Digitalization is fundamental to the development of Repsol’s strategy for the future. To meet emerging challenges, the business units have developed an ambitious program comprising multiple projects. Within Repsol’s Industrial Business, development of a refinery digital twin leads the digitalization program. The digital twin allows the business to maximize production while optimizing energy consumption. This session will explore the digital twin project objectives to improve the accuracy and scope of the Refinery LP model that the Programming and Planning departments use to make decisions regarding crude feedstock purchasing and refinery unit operations. It will also report on the context of the business goals achieved, the technology and architecture developed, and the connectivity deployed to communicate results. It will conclude with a description of how enhancements to existing technology work with new technologies to improve value.
3. Over 23,000 employees
in 24 countries (Main Headquarters
Madrid (Spain)
1. Repsol Worldwide
Industry challenges and how Repsol faces them
"Our future is bound to turn us into a global multi-energy provider, capable of supplying each
individual with the energy they need at any given time in a safe, competitive, and sustainable way"
Average net production:
~ 715,000 barrels of oil equivalent per
day. 750,000 barrels in 2020
SERVICE STATIONS:
+ 4,850 in Spain, Portugal, Italy, Peru
and Mexico
7 industrial facilities in Spain, Portugal
and Peru, with a refining capacity
of over 1 million barrels per day
We have incorporated low-emissions
power generation assets, with a
total installed capacity of 2,950 MW.
Target: 4,500 MW in 2025
Responsible mobility: 1,200 electric
vehicle charging stations and
797 Autogas (LPG) supply points
Distribution of petrochemical products
and lubricants in over 70 countries
CO2
reduction in 2018: 310 kton
CO2
eq emission in 2018: 10,3 Mton
CO2
reduction 2020: 2,1 Mton
190 Circular Economy
initiatives detected
2,500 M€ linked investment
to Energy transition
up to 2020
4. 23,500 /
84
People /
Nationalities
24 / 5
Present
in different
countries /
Continents
Upstream
Industrial / Commercial & Renewables
Both
1. Repsol Worldwide
Industry challenges and how Repsol faces them
5. “We are convinced that we must set more ambitious objectives
to fight climate change. We believe now it is the right time for
Repsol. We do it with the utmost confidence that we're investing
in the future, and addressing the significant challenges that lie
ahead with strategic clarity is what will enable us
to turn them into opportunities.“
Josu Jon Imaz, Repsol CEO
7. 2. Executive Summary
Why
How
What
Where
Who
Improve accuracy and frequency of updates for Planning
Models to take better decisions
Using a Digital Twin that combines Petro-SIM first principles
model with OSIsoft PI historian and dashboards
Digital Twin generation that allows to monitor reality vs.
model results (LP Vector & Simulation model), check health
and be able to update the LP vectors in an agile way
Bilbao’s FCC Unit in Spain
A combined Repsol*/KBC team developed the technology
* Refinery Planning Department, Business Department and Technology Lab.
8. 2. Digital Twin Solution - Objectives
Simplified workflow
Simplify model
evaluation
Simplify planning
model (LP vector
update
Model assurance
One version
of truth
•The Digital Twin
automates data
collection and
processing, so that
focus can be put in
analyzing results
instead of
generating data
•The Digital Twin
provides indicators
that are monitored
to check health of
LP Vectors and
Simulation model
comparing to
reality, at a glance
•If required, the
Digital Twin
provides a solution
to update the LP
Vectors based on
rigorous Petro-SIM
model
•The Digital Twin
provides early
detection and
notification of
relevant deviations
between actual
data and LP Vector
results
•The Digital Twin
provides access to
the process and
lab data but also to
derived indicators
that can be used
throughout the
organization
9. 3. Let’s Focus
A. The solution Architecture
B. Petro-SIM Model
C. PI – Vision Displays
D. Workflow
11. ▪ Simulation is a Petro-SIM calibrated first principle model used
in backcasting prediction mode, allowing:
• Calculation of critical operating parameters that allow a
better understanding and monitoring of the process. It is
sensitive to changes in feeds, operating conditions, catalyst
used and fractionation
• Updated calibration generation for digital twin itself
▪ Generates indicators to monitor input data quality, reality vs.
model results (LP Vector & Simulation model) and health of the
tool
▪ Generates new LP Vectors based on monitoring criteria if there
are deviations
▪ The model is run automatically as required – regular basis,
e.g. daily, weekly
▪ Petro-SIM model reads data from PI and write data back to PI
B. Petro-SIM Model
12. An accurate representation of the asset
over its full range of operation
Capture the full history and future of the asset
Automated, regular model runs.
Built-in to business workflows
Centralised single version of the truth, used by
everyone, outputs delivered directly to the business,
strong governance systems
B. Digital Twin vs. Traditional Simulation
Digital twin is a digital analog of the actual process that allows for
better operation and understanding of the facility
An accurate representation of a
particular operating case
Traditional Simulation Digital Twin
Provide a snapshot in time
Built on an ad-hoc basis to
answer a question
Owned and used by isolated groups
on an ad-hoc basis
13. ▪ Dashboards can incorporate results from
Petro-SIM alongside other PI data
▪ Displays are developed to help the users follow
the desired workflow.
• The Digital Twin generates a lot of information that
different stake-holders might use in different ways
• By using customized displays, we ensure that a
consistent procedure is used
▪ You can generate alerts to relevant deviations or
data quality
▪ Expert users can perform deeper analysis through
direct interaction with Petro-SIM, as needed
C. PI – Vision Displays
14. D. Workflow
?
Good LP match
Validated LP
model
Poor LP match
Update LP
sub-model
Poor simulation match
Update model
calibration
Scheduled
executions
Check DQPs ?
Bad input data
Check data
Dismiss results
Re-execute
historical runs
Validated input data
Check MPIs
15. 4. Digital Twin Solution - Implementation
Solution deployment
Model development
Workflow development
▪ Integration of historian data into the model: Identify missing data
▪ Testing based in historical data and evaluation of failed simulation runs: Identify systematic issues
▪ Development of error trapping mechanisms
▪ LP sub-model integration / Development of the LP update tool
▪ Coordinating between teams to agree on meaningful indicators
▪ Agreeing upon a workflow
▪ Developing displays to assist in those workflows
▪ Automating simulation runs
▪ Implementing the workflow
16. ▪ Apart from experience of subject matter experts (SMEs), most
decision-making activities related to improving plant profitability
(e.g. scheduling, planning, real time operations, retrofitting, etc.) rely
on a process model
• To change from traditional simulation to a digital twin solution assures
the best decision over time
▪ Digital twin model can accelerate identification and resolution of unit
issues and improved productivity
▪ Centralized solution provides information to all stakeholders
requiring information across the organization with no need for
advanced knowledge of the simulation model
• Digital twin provides a unified template from which the different teams
can discuss issues (model updates, bad lab data, etc). It constitutes a
single source of the truth
▪ The integration with PI allows modification of only one source and
leverage PI Vision dashboards
5. Summary and Conclusions
Real-Time
Data
SME
Domain
Knowledge
Production
Optimization
Digital Twin
/ Monitoring
Tools
Asset and
Supply Chain
Optimization
Asset and
Supply Chain
Models
Remote
Service
Support
17. 5. Summary and Conclusions
Potential Future Extensions
• First principles model may be supplemented with additional information and machine-learning models
to handle equipment and process states 🡪 HYBRID MODELS
• Monitoring economic performance of the unit
• Adding to the simulation model the energy KPI’s of another Project we are implementing
at the same refinery
18. The names of corporations, organizations, products and logos herein are either registered trademarks or
trademarks of Yokogawa Electric Corporation and their respective holders.
Digital Twins
Provide single source of the truth,
driving alignment of decisions and
actions across the value chain