Vous êtes sur la page 1sur 11

SPE-179994-MS

Data Mining and Analysis - Eagle Ford Formation, South Texas


Dario H. Romero, CPQ Energy LLC; Dr. Steve W. Poston, Emeritus Professor Texas A&M

Copyright 2016, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE/IAEE Hydrocarbon Economics and Evaluation Symposium held in Houston, Texas, USA, 17–18 May 2016.

This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents
of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect
any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written
consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may
not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract
The paper applies commonly available public information to outline well behavior for productive trend
modeling within a chosen area of the Eagle Ford shale in South Texas. Public records are often thought
to contain qualitatively insufficient information to provide a reasonable analysis of well and producing
character trend characteristics. They –Public records- contain both static and dynamic well data. Static
information consists of completion histories, locations, deviations surveys, core analyses, and well
property maps while oil, gas and water production records comprise dynamic records. Too often, this
freely available information is not utilized for regional studies because pressure, porosity, permeability
profiles are not included in these files. More descriptive data are usually only available in individual
company well files. This means for most cases, rate – time data represent the only information available
to the public.
The study of monthly production records involving multiple ownership properties can highlight regions
of no additional interest but also can highlight areas of potential additional economic value. Property value
could be increased by well cleanouts, workovers, and re-fracturing or new drill locations to increase
recovery from these high productivity areas. In any case, a seemingly intractable data problem could be
reduced to a rather straightforward procedure by the study of well records.
Cumulative production, remaining reserves or initial production rate estimates plotted on available
property maps could highlight areas for further review or in the majority of cases exhausted areas of no
additional interest.
Transforming production data is only the first step of the process, a comprehensive interpretation of the
results must follow. Property and maps obtained from public records provide a basis for plotting
production trends for a particular well set. Analysis of well tickets coupled with the oil, gas and water
production history supply considerable insight into interpreting a well history since erratic production
changes are numerous in the Eagle Ford trend.
Recovery estimates can only be determined after the data are plotted and studied to select the proper
production segment to analyze. Selection of transient and boundary dominated segments is of paramount
importance for interpreting Eagle Ford wells. Smoothing techniques such as material balance time and
normalized production can be applied if necessary. The ⬙goodness⬙ of the fracture system can be estimated
from the transient portion of the curve while estimated reserves from the boundary side. Of course, the
2 SPE-179994-MS

process will not apply if discrete systems cannot be identified. These results are plotted on a property maps
and interpreted.
A particular area within the south Texas area is identified and each well subjected to the data mining
process. The result of this study should help identify maturely developed areas and highlights areas of
interest.

Introduction
Public records contain both static and dynamic well data. The static information consists of completion
histories, locations, well property maps while oil, gas and water production records as a function of time
comprise dynamic well data. Too often, this freely available information is not utilized because pressure,
porosity, permeability profiles are not included in the records for more sophisticated engineering analysis.
These data are usually only available in individual company well files. Rate – time are the one data set
available for relating inter-well quality.
Monthly production records are the main source of information for a study involving analysis of
multiple ownership properties to determine production trends. Wells of interest could involve cleanouts,
workovers, and re-fracturing or new drill locations. In any case a seemingly intractable problem could be
reduced to a rather straightforward procedure by the study of well records. Eventually cumulative
production, remaining reserves or initial production rate estimates plotted on available property maps to
highlight areas for further review or in the majority of cases exhausted areas of no additional interest.

Sources
A public information services provider provides access to public information via company intranet or by
a monthly paid subscription. One can map, query, browse and export well & production information
common to designated properties. Records ranging from well location, deviation surveys, stimulation
fracing, core analysis and oil gas and water production can be accessed for various analyses.
Not all data are presented in usable form so analysis code must be written to transform some of the data
to a readable format for analysis and mapping. A Python program simplifies manipulation of the data to
Excel spreadsheets. An open source application for mapping allows mapping of lease, operator and well
locations for comparison purposes.

Fundamentals
The following discussion develops a method for applying basic decline curve studies to scan large
amounts of data when reconnoitering mature producing areas. An example lease of wells completed in the
Eagleford Shale is used as an example.
Eagleford wells are drilled horizontally for 3000 to 5000 feet and then serially fractured to materially
expand communication potential of the low permeability matrix face to the wellbore. The flow system
consists of two parts:
● Production initially migrates through the horizontal well bore (linear flow) after traveling through
the fracture system (bilinear) flow.
● The very low permeability matrix rock drainage area is often defined as the stimulated reservoir
volume (SRV). The drainage boundary is usually oval shaped in nature and defined in the ⬙x⬙ and
⬙y⬙ direction as the aspect ratio (AR)4.
The ⬙y⬙ component is a function of the extent of the SRV boundary. The aspect ratio defines
directional contribution of the flow system trending from bilinear (ultra-low) permeability SRV to
pseudo radial (low to moderate) permeability reservoir rock.
(1)
SPE-179994-MS 3

Fig. 1 presents the directional relation for the aspect ratio. Note the possibilities of contributions
of each component in the total flow system. Is the flow system mainly confined directionally along
and adjacent to the fracture (⬙x⬙ direction), or does fluid flow eventually spread out significantly
in the (⬙y⬙) direction? (Xe and Ye) define the half lengths of the drainage area in the (x) and (y)
directions.

Figure 1—The aspect ratio for the very low permeability matrix case when (SRV) boundary is represented by boundary in the ⴖyⴖ
direction and well bore length along the ⴖxⴖ axis. Changes in the aspect ratio range from all fracture contribution (AR > 20) to complete
square flow (AR ⬇ 1). Adapted from Ref. 4

For example:
AR ⫽ 1 when the drainage pattern is square, i.e., ⬙X⬙ and ⬙Y⬙ drainage distances are equal. This
shape will not occur for the very low permeability case.
AR ⬍⬍ 1 as the drainage length becomes greater along the ⬙X⬙ direction than in the ⬙Y⬙ direction.
For instance, a hydraulic fracture or a horizontal well penetrating a very low permeability interval.
Segments
Well histories seldom exhibit a smooth decline because mechanical or reservoir affects often alter the
relative volumes of the oil, water and gas production phases. Therefore most production histories should
be divided into ⬙producing segments⬙ when projecting future performance with any degree of certainty.
Future performance should be predicated on the latest characteristics not necessarily an average of the
total history.
Method of Analysis
Production records are plotted in four different coordinate systems for mutual analyses. Each plot may or
may not furnish a portion of the final cohesive picture. Comparison of plots often reinforces or dismisses
initial conclusions about well producing potential and number and meaning of interpreted segments.
Four basic production plot form the basis for the interpretation technique.
1. Semi log Rate vs. Time Performance (qo vs t) –Review performance history of the rate decline
history and also associated gas (GOR) and water (WOR) characteristics. Can I divide the history
into producing segments? Couple this analysis with well records review.
2. Log Log Rate vs. Time (log q vs log t) –If possible identify transient and boundary dominated flow
regimes.
4 SPE-179994-MS

3. - Aids in defining transient and boundary dominated flow for a linear system and if
identified, extent of formation damage.
Cinco Ley, et.al.5 presented a (constant pressure solution (pD) vs. dimensionless time (tD) solution
to illustrate transition from infinite acting to boundary dominated conditions in a dual porosity
system. The slope of the line becomes less curved as fracture conductivity decreases. The line
becomes straighter as fracture transmissibility becomes more like matrix transmissibility, i.e.,
matrix permeability increases.
Gonzales and Cinco-Ley6, analytically verified by Wattenbarger, et.al.7, showed the slope of a
plot provides an estimate of fracture flow properties for the infinite acting case. The
equation suggests the steeper the slope of the line the greater effect of matrix permeability on the
total flow system.
Additionally the authors showed the plot theoretically initiates at the origin. In actuality the curve
often begins at some positive value which eventually trends to the expected straight line. This
phenomenon is the result of formation damage effects.
4. Real Time (t) vs Material Balance Time (tmb/t) Ratio.
Horner8 applied an approximation technique to smooth erratic production fluctuations without the
need for superposition (tp). Palacio and Blasingame9 defined equivalent material balance time
(tmb), also called pseudo-producing time for smoothing erratic production behavior. The defini-
tions are identical.
The approximation is intuitive and does not have a theoretical basis. Many investigators have
applied this decline time function due to simplicity of the calculation procedure coupled with its
smoothing abilities.
(2)

Poe showed material balance time (tmb) will always be greater than dimensionless (tD)10. There-
fore, tmb/t ⬎ 1. The principal cause for this observation is changing producing relationships in the
dual permeability flow system as a function of time and permeability distribution.
Poe10 expressed dimensionless material balance time in terms of wellbore geometry (Lc).
(3)

Where characteristic length (Lc) is equal to one-half lateral extent of the well source/sink. This
relationship translates the radial flow case to a general equation able to describe multiple flow
systems. For example:
Lc ⫽ (rw) for an un-fractured vertical well.
Lc ⫽ the fracture half length for a vertical well.
Lc ⫽ half of the effective horizontal wellbore length.
Studies on the affect of well drainage geometry on the material balance time – real time
relationship were conducted by Poe10. Conditions such as fractured vertical well, effects of
formation damage, fracture conductivity and wellbore length were investigated. All results showed
the (tmb/t) ratio could be off by over 200% for the short horizontal well case depending on linear,
bilinear or pseudoradial condtions.

Dimensioned (tDmb/tD) ratio Relationship


Express the (tDmb/tD) ratio in dimensioned terms:
(4)
SPE-179994-MS 5

Studies were conducted on the effect of the Arps11 (b-exponent) value on the shape of the (tmb/ vs t)
plot, Fig. 2. Curves range from exponential, to boundary dominated, b – 0, to transient, b ⫽ 20 exponent
conditions. Note the dramatic change in shape ranging from exponential to transient. This shape defines
the overall well flow system.

Figure 2—Effect of (b-exponent) on relationship between (tmb) and real time (t) ratio. The exponential (b ⴝ 0) curves slightly concave
upward while the remaining curves all represent increasing transient flow regimes bend concave downward until the last (b ⴝ 20) which
is essentially flat.

Exponential – Well operates under pseudoradial flow which indicates a good communicating well-
bore, fracture system and slightly greater than normal matrix permeability, In other words an expanded
drainage volume.
Hyperbolic ranging (0.1⬍ b ⬍ 1) - Signifies minimal contribution of the matrix flow system either
from marginally effective fractures or low matrix oil permeability. Ilk12, et. al., showed through
simulation studies how the Arps curve flattens out as a function of increasing (b-exponent) value
which is similar to our work.
The flow system ranges in quality from extreme pseudoradial to extreme linear according to location
of the curve between the two boundaries.
Super Harmonic (b ⬎ 1) - Transient flow indicating fracture unloading with little inflow effect from
the surrounding matrix.

Case Study
Map 1 represents locations of the 17 wells drilled on the subject lease. Laterals were generally drilled in
a northwest – southeast direction. However, compare the direction and production of the EAU1 well to
the companion wells. Did the operator do something different?
6 SPE-179994-MS

Map 1—Well locations in subject lease. Note: predicted remaining reserves included beside well.

Color coding of the well track identifies particular operators. It is apparent some operators appear to
have done a better job than others. Completion efficiency appears to be a very important part of the
process.
Each well listing contained monthly oil, water and gas production information. These data were
subjected to the four previously discussed plotting techniques for analysis. Interpretation of character of
each well was derived from this analysis which is shown at the bottom of the page. Figs. 3 and 4 represent
the resulting data sheet for two of the seventeen wells. Note how conclusions derived from one plot may
reinforce or dismiss previous ideas generated during analysis by comparative analysis.
SPE-179994-MS 7

Figure 3—WELL EW111 – An example of production behavior displaying both transient and boundary dominated conditions.

Figure 4 —Well E1U – Example of production history encompassing two segment histories.
8 SPE-179994-MS

An example of the well by well summary is shown in Table 1. Transient and boundary dominated
conditions are apparent in two of the 17 lease wells with two additional wells seemingly displaying this
transition history. Data variability or lack of slope definition precluded similar conclusion in remaining
wells.

Table 1—Well Analysis Summary

Map 1 represents distribution of reserves for the wells subjected to this study. Ultimate recovery is a
function of cumulative production to date (11/14) plus expected future performance down to a 3 BOPD
economic limit.
SPE-179994-MS 9

The map indicates four major areas displaying different production characteristics.
1. North area – Marginal reserves potential.
2. EW111 – Mediocre well located at northwest periphery.
3. EAU1 – Single high value well located at the southeast periphery. Appraisal potential?
4. Southern portion of lease. Developed by S3, A1, S4, E4, E8 wells. E4 always a poor producer.
Completion problems?
5. Southern boundary defined by E5, E6 and E7 wells.
Fig. 5 – Refines our general knowledge about the Eagle Ford production in this area. The plot locates
decline rates as a function of ultimate recovery as a function of well location. Lease well decline rates run
the gamut of good to bad with the majority ranging from 25 to 60%. Unfortunately decline rates for the
two good wells lie within this same range.

Figure 5—Relation between credited productions with expected future performance. The EAU1 and E8 wells certainly require further
review. Well 1 was a good producer, why no additional reserves?

Fig. 6 – Comparison of production histories for area 4. Note the two outliers. Well A1 probably
suffered casing or sanded up around 1,200 days. Well E4 seemed to have suffered an ineffectual
completion and has always been a poor producer.
10 SPE-179994-MS

Figure 6 —Composite production plot for area of interest. Compare the recompletion potential of the S4 and E4 wells to the surrounding
wells on the property map.

Note how the good wells appear to produce nearly at a constant rate. Could reserves be underestimated?

Study Conclusions
Analysis of the figures and Summary of Results pages revealed these follow up questions.
● Offset the EAU1 area?
● Review workover potential of the two wells located in the southern area and possible extension of
the current productive northwest limit.
● In both cases geological studies must be incorporated into a more in depth study.

Conclusions
1. Information available on public oil and gas records is of sufficient quantity and quality to analyze
inter-well characteristics on an individual or regional basis.
2. Comparing production plots representing differing fundamental meanings provides a method for
determining individual well character.
3. These studies help identify by-passed reserves and additional remedial potential in this portion of
the Eagleford producing trend.
4. Areas of low potential and high remaining value were readily identified even though there was
considerable variation of well producing character.

References
1. IHS Enterdeq Browser, IHS Corporation: http://www.ihs.com
2. FracFocus – Chemical Disclosure Registry: http://fracfocus.org/
3. Python programming language - https://www.python.org/
4. Poe, B. D., & Poston, S. W. (2010, January 1). Evaluation of Time to Flow Stabilization and Effective Drainage Area
of Slow-Stabilizing Wells Using Production Decline Analyses. Society of Petroleum Engineers. doi:
10.2118/135475-MS
5. Cinco-Ley, H., & Meng, H.-Z. (1988, January 1). Pressure Transient Analysis of Wells With Finite Conductivity
Vertical Fractures in Double Porosity Reservoirs. Society of Petroleum Engineers. doi: 10.2118/18172-MS
6. Cinco L., H., Samaniego V., F., & Dominguez A., N. (1978, August 1). Transient Pressure Behavior for a Well With
a Finite-Conductivity Vertical Fracture. Society of Petroleum Engineers. doi: 10.2118/6014-PA
SPE-179994-MS 11

7. Wattenbarger, R. A., El-Banbi, A. H., Villegas, M. E., & Maggard, J. B. (1998, January 1). Production Analysis of
Linear Flow Into Fractured Tight Gas Wells. Society of Petroleum Engineers. doi: 10.2118/39931-MS
8. Horner, D. R. (1951, January 1). Pressure Build-up in Wells. World Petroleum Congress.
9. Palacio, J. C., & Blasingame, T. A. (1993, January 1). Decline-Curve Analysis With Type Curves - Analysis of Gas
Well Production Data. Society of Petroleum Engineers. doi: 10.2118/25909-MS
10. Poe, B. D., & Marhaendrajana, T. (2002, January 1). Investigation of the Relationship Between the Dimensionless and
Dimensional Analytic Transient Well Performance Solutions in Low-Permeability Gas Reservoirs. Society of Petro-
leum Engineers. doi: 10.2118/77467-MS
11. Arps, J. J. (1945, December 1). Analysis of Decline Curves. Society of Petroleum Engineers. doi: 10.2118/945228-G
12. Ilk, D., Perego, A. D., Rushing, J. A., & Blasingame, T. A. (2008, January 1). Integrating Multiple Production
Analysis Techniques To Assess Tight Gas Sand Reserves: Defining a New Paradigm for Industry Best Practices.
Society of Petroleum Engineers. doi: 10.2118/114947-MS
13. Poe, B. D. (2002, January 1). Effective Well and Reservoir Evaluation Without the Need for Well Pressure History.
Society of Petroleum Engineers. doi: 10.2118/77691-MS