SPE-196494-MS Sumatera Operation SMO Integrated Optimisation Decision Support Center IODSC Success in Embracing Digitalisation and Innovation To Deliver Business Results
SPE-196494-MS Sumatera Operation SMO Integrated Optimisation Decision Support Center IODSC Success in Embracing Digitalisation and Innovation To Deliver Business Results
SPE-196494-MS Sumatera Operation SMO Integrated Optimisation Decision Support Center IODSC Success in Embracing Digitalisation and Innovation To Deliver Business Results
This paper was prepared for presentation at the SPE/IATMI Asia Pacific Oil & Gas Conference and Exhibition held in Bali, Indonesia, 29-31 October 2019.
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
This paper will describe the success of Chevron Pacific Indonesia's (PT. CPI) Sumatera Operation (SMO)
to deliver improved business results through the implementation of an Integrated Optimization Decision
Support Center (IODSC). The SMO IODSC is focused on value optimization across the Rokan Production
Sharing Contract (PSC) by developing and implementing digital solutions ranging from the simple, such
as Short Message Service (SMS) gateway to monitor rig status, through to the more advanced, such as
Artificial Intelligence to identify failing equipment. Value creation was achieved by the systematic adoption
of digital solutions to reduce lost production opportunity (LPO), reduce expenses, improve energy efficiency
and optimize hydrocarbon production. Key to this systematic adoption is IODSC's Integrated Exception
Management (IEM) system. IEM supports CPI employees in their daily tasks by:
• Managing large amounts of data captured each day and using data science to change it to actionable
information.
• Capturing exception criteria based on Subject Matter Experts (SME) knowledge that automatically
identifies and prioritizes wells and equipment operating outside the desired condition (exception
signals).
• Housing workflows to enable review, action and close-out of the exception signals.
IEM provides "one-stop exception review" and "start to finish workflow" capabilities that documents
accountability for each step of the workflow and the time taken to complete. IEM allows for the
exception criteria to be defined by physical principles, subject matter expert (SME) knowledge, or Artificial
Intelligence (AI)-based techniques.
As a result of this systematic implementation of improved workflows and digital solutions, within 2 years
of commencement the IODSC exceeded it's 5 year value creation target.
2 SPE-196494-MS
Introduction
PT. CPI operates the Rokan Production Sharing Contract (PSC) on behalf of SKK Migas and the
Government of Indonesia. The Rokan PSC has a wide coverage area of about 6,264 km2 (see Figure 1) with:
■ ~ 9000 oil producers
■ ~ 1400 water and steam injection wells.
■ 85 active fields
■ 40 Gathering Stations
■ 50 Well Test Stations
■ 120 units of major rotating equipment (water injection pumps and gas compressors)
■ ~ 3000 km pipeline
■ 30 to 50 workover and well service rigs operating.
■ ~0.5 terabytes of digital data generated daily
Opportunity
The Sumatera Operation (SMO) Integrated Optimization Decision Support Center (IODSC) project was
initiated in September 2015 as part of Chevron IndoAsia Business Unit (IBU) Transformation project to
address the low oil price environment and improve business results.
Previously the Sumatera Operation has multiple, independent Decision Support Centers (DSC's) at
varying levels of maturity with different levels of focus and confined by asset boundaries. Value decisions
were made at the asset level and not at the Sumatera Operation portolio level, and different workflows, data
management tools and processes were employed for similar tasks.
The SMO IODSC project combined most of the existing Sumatra-based DSC's (see Figure 2) into one
central group (primarily well and facility monitoring DSC's), while also standing-up dedicated new teams
(Logistics Support Center and Analystic Support Center) to support the optimization and execution teams.
The objectives of this approach were to:
• Allow personnel to focus on value-adding optimization and efficiency tasks, while minimizing
time spent organizing or searching for data.
SPE-196494-MS 3
• Enable the utilization of subject matter experts (SME) across all Rokan PSC assets.
• Having dedicated work teams focused on improving key business metrics, such as well down time
and execution cycle times, compared to prior distributed model of responsibilities to act on these
metrics.
The dedicated project team worked for about 8 months (September 2015 – May 2016) to design and build
the workflows, tools and organization to support the implementation of the IODSC.
IODSC was staffed in May 2016 and formally launched on June 1, 2016. IODSC has 5 functional teams:
• 2 technical, analytic teams tasked with analyzing surveillance data and making recommendations
to reduce equipment failures, optimize production and energy consumption, and reduce lost
production.
○ Well Reliability and Optimization (WRO)
○ Surface Facility and Optimization (SFO)
• 2 Support teams
○ Logistic Support Center (LSC)
○ Analytic Support Team (AST)
• 1 Execution team
○ Rig Hub
Figure 3 documents the key roles and responsibilities for each IODSC team.
4 SPE-196494-MS
• recommending actions to reduce the downtime of surface equipment such as water injection pumps,
shipping pumps, gas compressors
• recommending actions to reduce surface production bottlenecks and optimize energy consumption.
• liaising with Operations, Maintenance and Facility Engineering teams to reduce lost production
due to surface equipment failures.
• ensuring well testing equipment is available to acquire surveillance data to support optimization
and investment decisions
The Analytic Support Team (AST) Team supports the 4 other IODSC teams by developing digital
solutions using readily available and low-cost platforms.
SPE-196494-MS 5
All 5 sub-teams in IODSC are located in central workspaces that promotes collaboration and cross-
functional approaches to achieve optimal results.
At a high level, the data that is gathered each day in Rokan PSC can be classified as either non-real
time data or real time data. The data written in the notebook/piece of paper or in engineer computer is less
valuable because other people can not use this data. Value can be obtained by collecting this data into a data
warehouse, so it can be correlated with other data and converted into information. IODSC leverages the
transformational impact of digitalization by "digitizing" SME rules and implements this as exception criteria
for well and equipment performance in IEM. With this approach, SME knowledge can be instantly and
consistently applied to incoming data streams to identify wells and equipment operating outside of expected
performance ranges. Capturing this SME knowledge in IEM has the additional benefit of mitigating the loss
of knowledge and experience when SME's are no longer in their roles.
In addition to SME rules and expert knowledge, Artificial Intelligence and predictive analytics modules
are also built in to IEM to identify exception conditions.
On a regular basis each day, exception signals are sent to the user based on priority and area responsibility,
so that users can focus their efforts and time on the highest impact tasks first. Exception signals are
prioritized by health, safety and environmental considerations, production impact and expense impact.
This IEM tool also acts as a workflow manager, so each step of the workflow is digitized and can be
tracked on completion time for accountability processes in certain date ranges with average BOPD impact
(see Figure 5). The bottleneck location in the workflow can be identified and improvement projects can be
initiated to reduce the time to complete and variance for the step.
6 SPE-196494-MS
IEM is the backbone tool of IODSC, and it has supported many improvement projects which have
delivered almost half the total IODSC value creation.
• consolidate with other supporting data (rig positioning Global Positioning System (GPS) and
logistic preparation status) to build a more comprehensive rig status monitoring dashboard
• enable rig scheduler to focus on well program prioritization
• improve collaboration between rig scheduler, logistic planner and transport tactical planner
SPE-196494-MS 7
Results:
Artificial Neural Network (ANN) Implementation on Short Cyclic Candidate Selection (SCCS).
Situation:
Short cyclic steam stimulation of oil producers is one of the most economic proactive well intervention
activities in the Duri field. The job execution has increased significantly in the last 5 years from 30 jobs
per month to 340 jobs per month, supported by the application of Artificial Intelligence to rapidly and
systematically identify candidates. Prior to the application of AI, the previous practice was to submit short
cyclic steam candidates from individual well reviews conducted by different Petroleum Engineers. As a
result, there was inconsistency between the Petroleum Engineers on selection criteria and time to identify
and submit candidates. In total, it took 2 to 3 hours to identify ~20 well candidates per day from the ~6000
active oil producer wells in the Duri field.
Solution (see Figure 7):
Figure 7—Short Cyclic Candidate Selection (SCCS) using Artificial Neural Network (ANN)
An Artificial Neural Network (ANN) using historical short cyclic steam stimulation results was used to
predict the result (‘success’ or ‘not success’) of potential future cyclic steam candidates, and then this signal
8 SPE-196494-MS
was captured in the Integrated Exception Management (IEM) System as a workflow to support the final
review and execution of the cyclic steam jobs.
Results:
• Reduce well candidacy cycle time from 2 to 3 hours to 15 minutes by adopting SME knowledge
in ANN architecture.
• Increase success ratio from 61% to 72% of executed jobs.
Utilize Artificial Intelligence (AI) and exception management to identify and proactively replace failing
Electrical Submersible Pump (ESP).
Situation:
Previously, there was no systematic early detection tool for failing Electrical Submsersible Pumps. Well
reviews had to be conducted individually to identify potential candidates. Early detection of failing pumps
can minimize lost production and well down time.
Solution (see Figure 8):
The AST working with WRO developed a Fuzzy Logic-based method using sparse surveillance data
from more than 2000 ESP wells to proactively identify failing ESP's and replace them prior to actual failure.
Wells that have a higher probability of operating near-failure condition will have a more negative Fuzzy
Confidence Index (FCI), and these wells will be prioritized for proactive ESP replacement. Similar to other
examples these failing wells are captured as exception signals in IEM with an associated workflow, so that
standard review, approval and execution steps can be tracked and followed-up to ensure completion.
Results (see Figure 9):
• Increase production above pre-job baseline level (average increase ~40 BOPD per job).
• Avoids potential 400 barrels oil of LPO from waiting on rig to replace failed pump.
Figure 9—AI Fuzzy Logic Results to identify and proactively replace failing ESP
SCADA Well Real Time Improvement using Intranet of Things (IoT) and Data Science.
Situation:
Previously welldown identification heavily relied on Field Operator manual inspections, which in some
cases had a lag time of 24 hours between visits to the well site.
A study conducted in January 2017 to track ESP motor state conditions using Supervisory Control
and Data Acquisition (SCADA) compared to Production information submission status confirmed the lag
time. There were production wells with SCADA, but this data was not fully utilized for rapid well down
identification.
Solution (see Figure 10):
• Cross functional team worked to install OLE for Process Control (OPC) to bridge SCADA
application with Historian server to make the real time data available within the PT. CPI network.
• Integrate and correlate motor state information from SCADA with production well information
status.
• Partner with field operations to monitor the production well status 24/7.
Figure 10—Connection of real-time data to wider network enabled significant reduction in well downtime.
Results
The response time to identify down wells reduced by 77% (from 1 day to few hours), and this resulted
in faster restoration of production.
10 SPE-196494-MS
Artificial Intelligence (AI) implementation for Weak Sucker Rod Pump Performance Identification.
One of the more recent applications of digital solutions in IODSC is the implementation of Case-Based
Reasoning to identify weak or failing sucker rod pumps (see Figure 11).
Figure 11—Process Overview Weak Sucker Rod Pump Identification using AI Case Based Reasoning
Situation:
Sumatera Operation (SMO) has over 2300 dyno cards collected daily from sucker-rod pump well online
monitoring. Because of limited resources, only about 10% of the dyno data collected were reviewed each
day based on pump fillage cut-off criteria. A cross-functional team was formed, consisting of AST, WRO
and subject matter experts to develop a digital solution to maximize the utilization of the data collected and
leverage this to increase production.
Results (see Figure 12):
• using case-based reasoning, all collected dyno cards can be analyzed everyday.
• 54% reduction in time to identify and execute the replacement of failing sucker rod pumps
• average 30% increase in weak pump replacement job submissions each month
Figure 12—AI Case Based Reasoning Result of Weak Sucker Rod Pump Identification
SPE-196494-MS 11
Results
As part of the business case for the IODSC, value creation targets were established by PT. CPI leadership
and within 2 years the IODSC had exceeded its 5 year value creation targets. This performance was achieved
because of the examples above, and other key achievements as listed below.
Optimized subsurface performance:
• automatic well tests in Heavy Oil increased by ~2500 tests per month.
Because of these successes, the IODSC was expanded to a larger scope in 2018:
• LSC Scope:
○ LSC was expanded to handle entire workover & well service fleet (40+ rigs)
○ Enhanced rig scheduling and material management processes and digital tools
• AST Scope:
○ Dedicated Artificial Intelligence (AI) and IEM Teams.
○ Expanded IEM to other teams outside of IODSC.
○ Key support role for digital solution development works outside of IODSC.
• WRO Scope:
○ Dedicated online well monitoring team
○ Expanded real time well surveillance
Conclusions
The key factors in IODSC's success in developing and deploying digital solutions to deliver business results
have been:
○ Full leadership support to embrace digital solutions.
○ Embracing "managing by exception" philosophy, documenting and improving business workflows
first, then engaging in software development to standardize workflows and capture the expert
knowledge.
12 SPE-196494-MS
○ Focus on developing digital solutions that integrate with the existing workflow manager (IEM).
○ In-house software development enables customization based on user needs.
○ Dedicated product owners from functional and execution teams willing to work closely with the AST
development team.
○ Dedicated and skilled AST development team collocated in one central workspace close to the end-
users (product owner).
○ Knowledgeable "digital translators". These are individuals who have functional expertise (e.g.
petroleum engineering, facility engineering, completion engineering), fluency with the application of
digital technologies in the oilfield (e.g. data science, automation, SCADA, information management
systems), and project management skills such that they can direct digital solution development
aligned with business needs
○ Sharing of best practices, knowledge and lessons learned from local and global subject matter experts.
○ Recognition and award systems to acknowledge successes and encourage further innovation.
Acknowledgments
The author thanks the Management of PT. Chevron Pacific Indonesia and SKK Migas for support to publish
this paper. Any opinions, findings, conclusions or recommendations expressed in this material are those of
the author and do not necessarily reflect the views of PT. Chevron Pacific Indonesia and SKK Migas.
Please note that substantial portions of this paper are from various internal PT. CPI documents and
presentations. Some images and portions of this paper have already been presented at PT. CPI meetings and
other related conferences and may be included in future presentations or publications.
The author would like to also recognize those IODSC team members that made significant contributions
to this paper:
• Mario Anggara Putra, Team Leader Material & Execution Management, LSC