You are on page 1of 24

34

HOW OPERATING

AND ENVIRONMENTAL

CONDITIONS

AFFECT EROSION

B.S. McLaury, S.A. Shirazi, J.R. Shadley, and E.F. Rybicki


The Erosion/Corrosion Research Center
Department of Mechanical Engineering
The University of Tulsa
Tulsa, OK 74104

ABSTRACT
Sand in producing wells can result in severe erosion even resulting in failure. This work
investigates the effect of a variety of parameters such as diameter, flow velocity, particle size, and fluid
properties on the severity of erosion for an elbow. The elbow is selected since it redirects the flow and
experiences erosion from the impact of sand particles that do no follow the streamlines of the flow and
impinge the pipe wall. A model previously developed for single-phase flow is presented and extended
in this work for use with multiphase flow. Results for the maximum penetration rate in an elbow are
examined for many parameters in both single-phase and multiphase flow. By assuming an allowable
penetration rate, the erosion prediction model can also be used to determine the threshold velocity that
would result in this penetration rate. The threshold erosional velocity curves are also generated for the
multiphase flow cases.
Kevwords:

erosion, single-phase flow, multiphase flow, mechanistic model


INTRODUCTION

The presence of sand in the produced fluid provides many challenges for oil and gas production
companies, and sand management has become increasingly important as high rate wells with sand
become more prominent. One of the primary concerns created by the presence of sand is erosion. The
sand impinges fittings and tubing resulting in loss of the wall material. The erosion damage can be
significant resulting in failure in a relatively short amount of time. Failures provide a potential safety

Copyright
@1999 by NACE International. Requests for permission to publish this manuscript in any form, in part or in whole must be made in writing to NACE
International, Conferences Division, P.O. Box 218340, Houston, Texas 77218-8340. The material presented and the views expressed in this
paper are solely those of the author(s) and are not necessarily endorsed by the Association. Pfinted in the U.S.A.

Giovanni Juzga Len - Invoice INV-849691-V5F2Z3, downloaded on 10/17/2014 4:06PM - Single-user license only, copying/networking prohibited.

Paper No.

BACKGROUND
Erosion results from the impingement of sand, and the impingements occur due to the transfer of
momentum of the fluid to the particle in the direction of the wall, In some cases such as the elbow, the
inefficiency of the fluid to redirect (change the momentum of) the particles results in impingements.
Accounting for the efficiency of the exchange of momentum is a key factor in predicting erosion. Once
again, the efficiency of momentum transfer depends on many factors; some of the most important being
the ratio of density of the particle and fluid and the particle size. To illustrate the effect of the efficiency
of momentum transfer, sample particle trajectories are determined for representative air and water flow
in a 1 inch elbow with standard radius of curvature. For both cases, the initial particle velocities are set
equal to the bulk fluid velocity and the particle size is 150 microns. Figure 1 shows the representative
particle trajectories in air at 50 psig with a bulk velocity of 50 ft/s. In air, the particles have sufficient
momentum to cross the streamlines and impinge the wall. The efficiency of exchange of momentum is
very low so the air does not alter the path of the particles, and the particles impinge with a velocity
similar to that of the bulk velocity of air. Figure 2 shows the representative particle trajectories for
water with a bulk velocity of 8 ft/s. For this case, the particles follow the streamlines much more closely
and only a few impingements occur. This behavior results since the density of the water is much larger
than the density of air and approaches the density of the particles.
Comparison of these figures also illustrates another important aspect of particle motion that
affects erosion. The particle trajectories for the water case demonstrate the effect of turbulence. The
particles are influenced by the instantaneous velocity of the fluid which is comprised of both the
fluctuating component resulting from turbulence and the mean component. There is also a fluctuating
velocity component associated with the air cases; however, due to the inefficiency of the transfer of
Some erosion such as in elbows occurs
momentum the particles do not react to the fluctuations.
primarily due to the mean component of velocity. The authors refer to this erosion mechanism as direct
impingement. However, in some geometries such as straight sections of tubing, particle impingements
only occur due to the fluctuating component of velocity. For these cases, the erosion mechanism is
random impingement.
Under most conditions, the erosion resulting from direct impingement is more
This is another reason that erosion damage in elbows is
severe than for the random impingements.
studied.
Previous guidelines for avoiding erosional damage by limiting production rates were based
For example, the recommended velocity limitation described by the
purely on empirical relations.
American Petroleum Institute Recommended Practice 14E (API RP 14E) gives a limiting production
velocity by the formula (1)

Giovanni Juzga Len - Invoice INV-849691-V5F2Z3, downloaded on 10/17/2014 4:06PM - Single-user license only, copying/networking prohibited.

risk for personnel as well as equipment. Oil and gas producers need to be able to predict the severity of
erosion so that the service life of fittings susceptible to erosion can be determined. If the prediction tool
has the capability of accounting for parameters such as production rate or flow velocity, producers can
also determine the highest production capacity that keeps the erosion damage within tolerable limits,
maximizing the economic potential of the well.
Erosion is complex and depends on a multitude of factors such as the fluid and sand properties,
production rate of produced fluid and sand, pipe size, and sand size. Therefore, developing predictive
tools for erosion poses a difficult task. Another factor that affects the severity of erosion is the type of
geometry. Most of the work that has been done on the development of erosion prediction models has
been for elbows and bends. This is a result of two primary factors. First, the elbow is a relatively
simple geometry, and second, the elbow experiences a significant amount of erosion.

(1)

JF

where pm is the fluid mixture density at flowing pressure and temperature, and in Equation (1) has the
units of lb/ft3; the units of the fluid velocity, Vc,is in ft/s. The API formula is very simple and easy to
use, but, as noted in the literature [2,3], the formula does not recognize many factors contributing to
erosiordcorrosion, and the use of this formula can result in unrealistically low production velocity limits
for preventing pipe damage in erosive service. The only physical variable accounted for in Equation (1)
is the fluid density. The formula suggests that the limiting velocity could be increased when the fluid
density is decreased. This does not agree with experimental observations for sand erosion, because sand
in gases with lower densities will cause higher erosion than liquids with higher densities.
Recent methods for predicting threshold velocities in producing wells are based on penetration
rates in an elbow geometry because this geometry is more susceptible to erosion damage than a straight
pipe section. A common procedure presented in the literature is to extrapolate a threshold flowstream
velocity relation from erosion rate data for an elbow geometry based on an allowable amount of erosion
(e.g., a penetration rate of 5 or 10 roils per year). Examples include work of Salama and Venkatesh [4],
Bourgo yne [5], Svedeman and Arnold [6]. Recently, some of these methods were also extended to
multiphase flow [3,7].
EROSION/CORROSION

RESEARCH CENTER (E/CRC) MODEL


(Application to Single-phase and Multiphase Flow)

It is well recognized by many researchers that predicting erosion is a very complex problem.
Current methods for erosion calculation are based on semi-empirical correlations that relate solid
particle impact velocity to pipe wall material loss rate. Thus, erosion calculation can be even more
complicated in multiphase flow systems where sand particles are entrained in a complex multiphase
flow and the particle impingement velocity and impingement location on the pipe wall must be
determined.
Recent models for predicting erosion do account for the flow effects and these models are
referred to as the generalized models because they can account for different pipe geometries. In the
generalized model, the erosion calculation is performed based on particle impingement angle and speed.
The generalized model requires significant flow modeling that are normally done by computational fluid
dynamics (CFD) codes which are still in their infancy for predicting multiphase flow. This approach is
also currently in progress at the E/CRC. The generalized model approach currently requires significant
computational effort and, in general, is not suitable for design calculations. Therefore, it is desirable to
devise simpler and more efficient mechanistic models, that are based on the generalized model and
empirical information, and do account for the important variables that affect erosion. Based on this idea,
a formula is proposed for computing penetration rates in elbows, tees and direct impingement
geometries. In this formula, an expression for computing the maximum penetration rate in steel material
is proposed of the form

w Vp
h = FM Fs Fp F,m
D2

(2)

Giovanni Juzga Len - Invoice INV-849691-V5F2Z3, downloaded on 10/17/2014 4:06PM - Single-user license only, copying/networking prohibited.

ve=

h
FM

= penetration rate, m/s (can be converted to mm/yr or mpy)


= an empirical constant that accounts for material hardness
For carbon steel materials, FM= 1.95 x 10-5/ B-059(for VL in m/s) where B is the
Brinell hardness factor
= empirical sand sharpness factor

= penetration factor for steel (based on 1 pipe diameter), m/kg

r/D

= penetration

w
VL

= sand production rate, kg/s


= characteristic particle impact velocity, rrds
ratio of pipe diameter in inches to a one inch pipe

D=

factor for elbow radius

The relation was developed based on extensive empirical information


gathered at the
Erosion/Corrosion Research Center and data gathered at the Texas A&M University [8,9]. A major
difference between the E/CRC model and the earlier work is that method was developed to find the
characteristic impact velocity of the particles on the pipe wall, VL. This characteristic impact velocity
of the particles depends on many factors including pipe geometry and size, sand size and density, flow
regime and velocity, and fluid properties.
Characteristic

Impact Velocity of Sand Particles

A method for computing the characteristic impact velocity of the particles with the pipe wall for
a tee and an elbow geometry has been presented in previous work [10, 11]; a brief description is given in
this section. The characteristic impact velocity of the particles is obtained by creating a simple model of
the stagnant layer in a pipe geometry. This is done by relating the erosion rates of complex geometries
to erosion occurring in a direct, or normal, impingement situation.
This procedure is presented
conceptual y in Figure 3. In order to impinge the target wall, the sand particles must penetrate the fluid
layer (so called stagnation zone) for each of the geometries that are shown in Figure 3. The behavior of
the particles in the stagnation region strongly depends on the pipe fitting geometry, fluid properties, and
sand properties. For two phase flow, this region can be assumed to be composed of gas and liquid
phases according to the flowing volumes of gas and liquid. Thus, the impact velocity of the sand
particles with a pipe wall is a strong function of the fluid properties as well as the amounts of gas and
liquid phases that are present in the stagnation zone or stagnation length through which the particles
must travel in order to strike the pipe wall. A characteristic length, called the equivalent stagnation
length (Figure 3), is used to represent this distance. A simplified particle tracking is used in this region
to determine the so called characteristic impact velocity of the particles.
The equivalent stagnation lengths for an elbow as well as a tee geometry were obtained by
erosion testing, flow modeling and particle tracking results [10, 11]. Figure 4 shows the results which
could be used for estimating the equivalent stagnation length for an elbow and a tee geometries. This
figure shows how the equivalent stagnation length varies with the pipe diameter, D.
A simplified particle tracking model [11] was used to compute the characteristic impact velocity
of the particles. This model assumes that the particle is traveling through a one-dimensional flow field
that is assumed to have a linear velocity in the direction of the particle motion and uses a simplified drag
coefficient model. The initial particle velocity was assumed to be the same as the flowstream velocity.
However, in a two phase gas-liquid flow, sand is normally entrained in the liquid phase. Thus, it is
reasonable to assume that sand particles have the same velocity as the liquid velocity (this is not the
same as the liquid superficial velocity) in the two-phase flow mixture. Assuming this equivalent or
characteristic liquid (and sand) flowstream velocity is known before the sand particles reach the

Giovanni Juzga Len - Invoice INV-849691-V5F2Z3, downloaded on 10/17/2014 4:06PM - Single-user license only, copying/networking prohibited.

where

Re

Pm

(3)

oI%n

where

V.

Pm

= density of fluid (mixture) in the stagnation layer, kg/m3

l-%
dp

= viscosity of fluid (mixture) in the stagnation layer, Pa-s (or N s/m2)


= diameter of particle, m

equivalent flowstream velocity, m/s

The density and viscosity of the fluid in the stagnation layer is computed based on the volume flow of
gas and liquid at the flowing conditions,
~m = QLPL + QGPG

. SLPL

QL + QG

+ SG~G

(4)

VSL + VsG

(5)
where
QL

= volume flow rate of liquid, m3/s

QG

= volume flow rate of gas, m3/s

SL

= superficial liquid velocity, m/s

SG

= superficial gas velocity, m/s

A dimensionless parameter @ which is proportional to the ratio of the mass of the fluid being displaced
by the particle to the mass of the impinging particle:

Pm
@=

(6)

dpPp
where
L

= equivalent stagnation length, m

PP

density of particle, kg/m3

Figure 5 contains much useful information about how various parameters affect the characteristic
For example, it shows how the
impact velocity of the particles, VL, and sand erosion rates.

Giovanni Juzga Len - Invoice INV-849691-V5F2Z3, downloaded on 10/17/2014 4:06PM - Single-user license only, copying/networking prohibited.

stagnation zone, a simplified particle tracking model, identical to the one used previously [10], can be
used to determine the characteristic impact velocity of the particles.
The results for the characteristic impact velocity of the particles with the pipe wall as a function
of several production parameters are shown in Figure 5. The impact velocity for two-phase flow
depends on: equivalent liquid flowstream velocity, a characteristic length scale describing geometry
and size, L, density of the fluid in the stagnation zone, viscosity of the fluid in the stagnation zone,
density of the particle, and diameter of the particle. As done previously for single-phase flow, these
parameters can be combined into three dimensionless groups related to one another as shown in Figure
5. The dimensionless groups are:
A particle Reynolds number, ReO, based on the equivalent flowstream velocity and particle diameter:

Jordan [7] developed a simple equation that represents


@/ReO <0.153,

the curves that are shown in Figure 5 for

(7)

For

@/ ReO >0.153,

the particles do not have enough momentum

to strike the pipe wall and the

characteristic particle impact velocity is zero. Figures 5 and 6 as well as Equation (7) are also valid for
single-phase flow. The mixture properties should be replaced by the appropriate property of the singlephase fluid. Additionally, the equivalent flowstream velocity should be replaced by the bulk velocity of
the single phase.
Summary of the E/CRC Procedure
In this section, the present method is presented in a step-by-step summary which demonstrates
how the method could be used to compute the penetration rates for an elbow or a tee geometry.

SEJm.
The first step in the procedure is to estimate the equivalent stagnation length, L. The semi-empirical
equivalent stagnation length depends on pipe geometry and is presented as a function of pipe diameter in
Figure 4 for tees and elbows. Note that the results in the figure are normalized with respect to a
reference stagnation length Lo which is the equivalent stagnation length for a 1 pipe diameter. The
values for Lo have been tabulated for several geometries in Table 1.
In order to estimate L, one obtains L/L. from Figure 4, and compute L using Lo value selected
from Table 1. Also shown in Table 1 are values of the penetration
discussed below.
-
The next step is to compute two dimensionless
Eq. (3), based on the equivalent flowstream

factor Fp used in Step #4 as

parameters, namely, the particle Reynolds number Reo,

velocity VO and particle diameter, and the dimensionless

parameter, 0, given by Eq. (6). For single-phase (gas or liquid) flow, one can assume that the
equivalent flowstream velocity is the same as the average flow velocity in the pipeline. For two-phase
flow, an ad hoc equation is used to calculate the equivalent liquid flowstream velocity based on the
superficial gas and liquid velocities.
When the superficial gas velocity is much larger than the superficial liquid velocity ( V~G >> V~L and
flow is annular), the liquid film thickness becomes a fraction of the particle size and can no longer
entrain the sand particles. Thus, the particles may be entrained in the gas phase and VO = V~~ should
be used.

Giovanni Juzga Len - Invoice INV-849691-V5F2Z3, downloaded on 10/17/2014 4:06PM - Single-user license only, copying/networking prohibited.

characteristic impact velocity of the particles is affected by fluid and sand properties.
Once the
characteristic impact velocity of the particles is determined, it is used in Equation (2) to compute erosion
and penetration rates for a specific geometry such as an elbow.

flowstream

velocity VO. Figure 6 presents the relationship

form tomakeit

easier toreadvaluesof

VL/

between @ and impact velocity in log-log

VO between O.01 and O.1. Alternatively,

used to compute the VL / Vo ratio. Then, compute the characteristic


v~ = (v~/vo)

Eq. (7) can be

impact velocity by

x V.

(8)

-.
Select particle sharpness factor, Fs, from Table 2, the penetration factor FP, from Table 1, and the
material factor, FM , from Table 3.
In order to obtain the elbow radius factor, Fr 1~, a general model that incorporates

a flow model,

particle tracking model, and erosion ratio model was used. This general model that is described in detail
by Wang et al. [12] was developed at the E/CRC. The model was verified for standard and long radius
elbows by showing excellent agreement with several sets of experimental data for a variety of conditions
provided by Bourgoyne [5], Tone and Greenwood [9], Eyler [13], and Bikbiaev [14]. Many simulations
of the general model were performed and various parameters effecting elbow erosion were studied.
Based on the analysis of particle motion in elbows, the following equation based on curve fitting of the
particle tracking results was recommended as a first-order approximation for elbow radius factor in long
radius elbow:

Fr/D=ex

0.1 ~4w~65 +o.015p:25+o.


d;.3

12

:c~~d
1

{[

(9)
1

where F~iDis the elbow radius factor for long radius elbows, c~~dis the r/D of a standard elbow (assumed
to be 1.5). Eq. (9) accounts for the elbow radius curvature effect in different carrier fluids and sand
particle size. It should be noted that this equation was based on the condition that the particle density is
approximately 165 lb/ft3.
-
Compute penetration rate by using Equation (2).

VERIFICATION

OF E/CRC MODEL

The ability to predict erosion in multiphase flow is a new addition to the E/CRC prediction
model.
Previously, the E/CRC erosion prediction model was verified for use with single-phase
flow[2, 15]. The ability of the model to account for a variety of parameters (particle diameter, fluid
density and viscosity, flow velocity, and particle shape) was verified through comparisons with
experimental data[ 16]. Another verification of the entire erosion prediction procedure for single-phase
flow in elbows (both standard and long radius) is shown here through experimental data. The
experimental data of Weiner and Tone, Tone and Greenwood, and Bourgoye is used for comparison.
Table 4 provides the test conditions along with the experimental and predicted results. Figure 7 is a
graphical representation of the predicted results versus the experimental results. This figure shows that
the present method provides accurate results over a broad range of erosion rates.

Giovanni Juzga Len - Invoice INV-849691-V5F2Z3, downloaded on 10/17/2014 4:06PM - Single-user license only, copying/networking prohibited.

Swt3.
Use @ and Re in Figure 5 or 6 to find the ratio of the particle impact velocity VL to the equivalent

EFFECT OF PARAMETERS

ON EROSION

The effect of several parameters such as flow velocity, pipe diameter, and sand size will be
shown for compressed methane and water for both single-phase and multiphase flow.
Single-Phase Flow
First, the effect of these parameters in single-phase flow is examined. The temperature of the
methane is 150 F and the pressure is 5000 psi; this corresponds to a density and viscosity of 8.28 lb/ft3
and 0.024 cp, respectively. The water also has a temperature of 150 W with a density of 61.7 lb/ft3 and a
viscosity of 0.432 cp. Unless varied to determine the effect of each parameter, the particle diameter is
150 microns and the diameter of the elbow is 4 inches with a standard radius of curvature (r/D = 1.5).
The flow velocity of the methane is 100 ft/s and the velocity of the water is 15 ft/s.
Figure 10 shows the effect of elbow diameter on the maximum penetration rate. For both the
compressed methane and water an increase in diameter decreases the severity of erosion. This occurs
since the impinging sand particles must pass through a larger stagnation region; therefore, there is more
time for the particles to decelerate. The effect of diameter is much greater for the water case. The
efficiency of transfer of momentum is larger for the water case, so the deceleration of the fluid in the
direction normal to the wall causes the particles to decelerate more efficiently.
Figure 11 examines the effect of sand size, and in both cases an increase in sand size results in
greater erosion. For the methane case, even the relatively small particles are able to pass through the
stagnation zone and impinge the elbow. However for the water case, the small particles travel with the
fluid and do not impinge the elbow wall. In water, the maximum penetration rate begins to decrease
rapidly with a decrease in particle size less than 150 microns (this is specific to the parameters elected
for this prediction).
The next parameter examined is flow velocity, which is shown on Figure 12. As expected, an
increase in flow velocity results in higher erosion rates. This figure displays similar behavior to Figure
11. The particles impinge resulting in erosion even at low velocities for the compressed methane, but
for the water case, the water is able to redirect the particles resulting in extremely low erosion for low
velocities. There is a sharp increase in penetration rate for the water case around 10 ftis; this appears to
be a threshold value for flow velocity for the water case.

Giovanni Juzga Len - Invoice INV-849691-V5F2Z3, downloaded on 10/17/2014 4:06PM - Single-user license only, copying/networking prohibited.

Data provided by Salama [3] and Bourgoyne [5] is also used to verify the present erosion
prediction model for multiphase flow. Table 5 provides the test conditions and the fluid and material
types. Figure 8 shows a summary of the results presented in Table 5, and these results indicate excellent
agreement between the experimental and predicted results over a broad range of erosion rates for
multiphase flow.
Another comparison of the present multiphase erosion prediction model is provided through
failure data obtained by Southwest Research Institute and Shell as reported by Jordan [7]. Since this is
field data and knowledge of the exact conditions is limited some assumptions had to be made in order to
apply the model. For this comparison, a sand size of 150 microns, pipe diameter of 2 inches, sand rate
of 0.06 ft3/day (10 lb/day), and a tolerable erosion rate of 5 mp y was selected. The density of the gas
and liquid at process conditions are 7.21 and 43.15 lb/ft3, respective y. The viscosity of the gas and
liquid are 0.0166 and 0.5369 cp, respectively. The fluid properties were suggested by Jordan [7]. The
material type selected for this comparison was AISI 1020. This information was used to predict the
threshold superficial gas velocity for a given superficial liquid velocity. Figure 9 shows the threshold
curve on the same plot as the failure data. The predicted threshold curve appears to capture the outline
of the shape that is created by the failure data. However, this is a first version of the erosion prediction
model for multiphase flow provided by E/CRC and plans are in place for further verification.

Multiphase Flow
Similar figures can also be created for erosion in multiphase flows in elbows. The values used
for these predictions are similar to those used for the single-phase cases. The primary difference is that
the compressed methane and water are now flowing together. The superficial liquid velocity for these
cases is set equal to 1 ft/s, and the superficial gas velocity is set equal to 25, 50, 100, and 150 ft/s.
Figure 14 examines the effect pipe diameter on erosion in the multiphase system. The behavior of this
family of curves is similar to the curve for compressed methane in Figure 10. In fact, the curve of
superficial gas velocity of 100 ftis of Figure 14 can be compared directly to the compressed methane
curve in Figure 10. The maximum penetration rates for the curve in Figure 14 are less than in Figure 10.
This demonstrates the effect of the presence of the liquid in the pipe. As the superficial gas velocity
decreases, the effect of the liquid becomes more prominent and the penetration rates decrease,
Figure 15 explores the effect of sand size. Once again, the effect of sand size in the multiphase
flow is similar to the effect in single-phase flow. As the sand size increases, the penetration rate also
increases. The curve for a superficial gas velocity of 25 ft/s demonstrates that a threshold particle
diameter exists for this case at around 50 microns. This behavior was also seen for the water case in
Figure 11, except the threshold particle diameter was larger.
The effect of flow velocity in terms of the superficial gas velocity is shown in Figure 16. For
this figure, the superficial gas velocity was varied but the superficial liquid velocity was maintained at 1
ft/s, This figure shows that an increase in superficial gas velocity increases the maximum penetration
rate. This occurs for two reasons. Primarily, the higher velocity provides the particles with greater
initial momentum.
Additionally at higher superficial velocities, the mixture properties for the flow
approach those of methane, which can not alter the trajectories of the particles as efficiently as water.
The effect of elbow diameter, sand size, and sand rate are also shown in Figures 17 through 19.
These figures show the threshold erosional velocity curve similar to that shown in Figure 9. In fact, the
fluid, particle, and material properties used to generate these figures are the same as for Figure 9. Figure
17 provides the theshold curve for elbow diameters of 1, 2,4, and 6 inches. Figures 10 and 14 showed a
decrease in erosion rate with increase in pipe diameter.
Since the penetration rate decreases with
increase in diameter, the threshold superficial gas velocity increases with an increase in diameter for a
given liquid superficial velocity. Effectively, an increase in diameter shifts the threshold curve to the
right allowing operation at higher superficial gas velocities.
Figure 18 is similar to Figure 17 but demonstrates the effect of particle diameter. Threshold
curves are shown for particle diameters of 50, 150, and 300 microns. Since smaller particles result in
less erosion, the threshold curves shift to the right for decreases in particle size. The effect of sand rate
is shown in Figure 19. Curves for sand rates of 1, 10, and 100 lb/day are shown. With less sand flowing
in the system, less erosion occurs. So for smaller sand rates, the threshold curve shifts to the right.
SUMMARY AND CONCLUSIONS
The effects of pipe diameter, sand size, flow velocity, and fluid properties on the severity of
erosion were examined for both single-phase and multiphase flows. The behavior or trend of the

Giovanni Juzga Len - Invoice INV-849691-V5F2Z3, downloaded on 10/17/2014 4:06PM - Single-user license only, copying/networking prohibited.

For Figures 10 through 12, the effect of fluid properties has been investigated by showing the
behavior for both compressed methane and water. Figure 13 examines the effect of fluid density for
compressed methane. These results are shown to display the effect of pressure on the penetration rate,
Table 6 shows the variation in the density and viscosity of methane with pressure using a temperature of
150~.
Themaximum penetration rate decreases with anincrease intensity (increase in pressure). As
the methane becomes more dense it increases the efficiency of momentum transfer, and the particles
follow the flow more closely.

ACKNOWLEDGEMENTS
The authors would like to thank the member companies
Center for their support and guidance.

of the Erosion/Corrosion

Research

REFERENCES
1.

API Recommended Practice for Design and Installation of Offshore Production Platfomn Piping
Systems, API RP 14E, American Petroleum Institute, Third Edition, Washington D. C.,
December 1981.

2.

Shirazi, S. A., McLaury, B. S., Shadley, J. R., and Rybicki, E.F., Generalization of the API RP
14E Guideline for Erosive Services, Journal of Petrolewn Technology (Distinguished Author
Series), August 1995, pp. 693-698.

3.

Salama, M.M., An Alternative to API 14E Erosional Velocity Limits for Sand Laden Fluids,
presented at the 1998 Offshore Technology Conference, Houston, OTC 8898, 1998.

4.

Salama, M.M. and Venkatesh, E. S., Evaluation of Erosional Velocity Limitations in Offshore
Gas Wells, 15th Annual OTC, Houston, Texas, May 2-5, OTC Number 4485, 1983.

5.

Bourgoyne, A. T., Jr., Experimental Study of Erosion in Diverter Systems Due to Sand
Production, paper presented at the 1989 SPE/IADC Drilling Conference, New Orleans,
SPIYIADC 18716, 1989.

6.

Svedeman, S .J. and Arnold, K.E., Criteria for Sizing Multiphase Flow Lines for
Erosive/Corrosive Service, paper presented at the 1993 SPE Conference, Houston, SPE 26569.

7.

Jordan, K., Erosion in Multiphase Production of Oil& Gas, Corrosion 98, Paper No. 58,
NACE International Annual Conference, San Antonio, April 1998.

Giovanni Juzga Len - Invoice INV-849691-V5F2Z3, downloaded on 10/17/2014 4:06PM - Single-user license only, copying/networking prohibited.

amount of erosion as a function of these parameters was similar for both single-phase and multiphase
flows. In fact, the predictions made by splitting the multiphase flow into its single-phase components
should provide the bounds for the predictions made for the multiphase cases. This is true since at low
liquid rates the annular film becomes extremely thin and does not provide any resistance for impinging
particles. At the other extreme if the gas rate is small, the gas will have little effect on the particle
motion and the particles will behave in a similar manner to cases with only liquid present.
For this work, the erosion was only examined in elbows.
It was shown that the erosion
prediction model must be able to capture the exchange of momentum between the fluid and the particles.
For elbows, much erosion occurs when the exchange of momentum between the fluid and the particles is
low. This is a result of the inability of the fluid to redirect the particles as they travel through the bend;
therefore, the particles travel in a relatively straight path and impinge the pipe wall. Any factor that
increases the efficiency of the exchange in momentum will decrease the erosion rate in elbows. This
includes decreasing the particle size or increasing the fluid density or viscosity.
Another way to
decrease the erosion rate is to increase the time allowed for the transfer of momentum.
This is
accomplished by increasing the diameter of the pipe or by using a long radius elbow. Finally, if the
initial momentum of the particles is less when entering the elbow, then there is less momentum that
needs to be overcome by the fluid. This explains why the flow velocity has such a tremendous effect on
the erosion rate.

Weiner, P.D. and Tone, G. C., Detection and Prevention of Sand Erosion of Production
Equipment, API OSAPER Project No. 2, American Petroleum Institute, Texas A&M Research
Foundation, March 1976.

9.

Tone, G.C. and Greenwood, D.R., Design of Fittings to Reduce Wear Caused by Sand
Erosion, API OSAPER Project No. 6, American Petroleum Institute, Texas A&M Research
Foundation, May 1977.

10.

Shirazi, S. A., Shadley, J.R., McLaury, B. S., and Rybicki, E. F., A Procedure to Predict Solid
Particle Erosion in Elbows and Tees, ASME PVP Vol. 259, Codes and Standards in a Global
Environment, 1993, pp. 159-167.

11.

McLaury, B. S., A Model to Predict Solid Particle Erosion in Oilfield Geometries,


Thesis, The University of Tulsa, 1993.

12.

Wang, J. and Shirazi, S. A., Shadley, J.R., and Rybicki, E. F., Application of Flow Modeling and
Particle Tracking to Predict Sand Erosion Rates in Elbows, in ASME FED Vol. 236, July 1996,
pp. 725-734.

13.

Eyler, R.L., Design and Analysis of a Pneumatic Flow Loop, M.S. Thesis, West Virginia
University, 1987.

14.

Bikbiaev, K. A., Krasnov, V. I., Maksimenko, M.I. Berezin, V. L., Zhilinskii, I.B., Main Factors
Affecting Gas Abrasive Wear of Elbows in Pneumatic Conveying Pipes, Chemical Petroleum
Engineering, Vol. 8, 1972, pp. 465-466.

15.

McLaury, B. S., Wang, J., Shirazi, S.A, Shadley, J. R., and Rybicki, E. F., Solid Particle Erosion
in Long Radius Elbows and Straight Pipe, paper 38842 presented at 1997 SPE Annual
Technical Conference and Exhibition, San Antonio, October 5-8.

16.

McLaury, B. S., Shirazi, S. A., Shadley, J. R., and Rybicki, E.F., Parameters Affecting Flow
Accelerated Erosion and Erosion-Corrosion, Corrosion 95, Paper No. 120, NACE International
Annual Conference, Orlando, April 1995.

M.S.

Giovanni Juzga Len - Invoice INV-849691-V5F2Z3, downloaded on 10/17/2014 4:06PM - Single-user license only, copying/networking prohibited.

8.

FACTORS
FP (for steel)

SHAPE

mm

Inch

mm/kg

in/lb

90 Elbow
Tee

30
27

1.18
1.06

206
206

3.68
3.68

Sharp Corners, Angular


Semi-Rounded, Rounded Corners
Rounded. !%herical Glass Beads

1.0
0.53
0.20

TABLE 3
MATERIAL PROPERTIES AND EROSION RATIO COEFFICIENTS
FOR NON-CARBON STEELS
Material
Yield
Tensile
Hardness
Material
Material
Type
Strength
Strength
Brinell
Factor* *
Factor*
Ksi
Ksi
B
FM X106
FM X107

1018

90.0
61
77
74
35
37

l-%%%%*

For VL in rnls

**

For VL in ft/s

++

99.5
105
93
111
85
91

210
190
180
217
183
160

0.833++
1.267
1.089
0.788
0.918
0.877

1.066
1.622
1.394
1.009
1.175
1.123

For carbon steel materials, FM= 1.95 x 10-5/ B-059(for VL in m/s) where B is the Brinell
hardness factor

Giovanni Juzga Len - Invoice INV-849691-V5F2Z3, downloaded on 10/17/2014 4:06PM - Single-user license only, copying/networking prohibited.

TABLE 1
SHAPE AND PENETRATION
Lo

Flow
Velocity
(m/s)

Note

OF EXPERIMENTAL DATA WITH PREDICTIONS


FOR SINGLE-PHASE FLOW
Sand
Size
(microns)

Elbow
Diameter
(mm)

Bend R

Measured
Erosion
mrnlkg

#Diameter

Pred.
Erosion
mmikg

1
1
1
1
1
1
1
1

9.15
12.2
15.25
18.3
21.35
24.4
27.45
30.5

300
300
300
300
300
300
300
300

52.5
52.5
52.5
52.5
52.5
52.5
52.5
52.5

1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5

2.14E-03
3.81E-03
7.52E-03
9.16E-03
1.22E-02
1.62E-02
1.80E-02
2.04E-02

2.66E-02
4.38E-02
6.45E-03
8.85E-O?
1.16E-02
1.46E-02
1.79E-02
2.14E-02

2
2

300
300

52.5
52.5

1.5
1.5

4.44E-03
1.56E-02

1.16E-02
2.14E-02

21.35
30.5
11.49

350

52.5

1.18E-06

2.12E-06

4
4
4
4
4
4
4
4
4
4

116
141
107
141
107
111
141
141
148
111

350
350
350
350
350
350
350
350
350
350

52.5
52.5
52.5
52.5
52.5
52.5
52.5
52.5
52.5
52.5

2.125
2.875
2.875
2.875
2.875
3.25
3.25
3.25
3.25
4.5

1.64E-01
1.75E-01
1.21E-01
1.74E-01
1.36E-01
1.12E-01
2.07E-01
1.91E-01
2.09E-01
5.26E-02

1.72E-01
2.18E-01
1.35E-01
2.18E-01
1.35E-01
1.37E-01
2.07E-01
2.07E-01
2.25E-01
1.15E-01

Notes:
1) Data from Weiner and Tolle8, Fluid is air at standard conditions, Material is assumed to be
carbon steel with a Brinell hardness of 109.
2) Data from Weiner and Tolle8, Fluid is air at standard conditions, Material is assumed to be
carbon steel with a Brinell hardness of 109, High sand rate
3) Data from Bourgoyne5, Fluid is drilling mud, Material is assumed to be carbon steel with a
Brinell hardness of 140.
4) Data from Bourgoyne5, Fluid is air at standard conditions, Material is assumed to be carbon steel
with a Brinell hardness of 140.

Giovanni Juzga Len - Invoice INV-849691-V5F2Z3, downloaded on 10/17/2014 4:06PM - Single-user license only, copying/networking prohibited.

COMPARISON

Superficial
Gas Vel.
(In/s)

Note

FOR MULTIPHASE FLOW


Superficial
Sand
Elbow
Bend R
Liquid Vel.
Size
Diameter
(m/s)
(microns)
(mm)
#Diameter

Measured
Erosion
mg

Pred.
Erosion
mm/kg

1
1
1
1
1
1
1
1
1
1

30
30
20
20
15
15
10
10
8
3.5

1
0.5
5.8
3.1
5
1
5
0.7
0.2
4

150
150
150
150
150
150
150
150
150
150

49
49
49
49
49
49
49
49
49
49

5
5
1.5
1.5
5
5
5
5
1.5
5

5.52E-04
2.46E-03
5. 19E-05
6.93E-05
6.38E-05
1.47E-04
1.35E-05
7. OIE-05
1.23E-04
4.60E-06

7.22E-04
1.27E-03
1.35E-04
1.58E-04
3.39E-05
9.63E-05
1.42E-05
4. 13E-05
1.49E-04
1.32E-06

2
2
2
2
2
2
2
2
2
2
2

9
14.4
14.6
34
35
34.3
37
38.5
44
51
52

6.2
1.5
1.5
2.1
1
0.5
0.7
0.5
1.5
0.6
0.7

250
250
250
250
250
250
250
250
250
250
250

26.5
26.5
26.5
26,5
26.5
26.5
26.5
26.5
26.5
26.5
26.5

5
5
5
5
5
5
5
5
5
5
5

1.80E-04
2.30E-04
4.20E-04
2.83E-03
6.56E-03
7.20E-03
8.03E-03
8.03E-03
1.05E-02
1.34E-02
1.33E-02

2.52E-04
4.99E-04
5. 13E-04
2.76E-03
4.44E-03
5.87E-03
5.98E-03
7.52E-03
6.00E-03
1.27E-02
1.25E-02

3
3
3
3
3
3
3
3

86
92
89
84
72
84
92
107

350
350
350
350
350
350
350
350

52.5
52.5
52.5
52.5
52.5
52.5
52.5
52.5

2.625
2.625
2.625
2.625
3,25
3.25
3.25
3.25

1.27E-01
1.21E-01
1.08E-01
9.34E-02
5.37E-02
7.51E-02
9.94E-02
1.05E-01

9.57E-02
1.08E-01
1.06E-01
9.19E-02
6.46E-02
8.46E-02
9.87E-02
1.28E-01

0.53
0.53
0.12
0.53
0.53
0.12
0.12
0.53

Notes:
1) Data from Salama3, Fluid is air and water at 2 bar, Material is carbon steel with a Brinell hardness
of 160.
2) Data from Salama3, Fluid is nitrogen and water at 7 bar, Material is Duplex stainless steel.
3) Data from Bourgoyne5, Fluid is air and water at standard conditions, Material is assumed to be
carbon steel with a Brinell hardness of 140.

Giovanni Juzga Len - Invoice INV-849691-V5F2Z3, downloaded on 10/17/2014 4:06PM - Single-user license only, copying/networking prohibited.

COMPARISON

TABLE 5
OF EXPERIMENTAL DATA WITH PREDICTIONS

Figure 1. Sample Particle Trajectories in Air.

Giovanni Juzga Len - Invoice INV-849691-V5F2Z3, downloaded on 10/17/2014 4:06PM - Single-user license only, copying/networking prohibited.

TABLE 6
EFFECT OF PRESSURE ON DENSITY
AND VISCOSITY
Pressure
I Density
I Viscosity
(psi)
(lblft)
(Cp)
14.7
0.036
0.0124
50
\
0.122
i
0.0124
100
0.245
0.0124
250
0.613
0.0126
500
1.25
0.0129
700
1.77
0.0134
1000
2.58
0.0139
2500
6.81
0.0191
5000
12.0
0.0239
7000
14.7
0.0318
10000
17.5
0.0350

Stagnation
Zone

Equivalent Stagnation

Length

*---

k
Particle initial
Position

Tee

Linear Fluid Velocity

Stagnation
Zone

Elbow

Figure 3. Concept of Equivalent Stagnation Length.

Giovanni Juzga Len - Invoice INV-849691-V5F2Z3, downloaded on 10/17/2014 4:06PM - Single-user license only, copying/networking prohibited.

Figure 2. Sample Particle Trajectories in Water.

Tee ---------------

3.0

- -----------

2.5
-/-

2.0

1.5

--

-.

....

= 1- 1.27 Tan-1(1,01

D-189)+D0129

~-- - ---- Lo=l,18jn -- - - - -

1,0

0,0

- - - ---

!~
... . . . . . . . . . . . . . . . . . . . . . . . . .

0.5

----

Giovanni Juzga Len - Invoice INV-849691-V5F2Z3, downloaded on 10/17/2014 4:06PM - Single-user license only, copying/networking prohibited.

3,5

---

Tee

Elbow

. .

Curve Fit Tee

1111111111

....

Curve Fit Elbow

1111111111

II

23456789

. ... ..

10

Inner

Diameter

11

(inches)

Figure 4. Stagnation Length versus PipeDiameter

forElbow.

1!0

09
Km

0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0

I I
1E-2

lE-1

lE+l

lE+O

[H

@=

Pm

dP

pP

1E+2

IE+3

Figure 5. Effect of Different Factors on Particle Impact Velocity.

12

0.10

i-!.!li
ReO=

0.01
lE-2

lE-1

lE+O

lE+l

1E+2

1E+3

[H

m=

Pm

dP

pP

Figure 6. Effect of Different Factors on Particle Impact


Velocity (log-log scale).
lE+O

lE-1

lE-2

lE-3
Perfect

Agreement

IE-4

1 E-5

IE-6

~ ~ , , 1,,!(,
IE-5
lE-6

n
1 E-4

Measured

IE-3

lE-2

lE-1

lE+O

Eroion Rate (mm/kg)

Figure 7. Comparison of Experimental Data with Predictions for Single-Phase

Flow.

Giovanni Juzga Len - Invoice INV-849691-V5F2Z3, downloaded on 10/17/2014 4:06PM - Single-user license only, copying/networking prohibited.

1.00

IE-I
Perfect

Agreement

IE-2

..0 . . .. . . . . . . . . . . . . .

IE-3

~l,,lrl,

a=
.................
8

IE-4

IE-5

lE-6

lE-6

IE-5

IE-4

Measured

~ -

--

lE-3

lE-2

Eroion

Figure 8. Comparison of Experimental


1E+2

Rate

IE-I

lE+O

(mm/kg)

Data with Predictions for Mulitphase Flow.

--------

~~~---

--

Threshold Erosional

<
=

lE+l
o

00

%
.
;

IE+O

2
v.

00

0
0

z
.
-1
~
0
~
a)

lE-1

: ----

~~~~~~~~~~-----0
A A

AA
Am

-0----

E-2:

A-A-

-----

m:,,,,,,,,,

,,,,,,,

lE-3
lE-I

Giovanni Juzga Len - Invoice INV-849691-V5F2Z3, downloaded on 10/17/2014 4:06PM - Single-user license only, copying/networking prohibited.

lE+O

lE+O
Superficial

lE+l
Gas

1E+2
Velocity

(ft/s)

Figure 9. Threshold Erosional Velocity Curve with Failure Data.

1E+3

1E+2

lE+l

lE+O

\.\.
r---;;;
. .\........
\
\ . -Methane

(5000

psi, 150 OF)

-.
------------ -- . - ..-..=...

--------.

----

D
lE-2

-.

4 ---
--
----

/
/

r
lE-1

:-

Water

(150

OF)

IE-3

1 E-4

(
2

Elbow Diameter

(Inches)

Figure 10, Effect of Elbow Diameter on Penetration Rate for Single-Phase


1E+3 ~... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

IE+2
Methane

IE+l

IE+O

(5000

psi, 150 OF) - - -----. . . ----~ . . . . . . . -------------------------------

1----

----------

/ .-

i-

P
r

IE-1

IE-2

-l
IE-3

Giovanni Juzga Len - Invoice INV-849691-V5F2Z3, downloaded on 10/17/2014 4:06PM - Single-user license only, copying/networking prohibited.

1E+3

~----------'"-`"/""--""""""""""""""""""""""-""""--"-"""""""""

lE-4
50

100

150
Sand

200
Size

250

300

(Microns)

Figure 11. Effect of Sand Size on Penetration Rate for Sinlge-Phase

Flow.

3-

1E+2 zMethane
IE+l

lE+O

(5000
.

----

psi, 150 OF)


-----.. . . . . . . . . . . . . .

/
/
.. .. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

.. . ..... .. . ...

lE-I
i
lE-2

:--

lE-3

:-

P
/

Water

(150 oF)

D
I

f
1 E-4

20

40

60
Flow

80

100

Velocity

120

140

1
160

(ft/s)

Figure 12. Effect of Flow Velocity on Penetration Rate for Single-Phase

Flow.

1E+3

1E+2

lE+l

10

Fluid Density

12

14

16

18

(lb/ft3)

Figure 13. Effect of Methane Density on Penetration Rate for Single-Phase

Flow.

Giovanni Juzga Len - Invoice INV-849691-V5F2Z3, downloaded on 10/17/2014 4:06PM - Single-user license only, copying/networking prohibited.

1E+3

/
.
1P
/------
/

1E+2

,K
*
.
.\

lE+l

. . . ..\

v
. . .. ..<~g

. . ..'
.

= 150 ftls
........................................

\-~.

-.

\-

\-

lE+O

-+
---

--

lE-1

1
I

Elbow

Diameter

(Inches)

Figure 14. Effect of Elbow Diameter on Penetration Rate for Multiphase Flow,
1E+2

---v~g= 150 ftls


--. . . ..-. . . . . . . . . . . . . . .
/
-------

lE+O

r.

--:---.-----<-------/

/-/

/
.

is;=

100 ft/s

----

. . . . . . . . . ....=. .. . . . . . . .. . .. .
-v~g= 50 ftls

1 E-1

---- 0-----------------/
6

V~g= 25 ftls

IE-2
/

r
/
lE-3

D
lE-4

Giovanni Juzga Len - Invoice INV-849691-V5F2Z3, downloaded on 10/17/2014 4:06PM - Single-user license only, copying/networking prohibited.

1E+3

~P
~
50

I
1

100

150

Sand Size

200

250

300

(Microns)

Figure 15. Effect of Sand Size on Penetration Rate for Multiphase Flow.

1
IE+2

4
4

lE+l

lE+O

. ...

...

I
4

lE-I

lE-2

~-

lE-3

1
,I

I
20

,
40

60

Superficial

80
Gas

100

120

Velocity

140

1
160

(ft/s)

Figure 16. Effect of Superficial Gas Velocity on Penetration Rate.


E+3
7"""" -"""" """"" """"" """"""""-"-""--"-"""""""""-""""""""""--"-""-"""-.
___
_.
\
-.
.
E+2 ~----

. . . .. . .---. . . . . . . . . . . . .-. . . . . . . . . . ..
-.

. ... . . ..\
\

D= 4,

D= 21:\
lE+l

..\ . . . . . . . . . ./. . ...!..

3-

J
I

D=l

. . .

//

6
I
~j:.D.=.
..
L.......
I
I
I
/
......
. ..G.
J. --------------------

lE+O

1 E-1

T
1
1

I
I

-1-------I
I

I 1

100

10
Superficial

II

I
1

I----J

1 E-2

lE-3

. . . . . ..\ . . . . . . . . .
\*

Gas

Velocity

(ft/s)

Predicted Threshold Erosional Velocity Curves for Different Elbow Diameters,

Giovanni Juzga Len - Invoice INV-849691-V5F2Z3, downloaded on 10/17/2014 4:06PM - Single-user license only, copying/networking prohibited.

1E+3

~--dP= 50 micron

1E+2

~---

lE+l

:----

--f-

. .\ . . . . . . . . . . . .

dP= 300 micronsI


\
#

i
~----

lE+O

. ..jt...jl

. . . . . . . . .
/

/
lE-1

..-.r- ..

. .

..J. J. . . . . . . . . . .

I
I

lE-2

. .

~----

lE-3

1:
II
I , ( , r,

100

10
Superficial

Gas

Velocity

(ft/s)

Figure 18. Predicted Threshold Erosional Velocity Curves for Different Sand Sizes.

Sand

Sand

rate = 1 lb/day

rate = 10 lb/day

i ----:-----
Sand

. . . . . ..\ . . . . . . . . . . . . .

rate = 100 lb/dayK -


\
#

.. .
/

/
I

. . . . . . . . . . . . . ..i . . ...z..<.l.

,.1
lE-2

1
. . ... . . .

..!..

lE-1

I
I

1-----------:-----------I
I

100

10
Superficial

Gas

Velocity

(ft/s)

Figure 19. Predicted Threshold Erosional Velocity Curves for Different Sand Rates.

Giovanni Juzga Len - Invoice INV-849691-V5F2Z3, downloaded on 10/17/2014 4:06PM - Single-user license only, copying/networking prohibited.

1E+3

You might also like