7/23/2019 12691 Lovell Presentation
1/39
Near Wellbore PhysicsNon-isothermal flow into awellbore
Oct 13th, 2011
John LovellSchlumberger
Excerpt from the Proceedings of the 2011 COMSOL Conference in Boston
7/23/2019 12691 Lovell Presentation
2/39
Near Wellbore Physics in Oilfield Exploration
Industry background
Simplistic approach to determining flow profiles
Actual data Role of multi-physics modeling
Some initial results and future research directions
Conclusion
7/23/2019 12691 Lovell Presentation
3/39
Industry Background Exploration and Production
Wellbores are drilled into the ground to extract hydrocarbon(gas/oil) from within the pores and fractures of a reservoir,
usually with water in the same pores
The zones will be of different permeabilities and pressures,
and the pressures will drop over time as the reservoirdepletes
Only about 20-30% of the oil is ultimately recovered, even
with hundreds of wells drilled into the same reservoir
Finding new reserves is important, but improving production
ratio even more so.
7/23/2019 12691 Lovell Presentation
4/39
Industry Background Exploration and Production
Production improvement by increasing permeability
Hydro-Fracking, Acid Stimulation,etc
Production improvement by increasing pressure
Drill nearby well and inject water, gas, or both
Production improvement by increasing viscosity of oil
Drill very-nearby well and pump steam to heat the oil
Difficult to improve production without knowing what fluidsare being produced, and from where along the wellbore
7/23/2019 12691 Lovell Presentation
5/39
Industry Background Exploration and Production
Production improvement by increasing permeability
Hydro-Fracking, Acid Stimulation,etc
Production improvement by increasing pressure
Drill nearby well and inject water, gas, or both
Production improvement by increasing viscosity of oil
Drill very-nearby well and pump steam to heat the oil
Difficult to improve production without knowing what fluidsare being produced, and from where along the wellbore
7/23/2019 12691 Lovell Presentation
6/39
Role of Temperature Profiling
7/23/2019 12691 Lovell Presentation
7/39
Role of Temperature Profiling
7/23/2019 12691 Lovell Presentation
8/39
Role of Temperature Profiling
7/23/2019 12691 Lovell Presentation
9/39
Role of Temperature Profiling
7/23/2019 12691 Lovell Presentation
10/39
Role of Temperature Profiling
7/23/2019 12691 Lovell Presentation
11/39
Role of Temperature Profiling
Measure/Estimate/Model incoming fluid temperature
Measure fluid temperature along wellbore
Flow profile identify major flowing zones
Evaluate effectiveness of stimulation/injection/etc
How to measure temperature?
7/23/2019 12691 Lovell Presentation
12/39
Role of Temperature Logging
7/23/2019 12691 Lovell Presentation
13/39
Subsea Data from Array of Sensors
Top zone cleaned successfully
Bottom zone barely starting
Bottom zone still had brine had absorbed lot of the
fluid during stimulation
When this well is brought online into production
facilities, need to plan for some initial some fluid
production before the dry gas
Clean-up of gas well
7/23/2019 12691 Lovell Presentation
14/39
Subsea Data from Array of Sensors
Fluid migrating from lower zone to upper.
Well had been completed many months earlier,
yet cross-flow still stable.
Indicative of relatively large reservoirs and
anomalous pressured zones.
Led operator to re-evaluate drill-stem test data.
Cross-Flow Before Production
7/23/2019 12691 Lovell Presentation
15/39
Subsea Data from Array of Sensors
Fewer than 30 sensors across sandface
Some sensors directly in front of gas
impingement reading much colder than theaverage wellbore temperature
Similar phenomenon have been since been
seen in DTS in fractured shale gas wells.
Not all wells so simple to interpret
7/23/2019 12691 Lovell Presentation
16/39
Azeri Sandface Data
7/23/2019 12691 Lovell Presentation
17/39
Azeri Sandface Data
7/23/2019 12691 Lovell Presentation
18/39
Azeri Sandface Data
7/23/2019 12691 Lovell Presentation
19/39
Azeri Sandface Data Pressure Depletion
7/23/2019 12691 Lovell Presentation
20/39
What temperature does DTS measure?
Temperature on outside of screens not same as average
wellbore temperature
Temperature on screens depends on incoming flow rate
advection versus conduction
Temperature on the inside of a screen not the same as
temperature on the outside of screen
Temperature on the inside of a screen not the same as theaverage wellbore temperature
7/23/2019 12691 Lovell Presentation
21/39
Velocity vs Temperature Profile in Wellbore
Characteristic velocity
shape for turbulent flow.
Fast drop to zero at wall.
Characteristic temperature
curves versus radius.Much faster drop to screen
temperature at interface
7/23/2019 12691 Lovell Presentation
22/39
Go to equations Wellbore
Inside wellbore assume instantaneous change versus time
(not transient formulation)
2D flow (both radial and axial flow)
Model turbulence using standard CFD formulation, k-e
Boundary conditions need either radial and axial velocity, or
radial and axial stress (i.e. two conditions, not one)
7/23/2019 12691 Lovell Presentation
23/39
Go to equations 2D Porous Media
Almost exactly same as Darcys equation except the term
(m/k)u. This term is only important near boundaries, but it
changes the nature of the equation
Boundary conditions need either radial and axial velocity, or
radial and axial stress (i.e. two conditions, not one)
I.e., same as the Navier-Stokes, so we can match them upconsistently
7/23/2019 12691 Lovell Presentation
24/39
COMSOL Multiphysics
2D Darcy-Brinkmann
N-S
2DBrinkmann uses NS computed stress as boundary condition
NS uses Brinkmann computed velocities as boundary condition
7/23/2019 12691 Lovell Presentation
25/39
COMSOL Multiphysics Stabilization
2D Darcy-Brinkmann
N-S
2DBrinkmann uses NS computed stress as boundary condition
NS uses Brinkmann computed velocities as boundary condition
Add a weak term to constrain divergence
7/23/2019 12691 Lovell Presentation
26/39
Go to equations Heat Transfer
Term in red is Joule-Thomson effect
E.g., if pure radial flow (w=0) and conduction not important
i.e. DT = J DP where J = (aT 1)/(rCp)
Note that J can be positive or negative (gas vs water vs oil)
7/23/2019 12691 Lovell Presentation
27/39
Challenges
Mesh needs to be very fine in wellbore:
6 diameter, hundreds of meters longMesh fairly dense in annulus
Mesh coarse in the rock
Results in many elongated elements
Slows convergence
7/23/2019 12691 Lovell Presentation
28/39
Challenges
Mesh needs to be very fine in wellbore
Mesh fairly dense in annulusMesh coarse in the rock
Results in many elongated elements
Slows convergence
7/23/2019 12691 Lovell Presentation
29/39
7/23/2019 12691 Lovell Presentation
30/39
Challenges
Mesh needs to be very fine in wellbore
Mesh fairly dense in annulusMesh Coarse in the rock
Results in many elongated elements
Slows convergence
~150,000 unknownsNonlinear equation for P
Weakly nonlinear in T
k-e convergence often very painful
To get CFD to converge, started with
uniform flow and used that as initial guess
~10 mins to solve one configuration
7/23/2019 12691 Lovell Presentation
31/39
COMSOL Results
10
15
20
25
30
35
40
45
50
55
60
60.6 60.7 60.8 60.9 61 61.1 61.2 61.3 61.4
MD(m)
Temperature (degC)
Axial Temp
Screen OD Temp
Fluid In TempExample shown is 50m
section with highly stratifiedflow. Hotter fluid is entering
from below, gets cooled
somewhat by fluid over this
interval.
Zones dominated by radial flow
measure the incoming fluid
temperature.
Zones dominated by conduction
measure close to the axial
temperature.
7/23/2019 12691 Lovell Presentation
32/39
Some results Large Sand Bodies
85 deg well, flow from largesand with varying
permeability.
30m spacing and 10m
spacing
Coarse spacing captures
overall temperature trend
10m preferable
4800
4900
5000
5100
5200
5300
5400
5500
5600
5700
5800
59.5 60.5 61.5
MeasuredDepth(m)
Temperature (degC)
4800
4900
5000
5100
5200
5300
5400
5500
5600
5700
5800
59.5 60.5 61.5
MeasuredDepth(m)
Temperature (degC)
7/23/2019 12691 Lovell Presentation
33/39
Some results Small Sand Bodies
85 deg well, flow fromsand blocks with
different perms and
separated by shales
Open-hole gravel pack
10m sensor spacing
Inverted flow rate based
on the shale data only
4950
5000
5050
5100
5150
5200
5250
5300
5350
60 60.5 61
MD(m)
Temperature (oC)
4990
5040
5090
5140
5190
5240
5290
0 1000 2000 3000 4000
MD(m)
Downhole Rate (m3/d)
Actual Rate
Inverted Rate
7/23/2019 12691 Lovell Presentation
34/39
Some results Flow from Vugs/Fractures
85 deg well, flow from veryhigh perm vugs and
fractures
Open-hole gravel pack
10m sensor spacing
Inverted flow rate vs actual
4950
5000
5050
5100
5150
5200
5250
5300
5350
59 60 61
MD(m)
Temperature (o
C)
4950
5000
5050
5100
5150
5200
5250
5300
5350
0 500 1000 1500
MD(m)
Downhole Rate (m3/d)
Actual Rate
Inverted Rate
7/23/2019 12691 Lovell Presentation
35/39
Some results Flow from Vugs/Fractures
85 deg well, flow from veryhigh perm vugs and
fractures
OHGP
5m sensor spacing
Inverted flow rate vs actual
4950
5000
5050
5100
5150
5200
5250
5300
5350
59 60 61
MD(m)
Temperature (degC)
4950
5000
5050
5100
5150
5200
5250
5300
5350
0 500 1000 1500
MD(m
)
Downhole Rate (m3/d)
Actual Rate
Inverted Rate
7/23/2019 12691 Lovell Presentation
36/39
New Interface Condition at Wellbore
Combination of k-eand Brinkmann-Darcy
gives a new interface
condition for heat-
transfer coefficient from
reservoir to wellbore
It lies in-between
Sleicher-Rouse and
Dittus-Boulter
7/23/2019 12691 Lovell Presentation
37/39
Future directions
Improved physics at the interface:
Beavers/Joseph, etc
Effective viscosity of the Brinkmann equation vs fluid viscosity
Integration of empirical PVT formula into the COMSOL set-up
Stabilization of the k-e (new empirical model for e in a tube?)
3D is the temperature on one side of the wellbore, the same as the
other??
7/23/2019 12691 Lovell Presentation
38/39
Future directions 3D
Inject fluid from just one side of the wellbore
Almost uniformly parabolic
7/23/2019 12691 Lovell Presentation
39/39
Conclusion
Multiphysics FEM code can solve combination of porous media,
wellbore CFD, advected temperature and temperature conduction
Reproduces temperature phenomenon seen in sandface data
Has been used in commercial applications to determine sensor
spacing needed for accurate flow-profile inversion
Has been used to validate when simplistic models can be used, and
when they cannot
Opportunities for new algorithms that take advantage of the difference
between real temperature and simple models
Understanding the near wellbore physics has proven crucial to
accurate interpretation of temperature data
Top Related