ClayMinerals (1998) 33, 171-183

Mechanical properties of North Sea

Tertiary mudrocks" investigations by
triaxial testing of side-wall cores
L. WENSAAS l , P. A A G A A R D , T. BERRE* AND E . R O A L D S E T t

Departmentof Geology, Universityof Oslo, P.O.Box 1047Blindern, N-0316Oslo, *NorwegianGeotechnicalInstitute,

P.O. Box 3930 UllevhlHageby, N-0806 Oslo, and tDepartment of Geology and Mineral ResourcesEngineering,
Norwegian Universityof Science and Technology (NTNU), N-7034 Trondheim, Norway

(Received 28 August 1996; revised 2 August 1997)

A B S T R A C T : In the North Sea Tertiary section, wellbore instability problems are frequently
reported in Palaeocene-Early Oligocene smectite-rich mudrocks. Analysis of the mechanical
properties of these Tertiary mudrocks is generally hampered by the lack of suitable core material.
This study represents an attempt to study the geomechanical behaviour of mudrocks by triaxial tests
of side-wall cores obtained from the borehole wall. The tests performed include measuring the
changes in pore pressure during shearing and undrained shear strength in specimens initially
consolidated to in situ effective stress levels. The coefficients of permeability (kf), estimated from the
consolidation time behaviour range from 2.6 x 10 - n to 2.4 x 10-12 m/s. The tested cores behaved
like slightly overconsolidated to normally consolidated materials with an initial near constant volume
(elastic behaviour) for low deviatoric load followed by an increasingly contractant behaviour
approaching failure. Compared with results from onshore analogues, the strength properties of the
investigated mudrocks appear to be related to their content of expandable clay minerals. A wellbore
stability chart to forecast adequate drilling fluid pressures for future wells has been developed by the
use of linear (Mohr-Coulomb) failure criteria based on the peak strength data. It is demonstrated that
side-walt cores can provide satisfactory test materials for rock mechanical analysis, and their use may
serve to improve our knowledge of the rock mechanical behaviour of typically troublesome
mudrocks for which no conventional cores are available.

Drilling problems caused by compressive or tensile drilling fluids. The latter includes the use of high-
failure o f the borehole wall often result in concentration KC1 water-based drilling muds (e.g.
substantial loss of drilling time and increases in Steiger, 1982; Chenevert, 1989), or the use of
rig time costs. Poor hole conditions (tight hole, hole various water soluble semi-synthetic or synthetic
enlargement, hole collapse, hole fill etc.) may polymers (e.g. Bruton & McLaurine, 1993). The
create problems in running the casings and reduce various inhibitor systems have generally been
the quality of petrophysical log data. Drilling and proven to reduce wellbore instability. Problems
borehole stability problems are related both to may still be experienced even with the most
chemical and mechanical processes. Chemically inhibitive drilling fluids, suggesting a mechanical
influenced swelling of clayey rocks can be nature of the instability problems.
reduced by the use of oil-based drilling fluids or The main objective of wellbore stability analysis
by modifying the chemical activity of water-based is to determine the critical upper and lower drilling-
fluid pressure limits ('drillability window') to avoid
compressive and tensile failure around the borehole
1 Present address: Statoil Research Centre, Postuttak, wall. Ideally, borehole stability should be handled
7005 Trondheim, Norway. as a combined approach including both chemical

9 1998 The Mineralogical Society

172 L. Wensaas et al.

and mechanical processes. Physicochemical dominated by ash-fall deposits (Balder Formation)

processes include the interchange of fluid and ions derived from the volcanic activity associated with
between the drilling fluid and the surrounding rock the opening stage of the Norwegian Greenland Sea.
which may modify both the stress conditions and Renewed Eocene subsidence formed deep depocen-
the mechanical properties of the mudrocks (Bol et tres along the Viking Graben (Jordt et al., 1995)
al., 1994; Mody & Hale, 1993). The transport which were filled gradually by Middle-Late Eocene
processes (e.g. osmotic flow, diffusion, hydraulic hemipelagic muds mixed with minor sands (Lower
flow) between the formation and the drilling fluid Hordaland Group) gradually changing to Early
are complex. Present conceptual models are Oligocene interbedded silty/sandy mudrocks and
weakened by their qualitative nature, where muddy diatomites towards the top of the studied
several processes have to be calibrated empirically interval (Middle-Upper Hordaland Group).
(Warpinsky, 1992).
In the North Sea Tertiary section, instability
SAMPLING AND ANALYTICAL
problems are frequently reported in Palaeocene-
TECHNIQUES
Early Oligocene smectite-rich mudrocks, however,
the necessary input data for modelling of physico- The investigated side-wall cores were sampled by
chemical processes in typical heterogeneous North shooting a small hollow bullet into the borehole
Sea Tertiary mudrocks are not yet available. A wall. A small cylindrical piece of rock with an axis
better understanding of the mechanical properties of normal to the well is thereby retrieved using a
these rocks has become increasingly more important wireline tool. Side-wall cores selected for triaxial
due to the extensive drilling of deviated wells. But strength tests (Fig. lb) were cleaned gently,
due to the lack of adequate core material, very few wrapped into cellophane and covered in wax
rock mechanical tests on North Sea Tertiary (offshore) to preserve their water content. The
mudrocks have been undertaken and even fewer samples were stored in glass jars at a temperature of
are reported in the literature. -4~ until the testing took place. Opening, sample
The main objectives of the paper are twofold: description, and sample preparation (trimming)
(1) to determine whether carefully selected side-wall were performed at high humidity (>98%) condi-
cores can represent appropriate test material for rock tions. The tests were performed using a triaxial cell
mechanical testing, and (2) to evaluate whether (70 MPa cell pressure capacity) modified as part of
calculated wellbore stability from the obtained peak this study to handle cores with a diameter of 25 mm
strength data and a simple linear (Mohr-Coulomb) (by T. Berre, Norwegian Geotechnical Institute).
mechanical approach can be correlated with the The test cell had both external and internal load
observed drilling problems in the studied mudrock transducers, two axial strain measuring devices, two
sequence. The test results are compared with radial strain measuring callipers, units for measure-
published results from triaxial tests of onshore ment and control of cell pressure, and a unit for
mudrocks (Weald Clay, London Clay and Fullers volume change measurement and for measuring and
Earth) and the rock strength behaviour has been control of back pressure. During the experiments,
considered in relation to mineralogical composition. axial load, confining pressure, radial and axial
displacement, pore pressure and back pressure, were
continuously monitored and recorded with an
GEOLOGY OF T H E STUDY AREA
automatic data-acquisition system.
Well 25/7-2 is located close to the eastern boundary The triaxial tests were run as isotropically
of the southern Viking Graben on the north-western consolidated undrained (CIU) tests, equivalent to
flank of the Utsira High (Fig. 1). The studied published triaxial tests on shales (e.g. Steiger &
Palaeogene sequence (c. 1200-2600 mKB) consists Leung, 1992) and onshore analogous tests (Nakken
of large Palaeocene submarine-fan complexes (Ty et al., 1989; Marsden et al., 1989, 1992). The CIU
and Heimdal Formations) interfingering with tests are divided into two parts: (1) drained
marine hemipelagic mudrocks (Lista and Sele isotropic consolidation to the level of the calculated
Formations) derived from deltaic progradation in s i t u stress (Table 1); and (2) undrained
eastward from the Shetland Platform into the (deviatoric compression) with a recording of the
Viking Graben (Isaksen & Tonstad, 1989). Late pressure build-up. During the isotropic consolida-
Palaeocene to earliest Eocene deposits were tion the samples were drained from both ends
 Mechanical properties of mudrock~ 173

D DD rlDDOOOO0 [] DD [] 13 ~ n~ ~ [] O n O0 [] [] D [] [] []

;;44 ,I, ,~ ;
I#,1 u~ I I LU ----.INI~I E: L~I ~ 191 0 0 '''~

e~
e,i'~
, ~
r ' - ,.~

9 9 I" I-I'1

o o o o o o o o o o
e~

(83~u) q),dao

z C~ ~

[] ::::::::::::::::::::::::::::::::::::: ,
::::::::::::::::::::::::::::::::::::::::::
:: :::::::::::::::::::::::::::::::::::::

~1 [~ :m:m:m:m:::m:::r ::::::!:K:!:!:1%!:K:!%1:
i:i~ [i:!:!:!:!:!:!:i:!:!:!:!:i:i:!:!:!:!:!:i:i:!:!:i:!:!:i:i:!:i:!:~
::::~~!~#~ie~i~i~i~i~i~i~i~i~i~i~i~i~i~i~i?i~ih~i?i~i(~i~iiiiiiiiiiiiiiiiiiiiiiiiii~
<~i~h 1 ~[ ~ |:F~<ilFiiiiiii%ii!i

1 t~ !
: ::: ....... ~ ~:::::::::::::~::::::::::::::::::::: ~::<:::::::: ::::::::::g::::::::::~:~::::~:::::::::::::: :::::: ::: :::::::::::::::::::::::::::::::::::::::::::::::::::
:: ::::::: ::::::::::::::::::::::::::::::::::

:i~i~...........
i~ ~i~i~i~~ ~;;;;~;m~i~i~i~i~!~i~i~i~L>~i~i~K~i~ .........~ ! ~ : ~ : ; ~ : <:~;;~m~m~m~:~;~ ......... ~=::~:~=::~:~:~:~:~::~:~:~:~:~:~::
........... ~:::~:~::::i:
~.... ~ ~iiiiiiiii~i~i,i,i~i~i~i,~,i,i~i~ ~::~:::~:, .............~,~,~,~,~,~,~,~,~,~, ~<:~i~i~i~%i~i~i~i~i:i~i~::~i~i~i~<:~iiiiiiii~i~/:i~ii:i~iii

~~:~:iiii ~iiL ~iiii~iiiiiii~i~iiiiiiiiii~i~i C~ !:i:i :!iiii!i~iL :.......... qi!i!iiii iiiiiiiiiiiiiii!ii ...............iiiiiiL !~:::::::ii:iiii~
~ ~:Ui~iiiiiiiiiii~ii~i~:iii~iii!~i~ii~i~<~ii~i~i~i~iiii~i~iiiiiiiiiii
ijiiiii!i~ii!i!i~iiii!!iiii

k.._,
NO
.....::u<::3::i:;i:i:iiil :::: ~ i ~ :: : : -:::::::!: ~ iiiiiii:iiiiiiii::iijjiiiiiiiiiiiiiii:iiiiiiiiiiii:iiiiiiiii:::::i::
: ii::::i:::: ...................

r-u.K. ~ ....
! ~ ~ ~ ~

I~ ~ ~_/14 ~lOl~l ~ :~:::: :#:~:~:~

~'~-:; ..... ~.'
174 L. Wensaas ct al.

0
o
o

o
i

_ ~ ~ d d d d
o

r~

~'~

oo

~g c~ r~ " ~ ~ 0

.~.~ :8oo 0 0 0 ,~
 Mechanical properties of mudrocks 175

against a constant backpressure (6 MPa). Side side-wall cores is presented elsewhere (Tyridal,
drains to assist the drainage were not used. The 1994) and only a summary for the tested specimens
consolidation serves three purposes: to rescind the is included here. The XRD analysis confirmed the
effects of sample disturbance (e.g. micro-cracks): to mudrocks to be homogeneous containing the
consolidate the sample to saturation: and to dissolve following: 71-82% (average 75%) clay, 5 - 1 4 %
any air inside the pore-space (Steiger & Leung, (average 7%) quartz, 0 - 4 % (average 2%) K-
1988, 1992). All specimens were saturated with feldspar, 2 - 7 % (average 4%) plagioclase, 1-10%
weak brine (52.0 g/1 KC1) at the consolidation (average 6%) pyrite, 3 - 8 % (average 5%) calcite,
stage, in order to prevent swelling during consolida- and 0 - 2 % dolomite and siderite. The fine fraction
tion. The degree of initial saturation (Swi) was (<2 ~tm) contained between 35-65% (average 46%)
proved by performing a 'B' value test (Bishop, smectite and mixed-layer I-S, 10-30% (average
1973). The strain rate used during the shearing 16%) illite, 0 - 5 0 % kaolinite and 0 - 2 5 % chlorite.
phase was calculated using the results of the
consolidation data, according to the recommended
Consolidation
procedure given by the British Standards Institution,
London (BS 1377: Part 8: 1990). Peak strength was During the drained isotropic consolidation, an
generally achieved after a testing time (tf) of equal all-round effective stress produces a change
14-40 h of deviatoric loading. All tests were run in sample volume, and for fully saturated samples,
at room temperature. An overview of the sample this volume change is directly related to the drained
data and the testing programme is given in Table 1. volume of water and hence the change in porosity
To obtain some broader information about the test (Table 1). High volumetric strains (EvoD and
material, sample 8 (1540 mKB) was consolidated to octahedral strains (Coot) (14-23%) during consoli-
only 10% of its in situ stress level and sample 9 dation indicate that the specimens have been
(1361 mKB) was run as an extension test. The affected by relaxation prior to analysis (e.g.
permeability of the specimens was estimated from swelling/microfracturing). Values for drained bulk
the consolidation behaviour according to the modulus (B) for the isotropic consolidation were
method of Head (1986). obtained at maximum effective confining pressure.
The mineralogy of the bulk samples and also of The coefficients of permeability (kf) estimated from
the <2 ~tm fractions was determined by X-ray the consolidation time behaviour (Table 1) range
diffraction (XRD) analysis of unoriented (powder) from 2.6 x 10 - l l to 2.4 x 10 -12 m/s (2.7 x 10 -6
mounts and oriented specimens (inverted Millipore to 2.5 x 10-7 Darcy), which is within the micro-
slides), respectively. The fine fractions were nannodarcy range reported by Steiger & Leung
scanned after the following treatments: (1) Mg- (1990). Similar permeabilities are also reported
saturation; air dried; (2) ethylene glycolation; and from fractured shales ( 1 0 - s - 1 0 -s Darcy), whose
(3) preheating (550~ h). All analyses were permeability of the intact rock is several orders of
performed on a Phillips PM 1700/1710 diffract- magnitude lower (Chenevert & Sharma, 1991).
ometer with Cu-Kct radiation (1.54060 A), Thus, if the measured high strains during consolida-
connected to a PDP 11/53 computer system with tion reflect closing of micro-cracks created by
a Phillips APD software. previous sample relaxation, the estimated perme-
abilities may have been overestimated. However,
large consolidation strains and subsequent large
RESULTS
reductions in porosity have also been reported from
CIU triaxial tests of onshore analogous clays
Physical properties and mineralogy
(London Clay and Fullers Earth) at offshore stress
All results of the geomechanical testing are conditions (Marsden et al., 1992). Hence, large
summarized in Table 1. The initial water content strains during consolidation seem to be typical for
(wi) ranges between approximately 38 and 20 wt%, soft, smectite-rich mudrocks.
decreasing with increasing burial depth. The
corresponding calculated initial (unstressed)
Shearing
porosity (ni) ranges between 57-36 vol%. The
initial degree of saturation (Swi) is between Figure 2 shows the deviatoric stress-strain
83-95%. The mineralogy of drill bit cuttings and response and the change in pore-fluid pressure (u)
176 L. Wensaas et al.

(edl~l) e J n s s e J d e J 0 d
0
~ o= ~
, i 0 i , 9 i i , 9 . i . 9 , i , , . i , , o
o i
~, ~,
i 9 9 ' 9 , 9 | 9 9 9

. =o ~
.,,:1- to A

, ~

q
\

/
en J o g~.=
NOISS~ddlAIOD NOISN~IX~ ~o

0
0
. . . . . 0 . . .
'L ~0 ~3

~. r
- I 2;
tO "-" - .= "~

O3
~ / / I I \ -

m ~, ~ . ~
c -

J, o
.- ~ ~ ;~
. m
Od ~ o~ u
'~ 0
.i I

o
NOISS~PddlAIOD NOISN~tlX~/
' .o~
o.

( \(( I \
":9. .m
t'--

..i.--,
~ : (/-)

O *~ / 0
"t-
O ~. "~
"r"
9 0 , - - . . . .
O
,

(edN) s s e J 1 S o!JOle!AeO
~,.~
 Mechanical properties of mudrocks 177

during the undrained shearing phase. The stress- correspond to shortening of the sample). As the
strain behaviour of all samples during deviatoric pore pressure increases throughout the tests
loading is basically the same; any difference is in (Fig. 2b), the stress paths get curved to the left
magnitude rather than nature. At low deviatoric (Fig. 3a). Pore pressure increases in undrained tests
stresses (q = Ol - or3) and low strains the because the specimen would reduce its volume due
specimens show a close to linear behaviour. The to the contractant behaviour, but contractancy is
curves show a gradual transition into an upward prevented by the test configuration. Initially, the
convex shape with a plateau of near constant stress mean effective stress p ' appears constant (i.e. path
in the post-failure region. Such overall stress-strain close to vertical) as expected for undrained constant
performance is typical of plastic/ductile behaviour. volume shearing of an isotropic elastic material.
The undrained shear strength is reached after low Based on the calculated pore pressure parameter A
(2.5-3.5%) axial strain (1.4-2.0% lateral strain). (Skempton, 1954) and dilatancy (D) (Janbu, t985)
The pore pressure in undrained compression tests at failure (Af, Dr in Table 1), the mudrocks tested in
(Fig. 2) continued to show a small but steady compression behave like normally consolidated to
increase in the post-failure region. The extension lightly overconsolidated materials (Af between 0.56
test showed a minimum pore pressure at about the and 0.79) with an initial near constant volume and
peak stress stage, from which it increased towards elastic behaviour for low deviatoric load followed
zero during post-failure deformation. Values of the by an increasingly contractant behaviour (De < O)
secant Young's modulus (E89 and Poisson's ratio approaching failure.
(~tlf) from deviatoric stress vs. axial strain curves Values for the cohesion e' (1.1 MPa) and the
obtained at half the peak deviatoric stress are given angle of friction qb' (17 ~ were obtained by fitting
in Table 1. The samples still appear to be within the the Mohr-Coulomb criterion to a linear regression
linear range at half their peak load. Young's of the peak strength data in the p'-q diagram
modulus (stiffness) represents the slope of the (Fig. 3b). A non-linear failure criterion proposed by
change in deviatoric stress against change in axial Marsden et at. (1989) is shown for comparison.
strain, while the dimensionless Poisson's ratio (~t) is
the negative ratio of the horizontal and vertical
APPLICATION TO WELLBORE
strain for the given stress range. Young's modulus
STABILITY
increases with increasing effective mean stress,
while Poisson's ratio shows no variation with stress. The following discussion is based on the assump-
The undrained shear modulus (G) represents the tion that the wellbore instability in the studied well
samples' resistance against shear deformation is purely mechanical in nature. Evaluation of
(modulus of rigidity). The initial values (Gi) mechanical stability of the wellbore consists of
measured during undrained shearing phase, vary two parts; calculation of the stress state around the
from 0.2 to 2.1 GPa (Table 1). Despite the change borehole and consideration of a failure criterion.
from an initially stiff to a gradually more soft The vertical overburden stress (Fig. 4a) was
behaviour approaching failure, the material largely determined from integration of the rig bulk
retains its ability to carry a load in the post-failure density measurements on drill bit cuttings. The
situation. Furthermore, high shear stresses in the 'best empirical solution' of the minimum horizontal
post-failure region suggest that the frictional stress (sh) was taken as the 'minimum envelope' of
resistance along the failure surface is high for the leak-off-test (LOT) data from 21 nearby wells
entire stress range studied (i.e. for stress conditions (Fig. 4a). No borehole breakout data were available
corresponding to typical wellbore situations). for determination of horizontal stress anisotropy and
the two horizontal stresses (O" h and ~rt) were
assumed to be similar in magnitude. The pore-
Stress paths and p e a k strength
fluid pressure estimated from wire line logs and
Figure 3a illustrates the stress paths in the q-p' drilling parameters indicate a maximum absolute
diagram for the triaxial compression tests. The overpressure in the investigated Tertiary section of
failure stress (undrained strength) is taken as the -4.7 MPa.
maximum principal stress difference qf = ( ~ 1 - - O'3)f Failure normally implies the condition in which
(Table 1), or the minimum stress difference in the the specimen can sustain no further increase in
extension test (positive deviatoric stress values stress. The linear Mohr Coulomb criterion (Fig. 3b)
178 L. Wensaas et al.

:>,.~

Q..

u') b'7
u) 0
-}-

M~ u')

._>_ II ,=i

U
r
0.)
t'-

0.) u~

I i
.Q I I I I I O
O I~ O ~r) 0 i~ O
u~

(BclIN) b sseJls o!Jole.~ea

0
c'O

CO
C'J
o~
",0

C~I

ao
...-.. o.~
,,0

"::::i"

0.) = ~.,
O4 .> -~
Oo~OOO.O ~ e [] 0

~ ~ I~
E
(D
c:)
v ',,0

~u
,.._:
v
co
C~
I I I I I I I i
oO (:2)

(odiN) b ssOJ:lSoiJo:lg.U~e(] 6~
 Mechanical properties of mudrocks 179

.~o-~
" ~ o
._= = ~ i o ~ ~ ~

-~S o
0 J j =~ ' 0 . ~j 0
,~ -- i1~ ..'~i::iii: '
o, o. I ,.",~iiiiiiiiiii~ I /
.o o ., :::::::::::::::::::::::::::::::::::::::::::::: / / \\v

o,, o~. v.
~.. :::::::::::::::::::::::k l ,9~ \~ ~ . v
..............................................
~. :::::::::::::::::::::::::::::::::::::::::::::::::::
v.,~,~
_~ 9 I ,~," ::::::::::::::::::::::::::::::::::::::::::::::::::
{~. ~ ~./ 9 ::::::::::::::::::::: t-}
i,a .- ~$!:!:~:i:!:i:i:i:i::~i:i:::~
9

1""" x ~z~::~::~i
N ~::~ii~:::~~ ~
: , . ~.~....!~.~,,
o _~ ~

I I "' "':::::::~ ~ I ~"

/ ,~i::iiiiiiii::~~*~ ...~'~" ~

./,., ./~.~w I
~'=~ ~
e~
~ ~.~
e~
o .... ~ 1; .... ; .... ;t ; .... o o
_ 0 0 0 0 0

(8)t'") q;dea
=-~ c~'~

-~~
IO.i=, I ' ~ ./ ..,'F ~~
,~0 ~ .~ ,

II ,-,'~" _..,-
~.-
..,'
._" ,.-'" :~ " ,>//
i1.,
~ o.gr3~ _~ -
~_
m
;"~" -:' .
I ~ "-"-'.--'~"" F
I ,a ~ 1 7 6
./
f
t~
0 0 0 0 0 0 0 0
0 0 0 0 0 0 0
~.~ ..

(8Ntu) qldeo
180 L. Wensaas et al.

incorporates the maximum (g~) and minimum (or3) The results of the stability calculations are given
principal stresses, and because of its simplicity it in Fig. 4. Although stability models based on linear
has been extensively used in the petroleum industry behaviour, typically tend to provide a conservative
(Aadnoy, 1988; Aadnoy & Chenevert, 1987; estimate of the 'drillability window', there is a good
McLean & Addis, 1990a,b). A failure criterion agreement between the results of the stability
defines the critical state boundary line between calculations and the reported hole problems
possible and impossible stress states (Fig. 3b). between 1000 and 1900 mKB (e.g. twist off, tight
Major shear deformation will only occur if the hole, pipe stuck, hole pack off). In this interval the
material is loaded in such a manner that it attains drilling fluid pressure is slightly lower than the line
this critical state. Here plastic shear deformation defining the calculated lower pressure limit
may continue without change in volume (Fj~er et (Fig. 4b). After an increase in the drilling fluid
al., 1992). In assuming a linear relationship, the pressure at 1900 mKB it entered the stable region
error appears negligible, at least for stress ('drillability window') and hole stability was
conditions corresponding to typical wellbore situa- reported to improve. The results also imply that
tions (<14 MPa; McLean & Addis, 1990a). the drilling-fluid pressure can be further increased
Failure of the borehole wall is induced by drilling in the troublesome Tertiary interval without any
fluid pressure which is either too high (tensile major risk of tensile failure. Moreover, drilling with
failure) or too low (compressive shear failure). In a low (1.07 g/cm 3) drilling-fluid weight down to
this paper, calculations of the upper and lower limits 1500 mKB in a nearby well resulted in stuck pipe
of drilling-fluid pressures were performed according and the hole being packed off at 1133 mKB and
to Fjmr et al. (1992) (eqns. 1 to 4) assuming a then having to be side-tracked. After increasing the
vertical hole with impermeable walls and a poro- drilling fluid weight to a similar level as in the
elastic material behaviour. Provided that isotropic studied well (1.32 g/cm 3) the well was successfully
stress conditions prevail, the stresses around a drilled throughout the Tertiary interval.
borehole can be expressed as (Fj~er et aL, 1992):

err =Pw DISCUSSION

(I0 = 2Crh -- Pw (1) As eqns. 2 and 4 indicate, the calculated wellbore
(~z = ~ v
stability is very sensitive to the uncertainty in the
where c~0, clz, and ~r represent the tangential, input parameters such as field stresses and pore-
vertical and radial stresses, respectively, around the fluid pressure. Pore-fluid pressures of 1.3 g/cm 3
borehole, and Pw the pressure of the drilling fluid MWE (mud weight equivalents) (12.8 kPaJm) are
(well pressure). With the relative magnitudes of the typically reported in the Lower Tertiary shales.
stresses being; c~0 < ~z < clr, the lower critical Quantification of shale pore-pressures (by indirect
drilling-fluid pressure (i.e. compressive shear methods) has been shown to be strongly influenced
failure) becomes: by lithological and mineralogical variations
(Wensaas et al., 1994), and these variations may
Pw ~ [2~h + ~xpf(tan2~ -- 1) -- Co]/(tan2~ + l) (2)
introduce an uncertainty in the estimated shale
where cz is the Blot's constant (assumed = 1), Pr is pressures. Calculations assuming hydrostatic pres-
the formation fluid pressure, 13 the angle of the sure conditions cause the lower critical pressure line
failure plane to the major principle stress (c~e), and to decrease (max. 5%) and the drilling fluid
where the uniaxial compressive strength Co is pressure tends to enter the 'drillability window'.
defined as: This suggests that the reported drilling problems
may essentially be driven by overpressuring.
Co = 2c 'tan~3 = 2c 'tan[(~b/2) + 45 ~ (3)
The oh trend (Fig. 4a) is very similar to the trend
The upper critical drilling-fluid pressure (i.e. o b t a i n e d from the C e n t r a l N o r t h Sea b y
tensile failure) becomes: Gaarenstroom (1993). Calculations with a higher
oh trend as proposed by Breckels & van Eeklen
Pw ~> 2crh - cxpf+ To (4) (1982) or Kwakwa et al. (1991), cause the lower
where To is the tensile strength, which due to small critical pressure line to increase. This will result in
fractures and cracks at the borehole wall, is a conservative estimate relative to the field
assumed to be zero. observations. The uncertainty related to the
 Mechanical properties o f mudrocks 181

60

1)
50
2)
0 40 1) Weald Clay
t3_

3) 2) Weald Clay

r 30 3) Weald Clay
o
j/ \ SMECT,TECO.TE.T 4) SWC
-~ 20 5) London Clay
C3
6) Fullers Earth
]0

I I I

0 10 20 30 40

Mean effective stress p' (MPa)

FIG. 5. A comparison with peak strength data from onshore analogues; (1), (2), (3) Weald Clay; smectite content
0-1% (Nakken et al., 1989; Marsden et al., 1989; 1992); (4) North Sea Tertiary mudrocks (this study); smectite
content 40-60%; (5) London Clay; smectite content 35-45% (Marsden et al., 1992); and (6) Fullers Earth;
'pure' smectite (Marsden et al., 1992).

determination of the Oh trend appears to be more CIU triaxial tests of various 'pure' clay mineral
critical to the stability forecast than the uncertainty samples, Wang et al. (1980) found that the peak
related to overpressure. strength of montmorillonite (smectite) was about
The undrained strength of shales and mudrocks is half of the strength of other clays at a given
not a unique parameter, but depends upon several confining pressure.
factors, where we regard the most important to be By performing the test exclusively on a set of
material type (e.g. mineralogy, texture, fabric, water carefully selected, full recovery cores (without any
content), present stress level, previous stress visible anisotropy), we anticipate that any influence
history, anisotropy and testing conditions. By from sampling distortion (e.g. micro-fracturing,
comparing the side-wall core data with data from swelling) on the obtained rock strength data is
onshore analogues, a correlation between rock kept to a minimum. Interestingly, both the recovery
strength and smectite content could be anticipated and quality of the side-wall cores were found to
(Fig. 5). The strength of the samples decreases with correlate broadly with variations in lithology and
increasing smectite content, from Weald Clay with mineralogy. Low recovery cores are typically
negligible smectite content to Fullers Earth with fractured and consist of silty and/or calcareous
close to 100% smectite. Similar correlations have mudrocks, whereas full recovery cores, with no
been reported by Shell and Exxon between low visible cracks, consist of fine-grained smectite-rich
shale strength and high water contents (Mody & mudrocks. Mineral reactions which tend to
Hale, 1993), or high surface areas (Steiger & strengthen the mudrocks, e.g. smectite diagenesis
Leung, 1990; Ewy et al., 1994), both being typical (>1800 mKB), carbonate cementation and Opal CT
properties of smectite-rich rocks. Moreover, from to quartz appears to be poorly represented in the
182 L. Wensaas et al.

selected test material (according to unpublished Petroleum Geology. Constructive criticism from V.
p e t r o g r a p h i c a l data). This implies that the Feeser and F. Madsen on an earlier version of the
instability calculations based on the selected core manuscript is much appreciated.
specimens are probably only representative of the
most fine-grained and smectite-rich formations (i.e. REFERENCES
'worst case scenario').
Aadn~y B.S. ( 1988) Modelling of the stability of highly
inclined boreholes in anisotropic rock formations.
CONCLUSIONS SPE Drilling Engineering 3, 259-268.
Aadnoy B.S. & Chenevert M.E. (1987) Stability of
(1) Triaxial tests of side-wall cores have been highly inclined boreholes. SPE Drilling Engineering
shown to represent a suitable method for determina- 2, 364-374.
tion of the mechanical properties and stress-strain Bishop A.W. (1973) The influence of an undrained
behaviour o f weak, smectite-rich shales and change in stress on the pore pressure in porous media
mudrocks. No tests were performed on hard of low compressibility. Geotechnique, 23, 435-442.
(cemented) or silty Tertiary shales. Bol G.M., Wong S.W., Davidson C.J. & Woodland D.C.
(2) The rock mechanical properties of the (1994) Borehole stability in shales. SPE Drilling &
Completion, 87-94.
Tertiary mudrocks from Well 25/7-2 identify them
Breckels I.M. & van Eekelen H.A.M. (1982)
as weak rocks with low values of the strength
Relationship between horizontal stress and depth in
parameters c ' (cohesion) and ~)' (angle of friction). sedimentary basins. J. Petr. Tech. September, 1982,
Within the test conditions applied, the mudrocks 2191-2199.
exhibit a ductile (non-brittle) behaviour. Breton J.R. & McLaurine H.C. (1993) Modified poly-
(3) The measured rock strength increases almost amino acid hydration suppressant proves successful
linearly with increasing confining stress. in controlling reactive shales. SPE Paper 26327,
(4) Compared with the results from onshore Proc. SPE flnn. Tech. Conf Exh., Drilling, Houston,
analogues, the strength properties of the investi- Texas.
gated mudrocks are most likely to be related to Chenevert M.E. (1989) Glycerol mud additive provides
shale stability. Oil Gas J. July 1989, 60-64.
their content of expandable clays.
Chenevert M.E. & Sharma A.K. 0 9 9 l ) Permeability
(5) Wellbore stability calculations based on a and effective pore pressure in shales. Paper IADC/
linear peak strength criterion (Mohr-Coulomb) and SPE 21918.
estimated field parameters (stresses and pore-fluid Ewy R.T., Ross G.D., Gast M.R. & Steiger R.P. (1994)
pressure) were used to calculate the drilling-fluid North Sea case histories of wellbore stability
pressure required to prevent compressional shear and predictions for successful high-angle Nelson Field
tensile failure of the borehole wall. The predicted wells. IADC/SPE 27496 paper, 1994 IADC/SPE
instability generally shows a good correlation with Drilling Conference, Dallas, 1994, 589-598.
reported drilling problems. In forecasting, the upper Fj~er E., Holt R.M., Horsmd P., Raaen A.M. & Risnes R.
and lower critical drilling fluid pressures appear to be (1992) Petroleum Related Rock Mechanics.
very sensitive to variations in the assessment of the Developments in Petroleum Science 33. Elsevier,
Amsterdam.
horizontal minimum stress trend. Overall, the stress
Gaarenstroom L., Tromp R.A.J, de Jong M.C. &
conditions should be better defined because
Brandenburg A.M. (1993) Overpressures in the
instability models are highly affected by the accuracy Central North Sea: implications for trap integrity
of field stress data. and drilling safety. Pp. 1305-1314 in: Petroleum
Geology of the Northwest Europe: Proc. 4th Conf.
(LR. Parker, editor). Geological Society, London.
ACKNOWLEDGMENTS
Head K.H. (1986) Manual of Soil Laboratory Testing 3:
We would like to thank the participants in licence PL Effective Stress Tests. Pentech Press, London.
103 (Conoco, Statoil and Amerada Hess) for allowing Isaksen D. & Tonstad K. (1989) A revised Cretaceous
us to publish these results. The authors acknowledge and Tertiary Iithostratigraphic nomenclature for the
the financial support provided by the Conoco Norway Norwegian North Sea. Norwegian Petroleum
INC. project CNRD 25-6 on Tertiary Claystones on the Directorate Bulletin No. 5.
Norwegian Shelf. This study is a contribution to the Janbu N. (1985) Soil models in offshore engineering,
joint University of Oslo/University of Trondheim R&D 1985. Rankine Lecture. Geotechnique, 35, 241-281.
Programme on Clays Claystones and Shales in Jordt H., Faleide J.I., Bjorlykke K. & lbrahim M.I.
 Mechanical properties of mudrocks 183

(1995) Cenozoic sequence stratigraphy of the central Foumaintraux, editors). Balkema, Rotterdam.
and northern North Sea Basin: tectonic development, Skempton A.W. (1954) The pore pressure coefficients A
sediment distribution and provenance areas. Mar. and B. Geotechnique, 4, 143-147.
Petrol. Geol. 12, 845-879. Steiger R.P. (1982) Fundamentals and the use of
Kwakwa K.A., Batchelor A.S. & Clarc R.C. (1991) potassium/polymer drilling fluids to minimize dril-
Assessment of the mechanical stability of high-angle ling and completion problems associated with
wells in block 22/11, Nelson Field discovery. SPE hydratable shales. JPT, 1661- 1670.
Drilling Engineering, 25-30. Steiger R.P. & Leung P.K. (1988) Quantitative determi-
Marsden J.R., Wu B., Hudson J.A. & Archer J.S. (1989) nation of the mechanical properties of shales. SPE
Investigations of peak rock strength behaviour for Paper 18024, SPE Ann. Tech. Conf. Exh., 1988.
wellbore stability prediction. Pp. 753-760 in: Rock Steiger R.P. & Leung P.K. (1990) Lecture: Predictions
at Great Depth (V. Maury & D. Foumaintraux, of wellbore stability and shale formations at great
editors). Balkema, Rotterdam. depths. Pp. 1209-1218 in: Rock at Great Depth (V.
Marsden J.R., Holt R.M., Nakken S.J. & Raaen A.M. Maury & D. Foumaintraux, editors). Balkema,
(1992) Mechanical and petrophysical characteriza- Rotterdam.
tion of highly stressed mudstones. Pp. 51-56 in: Steiger R.P. & Leung P.K. (1992) Quantitative determi-
Rock Characterization, Proc. ISRM Conf. Europe 92 nation of the mechanical properties of shales. SPE
(J.A. Hudson, editor). Chester, England. Drilling Engineering, 181 185.
McLean M.R. & Addis M.A. (1990a). Wellbore stability Tyridal D.S. (1994) Litologisk og mineralogisk sam-
analysis: A review of current methods of analysis mensetning av slamsteiner i relasjon til loggrespons.
and their field application. IADC/SPE Paper 19941, Cand. Scient. (MSc) Thesis, Univ. Oslo, Norway.
IADC/SPE Drilling Conf., Houston, Texas, 1990. Walker G.R. (1961) Vermiculite minerals. Pp. 279-324
McLean M.R. & Addis M.A. (1990b) Wellbore stability: in: The X-ray Identification and Crystal Structures of
The effect of strength criteria on mud weight Clay Minerals (G. Brown, editor). Mineralogical
recommendations. SPE Paper 19941, 65th Ann. Society, London.
Tech. Conf. Exhibition Soc. Petrol. Engineers. New Wang C., Mao N. & Wu F.T. (1980) Mechanical
Orleans, 1990. properties of clays at high pressure. Geophys. Res.
Mody F.K. & Hale A.H. (1993) Borehole stability model 85, 1462-1468.
to couple the mechanics and chemistry of drilling Warpinski N.R. (1992) Rock mechanic issues in
fluid shale interaction. SPE/IADC 25728 paper 1993 completion and stimulation operations. Proc. 33rd
SPE/IADC Conf. Amsterdam, 1993, 473-490. U.S. Symposium, Rock Mechanics, 375-386.
Nakken S.J., Christensen T.L., Marsden J.R. & Holt Wensaas L., Shaw H.F., Gibbons K., Aagaard P. &
R.M. (1989) Mechanical behaviour of clays at high Dypvik H. (1994) Nature and causes of overpressur-
stress levels for wellbore stability applications. Pp. ing in mudrocks of the Gulifaks area, North Sea.
141-148 in: Rock at Great Depth (V. Maury & D. Clay Miner., 29, 439-449.