You are on page 1of 20

Fluid Dynamics in Well Control 251

Summary
Surface
Old Pressure after Surface
Volume Surface Adding Kill Pressure after
Pumped Pressure Mud Bleeding
Initial -0- 1420
Conditions
1st Stage 20 1420 2031 966
2nd Stage 19 966 1633 513
3rd Stage 16 513 1227 120
4th Stage 10 120 920 -0-
By following this schedule, the well can be killed safely without
violating the casing seat and losing more mud and without permitting any
additional infux of formation gas into the wellbore.
In summary, the well is controlled by pumping kill-weight mud into
the annulus and bleeding dry gas out of the annulus. The volume of kill
fuid that can be pumped without fracturing the casing seat is determined
from Equation 5.3. Logically, the surface pressure should be reduced from
the value prior to lubricating kill mud by the additional hydrostatic con-
tributed by the kill-weight mud. However, for obvious reasons, the effective
hydrostatic cannot be less than the reservoir pressure. At the lower surface
pressures, the gas hydrostatic gradient is less and the fnal surface pressure
must be higher to refect the difference in gas gradient and prevent addi-
tional infux. The analysis is continued until the well is dead. Rudimentary
analysis suggests higher mud weights for well control.
DYNAMIC KILL OPERATIONS
Simply put, to kill a well dynamically is to use frictional pressure
losses to control the fowing bottomhole pressure and, ultimately, the static
bottomhole pressure of the blowout.
Dynamic Kill implies the use of a kill fuid whose density results
in a hydrostatic column which is less than the static reservoir pressure.
Therefore, the frictional pressure loss of the kill fuid is required to stop
252 Blowout and Well Control Handbook
the fow of reservoir fuids. Dynamic Kill in the purest sense was intended
as an intermediate step in the well control procedure. Procedurally, after
the blowout was dynamically controlled with a fuid of lesser density, it
was ultimately controlled with a fuid of greater density, which resulted in
a hydrostatic greater than the reservoir pressure.
Generically and in this text, Dynamic Kill includes control proce-
dures utilizing fuids with densities less than and much greater than that
required to balance the static reservoir pressure. In reality, when the den-
sity of the kill fuid is equal to or greater than that required to balance the
static reservoir pressure, the fuid dynamics are more properly described
as a Multiphase Kill Procedure.
Dynamically controlling a well using a Multiphase Kill Procedure
is one of the oldest and most widely used fuid control operations. In the
past, it has been a 'seat-of-the-pants¨ operation with little or no technical
evaluation. Most of the time, well control specialists had some arbitrary
rules of thumb, for example, the kill fuid had to be 2 pounds per gallon
heavier than the mud used to drill the zone, or the kill fuid had to achieve
some particular annular velocity. Usually, that translated into manifolding
together all the pumps in captivity, weighting the mud up as high as possi-
ble, pumping like hell, and hoping for the best. Sometimes it worked and
sometimes it didn`t.
There are many applications of the Dynamic and Multiphase Kill
procedures and any well control operation should be studied from that
perspective. The most common application is when the well is fowing out
of control and the drillpipe is on bottom. The kill fuid is pumped to the
bottom of the hole through the drillpipe, and the additional hydrostatic of
the kill fuid along with the increased friction pressure resulting from the
kill fuid controls the well.
Multiphase Kill operations were routine in the Arkoma Basin in
the early 1960s. There, air drilling was popular. In air drilling operations,
every productive well is a blowout. Under different circumstances, some of
themwould have made the front page of the NewYork Times. At one location
just east of McCurtain, Oklahoma, the fre was coming out of an 8-inch
blooie line and was almost as high as the crown of the 141-foot derrick.
The blooie line was 300 feet long, and you could toast marshmallows on
the rig foor. The usual procedure was the seat-of-the-pants Multiphase Kill
and it killed the well.
Fluid Dynamics in Well Control 253
The most defnitive work on the pure Dynamic Kill was done by
Mobil Oil Corporation and reported by Elmo Blount and Edy Soeiinah.
2
The biggest gas feld in the world is Mobil`s Arun Field in North Sumatra,
Indonesia. On June 4, 1978, Well No. C-II-2 blew out while drilling
and caught fre. The rig was immediately consumed. The well burned
for 89 days at an approximate rate of 400 million standard cubic feet
per day.
Due to the well`s high deliverability and potential, it was expected
to be extremely diffcult to kill. The engineering was so precise that
only one relief well was required. That a blowout of this magnitude
was completely dead one hour and 50 minutes after pumping operations
commenced is a tribute to all involved. One of the most signifcant contri-
butions resulting from this job was the insight into the fuid dynamics of a
Dynamic Kill.
The engineering concepts of a Dynamic Kill are best understood
by considering the familiar U-tube of Figure 5.6. The left side of the U-tube
may represent a relief well and the right side a blowout, as in the case just
discussed, or the left side may represent drillpipe while the right side would
correspond to the annulus, as is the case in many well control situations.
The connecting interval may be the formation in the case of a relief well
with the valve representing the resistance due to the fow of fuids through
the formation. In the case of the drillpipe-annulus scenario, the valve may
represent the friction in the drill string or the nozzles in the bit, or the
situation may be completely different. Whatever the situation, the technical
concepts are basically the same.
The formation is represented as fowing up the right side of the
U-tube with a fowing bottomhole pressure, T
ßow
, which is given by the
following equation:
T
ßow
 T
a

fr

h
(5.11)
Where:
T
a
 Surface pressure, psi
fr
 Frictional pressure, psi
h
 Hydrostatic pressure, psi
The capability of the formation to deliver hydrocarbons to the
wellbore is governed by the familiar back-pressure curve illustrated in
254 Blowout and Well Control Handbook
Q
P
flow
P
fr
P
h
P
s
F/gare 5.6 Dynamic Kill Schematic.
Figure 5.7 and described by the Equation 5.12:
Q  C(T
2
b
 T
2
ßow
)
n
(5.12)
Where:
Q  Flow rate, mmscfpd
Fluid Dynamics in Well Control 255
Log Q
F/gare 5.7 Typical Open Flow Potential Curve.
T
b
 Formation pore pressure, psi
T
ßow
 Flowing bottmhole pressure, psi
C  Constant
n  Slope of back-pressure curve
 0.5 for turbulence
 1.0 for laminar fow
Finally, the reaction of the formation to an increase in fowing
bottomhole pressure, P
ßow
, is depicted by the classic Horner Plot, illus-
trated in Figure 5.8. The problem is to model the blowout considering
these variables.
In the past, shortcuts have been taken for the sake of simplicity.
For example, the most simple approach is to design a kill fuid and rate
such that the frictional pressure loss plus the hydrostatic is greater than the
shut-in bottomhole pressure, P
b
. This rate would be that which is suffcient
to maintain control. Equations 4.13 through 4.14 for frictional pressure
losses in turbulent fow can be used in such an analysis.
256 Blowout and Well Control Handbook
Log T
P
f
P
b
F/gare 5.8 Typical Pressure Build Up Curve.
Consider the Example 5.4:
Example 5.4
Given:
Depth, D  10,000 feet
Kill fuid,
1
 8.33 ppg
Kill-fuid gradient,
m1
 0.433 psi/ft
Bottomhole pressure, T
b
 5200 psi
Inside diameter of pipe, D
i
 4.408 inches
Surface pressure, T
a
 atmospheric
Figure 5.6
Fluid Dynamics in Well Control 257
Required:
Determine the rate required to kill the well dynamically with water.
Solution:
T
ßow
 T
a

fr

h
h
 0.433(10,000)
h
 4330 psi
fr
 5200  4330
fr
 870 psi
Rearranging Equation 4.14 where
fr
 T
fti
:
Q 
T
fti
D
4.8
i
7.7(10
5 .8
1
TV
.2
/
1
1.8
Q 
(870)(4.408)
4.8
7.7(10
5
)(8.33)
.8
(1)
.2
(10,000)
1
1.8
Q  1011 gpm
Q  24.1 bpm
Therefore, as illustrated in Example 5.4, the well would be dyna-
mically controlled by pumping fresh water through the pipe at 24 bpm. The
Dynamic Kill Procedure would be complete when the water was followed
with kill mud of suffcient density (10 ppg in this example) to control the
bottomhole pressure.
The rate required to maintain control is insuffcient in most
instances to achieve control and is considered to be the minimum rate for
a Dynamic Kill operation. Kouba, et al. have suggested that the rate suf-
fcient to maintain control is the minimum and that the maximum rate is
258 Blowout and Well Control Handbook
approximated by Equation 5.13:
3
Q
k max
 A
2g
c
D
tvd
D
h
fD
md
1
2
(5.13)
Where:
g
c
 Gravitational constant,
/b
m
 ft
/b
f
 sec
2
A  Cross sectional area, ft
2
D
h
 Hydraulic diameter, feet
D
tvd
 Vertical well height, feet
D
md
 Measured well length, feet
¡  Moody friction factor, dimensionless
Consider Example 5.5:
Example 5.5
Given:
Same conditions as Example 5.4
Required:
Calculate the maximum kill rate using Equation 5.13.
Solution:
Solving Equation 5.13 gives
Q
k max
 A
2g
c
D
tvd
D
h
¡ D
md
1
2
Q
k max
 0.106
2(32.2)(10,000)(.3673)
.019(10,000)
1
2
Q
k max
 3.74
ft
3
sec
Q
k max
 40 bpm
As illustrated in Examples 5.4 and 5.5, the minimum kill rate is
24 bpm and the maximum kill rate is 40 bpm. The rate required to kill the
Fluid Dynamics in Well Control 259
well is somewhere between these values and is very diffcult to determine.
Multiphase fow analysis is required. The methods of multiphase fow
analysis are very complex and based upon empirical correlations obtained
fromlaboratory research. The available correlations and research are based
upon gas-lift models describing the fow of gas, oil, and water inside small
pipes. Precious little research has been done to describe annular fow, much
less the multiphase relationship between gas, oil, drilling mud, and water
fowing up a very large, inclined annulus. The conditions and boundaries
describing most blowouts are very complex to be described by currently
available multiphase models. It is beyond the scope of this work to offer
an in-depth discussion of multiphase models.
Further complicating the problemis the fact that, in most instances,
the productive interval does not react instantaneously as would be implied
by the strict interpretation of Figure 5.7. Actual reservoir response is illus-
trated by the classical Horner Plot illustrated in Figure 5.8. As illustrated in
Figure 5.8, the response by the reservoir to the introduction of a kill fuid
is non-linear.
For example, the multiphase frictional pressure loss (represented
by Figure 5.8) initially required to control the well is not that which will
control the static reservoir pressure. The multiphase frictional pressure loss
required to control the well is that which will control the fowing bottom-
hole pressure. The fowing bottomhole pressure may be much less than the
static bottomhole pressure.
Further, several minutes to several hours may be required for the
reservoir to stabilize at the reservoir pressure. Unfortunately, much of the
data needed to understand the productive capabilities of the reservoir in a
particular wellbore are not available until after the blowout is controlled.
However, data from similar offset wells can be considered.
Consider the well control operation at the Williford Energy Com-
pany Rainwater No. 2-14 in Pope County near Russellville, Arkansas.
4
The wellbore schematic is presented as Figure 5.9. A high-volume gas
zone had been penetrated at 4620 feet. On the trip out of the hole, the
well kicked. Mechanical problems prevented the well from being shut in
and it was soon fowing in excess of 20 mmscfpd through the rotary table.
The drillpipe was stripped to bottom and the well was diverted through the
choke manifold. By pitot tube, the well was determined to be fowing at a
rate of 34.9 mmscfpd with a manifold pressure of 150 psig. The wellbore
260 Blowout and Well Control Handbook
Casing Set at 560 feet
4 -inch 16.60 #/ft
End of Drillpipe at 4150 feet
Drill Collars - 16 6 inch 2 inch
480 feet
Bit at 4630 feet
TD 4700 feet
8 -inch hole
4012 feet
1
2
1
4
1
2
7 -inch hole
7
8
F/gare 5.9 Williford Energy Company, Rainwater No. 1, Pope County,
Arkansas.
schematic, Open Flow Potential Test, and Horner Plot are presented as
Figures 5.9, 5.10 and 5.11, respectively.
In this instance, the Orkiszewski method was modifed and utilized
to predict the multiphase behavior.
5
The well was successfully controlled.
Fluid Dynamics in Well Control 261
VOLUME RATE OF FLOW, Q (1000) MSCFPD
COFP
36
MMSCFPD
100
4000
1000
100 10 1
F/gare 5.10 Calculated Open Flow Potential Test, Williford Energy
Company-Rainwater No. 2-14.
The technique used in the Williford Energy example is very
conservative in that it determines the multiphase kill rate required to con-
trol the shut-in bottomhole pressure of 1995 psi. Analysis of Figure 5.11
indicates that the static bottomhole pressure will not be reached in the
blowout for more than 100 hours. The fowing bottomhole pressure when
the kill procedure begins is only 434 psi and is only 1500 psi approximately
20 minutes after the fowing bottomhole pressure has been exceeded. Of
course, in this case the kill operation is fnished in just over 10 minutes.
Including all these variables is more complex but well within the capa-
bilities of modern computing technology. Based on this more complex
analysis, the kill rate using 10.7-ppg mud was determined to be 10 bpm.
The model is further complicated by the fact that the Calculated
Open Flow Potential curve (COFP) presented as Figure 5.10 is a very
optimistic evaluation of sustained productive capacity. An Actual Open
FlowPotential curve (AOFP) based on sustained production would be more
262 Blowout and Well Control Handbook
2000
1750
1500
1250
1000
750
500
250
0
.1 1 10 100
Elapse time - hours
F/gare 5.11 Bottomhole Pressure Build Up Test, Williford Energy
Company-Rainwater No. 2-14.
appropriate for modeling actual kill requirements. A kill procedure based
on the COFP is the more conservative approach. An AOFP curve more
accurately refects the effect of the pressure draw down in the reservoir in
the vicinity of the wellbore.
THE MOMENTUM KILL
The Momentum Kill is a procedure where two fuids collide and
the one with the greater momentumwins. If the greater momentumbelongs
to the fuid fromthe blowout, the blowout continues. If the greater momen-
tum belongs to the kill fuid, the well is controlled. The technology of the
Momentum Kill procedure is the newest and least understood of well con-
trol procedures. However, the technique itself is not new. In the late ffties
and early sixties, the air drillers of eastern Oklahoma thought nothing of
pulling into the surface pipe to mud up an air-drilled hole in an effort to
avoid the hazards associated with the introduction of mud to the Atoka
Shale.
Momentum Kill concepts are best illustrated by Figures 5.12 and
5.13. Figure 5.12illustrates a situation in which the outcome would never be
Fluid Dynamics in Well Control 263
F/gare 5.12
F/gare 5.13
in doubt. The most fundamental reasoning would suggest that the occupant
of the car is in greater peril than the occupant of the truck. It is most likely
that the momentum of the truck will prevail and the direction of the car
will be reversed. Conceptually, fuid dynamics are not as easy. However,
consider the men in Figure 5.13. They have forgotten the fre and turned
their attention to each other. Obviously, the one with the least momentum
is destined for a bath.
The dynamics of a blowout are very much the same as those illus-
trated in Figure 5.13. The fuid fowing from the blowout exhibits a
defnable quantity of momentum. Therefore, if the kill fuid is introduced
at a greater momentum, the fow from the blowout is reversed when the
264 Blowout and Well Control Handbook
fuids collide. The governing physical principles are not signifcantly dif-
ferent from those governing the collision of two trains, two cars, or two
men. The mass with the greatest momentum will win the encounter.
Newton`s Second Law states that the net force acting on a given
mass is proportional to the time rate of change of linear momentum of that
mass. In other words, the net external force acting on the fuid within a
prescribed control volume equals the time rate of change of momentum of
the fuid within the control volume plus the net rate of momentum trans-
port out of the surfaces of the control volume.
Consider the following development with all units being basic:
Momentum:
M 
m:
g
c
(5.14)
and the mass rate of fow:
 vA  (5.15)
and, from the conservation of mass:
vA 
i
:
i
A
i
(5.16)
Where:
m  Mass, lb
m
:  Velocity,
ft
sec
g
c
 Gravitational constant,
lb
m
 ft
lb
f
 sec
2
 Density,
lb
m
ft
3
q  Volume rate of fow,
ft
3
sec
 Mass rate of fow,
lb
m
sec
Fluid Dynamics in Well Control 265
i  Conditions at any point
A  Cross sectional fow area, ft
2
All variables are at standard conditions unless noted with a subscript, i.
The momentum of the kill fuid is easy to compute because it is
essentially an incompressible liquid. The momentum of the kill fuid is
given by Equation 5.14:
M 
mv
g
c
Substituting
: 
q
A
m  
results in the momentum of the kill fuid, Equation 5.17:
M 
2
g
c
A
(5.17)
Since the formation fuids are compressible or partially so, the
momentum is more diffcult to determine. Consider the following devel-
opment of an expression for the momentum of a compressible fuid.
From the conservation of mass, Equation 5.16:
vA 
i
:
i
A
i
And, for a gas, the mass rate of fow from Equation 5.15 is
 vA 
Substituting into the momentum equation gives the momentum of
the gas as:
M 
qv
i
g
c
266 Blowout and Well Control Handbook
Rearranging the equation for the conservation of mass gives an
expression for the velocity of the gas, v
i
, at any location as follows:
:
i

q
i
A
:
i

i
A
From the Ideal Gas Law an expression for
i
, the density of the
gas at any point in the fow stream, is given by Equation 5.18:
i

S
g
M
a
T
i
z
i
T
i
R
(5.18)
Where:
S
g
 Specifc gravity of the gas
M
a
 Molecular weight of air
T
i
 Pressure at point i,
lb
¡
ft
2
z
i
 Compressibility factor at point i
T
i
 Temperature at point i, Rankine
R  Units conversion constant
Substituting results in an expression for v
i
, the velocity of the gas
at point i as follows:
:
i

qz
i
T
i
R
S
g
M
a
T
i
A
(5.19)
Making the fnal substitution gives the fnal expression for the
momentum of the gas:
M 
2
z
i
T
i
R
S
g
M
a
T
i
g
c
A
(5.20)
In this development, all units are BASIC! That means that these
equations can be used in any system as long as the variables are entered
Fluid Dynamics in Well Control 267
in their basic units. In the English system, the units would be pounds, feet,
and seconds. In the metric system, the units would be grams, centimeters,
and seconds. Of course, the units` conversion constants would have to be
changed accordingly.
Consider the example at the Pioneer Production Company Martin
No. 1-7:
Example 5.6
Given:
Figure 5.14
Bottom pressure, T
b
 5000 psi
Volume rate of fow, q  10 mmscfpd
Specifc gravity, S
g
 0.60
Flowing surface pressure, T
a
 14.65 psia
Kill-mud density,
1
 15 ppg
Tubing OD  2
3
8
inches
Tubing ID  1.995 inches
Casing ID  4.892 inches
Temperature at 4000 feet, T
i
 580 Rankine
Flowing pressure at 4000 feet, T
i
 317.6 psia
Compressibility factor at 4000 feet, z
i
 1.00
Required:
Determine the momentum of the gas at 4000 feet and the rate
at which the kill mud will have to be pumped in order for the
momentum of the kill mud to exceed the momentum of the gas
and kill the well.
268 Blowout and Well Control Handbook
8-
"
CSG
AT 2000 FT.
END OF TUBÌNG
AT 4000 FT.
5-
"
CSG
AT 10,000 FT.
5
8
1
2
F/gare 5.14 Pioneer Production Company-Martin No. 1-7.
Solution:
The momentum of the gas at 4000 feet can be calculated from
Equation 5.20 as follows:
M 
2
z
i
T
i
R
S
g
M
a
T
i
g
c
A
Fluid Dynamics in Well Control 269
Substituting in the proper units,
 0.0458
lb
m
ft
3
q  115.74
ft
3
sec
z
i
 1.00
T
i
 580 Rankine
R  1544
ft  lb
f
R  lb
m
S
g
 0.60
M
a
 28.97
T
i
 45,734
lb
¡
ft
2
g
c
 32.2
lb
m
 ft
lb
¡
 sec
2
A  0.1305 ft
2
M 
(0.0458)(115.74)
2
(1.00)(580)(1544)
(0.60)(28.97)(45,734)(32.2)(0.1305)
M  7.53 lb
f
The rate at which the kill mud must be pumped can
be determined be rearranging Equation 5.17, substituting in the
proper units, and solving for the volume rate of fow as follows:
q 
Mg
c
A
1
2
270 Blowout and Well Control Handbook
Where:
M  7.53 lb
f
 112.36
lb
m
ft
3
A  0.1305 ft
2
q 
(7.53)(32.2)(0.1305)
112.36
1
2
q  05307
ft
3
sec
q  57 bbl/min
The preceding example is only one simple instance of the use of the
MomentumKill technology. The momentumof the wellbore fuids is more
diffcult to calculate when in multiphase fow. However, the momentum of
each component of the fow stream is calculated and the total momentum
is the sum of the momentum of each component.
References
1. Grace, Robert, D., 'Fluid Dynamics Kill Wyoming Icicle.¨ World
Oil, April 1987, page 45.
2. Blount, E.M. and Soeiinah, E., 'Dynamic Kill: Controlling Wild
Wells a New Way,¨ World Oil, October 1981, page 109.
3. Kouba, G.E., et al. 'Advancements in Dynamic Kill Calculations
for Blowout Wells,¨ SPEDrilling andCompletion, September 1993,
page 189.
4. Courtesy of P.D. Storts and Williford Energy.
5. Orkiszewski, J., 'Predicting Two-Phase Pressure Drops in Vertical
Pipe,¨ Journal of Petroleum Technology, June 1967, page 829.
6. Courtesy of Amoco Production Company.