forensic analysis why did this field die? presented at western australia section of spe june 19,...
TRANSCRIPT
Forensic AnalysisWhy Did This Field Die?
Presented at
Western Australia Section of SPEJune 19, 2012
Perth, Australia
Dr. Bill CobbWilliam M. Cobb & Associates, Inc.
Petroleum Engineering & Geological ConsultantsDallas, Texas
Every Field is Different!
Reservoir Drive Mechanism PRIMARY
Rock & Liquid Expansion Above the Bubble Point SOLUTION GAS DRIVE Initial Gas Cap Expansion - Size Aquifer Influx – Size
WATERFLOODING CO2, Steam, Polymer, Other
Why Did This Oil Field Die?
PRIMARY RECOVERY VS WF
Solution gas drive requires the reservoir pressure to be constantly decreasing
WF is a displacement process and is most efficient when reservoir pressure is maintained or increased
When converting from primary to waterflooding or any type of fluid injection, the reservoir recovery mechanism changes.
Consequently reservoir evaluation and reservoir management procedures generally need to be changed
Can we learn lessons that may help us resuscitate the WF or guide us if we move to a tertiary project?
Why did the project die? Where do we start? What does the evidence say?
THIS PRESENTATION IS THE FORENSIC ANALYSIS OF A DEAD OR DYING WATERFLOOD WHERE PRIMARY
PRODUCTION WAS FROM SOLUTION GAS
WHAT ARE THE KEY FACTORS THAT DRIVE THE OUTCOME OF A WATER INJECTION
PROJECT?
NP = Cumulative Waterflood Recovery, BBL.
N = Oil in Place at Start of Injection, BBL.
EA = Areal Sweep Efficiency, Fraction
EV = Vertical Sweep Efficiency, Fraction
ED = Displacement Efficiency, Fraction
WATERFLOOD RECOVERY FACTOR
EA = f (MR, Pattern, Directional Permeability, Pressure Distribution, Cumulative Injection &
Operations)
EV = f (Rock Property variation between different flow units, Cross-flow, MR)
EVOL = Volumetric Sweep of the Reservoir by Injected Water
ED = f (Primary Depletion, So, So, Krw & Kro, μo & μw)
RF N
Np DVA E
E
EE
VOL
**RF
TO MAXIMIZE WATERFLOOD OIL RECOVERY
Recognize the recovery mechanism has changed.
The driver of the waterflood is the injection well.
The injector is the Quarterback of the operation.
Efficient utilization of the injector maximizes volumetric sweep. (The fraction of the reservoir contacted by the injected water).
THE QUARTERBACK OF ALL INJECTION PROJECTS IS THE INJECTION WELL
Properly located injection wells are A significant key to A successful waterflood:
They deliver the water where it needs to be injected
They deliver the water at the correct time They deliver the water in the proper volume Effective utilization of injection wells is an
important key to optimizing the WF by allowing EA and EV values and RF to be maximized
Failure to properly locate and manage the injectors will result in a less efficient injection project and could lead to failure.
DID THIS WATERFLOOD DIE AND FAIL TO PRODUCE EXPECTED WF OIL?
Did this project meet a premature high water cut death? If the answer is yes, some of the reasons for failure are likely due to:
Lack of significant movable oil saturation at the start of injection – low ED
High initial free gas saturation leading to a small or negligible oil bank and long gas fill-up time
Floodable net pay is overstated (same porosity cutoff used for WF as used for primary depletion)
Poor volumetric sweep – low EVOL
NORTH AMERICALIQUID EXPANSION - SOLUTION GAS DRIVE
Pi = 4400 Psi Pbp = 4000 Psi P = 400 Psi
Sg = 36% RF = 1% RF = 19%
So = 76% So = 76%
So = 40%
Swc = 24% Swc = 24% Swc = 24%
Boi = 1.75 Bobp = 1.78 Bo = 1.15
OOIP = 100 MMSTBO OIP = 80 MMSTBO *L.P. Dake – Fundamentals of Res Eng. – Eq. 3.22 Pg. 86
NET PAY
Static OOIP Dynamic OOIP
Drive mechanism – primary vs secondary Waterflood net pay cutoffs controlled by:
Water Cut Economic Limit Permeability Distribution between Flow Units
(Dykstra-Parson Coefficient) Oil/Water Relative Permeability Mobility Ratio (Oil and Water Viscosity) Fluid Saturations at Start of Injection (So, Sg, Swc)
* See SPE #48952 and SPE #123561
Cont'd – WATERFLOOD FAILURES MAYBE DUE TO:
VRR less than 1.0 and pressure declines High variation in permeability between geological
layers causing the high permeability layers to water out the producing wells before the lower perm layers Contribute meaningful waterflood production
Failure to develop a pattern Failure to recognize and honor KX and KY trends Failure to keep fluid levels pumped off in producing
wells Mechanical integrity of injection wells to control
vertical distribution of injected water
ASIAN WATERFLOOD
SOLUTION GAS DRIVE (WEAK WATER INFLUX)
Pi = Pbp = 2250 Psi P = 2100 Psi - At Start Of Injection
Rsi = 550 SCF/STBO Swc = 29%
Boi = 1.39 RB/STB Sg = 3%
µoi = 0.44 CP MR = 0.30
PRF W/O H2O Current RF EUR VRR Since
AREA % % % Start of Inj.
1 15-18 18 27 0.51
2 15-18 21 31 0.63
3 15-18 25 33 0.71
4 15-18 31 44 1.09
WATER INJECTION RESPONSE
WF WAS NOT SUCCESSFULWHAT SHOULD BE CONSIDERED?
Compute So and Sg at start of waterflood Compute WF volumetric sweep efficiency Re-evaluate net pay cutoffs Check for high permeability thief zones Check VRR since start of injection Consider infill drilling Realign pattern – this could violate the 11th
commandment Is it a candidate for horizontal producers?
WATERFLOOD LESSONS LEARNED
1) Every field is different2) There are almost no analogy floods3) Need high movable oil saturation4) Need low free gas from solution5) Prefer low oil viscosity (high viscosity oils can be
produced by cycling large volumes of water).6) Pattern highly desirable7) Honor KX/KY 8) Effective VRR since injection start equals 1.0 or more for
each pattern and key geological zones
LESSONS LEARNED Cont'd
9) Careful attention given to selection of net pay cutoffs. NET PAY FOR WATERFLOOD IS LESS THAN NET PAY FOR PRIMARY DEPLETION.
10) Keep fluid levels in producing wells pumped off11) Accurately test each producing well monthly for oil,
water and gas.12) Maintain production plots on a well by well basis.
Combining wells into groups is equivalent to combining “The Good, The Bad and The Ugly”.
LESSONS LEARNED Cont'd
13) Forecasting the future production should be performed on a well by well basis. Typical plots of oil rate vs time or oil rates vs cumulative
oil produced should be used with extreme caution because oil rates are directly related to injection rates and injection rate changes and stratification.
Semi-log plots of WOR vs cumulative oil produced are likely to be more reliable because WOR is largely independent of injection rate and injection rate changes.
Analysis of WOR plots is most reliable when forecasts are made using WOR values which exceed 2.0 to 3.0
14) Reliable well forecasts require accurate well tests
LESSONS LEARNED Cont'd
15)Conduct regular injection profiles to monitor the amount of water going into various zones.
16)Thief zones must be identified and isolated or oil production will occur at high water cuts.
17) REMEMBER THE QUARTERBACK
HONOR THE DATA
Every field is different The reservoir speaks to us – are we listening? Data costs money Lack of quality geological, reservoir and production data
usually leads to inefficient operations or early failure. Lack of relevant data puts the geoscientist, reservoir or
operational analyst into a position where he/she must trust experiences, instincts, hunches and wishful thinking.
“People are not always truthful – the evidence does not lie.” – from Gil Grissom of CSI
Become an OFI – Oil Field Investigator – follow the evidence
DON’T CHANGE ITDON’T CHANGE IT
THEN, THERE IS NO PROBLEMTHEN, THERE IS NO PROBLEM
YES NO
DID YOU TRY TO FIX IT?
DID YOU TRY TO FIX IT?
YOU IDIOT!YOU IDIOT!
DOES ANYBODYKNOW ABOUT IT?
DOES ANYBODYKNOW ABOUT IT?
HIDE ITHIDE IT
YES
YOU POOR BASTARD!YOU POOR BASTARD!
YES
NOCAN YOU
BLAME SOMEONEELSE?
CAN YOUBLAME SOMEONE
ELSE?
NO
YES
ARE YOUGOING TO BE IN
TROUBLE?
ARE YOUGOING TO BE IN
TROUBLE?
NO
YES
PRETEND YOU DON’T KNOW ABOUT
IT!
PRETEND YOU DON’T KNOW ABOUT
IT!
NO
WHEN ALL ELSE FAILS, A POTENTIAL WATERFLOOD ANALYSIS METHOD IS:
DOES IT FUNCTION?
DOES IT FUNCTION?
Forensic AnalysisWhy Did This Field Die?
Presented at
The Dallas E & P ForumDecember 13, 2011
Sponsored By SIPES and SPE
Dr. William M. CobbWilliam M. Cobb & Associates, Inc.
Petroleum Engineering & Geological ConsultantsDallas, Texas
Assume Gas fill-up has been achieved
(reservoir Contains oil and water Reservoir pressure is approximately constant
(Bo is constant) Steady state flow prevails (approximately)
Conclusion Effective Water Injection = Liquid Production
(at Reservoir Conditions)
DECLINE CURVE ANALYSIS
DECLINE CURVE ANALYSIS
* (1 )
*
w o w wo
o o
w ww
w
i f i fq
B B
i fq
B
ConclusionOil and Water Production Rates are directly related to injection rates. Therefore, DCA of qo vs t or qo vs NP must be evaluated only after giving consideration to historical and projected water injection rates.
Jan-
1969
Jan-
1971
Jan-
1973
Jan-
1975
Jan-
1977
Jan-
1979
Jan-
1981
Jan-
1983
Jan-
1985
Jan-
1987
Jan-
1989
Jan-
1991
Jan-
1993
Jan-
1995
Jan-
1997
Jan-
1999
Jan-
2001
Jan-
2003
Jan-
2005
Jan-
2007
Jan-
2009
Jan-
2011
Jan-
2013
Jan-
2015
100
1000
10000
WATERFLOOD EXPONENTIAL DECLINE
BO
PD
EL
Start Water Injection
0 5000 10000 15000 20000 25000 30000 35000 40000 45000 50000 55000 600000
1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
OIL RATE VS CUMULATIVE OIL PRODUCED
Cumulative Oil Production (MMBbls.)
BO
PD
EUR 49 MMBO
Start Water Injection
EUR 53 MMBO
0 5000 10000 15000 20000 25000 30000 35000 40000 45000 50000 55000 600000
1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
OIL RATE VS CUMULATIVE OIL PRODUCED
Cumulative Oil Production (MMBbls.)
BO
PD
EUR 49 MMBO
Start Water Injection
EUR 53 MMBO
WOR IS INDEPENDENT OF INJECTION RATE
0
. .
*
* (1 )
(1 )
( ) *(1 )
w
w w
w w
w
w
w oSTD COND
w w
qWOR
q
i fWOR
i f
fWOR
f
f BWOR
f B
Conclusion WOR is independent of injection rate
Other Points WOR should be applied to individual wells and
not field WOR should be applied using values greater
than 2.0