American Oil and Gas Reporter - July 2021 - 51

wells/rigs using a " best driller " algorithm.
The algorithm is trained to automatically
optimize drilling operations from spud to
total depth. " That will definitely help us
tighten up consistency across all our
drilling programs, and make step changes
from there, " Lowe details, adding that for
Devon, rig days is still the primary yardstick
for measuring drilling performance.
However, keeping the wellbore in the
planned target zone is critical to maximizing
completion quality and overall
well productivity, Lowe points out.
" Geosteering windows will get tighter
and we are going to see a much more dynamic
geosteering interpretation based
on data acquired while drilling, " he states.
" Not only will we adjust formations up
and down based on what we see as drilling
proceeds, but we will also adjust more
complex measurements, such as porosity
and rock stress, while drilling to make
better decisions.
" The industry has become so good at
throwing the lateral dart right on the bull's
eye two miles away. It is really amazing, "
Lowe goes on. " But if we can improve the
real-time interpretation of what the bit is
encountering down hole, we will be able
to make even more fine-tuned adjustments
to landing and steering laterals. "
Changing The Game
Real-time adjustments is also the operative
term for completion design, according
to Lowe. At present, the industry
makes real-time decisions on frac jobs
based on operational data, such as pump
rates and proppant densities. Devon's
objective is to apply information from
offset wells to make those same types
of optimization decisions in real time
based on the actual geometry of the
fracture being created. " No one in the
industry is doing that yet, but we are
really close, " he enthuses.
The company took a major step toward
realizing that objective when it developed
a patented low-cost, nonintrusive method
to evaluate and quantify fracture growth
rates and fracture-driven interactions during
stimulation using only a surface pressure
gauge on a monitor well (see " Breakthrough
Technology uses Pressure Monitoring
to Monitor Cluster Efficiency, "
by Kyle Haustveit, Wolfgang Deeg, Brendan
Elliott and Jackson Haffener, in
AOGR's April 2020 issue).
" We think sealed wellbore pressure
monitoring will allow us to use offset
well information to make live fracturing
decisions that dramatically improve completion
quality and cost, " Lowe relates.
" The ability to make operational decisions
based on a subsurface understanding of
the fractures in real time is one of the
things we are most excited about. It could
be a game changer. "
Here again, machine learning is the
enabling technology behind Devon's ambitious
vision of the future completion
operation. Most adjustments to fracturing
parameters are today made between stages,
typically after something has occurred to
cause the completion design to be changed
before pumping the next stage.
" We are not far from being able to
teach a machine to make interpretations
on the fly so that we can make decisions
during a stage, " he says. " That is the
next big leap for us with respect to optimizing
fracturing designs, cutting costs
and making better wells. "
How would the ability to modify completion
design parameters in-stage using
subsurface fracture geometries open the
possibility of not only changing operational
variables such as pump rates and pressures,
but also more dynamic factors such as
proppant density, fluid viscosity, chemical
additive concentrations, and even proppant
grain size during a stage treatment?
" All of these things are coming. We
will step into it one piece at a time, "
Lowe responds. " Down the road, I expect
it will become fairly routine for us to use
real-time measurements to optimize fracture
geometry and fracture lengths by
making on-the-fly adjustments to proppant
volume and size or fluid viscosity. "
That kind of value-adding capability
has obvious and profound bottom-line
implications for Devon across its unconventional
asset base. " Real-time
digital solutions to optimize drilling and
fracturing may allow us to increase rates
of return by adjusting well spacing to
drill fewer wells with larger frac designs,
or by placing more wells closer together
but with smaller frac sizes. Either way,
these are the types of decisions that
could ultimately result in hundreds of
millions of dollars to Devon's bottom
line. That is the big win. We make investments
for that type of return on our
capital program. "
And high-frequency data is not confined
to only measuring and adjusting
what happens when the frac pumps are
on; it also is being used to optimize efficiency
when the pumps are off. " We have
seen a 100%-120% improvement in pump
times, going from pumping about 40%
of the day five years ago to pumping
more than 80% of the day on average. "
Lowe reports. " We recently set a record
of 22.7 hours of continual pumping operations
in a 24-hour period on one of
our wells in the Rocky Mountain region.
Layering In Video Data
Even before it entered the Barnett
Shale nearly 20 years ago, Devon systematically
applied technology to its producing
assets to maximize run times and
improve productivities and efficiencies.
But even after decades of investing in
An early adopter of real-time data analysis in its drilling and completions operations, Devon
began streaming high-frequency live data from rigs and frac spreads in 2012. This approach
has been instrumental to the company's rise during the past five years to the top of its peer
group for drilling performance and operational execution metrics. One of the next big steps
for Devon is automating drilling operations using machine learning and edge computing
based on " best driller " practices from top-performing wells.
JULY 2021 51

American Oil and Gas Reporter - July 2021

Table of Contents for the Digital Edition of American Oil and Gas Reporter - July 2021