15 Well Problem

Identification

Contents
15.1 Introduction 15/228
15.2 Low Productivity 15/228
15.3 Excessive Gas Production 15/231
15.4 Excessive Water Production 15/231
15.5 Liquid Loading of Gas Wells 15/231
Summary 15/241
References 15/241
Problems 15/242
15/228 PRODUCTION ENHANCEMENT

15.1 Introduction . Reservoir heterogeneity (faults, stratification, etc.)

. Completion ineffectiveness (limited entry, shallow per-
The engineering work for sustaining and enhancing oil and
forations, low perforation density, etc.)
gas production rates starts from identifying problems that
cause low production rates of wells, quick decline of the . Restrictions in wellbore (paraffin, asphaltane, scale, gas
desirable production fluid, or rapid increase in the undesir- hydrates, sand, etc.)
able fluids. For oil wells, these problems include The first five factors affect reservoir inflow performance,
that is, deliverability of reservoir. They can be evaluated
. Low productivity on the basis of pressure transient data analyses.
. Excessive gas production The true production profile from different zones can be
. Excessive water production obtained based on production logging such as temperature
. Sand production and spinner flow meter logs. An example is presented in
For gas wells, the problems include Fig. 15.1, which shows that Zone A is producing less than
10% of the total flow, Zone B is producing almost 70% of
. Low productivity the total rate, and Zone C is contributing about 25% of the
. Excessive water production total production.
. Liquid loading The last factor controls well deliverability. It can be
. Sand production evaluated using data from production logging such as
flowing gradient survey (FGS). The depth interval
Although sand production is easy to identify, well testing with high-pressure gradient is usually the interval where
and production logging are frequently needed to identify the depositions of paraffins, asphaltanes, scales, or gas
the causes of other well problems. hydrates are suspected.

15.2 Low Productivity 15.2.1 Pressure Transient Data Analysis

Pressure transient testing plays a key role in evaluating
The lower than expected productivity of oil or gas well is
found on the basis of comparison of the well’s actual exploration and development prospects. Properly designed
production rate and the production rate that is predicted well tests can provide reservoir engineers with reservoir
by Nodal analysis. If the reservoir inflow model used in the pressure, reserves (minimum economic or total), and flow
Nodal analysis is correct (which is often questionable), capacity, all of which are essential in the reservoir evalu-
ation process. Some of the results one can obtain from
the lower than expected well productivity can be attributed
pressure transient testing include the following:
to one or more of the following reasons:

. Overestimate of reservoir pressure . Initial reservoir pressure

. Overestimate of reservoir permeability (absolute and . Average reservoir pressure
relative permeabilities) . Directional permeability
. Radial effective permeability changes from the wellbore
. Formation damage (mechanical and pseudo skins)
. Gas condensate fallout effect on flow

Temperature
Depth

150 160
sp % flow from spinner flowmeter
g - ray 0 25 50 75 100

7100

Zone A 7150

7200
Zone B

7250

Zone C

7300

Fill
7350

Figure 15.1 Temperature and spinner flowmeter-derived production profile (Economides et al., 1994).
 WELL PROBLEM IDENTIFICATION 15/229

. Near wellbore damage/stimulation rw ¼ wellbore radius

. Rate-dependent skin S ¼ total skin factor.
. Boundary identification
. Partial penetration effect on flow Horizontal Linear Flow. For hydraulically fractured
. Effective fracture length wells, the horizontal linear flow can be mathematically
. Effective fracture conductivity described in consistent units as
. Dual-porosity characteristics (storativity and transmis- "sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi #
sivity ratios) qBm pky t
pwf ¼ pi " þ S , (15:2)
The theoretical basis of pressure transient data analysis is 2pky h fm ct x2f
beyond the scope of this book. It can be found elsewhere
(Chaudhry, 2004; Horne, 1995; Lee et al., 2003). Modern where xf is fracture half-length and ky is the permeability
computer software packages are available for data anal- in the direction perpendicular to the fracture face.
yses. These packages include PanSystem (EPS, 2004) and Vertical Radial Flow. For horizontal wells as depicted
F.A.S.T. WellTest (Fekete, 2003). The following subsec- in Fig. 15.2, the early-time vertical radial flow can be
tions briefly present some principles of data analyses that mathematically described in consistent units as
lead to deriving reservoir properties directly affecting well # ! " $
qBm kyz t
productivity. pwf ¼ pi " ln þ 2S þ 0:80907 , (15:3)
Reservoir Pressure. Reservoir pressure is a key param- 4pkyz L fm ct r2w
eter controlling well deliverability. A simple way to deter- where L is the horizontal wellbore length and kyz is the
mine the magnitude of initial reservoir pressure may be geometric mean of horizontal and vertical permeabilities,
the Horner plot of data from pressure buildup test if that is,
the reservoir boundary was not reached during the test. pffiffiffiffiffiffiffiffiffi
If the boundary effects are seen, the average reservoir kyz ¼ ky kz : (15:4)
pressure can be estimated on the basis of the extrapolated Horizontal Pseudo-Linear Flow. The pseudo-linear
initial reservoir pressure from Horner plot and the MBH flow toward a horizontal wellbore can be mathematically
plot (Dake, 2002). described in consistent units as
Effective Permeability. The effective reservoir perme-
ability that controls the well’s deliverability should be "sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi #
derived from the flow regime that prevails in the reservoir qBm 4pky t
pwf ¼ pi " 2
þS : (15:5)
for long-term production. To better understand the flow 2pky ðh " Zw Þ fm ct L
regimes, the commonly used equations describing flow in oil
reservoirs are summarized first in this subsection. Similar Horizontal Pseudo-Radial Flow. The pseudo-radial
equations for gas reservoirs can be found in Lee et al. (2003). flow toward a horizontal wellbore can be mathematically
Horizontal Radial Flow. For vertical wells fully pene- described in consistent units as
trating nonfractured reservoirs, the horizontal radial flow # ! " $
qBm kh t
can be mathematically described in consistent units as pwf ¼ pi " ln þ 2S þ 0:80907 : (15:6)
4pkh h 2
fm ct rw
# ! " $
qBm kh t For vertical wells fully penetrating nonfractured reser-
pwf ¼ pi " ln 2
þ 2S þ 0:80907 , (15:1)
4pkh h fm ct rw voirs, it is usually the average (geometric mean) of hori-
where zontal permeabilities, kh , that dominates long-term
pwf ¼ flowing bottom-hole pressure production performance. This average horizontal perme-
pi ¼ initial reservoir pressure ability can be derived from the horizontal radial flow
q ¼ volumetric liquid production rate regime. For wells draining relatively small portions of
B ¼ formation volume factor hydraulically fractured reservoir segments, it is usually
m ¼ fluid viscosity the permeability in the direction perpendicular to the frac-
kh ¼ the average horizontal permeability ture face that controls long-term production performance.
h ¼ pay zone thickness This permeability can be derived from the horizontal lin-
t ¼ flow time ear flow regime. For horizontal wells draining relatively
f ¼ initial reservoir pressure large portions of nonfractured reservoir segments, it is usu-
ct ¼ total reservoir compressibility ally again the geometric mean of horizontal permeabilities

z L

x h
Zw
y

y h
Zw
x

Figure 15.2 Notations for a horizontal wellbore.

15/230 PRODUCTION ENHANCEMENT

that dominates long-term production performance. This qBm

k¼" : (15:17)
average horizontal permeability can be derived from the 4pHR mR
pseudo-radial flow regime. For vertical wells partially pene- For any types of linear flow, Eqs. (15.2) and (15.5) indicate
trating nonfractured reservoirs, both horizontal and vertical that plotting of the bottom-hole pressure versus the
permeabilities influence long-term production performance. square-root of time data will show a trend with a constant
These permeabilities can usually be derived from the hemi- slope mL , where
spherical flow regime. rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Flow regimes are usually identified using the diagnostic qB m
mL ¼ " , (15:18)
pressure derivative p0 defined as HL XL pfct ky
dDp dDp where HL ¼ h and XL ¼ 2xf for linear flow, and
p0 ¼ ¼t , (15:7) HL ¼ h " Zw and XL ¼ L for pseudo-linear flow, respec-
d ln (t) dt
tively. Then the permeability in the flow plane can be
where t is time and Dp is defined as estimated by
Dp ¼ pi " pwf (15:8) ! "2
m qB
for drawdown tests, where pi and pwf are initial reservoir ky ¼ : (15:19)
pfct mL HL XL
pressure and flowing bottom-hole pressure, respectively.
For pressure buildup tests, the Dp is defined as If a horizontal well is tested for a time long enough to
detect the pseudo-radial flow, then it is possible to estimate
Dp ¼ psw " pwfe , (15:9) other directional permeabilities by
where pws and pwfe are ship-in bottom-hole pressure and k2h
the flowing bottom-hole pressure at the end of flow (before kx ¼ (15:20)
ky
shut-in), respectively.
For any type of radial flow (e.g., horizontal radial flow, and
vertical radial flow, horizontal pseudo-radial flow), the
k2yz
diagnostic derivative is derived from Eqs. (15.1), (15.3), kz ¼ : (15:21)
and (15.6) as ky
dDp qBm Although kx and kz are not used in well productivity
p0 ¼ ¼ , (15:10) analysis, they provide some insight about reservoir anisot-
d ln (t) 4pkHR
ropy.
where k is the average permeability in the flow plane (kh or Skin Factor. Skin factor is a constant that is used to
kyz ) and adjust the flow equation derived from the ideal condition
pffiffiffiffiffiffiffiffiffiffiffi
kh ¼ kx ky (homogeneous and isotropic porous media) to suit the
applications in nonideal conditions. It is an empirical fac-
HR is the thickness of the radial flow (h or L). Apparently, tor employed to consider the lumped effects of several
the diagnostic derivative is constant over the radial flow aspects that are not considered in the theoretical basis
time regime. The plot of p0 versus t data should show a when the flow equations were derived. The value of the
trend of straight line parallel to the t-axis. skin factor can be derived from pressure transient test
For linear flow (e.g., flow toward a hydraulic fracture), analysis with Eqs. (15.1), (15.2), (15.3), (15.5), and (15.6).
the diagnostic derivative is derived from Eq. (15.2) as But its value has different meanings depending on flow
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi regime. A general expression of the skin factor is
dDp qB mt
p0 ¼ ¼ : (15:12) X
d ln (t) 4hxf pfct ky S ¼ SD þ SCþu þ SP þ SPS , (15:22)
For pseudo-linear flow (e.g., flow toward a horizontal well), where SD is damage skin during drilling, cementing, well
the diagnostic derivative is derived from Eq. (15.5) as completion, fluid injection, and even oil and gas produc-
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi tion. Physically, it is due to plugging of pore space by
dDp qB mt
p0 ¼ ¼ : (15:13) external or internal solid particles and fluids. This com-
d ln (t) 2L(h " zw ) pfct ky ponent of skin factor can be removed or averted with well
stimulation operations. The SCþu is a skin component due
Taking logarithm of Eqs. (15.12) and (15.13) gives
! rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi" to partial completion and deviation angle, which make the
1 qB m flow pattern near the wellbore deviate from ideal radial
logðp0 Þ ¼ logðtÞ þ log (15:14)
2 4hxf pfct ky flow pattern. This skin component is not removable in
water coning and gas coning systems. The SP is a skin
and component due to the nonideal flow condition around the
! rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi"
1 qB m perforations associated with cased-hole completion. It
logðp0 Þ ¼ logðtÞ þ log : (15:15) depends on a number of parameters including perforation
2 2L(h " zw ) pfct ky
density, phase angle, perforation depth, diameter, com-
Equations (15.13) and (15.14) indicate that the signature of pacted zone, and others. This component can be mini-
the linear flow regime is the 1⁄2 slope on the log-log plot of mized with optimized perorating technologies. The SSPS
diagnostic derivative versus time. represents pseudo-skin components due to non–Darcy
Once the flow regimes are identified, permeabilities flow effect, multiphase effect, and flow convergence near
associated with the flow regime can be determined based the wellbore. These components cannot be eliminated.
on slope analyses. For any types of radial flow, Eqs. (15.1), It is essential to know the magnitude of components of
(15.3), and (15.6) indicate that plotting of bottom-hole the skin factor S derived from the pressure transient test
pressure versus time data on a semilog scale will show a data analysis. Commercial software packages are available
trend with a constant slope mR , where for decomposition of the skin factor on the basis of well
qBm completion method. One of the packages is WellFlo (EPS,
mR ¼ " : (15:16) 2005).
4pkHR
Then the average permeability in the flow plane (kh or kyz ) Example Problem 15.1 A horizontal wellbore was placed
can be estimated by in a 100-ft thick oil reservoir of 0.23 porosity. Oil formation
 WELL PROBLEM IDENTIFICATION 15/231

volume factor and viscosity are 1.25 rb/stb and 1 cp, The channeling behind the casing and gas coning prob-
respectively. The total reservoir compressibility factor is lems can be identified based on production logging such
10"5 psi"1 . The well was tested following the schedule as temperature and noise logs. An example is depicted
shown in Fig. 15.3. The measured flowing bottom-hole in Fig. 15.12, where both logs indicate that gas is being
pressures are also presented in Fig. 15.3. Estimate directional produced from an upper gas sand and channeling down to
permeabilities and skin factors from the test data. some perforations in the oil zone.
Excessive gas production of an oil well could also be due
Solution Figure 15.4 presents a log-log diagnostic plot of to gas production from unexpected gas zones. This can be
test data. It clearly indicates a vertical radial flow at early identified using production logging such as temperature
time, a pseudo-linear flow at mid-time, and the beginning and density logs. An example is presented in Fig. 15.13,
of a pseudo-radial flow at late time. where both logs indicate gas production from the thief
zone B.
The semi-log analysis for the vertical radial flow is
shown in Fig. 15.5, which gives kyz ¼ 0:9997 md and
near-wellbore skin factor S ¼ "0:0164.
The square-root time plot analysis for the pseudo-linear 15.4 Excessive Water Production
flow is shown in Fig. 15.6, which gives the effective well- Excessive water production is usually from water zones,
bore length of L ¼ 1,082:75 ft and a skin factor due to not from the connate water in the pay zone. Water enters
convergence of S ¼ 3:41. the wellbore due to channeling behind the casing
The semi-log analysis for the horizontal pseudo-radial (Fig. 15.14), preferential flow through high-permeability
flow is shown in Fig. 15.7, which gives kh ¼ 1:43 md and zones (Fig. 15.15), water coning (Fig. 15.16), hydraulic
pseudo-skin factor S ¼ "6:17. fracturing into water zones, and casing leaks.
Figure 15.8 shows a match between the measured and Figure 15.17 shows how to identify fracture height using
model-calculated pressure responses given by an optimiza- prefracture and postfracture temperature logs to tell whether
tion technique. This match was obtained using the following the hydraulic fracture has extended into a water zone.
parameter values: In addition to those production logging tools that are
kh ¼ 1:29 md mentioned in the previous section, other production log-
kz ¼ 0:80 md ging tools can be used for identifying water-producing
S ¼ 0:06 zones. Fluid density logs are especially useful for identify-
L ¼ 1,243 ft: ing water entries. Comparison between water-cut data and
spinner flowmeter log can sometimes give an idea of where
To estimate the long-term productivity of this horizontal the water is coming from. Figure 15.18 shows a spinner
well, the kh ¼ 1:29 md and S ¼ 0:06 should be used in the flowmeter log identifying a watered zone at the bottom of
well inflow equation presented in Chapter 3. a well with a water-cut of nearly 50%.

15.3 Excessive Gas Production 15.5 Liquid Loading of Gas Wells

Excessive gas production is usually due to channeling be- Gas wells usually produce natural gas-carrying liquid
hind the casing (Fig. 15.9), preferential flow through high- water and/or condensate in the form of mist. As the gas
permeability zones (Fig. 15.10), gas coning (Fig. 15.11), and flow velocity in the well drops because of reservoir pres-
casing leaks (Clark and Schultz, 1956). sure depletion, the carrying capacity of the gas decreases.

Figure 15.3 Measured bottom-hole pressures and oil production rates during a pressure drawdown test.
15/232 PRODUCTION ENHANCEMENT

1,000

Delta P and P-Derivative (psi)

Begging
of Pseudo
100
Radial Flow

Pseudo Linear
Vertical Radial Flow
Flow

10
0.001 0.01 0.1 1 10 100 1,000
Elapsed Time (hours)

Figure 15.4 Log-log diagnostic plot of test data.

Radial Flow Plot

6,000

Model results
Two no-flow boundaries-homogeneous
Infinitely acting
5,500 Kbar = 0.9997 md
S = −0.0164

5,000
Pressure (psia)

4,500

4,000

3,500
0.001 0.01 0.1 1 10 100 1,000
Elapsed Time (hours)

Figure 15.5 Semi-log plot for vertical radial flow analysis.

When the gas velocity drops to a critical level, liquids begin underlying disposal zone is another option. However,
to accumulate in the well and the well flow can undergo an liquid-loading is not always obvious and recognizing the
annular flow regime followed by a slug flow regime. The liquid-loading problem is not an easy task. A thorough
accumulation of liquids (liquid loading) increases the diagnostic analysis of well data needs to be performed.
bottom-hole pressure, which reduces gas production rate. The symptoms to look for include onset of liquid slugs
A low gas production rate will cause gas velocity to drop at the surface of well, increasing difference between the
further. Eventually, the well will undergo a bubbly flow tubing and casing pressures with time, sharp changes in
regime and cease producing. gradient on a flowing pressure survey, and sharp drops in
Several measures can be taken to solve the liquid-loading production decline curve.
problem. Foaming the liquid water can enable the gas to lift
water from the well. Using smaller tubing or creating a
lower wellhead pressure sometimes can keep mist flowing. 15.5.1 The Turner et al. Method
The well can be unloaded by gas-lifting or pumping the Turner et al. (1969) were the pioneer investigators who
liquids out of the well. Heating the wellbore can prevent analyzed and predicted the minimum gas flow rate to
oil condensation. Down-hole injection of water into an prevent liquid-loading. They presented two mathematical
 WELL PROBLEM IDENTIFICATION 15/233

Figure 15.6 Square-root time plot for pseudo-linear flow analysis.

Figure 15.7 Semi-log plot for horizontal pseudo-radial flow analysis.

models to describe the liquid-loading problem: the film- to the critical Weber number of 30. Turner et al.’s terminal
movement model and the entrained drop movement slip velocity equation is expressed in U.S. field units as
model. On the basis of analyses on field data, they con- & '1=4
cluded that the film-movement model does not represent 1:3s1=4 rL " rg
the controlling liquid transport mechanism. vsl ¼ 1=4 1=2
: (15:23)
Turner et al.’s entrained drop movement model was de- Cd r g
rived on the basis of the terminal free settling velocity of According to Turner et al., gas will continuously remove
liquid drops and the maximum drop diameter corresponding liquids from the well until its velocity drops to below the
15/234 PRODUCTION ENHANCEMENT

Figure 15.8 Match between measured and model calculated pressure data.

Well bore High pressure

Gas zone

Casing leak

Bad cement job

Oil zone

Figure 15.9 Gas production due to channeling behind the casing (Clark and Schultz, 1956).

terminal slip velocity. The minimum gas flow rate (in attributed to several facts including the use of drag coeffi-
MMcf/D) for a particular set of conditions (pressure and cients for solid spheres, the assumption of stagnation velo-
conduit geometry) can be calculated using Eqs. (15.23) and city, and the critical Weber number established for drops
(15.24): falling in air, not in compressed gas.
The main problem that hinders the application of
3:06pvsl A
QgslMM ¼ (15:24) Turner et al.’s entrained drop model to gas wells comes
Tz from the difficulties of estimating the values of fluid den-
Figure 15.19 shows a comparison between the results of sity and pressure. Using an average value of gas-specific
Turner et al.’s entrained drop movement model. The map gravity (0.6) and gas temperature (120 8F), Turner et al.
shows many loaded points in the unloaded region. Turner derived an expression for gas density as 0.0031 times the
et al. recommended the equation-derived values be pressure. However, they did not present a method for
adjusted upward by approximately 20% to ensure removal calculating the gas pressure in a multiphase flow wellbore.
of all drops. Turner et al. believed that the discrepancy was
 WELL PROBLEM IDENTIFICATION 15/235

Well Bore
Gas-Oil
Contact
Intermediate
permeability

Low
permeability

High
permeability

Low
permeability

Intermediate
permeability

Figure 15.10 Gas production due to preferential flow through high-permeability zones (Clark and Schultz, 1956).

Well Bore

Gas Cap

Oil Zone

Figure 15.11 Gas production due to gas coning (Clark and Schultz, 1956).

The spreadsheet program TurnerLoading.xls has been four-phase flow model resulted in a closed-form analytical
developed for quick calculation associated with this book. equation for predicting the minimum gas flow rate.
Turner et al.’s entrained drop movement model was later
modified by a number of authors. Coleman et al. (1991)
15.5.2.1 Minimum Kinetic Energy
suggested to use Eq. (15.23) with a lower constant value.
Kinetic energy per unit volume of gas can be expressed as
Nosseir et al. (2000) expanded Turner et al.’s entrained drop
model to more than one flow regime in a well. Lea and rg v2g
Nickens (2004) made some corrections to Turner et al.’s Ek ¼ : (15:25)
2gc
simplified equations. However, the original drawbacks
(neglected transport velocity and multiphase flow pressure) Substituting Eq. (15.23) into Eq. (15.25) gives an expres-
with Turner et al.’s approach still remain unsolved. sion for the minimum kinetic energy required to keep
liquid droplets from falling:
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
'ffi
u &
15.5.2 The Guo et al. Method us r " r
t L g
Starting from Turner et al.’s entrained drop model, Guo Eksl ¼ 0:026 (15:26)
Cd
et al. (2006) determined the minimum kinetic energy of gas
that is required to lift liquids. A four-phase (gas, oil, water, If the value of drag coefficient Cd ¼ 0:44 (recommended by
and solid particles) mist-flow model was developed. Turner et al.) is used and the effect of gas density is
Applying the minimum kinetic energy criterion to the neglected (a conservative assumption), Eq. (15.26) becomes
15/236 PRODUCTION ENHANCEMENT

Temperature
Noise Amplitude
> 600 Hz

Gas
zone

Gas
Gas

Oil
Oil
production Oil
zone

Figure 15.12 Temperature and noise logs identifying gas channeling behind casing (Economides et al., 1994).

Temperature (!F)
Depth

Fluid density (g/cc)

Figure 15.13 Temperature and fluid density logs identifying a gas entry zone (Economides et al., 1994).

pffiffiffiffiffiffiffiffiffi
Eksl ¼ 0:04 srL : (15:27) droplets from falling) plus the transport velocity of the
In gas wells producing water, typical values for water–gas droplets, that is,
interfacial tension and water density are 60 dynes/cm and vgm ¼ vsl þ vtr : (15:28)
65 lbm =ft3 , respectively. This yields the minimum kinetic
energy value of 2:5 lbf -ft=ft3 . In gas wells producing conden- The transport velocity vtr may be calculated on the
sate, typical values for condensate–gas interfacial tension basis of liquid production rate, geometry of the conduit,
and condensate density are 20 dynes/cm and 45 lbm =ft3 , and liquid volume fraction, which is difficult to quantify.
respectively. This yields the minimum kinetic energy value Instead of trying to formulate an expression for
of 1:2 lbf -ft=ft3 . the transport velocity vtr , Guo et al. used vtr as an empir-
The minimum gas velocity required for transporting the ical constant to lump the effects of nonstagnation ve-
liquid droplets upward is equal to the minimum gas ve- locity, drag coefficients for solid spheres, and the critical
locity required for floating the liquid droplets (keeping the Weber number established for drops falling in air. On the
 WELL PROBLEM IDENTIFICATION 15/237

Low pressure
Oil reservoir

Water channel along

Bad cement job
Casing leak

High pressure
water sand

Figure 15.14 Water production due to channeling behind the casing.

Well Bore

Low permeability

High permeability

Intermediate
permeability

Low permeability

Figure 15.15 Preferential water flow through high-permeability zones.

basis of the work by Turner et al., the value of vtr Expressions for rg and vg can be obtained from ideal gas
was taken as 20% of vsl in this study. Use of this value law:
results in 2:7Sg p
vgm & 1:2vsl : (15:29) rg ¼ (15:31)
T
Substituting Eqs. (15.23) and (15.29) into Eq. (15.25) "2 TQG
vg ¼ 4:71 ' 10 (15:32)
results in the expression for the minimum kinetic energy Ai p
required for transporting the liquid droplets as Substituting Eqs. (15.31) and (15.32) into Eq. (15.25) yields
pffiffiffiffiffiffiffiffiffi
Ekm ¼ 0:0576 srL : (15:30) Sg TQ 2G
Ek ¼ 9:3 ' 10"5 : (15:33)
For typical gas wells producing water, this equation A2i p
yields the minimum kinetic energy value of 3:6 lbf -ft=ft3 . Equation (15.33) indicates that the gas kinetic energy
For typical gas wells producing condensate, this equation decreases with increased pressure, which means that the
gives the minimum kinetic energy value of 1:73 lbf -ft=ft3 . controlling conditions are bottom-hole conditions where
These numbers imply that the required minimum gas pro- gas has higher pressure and lower kinetic energy. This
duction rate in water-producing gas wells is approximately analysis is consistent with the observations from air-
twice that in condensate-producing gas wells. drilling operations where solid particles accumulate at
To evaluate the gas kinetic energy E k in Eq. (15.25) at a the bottom-hole rather than at the top-hole. However,
given gas flow rate and compare it with the minimum this analysis is in contradiction with the results by Turner
required kinetic energy Ekm in Eq. (15.30), the values of et al., which indicated that the wellhead conditions are in
gas density rg and gas velocity vg need to be determined. most instances, controlling.
15/238 PRODUCTION ENHANCEMENT

Well Bore

Oil Zone

Cone

Water

Figure 15.16 Water production due to water coning.

8,800 2,690

Pre Frac
9,000 Profile 2,750
Thermal
Conductivity
Static
Effects
9,200 Log 2,810

2,870
Hole Depth (ft)

9,400

Hole depth (m)

Post Frac
9,600 Profile 2,930

9,800 2,990

10,000 Fracture 3,050

Top Profiles
Separate
10,200 3,110

10,400 3,170
175 200 225 250 275
Temperature (!F)
80 !C 93 !C 108!C 121!C 135!C

Figure 15.17 Prefracture and postfracture temperature logs identifying fracture height (Dobkins, 1981).

15.5.2.2 Four-Phase Flow Model where

To accurately predict the bottom-hole pressure p in 15:33Ss Qs þ 86:07Sw Qw þ 86:07So Qo þ 18:79Sg QG
Eq. (15.33), a gas-oil-water-solid four-phase mist-flow a¼
103 Tav QG
model was developed by Guo et al. (2006). According to the
four-phase flow model, the flowing pressure p at depth L can ' cos (u), (15:35)
be solved numerically from the following equation: 0:2456Qs þ 1:379Qw þ 1:379Qo
* * b¼ , (15:36)
( ) 1 " 2bm ** ð144p þ mÞ2 þ n ** 103 Tav QG
144b p " phf þ ln*( *
* 144phf þ m)2 þ n*
"6
2 6:785 ' 10 Tav QG
c¼ , (15:37)
* * Ai
( ) 1 " 2bm ** ðP þ mÞ2 þ n ** Qs þ 5:615ðQw þ Qo Þ
b P " Phf þ ln*( *
2 * Phf þ m)2 þ n* d¼
600Ai
, (15:38)
# ! "
m þ bc n " bm2 144p þ m 6f
" pffiffiffi tan"1 pffiffiffi e¼ , (15:39)
n n gDh cos (u)
! "$
144phf þ m 2 32
" tan"1 pffiffiffi 1
n fM ¼4 & '5 , (15:40)
0 2"
( 2
) 1:74 " 2 log Dh
¼ a 1 þ d e L, (15:34)
 WELL PROBLEM IDENTIFICATION 15/239

Spinner Speed
0 RPS 10

Production Profile
0 B/D 4,860

0.0%
12,700
4.1%

21.0%
12,800

12,900
14.4%

13,000
12.1%

13,100

13,200
48.4
%

13,300

13,400

Figure 15.18 Spinner flowmeter log identifying a watered zone at bottom.

15/240 PRODUCTION ENHANCEMENT

12,000

10,000

Test Flow Rate (Mcf/D)

8,000

6,000

4,000 Unloaded
? Nearly loaded up
2,000 Loaded up
Questionable
0
0 2,000 4,000 6,000 8,000 10,000 12,000
Calculated Minimum Flow Rate (Mcf/D)

Figure 15.19 Calculated minimum flow rates with the Turner et al. model and test flow rates.

cde Substituting Eq. (15.44) into Eq. (15.34) results in

m¼ , (15:41)
1 þ d2e
and 1 " 2bm m þ bc n " bm2
144ba1 þ ln a2 " pffiffiffi
2 2 n (15:45)
c e + "1 ,
n¼ 2
, (15:42) "1
' tan b1 " tan b2 ¼ g,
ð1 þ d 2 eÞ
where where
A ¼ cross-sectional area of conduit, ft2
Dh ¼ hydraulic diameter, in. Sg Tbh Q2gm
a1 ¼ 9:3 ' 10"5 " phf , (15:46)
fM ¼ Moody friction factor A2i Ekm
g ¼ gravitational acceleration, 32:17 ft=s2 & '2
S T Q2
L ¼ conduit length, ft 1:34 ' 10"2 gA2bhE gm þ m þ n
km
p ¼ pressure, psia a2 ¼ (
i
)2 , (15:47)
phf ¼ wellhead flowing pressure, psia 144phf þ m þ n
QG ¼ gas production rate, Mscf/day S T Q2
1:34 ' 10"2 gA2bhE gm þ m
Qo ¼ oil production rate, bbl/day b1 ¼ pffiffiffi i
km
, (15:48)
Qs ¼ solid production rate, ft3 =day n
Qw ¼ water production rate, bbl/day 144phf þ m
Sg ¼ specific gravity of gas, air ¼ 1 b2 ¼ pffiffiffi , (15:49)
n
So ¼ specific gravity of produced oil,
freshwater ¼ 1 and
Sw ¼ specific gravity of produced water, ( )
freshwater ¼ 1 g ¼ a 1 þ d 2 e L: (15:50)
Ss ¼ specific gravity of produced solid,
freshwater ¼ 1 All the parameter values should be evaluated at Qgm .
Tav ¼ the average temperature in the butting, 8R The minimum required gas flow rate Qgm can be solved
«0 ¼ pipe wall roughness, in. from Eq. (15.45) with a trial-and-error or numerical
u ¼ inclination angle, degrees. method such as the Bisection method. It can be shown
that Eq. (15.45) is a one-to-one function of Qgm for
15.5.2.3 Minimum Required Gas Production Rate Qgm values greater than zero. Therefore, the Newton–
A logical procedure for predicting the minimum required Raphson iteration technique can also be used for solving
gas flow rate Qgm involves calculating gas density rg , gas Qgm . Commercial software packages such as MS Excel can
velocity vg , and gas kinetic energy Ek at bottom-hole con- be used as solvers. In fact, the Goal Seek function built
dition using an assumed gas flow rate QG , and compare into MS Excel was used for generating solutions presented
the Ek with Ekm . If the Ek is greater than Ekm , the QG is in this chapter. The spreadsheet program is named
higher than the Qgm . The value of QG should be reduced GasWellLoading.xls.
and the calculation should be repeated until the Ek is very
close to Ekm . Because this procedure is tedious, a simple
equation was derived by Guo et al. for predicting the Example Problem 15.2 To demonstrate how to use
minimum required gas flow rate in this section. Under Eq. (15.45) for predicting the minimum unloading gas
the minimum unloaded condition (the last point of the flow rate, consider a vertical gas well producing 0.70
mist flow regime), Eq. (15.33) becomes specific gravity gas and 50 bbl/day condensate through
2
a 2.441-in. inside diameter (ID) tubing against a
Sg Tbh Qgm wellhead pressure of 900 psia. Suppose the tubing string
Ekm ¼ 9:3 ' 10"5 , (15:43)
A2i p is set at a depth of 10,000 ft, and other data are given in
which gives Table 15.1.
2
Sg Tbh Qgm Solution The solution given by the spreadsheet program
p ¼ 9:3 ' 10"5 : (15:44)
A2i Ekm GasWellLoading.xls is shown in Table 15.2.
 WELL PROBLEM IDENTIFICATION 15/241

Summary
Table 15.1 Basic Parameter Values for Example
Problem 15.1 This chapter presents a guideline to identifying problems
commonly encountered in oil and gas wells. Well test
Gas-specific gravity 0:7 (air ¼ 1) analysis provides a means of estimating properties of indi-
Hole inclination 0 degrees vidual pay zones. Production logging analysis identifies
Wellhead temperature 608 fluid entries to the wellbore from different zones. The
Geothermal gradient 0.01 8F/ft Guo et al. method is more accurate than the Turner et al.
Condensate gravity 60 8API method for predicting liquid-loading problems in gas pro-
Water-specific gravity 1:05 (water ¼ 1) duction wells.
Solid-specific gravity 2:65 (water ¼ 1)
Interfacial tension 20 dyne/cm
Tubing wall roughness 0.000015 in. References
chaudhry, a.c. Oil Well Testing Handbook. Burlington:
Table 15.2 Result Given by the Spreadsheet Program Gulf Professional Publishing, 2004.
GasWellLoading.xls clark, n.j. and schultz , w.p. The analysis of problem
wells. Petroleum Engineer September 1956;28:B30–
Calculated Parameters B38.
Hydraulic diameter 0.2034 ft coleman, s.b., clay, h.b., mccurdy, d.g., and norris,
Conduit cross-sectional area 0.0325 ft2
l.h., iii. A new look at predicting gas well loading-up.
Average temperature 570 8R
Minimum kinetic energy 1.6019 lb-ft/ft3 JPT (March 1991), Trans. AIME 1991;291:329.
a¼ 2.77547E-05 dake, l.p. Fundamentals of Reservoir Engineering.
b¼ 1.20965E-07 Amsterdam: Elsevier, 2002.
c¼ 875999.8117 dobkins, t.a. Improved method to determine hydraulic
d¼ 0.10598146 fracture height. JPT April 1981:719–726.
e¼ 0.000571676 economides, m.j., hill, a.d., and ehlig-economides, c.
fM ¼ 0.007481992 Petroleum Production Systems. New Jersey: Prentice
m¼ 53.07387106 Hall PTR, 1994.
n¼ 438684299.6
E-Production Services, Inc. FloSystem User Manual.
Solution Edinburgh: E-Production Services, Inc., 2005.
Critical gas production rate 1,059 Mscf/day E-Production Services, Inc. PanSystem User Manual.
Pressure ( p) ¼ 1,189 psia Edinburgh: E-Production Services, Inc., 2004.
Objective function f(Qgm ) ¼ "1:78615E-05 fekete., f.a.s.t. WellTest User Manual. Calgary: Fekete
Associates, Inc., 2003.
guo, b., ghalambor, a., and xu, c. A systematic approach
to predicting liquid loading in gas well. SPE Produc-
15.5.3 Comparison of the Turner et al. and the Guo tion Operations J. February 2006.
et al. Methods
horne, r.n. Modern Well Test Analysis: A Computer-Aided
Figure 15.20 illustrates Eq. (15.45)–calculated minimum
flow rates mapped against the test flow rates for the Approach. New York: Petroway Publishing, 1995.
same wells used in Fig. 15.19. This map shows six loaded lea, j.f. and nickens, h.v. Solving gas-well liquid-loading
points in the unloaded region, but they are very close to problems. SPE Prod. Facilities April 2004:30.
the boundary. This means the Guo et al. method is more lee, j.w., rollins, j.b., and spivey, j.p. Pressure Transient
accurate than the Turner et al. method in estimating the Testing. Richardson: Society of Petroleum Engineers,
minimum flow rates. 2003.

12,000

10,000
Test Flow Rate (Mcf/D)

8,000

6,000

4,000 Unloaded
? Nearly loaded up
2,000 Loaded up
Questionable
0
0 2,000 4,000 6,000 8,000 10,000 12,000
Calculated Minimum Flow Rate (Mcf/D)

Figure 15.20 The minimum flow rates given by the Guo et al. model and the test flow rates.
15/242 PRODUCTION ENHANCEMENT

nosseir, m.a., darwich, t.a., sayyouh, m.h., and sallaly, 15.2 Consider a gas well producing 50 bbl/day of water
m.e. A new approach for accurate prediction of loading and 0:2 ft3 of sand through a 2.441-in. ID tubing
in gas wells under different flowing conditions. SPE against a wellhead pressure of 600 psia and tempera-
Prod. Facilities November 2000;15(4):245. ture of 80 8F. Suppose the tubing string is set at a
depth of 9,000 ft and geothermal gradient is
turner, r.g., hubbard, m.g., and dukler, a.e. Analysis 0.01 8F/ft, estimate the minimum gas production
and prediction of minimum flow rate for the continu- rate before the gas well gets loaded.
ous removal of liquids from gas wells. JPT November 15.3 Consider a gas well producing 80 bbl/day of water
1969, Trans. AIME 1969;246:1475. and 0:1 ft3 of sand through a 1.995-in. ID tubing
against a wellhead pressure of 400 psia and tempera-
ture of 70 8F. Suppose the tubing string is set at a
Problems depth of 7,000 ft and geothermal gradient is
0.01 8F/ft, estimate the minimum gas production
15.1 Consider a gas well producing 50 bbl/d of condensate rate before the gas well gets loaded.
and 0.1 cubic foot of sand through a 2.441-in. I.D. 15.4 Consider a gas well producing 70 bbl/day of oil and
tubing against a wellhead pressure of 500 psia. Sup- 0:1 ft3 of sand through a 1.995-in. ID tubing against a
pose the tubing string is set at a depth of 8,000 ft, use wellhead pressure of 600 psia and temperature of
the following data and estimate the minimum gas 80 8F. Suppose the tubing string is set at a depth of
production rate before the gas well gets loaded. 6,000 ft and geothermal gradient is 0.01 8F/ft, esti-
mate the minimum gas production rate before the gas
Gas-specific gravity : 0:75 (air ¼ 1) well gets loaded.
Hole inclination : 0 degrees
Wellhead temperature : 60 8F
Geothermal gradient : 0.01 8F/ft
Condensate gravity : 60 8API
Water-specific gravity : 1:07 (water ¼ 1)
Solid-specific gravity : 2:65 (water ¼ 1)
Oil–gas interface tension : 20 dyne/cm
Tubing wall roughness : 0.000015 in.