Getting digital twin to enhance your assets
A user-focused and value-driven approach
Dr Jingfeng Xu
22/03/2023 – Australian Smart Water Utilities 2023

What I am going to cover?
• What is a digital twin?
• The digital twin methodology (not a technology)
• Integrating digital twin - user-focused and value-driven
• Exploration of Arup studies for flood forecasting, water quality, and
sewers

Digital Twin
Methodology not Technology

Methodology not Technology
• ‘Digital twin’ means many things
– Sometimes a misunderstood concept
– High on hype cycle (everything is a “digital
twin”)
• We know that it is a methodology, not a
technology
– A way of working, underpinned by full-
lifecycle information & data management
• We focus on the use cases & incremental value
of adopting and implementing a digital twin
approach
Image courtesy of the cdbb

A digital twin is a digital replica of
a physical thing.
The feedback loop of physical to
digital and digital to physical is
what makes a twin a twin, rather
than just a model or system.
Digital twins offer unparalleled
levels of insight into everything
from energy use, asset performance
to imminent maintenance
requirements.
What is a digital twin?
Digital twins
Data
Right-time,
not real time
Insights &
Interventions
Physical
Asset
Digital
Asset

What is a digital twin?
Water Treatment Filter Optimisation Case Study
Data
Water Quality
Environment Data
Maintenance Schedule
Insights & Interventions
Dashboarding & Informed decision making
Filter cleaning optimised
Achieve targets for carbon emissions
reduction
Physical Asset
Water Treatment Filter
Requires washing to
maintain water quality
Digital Asset
Digital twin is trained on
historic performance to
optimise operations
Benefits
Improved water quality
Reduced energy use
Reduced water usage

Digital twins can be used to solve
many different problems, the key is
that each digital twin is specific to its
use case and is developed according
to the available data.
It is about the process and
methodology, underpinned by full-
lifecycle information & data
management. That's why we refer to
it as a "digital twin approach".
The aim is that a digital twin is about
using a constellation of technologies
that work together and connect to
create a ‘single version of truth’ of the
system.
Why twin at all?
M AN AG IN G ASS ET S
Asset Control
Asset Performance
Asset Optimisation
ASS U R A N C E
Resilience
Safety
Regulatory Compliance
ST RAT EGY &
PLAN NI NG
Investment Planning
Scenario Planning
Twin Integration

Predictive
maintenance and
reduced maintenance
down time
Improved
regulatory
compliance
Planning
future
investments
and improving
the design of
new assets
Increased asset
flexibility and
optimisation
Improved scenario
planning and running
‘what if’ simulations
and scenarios
Provide valuable
insights on existing
asset performance
Improved user
experience,
increased
productivity and
usage
Benefits of twinning
Reduced energy
consumption and
carbon production

Our Water Sector Examples
Client-driven, purpose-built solutions
Sewer level
prediction
Our value proposition
A Digital Twin enabled platform combined with industry leading
advisory services.
We help our clients solve complex problems with their data, by
combining AI/ML tools with deep domain expertise.
River level
prediction
Water quality
prediction
Water treatment works
optimisation

Digital Twin Approach
Arup’s framework
MANAGE &
MAINTAIN TWIN
CREATE
BASELINE
DIGITAL TWIN
DEVELOP
USE CASE
• Help our clients
understand if asset
digitalisation is right
for them
• Determine use cases
& cost/benefit
analysis
• Capture asset (if
needed)
• Create a baseline
model
• Connect to relevant
persistent information
& BIM metadata
• Handover and help our
clients use their twin &
realise value / benefits
• Ongoing ‘as a service’
• Manage and maintain,
keeping the twin
current & value-adding
• Analyse and assess
• Provide engineering
insights and services
• Creating the twin road
map, including
• Systems architecture,
Data model
• Security Model
• Partnership Approach
• Bringing on “other
twins”
CREATE
ROAD MAP

Digital Twin Approach
Arup’s framework
MANAGE &
MAINTAIN TWIN
CREATE
BASELINE
DIGITAL TWIN
DEVELOP
USE CASE
• Help our clients
understand if asset
digitalisation is right
for them
• Determine use cases
& cost/benefit
analysis
• Capture asset (if
needed)
• Create a baseline
model
• Connect to relevant
persistent information
& BIM metadata
• Handover and help our
clients use their twin &
realise value / benefits
• Ongoing ‘as a service’
• Manage and maintain,
keeping the twin
current & value-adding
• Analyse and assess
• Provide engineering
insights and services
• Creating the twin road
map, including
• Systems architecture,
Data model
• Security Model
• Partnership Approach
• Bringing on “other
twins”
CREATE
ROAD MAP
No technology!

So how do we get started?
Integrating digital twin –
user-focused and value-driven

Start with the Users
We have discussed that people and
process are the core element
Think about how they are going to
succeed in achieving the outcomes
So start with the users, processes and
outcomes
What are their needs and objectives,
how do they work together?

An Example:
Next Generation Flood
Resilience
Exploring resilient flood forecasting

Project Vision
To enable greater flood resilience in our six communities by delivering
technologically innovative and enhanced flood warning systems
capable of providing a specific and timely warnings in mid-sized & fast
response catchments.
From our engagement responses, resilience is…
'Being aware that flooding is a possibility; being prepared in case of flooding; knowing when and where
flooding is likely; responding quickly and appropriately to mitigate the worst effects when flooding happens.’
'Good honest communication with at risk areas.'

The Problem
– Warning is required to prepare for flood
events
– Industry has empirically modelled the
‘mechanics’of water
– Over time models have become more
computationally expensive
– Responders use catchment experience
Northumberland
during
Storm
Desmond
(2015)
Source:
Hexham
Courant

Our Objectives
1. Provide means of flood warning
2. To enable communities to make informed decisions
3. Create a prototype for national implementation
Objectives

Interviews
“People receive a Yellow
Met office alert and simply
to not know what it
means”
Community Flood Warden
“There is a lack of awareness
and understanding from local
community members, of how
flooding actually works, which
can put them at risk”
EA Field Team Member
NCC Response Coordinator
“There is an issue of
knowledge retention as
there is a high level of staff
turnover for response
coordinators”

Personas
Flood Wardens
Local Community
Members
Response
Coordinator
Lead Responder
Responder
Flood Response
Community Liaison
Flood Warning
Duty Officer

= depth x
velocity x
debris risk
= flood
routing
= flood
extent
Flood Wardens
Local Community
Members
= when
and
duration
Technical
Definitions
Who we asked about
the type of service:
Exploring solutions

Solution Blueprint
72 to 48 hours 24 to 12 hours 8 hours 4 to 2 hours Flooding event
Flooding Threshold
Goal of warning
Raise awareness about a
specific location
Who for
Local Community
Members
Responders
Flood Wardens
Goal of warning
Understand likelihood of
impact
Who for
Goal of warning
Understand what actions to
take
Who for
Goal of warning
Near real time forecast to
take actions
Who for
Responders
Flood Wardens Local Community
Members
Responders
Flood Wardens
Local Community
Members
Responders
Flood Wardens

Next phase: Alpha
Two clear assumptions to test:
1. Forecasting Main Assumption: It is possible to forecast flash flooding in the time
and at the level of detail necessary to inform local response.
2. Service Main Assumption: It is possible to use forecast information to issue a
warning that local communities can understand/act upon.

So how do we approach different use cases?
Incremental design approach,
Focus on high impact high value use cases

The power of a discovery phase
Design the Right Twin
Research:
• The users and what they’re trying
to achieve
• Any constraints you face in
creating the service, e.g.
technology, legislation, costs
• Opportunities for improvement
from existing baselines
RESEARCH DESIGN
Discover Define Develop Deliver
Design the right thing Design the thing right
Double Diamond Approach
UK GDS Agile Framework

Use
case
1
Use
case
2
Use
case
3
Partnership approach
Business processes and architecture
Information processes and architecture
Technical processes and architecture
Operations and maintenance processes
People processes, training and implementation
Security requirements and processes
Journey of adopting a digital twin approach
The first ‘use case’ across the layers will inform
future adjacent slices
Current
use
Desired
future
state
(digital
twin)
Focus on value, Build incrementally
Design the Right Twin
• When considering digital twin use
cases, use a ‘thin slice’ approach
• Develop the right strategy and
implementation plan for delivering
tangible use cases based on value
and impact
• Don’t get too focused on one area

An Example:
Digital Twin for Water Quality
Exploring water treatment

Time
Output
Turbidity
Performance Threshold
Winter
Summer
Example - Rapid Gravity Filter Optimisation
• Arup were asked to deploy AI with at
an existing works
• Arup have developed an AI that can
optimise Rapid Gravity Filter
performance
• AI can optimise backwashing to
reduce energy and prolong the life of
the filter
• Able to optimise for throughput or
for energy reduction based on
user/site needs

Lets step back – explore the problem space
• RGF optimisation assumed the
correct solution
• However what are the site
operator / users telling us?
Initial
Solution

Challenges in the wider treatment process
• RGF optimisation assumed the
correct solution
• However what are the site
operator / users telling us?
• Users tell us the big challenge is
spikes in nitrates and algal
blooms – site always on standby
to counter this
Initial
Solution
Revised
Focus
Upstream
Catchment

Out in the catchment – what is driving this?
• Biggest opportunity at this works
could be outside of the works!
• Can we forecast water quality
indicators in the catchment?
• Using Catchment features, AI/ML
provided us a forecast up to 2 days
in advance
• Could this unlock performance on
all parts of the treatment train?
Actual
Predicted
Algal Bloom
1 Week Horizon*
*Results
are
indicative
and
should
not
be
relied
upon
R2 – 0.65

Can we forecast the catchment load?
• Yes this likely the case!
• There is value in stepping back
and using a discovery exercise to
take a wider look
• Talk to your users – understand
the processes – assess their needs
• Next question - how do you phase
your digital twin rollout?
Actual
Predicted
Nitrate
1 Day Horizon*
R2 – 0.95
*Results
are
indicative
and
should
not
be
relied
upon

A Wastewater Example
• Just to show this isn’t
limited to rivers and
water treatment…
• This is the same
principle applied in
sewers
• Have been able to
achieve up to 5 day
forecasts

Closeout & Key Points

Key Messages
Listen to your users, build your
roadmap around them
Your biggest opportunity may
be outside of your focus area
Ultimately your users create the value, not
your digital twin!
The most valuable data to your asset may not
be your data!
Digital Twin Methodology not
Technology
Discover your problem space –
you won’t regret it
Technology is an important enabler, but
putting your focus there wont create value
alone.
Map your use cases, understand who your
users are. What are the outcomes your are
hoping to achieve, how will you unlock them?

How can Arup help
Thought Leadership
Supporting decision
makers to enable
digital twins.
Define & Discover
Helping clients to
discover the potential
for digital twins.
Develop & Deliver
Developing time-series
twin pilots to address
client needs.
Image credits CDBB
Evolve
Maintaining and updating
twins to evolving needs.

Contact
Jingfeng.xu@arup.com
Dr Jingfeng Xu