Oilfield Technology - May/June 2020

29 / 56

May/June 2020 Oilfield Technology | 27
processes to added featureswas fuelled in part by end-users and feedback
fromoperational personnel.
Usecases
Following a KPI meeting, a customer requested certain changes to cost
benchmarks that would better illustrate their specific concerns. The
platform’s flexibilitymeant that the updatesweremade inminutes across
all wells for an easy comparison of the entire drilling campaign’s history.
While data analytics platforms continue to claimheadline-worthy
gains, most of the efficiencies are in the details. For example,
AES ANALYTICS’ geo-fencing capabilities offer quick insight intowell
risks based uponGPS coordinates andwellswithin a specific radius.
The targets are adjusted by depth, which is critical when considering
many unconventional developments in the same area target different
formations. When risks aremitigated in advance, it is difficult to capture
the true cost savings of non-events until they accumulate acrossmultiple
wells (Figure 5).
In one basic example, a drilling consultant in the field requested the
addition of morewetting agent than programmed. Using the platform,
the accountmanager overseeing the operationwas able to demonstrate
that the current treatment regimematched all of the nearbywells. This
data-driven answer prevented a costly overtreatment, saving the customer
nearly US$6000.
In another example, the customer asked for a reviewof barite
consumption during ameeting. Typically, thiswould require the data
to be gathered and a separatemeeting scheduled to discuss
its implications. Instead, the accountmanager opened the
platform, presented the data, and an informed decisionwas
made on the economics of barite recovery for a solids control
set-up.
Whilemany visuals are best observed on a single or dual
monitor set-up, the platform is available formobile devices
(Figure 6). This allows for connectivity almost anywhere;
whether at the rig site or in an impromptu discussion, there is no
lag between information and informed decisions.
Its capabilities, fromcore processes to added features, are
nowbeing expanded, as a result of feedback fromend-users
and operational personnel. New features expand the original
data sets to even greater detail, alongwith newopportunities for
usage of the data.
In the Permian Basin, US, mudweight selection is extremely
challenging. Highly variable pressure regimes and unexpected
loss zones create significant uncertainty. To assist customers
with better information, a new tool provides a statistical
breakdown of mudweights in the area and subsurface losses
encountered. The distribution helps to determine themud
weightmost likely to be effective for drilling a target formation.
As the systemmatures, new features providemore and
more of the answers to questions that were not easy to answer
without first taking time to investigate and respond. Faster,
more informed decisions yield better results, instead of isolated
information or cases frommemory.
The benefits realised by users continue to lead tomore
features and greater adoption. Data-driven solutions are now
the standard as the evolution of decision-making processes and
simplification of data-gathering tasks takes place.
The aggregate deliverables of the systemare starting to
deliver noteworthy value, but it began through collaboration
between experts and data scientists, and a basic objective. With
a robust, standardised data structure, the future of data science
extends beyond just visual analytics. With the foundation set,
machine learning provides the potential to automatically review
reports and suggest responses. Predictivemodelling can identify the
greatest risk towells and determine themost cost-effective risk-mitigation
techniques.
Someday in the not so distant future, it will be possible to turn some
decisions over to an AI enginewithin the platform. Consider awell-control
event where themud engineer is on the pitsweighting up the system for
a killing operation. Barite consumption is noted, alongwith the event in
real time, and an order is placed. Trucks are on the road, without having to
leave a critical activity to place an order, and arrive in plenty of time.
Conclusion
Big Data is the future, and its boundless possibilitiesmake it difficult
to knowwhere to start. The promise of identifying serious bottlenecks
and rapidly delivering prescriptive actions to ensure optimal operating
conditions has left everyone racing to implement new tools. As companies
throughout the industry continue to strugglewith large amounts of
valuable data, one thing remains clear: a data analytics platformbuilt on a
sound foundational data structurewill deliver pragmatic, concrete results.
This positions the technology for future developments inmachine learning
and AI.
Reference
1. Accenture, ‘Industrial Internet Insights Report for 2015’,
acnmedia/Accenture/next-gen/reassembling-industry/pdf/Accenture-Industrial-
Internet-Changing-Competitive-Landscape-Industries.pdf (2015).
Figure 5. Dashboard visualisation comparing various costmetrics across amulti-well
drilling campaign.
Figure 6. Drilling fluids engineer reviewing offset fluids data through themobile app.