontal. Subsequently, the
pullback forces at different points can be calculated by Eq. (9).
⎧TA = e μg α μ
g wp (L1 + L 2 + L3 + L4 )
⎪T = e μb α [T + T + μ |w |L + w H −e μg αμ w L ]
⎪ B A HK b b 2 b g p 2
⎨TC = TB + THK + μb |wb |L3−e μb α (e μg αμg wp L3)
⎪ borehole/ground surface, mud drag force and capstan effect, and pro-
⎪TD = e μb β [TC + THK + μb |wb |L4 −wb H −e μb α (e μg αμg wp L4 )]
⎩ duces a series pullback forces for different pullback process. This model
where TA, TB, TC, TD are the pullback load at point A, C, D, F
respectively, L1 is the additional length of pipe, L2, L3, L4 are the
horizontal length of AB, BC, CD respectively, α and β are the entry and
exit angle, respectively and H is the maximum burial depth of pipe.
2.1.2.2. Pipeline Research Council International (PRCI) model. Assuming
the maximum pullback load occurs at the end of the pullback process,
and accounting for the friction force, bending effect and mud drag
force, Huey et al. (1996) suggested that the pullback force is the sum of
individual forces. They balanced the resistance forces at straight and
curved sections in the borehole and established a closed form
estimation model that was used by the Pipeline Research Council
International (PRCI) and is widely used in practice today. As shown in
Figs. 4 and 5, the model can be expressed as Eq. (10):
⎧ (Tb)i = (Tb)i − 1 + |Ff | + DRAG ± wb Lbi sinβbi (straight section)
⎪
⎪ (Tb)i = (Tb)i − 1 + 2|Ff | + DRAG ± wb Lbc sinβbi (curved section)
k
⎨
⎪T = ∑ ((Tb)i−(Tb)i − 1) (total pullback load )
⎪
⎩ 1
(10)
where Ff is the friction between pipe and borehole wall, DRAG is the
mud drag force, Lbc is the arc length of pipe.
It should be noted that the PRCI model does not consider the impact
of ground surface friction, which may underestimate pullback load. As a
result, it was modified in the Driscopipe and Drillpath models
(Driscopipe, 1993; Drillpath, 1996). However, the Driscopipe model
was intended for plastic pipe, assumed that the borehole is polyline,
and ignored the capstan effect, bending effect and mud drag force.
While the Drillpath model provides a 3-D method similar to the PRCI
model and considers the capstan effect, it neglects the bending effect
and drag force.
2.1.2.3. GB50424-2015 model. The GB50424-2015 model was
developed to provide a simple calculation for steel pipe installations
using HDD for oil and natural gas as shown in Eq. (11):
Tunnelling and Underground Space Technology 72 (2018) 162–173
2
⎡ Dop δ ⎤
T = Lμ ⎢π ρm g−π (2Dop−δ ) ρp g−Wb w⎥ + πDop Lk v
⎣ 4 4 ⎦ (11)
where L is the total length of pipe, μ is the friction factor (0.3), g is the
acceleration of gravity, ρm is the mud density, ρp is the density of pipe, δ
is the thickness of pipe, Wbw is the balance weight of pipe per unit
length, kv is the coefficient of viscosity (0.175 kN·s/m2).
This model, however, neglects several factors, resulting in an un-
derestimation of actual pull load. Therefore, the standards suggest
multiplying the calculated results by 1.5–3 times for more accuracy.
2.1.2.4. GB50268-2008 model. The GB50268-2008 model was
developed as guidance for installation of municipal pipelines using
HDD. In this model, the pullback load is assumed to be the sum of the
resistance force on the reamer and the friction force of the pipe as
shown in Eq. (12):
T = T1 + TF
⎧
⎪
π
TF = 4 Db2 Ra
⎨
⎪T1 = πDop Lμ1 (12)
⎩
where TF is the resistance force on reamer, T1 is the friction force of the
pipe, Ra is the front soil pressure (50–60 kPa in clay, and 80–100 kPa in
sand), μ1 is the friction force of pipe wall per unit area (0.3–0.4 kPa in
clay, and 0.5–0.7 kPa in sand).
The ASTM F 1962 model considers the friction between the pipe and
(9)
can provide an accuracy and reliable estimation of pullback force for
vary diameter PE pipes. Compared with the ASTM F 1962 model, the
PRCI model considers the bending effect by using an empirical equa-
tion, and produce a reliable but rough calculation. The GB50424 model
assumes the drillpath as a straight line by ignoring the curve section
and the friction between pipe and ground surface, and provides rough
estimation for steel pipe only. Additionally, the GB50424 model con-
siders that the maximum pullback occurs at the end of pullback process
which is a risk assumption for HDD crossing project. The GB50268 is
designed for small diameter (less than 400 mm) pipes which have
negligible mud drag force. Therefore, the accuracy of this model is the
lowest among these models.
To date numerous models for estimating pullback load for different
pipe materials and installation conditions have been proposed and used
in practice. Advancements in computer technology has resulted in
commercial software that can more accurately predict pullback loads
during the project design phase, and provided more informed sugges-
tions regarding equipment selection and construction budget.
2.2. Borehole instability
During HDD crossing projects, there are two fluid pressure limita-
tions that are a concern: (1) minimum fluid pressure to main flow and
return cuttings back to the surface, and (2) maximum fluid pressure to
maintain borehole integrity. Models for estimating minimum fluid
pressures assume laminar flow (Ariaratnam et al., 2003; Bennett and
Wallin, 2008). Eq. (13) is a model for calculating minimum fluid
pressure.
μp v τy
Pmin = ρm gH + L ⎡ + ⎤
⎢ (Db−Dop)2 (D − D ) ⎥
⎣ b op ⎦ (13)
where μp is the plastic viscosity, τy is the yield point of mud.
As the fluid velocity in the annular space is directly proportional to
pumping pressure, it is common practice to maintain higher mud ve-
locities to effectively flush out the cuttings (Shu et al., 2011, 2014; Shu
and Ma, 2015; Yan et al., 2016). However, attention must be taken to
ensure borehole integrity. Thus, researchers have studied soil failure
mechanism based on the Mohr-Coulomb criterion under maximum mud
165
X. Yan et al. Tunnelling and Underground Space Technology 72 (2018) 162–173

pressure in the borehole (Mori and Tamura, 1987; Panah and

Yanagisawa, 1989). It was discovered that borehole instability tends to
occur when shear failure is initiated in the surrounding soil.
Keulen (2001) employed the cavity expansion theory (Vesic, 1972)
to develop a new method to describe the soil behavior around the
borehole, and Andersen et al. (1994) presented this method to calculate
the maximum allowable mud pressure, which was often referred to as
the “Delft Equation” and is widely used today. The Delft equation has
three basic assumptions: (1) the soil medium is homogeneous and iso-
tropic; (2) it neglects gravity and is of infinite dimension; and (3) the
soil behaves linearly elastic in the elastic zone as determined by the
Mohr-Coulomb failure criterion. In the plastic zone, the soil behaves as
a compressible plastic solid with a constant ratio between the volume
change and the shear deformation. Then the Delft equation for calcu-
lated maximum allowable fluid pressure is shown in Eq. (14).

Pbm = [σ ′ (1 + sinφ) + c cosφ + c cotφ]

−sinφ Fig. 9. Stresses state around borehole.
2
⎡ R (σ ′sinφ + c cosφ) ⎤1 + sinφ
× ⎢ ⎜⎛ 0 ⎟⎞ + 0 ⎥ −c cotφ + u
R G
⎣ ⎝ pm ⎠ ⎦ (14) ⎧ σr = P
τrθ = 0 (springline)
where Pbm is the maximum allowable fluid pressure in the borehole, σ0′ ⎨
σ = 3σ −σ −P = (3−K ) σ −P
⎩ θ v h 0 v (16b)
is the initial effective stress, φ is the internal friction angle, c is the
cohesion, R0 is the initial radius of the borehole, Rpm is the maximum
⎧ σr = P
allowable radius of the plastic zone, G is the shear modulus, and u is the τrθ = 0 (invert)
in-situ pore pressure. ⎨
⎩ σθ = 3σh−σv −P = (3K 0−1) σv −P (16c)
Today, the Delft equation is widely used in practice for calculating
maximum allowable fluid pressures. Based on its assumptions, the Delft where σr is the radius stress, τrθ is the shear stress, is the hoop stress, σv
equation considers the soil as homogenous and in an isotropic medium, is the vertical stress of soil, σh is the horizontal stress of soil (σh = K 0 σv ),
which means that the impact of gravity is neglected and the soil exhibits P is the mud pressure in the borehole.
the same stress in different directions. However, considering the load For purely cohesive soil (clay), it is assumed that: (1) the gradient of
history and soil characteristics, the soil usually acts as anisotropic or earth pressures and mud pressure across the borehole can be neglected;
transverse isotropic medium. The coefficient of lateral earth pressure at and (2) the tensile strength of purely cohesive soil is zero. Hence, the
rest expresses the ratio between the horizontal and vertical stresses soil surrounding a borehole becomes homogenous, but anisotropic.
(Jaky, 1944; Mesri and Hayat, 1993; Michalowski, 2005). For con- They considered the maximum mud pressure would occur when the
venience, the empirical equation proposed by Jaky (1944) to calculate tensile failure initiates, and the critical equations are as follows
this coefficient in normal consolidated soils is shown as follows: (Kennedy et al., 2004a, 2004b):

K 0 = 1−sinφ (15) PTi = P0 (3K 0−1) (K 0 < 1) (17a)

where K0 is the coefficient of lateral earth pressure at rest. K0 is usually
less than 1 for normal consolidated soil (sand: 0.35–0.50, clay:
0.50–0.70), and larger than 1 for some overconsolidated soil in general
(Holtz and Kovacs, 2010). Hence, the Delft equation could over-esti-
mate the strength of normal consolidated soil, and provide a risk esti-
mation of allowable fluid pressure during HDD (Duyvestyn, 2004;
Moore, 2005; Wang and Sterling, 2007; Kennedy et al., 2004a, 2004b;
Kennedy, 2005; Kennedy et al., 2006a, 2006b; Xia and Moore, 2006,
2007; Elwood et al., 2007; Elwood, 2008; Elwood and Moore, 2009).
Additionally, soil properties and behaviors are quite different be-
tween clay and sand. The sand is purely frictional material (considered
tensile strength is 0) with large grain and high hydraulic conductivity,
and the mud can intrude and filtrate into the sand around the borehole.
Therefore, it is necessary to analyze borehole instability mechanisms of
clay and sand separately. The Delft equation attempts to predict max-
imum allowable pressure with a uniform equation for all soil mediums.
Considering the geometry of borehole path and borehole profile, the
stress status of soil surrounding the borehole can be simplified as plane
strain problem illustrated in Fig.9 (Kennedy et al., 2004a, 2004b).
Kennedy et al. (2006a, 2006b) presented an elastic σθ solution for
soil at the borehole crown, springline and invert using total cir-
cumferential stresses to determine the tensile failure mechanics as
shown in Eq. (16a)–(16c):
⎧ σr = P
τrθ = 0 (crown)
⎨ covered in Kennedy’s research with the following assumptions: (1) the
⎩ σθ = 3σh−σv−P = (3K 0−1) σv−P (16a)
PTi = P0 (3−K 0) (K 0 > 1) (17b)
where PTi is the tensile failure initiation pressure, P0 is the initial
overburden pressure (P0 = σv ).
The hydro-fracture initiated at crown when K0 < 1, and at
springline when K0 > 1 (Kennedy et al., 2006a, 2006b). In this model,
Kennedy et al. (2006a, 2006b) assumed that tensile fracture does not
occur once shear failure is initiated, and did not estimate failure for that
part of K0. However, if the shear failure initiated before the hydraulic
fracturing, the blow-out most likely control the failure mechanism (Xia
and Moore, 2006), and the furthest shear point in the plastic zone
would located above the crown when K0 < 1 (Moore, 2005). Ac-
cording to the Mohr-Coulomb failure criterion, the stress situation at
the furthest shear point should meet the following equation:
σ1−σ3 = σr −σθ = 2cu (18)
where cu is the undrained cohesion strength of soil. The shear failure
initiation pressures at the crown and springline are as follows:
(3K 0 − 1) σv
⎧ PSi = cu + 2
(crown)
⎨ PSi = cu + (3 − K 0) σv
(springline )
⎩ 2 (19)
where PSi is the shear failure initiation pressure.
Considering the shear failure in clay, Xia and Moore (2006) em-
ployed the cavity expansion theory and developed a new estimation
model as a supplementary method to estimate the values of K0 not
plastic zone was assumed to be concentric circles with the outer radius

166
X. Yan et al. Tunnelling and Underground Space Technology 72 (2018) 162–173

equal to the radius of maximum plastic range; and (2) the point dis- where σθ′ is the effective hoop stress, K 0′ is the lateral earth pressure
placement on the elastic-plastic interface was assumed to have the same coefficient (K 0′ = σh′/ σv′), σh′ is the effective horizontal earth pressure, σv′
value, dependent on the maximum plastic radius and independent of is the effective vertical earth pressure (σv′ = σv−u 0 ), u 0 is the hydrostatic
the position angle. The new, revised model can be expressed as follows: water pressure, and Δu is the excess pore pressure.
3
Lan and Moore (2016) considered the initiation of shear failure and
2
1 ⎧ R ⎡ cu + 2 P0 (K 0−1) ⎤ ⎫ developed a criterion to define the two failure mechanisms by PTi′ and
Pbm = cu + P0 (3K 0−1)−culn ⎜⎛ 0 ⎟⎞ + ⎢ ⎥ ⎬ (K 0 < 1)
2 ⎨ ⎝ Rpm ⎠ G PSi . If mud pressure in the borehole reaches PTi′ first, then tensile failure
⎩ ⎣ ⎦⎭
occurs. Conversely, if it reaches PSi first, then shear failure initiates.
(20a) These are shown in Eq. (24).
2 3
1 ⎧ R ⎡ cu + 2 P0 (1−K 0) ⎤ ⎫ ⎧ 2cu/ σv′−K 0′ > 1 (tensile failure )
Pbm = cu + P0 (3−K 0)−cu ln ⎜⎛ 0 ⎟⎞ + ⎢ ⎥ ⎬ (K 0 > 1)
2 ⎨ ⎝ Rpm ⎠ G
⎩ ⎣ ⎦⎭ ⎩ 2cu/ σv′−K 0′ ⩽ 1 (shear failure )
⎨ (24)
(20b) Lan and Moore (2016) indicated that tensile or shear failure usually
where; G is the shear modulus, Rpm is the maximum plastic radius. occurs in purely cohesive soil (such as clay), while shear failure controls
Pure sand is a cohesionless porous medium, thus drilling fluid the borehole instability in purely cohesionless porous medium (such as
during the HDD process would erode the sand surrounding a borehole. sand). Previous researchers indicated that the Delft equation was con-
Kennedy et al. (2006a, 2006b) established the analyses for sand and servative and overestimates maximum allowable fluid pressure. In ac-
sand-filtercake system by employing the elastic plate theory, plasticity tual HDD projects, the results calculated by the Delft equation and other
theory and Finite Element Method (FEM). Assumptions for sand-filter- models do not match actual job site records (Wallin et al., 2010; Neher,
cake system are: (1) the filter cake exhibits undrained response with no 2013), especially in large diameter crossing projects (Yan et al., 2016).
volume change, the filter cake acts as purely cohesive material, and the The actual soil condition on the job site is much more complex than
shear stress become tensile stress in this medium; and (2) the undrained assumed (neither homogeneous nor isotropic), and usually there are
elastic shear modulus of filter cake equals the drained shear modulus of different stratums above the borehole. Therefore, there still remains
the sand. Finite Element Method results show that, for intermediate, the more research that needs to be conducted to better model theory and
elastic plate theory calculations cannot match the elastoplastic results actual results.
because of the shear failure near the borehole changed the crown and
invert stress. Xia (2008) provided an estimating equation for sand 2.3. Prediction of fluid pressure in the borehole
shown in Eq. (21) using the parameterization theory.
H
Drilling fluid is pumped to the drill bit (or reamer) via the drill rod,
⎧ Pbm is released into the borehole through nozzles, and then flows back to
( ) ⎡ 12.9K 0
( 1.23K 0
⎪ P0 φ = ⎢17.57− K 0 − 0.084 + log K 0 + 0.022
⎪ ⎣
)( 1.23K 0
K 0 + 0.022 ) D⎤
⎥
⎦
×
the entry or exit point with cuttings through the annular space (Shu and
⎪ 0.001 Ma, 2016). Therefore, the maximum fluid pressure in the borehole
(φ − 30°
⎨⎡1.46e K0− 0.084 10°
⎪⎣
) (
+ 10° ⎤ )
40° − φ

⎦ occurs near the drill bit (or reamer), and decreases gradually along the
borehole. Since the maximum fluid pressure may result in collapse of
⎪ P = γH Pbm
⎪
⎩
bm ( )
P0 φ
(21) the borehole during the drilling process, it is important to predict the
highest pressure along the borepath.
where H is the distance from the borehole to the ground surface During drilling process, fluid in the drill rod or borehole exhibits
(Rpm = H/2 in clay and Rpm = 2H/3 in sand, suggested by Andersen two flow regimes: (1) laminar flow and (2) turbulent flow. Bourgoyne
et al., 1994; Van Brussell and Hergarden, 1997). D is the borehole et al. (1986) established three equations for three widely used flow
diameter (D = 2R0), and γ is the unit weight of the overburden soil. models (Newtonian flow, Bingham Plastic flow and Power Law flow)
The failure models previously mentioned were all based on the total when they behave under laminar flow conditions:
stresses in the soil; however, Soga and Mitchell (2005) indicated that it
is the minor principal effective stress that reaches the tensile strength of 48μ Lv
⎧ P = (D −ad)2 (Newtonian Flow )
soil and results in tensile failure. Wang and Sterling (2007) considered ⎪
⎪ 48μp Lv 6Lτy
that the mud filtrates into loose sand and increases the pore pressure, P=
(D − d)2
+ (D − d) (Bingham Plastic Flow )
⎨
which reduces the effective stresses in the soil and leads to liquefaction ⎪ 4v (2n + 1) n 4KL
and borehole instability. They focused on the importance of filter cake ⎪ P = ( (D − d) · n ) · (D − d) (Power Law Flow ) (25)
⎩
for reducing mud filtration. Two numerical models were developed to
where P is the mud pressure loss, μa is the apparent viscosity of drilling
support their opinions: (1) with filter cake surrounding the borehole;
mud, L is the flow length of drilling mud, v is the average velocity of
and (2) without filter cake. Simulation results indicated that the
drilling mud, D is the borehole diameter, d is the diameter of drill rod,
minimum fluid pressure supports the soil thus avoiding collapsing, the
μp is the plastic viscosity, τy is the yield point.
filter cake “holds” loose soil grains, and the maximum fluid pressure
If the fluid exhibits turbulent flow, only Eq. (26) is used to calculate
equals the soil earth pressure surrounding the borehole.
the pressure loss for different flow models (Bourgoyne et al., 1986):
Lan and Moore (2016) emphasized the importance of pore pressure
in borehole instability when crossing in saturated clay and used effec- 2fρLv 2
tive stresses to establish a new solution for soil hoop stresses at the P=
D−d (26)
crown and spring line shown in Eq. (22):
where f is the friction factor.
⎧ σθ′ = σθ−Δu−u 0 = 3σh−σv−P−(σh−σv )−u 0 = 2σh−u 0−P (crown) Because of the high ratio of Dborehole/Drod in HDD crossing projects,
⎩ σθ′ = σθ−Δu−u 0 = 3σv−σh−P−(σv−σh)−u 0 = 2σv−u 0−P (springline )
⎨ drilling fluid moves as laminar flow in the borehole, and the rotation of
drill rod has minimal impact on the fluid flow behavior. Subsequently,
(22)
pressure loss in the borehole can be calculated, which is also the
Eqs. (23a) and (23b) show the initiation of tensile failure in soil. maximum mud pressure around the drill bit (or reamer) (Shu and Ma,
PTi′ = 2σh−u 0 (K 0′ < 1) (23a) 2016). However, the borehole profile is often not a concentric circular
shape and the roughness of borehole wall is difficult to determine. Yan
PTi′ = 2σv−u 0 (K 0′ > 1) (23b) et al. (2016) established a new method to calculate the maximum

167
X. Yan et al. Tunnelling and Underground Space Technology 72 (2018) 162–173

pressure in the borehole considering that the pumping pressure bal-

ances the circulating pressure and keeps the mud flowing, which can be
measured on the ground, and that the circulating pressure is the sum of
pressure losses in the drill rod through the bit nozzles and in the an-
nulus.

⎧ Pb = Pp−Pdr −Pbn
⎪ fρQ2
Pdr = 2 5 L
π R1
⎨
⎪ Pbn = ρ Q 2
⎩
( )
2 CA
where Pb is the mud pressure in annulus, Pp is the real pumping pres-
sure, Pdr is the pressure loss in the drill rod (about 20–45% of the cir-
culating pressure), Pbn is the pressure loss through the bit nozzles
(Schuh, 1964), ρ is the unit weight of the fluid, Q is the volumetric flow
rate, R1 is the inner radius of the drill rod, L is the length of drill rod, f is
the Fanning friction factor, and it is mainly depended on fluid proper-
ties, flow regimes, Reynolds numbers and boundary conditions
(Appendix A), C is the nozzle constant (taken as 0.95 in general), A is
the total area of the nozzles.
Calculation of pressure losses in the drill string is accurate due to the
fact that the drill rod and bit nozzles have circular cross-sections.
Therefore, the method to predict maximum fluid pressure in the bore-
hole by pumping pressure and pressure loss in the drill string provides a
more accurate estimation than other approaches.
3. New innovations in HDD
3.1. Reverse circulation
During the middle of last century, reverse circulation was used in
pile foundation engineering in Germany and The Netherlands. Based on
fundamental research, reverse circulation was developed into pump
suction reverse circulation, jet flow reverse circulation and airlift re-
verse circulation, and was soon employed by geological and oil drilling
(Bhutada and Pangarkar, 1987; Morrison et al., 1987; Yasashin et al.,
1996; Kumar and SPE, 2005; Li et al., 2010; Zhu et al., 2015). Liu and
Xia (2009) used the pump suction reverse circulation in an HDD rock
crossing project. The pump rate required by reverse circulation was
found to be much lower than positive circulation, while the cuttings
transportation performance (4132 m3/h) was several times higher than
positive circulation. Wei (2011) developed the pump suction reverse
circulation for HDD and provided the basic operation parameters for
application by contractors. Kong and Ma (2015) studied the concept of
jet flow reverse circulation and designed a special backreamer for this
new technology as illustrated in Fig. 10. Though reverse circulation was
introduced to the HDD industry to address the problem of effectively
transporting cuttings during large diameter crossing projects, it can be
used in both large and small diameter crossings applications. Currently,
this new technology is starting to gain limited accepted in the HDD
industry; however, there remains numerous unknown mechanics of
reverse circulation requiring further detailed theoretical analysis and
experiments.
3.2. Push hole-opening
Push Hole-Opening in HDD is most often used when drilling into
landfalls or sea outfalls where the exit point is inaccessible. Different
from traditional backreaming, the Push Hole-Opening method pushes a
hole-opener from the entry point to the exit point several times until the
final diameter requirements of borehole as achieved as illustrated in
Fig.11. This is a high-risk method in sand, silt and clay conditions and is
best suited through rock; however, the likelihood of success decreases
with increased length, borehole diameter, and unconfined compressive
strength of rock. Similar to traditional oil field drilling, this method
may require additional in-string stabilizers to avoid drill string
(27)
Fig. 10. Reverse circulation.
instability when drilling larger diameter boreholes. Strength is the key
parameter in designing the geometry of the hole-opener and it is dic-
tated by the strength of its cutters and body. Therefore, in order to
minimize risk, it is recommended to drill a large diameter pilot hole
(305 mm in soft rock and 406 mm in hard rock). LMR Drilling utilized
the Push Hole-Opening technique to successfully install the Sines to
Setubal natural gas pipeline owned by Transgas that included four
sections as presented in Table 1.
3.3. Direct Pipe®
At first, the thruster was invented as an auxiliary equipment to assist
HDD contractors in installing pipes during the pullback process. A few
years ago, Herrenknecht combined the thruster with microtunneling to
develop a novel pipeline installation technique called Direct Pipe®
(Fig.12). In 2007, Direct Pipe® was first used on the Rhine River
crossing project in Germany. Today, it has been used on dozens of
projects worldwide. Direct Pipe® was developed as an optional method
to drill through soils where traditional HDD is not applicable or risky
(such as weak soil, heterogeneous formation and hard rock). The
method is a one-pass installation using microtunneling and a thrusting
unit to simultaneously install a pipeline into the excavated borehole.
Prior to installation, the entry location should be prepared to place the
thrusting unit. During installation, the soil is excavated from the tunnel
face in the microtunneling machine and transported back to the ground
via slurry circulation lines inside the pipe. The thrusting unit provides
the main jacking force and pushes the pipe with machine along the
designed borepath. A special clamping unit is designed to protect the
pipe coating during the pushing operation. Direct Pipe has a number of
benefit as it requires lower pump volume and pressure thereby reducing
the risk of fluid loss, a smaller working space, and a smaller annular
space (less than 100 mm).
3.4. New pipe material
Steel and HDPE pipe have been used for HDD crossing projects for
decades. The introduction of retrained joint ductile iron pipe (DIP) and
fusible polyvinylchloride pipe (FPVCP™) have provided owners and
designers with feasible alternative material options. DIP requires a
smaller curvature radius and site space than other pipes due to the
ability to cartridge load the drill string rather than string out the entire
fused or welded pipe string (Ariaratnam et al., 2008a, 2008b).

168
X. Yan et al. Tunnelling and Underground Space Technology 72 (2018) 162–173

Fig. 11. Push Hole-Opening method.

Table 1 Table 2
Sines to Setubal natural gas pipeline project information. Sampling of ductile iron pipe projects installed by HDD.

Section Site condition Length (m) Final Borehole Diameter (mm) Location Diameter (mm) Length (m)

1 Land to land 650 1120 San Marcos, Texas USA 762 369
2 Landfall 1400 1222 Valencia, Spain 900 456
3 Sea to sea 1300 1120 Moscow, Russia 300 700
4 Landfall 1250 1222 Berlin, Germany 250 600
Meridian, Mississippi USA 900 500
Belarus 400 240
Applications of DIP have increased due to comfort levels of owners with Lianyuan, China 600 192
Pasco County, Florida USA 900 530
iron pipe in the their systems and ease of assembling (Rowell, 2003).
Xiangyang, China 500 204
The diameters of DIP installed by HDD range from 100 mm to 900 mm
with achieved crossing distances over 500 m as shown in Table 2
(Ariaratnam and Carpenter, 2004; Carpenter et al., 2005; Ehrin and of no increase in profile at the fusion joint, a smaller diameter borehole
Carpenter, 2006; Pontesso and carpenter, 2007; Dorwart and Carpenter, can be achieved, resulting in a smaller diameter backream and use of
2010; Cherkashin, 2016). There are numerous DIP joints that can less drilling fluid. The high tensile strength of joints leads to a higher
produce uniform stresses in the pipe body (Cherkashin, 2016). Subse- allowable pullback load, which is determined by the minimum tensile
quently, joint strength is the key parameter in determining the max- stress of PVC (48.26 MPa) and a safety factor of 2.5 (Ariaratnam et al.,
imum allowable pullback load of DIP; however, there currently exists 2008a, 2008b, 2010). Ariaratnam et al. (2010) developed a model to
no theoretical solution for different joints to accurately estimate pull- predict pullback loads for FPVC™ based on ASTM F1962-05. Predicted
back loads. values were compared to actual field results. It maintains properties of
Different from the classic PVC pipe, which is connected by bell or PVC pipe such as corrosion resistance, high resistance to chemicals,
spigot joints increasing the diameter at the connection parts compared high Hydrostatic Design Basis (HDB) and tensile strength, and no ma-
to the pipe body, FPVC™ provides restrained joints without a diameter terial relaxation time. Compared to HDPE, FPVC™ requires a larger
change and the same tensile capacity of the pipe post-fusion. As a result bend radius (250 times of the outer diameter) (Najafi, 2005; Lim et al.,

Fig. 12. Direct pipe.

169
X. Yan et al. Tunnelling and Underground Space Technology 72 (2018) 162–173

Table 3 Table 4
Sampling of FPVC™ pipe projects installed by HDD. Comparison of recorded and estimated pullback loads (Wu et al., 2014).

Location Diameter (mm) Length (m) Pipe Diameter Estimated pullback load (t, calculated Recorded pullback load
by GB50424-2007) (t)
Beaufort, South Carolina 900 396
Gloucester, Massachusetts 500 690 406 mm-C 160 < 60
Ambridge, Pennsylvania 500 1524 711 mm-B > 300 < 180
St. Petersburg, Florida 762 548 711 mm-A > 300 < 160
Wakpala, South Dakota 600 1433
Macon, Georgia 600 1311
Hillsborough, Florida 400 2133 installations have been widely utilized as an alternative to traditional
open-cut methods in Mainland China (Ma and Najafi, 2008). There
continues to be challenging, record-setting river crossings using HDD in
China.

4.1. Jiaozuo irrigation pipeline Qin River HDD crossing project

In 2013, the local government invested to build a new irrigation

Fig. 13. Drilling operation and reamers used in the Qin River crossing project.
2010). To date, the diameter range of FPVC™ installed by HDD ranges
from 100 mm to 910 mm and maximum installation length of 2133 m
(see Table 3).
pipeline to alleviate the water shortage in Jiaozuo, Henan Province of
China. The Qin River crossing project is the key section of a new pi-
peline project that includes two parallel crossings; each with 1016 mm
diameter steel pipe and drive lengths of 1750 m, which are the longest
large diameter HDD projects completed with single crossing tech-
nology.
There was concern about soil failure and mud loss due to the main
crossing area being in a weak clay formation. The drill path was de-
signed with a 7° entry angle, 8° exit angle, 2290 m curvature radius,
33.5 m maximum depth and 29.78 m minimum depth under the riv-
erbed. When drilling the pilot hole, in order to reduce the total reaming
process, a pre-reaming method was used that involved a 325 mm dia-
meter tri-cone bit followed by a 600 mm diameter reamer. Guidance
was provided by TruTracker technology.
Only one drill rig was used and there was no auxiliary drill rod
connected behind the reamer during backreaming (Fig. 13). Upon
completion of one reaming stage, the drill rig pushed the reamer and
drill rods back to the exit point to prepare for the next stage of the
reaming process. The drill maintained rotation during push back to
achieve better guidance. In order to push the reamers smoothly and
obtain a circular borehole that is more stable and better suited for the
pullback process, a series of lightweight reamers (900 mm, 1300 mm,
and 1500 mm) were designed. Each reamer has a large hollow body
that makes it travel concentric with the borehole during reaming and
pushing (Fig. 13). In addition, this project utilized the largest and most
powerful HDD drill rig currently on the market with 1000 t of pulling
force capacity. The contractor used the same drill rig to complete a
1800 mm diameter heat pipe river crossing project with a drive length
of 962 m, making it the largest diameter pipe installed to date by HDD
technology.
4.2. Jiangdu-Rudong natural gas pipeline Yangtze River HDD crossing
project

4. HDD case studies In 2014, China Petroleum Pipeline Bureau completed a challenging
3300 m Yangtze River HDD crossing project for installing oil and nat-
With the development of HDD technology, large diameter ural gas pipelines. Three parallel pipelines were installed includes two

Fig. 14. The Rotary magnet orientation intersect tech-

nology.

170
X. Yan et al. Tunnelling and Underground Space Technology 72 (2018) 162–173

Table 5
Effective viscosity in drill rod of different flow models.

Flow model Equation

Newtonian μed = μa
Bingham Plastic μed = μp
Power Law μed = K (
3n + 1 n 4Q n − 1
) ( 3)
4n πR1
Hersehel Bulkley μed = μy + K (
3n + 1 n 4Q n − 1
) ( 3)
4n πR1

Where; μa is the apparent viscosity, μp is the plastic viscosity, μy is the yield

viscosity (see Table 6).

Table 6
Effective Viscosity in Drill Rod of Different Flow Models.

Flow model Reynolds Regime Fanning friction

Newtonian < 2100 Laminar f=

16
NRe
> 3000 Turbulent Fig.16**
Non-Newtonian Fig.15* Laminar f=
16
NRe
Fig.15* Turbulent Fig.15**

* The critical Reynolds number of Non-Newtonian fluid depend on the flow behavior index n (Fig. 15, Dodge and Metzner, 1959).
** For turbulent flow, the fanning friction factor of Newtonian fluid can be found in Fig. 16 (Bobo, 1958), and of Non-Newtonian Fluid can be
found in Fig. 15.

Fig. 15. Relationship about f, NRe and n for Non-

Newtonian fluid.

Fig. 16. Relationship between Fanning Friction Factor

and Reynold’s Number for Newtonian fluids.

171
X. Yan et al.
711 mm natural gas pipelines and one 400 mm product oil pipeline
crossing the Yangtze River from Jingjiang to Jiangyin. It is the key
project of the Jiangdu-Rudong natural gas pipeline, and the longest
large diameter pipeline crossing project in the world to date. The main
crossing area consisted of clay, silt, and fine sand with a small amount
of gravels (diameter ranges from 5 mm to 70 mm).
Due to the long length of the designed crossing, the project could
not be completed using a traditional single crossing HDD technology.
Additionally, the main crossing area was cohesive soil, which would
result in extremely high reaming torque (the estimated maximum
torque was over 40,000 N·m). The pressure loss in the borehole would
be much higher than normal project, and critical challenges of cutting
transportation and borehole stability were concerns. The pullback load
was projected to increase linearly with each increment of drill length.
The advanced rotary magnet orientation intersects (RMOI) tech-
nology was employed in the project to deal with the long-distance pilot
hole drilling issue as shown in Fig. 14. A considerable number of
technical difficulties (such as the intersection accuracy, type of the
external guiding magnetic field, and pushing process after intersection)
needed to be resolved because the RMOI technology was being used for
the first time on a large river crossing project. A submarine artificial
magnetic field was successfully used to provide a stable, powerful and
accuracy guidance signal source for the RMOI. When intersection was
accomplished, the connection part was pushed out to the ground. Two
types of high-strength drill rods (193.7 mm S-135 with maximum pull
and push strength of 90 t, 168.3 mm V-150 with maximum torsional
strength over 70,000 N·m) were designed to satisfy the pilot hole and
backreaming demands. When pulling back, a water ditch and lubrica-
tion fluids were used to reduce the resistance forces, which were much
lower than the estimated pullback loads as presented in Table 4.
5. Conclusions
In this state-of-the-art review of Horizontal Directional Drilling
(HDD), four common estimation models for calculating pullback loads,
Appendix A
The Reynolds number is used to determine the flow regime (laminar flow / turbulent flow) of mud, and it can be calculated by the following
equation:
2Qρ ⎧ NRe < NRe − crit Laminar Flow Regime
NRe = ,
⎩ NRe > NRe − crit Turbulent Flow Regine
πR1 μed ⎨
where NRe-crit is the corresponding critical Reynolds number, NRe is the Reynolds numbers, μed is the effective viscosity in the drill rod. For different
fluid flow type, the effective viscosity can be calculated using equations shown in Table 5 (Ariaratnam et al., 2007):
References
Andersen, K.H., Rawlings, C.G., Lunne, T.A., By, T.H., 1994. Estimation of hydraulic
fracture pressure in clay. Can. Geotech. J. 31 (6), 817–828.
Ariaratnam, S.T., Lueke, J.S., 2002, April. Post construction evaluation of the annular
space region in horizontal directional drilling. In Proc NASTT No-Dig2002 Conf,
Montreal, Quebec. North American Society for Trenchless Technology, Arlington, VA.
Ariaratnam, S.T., Stauber, R.M., Bell, J., Harbin, B., Canon, F., 2003. Predicting and
controlling hydraulic fracturing during horizontal directional drilling. In: New
Pipeline Technologies, Security, and Safety, pp. 1334–1345.
Ariaratnam and Carpenter, 2004. Field Evaluation and Testing of Ductile Iron Pipe for
Trenchless Applications in Municipal Engineering, No-Dig 2004. New Orleans LA.
Ariaratnam, S.T., Harbin, B.C., Stauber, R.L., 2007. Modeling of annular fluid pressures in
horizontal boring. Tunn. Undergr. Space Technol. 22 (5), 610–619.
Ariaratnam, S.T., Bonds, R., Crabtree, D., 2008. A Field Study of HDD Installed Ductile
Iron Pipe. In: Pipelines 2008: Pipeline Asset Management: Maximizing Performance
of our Pipeline Infrastructure, pp. 1–10.
Ariaratnam, S.T., Marti, T., Botteicher, R.B., Finnsson, S., Ghosh, A., 2008. Determination
of safe pulling loads on fusible pvc pipe during horizontal directional drilling in-
stallations. The North American Society (NASTT) and the International Society for
Trenchless Technology (ISTT) International No-Dig Show 2008.
Ariaratnam, S.T., Botteicher, R.B., Ghosh, A., Piratla, K., Vilfrant, E., 2010. Predictive
modeling of installation loads for directional drilled fusible PVC pipe. Tunn. Undergr.
Space Technol. 25 (6), 766–772.
172
Tunnelling and Underground Space Technology 72 (2018) 162–173
borehole stability equations, and two calculation methods for drilling
fluid pressure in the borehole were discussed. Additionally, new in-
novations in HDD applications were described. HDD has been proven
itself to be environmentally friendly with high productivity compared
to traditional open-cut methods.
Pullback load composed of the resistance force between the pipe
and ground surface and borehole wall, with additional resistance force
at the curves along the borepath and drag force of the drilling fluid.
Based on these components, several estimation models were established
with different assumptions. To date, with the help of computer soft-
ware, pullback load can be predicted fairly accurately. Borehole stabi-
lity is a big challenge for the HDD industry. The goal of any borehole
stability study is to obtain the maximum allowable fluid pressure in the
borehole. The Delft equation was established and is widely used as a
closed form model based on elastic theory and several assumptions.
With the consideration of soil anisotropic, other models were presented
that can provide more accurate results in theory. However, there are
still some unknown factors in estimation of borehole stability requiring
further research.
Reverse Circulation, Push Hole-Opening, and Direct Pipe are new
technologies developed to expand the application of HDD. The main
advantages of Reverse Circulation are its high cuttings transportation
performance and low mud requirement. The Push Hole-Opening makes
possible pipeline installations at landfall or outfall, where the exit point
is inaccessible. Direct Pipe greatly overcomes the limitation of complex
soil conditions and minimizes loss of drilling fluid. New pipe materials
presented provide designers and owners will options. Both restrained
joint ductile iron and fusible PVC have a proven record of installation
using HDD.
Changes are inevitable in the HDD industry. Further enhancements
will occur as technological advancement continue to drive designers to
push the envelope in terms of length and diameters of pipeline in-
stallations. Research efforts will continue, as new models are developed
and existing models are modified to provide more accuracy between
estimated and calculated results.
(28)
ASTM, F., 1962–99. Standard Guide for Use of Maxi-Horizontal Directional Drilling for
Placement of Polyethylene Pipe or Conduit Under Obstacles, Including River
Crossings. ASTM F.
ASTM, F., 1962–05. Standard Guide for Use of Maxi-Horizontal Directional Drilling for
Placement of Polyethylene Pipe or Conduit Under Obstacles, Including River
Crossings. ASTM F.
Baumert, M.E., Allouche, E.N., 2002. Methods for estimating pipe pullback loads for
horizontal directional drilling (HDD) crossings. J. Infrastruct. Syst. 8 (1), 12–19.
Baumert, M.E., Allouche, E.N., Moore, I.D., 2004. Experimental investigation of pull loads
and borehole pressures during horizontal directional drilling installations. Can.
Geotech. J. 41 (4), 672–685.
Baumert, M.E., Allouche, E.N., Moore, I.D., 2005. Drilling fluid considerations in design
of engineered horizontal directional drilling installations. Int. J. Geomech. 5 (4),
339–349.
Bennett, D., Wallin, K., 2008. Step by step evaluation of hydrofracture risks for horizontal
directional drilling projects. In: Pipelines 2008: Pipeline Asset Management:
Maximizing Performance of our Pipeline Infrastructure, pp. 1–10.
Bobo, R.A., 1958. Keys to Successful competitive drilling. Gulf Publishing Company.
Bourgoyne, A.T., Millheim, K.K., Chenevert, M.E., Young, F.S., 1986. Applied drilling
engineering. Society of Petroleum Engineers (SPE), Richardson, TX.
Bhutada, S.R., Pangarkar, V.G., 1987. Gas induction and hold-up characteristics of liquid
jet loop reactors. Chem. Eng. Commun. 61 (1–6), 239–258.
Carpenter, R., Schwarzlose, R., Whitaker, K., 2005. Horizontal Directional Drilling Using
Ductile Iron Pipe—A Successful Paradigm Shift. In: Pipelines 2005: Optimizing
Pipeline Design, Operations, and Maintenance in Today's Economy, pp. 16–25.
X. Yan et al.
Chehab, A.G., 2008. Time Dependent Response of Pulled-in-Place HDPE Pipes. Queen’s
University, Kingston, Ontario, Canada.
Cheng, E., Polak, M.A., 2007. Theoretical model for calculating pulling loads for pipes in
horizontal directional drilling. Tunn. Undergr. Space Technol. 22 (5), 633–643.
Cherkashin, S., 2016. Installation of the Pipelines Made of Ductile Iron (DI) With the
usage of Horizontal-directional Drilling Technique (HDD) For Water Supply
Treatment Service and Sewerage Pipelines Construction and Reconstruction. Proc.
Eng. 165, 717–725.
Chevron Chemical Co., LLC., 1999. Technical Note: Horizontal Directional Drilling
(Guided Boring) with PLEXCO Pipe, pp. 1–39.
Chin, W.C., 2001. Computational Rheology for Pipeline and Annular Flow: Non-
Newtonian Flow Modeling for Drilling and Production, and Flow Assurance Methods
in Subsea Pipeline Design. Gulf Professional Publishing.
Dareing, D.W., Ahlers, C.A., 1991. Tubular bending and pull-out forces in high-curvature
well. J. Energy Res. Technol. 113, 133.
Dodge, D.W., Metzner, A.B., 1959. Turbulent flow of non-Newtonian systems. AIChE J. 5
(2), 189–204.
Dorwart, B., Carpenter, R., 2010. HDD Crossing of Lake Austin Generates Data and New
Model for Calculating Pull Force for Ductile Iron Pipe. In: Pipelines 2010: Climbing
New Peaks to Infrastructure Reliability: Renew, Rehab, and Reinvest, pp. 1205–1216.
Drillpath, 1996. Drillpath Theory and User’s Manual. Infrasoft L.L.C, Houston.
Driscopipe, 1993. Technical expertise: Application of driscopipe© in directional-drilling
and river-crossings. Technical Note #41, pp. 1–40.
Duyvestyn, G.M., 2004. Field and numerical investigation into pipe bursting and hor-
izontal directional drilling pipeline installation ground movements. University of
Waterloo.
Ehrin, J., Carpenter, R., 2006. Horizontal Directional Drilling Utilizing 36-Inch Flexible
Restrained Joint Ductile Iron Pipe—A Successful First Experience. In: Pipelines 2006:
Service to the Owner, pp. 1–8.
Elwood, D., Moore, I., Xia, H., 2007. Hydraulic fracture experiments in a frictional ma-
terial and approximations for maximum allowable mud pressure. Ottawa Geo 2007,
1681–1688.
Elwood, D.E., 2008. Hydraulic Fracture Experiments in a Frictional Material and
Approximations for Maximum Allowable Mud Pressure. Queen’s University,
Kingston, Ontario, Canada.
Elwood, D., Moore, I., 2009. Hydraulic fracturing experiments in sand and gravel and
approximations for maximum allowable mud pressure. The North American Society
(NASTT) and the International Society for Trechless Technology (ISTT) International
No-Dig Show, 2009.
El Chazli, G., 2005. Experimental Investigation of Friction Factors in Horizontal
Directional Drilling Installations. University of Western Ontario, London, Ontario,
Canada.
GB50424, 2015. Code for construction of oil and gas transmission pipeline crossing en-
gineering.
GB50268, 2008. Code for construction and acceptance of water and sewerage pipeline
works.
Haciislamoglu, M., Langlinais, J., 1990. Non-Newtonian flow in eccentric annuli. Trans.
ASME J. Energy Res. Technol. 112 (3), 163–169.
Hair, J.D., 2002. Coating requirements for pipelines installed by horizontal directional
drilling. In: Pipelines 2002: Beneath Our Feet: Challenges and Solutions, pp. 1–8.
Holtz, R.D., Kovacs, W.D., Sheahan T.C., 2010. An Introduction to Geotechnical
Engineering, second ed.
Huey, D.P., Hair, J.D., McLeod, K.B., 1996. Installation loading and stress analysis in-
volved with pipelines installed by horizontal directional drilling (No. CONF-
9603151–). North American Society for Trenchless Technology, Chicago, IL (United
States).
Jaky, J., 1944. The coefficient of earth pressure at rest. J. Soc. Hungarian Archit. Eng. 78
(22), 355–358.
Kennedy, M.J., Skinner, G.D., Moore, I.D., 2004. Elastic calculations of limiting mud
pressures to control hydrofracturing during HDD. Proc., No-Dig 2004.
Kennedy, M.J., 2005. Finite Element Calculations of Hydraulic Fracturing During
Horizontal Directional Drilling. Queen’s University, Kingston, Ontario, Canada.
Kennedy, M.J., Skinner, G.D., Moore, I.D., 2004. Limiting mud pressures to control hy-
drofracturing during HDD in an elastoplastic soil. In: Proc., 2004 Canadian
Geotechnical Society Conf.
Kennedy, M., Skinner, G., Moore, I., 2006a. Limiting slurry pressures to control hydraulic
fracturing in directional drilling operations in purely cohesive soil. Transp. Res. Rec.:
J. Transp. Res. Board 1976, 172–180.
Kennedy, M.J., Moore, I.D., Skinner, G.D., 2006b. Development of tensile hoop stress
during horizontal directional drilling through sand. Int. J. Geomech. 6 (5), 367–373.
Keulen, B., 2001. Maximum Allowable Pressures During Horizontal Directional Drillings
Focused on Sand. Delft Univ. of Technology, Delft, The Netherlands.
Kong, Y., Ma, B., 2015. Inside jet flow method for cuttings removal in HDD. China Patent,
CN 103527092 B.
Kumar, P.S., SPE, B., 2005. Coiled tubing reverse circulation-an efficient method of
cleaning horizontal wells in a mature pressure depleted field. SPE 92804, 5–7.
Lasheen, A., Polak, M.A. (2001). HDD loads imposed on a pipe. In: Proceedings of North
American No-Dig.
Lan, H., Moore, I.D., 2016. Practical Criteria for Borehole Instability in Saturated Clay
during Horizontal Directional Drilling. The North American Society (NASTT) and the
International Society for Trechless Technology (ISTT) International No-Dig Show
2016, Paper TM2-T2-01.
Li, J., Aitken, W.A., Jumawid, F., Alingig, G., 2010. Cleaning Horizontal Wellbores
Efficiently With Reverse Circulation Combining With Wiper Trip for Coiled-Tubing
Annulus Fracturing Application. In SPE/ICoTA Coiled Tubing and Well Intervention
Conference and Exhibition. Society of Petroleum Engineers.
Liu, Q., Xia, B.R., 2009. Research on reverse circulation technology of drill dregs removal
173
Tunnelling and Underground Space Technology 72 (2018) 162–173
in trenchless technology. Explor. Eng. (Rock and Soil Drilling and Tunneling) 2, 019.
Lim, L., Mainali, A., Rodrigue, L., Sonnenberg, R., Boring, E.D., Nisku, A.B., 2010. Cold
weather installation of fusible pvc using horizontal directional drilling. The North
American Society (NASTT) and the International Society for Trechless Technology
(ISTT) International No-Dig Show 2010. Paper D-1-04.
Ma, B., Najafi, M., 2008. Development and applications of trenchless technology in China.
Tunn. Undergr. Space Technol. 23 (4), 476–480.
Maidla, E.E., Wojtanowicz, A.K., 1987. Field Method of Assessing Borehole Friction for
Directional Well Casing. SPE Paper No. 15696. SPE Middle East Oil Show, Bahrain
85–96.
Mesri, G., Hayat, T.M., 1993. The coefficient of earth pressure at rest. Can. Geotech. J. 30
(4), 647–666.
Michalowski, R.L., 2005. Coefficient of earth pressure at rest. J. Geotech. Geoenviron.
Eng. 131 (11), 1429–1433.
Moore, I.D., 2005. Analysis of ground fracture due to excessive mud pressure. In: Proc.,
2005 North American No-Dig Conf., Orlando, C-1-03.
Morrison, G.L., Zeineddine, T.I., Henriksen, M., Tatterson, G.B., 1987. Experimental
analysis of the mechanics of reverse circulation air lift pump. Ind. Eng. Chem. Res. 26
(2), 387–391.
Mori, A., Tamura, M., 1987. Hydrofracturing pressure of cohesive soils. Soils Found. 27
(1), 14–22.
Najafi, M., 2005. Trenchless Technology: Pipeline and Utility Design, Construction, and
Renewal. McGraw Hill Professional.
Neher, M., 2013. HDD Hydrofracture: Roots of the Model We Use (But Don’t Understand).
The North American Society (NASTT) and the International Society for Trechless
Technology (ISTT) International No-Dig Show 2013. WM-T2-04.
Panah, A.K., Yanagisawa, E., 1989. Laboratory studies on hydraulic fracturing criteria in
soil. Soils Found. 29 (4), 14–22.
Polak, M.A., Lasheen, A., 2001. Mechanical modelling for pipes in horizontal directional
drilling. Tunn. Undergr. Space Technol. 16, 47–55.
Polak, M.A., Chu, D., 2005. Pulling loads for polyethylene pipes in horizontal directional
drilling: theoretical modeling and parametric study. J. Infrastruct. Syst. 11 (2),
142–150.
Pontesso, K., Carpenter, R., 2007. Twin 30-Inch Ductile Iron Pipe HDD Crossings of the
Historic San Marcos River. In: Pipelines 2007: Advances and Experiences with
Trenchless Pipeline Projects, pp. 1–11.
Puckett, JS., 2003. Analysis of theoretical versus actual HDD pulling loads In New
Pipeline Technologies. Security Safety 1346–1352.
Rabiei, M., Yi, Y., Bayat, A., Cheng, R., 2016. General method for pullback force esti-
mation for polyethylene pipes in horizontal directional drilling. J. Pipeline Syst. Eng.
Practice 7 (3), 04016004.
Rowell, RW., 2003. Using ductile iron pipe for horizontal directional drilling installations.
New Pipeline Technol. Security Safety 1505–1508.
Shu, B., Ma, B., Zeng, C., Lan, H., 2011. Preparation and recycling of drilling mud in large
scale Horizontal Directional Drilling projects in Mainland China. In: ICPTT 2011:
Sustainable Solutions For Water, Sewer, Gas, And Oil Pipelines, pp. 2049–2059.
Shu, B., Ma, B., Lan, H., 2014. Cuttings transport mechanism in a large-diameter HDD
borehole. J. Pipeline Syst. Eng. Practice 6 (4), 04014017.
Shu, B., Ma, B., 2015. Study of ground collapse induced by large-diameter horizontal
directional drilling in a sand layer using numerical modeling. Can. Geotech. J. 52
(10), 1562–1574.
Shu, B., Ma, B., 2016. The return of drilling fluid in large diameter horizontal directional
drilling boreholes. Tunn. Undergr. Space Technol. 52, 1–11.
Soga, K., Mitchell, J.K., 2005. Fundamentals of Soil Behavior. John Wileyand Sons,
Hoboken, New Jersey, USA.
Schuh, F.J., 1964. Computer makes surge-pressure calculations useful. Oil Gas J. 31, 96.
Van Brussell, G., Hergarden, H., 1997. Installation of pipelines beneath levees using
horizontal directional drilling. Delft Geotechnics SO-59407-701/2. 1–16.
Vesic, A.S., 1972. Expansion of cavities in infinite soil mass. J. Soil Mech. Found. Div. 98
(sm3).
Wang, X., Sterling, R.L., 2007. Stability analysis of a borehole wall during horizontal
directional drilling. Tunn. Undergr. Space Technol. 22 (5), 620–632.
Wallin, K., Wallin, M., Bennett, D., 2010. HDD crossing under environmentally sensitive
slough: mitigation of hydrofracture risk. The North American Society (NASTT) and
the Internationa Society for Trechless Technology (ISTT) International No-Dig Show
2010. Paper A-1-02.
Wei, L., 2011. The Research of Horizontal Directional Drilling Reverse Circulation
Technology. China University of Geosciences, Wuhan, China.
Wu, Yiquan, Lu, Guiming, Xu, Liangkui, Zhang, Panjun, Mao, Naibing, 2014. Technical
difficulties and innovation in the Jiangyin Yangtze River crossing project of 3300 m
HDD. Nat. Gas Industry B 1 (1), 119–124.
Xia, H.W., Moore, I.D., 2006. Estimation of maximum mud pressure in purely cohesive
material during directional drilling. Geomech. Geoeng.: Int. J. 1 (1), 3–11.
Xia, H., Moore, I.D., 2007. Discussion of limiting mud pressure during directional drilling
in clays. Ottawa GEO 2007, 1696–1701.
Xia, H.W., 2008. Investigation of Maximum Allowable Mud Pressure within Sand and
Clay during Horizontal Directional Drilling. Queen’s University, Kingston, Ontario,
Canada.
Yasashin, V.A., Makarov, N.G., Nazarov, A.M., Serikov, D.Y., 1996. New design of drilling
bit with centrifugally through-hardened cutter structure, for jetturbine drilling.
Chem. Pet. Eng. 32 (2), 115–118.
Yan, X., Ma, B., Zeng, C., Liu, Y., 2016. Analysis of formation fracturing for the Maxi-HDD
Qin River crossing project in China. Tunn. Undergr. Space Technol. 53, 1–12.
Zhu, L.H., Huang, Y., Wang, R.H., Wang, J.Y., 2015. A mathematical model of the motion
of cutting particles in reverse circulation air drilling. Appl. Math. Comput. 256,
192–202.