Вы находитесь на странице: 1из 16

Conductor Setting Depth in Sand

State of Practice and Ongoing Research

Brian Carlton, Norwegian Geotechnical Institute


April 29, 2016
OUTLINE
Introduction
State of the practice
Motivation
Failure mechanisms
Hydraulic fracturing
Available laboratory tests
Relevant Numerical Simulations
Conclusions
INTRODUCTION
NGI is frequently asked by oil/gas Mud discharge level Pressure

companies to give recommendations on

hd
Sea level

conductor setting depth in locations with


both clay and sand layers. Conductor

dw
Too deep will result in unnecessary costs
Too shallow will result in loss of
Seafloor

circulation, which will slow down drilling Mud pressure

and may even threaten the integrity of


the foundation soils Allowable mud pressure

dc
NGI developed a method to estimate
conductor setting in clay (Anderson et al, Conductor setting depth
u0

1994)

Depth
σv0
STATE OF THE PRACTICE IN SANDS
Schotman and Hospers (1992)
─ Assumes an impermeable mud cake on the wall of the borehole and that hydraulic
fracturing does not occur
─ Does not consider fluid viscosity or soil grain size
dc = conductor setting depth
γ m hd + γ m′ d w hd = height of discharge above sea level
dc ≥ dw = depth of seafloor
pL* γ ′ − γ m′ γm = unit weight of drilling fluid
γm′ = buoyant unit weight of drilling fluid
γ′ = buoyant unit weight of soil
pL* = 2 K 0 − K a where K a < K 0 < 1
1 + sin φ ′ pL* = dimensionless limiting yield
pressure coefficient
1
= K0 − where 1 < K0 < 1
1 + sin φ ′ 1 + sin φ ′ 1 − sin φ ′
Ka = coefficient of active earth pressure
= Kp where 1 < K0 < K p Kp = coefficient of passive earth pressure
1 − sin φ ′ φ′ = angle of internal friction
MOTIVATION
Current state of the practice methods have the following limitations:
1. No consideration of whether or not a mud cake forms at the borehole
periphery and how that influences the fracture mechanism
The change in tangential effective stress and the cavity expansion
pressure will depend on whether (1) the mud forms a low permeability
cake at the borehole wall, or if (2) the mud flows into the soil.
2. Assumes the fracture remains open once initiated
It is unknown whether a crack in sand will stay open if the effective
tangential stresses become negative, or if sand will fall into the crack.
3. No consideration of soil anisotropy or soil fabric/structure
4. No consideration of the (non-linear) stress strain properties of the soil
MOTIVATION CONTINUED
Sand layers are also
usually the limiting
factor
Therefore, there is a
need to develop a more
robust method for
sands
FAILURE MECHANISMS
Main mechanisms
─ Infiltration
─ Cavity expansion
─ Hydraulic fracture
Other mechanisms
─ Piping
─ Sloughing or caving
─ Sandout
HYDRAULIC FRACTURING
Hydraulic fracturing occurs when Mud discharge level Pressure

minor principal effective stress σ3’

hd
Sea level

equals zero
Conductor

This may occur for two scenarios:

dw
─ K0 < 1 Seafloor

σ h < σ mud (vertical fracture) Mud pressure

K 0σ v′ + γ w (d w + d c ) < γ m (hd + d w + d c )
Allowable mud pressure

dc
─ K0 > 1, horizontal fracture plane
σ v < σ mud (horizontal fracture) u0

σ v′ + γ w (d w + d c ) < γ m (hd + d w + d c ) Conductor setting depth

Depth
σv0
AVAILABLE LABORATORY TESTS
Reference No. of Tests Test Apparatus σ'v (kPa) σ'h (kPa)
Triaxial: H = 1.0 m,
Kosar and Been (1989) 1 600 400
D = 1.4 m
Triaxial: H = 1.0 m,
Been and Kosar (1991) 1 200 400
D = 1.4 m
NGI (1991) 1 Box: 0.5x0.5x0.5 m 50 20
Reed and Dusseault (1997) 8 Box: 1x1x1 m 7 -12 ??
Triaxial: H = 0.51m,
Bohloli and de Pater (2006) 18 800 - 31000 500 - 20000
D = 0.4 m
Jasarevic et al (2010) and Golovin
75 Box: 0.61x0.61x0.61 m 275 - 930 70 - 930
et al (2010)
Golovin et al (2011) 45 Box: 0.61x0.61x0.61 m 590 - 650 205 - 380

Triaxial: H = 0.56 m,
Hurt and Germanovich (2012) ?? 110 - 11032 55 - 5516
D = 0.3 m
Triaxial: H = 0.9 m,
de Pater and Brizard (2015) ?? 5000 3000
D = 0.4 m
Bohloli and de Pater (2006)
Performed laboratory hydraulic fracturing experiments in sand with:
─ a Newtonian fluid (viscasil oil 500), a Bentonite slurry, and a cross-linked gel
─ confining pressures ranging from 0.5 to 20 MPa radial and 1.3 to 28 MPa axial
Bohloli and de Pater (2006) cont.
Newtonian fluid: infiltration only and in two cases small fractures.
Bentonite: a combination of borehole expansion, infiltration, shearing and fracturing,
with infiltration and borehole expansion the dominant failure modes.
Cross-linked gel and quartz powder: hydraulic fracturing, even at radial stresses of 20
MPa and axial stress of 28.5MPa.

Bentonite slurry. (a) Shows borehole expansion plus shear


fractures while (b) displays mainly radial fractures.
RELEVANT NUMERICAL SIMULATIONS

Reference Type Parameters Studied

particle size, confining


Pruiksma and Bezuijen
DEM stress, injection rate,
(2002)
friction angle

confining stress,
Zhai and Sharma (2005) FDM
permeability
de Pater and Dong confining stress,
DEM
(2007) permeability, viscosity
de Pater and Brizard
FEM cohesion, permeability
(2015)
NUMERICAL SIMULATIONS AT NGI
Axisymmetric Plaxis 2D model
Dummy layer used to produce the corresponding initial stress conditions
Left, right and bottom boundaries are prescribed as roller conditions, while the top
boundary is set as free.
Set a thin layer outside the borehole wall with very small soil permeability to simulate
the mudcake formation during injection
NUMERICAL SIMULATIONS AT NGI

Hardening Soil Model Hypo-Plastic Model


CONCLUSIONS
Model to estimate conductor setting depth in sand should include:
─ Density of sand
─ Soil particle size / gradation
─ Viscosity of injection fluid
─ Confining pressure
Effect of confining pressure on failure mechanism:
─ High: infiltration
─ Medium: fracturing
─ Low: cavity expansion
As density or injection rate increase, or confining pressure decreases,
the ratio of injection pressure to confining pressure at failure increases
Possible topic for a JIP
@infoNGI

NORWEGIAN GEOTECHNICAL INSTITUTE


NGI.NO

Вам также может понравиться