Digital Transformation in the Aluminum Industry: Insights and Strategies
E4

Digital Transformation in the Aluminum Industry: Insights and Strategies

Luke (2): Welcome back to
the Smart Metals podcast.

Today, Denis Gontcharov and I, Luke van
Enkhuizen will be talking about Denis,

his background , and we're providing
you with a starter pack for digital

transformation in the aluminum industry.

Denis: Hi Luke, glad to be here again.

Luke (2): yes, yes, it's a bit of a
different episode where we're just

trying to see both of our stories and
backgrounds also to provide you with some

insights of why we're doing this and what
motivates us to, to share this with you.

So Denis, could you please start
by telling us a little bit about

your background and experience
in the field of aluminum?

Denis: Yeah, sure.

Let me just mention that the
timing could not have been better.

In fact, next week, that means on
the 22nd, 23rd of May, we have the

conference of the future aluminium forum.

where we will discuss industry 4.

0 and digital transformation
in the aluminium industry.

Right?

So my background, in fact, my background
can very shortly be summarized in just

two words, which are data and aluminium.

I began my career working as a process
engineer in aluminium smelting for

about two years, and then I switched
to a more data focused role as a data

engineer at various companies such as
Novelis, which is the world leader of

rolling and recycling of aluminium.

But I've also done quite a bit of
freelance work mostly as a contractor.

And at this present, I'm a data
consultant who helps aluminium

manufacturers break down data silos.

Luke (2): Nice.

Well, yeah, it sounds really like we're
hitting it right in the moment here.

So many things to cover here.

So first of all, maybe you can
describe how you've actually grown

an interest for data science and
aluminum before you reached the

point where you're at right now.

Denis: Yeah, that's a great question.

In fact, both my education and my first
job experience coming from my internship

were all focused on materials science,
which means I really studied the domain

knowledge side of things of aluminium.

So the real production processes,
the physics and the chemistry,

which I still find very interesting.

But via my first.

I really had to work with lots and
lots of data for my internship topic.

And that's how I unexpectedly fell in
love with computers, with coding, with

statistics, and later, data science.

And it is through this path that I
eventually decided, after about two years,

in the industry to refocus my career on
data science by joining a data science

consultancy company in Frankfurt, Germany.

Luke (2): Nice.

Well, yeah, it sounds really like
quite a journey you had there.

All right.

So you've seen quite some Developments
in your industry, as you've been

in various positions, most of them
revolve around digital transformation.

And so in what, in your opinion, are
these first steps that are normally

being taken by companies when it
comes to digital transformation?

Denis: Right.

So the first steps have been very
difficult in this industry because

in my experience, the aluminum
industry compared to, for instance,

the steel industry is really behind
in terms of data infrastructure.

By this, I mean that we have a lot of
legacy equipment, and that's why you will

always start your digital transformation
in a brownfield environment.

That makes it the more challenging.

Luke (2): So the first steps to
solve this, what would that be?

Denis: I think like in every
industry, it really boils down

to beginning with a solid.

digital strategy, which will of
course encompass a modernization

of the data infrastructure.

And as we will see later in the
discussion, this will mean both

upgrading existing systems.

It's no longer okay to hire retired
COBOL programmers just because

you're still using a COBOL as your
mainframe, but then also you will

have to invest into developing
additional digital infrastructure.

Maybe to add a bit to that
in terms of digital strategy,

companies have been doing quite well in
this industry, in my view, is they really

focus on what it brings for the business.

So they really begin with a business
problem or a technological problem

that when solved will lead to either
cost savings or new revenue streams.

So that should always be the cornerstone
of any digital transformation.

Luke (2): And so what are the biggest
challenges of pitfalls that you have

seen or that companies face when they are
making this change in your experience?

Denis: right?

So I'm going to be on
this conference next week.

Right.

And I'm really looking forward
because the last conference I attended

in October, I clearly saw that
companies really are not aware of the

latest technologies and approaches.

So my biggest fear and what I've seen
in the past at clients and previous

employers, was that they had taken a
very naive technology slash IT first

approach to digital transformation,
which leads to them spending a lot

of money on a lot of software that
doesn't generate any business value.

Luke (2): Okay.

And so maybe to make this a bit more
understandable or relatable for those that

don't have the data science background
what kind of things do you mean by that?

What is like it first?

And what are things that are,
that companies are focusing on

that, that do you think that maybe
don't add that much value to it?

Denis: Sure.

Let me give a concrete example.

So, one important step of digital
transformation and building a good

digital infrastructure is to collect
and organize your data in one source.

Lots of companies get this fact, right?

They realize that this is important.

So they go look for solutions and they
stumble on some external contractor or

consultant, usually a very large company
that tells them you need to build a

data warehouse and that's how they get
started and it's the right approach.

But one important key that dismisses
is that you're essentially just

moving existing data, the way it's
currently saved just to a different

location, which means you're not really
solving the problem of data access.

You're just putting all your data in a one
big messy pile somewhere on the internet.

Luke (2): All right.

So I think when, when the company
hears that they should build one

or they actually are being told
us to buy one, maybe even more

like buy this product right now.

So I guess there's a bunch of
software vendors in the industry

that provide, these products.

These solutions that they could choose
and therefore they kind of have more

like a monopoly on, on what's out
there because you relate to that.

That's at least what I
see in many industries.

Denis: Yeah, it's true that especially
the market for more complex analytics

tools such as data warehouses Is really
governed by the big three players, right?

So you have Microsoft Google and
Amazon Offering solutions, but

then you have a lot of let's say
service providers that build these

technologies using well, what's
Google and Amazon and Microsoft offer

I Guess the point I'm trying to make
here is that You Unfortunately, because

digital transformation is so new and
it's so far away from the core business

of producing or recycling aluminum,
that it's very hard to troubleshoot

your strategy and to realize, are you
really doing something right or wrong?

So you have to put a lot of faith in the
partners that go with you on this journey.

And it's very hard to select
these partners, in my opinion.

Luke (2): Why would that be?

Why, why, why is there
such a new thing for them?

Denis: well, I think you make a good
point and perhaps it wasn't that fair

to say that the whole concept is new.

I mean, after all, we've been
talking about industry 4.

0 for the last 20 years,
I think in Germany.

So in that sense, the concept
is definitely familiar.

I think what's by new is
more the, I refer more to the

technologies which are available.

Luke (2): Okay, so the newer
technologies are not known.

They know they should do
things differently, but maybe

not much has changed yet.

Maybe some initiative has been launched,
as I said, IT driven ones, buying a new

software package, ERP system, so forth.

But what you're aiming at here is really
the difference between an OT and an

IT solution and perhaps how to find
more of an overall strategy for this.

Denis: Yeah, that's a fair point, right?

I think we see digital transformation too
much as a technological transformation,

whereas in reality, it's really the
transformation of your entire business.

So really from a business perspective, and
this requires a very holistic approach.

So you will really have to change
the way you currently produce or

the way you interact with other
companies, the way you do business.

You cannot just go ahead and upgrade.

a particular system you have,
like the MES or the ERP system.

And I think too many managers are still
stuck in that view, that well, if we just

upgrade the systems we have, even though
they don't really want that because IT

or OT investments are seen as a cost,
then we will achieve the transformation.

But whereas in reality, it really
begins with the main knowledge.

Luke (2): Yeah.

Okay.

So domain knowledge is crucial
from the aluminum industry.

And but what makes it so unique in
its digital transformation needs, like

other specific environmental or cultural
factors or like legal requirements

and so forth, or other things that
are very distinct to the industry?

Denis: Yeah, I'm happy
you asked this question.

I wouldn't call it necessarily like
roadblocks or blockers or challenges, I

would call them more opportunities for
the aluminum industry because those are

really points that they need to improve
on and to give you a number of examples.

So big topics in the aluminum
industry are sustainability.

Companies need to become more
clean in CO2 emissions, less

fluorine emissions, and so on.

Okay.

But then you also have the whole story
of decarbonization and I'm talking now

about smelter specifically flexibility
where you want to modulate the current.

In essence, one big opportunity
I see specifically in developed

countries is the need to adapt your
power consumption to accommodate

more renewable energy on the grid.

So what am I, what do I mean by this?

We are essentially as a society
in Europe, trying to move away

from fossil fuels to renewable
energies, which is all good and well.

The problem is that , the
sun doesn't always shine and

the wind doesn't always blow.

Or sometimes they do moments that you
don't want, for example, at night.

So you have these excesses and.

shortages of energy that the grid really
doesn't like because the grid has been

designed to operate on a constant power.

So it would be very beneficial if you
have certain big energy consumers having,

for example, a service that allows
power providers to either shut them

off the grid or make them consume more.

In this way, they can absorb these
fluctuations in energy supply.

And the reason why I saw this
example is that If this is hard to do

technologically, but also this is where
I really see that the role of data

and digital can make a big difference.

And I'm talking about forecasting the
demand, but also just making sure that

the changes you do in your process
by Increasing your power consumption

or decreasing your power consumption
are sufficiently accommodated.

Luke (2): Right.

So this is a very distinct thing to know.

The ministry essence and they use
a lot of energy and this is one of

the primary cost drivers for their
production systems and plus they now

have this pressure to actually adapt
for environmental reasons as well.

If I can summarize this correctly and
then, all right, so if that's one of

the Use cases or purposes that you
want to streamline your operations

upon and make a digital project.

I assume that many companies are either
considering this or already launched one

or actively pursuing follow up on this.

Maybe you can maybe touch a little bit
on what kind of obstacles they face,

what kind of problems did I have when
they want to launch a digital project?

What kind of what kind of struggles
did I face generally speaking?

Denis: Right.

Sure.

So beside the first point about having
too much legacy equipment, I mentioned

common other problems really pertain to
culture and people more than technology.

I think as digital transformation
is really a holistic transformation

of your entire business,

The most difficult aspect of it is
the alignment of all the stakeholders.

You have to not only align OT
and IT to work together, but also

the business, accounting, and
everyone needs to be on board.

And this coordination exercise has
been proven to be rather difficult.

Luke (2): Okay, it's very difficult,
but what are the symptoms?

Like, how do you
recognize what's going on?

Denis: I think you will feel this
when you do your first, When you start

implementing your digital transformation
strategy, you're working on your first

project and very often we see the
first symptoms occur when you have

to go and get the data for something.

Say you want to train a machine learning
model, or you want to build some kind of

dashboard and you want to combine data.

Usually people budget about 20 percent
for this activity in the entire project,

but in practice very often these small
parts Balloons into 80 percent of the

allocated time for the entire project
because it's just so difficult to

get the data from the old systems.

So if you are consistently doing
projects that go over time and

over budget, you really need to ask
yourself whether you have your digital

strategy and more specifically, your
digital infrastructure straight.

That's the first symptom.

The second symptom I would mention is
that if you are past this phase and

you have indeed one or let's say two
successful use cases, applications

for example, that are running in one
plant, but now you want to expand these

solutions to the other plants, which
should technically be trivial because

you are copying an existing solution.

Well, very often companies get stuck
in so called pilot purgatory where they

are having a lot of use cases or pilot
projects that they just cannot scale or

cannot roll out over different plants
because again the digital infrastructure

is not uniform across the plants.

So what was built for one plant has
essentially be to be rebuilt in order

for it to work in a second plant.

Luke (2): Oh, that sounds like
a whole lot of work indeed.

And also frustrating indeed, if things
are getting over time and budget.

So this is actually a company that
actually blocks companies from

pursuing further projects as they
haven't really found a solution for

this thing spiraling out of control.

Or maybe they haven't found
the right technology with it.

They may be only interested in that.

Get offered, like I said, the data
warehouse that would require a whole

lot of consulting and perhaps still
not guarantee a solution that would

work for your entire enterprise.

Denis: Oh, absolutely.

And you mentioned the word
frustration and that that's actually

a very important symptom as well.

Employees that work at the
company just get burned out with

all the digital transformation.

I mean, there are plenty of engineers
in OT who just can't stand the word

digital transformation anymore because
they're just so tired of IT or someone

else, especially external people, coming
with this new idea that will, at the

end of the day, just mean extra work
for them, especially if those projects

get dragged over, over the time limit.

So I think companies really have
to be careful when they do digital

transformation that they do it smartly
without just creating more meaningless

work for their existing employees

Luke (2): Yeah, I imagine this definitely.

And of course, if you just keep it
vague, which is digital transformation

without talking about specific benefits
you want to achieve and drive it from

the practical needs, it's going to be
very hard to keep your team motivated.

And then I can imagine if they have all
these people pushing solutions to them.

You know, it might be really getting a
bit tiresome if you talk to the fifth

vendor in the week, all having their
promising cast in disguise, you know.

Denis: Yeah, absolutely.

I've had employees recently.

We refuse to work with external
vendors that just say, look,

I'm not going to do that.

But you also have reports from McKinsey.

I was reading today that whenever
your digital project has more

than 50 percent external workers,
That you have to be really careful

because the odds of that project
failing are increasing drastically.

Luke (2): Yeah, because then for
who am I building the project for

right like we're gonna have decision
makers about a project That all

don't work in the company themselves.

Like how Is that's kind of like a
pattern you see right now that's I

mean like maybe a bit of a tangent
right but isn't this whole story going

on about boeing for example right now
in the us where also there's a such

disconnect between the engineers and the
people that actually produce parts and

They don't talk to each other and the
management buys solutions and forces

plans upon the operations it's of
course they also use a lot of aluminum

though, you know coincidentally

But also I can imagine this is you know It
shows a little bit of the miscommunication

that could happen quickly and how this
can spy out of control, but also their

decision makers Are being brought in
on a management level that have nothing

to do with the daily operations And
thus it's going to be very hard to

find a suitable need for your business

Denis: Yeah, absolutely.

Maybe

Luke (2): Yeah

Denis: one more point I can throw
in is that the way companies

became very large, right?

Those big international enterprises
is mostly if you had takeovers.

So mergers and acquisitions,
it means that you essentially

have one company, one umbrella.

It owns a lot of plants that
are on a local level are

still very different, right?

So then again, you have to
push one uniform solution

down to a number of plants.

I think we talked about this in a
different episode that are quite

different themselves, which just makes
the whole process more difficult.

Luke (2): Yeah, exactly.

Well, maybe this is actually a
good conversation starter for

the second part of this episode
and where we can talk a bit about

the vision, how it could be done.

So we titled this episode
the starter pack for digital

transformation in the linear industry.

And now let's imagine
we have a couple sites.

That are all producing and they are
somewhat connected to each other

either to a group or a conglomerate
or whatever They might be connected in

some way the other and then All right.

So What do they do if
they want to get started?

What will be in your
experience the first steps?

And what do what should I have?

What should I strive for?

Yes

Denis: So maybe let's narrow down
the discussion on the assumption

that they already have a strategy.

The strategy is sound.

It involves creating a
digital infrastructure.

That's usually the first point.

So the question is really, how do you
create an infrastructure that will

support your digital transformation
across the entire enterprise, right?

Because what you don't want is.

In one company, having multiple teams
working on different solutions for

the digital infrastructure, because
you already know that you will have to

throw away everyone, but the last one.

Or the one that you select and to make
this episode more concrete, I think

we can already get started narrowly on
the unified namespace, which is just

a concept that I keep pitching as much
as I can, because I really believe

that this is the architecture that will
finally enable us to digitally transform.

Luke (2): Okay, great.

So before we dive into that part,
which I fully agree with you

mentioned they have a strategy.

If they have a strategy, they know
what is important to them, but I want

to initiate a project, as I said in
the beginning, a digital project.

So a project gets a title,
it gets you know, a purpose.

Why are we doing the projects?

But what is your, your example, the
outcome of a successful project, or

how are we, what are we dreaming about?

Like, what would you want to achieve?

Denis: Oh, that's a great question.

And I'm glad you asked it.

So.

At the end of the day, I believe
it's very hard to transform

the entire business overnight.

Or better said, it's hard to convince
the board of directors to say,

Hey, let's transform our business.

It's just too risky.

The first question they will ask is,
well, what will this bring financially?

Or will this work?

Or what's the benefit?

So I think what you really have
to be able to do is you have to

be able to demonstrate value.

And by value, I really mean dollars.

So you need to prove that the use
cases you implement lead to concrete

cost savings or to more revenue.

So speaking of the first use case
that you would start to work on the

first project, as you said, I think
the outcome of this project should

be a very clear financial estimate
of the cost savings or the additional

revenue that this project led to.

Yeah,

Luke (2): important.

So that's what this is what we start with.

And then we have to find a
solution to that out for it.

So again, going back to the starter pack.

So we have all these components here now.

So will be the next component in the list.

You want to dive into you and s

Denis: let's do that.

And that's in fact a good point to start
with because unfortunately the UNS is

the first project that you implement that
has no real financial benefit, right?

Or no direct benefit.

Benefits at the end of the day.

I always compare it to
having good plumbing, right?

No one really likes to
spend money on plumbers.

It's a super expensive But when your
toilet doesn't flush anymore or when

your washing machine is broken Then you
suddenly realize the value of having good

plumbing in my experience and I don't
know how your experience was, but it's

very hard to sell the unified namespace.

Just as an improvement on your
data infrastructure Boards and

decision makers need to see more.

They really need to see what this unified
namespace will bring which is why I always

try to focus on a use case and Kind of
squeezing the unified namespace Into

that use case as just something that has
to be done to accomplish the use case

Does that make sense?

Luke (2): Yeah, definitely.

It makes a lot of sense because you
do need to build like the foundation

first before you built upon something.

It's interesting.

You mentioned the plumbing example,
because you know, something

doesn't flush, that's one thing,
but what if a pipe breaks, then

we have a real big problem, right?

So I think this is also.

Good to mention.

'cause if a system breaks in operation
or if a SY data flow is not correct, or

you have some data loss or there is the
malfunction in and or very important

analytics tools of your is not working
consistently and you cannot debug it

because it's, it's just spaghetti.

It can be quite frustrating
to have those problems.

Right.

So

Denis: Yeah, absolutely.

Let me give you a different
example, a different metaphor

that can perhaps make more sense.

Imagine you are looking for a website.

And I'm telling you, well,
you need a better backend.

You just, you don't care about it.

You just want your clients to see
your website so they can buy stuff

from your website or find you.

You don't care about the back end or how
good it is, even though it's important.

Whereas what you see, the actual
website is called the front end.

It's the same thing with
the unified namespace.

At the end of the day, the
unified namespace is really just

a back end for your website.

Digital data, right?

But you cannot sell a backend
separately from your website.

So you have to package it nicely into
a nice front end that looks beautiful

and then sell the backend with it.

And that's what I mean by having
both the use case and the unified

namespace to power the use case.

Luke (2): Yeah, because the,
the owner will see that they

want to see the plant's data.

They want to see dashboards.

Probably they want to see certain
signals being sent, triggers, events

being notified when something happens.

I think that's the kind
of thing they want to see.

So when should I do that?

Consume more energy.

Are we using the energy as we expected
it to do, as you mentioned earlier?

So that's what they want.

And then you have to make a transition,
I guess, from what do you want and how

do we do this correctly with the biggest
chance of succeeding and not thinking

only about that one thing, but also
making sure you have a foundation for

the rest of the ongoing projects after
that, instead of having to do everything

again for each individual project.

And I feel that part is actually
the hardest with any Project in my

experience to making, to thinking a
little bit ahead and not just taking

off one project and calling it a day.

Denis: Exactly.

That hits the nail on the head.

That's exactly the problem.

Luke (2): Yeah.

Okay.

So, but we promised to make
a starting point on this one.

So what can we say about
this as some, yeah, some tips

to approach this correctly.

Conversations you should be having.

How do you deal with this?

Denis: Right.

So to make it very concrete in my
experience, the development of the use

case and the implementation of the unified
namespace that will support this use

case can be done in just three months.

Right.

And the example I always like to
cite because it's pretty generally

applicable is the calculation of
the overall equipment efficiency.

Imagine you have a machine, it can
be a cold mill that rolls aluminum

ingots into coils or into plates.

The machine has downtime.

The machine makes errors, makes scrapped
metal and machine needs to be maintained.

Right.

Essentially, you want to have a
number that tells you how effective

this machine is operating.

And to calculate this number, you
need some metrics on the machine,

and you need some information about
the orders that is being produced

by this machine coming from the MES.

So that's two systems,
the machine and the MES.

Well, if you integrate the data
from these two systems into your

unified namespace, you can quite
easily design a dashboard or an Excel

document or an email that notifies
you or that reports the OEE metric,

so the overall equipment efficiency.

And that's, in my opinion, is a
very good use case to start with.

Luke (2): Okay.

Excellent.

Yeah, it's this, I think this is
via the stuff that most people want

to know, like what can we achieve
by this and how long will it take?

What do we need by this?

Do we need to buy some expensive software?

What can you tell a bit about this?

There's, of course you
have open source solutions.

I think you work with a
very known one in the space.

Maybe you can talk a
bit more about that one.

Yeah.

Maybe shed some, some light on that as
well to close it down, make it practical.

Okay.

Denis: So let me preface by saying that
Unified Namespace is a philosophy, right?

It's the snapshot of your business.

It's a central repository
of all your data.

in real time in a way
that's clearly understood.

The way you build your unified
namespace is really up to you.

And it's not up to me to like give
you tell you what you need to do.

It really depends on your company.

But there is one company which you
mentioned that I personally really

like, and that's the unified namespace
solution provided by the German company

called United Manufacturing Hub.

So that's United Manufacturing Hub.

And the reason why I like their unified
namespace implementation is that,

because as you mentioned, it's open
source, that means it's free and I'm not

talking about a demo version, I'm talking
about the complete product is free.

You only pay a yearly license fee if
you wish to have their support, but

if you're just starting out, there's
nothing that's stopping you to try to

implement the simple use case I described,
the OEE, using their technology.

Luke (2): All right.

All right.

Sounds great.

So, so this is one side and
then there are other options

out there that could be bought.

Can you maybe describe a little
bit of differences between those

two in, in like an approach?

And why would one be more interesting
than the other for companies?

Of course, if you say this is part
of my consulting, I also understand,

but I'm just like, maybe you get

Denis: Now, I think this is very
interesting knowledge to share.

For instance the platforms I've heard
about the most are those of HiByte,

American company Ignition has in the
inductive automation, and then the, I

think the choice will depend on how big
your enterprise is, like the UMH really

focuses on enterprise customers and
really excels, in my opinion, at the data

handling part, the data infrastructure,
the databases, ignition is often preferred

by OT engineers, and it's very easy
to make interactive visual dashboards.

So you can check out Ignition if
that's one of your requirements.

And I've heard great things about HiByte
that they offer a great platform for

connectivity, having lots of connectors.

So you can check out them if
that's what's your requirement is.

Luke (2): All right.

Thanks.

This is great to hear.

If people want to know a bit more
about the UMH and what you're doing

and your consulting, where can I find
more about what you do, how you do it?

Denis: Well, if you want to find
out more about my services and my

thinking, I would invite you to check
out my website, which is goncharov.

eu.

for European Union.

And there you will have the opportunity to
subscribe to my weekly newsletter as well.

Luke (2): All right.

Excellent.

Do you have any closing words
you want to give to the listeners

before we close down this episode?

Denis: Yes, I think we mentioned a
lot of scary technologies a lot of

potential points of failure, but I think
it's very important to still embark

on your digital transformation journey
because the rewards are just so great.

In fact, if you don't, then
eventually you'll be left behind.

What do you think?

Luke (2): I think you worded it very well.

But also what really stuck with
me was the promise or the vision

that you could have your analytics,
your data, your dashboards, your

UNS in a very short amount of time.

If you work with somebody that comes
from the industry and works with the

Solution that it's like, you know, it's
just, it's quickly to be implement.

It doesn't require a lot of long sales
cycles because it's open source, right?

So it is a whole lot of
benefits for that one.

And the more that I think about that
part, the more I realize this is

actually a promising proposal here.

So I learned something new and
thanks for sharing your story.

It was really inspiring to hear it.

Denis: Yeah, great.

Look I really enjoyed this episode.

Thanks a lot.

Luke (2): See everybody.

Bye bye.

Denis: Bye bye.