“Carbon Neutral”

Driven by Digital Transformation

Loh Wei Hong

Senior Regional Manager

Ascendas Systems

1
2
AI for Predictive Maintenance by Digital Engineering
Korea Institute of Energy Research Develops AI-Based Predictive Maintenance Models for Offshore Wind Power

8,000+ Components 3,000+ Signals AI

Ref. user story of KIER 3

Digital Engineering Platform
Korea Institute of Energy Research Develops AI-Based Predictive Maintenance Models for Offshore Wind Power

“Despite having little previous

experience with AI... we completed
a diagnostics model in MATLAB
capable of detecting wind turbine
component failure with over 90%
accuracy.”

—Jung Chul Choi, senior researcher at

Korea Institute of Energy Research

Ref. user story of KIER 4

Integrate DX into Carbon Neutral

Digital Transformation Carbon Neutral

Engineering
Access Data Analyze Data Develop Deploy

Data
Sensors AI model Desktop apps
exploration

Algorithm Enterprise
Files Preprocessing
development systems

Domain-specific Modeling & Embedded

Databases
algorithms simulation devices

Operation

5
Carbon Neutral Driven by DX

Digital Transformation Carbon Neutral

Cloud + AI
KGI・KPI
Big Data
Process Management and Quality Improvement

Business Strategy
DX for Internal Operation DX for Customers

Process Electricity & Thermal Technology Service

6
Development Challenge and DX Utilization
Technology Readiness Level
System Test,
Prove Feasibility Technology Demonstration Deployment & Operations

1 2 3 4 5 6 7 8 9

Basic Research Technology Development System Development

Using untested technologies that Combining technologies to

Challenge
have not been proven out implement one system
Data Analytics/AI System Simulation
(e.g., Forecasting, Optimization) (e.g., Compliance with grid codes)

Ref. user story of Vestas

7
Development Challenge and DX Utilization
Technology Readiness Level
System Test,
Prove Feasibility Technology Demonstration Deployment & Operations

1 2 3 4 5 6 7 8 9

Basic Research Technology Development System Development

Using untested technologies that Combining technologies to

Challenge
have not been proven out implement one system

Digital Engineering Platform

Techno-Economic Optimization Detailed Technical Design Operational Optimization

8
Development Challenge and DX Utilization
Technology Readiness Level
System Test,
Prove Feasibility Technology Demonstration Deployment & Operations

1 2 3 4 5 6 7 8 9

Basic Research Technology Development System Development

Using untested technologies that Combining technologies to

Challenge
have not been proven out implement one system

Digital Engineering Platform

Techno-Economic Optimization Detailed Technical Design Operational Optimization

System models not shared or reused. Not clear what to change and what to keep the same
9
 Digital Engineering
Do things better
Optimization

Techno-Economic Optimization

10
Renewable Energy Microgrid with Hydrogen Production

11
Renewable Energy Microgrid with Hydrogen Production

Energy storage
charged only from
renewable energy

12
Renewable Energy Microgrid with Hydrogen Production

Electrolyzer supplied
only from renewable
energy and energy
storage

13
Renewable Energy Microgrid with Hydrogen Production

Load supplied from

utility grid,
renewable energy
and energy storage

14
 Optimization Objectives

• Determine power rating for

wind, solar and storage
• Determine energy rating
for storage
• Produce as much green
hydrogen as possible for a
given electrolyzer power
rating
• Minimize overall cost/
Maximize revenue over a
20-year period at one-hour
time-steps (assume basic
cost model)

15
 Optimization Formulation and Results
Optimization
• 8760 load, wind and solar data was taken from
NREL
• 8760 price data was taken from system operator
market price
• Optimization was formulated using linear
programming in Optimization Toolbox
• 8760 optimization solves in approx. 30 seconds.
• 20-year optimization at 1-hour time-steps
(175,200 time-stamps) solves in approx. 900
seconds.

Internal Carbon Price

Techno-Economic Analysis
16
 Digital Engineering
Do new things
Transformation

Advanced Energy Management

17
Transformation for Carbon Neutral
Unplugged: The New Way to Charge Electric Vehicles

Model-Based Design

Vehicle to Grid/Home

Ref. user story of Lume Freedom 18

 New Business Development (Energy Management System)

CEMS BEMS

Vgen: Virtual Power Plant BuildingIQ: HVAC Energy

Optimization

H2
HEMS FEMS
Fujitsu AT: Hydrogen generation by PV
Denso, Panasonic: Home Energy Management 19
Development Workflow of Energy Management System
VGEN Develops Virtual Power Plant with Deep Learning and Machine Learning

AI Optimization

Ref. user story of Vgen 20

Digital Engineering Platform
VGEN Develops Virtual Power Plant with Deep Learning and Machine Learning

“MATLAB enabled us to develop

machine learning models,
optimization algorithms, and user
interfaces in a single environment,
which reduced development time by
about 50%.”

Access Data Analyze Data Develop Deploy —Jung Chul Choi, senior researcher at
Korea Institute of Energy Research
Data
Sensors AI model Desktop apps
exploration

Algorithm Enterprise
Files Preprocessing
development systems

Domain-specific Modeling & Embedded

Databases
algorithms simulation devices

Ref. user story of Vgen 21

Factory Energy Management System
Musashi Seimitsu Industry develops factory energy management system

“With Simscape Electrical and Embedded

Coder, we were able to develop our factory
energy management system (FEMS) through
Model-Based Design and automate code
generation. This enabled us to validate and
test the system through simulation and
reduce manual processes,
development time, and costs.”
—Tomohisa Sukigara, Musashi Seimitsu Industry

Ref. user story of Musashi Seimitsu Industry 22

Digital Engineering
DX Platform

23
Systematic Use of Digital Twin

System System Functionality Subsystem Subsystem System Integration Operations and

Requirements and Architecture Design Implementation and Qualification Sustainment

Use Cases Behavior models Physics-based C,C++ Model-based V&V

Physics
Predictive
maintenance
Docs & models System Components VHDL, Verilog Code-based V&V

Software
Digital Twins
Architecture models Data-driven GPU code Certification
workflows
Structured text

Digital Thread

Predictive Maintenance
Techno-Economic model Detailed Design model System Level model model
24
Systematic Use of Digital Twin
Atlas Copco Minimizes Cost of Ownership Using Simulation and Digital Twins

Process
• Company-wide workflow
• Used throughout product lifecycle
• Optimized maintenance and Data
Analytics platform
• Continuously updated digital twins

People
• Collaboration platform for efficient
communication
• Standardized accurate configuration tool
used by global sales

Results
• 120k+ connected machines
• Quick implementation of upgrades
• Re-establishing Atlas Copco as
undisputed global market leader
Ref. user story of Atlas Copco 25
Carbon Neutral with DX

In Practice

26
“Carbon Neutral” Driven by Digital Transformation
Digital Transformation Carbon Neutral

Cloud + AI KGI・KPI
Big Data
Process Management and Quality
Improvement

DX for Internal Operation Business Strategy DX for Customers

Digital Engineering Platform

Techno-Economic Optimization Detailed Technical Design Operational Optimization

27
MATLAB as Digital Engineering Platform

System System Functionality Subsystem Subsystem System Integration Operations and

Requirements and Architecture Design Implementation and Qualification Sustainment
Techno-Economic Optimization Detailed Technical Design Operational Optimization

Digital Twin

Auto Code Generation

Deployment Option

Embedded Target Cloud/Server

AI Optimization Predictive Maintenance Data Analytics

….
28
Ecosystem with 100+ third-party tools and languages

Data platforms

NX Motion AGI STK

Continuous- Discrete-Time Physical State Discrete-

Time Models Models Models Machines Event Models

Virtual ECU
models
ModelCenter HMI development

29
Upskilling Support

Training Consulting

Two EVLO 500s, part of a 4-

MWh energy storage system
coupled with a PV solar farm
near Montreal, Canada.

In their design workflow, EVLO engineers

use Simulink to model both the EMS
controller and plant. To support their product
onboarding and streamline their workflow,
EVLO worked with MathWorks
application engineers and consultants.
The EVLO team is now fully self-sufficient.

30