American Oil and Gas Reporter - June 2016 - 25

TECHCONNECTIONS
Poorly Performed
Decline Curve Analysis
Won't Be Reliable
Although decline curve analysis is the oldest and most widely used practice for determining reserves, a paper presented at the
Society of Petroleum Engineers/International Association for Energy Economics Hydrocarbon Economics and Evaluation Symposium, held May 17-18 in Houston, suggests there still is much
to be understood about its practice and implementation.
In a paper titled, "The Practice of Decline Curve Analysis"
(SPE-179979-MS), author D.C. Purvis discusses decline curve
analysis as practiced in the oil industry being an established process.
But while decline curve analysis has been used for more than a
century, widespread misunderstanding and implementation
plague the industry. Purvis contends decline curve analysis
should be more than just a "curve-fitting exercise," but rather is
an exercise in "seeing, understanding and predicting reservoir dynamics."
As defined by IHS in a Fekete publication, decline analysis is
a reservoir engineering technique that extrapolates data from oil
and gas wells to, hopefully, accurately forecast the future production
rates and ultimate recoverable reserves in a particular reservoir.
Purvis and IHS look to J.J. Arps, an American geologist who
worked during World War II and published a mathematical relationship for the rate of production decline over time.
According to Purvis, decline curve analysis was used widely
in the early 1900s, but it was Arps' paper in 1945 that really advanced the process in a large way. In fact, Arps himself mentions
this in his review of the literature of the 40 years prior to his work,
in which he referenced the earliest published observation of this
process by authors Arnold and Anderson in their 1908 U.S. Geological Survey report, which observed a production drop over
time.
Additionally, the IHS site observes that while many papers have
attempted to modify Arps' work, his original method is still used
widely. According to Purvis, what made Arps' work stand out was
that he developed a more general hyperbolic equation to incorporate exponential and hyperbolic behaviors.
In his 1945 paper (SPE-945228), J.J. Arps states, "The basic
problems in appraisal work are determining a well's most probable future life and estimating its future production."
Arps goes on to say that volumetric calculations are good, but
may not be as valuable without additional data, and that the easiest and most reliable source of data is a producing well. The most
logical approach is to extrapolate production data and plot it against
time, then extend the curve until it reaches the economic limit,
at which point one can estimate the life of the well. The process
attempts to predict future life of a well or future oil recovery, assuming all trends or calculations remain the same over time.
As it turns out, Arps' assumption about trends and calculations
remaining constant is the thing that gets modern analysts into trouble. Purvis asserts that adequate analysis can be accomplished only
by drawing on general experience, examining the nearest analog
well while understanding operational changes as well as the uncertainty of the production data themselves. In other words, examining as much data as possible seems to be the key, but it is

"Doing the right thing the wrong way
is still wrong."
important to understand that not all data are created equal and
sometimes things happen that impact the data. These all play into
successfully estimating the life of a well.
Purvis says, "Doing the right thing the wrong way is still
wrong."
In other words, even if decline curve analysis is right for a given situation, if it is done incorrectly, it can be unreliable, at best.
Purvis highlights some examples of when decline curve analysis works best. These include things such as using quality graphs
and understanding the limitations of the data they show. He discusses the importance of graphs being instantly readable and presenting meaningful data. Purvis recommends using a consistent
frame of reference for the benefit of the end-user as well as the
analyst.
The author encourages analysts not to compress or expand
graphs, or show meaningless data. In addition, Purvis suggests
using multiple graphs for each well, and comparing graphs among
multiple wells with and extra eye for the older wells. Analysts
should consult multiple scales and types of graphs as they conduct their analyses. Sometimes it is tempting to rely on production data only, without regard for how operations have changed.
Only through reviewing operational changes can the analysis accurately determine reservoir trends.
Finally, Purvis recommends forming a functional understanding
of what the graphs, plus the geologic data, say about production
and the reservoir.
Purvis seems to be advocating a more thoughtful and consistent curve analysis, making sure to take note of as much data as
possible in an effort to develop forecasts that are more accurate
and reliable.
❒

JEREMY VISCOMI is the di-

rector of technology transfer
for the University of Kansas
Tertiary Oil Recovery Program and the PTTC Midcontinent Region office. He has
more than a decade of experience in developing and organizing technical conferences
and special events.
JUNE 2016 25

American Oil and Gas Reporter - June 2016

Table of Contents for the Digital Edition of American Oil and Gas Reporter - June 2016