Download as pdf or txt
Download as pdf or txt
You are on page 1of 9

Engineering Failure Analysis 22 (2012) 83–91

Contents lists available at SciVerse ScienceDirect

Engineering Failure Analysis


journal homepage: www.elsevier.com/locate/engfailanal

Numerical analysis of flow erosion on drill pipe in gas drilling


Hongjun Zhu a,⇑, Yuanhua Lin a, Dezhi Zeng b, Ying Zhou c, Juan Xie c, Yuanpeng Wu c
a
State Key Laboratory of Oil and Gas Reservoir Geology and Exploitation (Southwest Petroleum University), Chengdu, Sichuan 610500, China
b
CNPC Key Lab for Tubular Goods Engineering (Southwest Petroleum University), Chengdu, Sichuan 610500, China
c
School of Material Science and Engineering (Southwest Petroleum University), Chengdu, Sichuan 610500, China

a r t i c l e i n f o a b s t r a c t

Article history: The majority of this paper investigates the factors affecting the erosion rate of drill pipe
Received 24 September 2011 using the computational fluid dynamics (CFDs) methodology. A three-dimensional drill
Received in revised form 11 January 2012 pipe-casing annulus including drill pipe connector is selected as computational domain
Accepted 19 January 2012
in this investigation. CFD models are proposed to obtain the flow field distribution in annu-
Available online 30 January 2012
lus and predict the erosion rate of drill pipe, which include both a gas-cuttings two-phase
hydrodynamic model and an erosion model. The impacts of gas velocity and drill pipe
Keywords:
eccentricity on erosion rate have been reasonably captured. Based on the analysis of calcu-
Flow erosion
Drill pipe
lated flow characteristics, a structure optimized scheme by reducing the slope of drill pipe
Wear connector is put forward to decrease the erosion rate and increase the life of drill pipe.
Gas drilling Ó 2012 Elsevier Ltd. All rights reserved.
Numerical simulation

1. Introduction

For gas drilling, flow velocity in drill pipe-casing annulus should be high enough to carry cuttings to surface. However,
high-speed cuttings–carrying flow wears drill pipe continuously, especially for drill pipe connector. As shown in Fig. 1, cracks
and corrosion pits usually present after a short period of time, due to fatigue and stress corrosion. Then high-speed cuttings–
carrying flow would accelerate cracking and deepen pitting, easily leading to drill pipe leakage, drilling troubles, or even per-
sonnel injury. This effect is believed to be piping erosion wear [1,2], which is a major factor in shortening the life of drill pipe.
Just in 2010, such drill pipe failure accidents account for 62.9% in one oil field in China. Therefore, the study of erosion effects
caused by gas-cuttings two-phase flow is urgently needed in gas drilling engineering.
Several studies have been carried out to investigate erosion phenomena [3–6]. There is a general agreement that the three
most important factors governing erosion are the velocity of erosive particles, angle of impact and type of material [7]. And
some empirical or semi-empirical correlations have been built to predict the erosion rate of piping. However, the erosion
phenomena strongly depend on the piping structure, piping layout, and flow localized distribution, etc. So the erosion rate
cannot be obtained by the previous studies realistically [8]. In addition, experimental conditions are considerably different
from practical drilling conditions. Thus, in this paper, CFD modeling and simulation method is proposed to explore the ero-
sion failure of drill pipe in gas drilling. The flow field distribution in drill pipe-casing annulus and the erosion rate of drill pipe
are captured under several working conditions. And the main factors affecting the erosion rate are discussed. Finally, a struc-
ture optimized scheme by reducing the slope of drill pipe connector is put forward.

⇑ Corresponding author. Tel.: +86 02883032206.


E-mail address: ticky863@126.com (H. Zhu).

1350-6307/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.engfailanal.2012.01.007
84 H. Zhu et al. / Engineering Failure Analysis 22 (2012) 83–91

Nomenclature

A coefficient taken as aq4:14


Af projected area of particles in the wall
B coefficient taken as a2q :65
b1 constant taken as 0.186
b2 constant taken as 0.653
b3 constant taken as 0.437
b4 constant taken as 7178.741
CD drag coefficient
C1e empirical constant taken as 1.44
C3e empirical constant taken as 0.09
C2 empirical constant taken as 1.9
ds particle diameter
E time-averaged strain rate
ess restitution coefficient of particle collision
~
F body force per unit mass
~
F lift lift force per unit mass
~
Fvm virtual mass force per unit mass
~
g gravitational acceleration
g0,ss solid-phase radial distribution function
Gk production term of turbulent kinetic energy due to the average velocity gradient
Gb production term of turbulent kinetic energy due to lift
K momentum exchange coefficient
k turbulent kinetic energy per unit mass
m_ mass transfer rate from one phase to another phase
m_s mass rate of particles
Nsd number of particles
p pressure
Reds particle equivalent Reynolds number
v
~ velocity
vr end speed
YM impact of compressible turbulence inflation on the total dissipation rate
a volume fraction
as,max the maximum radial solid volume fraction
rk Prandtl number corresponding to turbulent kinetic energy
re Prandtl number corresponding to turbulent kinetic energy dissipation rate
e turbulent kinetic energy dissipation rate per unit mass
q density
l dynamic viscosity
t molecule kinetic viscosity
s pressure strain tensor
st particle relaxation time
Hs particle temperature

Subscripts
q gas phase
s solid phase
mx mixture

2. Numerical procedure

2.1. Governing equations

The flow condition within drill pipe-casing annulus belongs to compressible and two-phase flow. Therefore, Euler–Euler
multiphase hydrodynamic approach [9,10] is adopted to calculate the gas-cuttings two-phase flow field in annulus, which is
presented as follows:
@
v q Þ ¼ m_ sq
ðaq qq Þ þ r  ðaq qq~ ð1Þ
@t
H. Zhu et al. / Engineering Failure Analysis 22 (2012) 83–91 85

Fig. 1. Macro photographs of failed drill pipes: (a) crack in connector slope, (b) corrosion pits in drill pipe connector.

@
v s Þ ¼ m_ qs
ðas qs Þ þ r  ðas qs ~ ð2Þ
@t
@
@t
v q Þ þ r  ðaq qq~
a qq ~
ð q v q Þ ¼ aq rp þ r  sq þ aq qq~g
v q~
ð3Þ
þaq qq ð~
Fq þ ~
F lift;q þ ~ vs  ~
F v m;q Þ þ K sq ð~ v q Þ þ m_ sq~
v sq
@
@t
v s Þ þ r  ðas qs~
a qs ~
ð s v s Þ ¼ as rp  rps þ r  ss þ as qs~g
v s~
ð4Þ
þas qs ð~ Fs þ ~
F lift;s þ ~ vq  ~
F v m;s Þ þ K qs ð~ v s Þ þ m_ qs~
v qs
where Ksq is the momentum exchange coefficient between gas and cuttings expressed as:
3as aq qq
K sq ¼ vs  ~
C D j~ vqj ð5Þ
4v 2r;s ds

where,
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
v r;s ¼ 0:5ðA  0:06Res þ ð0:06Res Þ2 þ 0:12Res ð2B  AÞ þ A2 Þ ð6Þ

The pressure of solid phase is computed from:

ps ¼ as qs Hs þ 2qs ð1 þ ess Þa2s g 0;ss Hs ð7Þ

where,
"  1=3 #1
as
g 0;ss ¼ 1  ð8Þ
as;max
Due to the high velocity of cuttings–carrying gas flow, realizable k–e turbulence model [11,12] is employed in this study
and can be expressed as follows:
  
@ðqmx kÞ @ðqmx kui Þ @ lt @k
þ ¼ lmx þ þ Gk þ Gb  qmx e  Y M ð9Þ
@t @xi @xj rk @xj
  
@ðqmx eÞ @ðqmx eui Þ @ lt @ e e2 e
þ ¼ lmx þ þ qmx C 1 Ee  qC 2 pffiffiffiffiffi þ C 1e C 3e Gb ð10Þ
@t @xi @xj re @xj k þ te k
where,
 
g
C 1 ¼ max 0:43; ð11Þ
gþ5

k
g ¼ ð2Eij  Eij Þ1=2 ð12Þ
e
 
1 @ui @uj
Eij ¼ þ ð13Þ
2 @xj @xi
86 H. Zhu et al. / Engineering Failure Analysis 22 (2012) 83–91

After the mainstream flow field such as velocity and pressure distribution obtained by solving the above equations, tur-
bulent properties of solid particles can be described by a particle motion equation, especially the trajectory of cuttings, the
attack angle and velocity perturbation [13,14]. This particle motion equation is expressed as:
dv s C Ds Reds gðqs  qmx Þ q dðv mx  v s Þ
¼ ðv mx  v s Þ þ þ 0:5 mx ð14Þ
dt 24st qs qs dt
where,

qs d2s
st ¼ ð15Þ
18lmx

qmx ds jv s  v mx j
Reds ¼ ð16Þ
lmx

24 b3 Reds
C Ds ¼ ð1 þ b1 Rebds2 Þ þ ð17Þ
Reds b4 þ Reds
Then the erosion rate can be determined by the mass transfer rate of magnetite on the metal surface, which is formulated as
follow:
XN sd
1:8  109 m
_s
Rc ¼ ð18Þ
sd¼1
A f

In simulation, gas is seemed as continuous phase, and cuttings are added into continuous phase flow field as discrete
phase spherical particles. All computations are performed using the finite-volume commercial CFD code, FLUENT 13.0.
SIMPLE algorithm is applied to solve the pressure–velocity coupling. Second-order upwind scheme is employed to solve
the momentum equation. And discrete phase model (DPM) is used to simulate the discrete phase based on Lagrangian ap-
proach. The convergent criteria for all calculations are set as that the residual in the control volume for each equation is
smaller than 105.

2.2. Computational domain

API £127 mm drill pipe and £273 mm casing are employed in present study. Fig. 2 shows the schematic of solution do-
main for simulation, which is a three-dimensional drill pipe-casing annulus including drill pipe connector. The slope of API
£127 mm drill pipe connector is 18°. And the upstream and downstream lengths of connector are 1 m and 2 m, respectively.
Drill pipe and casing are concentric under normal working condition. If eccentricity presents, the computational domain
should be re-established. And the eccentric distance selected in present simulation is 25 mm or 50 mm.

(a) outlet (b)


cuttings

casing
d
L3

Ro
Rc

A-A

A A L2

?
drillpipe
connector

L1
drillpipe
inlet

Fig. 2. Computational domain: (a) schema of cuttings–carrying gas flow in annulus; (b) geometric model.
H. Zhu et al. / Engineering Failure Analysis 22 (2012) 83–91 87

cross section

Fig. 3. Grid distribution of computational domain.

GAMBIT 2.3 mesh-generator is employed to perform all geometry generation and meshing. As shown in Fig. 3, computa-
tional domain is divided into several blocks, and each block is discretized with hexahedron cells in order to control the grid
distribution. Progressive mesh is performed to discretize the flow channel-changing region. And boundary layers present
near the drill pipe and casing wall with four-layer grids (the cell height of the first layer is 1 mm and the growth factor is
1.3). The suitable grid density is reached by repeating computations until a satisfactory independent grid is found.

2.3. Boundary conditions

The velocity of gas and mass flow rate of cuttings are set at the inlet. In simulation, the velocity of gas is defined as
27.78 m/s, 34.72 m/s or 41.67 m/s, corresponding to 120 m3/min, 150 m3/min, and 180 m3/min injecting gas volume rates
respectively. The mass flow rate of cuttings is defined as 0.6 kg/s, corresponding to 10 m/h penetration rate. And the average
diameter of cuttings is assumed as 1 mm. Pressure at outlet is set as 0 Pa in order to facilitate comparative analysis. No-slip
boundary condition is imposed at the drill pipe and casing wall. And wall function method is adopted for the turbulent flow
behaviors change abruptly near the wall. In addition, drill pipe and casing wall are set as reflection boundaries for discrete
particles, namely that particles will bounce back to the mainstream after impinging wall.

3. Results and discussion

3.1. Impact of gas velocity

Fig. 4 shows flow field distribution of gas-cuttings two-phase flow and erosion rate of drill pipe under three different gas
inlet velocities (27.78 m/s, 34.72 m/s, and 41.67 m/s). Due to friction loss, there is a significant pressure drop along the flow
channel. And the larger the gas inlet velocity, the greater is the pressure drop, because friction loss is proportional to the
velocity of flow. Sudden change of pressure is very clear in drill pipe connector. Localized high pressure presents in the wind-
ward slope of connector, where gas flow velocity decreases rapidly with the impingement on the slope wall. And localized
low pressure presents in the leeward slope of connector, resulting in eddy scouring the downstream drill pipe wall repeat-
edly. The value of localized high pressure and the absolute value of localized low pressure both increase with the increase of
gas inlet velocity.
It is clearly seen in Fig. 4 that flow velocity changes with the change of flow cross-sectional area. High speed occurs in
annulus between the drill pipe connector and casing, where the flow cross-sectional area is small. As wall shear stress is pro-
portional to flow velocity, the value of wall shear stress is larger in drill pipe connector wall than that in drill pipe body. And
the maximum value of shear stress presents at the localized low pressure zone. So the washout of connector wall by gas-
cuttings flow is serious, leading to pits-deepening or cracks-accelerating. And the larger the gas inlet velocity, the more
severe is the washout.
When flow reaches the drill pipe connector, solid particles would continue to flow along the windward slope wall, ren-
dering severe erosion wear occur there. As shown in Fig. 4, the maximum value of erosion rate presents in the windward
slope of connector. Then turbulent flow winds up and impinges on the drill pipe connector wall. Flow channel expands in
the leeward slope of connector, resulting in flow direction change and then pipe washout by gas-cuttings flow in a certain
angle. Thus, the value of erosion rate in the downstream of leeward slope is the second largest. With the increase of gas inlet
88 H. Zhu et al. / Engineering Failure Analysis 22 (2012) 83–91

P(Pa) v(m/s) τ (Pa) E(kg/(m2·s))

27.78m/s:

P(Pa) v(m/s) τ (Pa) E(kg/(m2·s))

34.72m/s:

P(Pa) v(m/s) τ (Pa) E(kg/(m2·s))

41.67m/s:

Fig. 4. Simulation results under different gas inlet velocities (pressure and velocity distributions in annulus, and erosion rate and wall shear stress
distributions in drill pipe wall).

velocity, the value of erosion rate has an obvious rise. When gas inlet velocity is 41.67 m/s, the maximum value of erosion
rate is about 11.27% larger than that when inlet velocity is 27.78 m/s.
Therefore, under meeting enough cuttings–carrying capacity, reducing gas injection pressure and displacement as much
as possible can effectively reduce the flow velocity in annulus and then reduce the erosion wear.

3.2. Impact of drill pipe eccentricity

If the drill pipe is eccentric in borehole, the flow channel in annulus would be divided into wide-gap channel and narrow-
gap channel. Fig. 5 shows flow field distribution of two-phase flow and erosion rate of drill pipe with three different eccentric
distances (0 mm, 25 mm and 50 mm). The larger the eccentric distance, the greater is the pressure drop along the longitudinal
H. Zhu et al. / Engineering Failure Analysis 22 (2012) 83–91 89

P(Pa) v(m/s) τ (Pa) E(kg/(m2·s))

0mm:

P(Pa) v(m/s) τ (Pa) E(kg/(m2·s))

25mm:
W N W N
G G G G
S S S S

P(Pa) v(m/s) τ (Pa) E(kg/(m2·s))

50mm:
W N W N
G G G G
S S S S

Fig. 5. Simulation results under different drill pipe eccentricities (pressure and velocity distributions in annulus, and erosion rate and wall shear stress
distributions in drill pipe wall; WGS: wide-gap side and NGS: narrow-gap side).

direction of flow channel. When the eccentric distance is 50 mm, the pressure drop is about 19.13% larger than that in con-
centric annulus.
Velocity distribution is also influenced by eccentric drill pipe. The larger the eccentric distance, the greater is the maxi-
mum value of flow velocity in wide-gap channel, and the smaller is the maximum value of flow velocity in narrow-gap chan-
nel. However, velocity in narrow-gap channel near connector slope has a larger change, leading to a more uneven
distribution of wall shear stress. The maximum value of wall shear stress in narrow-gap side presents at the location near
the windward slope of connector. Comparatively speaking, the wall shear stress in wide-gap side shows a more even
distribution.
Impact of drill pipe eccentricity on erosion rate is clearly indicated in Fig. 5. The larger the eccentric distance, the more
intense is the flow field change and then the more serious is the erosion, especially in the windward slope of connector. So
centralizers should be placed in proper positions to reduce drill pipe eccentricity.
90 H. Zhu et al. / Engineering Failure Analysis 22 (2012) 83–91

3.3. Re-design of connector slope

Erosion wear sites mainly locate at the slope of drill pipe connector from the above analysis. Therefore, reducing erosion
in connector becomes the key problem to make drill pipe last longer and ultimately perform safer drilling operations. Apart
from reducing the gas speed and drill pipe eccentricity, modifying the geometry of drill pipe connector may also reduce
erosion. In present study, the slope of connector has been reduced to observe its effect on erosion rate.
Fig. 6 shows flow field distribution of two-phase flow and erosion rate of drill pipe with three different connector slopes
(18°, 15°, and 12°). Reducing the slope of connector can make flow channel tend to be more streamlined. The value of pres-
sure drop along the longitudinal direction of flow channel, the maximum value of flow velocity in annulus and the value of

P(Pa) v(m/s) τ (Pa) E(kg/(m2·s))

18°:

P(Pa) v(m/s) τ (Pa) E(kg/(m2·s))

15°:

P(Pa) v(m/s) τ (Pa) E(kg/(m2·s))

12°:

Fig. 6. Simulation results under different drill pipe connector slopes (pressure and velocity distributions in annulus, and erosion rate and wall shear stress
distributions in drill pipe wall).
H. Zhu et al. / Engineering Failure Analysis 22 (2012) 83–91 91

wall shear stress in drill pipe outer wall have both decreased with the reduction of slope. And there is also a significant
reduction in the value of localized high pressure and the absolute value of localized low pressure in the slope of connector.
All the above indicate that flow field has been improved by reducing the slope.
There is a reduction in particle impact angle for a small connector slope, resulting in little erosion, as shown in Fig. 6. The
maximum value of erosion rate in 12° windward slope is 1.03  104 kg m2 s1, which is 34.8% smaller than the maximum
value of erosion rate in 18° windward slope. So reducing the slope of drill pipe connector is an effective method of modifying
geometry of drilling tools to reduce erosion.

4. Conclusions

A CFD methodology is employed in this paper to investigate the flow erosion on drill pipe in gas drilling. The following are
the main conclusions of this research:

(1) Flow field distribution and erosion rate can be obtained using CFD modeling and simulation method.
(2) Drilling operating parameters such as gas injection pressure and displacement, and the structure of drill pipe and drill
pipe-casing annulus both have obvious impacts on erosion. Both high gas velocity and large eccentric distance can lead
to severe wear, especially in the windward slope of connector.
(3) Apart from reducing the gas speed and drill pipe eccentricity, a structure optimized scheme by reducing the slope of
drill pipe connector is put forward to reduce the erosion, which has been also confirmed by simulation results.

Acknowledgements

Research work was co-financed by the China National Natural Science foundation and Shanghai Baosteel Group Corpo-
ration (No.: 51074135), program for New Century Excellent Talents in University (No.: NCET-08-0907), and the Fundamental
Research Funds for the Southwest Petroleum University (No.: 2010xjz133). Without their support, this work would not have
been possible.

References

[1] Ferng YM. Predicting local distributions of erosion–corrosion wear sites for the piping in the nuclear power plant using CFD models. Ann Nucl Energy
2008;35:304–13.
[2] Ferng YM, Lin BH. Predicting the wall thinning engendered by erosion–corrosion using CFD methodology. Nucl Eng Des 2010;240:2836–41.
[3] Postlethwaite J, Dobbin MH, Bergevin K. The role of oxygen mass transfer in the erosion–corrosion of slurry pipelines. Corrosion 1986;42:514–21.
[4] Postlethwaite J, Lotz U. Mass transfer at erosion–corrosion roughened surfaces. Can J Chem Eng 1988;66:75–8.
[5] Naitoh M, Uchida S, Okada H, et al. Evaluation of flow accelerated corrosion by coupled analysis of corrosion and flow dynamics (I) major features of
coupled analysis and application for evaluation of wall thinning rate. In: Presented at the 13th international topic meeting on nuclear reactor thermal
hydraulics, Kanazawa City, Ishikawa Prefecture, Japan; September 27–October 2, 2009.
[6] Uchida S, Naitoh M, Okada H, et al. Evaluation of flow accelerated corrosion by coupled analysis of corrosion and flow dynamics (II) flow dynamics
calculations for determining mixing factors and mass transfer coefficients. In: Presented at the 13th international topic meeting on nuclear reactor
thermal hydraulics, Kanazawa City, Ishikawa Prefecture, Japan; September 27–October 2, 2009.
[7] Arefi B, Settari A, Angman P. Analysis and simulation of erosion in drilling tools. Wear 2005;259:263–70.
[8] Tang P, Yang J, Zheng JY, et al. Failure analysis and prediction of pipes due to the interaction between multiphase flow and structure. Eng Fail Anal
2009;16:1749–56.
[9] Sokolichin A, Eigenberger G. Dynamic numerical simulation of gas-liquid two-phase flows Euler/Euler versus Euler/Lagrange. Chem Eng Sci
1997;52:611–26.
[10] Deen NG, Solberg T, Hjertager BH. Large eddy simulation of the gas-liquid flow in a square cross-sectioned bubble column. Chem Eng Sci
2001;56:6341–9.
[11] Kimura I, Hosoda T. A non-linear k-e model with realizability for prediction of flows around bluff bodies. Int J Numer Methods Fluids 2003;42:813–37.
[12] Maele KV, Merci B. Application of two buoyancy-modified k–e turbulence models to different types of buoyant plumes. Fire Saf J 2006;41:122–38.
[13] Sun L, Lin JZ, Wu FL, et al. Effect of non-spherical particles on the fluid turbulence in a particulate pipe flow. J Hydrodyn 2004;16:721–9.
[14] Dickenson JA, Sansaloned JJ. Discrete phase model representation of particulate matter (PM) for simulating PM separation by hydrodynamic unit
operations. Environ Sci Technol 2009;43:8220–6.

You might also like