Investigations of pipe installation using directional drilling

Moore, I.D.1, Allouche, E.N.2, Baumert, M.2

1
GeoEngineering Centre at Queen’s – RMC, Queen’s University, Kingston, Canada
2
Trenchless Technology Center, Louisiana Tech, LA, USA

Abstract

A concerted research effort has been underway since 2002 to investigate the performance of steel,
HDPE, and PVC pipes installed using directional drilling. In particular, the axial forces needed to pull
the pipe into place, the magnitude and impact of mud pressures during drilling operations, and the
performance of pipe under those loads has been studied. This paper summarizes the outcome and
progress of various graduate research projects supported by a collaborative research grant.

First, load cells were developed to measure axial pulling forces. Measurements of pulling force for
HDPE, PVC and steel pipes have been used to examine existing design equations and propose a new
approach for estimating pulling force.

Second, mud pressures have been measured in the field, and factors influencing those mud pressures
evaluated. The ‘resistance’ side of the mud pressure problem has also been investigated, with design
equations developed to quantify the mud pressures that lead to tensile fracture of the soil around the
borehole, and others to quantify the ground resistance to ‘blowout’ (mud release following mobilization
of the shear strength of the overlying soil).

Finally, the time dependent response of HDPE is being measured and modeled so that long term axial
stresses can be estimated over the service life of the installed pipe.

Introduction to strategic research project

Horizontal Directional Drilling (HDD) is now widely used in practice internationally. The propagation of
equipment and contractor expertise in the 1990’s was accompanied by research work on this
technology. In particular, strategic research grant funding from the federal Canadian granting council
NSERC has permitted a team of 16 faculty members and graduate students to undertake an
integrated, coordinated research effort to investigate various factors influencing the success of ‘pulled
in place’ installation (covering both directional drilling and pipe bursting). Figure 1 summarizes the
seven graduate student projects that have been focused on the study of different aspects of directional
drilling. This article outlines the work undertaken to measure the pulling forces that develop (Baumert,
2003; El Chazli, 2005), the response of the pipe to those pulling forces both during and after
installation (work by Chehab and Cholewa), the drilling mud pressures that develop (Baumert, 2003),
and mud loss due to blowout or tensile fracture of the soil (Kennedy, 2004 and work underway by Xia
and Elwood).
 Field studies: measurement of Finite element analysis of
pulling force and mud pressures mud loss by tensile fracture
PhD: Baumert (2003) MSc: Kennedy(2004)

Laboratory study of mud-

pipe friction. Mud loss from blowout and
MSc: El Chazli (2005) fracture in cohesive soils
PhD: Xia (underway)

HDPE pipe response during

and after installation: testing Mud loss from blowout and
PhD: Cholewa (underway) fracture in granular soils: testing
MSc: Elwood (underway)
HDPE pipe response during and after
installation: modeling
PhD: Chehab (underway)
Figure 1. Graduate research projects on Horizontal Directional Drilling

Field monitoring of pulling force and mud pressure

Two aspects of installation practice that have not been fully understood and which still cause problems
in the field are 1. the estimation of expected pulling forces so that installation equipment can be
reliably chosen, and 2. the prevention or control of mud loss as a result of soil failure around the
borehole.

The first key element of the project was the development and use of load monitoring cells to record
axial force histories for loads applied to pipes during pullback operations (when the pipe is inserted
into the prepared borehole). Three load cells were developed by Baumert and Allouche (2003), Figure
2a, and these were used to monitor the installation of HDPE (high density polyethylene), PVC
(polyvinyl chloride) and steel pipes, Baumert et al. (2004). Figure 2b shows a typical pulling force
record measured by a load cell placed between the end of the drill string and the new pipe being
pulled into place. Also shown are measurements of mud pressures at that location.

The effect of pressurized drilling mud on borehole stability, efficient removal of cuttings (during both
pilot borehole drilling, and reaming), and the pulling forces needed to install the new pipe has also
been explored, Baumert et al (2004). In particular, the mud pressures contribute to the pulling forces
measured by the load cell. Figure 2c shows the measurements of total pulling force, as well as the
force calculated as acting on the end of the new pipe (after the effect of mud pressures on the auger
has been removed). Baumert et al. (2005) have explored the relationship between borehole
geometry, mud characteristics, and the mobilized mud pressures using information collected from the
oil and gas drilling communities.
 30 500
TOTAL LOAD MUD PRESSURE
450
25
400

350

Mud Pressure (kPa)

20

Total Load (kN)

300

15 250

200
10
150

100
5
50

0 0
0 20 40 60 80 100 120
Length of Pipe in Bore (m)

a. Load cells b. Record of pulling forces and mud pressures

30

TOTAL LOAD IN-BORE RESISTANCE

25

20
LOAD (kN)

15 REGRESSION TRENDLINE
y = 0.04x + 4.98
2
R = 0.50, σx = 1.26
10 +σx boundary

5
Regression Exit Load
-σx boundary 9.6 kN
0
0 20 40 60 80 100 120
LENGTH OF PIPE IN BORE (m)

c. Pulling force after accounting for the effect of mud pressure.

Figure 2. Equipment and field data after Baumert and Allouche (2002), Baumert et al. (2004).

Table 1. In-bore pipe resistance (200-300 mm OD)

Soil Type
Construction effort Gravelly Sandy Silty Clayey
High
Medium 0.4 to 1.2 kN/m 0.4 kN/m 0.26±.03+ kN/m
Low
*Proposed framework only – values shown based on data collected to date
+
Based on 12 installations in Silty clay

The established pulling force equations have been found to provide poor quality estimates of actual
pulling forces, Baumert and Allouche (2002). Therefore, the pulling force data collected by Baumert
(2003) has been used to propose a new approach for estimating pulling forces, Baumert et al (2004).
This employs a new system reflecting three installation quality classifications:

1) high level, where effort is made to ensure a clean, stable bore, minimizing the solids remaining in
the bore (usually arising with high cost projects or where the absence of a high level of
construction effort may lead to pull loads that exceed rig capacity).

2) medium level, where effort is made to ensure an economic bore; this permits the contractor to
reduce drilling and back reaming times, and the volume of drilling mud (it typically occurs where
there is little risk of exceeding rig capacity or of causing surface heave above the borepath).
3) low level, where the focus is minimization of drilling and back reaming times and the volume of
drilling mud. High solids content in the bore and high installation load levels per unit length of pipe
installed are then considered acceptable. Little effort is made to minimize the risk of surface heave
(this is generally used for shorter installations and should be limited to green-field conditions).

Table 1 provides typical values of force per unit length for installations in soils classified as: gravels,
sands, silty soils, and clayey soils. Further values should be provided as additional projects are
instrumented.

Limiting mud pressures

Drilling mud plays a key role during the drilling process to return cuttings to the ground surface,
stabilize the borehole, and lubricate the pipe being pulled into place. Unfortunately, the pressurized
mud will lead to expansion of the borehole, and if the strength of the neighboring soil is exceeded, the
mud-loss can occur. Both economic and environmental consequences then result. This is not just the
cost of drilling fluid since environmental costs need to be considered.

Kennedy et al. (2004, 2006a) recognized that there are two possible mechanisms of ground failure can
result in mud loss from a borehole:

I. shear failure (or blowout), Figure 3a; and

II. tensile fracture, Figure 3b.

The terms ‘hydrofracture’ or ‘frac-out’ are generally used to describe either of these possibilities.

Kennedy et al. (2004, 2006a) used finite element analysis to examine the conditions leading to tensile
fracture in a cohesive elastic-plastic soil. The limiting mud pressure causing fracture initiation Pfrac is
estimated by considering how the tangential stress in the soil at the crown or springline of the borehole
reduces down to zero from the initial compressive stress of P0 (making the conservative assumption
that the tensile strength of the soil is zero):

Pfrac/P0 = (3K0 -1) for Ko<1,

Pfrac/P0 = (3 –K0) for Ko>1

provided the fracture pressure is less than twice the undrained cohesion of the soil (Pfrac < 2 Cu). They
concluded that the initiation of a tensile fracture occurs at the crown of the borehole where K0 <1, and
at the springline where K0 >1.

Xia and Moore (2006) reported on a new approach to estimate the soil resistance to blowout, providing
allowable mud pressure considering growth of maximum plastic radius with increasing borehole
pressure. The new approach extends the use of purely axisymmetric analysis (e.g.Arends, 2003) to
include explicit consideration of coefficient of lateral earth pressure. They made comparisons between
estimations of the size of the plastic zone surrounding the borehole (where shear failure has occurred)
using their new approach, the so-called ‘Delft equation’ of Arends (2003), and finite element
calculations, demonstrating that their new closed form solution provides better estimates of plastic
radius.

a. Shear failure - blowout b. Tensile failure - hydraulic fracturing

Figure 3. Two mechanisms of soil failure leading to mud loss.
 3

2.5
Pressure ratio, Pmax /Po

2
Fracturing
Hydraulic

Hydraulic Fracturing
Blowout
1.5

Queen's equation, d=2m

Queen's equation, d=4m
1 Queen's equation, d=6m
Delft equation, d=2m
Delft equation, d=4m
Delft equation, d=6m
0.5
0.5 0.75 1 1.25 1.5 1.75 2 2.25 2.5
Lateral earth pressure coefficient at rest K0

Figure 4 Relationship between pressure ratio (applied mud pressure to initial overburden pressure)
and lateral earth pressure coefficient K0 (overconsolidated clay, Cu/P0=0.7).

Maximum mud pressure to prevent mud loss due to tensile fracture can be estimated using the
procedure described by Kennedy et al. (2004). However, if shear failure develops in the vicinity of the
borehole, blowout is the likely failure mechanism and the dependence of the maximum mud pressure
on K0 can be estimated using the Xia and Moore (2006) solution.

Kennedy et al. (2006) developed the following criteria for judging whether hydraulic fracturing or
blowout is critical for a given coefficient of lateral earth pressure at rest, K0:

F1 (K0,P0,Cu) = 0.5 ( 3 K0 - 1) P0 - Cu for K0<1

F2 (K0,P0,Cu) = 0.5 ( 3 - K0 ) P0 - Cu for K0<1

The sign of F depends on shear strength (Cu), borehole depth (P0), and coefficient of lateral earth
pressure, K0. If function F is positive, blowout is expected, whereas if negative, tensile fracturing is
expected.

Various values of cohesive strength and the coefficient of lateral earth pressure were used by Xia and
Moore (2006) to investigate conditions that produce tensile fracture and blowout. For example, Figure
4 shows results for Cu/P0=0.7 (a possible strength condition for an overconsolidated clay). Maximum
mud pressure values relative to initial vertical stress P0 are given for a range of K0 values. Tensile
fracture is expected in these heavily overconsolidated clays where K0>1.8. Blowout is expected at
somewhat higher mud pressures when K0<1.8. Other similar calculations reveal that blowout is most
likely in normally consolidated and lightly overconsolidated clays.

Kennedy et al. (2006b) conducted a finite element study examining fracture initiation in the filtercakes
that develop when drilling in sand. This work illustrates the likely significance of the filtercake cohesive,
the initial stress conditions, and the friction angle of the sand. Physical tests are needed to understand
how fracture initiates and propagates through the soil to the ground surface.
a. Pipe-mud-borehole system and regions of contact b. Pipe-soil friction tests
Figure 5 Investigation of pipe friction through borehole, El Chazli (2005).

a. Pulling

b. Recovery

c. Long term

Figure 6 Schematic of load path for HDPE pipe at end of and following installation.

a. Laboratory test specimen b. Internal pipe restraint

Figure 7 Testing of HDPE pipes to simulate installation (see also Cholewa et al., 2006).

20
20
E VE VP
15 15
True Stress (MPa)
True Stress (MPa)

10
10

5 Experiment
5
Model
0
0 0.02 0.04 0.06
0
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 -5
True Strain True Strain
a. Different components of strain b. Model performance for cyclic loading
Fig. 8 Modeling of HDPE to capture strain reversal, Chehab and Moore (2006).

Pipe response during pullback

El Chazli et al. (2005) examine the interaction of pipes sliding within the borehole, Figure 5a, and
developed filtercake surfaces in the laboratory to facilitate measurements of pipe-filtercake friction,
Figure 5b. Friction factors of from 0.2 to 0.5 were measured between steel and PVC pipes and the
filtercake. They conclude that time-dependant behaviour (rheology) of drilling mud is a major factor
affecting interface behavior, and that the interface behavior may, in part, account for the high variability
observed in measured pull-loads.

Figure 6 represents three different stages of HDPE pipe response during installation. In the top
illustration, the end of the installation process is being approached, with a large axial tension from the
drill string applied to the end of the pipe, so that it moves through the ground. The pulling force from
the drill rig is resisted by shear stresses that act on the external surface of the pipe. These actions
result in increases in length of the pipe (tensile axial strains).

After the pipe has been pulled right through the borehole, it is left without end restraint for some period
of time. As a result, the pipe begins to return to its original length. The axial contraction (strain
recovery) that results will lead to shear stresses between the exterior of the pipe and the surrounding
soil.

Finally, the ends of the pipe are attached to the fixtures that the pipe is intended to connect. From this
point, the length stays constant and the resistance imposed to prevent any additional axial contraction
leads to tensile axial stresses in the pipe.

These have been investigated in detail by Cholewa et al. (2006). Test fixtures have been developed to
permit application of the load histories like those measured in the field, Figure 7a, and an internal
restraint system and load cell permits long term measurements of axial stress history, Figure 7b.

Current HDPE pipe selection for use in directional drilling projects is based on the specification of pipe
wall thickness to ensure that the maximum pulling force is less than some fraction of the ultimate axial
load capacity of the pipe (e.g. 30%). The complex loading history discussed in the previous section
and the axial tensions that are expected to be acting over the life of the pipeline lead to questions
regarding the effect of those long term axial tensions on pipe durability. Furthermore, the loads
measured where the pipe is attached to the drill string are not representative of the loads acting along
the rest of the pipe during installation, since there are different shear stress regions, with shear
direction dependent on the direction of mud flow, and/or the direction of pipe wall movement relative to
the surrounding borehole.

Chehab and Moore (2006) describe the adaptation of the procedures of Moore and Hu (1996) and
Zhang and Moore (1997ab) to develop new viscoelastic-viscoplastic constitutive models for estimating
the cyclic response of the pipe materials, and the tensile stress history of the pipe during and after
installation. Figure 8a shows the three components of the response: elastic (‘E’), viscoelastic (‘VE’)
and viscoplastic (‘VP’), and the ability of the model to capture the cyclic response of the HDPE is
demonstrated in Figure 8. The model performance will be assessed further using laboratory data
obtained by Cholewa, and a one dimensional analysis developed to capture the complexities of pipe-
soil interaction and the dependence on axial force on time and location.

Acknowledgements

This research was funded by Strategic Project Grant No. 257858 from the Natural Sciences and
Engineering Research Council of Canada (NSERC) with industry support provided by Mr. Richard St.
Aubin of Ipex Industries and Mr. Trevor Fitzell and Mr. John Scholte of Golder Associates (Burnaby).
The contributions to this work on directional drilling of Dr Richard Brachman, Mr. Abdul Chehab, Mr.
John Cholewa, Mr. Ghassan El Chazli, Dr Sean Hinchberger, Mr. Matthew Kennedy, Dr Graeme
Skinner, and Mr. Hongwei Xia are gratefully acknowledged.
References

Arends, G. 2003 ‘Need and possibilities for a quality push within the technique of horizontal directional
drilling (HDD)’, Proceedings of 2003 No-Dig Conference, Las Vegas, Nevada, March 31-April 2.

Baumert, M.E. 2003. ‘Installation loads on pipes installed using horizontal directional drilling’, Doctoral
Thesis, Faculty of Engineering Science, The University of Western Ontario, London, ON, Canada

Baumert, M.E. and Allouche, E.N. 2002. ‘Methods for estimating pipe pullback loads for HDD
crossings’, J. Infrastructure Systems, American Society of Civil Engineers, 8(1), 12–19.

Baumert, M.E. and Allouche, E.N. 2003. Real-time monitoring for quality delivery of directional drilling
installations, J. Infrastructure Systems, American Society of Civil Engineers, 9(1), 35-43.

Baumert, M.E., Allouche, E.N. and Moore, I.D. 2004. ‘Experimental investigation of pull loads and
borehole pressures during horizontal directional drilling installations’, Can. Geotech. J., 41, 672-685.

Baumert, M., Allouche, E.N. and Moore I.D. 2005. ‘Drilling fluid considerations in design of engineered
horizontal directional drilling installations’, Int. Journal of Geomechanics, 5(4), 339-349.

Chehab, A.G., Moore, I.D. 2006. ‘A viscoelastic- viscoplastic constitutive model for high density
Polyethylene’, Pipelines 2006, ASCE, Chicago, July-August, 7pp.

Cholewa, J.A., Brachman, R.W.I., Moore, I.D. 2006. ‘Influence of HDD installation on the stress-strain
response of HDPE pipe’, International Society of Trenchless Technology, Brisbane, Qld, Australia.

El Chazli, G., Hinchberger, S.D., Baumert, M., Allouche, E. 2004 ‘A medium-scale apparatus and
procedure to measure interface shear strength for HDD design’, Proc. 2005 No-Dig Conf., New
Orleans, Florida April 24-27, 2005.

Kennedy, M.J. 2004. ‘Finite element calculations of hydraulic fracturing during horizontal directional
drilling’, MSc Thesis, Department of Civil Engineering, Queen’s University at Kingston, Ontario,
Canada.

Kennedy, M.J., Moore, I.D., Skinner, G. 2006. ‘Elasto-plastic response of a sand-filtercake soil system
during horizontal directional drilling’, Int. J. Geomechanics, ASCE, 6.

Kennedy, M.J., Skinner, G., Moore, I.D. 2006. ‘Limiting drilling slurry pressures to control hydraulic
fracturing during directional drilling operations in purely cohesive soil’, Trans. Res. Rec. (to appear).

Kennedy, M., Skinner, G., Moore, I.D. 2004 ‘Elastic calculations of limiting mud pressures to control
hydro-fracturing during HDD’, No-Dig 2004, April, New Orleans, LA, paper E-1-03, 10pp.

Moore, I.D. 2005. ‘Analysis of ground fracture due to excessive mud pressure’, Proc. 2005 No-Dig
Conf., New Orleans, Florida April 24-27, 2005.

Moore, I. D. and Hu, F. 1996. ‘Linear viscoelastic modelling of profiled high density polyethylene pipe’.
Canadian Journal of Civil Engineering, 23, pp. 395-407

Xia, H. and Moore, I.D. 2006. ‘Estimation of maximum mud pressure in purely cohesive material
during directional drilling’, Geomechanics and GeoEngineering, 1, 3-11.

Zhang, C. and Moore, I. D. 1997. ‘Nonlinear mechanical response of high density polyethylene. Part
I: Experimental investigation and model evaluation’. Polymer Engineering and Science,37(2),404-413.

Zhang, C. and Moore, I. D. 1997. ‘Nonlinear mechanical response of high density polyethylene. Part
II: Uniaxial constitutive modeling’. Polymer Engineering and Science, 37(2), 414-420.