MRST: MATLAB® Reservoir Simulation Toolbox

A rapid, adaptive solution for poroelastic multi-phase

fluid flow analysis and simulation of reservoir models
Chris R. Wells
Global Manager – Energy Solutions
MathWorks
October 2023

Artificial Big Data Deep Machine Reinforced Predictive Internet Process Process Process Value Chain
Intelligence Analysis Learning Learning Learning Analytics of Things Optimization Digitization Automation Integration

© 2023 The MathWorks, Inc.

1
Outline

▪ MathWorks® digital solutions

▪ MRST: MATLAB® Reservoir Simulation Toolbox
– MRST in a nutshell
– MRST advantages
– MRST examples
– MRST modules
– MRST resources

© 1994-2023 The MathWorks, Inc. 2

MathWorks® Digital Solutions
Fact Sheet Highlights
▪ MathWorks® is a private company founded in Massachusetts, USA in 1984 to:
– Provide the ultimate computing environment for technical computation, visualization,
design, simulation, and implementation
– Accelerate the pace of discovery, innovation, development, and learning in engineering
and science
▪ MathWorks® has developed major digital solutions for industry and academia:
– MATLAB®, the language of engineers and scientists, for algorithm development, data
analysis, visualization, and numerical computation.
– Simulink®, a block diagram environment for model-based design and simulation of
multidomain and embedded engineering systems; plus…
– …over 120 digital products for data analytics, image/signal processing, control systems,
robotics, deep learning, digital twins, and many, many more.

© 1994-2023 The MathWorks, Inc. 3

 MathWorks® Digital Transformation Solutions
Digital Transformation Solutions Ecosystem
Energy sector demands Digital technology enablers Digital business outcomes
Digital • Process optimization • AI/ML & HPC computing • Prompt big data-driven analysis
Transformation • Sustainability compliance • IoT, digital twins & robotics • Asset performance optimization
• Circular economy innovation • Process digitization (modeling, • Value chain automation
Process • Energy transformation simulation, predictive analysis) • Process optimization & integration

AI/ML & HPC Integration Computational Modeling & Simulation* Process Automation

MathWorks®
Digital digital twin

Transformation
Solutions
predictive
analytics

Key transformational challenges Value of digital technology ROI as a digital business*

Challenges • Organizational barriers • Accelerated decision making • Upstream: $5/boe incr. ROI
& • Unclear ROI from digital • Preventive decision making • Downstream: $1/boe incr. ROI
Outcomes • New digital skills and expertise • Agile production optimization • Up to 20% in extra reduction of
• Siloed vs integrated digitization • Adaptive integrated platforms carbon emissions (scope 1 & 2)

* Sources: Hassmann et al (2022) – Technology transformation in oil and gas. McKinsey& Co.
Lie et al (2021) – Advanced modeling with MATLAB Reservoir Simulation Toolbox. Cambridge University Press
© 1994-2023 The MathWorks, Inc. 4
MathWorks® in Energy Resources
Customizable Digital Solutions for Upstream & Downstream
MathWorks® Process Process
Solutions Modeling & Simulation Optimization & Automation

SeReM + MRST
MATLAB® Reservoir MATLAB® MATLAB® MATLAB®
Modeling & Simulation Production Cloud Quantum
Subsurface Server Service Computing

Drilling
Digital Twin Parallel GPU
Simulink® Surface Simulink®
Computing Computing
Simscape
Oilfield Drilling Automation Toolbox Toolbox

Plant
Digital Twin
MATLAB® & Simulink®
Image/Signal Computer Machine Deep
Plant Predictive Maintenance Learning Learning
Processing Vision
Toolboxes Toolbox Toolbox Toolbox

5
MathWorks® in Energy Resources
Data Science Workflow for Big Data Analysis
Data Processes Technical Requirements Business Applications

Subsurface UPSTREAM
data (Tb-Pb) Subsurface Subsurface

Database size
Process Modeling Process Simulation
Monitoring
Big-Data data (Gb-Tb)
Structural Modeling

Gathering
Production MATLAB® MATLAB® MATLAB®
data (Mb-Gb) Production Cloud Quantum
Server Service Computing
months Acquisition frequency hours
Facies Modeling

Quantum
Computing capacity

Computing

High-Power Cloud
Computing
Computing Parallel GPU
Simulink®
Computing Computing
Parallel GPU Toolbox Toolbox
Simscape
Computing Computing
DOWNSTREAM
hours Processing turnaround seconds
Dynamic Process Modeling & Simulation
Process Complexity

Image/Signal
Processing
Computer
Predictive Vision
Analytics Machine
Image/Signal Computer Machine Deep
Learning Learning Learning
Processing Vision
Deep Toolbox Toolbox Toolbox
Toolboxes
Learning
low Data Predictability high

© 1994-2023 The MathWorks, Inc. 6

MathWorks® – Digital Subsurface Toolset (v2023)

SeReM MRST AI HPC IPCV

Subsurface Geoscience Subsurface Engineering Machine & Deep Learning CPU | GPU | Cloud | Quantum Image Proc. | Comp. Vision

STEM BDIA
Digital Geoscience & Engineering Analysis Big Data & Image Analytics

MATLAB®
Key technology differentiators
▪ Customizable STEM and BDIA toolboxes developed and fully interconnected on MATLAB® platform
▪ Model-based and data-driven geoscience & engineering workflows to maximize data & image usage
▪ MathWorks® support, training, and development of data science, engineering, and analytics solutions
▪ Adaptive digital solutions to assess and integrate new energy processes using high-end technologies
▪ Low-cost, high-quality software solution to maximize technical expertise, IT infrastructure, and budget
▪ 200+ energy companies globally currently use MATLAB® solutions across upstream and downstream
© 1994-2023 The MathWorks, Inc. 7
MRST: MATLAB® Reservoir Simulation Toolkit

▪ MRST is a MATLAB®-based toolbox with building blocks necessary to

design, prototype, and build simple and complex dynamic reservoir models

▪ Developed by the Computational Geosciences group at SINTEF Digital, a

research consortium based in Norway sponsored by industry leaders (CVX,
XOM, RDS, SLB, ECP, ENI), renowned academia (TNO, TU Delft, Heriot-
Watt, NTNU), and more than 25,000 users worldwide.

▪ Enables rapid prototyping of reservoir simulation concepts

▪ Offers a wide range of data structures and algorithms that can be combined
to create customized simulation workflows

© 1994-2023 The MathWorks, Inc. 8

Key Advantages of MRST

▪ MATLAB®-based modular design compatible with ECLIPSE decks

▪ Customized black-oil and compositional model solvers

▪ Exhaustively validated against industry-standard benchmark cases

▪ Automatic Differentiation (AD) framework with extensive diagnostic tools

▪ Optimized linear and nonlinear solvers

© 1994-2023 The MathWorks, Inc. 9

MRST in a nutshell
Reservoir simulation grids, operators, diagnostics, and models

© 1994-2023 The MathWorks, Inc. 10

MRST highlights
Structured, unstructured, and optimized grids including perpendicular bisector (PEBI)
and conformable, adaptive facies/fault/fracture/borehole-centric CFD grids
structure facies baffles porosity permeability

sedimentary beds faulted grid composite PEBI grid (1) composite PEBI grid (2) composite PEBI grid (3)

© 1994-2023 The MathWorks, Inc. 11

MRST highlights
Black-oil and compositional simulation and diagnostics on compressibility, thermal
expansion, and miscibility fluid flow via automatic-differentiation (AD-OO) (1)
Incompressible, single-phase Compressible, thermal flow Multi-phase flow Incompressible & inmiscible
simulation (pressure, ITT) (rapid prototyping) and CO2 displacement Water + CO2 (WAG) flow
(homogeneous & heterogeneous)

© 1994-2023 The MathWorks, Inc. 12

MRST highlights
Black-oil and compositional simulation and diagnostics on compressibility, thermal
expansion, and miscibility fluid flow via automatic-differentiation (AD-OO) (2)
Black Oil AD-OO Modeling, Upscaling, and Diagnostics Flow simulation diagnostics
(MRST vs. Eclipse) (SAIGUP model example)
fluid distribution degree of saturation

water influx (900d) gas dissolved (900d)

Permeability upscaling Transmissibility upscaling

flow distribution diagnostics

© 1994-2023 The MathWorks, Inc. 13

MRST highlights
Multi-phase, fractured, geothermal, or unconventional reservoirs with poroelastic
processes including unsaturated, EOR/CCS, coupled geomechanical modeling
Fracture-bound compositional (CO2+CH4+H2O) flow Miscible, compositional Discrete fracture modeling
Fault-bound saturation profiles (511d) (CO2+CH4+CH10) modeling (pressure)

Fault-bound hydrocarbon saturation profiles (600+d)

© 1994-2023 The MathWorks, Inc. 14

MRST Example:
Dynamic model building of Black Oil+CO2 fluid flow

© 1994-2023 The MathWorks, Inc. 15

MRST Example – Nonlinear Production Optimization
Solve IPR/VLP problem to optimize production from multiple wells

Gas
Oil
Water

Key takeaways:
• MATLAB is efficient and robust solving complex nonlinear optimization problems
• MATLAB solved the IPR/VLP problem to optimize production from multiple wells
© 1994-2023 The MathWorks, Inc. 16
MRST Example – Sequential Geomechanical Simulation
Solve for realistic fractured unconventional reservoir production
Fluid Flow Model Matrix & Fracture Permeability Hydro-Mechanical Solver Production Metrics
(PVT properties, ode23 solver) (sequential geomechanics) (sequential geomechanics) High Matrix Permeability
Low SRV conductivity

Pressure Distribution Permeability Retention

Key takeaways:
• Pressure-drawdown is the most significant early impact on the productive life of a well.
• Geomechanics can delay, or suppress, impact of long-term production of released gas.
© 1994-2023 The MathWorks, Inc. 17
MRST Example – Groundwater Geothermal Simulation
(Geneva Basin, Switzerland | Alcanie et al, 2021)
3D Geological Model Geothermal Gradient & Heat Flow Models Final Geothermal Simulation Models
(0.5M cells) (4,000 time stamps, 3 formations) (thermal, structural, and petrophysical parameters)

© 1994-2023 The MathWorks, Inc. 18

MRST modules usage (2022)

▪ Unstructured grids MRST usage worldwide

▪ ECLIPSE I/O interface (EOY2022)

▪ Multi-scale solvers
▪ Black oil & compositional
▪ Fractured media
(DFM, EDFM, DPDP)
▪ Geothermal,
geochemical,
geomechanical modeling
▪ Unsaturated media
▪ Unconventional
segments
▪ CO2 process
(WAG/CCS)
▪ Flow diagnostics
© 1994-2023 The MathWorks, Inc. 19
MathWorks MRST Resources
Introductory Guide to Advanced Modeling with
MATLAB® Reservoir MATLAB® Reservoir
Simulation Toolbox Simulation Toolbox
(MRST) (MRST)

MathWorks® technical & sales support:

MathWorks® Technical Support SINTEF Technical Support on MRST CES Support in Middle East
+1-508-647-7000 (Plano, TX USA) +47-930-58-721(Trondheim, Norway) +971-4-427-3663 (Dubai, UAE)
www.mathworks.com www.sintef.no www.cesimulations.com

© 1994-2023 The MathWorks, Inc. 20

MathWorks® in Energy Resources
MATLAB® HPC Computing & Integration with Enterprise IT systems

MATLAB® Compiler MATLAB® Compiler SDK MATLAB® Coder Simulink®

GPU HDL
PLC Coder
MATLAB® Coder Coder
Excel Spark Python
Hadoop C/C++ Java .NET
Add-in Production
Server
Standalone C, C++ CUDA HDL PLC
Applications

MATLAB® Production Server HPC capabilities: Parallel | GPU | Cloud | Quantum Computing

© 1994-2023 The MathWorks, Inc. 21

 Artificial Big Data Deep Machine Reinforced Predictive Internet Process Process Process Value Chain
Intelligence Analysis Learning Learning Learning Analytics of Things Optimization Digitization Automation Integration

Accelerating the pace of engineering and science

22