How Successful Digital Mining Operations Transform

7 seconds ago - UPDATED. Hello, ever wondered what you should do in your organization to go from traditional to more modern processes? Not all mineral and metals industry leaders operation the same way. We know this. Some of the ones who acquired a greater success rate than others have approached digital transformation as if it was a mission. Here we are going to learn how to set new KPI's that make sense.

What is Digital Mining?

 I'm here to talk about the success in

mining digital transformation I don't

know if a lot of you know about uptake

but just a short background uptake is

five years old it's a company based in

Chicago and we do predictive analytics

across different industrial companies so

we take data coming from assets single

data we also take work order data.

Note: Get the outlook 2020 pdf here

Any other sort of contextual data so that

can be weather data we bring it all

together and we develop data science

models those models their goal is to

flag anomalies ahead of time or predict

failures before they happen so it's all

about making a more proactive

maintenance towards your assets and

being less reactive so one of the things

that I'm going to talk today is for us

success is when the insights we generate

so those outputs from the data science

models actually translate to value so

that there's someone on the ground

taking action on it so that's for us

success when you actually generate that

value and I only have ten minutes but

I'm gonna show a quick demo on how we

immerse ourselves in that workflow of

the day-to-day of an operator of

reliability engineer of a condition

monitoring analyst and how we really

make our insights generate the impact

that we're after.

Digital Transformation Starts with your Inspections: in 2020, Every Leader Ditch's the Paper Form.

So, solving those

problems that generate downtime and at

the end what we want to do is make sure

the machines are up and running that you

know what's gonna happen in a couple

months

you can plan ahead of time so instead of

just running around putting out fires

getting all this million alerts you know

exactly what you need to be doing so for

this I'll jump into the demo and there

are basically three things that I want

to highlight one is we want to make it

very easy for the final user to

understand what they have to focus on

second is when we provide the insight we

want them to understand what is the

contextual information what does these

things I tell me I may convince

by this insight and three how quickly

can I take action on it so if I jump

into the demo this is just a

configuration or for after your

application this is one way we surface

the insights to the final user another

way could be if your team is already

using another tool.

The future of electronic forms is here for all serious operators.

We can send the insights there but this is just one way

that we present those so the first thing

is how can we direct the focus towards

what you have to focus on that day we

work with a lot of customers they have

all these high-low rules they get

literally 500 emails a day about a pump

it's very difficult for them to sort

through the noise understand what's

going on and be really proactive around

the essence so the first thing is not

only can see all the information on

what's going on what's relevant for me

that what they have to focus on that day

and in this case when you see a map or

it can even be a list some heat

highlights and gets your attention and

that's like the red block there so

there's something going on on the

concentrator plan so me as a user and

imagine I'm for example a condition

monitoring analyst I know that today I'm

going to look into this particular area

because there are some insights in this

area that really require my attention

when I go into the concentrator area I

would see all the assets that ab c-- is

monitoring and i quickly see what is a

prison in this case this example is a

segment so if i go deeper i can see all

the information surrounding the asset

and app type isn't just about this

information right signal data work

orders alert

so we displayed in a way that the user

really understands what's happening so

the

you see here I just the things that I

look at in a databases can be something

different and this that's computable

right is a logic in the back we're

getting the data already so this is just

for me as a user to understand

everything about the Atman how does

import working in the last seven days or

maybe an interesting that's 24 hours and

we would still display whatever I wanted

to see and understand it seemed down

it's been running and in the case of the

segment that we develop a year ago it

was a lot around the liner world so a

customer in particular was stopping the

sack mill every three weeks or months at

the beginning for five hours making this

section measuring the liner with and

then they were doing this progression

line to understand what is it gonna wear

out and while the time comes at the end

of the liner useful life they would do

the inspections more often and more

often or more often and they were

getting a lot of data from the asset so

the problem we wanted to solve was can

we use the data that's already there to

remove those inspections instead of

having to do

ten inspections to do over the life of

the liner now we just alert you at a

point in time where you should start

doing inspections so you remove some of

that downtime at the beginning so in

this case and the user looking at the

stack mill and the inside is really what

comes from the data science models the

insights are the outputs of the models

that we've developed some of them are

focused on being predictive in nature

like we're predicting a failure in this

case where we fly us an anomaly in the

liner world so as the user I want to

understand what the insight means this

is the second part really making it

consumable for the user so it's not only

about saying oh detective the liner

started to wear out it's understanding

really what surrounding the same time so

by working with a lot of customers at

the beginning we just provided the title

but then we realized they were dead in

action

so we asked what do you need to make

sure to be convinced

so some test things were things like

sing applause so what are the most

relevant signals that the model is

looking at and what do I need to see in

order to be convinced and we are

is happening and that's the beauty of AI

at the background so that's able to tell

that there's an academy

however they users still wanted to see

some single bond so we're okay but we

give you that true

they are mostly you contextual

information so what are the others that

you have been in the past

recent events corporate data so this is

a way for me as a user I don't have to

go into a different system to take what

happened recently I don't have to call

my 9:15 to understand what happened

yesterday I have everything in one place

and now that I've concealed inside the

thirdly how do I take action and there

are different things that can happen

right I mean

I feel it through and be like you know

what we change the patterns like

recently and these brought me the

pressure something that I would expect

so I'm not going to send anyone to

inspect or under gonna plan anything

right now so I'm going to disregard

inside so through the application you

can disregard and inside when you do

that it gets fed back into the platform

so the model is machine learning it

learns that something was not right I'm

not predicting something that is

actionable yet so it knows why that

happens on the other hand okay this is

something that we've seen in the past

I played me saying that normally on

average these and nobody starts

triggering in two months ahead of the

end of the liner word useful life so I'm

gonna do something I'm going to plan an

inspection maybe

well wait two weeks whenever someone

takes an action you can create a test

that tests can be sent to my work order

systems introducing ASAP IBM maximum can

be sent there so it's a really easy way

to translate

into an actual action and someone from

my team is going to do something about

it and also since we're reading the word

partners we're getting the feedback so

whenever someone goes out maybe some

third-party company but they turn the

wrench they do something about it they

expected they see that the line areas

were or not that gets feedback games or

assistance so the model says many more

and more yes this happened I guess

someone took action and start saying

what actions did they take so in this

case maybe put just a new section in

some other case maybe they went to

Greece the bearing and that resolved the

issues so it learns that was doesn't

that they did in the past so next time

I'm gonna recommend that they start

doing that so it's all about also

keeping tribal knowledge in house not

only in the minds of the of the users

and I'm out of time that mainly what I

wanted you to get away with is that

success for us crafting is really making

sure that the insights are translated

into actions

waverer users and you're in a booth in

front the Grand Ballroom if you want to

talk more about it