Last updated
Was this helpful?
Everyone is talking about digital twins.
But I find so many different definitions that I thought I'd use myself as an example and explain how digital twin applications for real-time operations can help organizations improve asset performance and help find those unexpected events that typically increase risk and cost.
Learn how XMPro can help you create Real-Time Digital Twin Applications at http://xmpro.com
SUBSCRIBE to learn more about Digital Twins and Real-Time Operational Intelligence: http://www.youtube.com/user/xmprobpm?sub_confirmation=1
everyone's talking about digital twins
but I find so many different definitions
and different understandings and I
thought I'd use myself as an example and
explain how digital twin applications
for real time operations can really help
organizations improve asset performance
and help find those unexpected events
that typically increase the risk and
cost us money so if you look at a
digital twin for myself so this is me
well this is a photo of me but this is
me and a couple of years ago at
Microsoft event Microsoft expire inspire
event they had a great set up where you
could stand on a rotating disc and they
were a couple of kinetic cameras and as
the disc spun it would take a 3d image
of you which they handed out so now I
have a 3d model of myself which I could
potentially use for a number of
different use cases I could I could
print a 3d model of myself I could
potentially fit some clothes on this 3d
model and see what brands would fit me
or what different colors would look like
but you know is this really a digital
twin of myself and looking at how
digital twins has evolved over the last
couple years I kind of came to the
conclusion that my smartphone is
probably a better representation of a
digital twin or myself then that 3d
model now if you look at the history of
digital twins it typically started off
as 3d or CAD based models where we try
to create a visual representation or
digital representation of a physical
object as as digital twins own evolving
we're starting to see that it's more
than just a 3d model but it actually
encapsulates a lot of the things that I
do it knows a lot about me in a number
of different areas now the smart phone
is a great example of this way there are
a number of different applications on my
phone that knows quite a lot about me so
if you look at the photos my phone knows
where I've been
what people I I've I've been in contact
with my families if it looks at Google
map data it will know where I've
traveled we have been like wise with
Instagram or Skype it'll mark
conversations well hopefully it doesn't
know all the the detail of the
conversation but it's a bit creepy that
when I when I get in my car it even
knows where I'm going so Google is
telling me you know you're heading there
and it'll take you this long to drive
there which again is a little bit creepy
in but in terms of predictive capability
of digital twins it kind of shows the
future of ways some of these things are
going and to be quite honest my phone
knows a lot more about me than anyone
else including my wife and close family
my my my phone knows absolutely
everything you know if we look at a
couple of examples around my my health
for example there are different number
of health apps and depending on what I'm
but what I want to use it for it's got
different sets of data that it that it
collects either real time or when I when
I'm using that specific app likewise
with finances it knows everything about
my investments my banks but my bank
account what I'm spending on my credit
card
the the the stocks that I'm invested in
and in terms of my travel where I've
been so in if you're looking at the
phone as a repository or as a single
front-end it's not a single database
with all the information in each
application has its own data source and
it can it can interchange so if you're
looking at the health apps it can
interchange some of the data so but but
but each app has got it its own data
source we its own way of collecting its
own analytics and it's only use guys
that it applies to but it's all
accessible from one single user
interface and it it's got the best
representation
of of knowing who I'm talking to
socially we have been what my financial
state is and also what my health is so
if I look at health in a little bit more
detail if you go into for example the
Apple health app it'll show you in terms
of your current performance in terms of
where you're in terms of activity and
things that you might have done it will
show you some of your records now some
of these are historical records anything
from and again this is whether it can
come from a sensor it could come from
doctors records there could be different
places where all this data comes from
but this is the single place where I can
see and this is the best digital
representation of myself it knows more
about my health than anyone else in one
single area so whether it is it is heart
health health records reproductive
health there's a broad range of of
different data sources and depending
again on the use case that I'm that I
want to use this health data for so in
my instance I'm a aspiring triathlete so
for example if I use Strava and one of
the things that I do I'd like to try and
based on on keeping my heart right at a
certain level
so that's real-time monitoring and
information that I'm that I need to know
and what does digital twin does he shows
me where I've been it gives me some of
the historical data for a specific
activity so I can look into that but I
can also see that real-time and you know
I can even have it on on a different
user interface like Apple watch where I
can actually see it real-time as I'm
doing the exercise and the other use
case for the same information is that my
family so when I'm at a race my family
can monitor me in real-time without
looking at the physical me if I just
look at
my twin or proxy on the internet so they
don't have to look at me they just look
at this there's avatar that I have on on
on the Internet and it uses the same
sensors that that feed the different you
eyes through the same IP eyes so each
app doesn't have to have its own sensor
data my phone for example the GPS is
used by multiple different applications
for their own use but it's the same GPS
location the same with any one of the
other sensors in the phone the other
interesting thing with this is I can
swap out equipment the underlying sense
the equipment so I can change my Fitbit
to a Garmin or have a new model of a
Garmin or whatever new and lightest fad
there is worth with the sensor based
technology that I use to collect this
data and it doesn't even blip on the app
the app doesn't even know because it's
just collecting the data through a
standard integration user interface and
if I swap out the underlying data source
the application and what I use this for
in this example my heart rate training
it does not even know that I've swapped
out the underlying equipment so looking
at this this is the best representation
of me if you wanted to know if from a
digital twin perspective again it knows
everything about me in terms of my
physical health financial health
my social life who I know what
conversations I'm I mean who I'm
connected to so this is the best digital
representation of myself and if we look
at our factory or plant or mine so how's
this different than just having my my
iPhone or my Android or whatever device
you're using as the as the the place
that serves all the applications well it
isn't really different so if
look at how we see it from my XM Pro
perspective we have all these different
apps so whether this is my Google Maps
or Skype or Outlook or whatever it is
that I'm that I'm using all my health
apps so let's say for example the
equipment maintenance that you see there
similar to do what you saw on the health
app if I go one level deeper into
equipment maintenance I can see the
different applications that I can that I
can set up as digital twin
representations so for example I can
create a centrifugal pump condition
monitoring or predictive maintenance
model the first is I two different
applications and they use the same
underlying data source and the way that
we set this up inside XM Pro is here's
my data flow or how I wire up my digital
twin from different data sources so in
this instance it's using OSI soft and
OPC UA and it's going through some logic
some analytics the same as what my phone
is doing when it's predicting where I'm
heading if you look at what Google Maps
are doing or what my time is going to be
if I run or and in this instance we're
predicting a certain equipments going to
file and then we are creating work
orders and sending our recommendation
and putting it on a dashboard for you to
see the interesting thing is I can swap
out OPC or or the OSI data source with a
different data source and it wouldn't
change what I'm seeing at the end what
might digital twin representation is the
thing that I'm getting at the end now
what this could look like again this
could be a dashboard similar to what
you've seen on this driver 1 + X employ
dashboards can have some historical data
some real-time status what is happening
right now coming from that data stream
the historical two ones could be looking
at at historical records similar to what
you've seen on the health app some
recommendations around what I should be
doing
and then current open tasks so looking
at this hydrocracking unit the digital
twin for this hydrocracking unit is
telling me quite a lot of information
around that so I don't have to go out to
the to the physical unit I can just look
at this if I look at a maintenance a
pump where a certain condition occurred
down that flow again a different style
of dashboard so every app can have its
own look and feel in terms of what the
dashboards look like inside exemplar but
I can see also other information coming
from my si B or Maximo or whatever
EAM system what current maintenance
we've got planned what did we do
recently so bringing in some of that
transactional information again like the
health records and we're not replacing
it so this is not one mega uber big
digital twin it actually connects to the
data that sits in the different places
it's not replacing it it's also not
replicating it it is merely showing it
you in the context of the application
that you're trying to use it in in this
instance I want to either on the right
hand side and select a certain action
whether it's grading work orders
requesting space escalating into
engineering to have a look at this and
review it all the different use cases
and yes I can have the 3d models
embedded as part of that as part of this
decision support in this digital twin so
as you can see there's no difference
between how I use my smart phone as my
digital twin representation there is no
difference in terms of how you can set
up digital twin applications for
real-time operations monitoring and
operations management
now just to explain a little bit in
terms of at what level does a digital
twin sit well it can actually sit at any
level it really depends on what your use
case is so it might be that I've got
that I specifically want to look at this
gear box on the Bommel and that might be
a simple twin so I've got a simple
dashboard simple monitoring
or it might be a composite twin which is
assembly of a number of different
individual twin components and this
creates my ball more unit so this is for
one ball more all my use case might be I
need to look at the total throughput
throughput in my whole processing plant
and for that this ball mill is just one
of the components so again this digital
twin can fit into a broader or bigger
application which we call kind of a
system based composite and digital twin
and then just in terms of the proxy as I
said when I was running the rice my
family didn't have to look at me
physically they could look at my proxy
which is really that digital twin and I
could see where I am and what my
performance is life without a digital
twin if for all the different use cases
that you have operations maintenance
safety and like a whole number of
different ones quite often you end up
having different or senses that
duplicate or replicate the same
information and every kind of use case
has got its own sensors on operations
have their own sensors and information
that they're collecting and it creates
these silos of different information the
benefit is if you're using this digital
twin approach now I just think of myself
with my phone in the middle and the
different use cases that I have like
health finance and whatever the case
might be where the phone rate with all
these little blocks represent the
different applications on my phone
likewise on a plant if I've got a well
platform and I use the concept of a
digital twin which can contain physics
based models analytics models so the
physics based models classically where a
lot of these things thought it with
finite element methods and some of the
engineering things like thermodynamics
analytical model some of the newer
things that we're seeing around
predictive capability all the time
series data and historical data that we
are collecting around time time values
and then is a transactional data sitting
in your piece around maintenance master
data around what is the equipment when
did we buy it and visual models like
like the CAD and 3d and and some of the
others so as you can see there is no
difference between myself and my iPhone
versus your plant and its digital twin
and it's in different use cases
