Mathematical Modeling of Fluid Flow through Channels behind Well Casing

Authors: Guo B, Cai X, Lee J

Journal: Petroleum & Petrochemical Engineering Journal. Article type: Research Article. Published: 2020-08-10. DOI: https://doi.org/10.23880/ppej-16000231

Abstract

Oil and gas leakage through channels in the annular space of oil and gas wells has long been a problem responsible for sustained casing pressure and environmental issues. This type of channels forms due to low efficiency of cement placement during well cementing processes, gravity segregation after cementing horizontal wells, formation fluid invasion during the setting of cement slurry, and corrosion of formation gas, such as CO2 and H2S over the life of well service. Successful sealing of the channels requires a rigorous hydraulics model for simulating the friction of sealants in the channel. The challenge is from the irregular shape of cross sections of the channels. Owing to the nature of circular shapes of wellbore and casing, bow-shape cross sections are assumed to represent these irregular shapes of cross sections. An analytical model is presented to describe laminar flow in channels of bow-shaped cross sections. The model is validated by a comparison with a traditional rectangular slot model for narrow cross sections. Result of model analysis indicates that use of rectangular cross sections to approximate bow-shaped cross sections will under-estimate the pressure gradient in narrow cross sections and over-estimate the pressure gradient in wide cross sections. A case analysis of a cement squeezing operation shows that the newly developed hydraulics model for fluid flow in channels of bow-shaped cross sections is easy to use in engineering applications.

Keywords: Laminar flow, Slot, Cross section, Channel, Pressure, Model

Introduction

Fluid flow through channels in the cement sheath behind casing in oil and gas wells has long been a problem responsible for gas and oil leakage from gas and oil pay zones to surface or other formation zones, threatening surface and/or environment [1]. This type of channels forms due to low efficiency of cement placement during well cementing processes, gravity segregation after cementing horizontal wells, formation fluid invasion during the setting of cement slurry, and corrosion of formation gas, such as CO2 and H2S over the life of well service [2,3]. Detection and mitigation of the problem have been studied for a decade [4-7].

Squeeze cementing technology has long been available that allowed the first attempt success in mitigation of the flow-behind-casing problem [8]. The squeezing efforts can take days when multiple attempts are necessary. Experiments using various techniques to remedy the problem of flow behind casing have been costly and have consistently had low rates of success. Wireline logging, high density perforating and cement/resin squeezing have been applied. Cement squeezing without damaging formation is virtually impossible due to large cement particles bridging and hydration [9]. The repair of the problem is a non-revenue generating exercise with significant expenditures. The process involves a) identifying the fluid source or sources that are responsible for the problem and b) communicating with this fluid source in a manner that enhances a remedial cementing activity [10]. There are essentially three areas of squeeze cementing that are generally misunderstood and misapplied. They are 1) injection rates are conducted at excessively high rates and pressures and either create or perpetuate damage, 2) proper design of cement slurry properties is often neglected, and 3) the slurry is placed downhole at too high a rate because of the fear of cementing up a workstring in limited thickening time, which causes excessive formation breakdown, losing cement to the formation. These issues are inter-related through the slurry flow hydraulics in the channels behind the casing.

Effective mitigation of the problem requires adequate mathematical modeling of fluid movement behind casing [11]. However, such mathematical modeling work has not been done. Result of modeling should allow for selection of choice of treatment method, properties of blocking agent/ material, volume of blocking agent/material to be used, and placement method.

Owing to the nature of radial geometry of wellbore, it is expected the cross section of the channels in general takes an arc shape in one side and a near-straight shape in the other side, or bow-shape. Figure 1 presents images of channels in cement sheath observed by Goodwin and Crook in laboratory [12]. It indicates that the cross-section of wide channels such as that shown in Figure 1(a) can be approximated by a bow shape, while that of narrow channels shown in Figure 1(b) may be approximated by slots.

(a)

2 (b) Figure 1: Channels in cement sheath observed in laboratory [12].

Figure 2 shows a schematic of an idealized cross section of a water channel in the cement sheath. Mathematical modeling of fluid flow in channels with bow-shape cross-sections is not available from literature. But a method of solving similar flow problems with other irregular cross sections is found. Tamayol and Bahrami [13] presented analytical solutions for a fully developed laminar flow through hyper-elliptical and polygonal microchannels. The presented model enables the prediction of velocity distribution and pressure drop for several common fabricated geometries for industrial applications including circular, elliptical, rectangular, rhombus, triangular, and hexagonal ducts. Model-predicted values were successfully compared with experimental data collected by others for rectangular channels. Wu [14] performed a comprehensive investigation of the influence of cross-sectional shapes on channel flow. The cross-sectional shapes that he studied are circle, rectangle, ellipse, eccentric annulus, half-moon, circular sector, equilateral triangle, and limacon. The most popular model used in the oil and gas industry is still Bourgoyne, et al. [15] model for channels with cross sections of rectangular slot-shape, which is in question for accuracy when applied to channels with cross sections of bow-shape.

Figure 2: Cross-section of a bow-shaped water channel in oil and gas wells.

No study is found from the literature on the fluid flow in channels with bow-shaped cross-sections. This work presents an analytical model for the laminar flow of Newtonian fluids through channels with bow-shaped cross- sections and extends the model to non-Newtonian fluids using apparent Newtonian viscosity. The cross-section is a bow-segment of an oval/ellipse so that bow-segments of circles (_a=b_) and narrow slots (_a>>b_) are covered.

Mathematical Model

The standard equation to describe a planar ellipse shown in Figure 3 is written as

2 2 1 2 2 _x_ _y_

_a_ _b_ + = (1)

_y_ _b_ _x_ _a_ _o_

Figure 3: A general bow-shaped area as a segment of an ellipse.

The shaded area presents a bow-segment of the ellipse with width _w_ and height _h_. **Appendix A** shows that the area of the bow-segment takes the following form:

$$

A = \frac {a}{b} \left[ \frac {\pi}{2} b ^ {2} + (h - b) \sqrt {2 h b - h ^ {2}} + b ^ {2} \arcsin \left(\frac {h - b}{b}\right) \right]

$$ (2) For a special case where _a=b,_ Equation (2) degenerates to describing a bow-segment of a circular area. Although Equation (2) is derived for a bow-shape segment, it is valid for other shapes of areas as long as the parameter _h_ takes values between _0_ and _2b_. Table 1 lists some special shapes of areas with _h=b/2_, _h=b_, _h=3/2b_, and _h=2b_ for _a=b_, _a=2b_, and _a>>b_.

Table 1: Some Special Shapes of Segment Areas for Various a and h Values.

Table 1: Some Special Shapes of Segment Areas for Various a and h Values.

For the total volumetric flow rate _Q_, the mean flow velocity through the cross section is expressed in consistent units as _Q_ _b_ _vav_ _A_ _h_ _b_ _a_ _b_ _h_ _b_ _hb_ _h_ _b_ _b_ π = =  −    + − − +         ( ) 2 2 2 2 arcsin 2 (3) The expression for the frictional pressure gradient of laminar flow through the channel is derived in Appendix B. The resultant solution takes the following form:

_dp_ _v_ _f_ _av_ _dz_ _h_

µ = (4)

2 1,000 where _dp__f_ _/dz_ is frictional pressure gradient in psi/ft, µ is Newtonian viscosity in cp, _v__av_ is the mean flow velocity in ft/s, and _h_ is height of cross section in inch. Equation (5) is valid for Newtonian fluids only. For non-Newtonian fluids, the Newtonian viscosity µ is replaced by the apparent Newtonian viscosity _m__a_ [16]. For Bingham plastic fluids, the apparent Newtonian viscosity is expressed as

5 _h_ _y_ _a_ _p_ _vav_

$$

\mu_ {a} = \mu_ {p} + \frac {5 \tau_ {y} h}{v} (5)

$$

where _m__a_ is apparent viscosity in cp, _m__p_ is plastic viscosity in cp, and _t__y_ is yield point in lbf/100ft2. For Power Law fluids, the apparent Newtonian viscosity is expressed as $$

\mu_ {a} = \frac {K h ^ {1 - n}}{1 4 4 v _ {a v} ^ {1 - n}} \left(\frac {2 + 1 / n}{0 . 0 2 0 8}\right) ^ {n}

$$

1 2 1/ 1 0.0208 144

(6) where Κ is consistency index in cp equivalent and _n_ is flow behavior index.

Comparison with Other Model

The newly developed model for a channel of bow-shape cross section is compared here with the model presented by Bourgoyne, et al. [15] for a channel of rectangular slot-cross section. The height of the rectangular slot section is taken as the same of the bow-shape cross section _h_. The width of the rectangular slot section _w_ is taken as the width value of the bow-shape section at height _h_, according to Equation (7):

The difference between the two models is expressed by the resistance ratio (_R__R_) defined by:

Pressure gradient given by Equation (4) for bow-shape cross section Pressure gradient given by Bourgoyne, et al. [15] model for slot-cross section _RR_ = (8) where mathematically expressed as:

2 2 2 _a_ _bh_ _h_ _w_ _b_ _RR_ _w_ _a_ _h_ _b_ _eq_ _b_ _h_ _b_ _b_ _h_ _b_ _b_ _hb_ _b_ π $$

\mathrm {2} = \frac {w}{w _ {e q}} = \frac {\frac {2 a}{b} \sqrt {2 b h - h ^ {2}}}{\frac {a}{h b} \left[ \frac {\pi}{2} b ^ {2} + (h - b) \sqrt {b ^ {2} - (h - b) ^ {2}} + b ^ {2} \arcsin \left(\frac {h - b}{b}\right) \right]}

$$

2 2 2 2 arcsin 2

( ) ( ) (9) or simplified to:

(10) It is noticed that the ratio _R__R_ is independent of _a_ and _b_ but _h/b_ only. Table 2 and Figure 4 present the _R__R_ data in the full range of _b_ _h_/ . They indicate that _R__R__=_1 occurs at _h/b_=1.54. Figure 5 shows the two cross sections for _h/b_=1.54. Since _R__R__=_1 does not hold true at other _h/b_ ratio values, Bourgoyne, et al. [15] slot cross-section model is not valid for channels with bow-shaped cross sections. The error of Bourgoyne et al.’s model is expressed as

1 1 1 1 _p_ _p_ _slot_ _bow_ _EB_ _p_ _p_ _R_ _bow_ _R_ _bow_ _pslot_

$$

= \frac {\Delta p _ {s l o t} - \Delta p _ {b o w}}{\Delta p _ {b o w}} = \frac {1}{\underline {{\Delta p _ {b o w}}}} - 1 = \frac {1}{R _ {R}} - 1

$$ (11) ∆ where _E__B_ is the relative error of Bourgoyne et al.’s model, D_p__slot_ is the pressure gradient given by Bourgoyne et al.’s model [15], and D_p__slot_ is the pressure gradient given by Equation (4). For channels with flat cross sections where _h/_ _b_

Case Analysis

This section shows an application of the newly developed hydraulics model in a cement squeezing operation. The objective is to seal a channel behind casing without break down the formation.

A primary cement job has been performed on a 7” production casing. After 48 hours wait on cement time, a cement bond log was run. Result of log interpretation indicated that a long channel exists in the annulus. The volume of the channel is about 5% of the annulus space over the logged interval. A cement squeezing operation is designed as shown in Figure 6. The cement slurry is designed to have a weight of 15.8 ppg with flow consistency _K_=30 cp equivalent and flow behavior index _n_=0.95. The placement efficiency of cement in the channel depends on cement properties and cement slurry flow velocity during cement squeezing. It is essential to know the maximum permissible pumping rate that will not induce friction pressure to break down the formation. It is also essential to know the peak surface pumping pressure corresponding to the formation fracturing pressure at bottom hole so that the pumping rate is monitored and controlled to avoid formation break down.

10 ppg polymer drilling fluid

10-3/4 in. OD, 40.5 lb/ft casing

9,350 ft

10 ppg NaCl brine

25 bbl Spacer

2-7/8 in. EUE tubing

9,325 ft

9-5/8 in. borehole Packer

Channel behind casing

5 bbl spacer

20 bbl 15.8 ppg cement slurry

9,800 ft Fracture gradient 0.8 psi/ft Perforations:4 shots/ft,

360 deg. phasing

9,850 ft

7-in. OD, 26-lb/ft casing

10,000 ft Figure 6: Schematic of a cement squeezing design.

The annulus area is

$$

A _ {a} = \frac {\pi}{4} \left[ \left(9. 6 2 5\right) ^ {2} - \left(7\right) ^ {2} \right] = 3 2. 2 6 \mathrm {i n} ^ {2}

$$ The cross sectional area of the channel is estimated to be (0.05)(34.26) = 1.71 in2 Assuming bow-shape cross section of the channel, the height of the cross section can be estimated using Equation (2):

( ) ( ) ( ) ( )

9.625 / 2

9.625 / 2

$$

1 = \frac {9 . 6 2 5 / 2}{9 . 6 2 5 / 2} \left[ \frac {\pi}{2} \left(9. 6 2 5 / 2\right) ^ {2} + \left(h - 9. 6 2 5 / 2\right) \sqrt {2 h \left(9. 6 2 5 / 2\right) - h ^ {2}} + \left(9. 6 2 5 / 2\right) ^ {2} \arcsin \left(\frac {h - 9. 6 2 5 / 2}{9. 6 2 5 / 2}\right) \right]

$$

which gives h = 0.57 in.

The formation fracturing pressure can be predicted using formation fracture pressure gradient 0.8 psi/ft and depth at the bottom of channel:

$$

p _ {F} = (0. 8) (9, 8 5 0) = 6, 3 0 4 \mathrm {p s i}

$$

The flowing cement pressure at the bottom of the channel is:

(0.052)(10)(9,350) (0.052)(15.8)(9,850 9,350) _p_ _p_ _C_ _f_ _p f_ $$

= (0. 0 5 2) (1 0) (9, 3 5 0) + (0. 0 5 2) (1 5. 8) (9, 8 5 0 - 9, 3 5 0) +

$$

5,273 = + Equating _P__F_ and _P__C_ yields maximum permissible friction pressure of cement slurry _p__f_ = 1,031 psi, which gives pressure gradient of _dp__f_ _/dz_ = (1,031)/(9,850-9,350) = 2.06 psi/ft. Table 3 presents a summary of calculations given by Equation

(4). It shows that the maximum permissible pumping rate is 4.33 bpm. For safe operations the cement pressure should be controlled to be lower than the formation breakdown pressure by at least 500 psi. Therefore a pumping rate of 2 bpm is recommended for this operation. This pumping rate is expected to create a frictional pressure of 495 psi in the channel, as shown in Table 3.

Table 3: Result of Calculations Given by Equation (4).

When the frictional pressure of 495 psi is added to the hydrostatic pressure in the channel, the pressure at the bottom of the channel becomes 5,273+495=5,768 psi. This corresponds to a surface squeezing pressure of 5,768- 0.052(10)(9,850) =646 psi. The surface “breakdown” pressure is calculated to be $$

p _ {b d} = (0. 8) (9, 8 5 0) - (0. 0 5 2) (1 0) (9, 8 5 0) = 2, 7 5 8 \mathrm {p s i}

$$ This pressure is the optimum breakdown pressure. Breakdown can occur at lower pressures depending on the number of opened perforations, increased efficiency with higher viscosity, etc.

Conclusions

A new analytical model was developed in this project to describe fluid flow in channels behind well casing. The model assumes general bow-shaped cross sections of flow channels. The following conclusions are drawn. 1. The new model is equivalent to the Bourgoyne, et al. [15] slot model only for channel parameter _h/b_=1.54, indicating a limitation of Bourgoyne, et al. model to channels with special geometries. For channel parameter _h/b_1.54, use of Bourgoyne, et al. model should over-estimate pressure gradient in channels [15]. 2. For channels with narrow cross sections (_h/b_

Nomenclature

_A_ = area of bow-shaped cross-sections, in2 _a_ = half-length of the long axis of the ellipse, in _b_ = half-length of the short axis of the ellipse, in _F__1_ = force applied by the fluid pressure at Point 1, pound _F__2_ = force applied by the fluid pressure at Point 2, pound _F__3_ = frictional force exerted by the adjacent layer of fluid below the fluid element of interest, pound _F__4_ = frictional force exerted by the adjacent layer of fluid above the fluid element of interest, pound _h_ = height of bow-shaped cross-section, in Κ = fluid consistency index, cp equivalent _n_ = flow behavior index _p_ = pressure, pound per square inch (psi) _p__f_ = frictional pressure, psi _p__F_ = formation fracturing pressure, psi _p__C_ = flowing cement pressure at the bottom of the channel, psi _p__bd_ = surface “breakdown” pressure, psi _Q_ = volumetric flow rate, ft3/s _R_ = radius of the circle, in _R__R_ = resistance ratio _v_ = flow velocity, ft/s _v__av_ = average velocity, ft/s _v__0_ = velocity at boundary, ft/s _w_ = width of cross-section, in _w__eq_ = equivalent width of cross section, in _x_ = horizontal coordinate of ellipse _y_ = vertical coordinate of ellipse _D__y_ = thickness of fluid element, in _z_ = horizontal coordinate in the flow direction _D__z_ = length of the fluid element, in Greeks γ = shear rate, s-1 m = Newtonian viscosity, cp _m__a_ = apparent viscosity, cp _m__p_ = plastic viscosity, cp τ = shear stress, lb/ft2 _t__0_ = shear stress at boundary, lb/ft2 _t__y_ = yield point, lb/100 ft2

Acknowledgment

This research was supported by the U.S. - Israel Center of Excellence in Energy, Engineering and Water Technology through the project “Safe, sustainable, and resilient development of offshore reservoirs and natural gas upgrading through innovative science and technology: Gulf of Mexico – Mediterranean.”

References

1. Uswak G, Howes E (1992) Direct Detection of Water Flow behind Pipe Using A Transient Oxygen Activation Technique. Journal of Canadian Petroleum Technology 31(4): 9.

2. Dossary A, Al-Majed AA, Hossain E, Rahman MK, Jennings SS, et al. (2011) Cementing at High Pressure Zones in KSA “Discovering the Mystery Behind the Pipe.” SPE Middle East Oil and Gas Show and Conference, Society of Petroleum Engineers, Manama, Bahrain, pp: 13.

3. Duguid A, Guo B, Nygaard R (2017) Well integrity assessment of monitoring wells at an active CO2-EOR flood. 13th International Conference on Greenhouse Gas Control Technologies (GHGT-13), Lausanne, Switzerland, pp: 5119-5138.

4. Marca C (1990) 13 Remedial Cementing. Developments in Petroleum Science 28: 13-1-13-28.

5. Bakulin A, Korneev V (2007) Acoustic Signatures of Cross-flow Behind Casing: Downhole Monitoring Experiment At Teapot Dome. Society of Exploration Geophysicists, San Antonio, Texas.

6. Perrier S (2011) Detection of Cross-flows Behind Casing before Perforations, and Cement Isolation Diagnosis Based on Temperature Analysis (TUNU field, Indonesia). International Petroleum Technology Conference. Bangkok, Thailand, pp: 19.

7. Croon M, Huber P, Wright J (2019) Enhanced Well Remedial Decisions from Exact Location of Fluid Movement Behind Casing Identification. Abu Dhabi International Petroleum Exhibition & Conference, Society of Petroleum Engineers, Abu Dhabi, UAE, pp: 16.

8. Grant WH, White RL, Smith RC, Miller AG (1990) Successful Squeezing of Shallow and Low-Pressure Formations. IADC/SPE Drilling Conference, Society of Petroleum Engineers, Houston, Texas, pp: 8.

9. Saponja J (1999) Surface Casing Vent Flow And Gas Migration Remedial Elimination-New Technique Proves Economic And Highly Successful. Petroleum Society of Canada, Calgary, Alberta, pp: 17.

10. Slater HJ (2010) The Recommended Practice for Surface

Casing Vent Flow and Gas Migration Intervention. SPE Annual Technical Conference and Exhibition, Society of Petroleum Engineers, Florence, Italy.

11. Jabarov KA (2019) Mathematical modeling the processes of behind-casing fluid movement in the wells during waiting on cement (Russian). Oil Industry Journal 2019(5): 5.

12. Goodwin KJ, Crook RJ (1992) Cement Sheath Stress

Failure. SPE Drill Eng 7(4): 291-296.

13. Tamayol A, Bahrami M (2010) Laminar Flow in Microchannels With Noncircular Cross Section. Journal of Fluids Engineering 132(11): 111201.

14. Wu B (2014) The influence of the cross section shape on channel flow: modeling, simulation and experiment. Modeling and Simulation. Université de Grenoble, France, pp: 252.

15. Bourgoyne AT, Millheim KK, Chenevert ME, Young

FS (1986) Applied Drilling Engineering. Volume 2, SPE textbook series, Society of Petroleum Engineers, Richardson, Texas, USA, pp: 137-144.

16. Guo B, Liu G (2011) Applied Drilling Circulation Systems_._

Gulf Professional Publishing_,_ Elsevier, pp: 272.

Figure 1: Channels in cement sheath observed in laboratory [12].

Figure 2: Cross-section of a bow-shaped water channel in oil and gas wells.

Figure 3: A general bow-shaped area as a segment of an ellipse.

Some Special Shapes of Segment Areas for Various a and h Values.

Some Special Shapes of Segment Areas for Various a and h Values.

Model-calculated data of _h/b_ versus _R__R__._

Result of Calculations Given by Equation (4).

References

  1. Uswak G, Howes E (1992) Direct Detection of Water Flow behind Pipe Using A Transient Oxygen Activation Technique. Journal of Canadian Petroleum Technology 31(4): 9.
  2. Dossary A, Al-Majed AA, Hossain E, Rahman MK, Jennings SS, et al. (2011) Cementing at High Pressure Zones in KSA “Discovering the Mystery Behind the Pipe.” SPE Middle East Oil and Gas Show and Conference, Society of Petroleum Engineers, Manama, Bahrain, pp: 13.
  3. Duguid A, Guo B, Nygaard R (2017) Well integrity assessment of monitoring wells at an active CO2-EOR flood. 13th International Conference on Greenhouse Gas Control Technologies (GHGT-13), Lausanne, Switzerland, pp: 5119-5138.
  4. Marca C (1990) 13 Remedial Cementing. Developments in Petroleum Science 28: 13-1-13-28.
  5. Bakulin A, Korneev V (2007) Acoustic Signatures of Cross-flow Behind Casing: Downhole Monitoring Experiment At Teapot Dome. Society of Exploration Geophysicists, San Antonio, Texas.
  6. Perrier S (2011) Detection of Cross-flows Behind Casing before Perforations, and Cement Isolation Diagnosis Based on Temperature Analysis (TUNU field, Indonesia). International Petroleum Technology Conference. Bangkok, Thailand, pp: 19.
  7. Croon M, Huber P, Wright J (2019) Enhanced Well Remedial Decisions from Exact Location of Fluid Movement Behind Casing Identification. Abu Dhabi International Petroleum Exhibition & Conference, Society of Petroleum Engineers, Abu Dhabi, UAE, pp: 16.
  8. Grant WH, White RL, Smith RC, Miller AG (1990) Successful Squeezing of Shallow and Low-Pressure Formations. IADC/SPE Drilling Conference, Society of Petroleum Engineers, Houston, Texas, pp: 8.
  9. Saponja J (1999) Surface Casing Vent Flow And Gas Migration Remedial Elimination-New Technique Proves Economic And Highly Successful. Petroleum Society of Canada, Calgary, Alberta, pp: 17.
  10. Slater HJ (2010) The Recommended Practice for Surface Casing Vent Flow and Gas Migration Intervention. SPE Annual Technical Conference and Exhibition, Society of Petroleum Engineers, Florence, Italy.
  11. Jabarov KA (2019) Mathematical modeling the processes of behind-casing fluid movement in the wells during waiting on cement (Russian). Oil Industry Journal 2019(5): 5.
  12. Goodwin KJ, Crook RJ (1992) Cement Sheath Stress Failure. SPE Drill Eng 7(4): 291-296.
  13. Tamayol A, Bahrami M (2010) Laminar Flow in Microchannels With Noncircular Cross Section. Journal of Fluids Engineering 132(11): 111201.
  14. Wu B (2014) The influence of the cross section shape on channel flow: modeling, simulation and experiment. Modeling and Simulation. Université de Grenoble, France, pp: 252.
  15. Bourgoyne AT, Millheim KK, Chenevert ME, Young FS (1986) Applied Drilling Engineering. Volume 2, SPE textbook series, Society of Petroleum Engineers, Richardson, Texas, USA, pp: 137-144.
  16. Guo B, Liu G (2011) Applied Drilling Circulation Systems_._ Gulf Professional Publishing_,_ Elsevier, pp: 272.

PDF of this article

Related articles