Using
Using
Using
Walt Aldred
John Cook
Cambridge, England
Peter Bern
BP Exploration Operating Company Ltd.
Sunbury on Thames, England
Bill Carpenter
Mark Hutchinson
John Lovell
Iain Rezmer-Cooper
Sugar Land, Texas, USA
Pearl Chu Leder
Houston, Texas
40
Oilfield Review
Pressure
Winter 1998
Fracture
gradient
Pore
pressure
Annular
pressure
Depth
Kick
tolerance
2.4
Fracture
gradient
2.0
Stable
1.6
Mud weight, sg
Collapse
gradient
1.2
0.8
0.4
0.0
0
20
40
60
Well deviation, degrees
80
Wellbore Stability
Successful drilling requires that the drilling fluid
pressure stay within a tight mud-weight window
defined by the pressure limits for wellbore stability. The lower pressure limit is either the pore
pressure in the formation or the limit for avoiding
wellbore collapse (above). Normal burial trends
lead to hydrostatically pressured formations,
where the pore pressure is equal to that of a
water column of equal depth. If the drilling fluid
pressure is less than the pore pressure, then formation fluid or gas could flow into the borehole,
with the subsequent risk of a blowout at surface
or underground.
The upper pressure limit for the drilling fluid
is the minimum that will fracture the formation. If
the drilling fluid exceeds this pressure, there is a
risk of creating or opening fracturesresulting
in lost circulation and a damaged formation. In
the language of drilling engineers, pressures are
often expressed as pressure gradients or equivalent fluid densities. The upper limit of the pressure window is usually called the formation
fracture gradient, and the lower limit is called the
pore pressure, or collapse, gradient.
41
42
CDR tool
Gamma ray
Pressure port
Resistivity
6.5 ft
Oilfield Review
Pore pressureOne ongoing oilfield challenge is determining the pore pressure in shales,
and almost all pore pressure prediction is based
on correlation to other measured properties of
shales. Shales start their life at the surface as
clay-rich muds, and water is expelled from them
as they are buried and subjected to increasing
loading from the overburden above them. If the
burial is sufficiently slow, and there is an escape
route for the water, the pressure in the pore fluid
remains close to hydrostatic, and the overburden
is supported by increased stresses in the solid
parts of the rock. The water content, or porosity,
decreases, and this variation of porosity or other
water-dependent properties with depth is known
as the normal compaction trend.
However, if burial is very rapid, or the fluid
cannot escapebecause of the low permeability
of shalesthe increasing overburden load is supported by the increasing pore pressure of the fluid
itself. The stress in the solid parts of the rock
remains constant, and the water content, or
porosity, does not decrease. After rapid burial, the
shale is not normally compacted; its pore pressure is above hydrostatic, and its water content is
higher than it would be for normallycompacted shale at that depth. The shale
becomes overpressured as a result of undercompaction. Detecting overpressured zones is a major
concern while drilling, because water or gas
influx can lead to a blowout.
Fracture gradientsFracture gradients are
determined from the overburden weight and lateral stresses of the formation at depth and from
local rock properties. Density and sonic logging
data help provide estimates of rock strengths.3
Calculating offshore fracture gradients in deep
water presents a special problem. The uppermost
formations are replaced by a layer of water,
which is obviously less dense than rock. In these
wells, the overburden stress is less than in a
comparable onshore well of similar depth. This
results in lower fracture gradients and, in general, fracture gradients decrease with increased
water depth. Thus, increasing water depth
reduces the size of the margin between the
mud weight required to balance formation
pore pressures and that which will result in
formation breakdown.
Downhole Pressure
After the wellbore stability pressure window has
been determined, the driller has more to do than
keep the drilling fluid within these limits. To correctly interpret the response of a downhole annular pressure measurement, it is important to
appreciate the physical principles upon which it
depends. The downhole annular pressure has
two components. The first is a static pressure
due to the density gradients of the fluids in the
borehole annulusthe weight of the fluid vertically above the pressure sensor. The density of
the mud column including solids (such as cuttings) is called the equivalent static density
(ESD), and the fluid densities are pressure- and
temperature-dependent.
Second is dynamic pressure related to pipe
velocity (swab, surge and drillpipe rotation),
inertial pressures from string acceleration or
deceleration when tripping, excess pressure to
break mud gels, and the cumulative pressure
losses required to move drilling fluids up the
annulus. Flow past constrictions, such as cuttings
beds or swelling formations, changes in hole
geometry, and influxes or effluxes of liquids and
solids to or from the annulus all contribute to the
dynamic pressure. The equivalent circulating
density (ECD) is defined as the effective mud
> Flow regimes. In laminar flow the annular pressure losses decrease with increasing pipe rotation,
because azimuthal stresses reduce the effective viscosity of the drilling fluid. Once the Taylor number
(a condition for rotational flow instability) is exceeded, vortices will be formed, which extract energy
from the mean axial flow, and yield a turbulent-like pressure drop. As the axial flow rate increases,
full turbulence will occur and the axial pressure drop will then increase with increasing rotational
rate. Similarly, increases in the rotation rate can also assist in the transition from laminar to turbulent
flow and can lead to an increase in the axial pressure drop.
Winter 1998
Flow rate
Turbulent
Laminar
Laminar with vortices
Rotation rate
Pressure increasing
43
2500
2000
Low flow
Medium flow
High flow
1500
Experimental conditions
Hole size = 4.9 in.
Pipe outer diameter = 3.5 in.
Plastic viscosity = 3.4 cp
Yield point = 3.8 lbf/100 ft2
1000
500
0
0
50
100
150
200
250
300
350
400
450
500
Rotation, rpm
> Laboratory-measured annular pressure losses. Experimental comparisons of the effect of rotation
on annular pressure losses with a clean drilling fluid (no solids) at low, medium and high flow rates
highlight the flow behavior in different flow regimes. These measurements confirm that under some
conditions rotation acts to increase axial pressure losses, whereas under other conditions it
decreases the losses.
1.13
1.11
600 gal/min
1.09
1.07
1.05
1.03
400 gal/min
1.01
0.99
0.97
200 gal/min
0.95
0
10
50
100
Rotation, rpm
120
130
> Effect of rotation on equivalent circulating density (ECD). In addition to distinct effects on holecleaning efficiency, rotation also affects fluid behavior in an unloaded annulus. In this field experiment
of drillpipe rotation (at different flow rates), the ECD increases from 1.092 sg to 1.114 sg as the rotation
rate increases to 130 rpm with a 1.0 sg mud weight circulating at 600 gal/min [2300 L/min] in an 8-in.
[20-cm] casing section. The increment of 0.022 sg is equivalent to 50 psi [350 kPa]. No cuttings were
in suspension as this test was performed just prior to drilling.
Drillpipe
eccentricity
Cuttings
density
Little
influence
on cuttings
transport
Mud
rheology
Hole size
and angle
Mud
weight
Flow
rate
Rate of
penetration
Drillpipe
rotation
Cuttings
size
Hard to control
44
Hole cleaning.
Some factors are
under the control
of the driller, such
as surface mud
rheology, rate of
penetration (ROP),
flow rate and hole
angle. Others,
including drillpipe
eccentricity, and
cuttings density
and size, cannot be
controlled as easily.
>
Large
influence
on cuttings
transport
Easy to control
Oilfield Review
High ECD
Asymmetric
suspension
Rolling
Stationary
Symmetric
suspension
Low ECD
>
0
Time
Surface rotation
Total pump flow
rpm
gal/min 1000 0
150 0
Surface torque
Standpipe pressure
psi
kft-lbf
30 0
4000 10
Temperature
C
ECD/ESD
lbm/gal
>
200
14
01:00
02:00
>
Effect of drillpipe
rotation on hole
cleaning. The ECD (red
curve), shown in track
6, increasesindicating cuttings are resuspended in the drilling
fluidat 16:15 as
rotation recommences
after a slide-drilling
interval is completed.
Depth
ft
Block
speed
-10 ft/s 10 0
Hookload
klbf
500
Time
Surface torque
kft-lbf
Total
pump flow 0
Surface
rotation
0
1000
gal/min
0
rpm 200
13.2
15:00
Temperature
C
100
Standpipe pressure
5000
psi
ECD / ESD
14.2
lbm/gal
Rotation
stops
Sliding
interval
16:00
Rotation
starts
Winter 1998
Drillpipe rotationAnother example demonstrates the effect of pipe rotation on hole cleaning (above). At 15:00, pipe rotation was stopped
to enable drill-bit steering. The ECD decreased
for 20 minutes as the cuttings fell out of suspension. A few swab-and-surge spikes were
observed. These pressure spikes were introduced
as the pipe was moved up and down to adjust
mud motor orientation. After steering for a total
45
Drilling cycle 1
Block
position
0
m 50
ROP
Surface
weight on bit
Surface torque
200 m/hr 0
klbf
80
Hookload 0
20
kft-lbf 60
0 klbf 400
Downhole
Surface
Downhole
weight on bit
rotation
torque
Bit depth
Time
value, m
kft-lbf 20 0 rpm 200
0
klbf
80 19:00 0
ECD
1.65
sg
1.75
Annulus temperature
Turbine
rotation 0
C
200
Annulus pressure
0 rpm 5000
Total pump 0
psi
4000
flow
Standpipe
pressure
gal/min
psi
0
1500 0
4000
Pumps on
Rotary drilling
Hole swabbed
Hole surged
Slide drilling
Cuttings
settling out
Kelly down
20:00
Pipe
reciprocating
Surge and
swab
Pumps off
Drilling cycle 2
21:00
> Downhole pressure monitoring to improve hole cleaning. Drilling fluid pumps start at 19:05, shown by pump flow rate in track 5. Pipe rotation starts a few
minutes later, seen by the increase in surface rotation rate shown in track 4. The instantaneous increase in standpipe pressure (green curve) and the
delayed downhole ECD (black curve) measurement can be seen in track 6. Rotary drilling stops and slide drilling starts at 19:27, shown by surface rotation
rate (track 4) and weight on bit (track 2). The immediate effect of slide drilling on the downhole ECD (black curve) can be seen in track 6. After Kelly down,
shown by the block position in track 1, the driller starts hole cleaning by reciprocating the pipe in and out. After the hole cleaning is completed, the driller
makes a new connection and starts the next drilling cycle at 20:30.
Improving Efficiency in
Extended-Reach Drilling
BP encountered severe wellbore instability problems when drilling development wells in Mungo
field in the Eastern Trough Area Project (ETAP) of
the North Sea. These instability problems were
due in part to large cavings formed while drilling
the flanks of salt diapirs. Long S-shaped 1214-in.
[31-cm] sections are generally the most problematic. The volume of cavingscoupled with highly
inclined wellbore trajectoriesresults in poor
hole-cleaning conditions. The main cause of the
poor hole cleaning is believed to be the formation
of cuttings and cavings beds on the highly
inclined 60 section. These beds are manageable
while drilling, but present a major hazard when
46
Oilfield Review
Winter 1998
Block position
Hookload
50 0
m
500
klbf
Surface torque
ROP
0
50
kft-lbf
100 0
m/hr
ECD
Standpipe pressure
3000
4000 1.2
sg
1.3
psi
Annulus pressure
Total pump flow
2000
psi
3000 0
gal/min 2000
12:00
Drilling
break
1:00
4:00
5:00
Standpipe
pressure
increase
5:00
ECD
increase
> Preventing packoff events. The ECD, shown in track 4, risesdue to cuttings accumulation entering
the vertical section of an extended-reach wellabout five hours after a drilling break.
6. Allen F, Tooms P, Conran G and Lesso B: ExtendedReach Drilling: Breaking the 10-km Barrier, Oilfield
Review 9, no. 4 (Winter 1997): 32-47.
47
48
Slim hole
20
5000
12
4800
8
16
4600
Shut-in Kill
Shut-in Kill
800
2000
600
400
1000
200
0
0
10
20
Time, min
40 0
30
Pit gain
10
Standpipe pressure
20
Time, min
40
30
Typical hole
Kick Detection
The influx of another fluid into the wellbore due
to unexpected high formation pressure is one of
the most serious risks during drilling. The character of the fluid influx will depend primarily
upon influx fluid density, rate and volume,
drilling fluid properties and both borehole and
drillstring geometry (right). Simulations performed by The Anadrill SideKick software model
are frequently used to understand the pressure
responses expected downhole and at the surface due to gas influxes. (see Simulating Gas
Kicks, page 50).7 During gas kicks, ECD
responses for typical boreholes and slim wellbore geometries are dominated by two phenomenareduced density of the mud column as
heavier drilling fluid is replaced by less dense
gas, and increased annular pressure loss due to
friction and inertia when accelerating the mud
column above the gas influx.
The reduced annular gap in slimhole wells
can cause unique drilling problems.8 For example,
in slim holes the acceleration of the kick fluid
into the wellbore can lead to a sudden increase
in frictional pressure loss in the annulus due to
acceleration of the mud ahead of the kick fluid. In
addition, evidence of the influx may not be seen
until the pumps are shut down. In typical hole
sizes, the hydrostatic imbalance between the
drillpipe and the annulus outweighs any frictional
losses, and a decrease in the bottomhole annular
pressure is evident.
Constant monitoring of all available drilling
data is critical in detecting a downhole kick
event. In an example of a gas kick, an operator
was drilling a 1214-in. hole section in a well in
the Eugene Island field in the Gulf of Mexico
(next page). The formations were sequences of
shales and target sands, and several of the
sands were likely to be depleted by previous
production. In offset wells, the low-pressure
sands led to problems including stuck pipe,
twist-offs and stuck logging tools.
Maintaining a minimum mud weight was
required to avoid differential sticking in the
depleted sands. Due to faulting in the area, zonal
communication was uncertain and the pore pressure limits were difficult to anticipate. Anadrill
was using the CDR Compensated Dual Resistivity
tool for formation resistivity and the Multiaxis
Vibrational Cartridge (MVC), Integrated Weighton-Bit (IWOB) tool and APWD sensors for monitoring drilling performance. The plan was to set a
liner below a normally pressured zone before
drilling into the underpressured sand beds.
Annulus pressure
> Kick detection. In a typical wellbore geometry (top left), the annular pressure (orange curve) can be
seen to decrease as the displacement of heavier drilling fluids by a gas influx dominates the pressure
response. For slimhole geometry (top right) the annular pressure (orange curve) can increase initially
during a gas influx as the inertia of the mud column dominates the response. One major benefit of
downhole annular pressure monitoring is early kick detection. Mud-pit gain (red curves in upper plots),
standpipe pressure (green curves in lower plots), and frictional pressure loss (yellow curves in lower
plots) help the driller identify gas kicks.
Annulus
temperature
Block height
0
ft
120 200
F 300
Time
08:00
13
ECD
lbm/gal
Standpipe
pressure
18 3000 psi 5000
09:00
Rack back
stand of pipe
10:00
Temperature rises,
ECD drops
11:00
Flow check
and close in
12:00
> Gas influx. When gas mixes with drilling fluid, the density of the drilling fluid decreases. Fifty minutes
after the ECD (blue curve), shown in track 3, started to decrease, a flow check confirmed that a small
gas influx had occurred. Note the increase in annular temperature, shown in track 2, as the formation
fluid warmed the borehole.
Oilfield Review
Winter 1998
Annulus temperature
100
300
Standpipe pressure
Block speed
-2
ft/s
Axial vibration
2 4
ROP
500 ft/hr
Bit depth
0
ft
klbf
4000
klbf
60
60
ft-lbf
lbm/gal
Time
kft-lbf
25
Downhole torque
kft-lbf
5000
ECD
0
Surface torque
100 0
psi
Torsional vibration
G
Surface weight on bit
0 0
8 0
Total
pump
flow
gal/min
1500
11
psi
10000
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
> Kick alert in the Gulf of Mexico. A sudden increase in the rate of penetration (ROP) (blue curve),
shown in track 1, at 17:10 alerted the driller that the bit had entered a sand zone and that an influx
was possible. Drilling restarted after having seen no evidence of flow in the mud-flow measurements
or pit volume. However, as drilling progressed into the sand zone, the ECD (pink curve), shown in
track 5, started to decrease slowly at 18:10 and continued until 19:20. At this time, the rate of decrease
suddenly increased. After drilling ahead for 30 minutes with rapidly decreasing ECD and increasing
pit volume, the driller recognized that an influx had occurred and the well was shut in.
49
1000
Static
Circulating
900
Unsafe
800
700
Safe
600
500
400
0
10
20
Pit gain, bbl
30
40
> SideKick-PC kick tolerance. The SideKick-PC program computes separate kick tolerances for the shut-in
and kill periods of a simulation. The kick tolerance plot is used to differentiate kicks that can be safely
shut in (static) from those that can be safely killed (circulating). The determination depends on many
factors such as pressures in the well, gas migration, circulating friction and kill-mud hydrostatic pressure.
Kicks in the region to the left and below each curve are considered safe, and those severe enough to be in
the region above and to the right of each curve may cause lost circulation.
50
Oilfield Review
Casing, in.
Deep-Water Wells
Unconsolidated sediments typically encountered in deep-water formations tighten the
wellbore stability window between pore pressure and formation fracture pressure. At a given
depth, fracture gradient decreases with increasing water depth, and can result in a very narrow
pressure margin.9
Additionally, cooling of the mud in the deepwater riser can cause higher mud viscosity,
increased gel strength, and high frictional
pressure losses in choke and kill lines during
well-control procedures. Combined, these factors increase the likelihood of lost-circulation
problems, and drilling engineers must take
appropriate steps to avoid exceeding formation
fracture gradients.
Staying within the pressure window
Keeping the ECD within the pressure window is
a constant struggle, especially in deep water and
HPHT applications. In a well in the Gulf of
Mexico, EEX Corporation experienced a kick
while drilling at near-balance conditions in Zone
A (right). After the kick was taken and the well
was under control, increased mud weight was
needed to continue safely. A 13 38-in. [34-cm] casing string was set because the heavier mud
weight exceeded the previous leakoff test.
The next two hole sections were drilled
without incident. However, as drilling proceeded deeper into the third section, the
increasing pore pressure eventually approached
the pressure exerted by the heavier mud and
another kick was experienced in Zone B. A
958-in. [24-cm] casing was needed to permit
another increase in mud weight. As drilling continued, increases in the cuttings load caused the
mud pressure to exceed the overburden pressure in Zone C, resulting in some lost circulation
over a period of several days. Lost-circulation
material helped minimize mud losses, and
drilling continued successfully thereafter. At the
narrowest point shown in this example, the
pressure window was only 700 psi [4827 kPa].
Dynamic kill procedureReal-time analysis
of downhole annular pressure helped BP
Exploration monitor a dynamic kill procedure
used to stop an underground flow in a deepwater well in the Gulf of Mexico. Drilling unexpectedly entered a high-pressure zone, where a
Winter 1998
20
16
Zone A
Kick
133/8
113/4
Zone B
Kick
95/8
Zone C
75/8
10.00
17.00
10.00
17.00
10.00
ECD, lbm/gal
17.00
10.00
17.00
> Staying within the pressure window. A gas kick was observed in Zone A, where the ECD (blue curve)
dropped significantly below the pore pressure gradientestimated from resistivity logs (red curve) or
seismic time-to-depth conversions (black curve). The well was brought under control with an increase
in mud weightshown by the increased ECD. However, a second kick was experienced in Zone B
as pore pressure again increased above the ECD in this deeper section of the well. After another
increase in mud weight, some mud losses were experienced in Zone C, where the ECD increased
slightly above the overburden gradient (purple curve).
51
Without riser
With riser
water
mud
Sand
mud
Sand
> Riserless operations. During typical offshore drilling (left), drilling mud is circulated
through a riser back to the surface and the APWD tool measures an average ECD for the
entire interval. During riserless operations (right), the pumped drilling fluid does not return
to the surface, but rather carries its drilling solids only as far as the seafloor.
ECD
12
Block speed
2
500
ft/s
ROP
ft/hr
ft
Standpipe pressure
psi
4000
Annulus temperature
0
Bit depth
0
2000
13
lbm/gal
50
Hookload
100 400
klbf
Time
600 18:00
Surface torque
0
kft-lbf
F
Annulus pressure
150
psi
9000
Kill starts
19:00
Kill stops
> Monitoring dynamic kill procedure. A water influx was encountered in a Gulf of Mexico deep-water well that was strong enough to fracture the
casing shoe, resulting in an underground flow. In track 6, both the standpipe pressure (green curve) and downhole annulus pressure (purple curve)
showed a steady increase at 18:30 while the kill mud was being circulated in the wellbore.
52
Oilfield Review
Winter 1998
500
0
Rate of penetration
ft/hr
0
0
Gamma ray
Depth
m
150
API
0
Annulus temperature
Phase-shift resistivity
ohm-m
F
2 50
100
Phase-shift resistivity
Annulus pressure
ohm-m
psi
10 2000
3000
Attenuation resistivity
ECD
ohm-m
lbm/gal
10 8
9
X000
Water influx
X100
B-upper
B-lower
X200
Water influx
X300
X400
C
X500
X600
X700
X800
D
X900
Water influx
> Shallow water flow in a deep-water well. Sand zones at A, B, C and D are indicated by decreasing
gamma ray (pink curve), shown in track 1, and resistivity responses shown in track 2. Increasing
annular pressure (green curve) and ECD (blue curve), shown in track 3, indicate that a water influx
occurred in three of these sands.
53
Pumping-up
phase
Leakoff
phase
600
B
Leakoff test
520
500
480
460
445
435
430
430
400
420 415
410 408 405
400
350
Formation taking
drilling fluid
300
260
200
165
100
80
0
1
2
3
Volume, bbl
1 2 3 4
5 6 7 8 9 10
Time, min
Surface
rotation
rpm
Total pump
flow
9
gal/min
0
1500 0
ECD
12
lbm/gal
Annulus pressure
psi
10000
15:00
16:00
> Leakoff testing. A leakoff test was conducted in a deep-water well in the Gulf of Mexico. During
the pumping-up phase, the standpipe pressure increases linearly as the pump volume increases
(top). At point A, the formation fractures and starts to take on some of the drilling mud. After the
pumping stops at point B, the standpipe pressure decreases rapidly at first, then more slowly as the
formation fractures close. The ECD log (bottom) from the APWD measurements, shown in track 6,
increases from the hydrostatic pressure to 10.9 lbm/gal [1.31 g/cm3] during the pump-up phase. After
pumping stops, the pressure starts to fall, and the ECD drops back.
54
Oilfield Review
Event or procedure
ECD change
Other indications
Comments
Mud gelation /
pump startup
Sudden increase
possible
Cuttings pick-up
Cuttings at surface
Plugging annulus
Intermittent surge
increases
Standpipe pressure
Surge increase?
Torque/RPM fluctuations
High overpulls
Packoff may
blow-through
before formation
breakdown
Gradual increase
Sudden increase as
packoff passes sensor
none if packoff remains
below sensor
High overpulls
Steady increase in
standpipe pressure
Gas migration
Increase if well is
shut-in
Running in hole
Increase magnitude
dependent on gap,
rheology, speed, etc.
Effect enhanced if
nozzles plugged
Decrease magnitude
dependent on gap,
rheology, speed, etc.
Effect enhanced if
nozzles plugged
Making a connection
Decrease to static
mud density
Pumps on/off
indicator
Pump flow rate lag
Barite sag
Decrease in static
mud density or
unexplained density
fluctuations
Gas influx
Decreases in
typical size hole
Initial increase in
pit gain may be masked
Liquid influx
Decreases if lighter
than drilling fluid
Increases if influx
accompanied by solids
> Interpretation guide. Monitoring ECD with downhole annular pressure measurements along with
other drilling parameters helps the operator know what is happening downhole in the wellbore. Some
of the known, clearly identifiable, and repeatable signatures of ECD changes are shown along with
secondary or confirming indications, such as those seen in surface measurements.
Winter 1998
55