Transcript No Slide Title

SPE 71514
A Semianalytical p/z Technique
for the Analysis of
Abnormally Pressured Gas Reservoirs
Ronald Gunawan Gan,
VICO Indonesia
and
T. A. Blasingame,
Texas A&M University
Objective
To present a new technique that can be
used to :
 Calculate gas-in-place for an abnormally pressured gas reservoir using
only average reservoir pressure and
cumulative production data.
 Calculate pore volume compressibility as a function of reservoir pressure.
Presentation Outline
 Introduction
 Overview of Existing Methods
 New Method
 Field Examples
 Conclusions
Introduction

p/z schematic for a normally-pressured
volumetric gas reservoir
p pi  G p 

 1 
z zi 
G 
p/z
Gp
G
Introduction

p/z schematic for an abnormally-pressured
gas reservoir
pi  G p 
p
1   ( pi  p)  1  
z
zi 
G 
p/z
Gp G
Gapp
Introduction
 Reasons for the non-linear p/z behavior:

Rock and water compressibility effects —
"rock collapse theory" (Hawkins, 1969)

Shale water influx (Bourgoyne, 1989)
Existing Methods
 Methods
based on presumed knowledge of
system compressibility:

Hammerlindl (Constant Compressibility), 1971

Ramagost & Farshad (Constant Comp.), 1981
p( S wcw  c f )  pi 
Gp
p
1 
 
1 
z
(1  S w )
G
 zi 




Yale et al. (Variable Compressibility), 1993
 Methods
based on presumed knowledge of
system compressibility (continued)

Fetkovich, Reese, and Whitson - 1991
- Derived General Material Balance Eq.
- Define cumulative effective compressibility,
ce ( p) 
S wi ctw ( p)  c f ( p)  M [ctw ( p)  c f ( p)]
1  S wi
- ce represents the cumulative change in
hydrocarbon PV caused by compressibility effects (and water influx).
 Methods which do not require a prior
knowledge of system compressibility

Roach - 1981
- very sensitive to initial pressure.
- method sometimes doesn’t exhibit
a negative intercept (which is not possible).

Bernard - 1985
- using Least Squares approach.
- very sensitive to data scatter.

Ambastha - 1991: Type Curve Approach
- non-uniqueness problems.
New Method
 Satisfies both "rock collapse" and
"shale
water influx" theories
 Develops 2 new plotting functions:
1. ce ( pi  p) versus ( p/z )/( pi /zi )
2. ( p/z )/( pi /zi ) versus Gp /G
 Requires production data only (p and Gp)
New Method
 Uses general material balance equation
(proposed by Fetkovich, et al.)
pi  G p 
p
1  ce ( pi  p)  1  
z
zi 
G 
 Rearranging, we obtain
 pi / zi  G p 

1 
ce ( pi  p)  1  
G 
 p / z 
New Method
 Calculate the ce(pi-p) function for
each p/z versus Gp trend
ce(pi-p) = ???
p/z
ce(pi-p) = ???
Gp
G
Gapp
New Method
 For early time data (1st straight line) :
Gapp 1
Gapp
1
ce ( pi  p)  1  p pi 

p pi
( z / zi ) G ( z / zi ) G
 For late time data (2nd straight line) :
 G pA 
1 

ce ( pi  p)  1  ( p / z )
A
G 
( pi / zi ) 
1
where: A is the inflection point
New Method
 Plot of log ce(pi-p) versus (p/z)/(pi/zi):
inflection point
G/Gapp=0.6
G/Gapp=0.7
h
G/Gapp=0.8
(p/z)/(pi/zi)
New Method
 Plot of log ce(pi-p) versus (p/z)/(pi/zi) :
inflection point
h
(p/z)/(pi/zi)
New Method
 ( p/z )/( pi /zi ) versus Gp /G
p/z
G
 1
pi /z i
Gapp
(p/z)/(pi/zi)
1
 Gp 
 
G
Infl. Point: GpA/G, (p/z)A /( pi /zi )
h
 Gp 
p/z
( p / z) A

 
pi /z i ( pi / zi )(1  G pA / G)  G 
0
Gp/G
1
New Method
 ( p/z )/( pi /zi ) versus Gp /G
1
(p/z)/(pi/zi)
G/Gapp=0.6
Inflection point
h
G/Gapp= 0.8
G/Gapp=1
0
Gp/G
1
New Method
 ( p/z )/( pi /zi ) versus Gp /G
1
(p/z)/(pi/zi)
Inflection point
h
G/Gapp=0.8
0
Gp/G
1
New Method
 ( p/z )/( pi /zi ) versus Gp /G

Dynamic Type Curve Matching.
Automatic Matching using SOLVER
m(Excel function for non-linear regression).

New Method
 Data required for analysis:
Fluid property data
 Initial Reservoir p and T


p and Gp data
New Method
 Computer program:
Visual Basic Application in MS Excel
 Only requires MS Excel
 Easy to use - especially for analysis

Data Analysis Sheet
Example 1: G is too low
Example 1: G is too high
Example 1: Correct G
Example 2: Long transition period
Example 3: Early time data
Example 4: Synthetic Dry Gas Case
Example 4: Backcalculated cf
 Procedure to calculate cf vs. p from
production data:
1. Get ce ( p) from type curve matching
2. Use the following equation to calculate c f ( p ) :
ce ( p) 
S wi ctw ( p)  c f ( p)  M [ctw ( p)  c f ( p)]
1  S wi
3. Calculate cf (p):
n
c f ( pi  pn )   c f j p j
j 1
Example 4: Backcalculated cf
Conclusions
 We have developed a straightforward
approach for analyzing p/z versus Gp
behavior for abnormally pressured gas
reservoirs — the approach considers
that two straight-lines must be observed on the p/z plot.
 The proposed method determines
gas-in-place without using system
compressibility data. Only p, Gp, and
fluid property data are required.
Conclusions (continued)
 Our approach of using ce(pi-p) versus
(p/z)/(pi /zi) and (p/z)/(pi /zi) versus Gp/G
as dynamic type curve matching functions has been shown to work extremely well.
 Using our new method, it is possible to
calculate rock compressibility as a function of pressure from p and Gp data
Conclusions (continued)
 The "dynamic type curve matching
technique" used for calculating gasin-place from production data is more
representative (and more stable) than
the non-linear optimization method
provided by SOLVER.
SPE 71514
A Semianalytical p/z Technique
for the Analysis of
Abnormally Pressured Gas Reservoirs
Ronald Gunawan Gan,
VICO Indonesia
and
T. A. Blasingame,
Texas A&M University