Innovate & Scale

A Blueprint for Deploying a Simio Digital Twin

Platform
Paul Glaser
07-MAR-2024

© 2024 Simio LLC 1

About Me
Director of Consulting Services at
Simio

10+ years of Simio model

development experience
o Simulation
o Scheduling
o Project Management

Started as a Simio customer!

o Background in Operations
Engineering / Strategy / Supply Chain

© 2024 Simio LLC 2

Value to Stakeholders

© 2024 Simio LLC 3

Customer Problem Statement

“Hey – we need a low-cost way to quickly get some views into a

distant place that is hard to reach. We hear people are building
drones and we want one!”

© 2024 Simio LLC 4

Digital Twins Need to Scale to Maximize Value Creation

© 2024 Simio LLC 5

Keys to Success

Agile
Development
Process

High Value
MVP

Build with a Product

Approach

© 2024 Simio LLC 6

Keys to Success

Agile
Development
Process

High Value
MVP

Build with a Product

Approach

© 2024 Simio LLC 7

The Digital Twin Model is a Product
Designed for users

Defined development process

Robust for day-to-day use

Scalable

Proper controls and governance

for updates

© 2024 Simio LLC 8

The Digital Twin Model is a Product
Who is the customer of your
digital twin?

What inputs do they provide?

What is their capability to run

simulation?

What outputs do they want to

see and what is their success
criteria?

© 2024 Simio LLC 9

The Simio Platform Enables the Digital Twin as a Product

App in
Developer App App User
App Store

© 2024 Simio LLC 10

Keys to Success

Agile
Development
Process

High Value
MVP

Build with a Product

Approach

© 2024 Simio LLC 11

Proven Scalable Archetypes

Process Factory Warehouse Deployment

© 2024 Simio LLC 12

Proven Scalable Archetypes

Process Factory Warehouse Deployment

Process Simulation
Archetype description
A single process or site (inside the 4 walls)
Can be manufacturing or business process that is start to finish
Typical questions
How can I improve the system KPI?
What if we change the system, how do KPIs change?
How robust is my system design?

© 2024 Simio LLC 13

Proven Scalable Archetypes

Process Factory Warehouse Deployment

Production Scheduling
Archetype description
A single process or site
Typically manufacturing plant
Key questions
What’s my schedule for tomorrow that reduces cost?
Do I need to add another machine?

© 2024 Simio LLC 14

Proven Scalable Archetypes

Process Factory Warehouse Deployment

Warehouse Management
Archetype description
A single warehouse
Shelving, bins, racks with pick and ship operations
Key questions
Can we improve KPIs by changing picking strategies?
Can we improve KPIs by adding or changing AGVs, Fork Trucks,
workers?

© 2024 Simio LLC 15

Proven Scalable Archetypes

Process Factory Warehouse Deployment

Deployment
Archetype description
Production sites and warehouse with multiple lanes and choices
to move product
Key questions
How can I meet service targets at the lowest cost?
What if I changed my sourcing strategy, would service change?

© 2024 Simio LLC 16

Digital Twin + Data = Workstream

Historical or Manually Created Data

Process Capacity Delivery &

Time to Pick
Design Planning Cost

Real Time Data

Process Site Warehouse Deployment

Scheduling Scheduling Scheduling Scheduling

© 2024 Simio LLC 17

Combine for a Network Level Digital Twin

Historical or Manually Created Data

Network Resiliency Network Capacity

Design Planning Expansion Planning

Real Time Data

Network Network Trade-off

Orchestration
Scheduling Balancing Analysis

© 2024 Simio LLC 18

Determining the best MVP Archetype + Workstream
Problem needs to be complex,
Complexity business struggling to solve

System must be known and able to

Known System be modeled with Discrete Event
Simulation (DES)

Data Data must be readily available &

understood

ROI ROI method must be known, target

defined to pay back investment

© 2024 Simio LLC 19

Getting The Right Team & Leadership Buy-in
CIO
o Data Engineering
o Data Science
o Cloud Engineering
o Project Management

COO
o Operations Engineering
o Industrial Engineering
o Planners / Supply Chain Leaders

CFO
o Financial Analyst
o S&OP Leader

© 2024 Simio LLC 20

Keys to Success

Agile
Development
Process

High Value
MVP

Build with a Product

Approach

© 2024 Simio LLC 21

Digital Twin Development Journey
10 to 26 weeks depending on scope & cadence

FRS Validation Model Enhancements

•Understand all requirements & •Stakeholders validate inputs & •Validation feedback incorporated
desired outcome outputs

Initial Data Refined Model Ready for Production

•Flat files of all needed data to •Begin running model with new •Model ready for day-to-day
model data sets planning

Draft Model with Initial Refined Data

Data •Build data pipeline
•Remove assumptions from POC •Cleaning data upstream
model
•Expand model to match needed
fidelity

2-4 weeks 4-10 weeks 4-12 weeks

© 2024 Simio LLC 22
Functional Requirement Specification Process
FRS is a structured method to set the
project up for success

Determine all the inputs

o Data

Scope the System

o Business Rules

Determine the outcome

o Verification & Validation methods
o Outputs, ROI
o Workflow

Setup the project management

o Alignment on resources, timeline, cost

© 2024 Simio LLC 23

Example of FRS Process

© 2024 Simio LLC 24

Digital Twin Product Development Process

Data Model Model Production Ongoing

Development Development Validation Workflow Maintenance

© 2024 Simio LLC 25

Data Science & Engineering
Data will make or break your
digital twin

Dev Data Pipeline

o Needs to be agile
o Ability to contextualize the data

Prod Data Pipeline

o Path needs to start early as it takes
a long time.
o Consider the broader impact of
Data Model Model Production Ongoing
the data pipeline when
Development Development Validation Workflow Maintenance considering ROI

© 2024 Simio LLC 26

Data Validation & Verification in Parallel
Data Validation Data verification
o Does the data match reality? o Once the data is as expected,
• Equipment run rates ensure new data sets maintain all
• Packaging sizes data rules
• Crew sizes
o May be necessary to create error
o Do you understand the data? logs and make transformations of
• If the data is coded, do you the data to keep the model
understand the codes and how to running but allow reporting of
use them data issues
• How to handle ‘duplicates’

© 2024 Simio LLC 27

Model Building Process
Agile / Sprint mentality

Think about the digital twin as a product

Make model design choices that create

scalability

Establish milestones

Establish RACI

Find the right level of model fidelity

o Complex enough to get the right answer
Data
Development
Model
Development
Model
Validation
Production
Workflow
Ongoing
Maintenance
o Simple enough to be understood

© 2024 Simio LLC 28

Agile Approach to Model Building
Utilize Software development tools
o Azure Dev Ops
o Microsoft Planner
o JIRA

Plan Sprints
o What’s achievable in next 2 weeks?
o What is needed from different shareholders
o Plan for release

© 2024 Simio LLC 29

Model Testing & QA
Align on measurements up front

Create dashboards that closely mimic

the reports and visualization used
today

Find creative ways to build trust in

digital twin results

Establish risks of moving forward vs

risks of status quo

Keep momentum, don’t overengineer

Data
Development
Model
Development
Model
Validation
Production
Workflow
Ongoing
Maintenance
or try to perfect too early

© 2024 Simio LLC 30

Visualization Development
Create dashboards for input data

Match dashboards to what users

view today.

Utilize Simio Dashboards on

Portal

Export data for use with Power

BI, Tableau, Looker

© 2024 Simio LLC 31

Model Verification & Validation Best Practices
Upfront targets, goals

Create conceptual models to

ensure logic works as expected

Compare with reality using

historical data and actual reports

© 2024 Simio LLC 32

Workflow Design & Automation
Understand your use case

What can be automated on

model runs and data creation?

Need flexibility for ad-hoc

Python, Node Red, other tools

can interact by API with Simio
Data
Development
Model
Development
Model
Validation
Production
Workflow
Ongoing
Maintenance
Portal

© 2024 Simio LLC 33

Maintenance and Governance
Version control for models &
data to ensure accurate results

Documentation & release notes

Design a process that will keep

the model performing and
upgrading to user’s needs
o Simio Portal for Model Control
o DBT / Github for Data Control
Data Model Model Production Ongoing
Development Development Validation Workflow Maintenance

© 2024 Simio LLC 34

Keys to Success
Business Expertise & Availability
o There are many facets to data both technical and business that must align for the project to be successful
o The right people with capacity & alignment drive speed in development.
o Good team chemistry is a catalyst.

Data Availability, Understanding, & Agility

o Simio needs access to real-time, clean data to create a data-based model
o All projects have a level of data cleaning and the ability to make changes to the data pipeline & master data drive
the timeline

Clear alignment on project scope & deliverables

o Different stakeholders may have different needs. Alignment and consistency is key.
o Scope creep happens as the team gets excited about what the model can accomplish. Having a good check on
needed fidelity and a backlog helps.

© 2024 Simio LLC 35

Think of Digital Twin Model as a Product
Building a Digital Twin Product is
step change project
Agile
Development
Process
Getting the right project,
sponsor, plan – MVP!
High Value
MVP

Build with a Product

Follow the Process
Approach

Scale & Grow!

© 2024 Simio LLC 36