Transforming Maintenance with Predictive Analytics · Transforming Maintenance with Predictive...

Formerly SPSS Ireland

Predictive Maintenance

Transforming Maintenance with

Predictive Analytics

Fixed Before It’s

Broken!

Today’s Topics and Speakers

2

Pierre Baviera

CEO for Presidion

Alan McMahon

Senior Analytics Consultant

linkedin.com/in/pierrebaviera

linkedin.com/in/alanpmcmahon

How Predictive Maintenance is a proven

capability that leaders have already

embraced to deliver significant RoI

How does Predictive Maintenance work?

‘I’ll have what she’s having’ – How Predictive Maintenance

borrows from data rich industries

How to succeed with

Predictive Maintenance.

Today’s topics

How to succeed with Predictive

Maintenance.

‘I’ll have what she’s having’ – How Predictive Maintenance

borrows from data rich industries

How does Predictive Maintenance work?

How Predictive Maintenance is a proven

capability that leaders have already embraced

to deliver significant RoI

3

INDUSTRY 4.0

4

Internet of Things ‘Cheap and Easy Sensoring’

Quality and Regulations ‘Ever increasing consumer expectations –

mass personalisation’

Use of Unstructured Data ‘data format is unimportant’

Predictive Maintenance Where Industry 4.0 Leaders are going

Predictive Maintenance The manufacturing world is going through a ‘digital revolution’

Pressure on Margins ‘Sweat the assets’

Industry Research Industries using Predictive Maintenance

Source: Frost & Sullivan 2015

Life Sciences (Pharma, Medical and Bio

Tech)

Automotive and

Aerospace

Pulp and Paper

Power Generation

Mining

Oil and Gas

Food and Beverage

Chemical Others (industrial equipment, hi-

tech and general manufacturing)

IT Spend as a Percent of Vertical Revenue

Big

Data

an

d A

na

lyti

cs

Ma

nu

fac

turi

ng

Op

po

rtu

nit

y

( B

ase

d o

n v

olu

me

an

d a

pp

lica

tions)

0% 1.5% 3.0%

LOW

HIGH

5

Industry Research How are they using Predictive Maintenance applications

Opportunity for Analytics at a Production Line Level

Opportunity for Analytics at a Plant Level

En

terp

ris

e/E

nd

Use

r

Pe

ne

trati

on

Le

ve

l

En

terp

ris

e/E

nd

Use

r

Pe

ne

trati

on

Le

ve

l

LOW

LOW

LOW

LOW

HIGH

HIGH

System Downtime

Asset Performance

Quality Assurance

Asset Utilization

Asset Availability

Resource Allocation

Material Wastage

HIGH

HIGH

Financial Analysis

Demand and Supply Forecasting

Product Lifecycle Management

Enterprise Visibility

Operational Downtime

Inventory Management

Production Line Level

Plant Level

6

Source: Frost & Sullivan 2015

Source: Frost & Sullivan 2015

7

Source: Roland Berger 2014

2

75% Reduction in breakdown

4 More data = More accuracy = More value

1 $18 /hp p.a.

$9 /hp p.a

$13 /hp p.a.

3

15% Time spent on Predictive Maintenance only

85%

15%

Reactive a

nd

Pre

ve

ntive

Predictive Maintenance Transforming of Maintenance Business Model

Government

Predictive

Maintenance

& Quality

Automotive

Travel & Transportation Electronics

Chemical & Petroleum

Mining & Construction

Aerospace & Defense

$75k

80%

$10M 98%

<1%

$1B $100k

5%

Energy & Utilities

Savings in capital plan expenditures

Level of accuracy predicting alarms

Reduction in equipment failure probability

Reduction in scrap rate in 12 weeks

Reduction in annual total warranty costs

Savings per oil platform

Reduction in aircraft-on-ground events

Reduction in fuel costs per turbine

Predictive Maintenance Various industries are realising significant value

8

Today’s topics

How to succeed with Predictive

Maintenance.

‘I’ll have what she’s having’ – How Predictive Maintenance

borrows from data rich industries

How does Predictive Maintenance work?

How Predictive Maintenance is a proven

capability that leaders have already embraced

to deliver significant RoI

9

Presidion A Pentagon of Principles

10

Don’t ignore your ‘gut’, just

make sure to test it

2

Analytics

helps to simplify not to over-complicate

3

Numerous Incremental Projects

= Radical overall benefit

4

Zero the

distance Between information and action

5 Information

Insight

Decision

Action

ANALYTICS IS

AWESOME (but show me the money!)

1

©Copyright Presidion 2016

Analysis & Monitoring

Predictive Analytics

Prescriptive Analytics

Reporting

What happened?

Why did it happen?

What will happen?

What should I do

next?

The right solution for the right business problem?

11

Analysis & Monitoring

Predictive Analytics

Prescriptive Analytics

Reporting

What happened?

Why did it happen?

What will happen?

What should I do

next?

The right solution for the right business problem?

12

Analysis & Monitoring

Predictive Analytics

Prescriptive Analytics

Reporting

What happened?

Why did it happen?

What will happen?

What should I do

next?

The right solution for the right business problem?

13

Predictive Maintenance What does it mean?

14 Manufacturing

Finding patterns in maintenance operations that could point to opportunities for

improvements

Identifying assets at risk of failure even when they have no previous failure history

Unearthing characteristics that lead to an increased frequency of failures

ASSETS

MAINTENANCE LOGS SENSORING

15

ASSETS

MAINTENANCE LOGS SENSORING

Predictive Maintenance is about

understanding the

patterns in data to determine the

areas of greatest risk and

directing resources before risk

becomes reality.

» Increase asset reliability and lifecycle

» Reduce interruptions in production/service

» Optimize resources scheduling and allocation

» Improve capital planning and financial

positioning

» Increase warranty recoveries and after-sales

support/customer satisfactions

Predictive Maintenance What does it mean?

16

ASSETS

SENSORING

MAINTENANCE LOGS

EXTERNAL INFORMATION

INTEGRATE RESULTS

ASSET HEALTH

Predictive Maintenance How does it work?

STATISTICAL MODELS

What if I could tell you that a specific

asset is 90% likely to fail within one

week for Reasons A, B and C?

2000 2005 2010 2015 2020

Bu

sin

ess

Val

ue

Ad

dit

ion

Data Reporting Data Monitoring

Data Analysis, Mining and Evaluation

Predictive Analytics

PHASE 1

PHASE 2

Maintenance Types

17

Evolution of Analytics in the Maintenance

Domain for Manufacturing

Reactive

Maintenance

Planned

Maintenance

Condition Based

Maintenance

PHASE 3 Predictive

Maintenance

PHASE 4

Predictive Maintenance What does it look like?

18

Predictive Maintenance What does it look like?

19

Predictive Maintenance What does it look like?

20

Predictive Maintenance What does it look like?

21

Predictive Maintenance What does it look like?

22

Today’s topics

How to succeed with Predictive

Maintenance.

‘I’ll have what she’s having’ – How Predictive Maintenance

borrows from data rich industries

How does Predictive Maintenance work?

How Predictive Maintenance is a proven

capability that leaders have already embraced

to deliver significant RoI

23

Predictive Maintenance Nothing New?

24

STATISTICAL MODELS

Asset Lifetime Estimated time to asset failure or maintenance

What is the expected time for

survival during treatment?

SURVIVAL ANALYSIS

Predictive Maintenance Nothing New?

25

STATISTICAL MODELS

Sensor Anomalies Highlight unusual sensor activity leading to or

indicating sub-optimal asset performance

ANOMALY DETECTION

Which claims appear fraudulent as they have a claim value vastly

different than expected?

Predictive Maintenance Nothing New?

26

STATISTICAL MODELS

Maintenance Insights Uncovered additional details added to

maintenance logs

What are the key topics raised and customer sentiments expressed for

call centre wrap notes?

TEXT MINING

Predictive Maintenance Nothing New?

27

Event Associations What log events occur together leading to

maintenance or failures

MACHINE LEARNING

Root Cause Analysis Key characteristics and factors leading to asset

failure or maintenance

STATISTICAL MODELS

28

UNDERSTANDING

Asset Data: Age, Material, Warranty

Environment Data: Weather, Location

Operations Data: Sensors, Temperature, Energy

Consumption

Interactions Data: Maintenance, Failure Logs

A

O

E

I

U

Predictive Maintenance Nothing New?

WHAT INFORMATION DO MANUFACTURING INDUSTRIES USE FOR PREDICTIVE MAINTENANCE?

Start where you can, with what you have. Expand Incrementally overtime...

Today’s topics

How to succeed with Predictive

Maintenance.

‘I’ll have what she’s having’ – How Predictive Maintenance

borrows from data rich industries

How does Predictive Maintenance work?

How Predictive Maintenance is a proven

capability that leaders have already embraced

to deliver significant RoI

29

30

What issues have the

highest impact on my business?

Do I have enough data and in the right

formats?

Can I execute and

operationalize results?

Predictive Maintenance Business Value Assessments

Which areas are costing me the most downtime/ maintenance costs?

31

What issues have the

highest impact on my business?

Do I have enough data and in the right

formats?

Can I execute and

operationalize results?

Predictive Maintenance Key tenets of success

32

Predictive Maintenance Data Readiness Assessments

Is my data good enough for Predictive Maintenance?

It is most likely good

enough to start

33

What issues have the

highest impact on my business?

Do I have enough data and in the right

formats?

Can I execute and

operationalize results?

Predictive Maintenance Organisational Readiness Assessments

When do I need results and in what

format?

34

What issues have the

highest impact on my business?

Do I have enough data and in the right

formats?

Can I execute and

operationalize results?

Lack of Focus

Low Adoption Poor Results

ACTIONABLE

INSIGHTS

Predictive Maintenance Key tenets of success

35

Business Domain Expertise

Data and Analytics Specialist

End User and Change Management

Predictive Maintenance Key tenets of success

Predictive Maintenance What does it look like?

36

Potential range of impact by employing Predictive

Maintenance

expense potential improvements*

annual cost of unplanned

downtime reduce by 60% to 90%

excess capacity required to

compensate for unplanned

downtime

reduce by up to 90%

scrap or re-work reduce by up to 50%

asset useful life extend life by 5% to 15%

* Based on Nucleus Research analysis 37

End to End Advanced Analytics Framework Approach

Learn & discover how

Advanced Analytics can transform your organisation

Prepare yourself for success in Advanced Analytics

Improve business

performance

Sustain improved

performance

Questions and Answers

39

Data House

79 Old Kilmainham Road

Kilmainham

Dublin 8, Ireland

TELEPHONE

+353 (0) 1 456 7896 EMAIL

info@presidion.com

www.presidion.com

Formerly SPSS Ireland

Thank you for attending

Follow Us and check out our Blog

www.presidion.com

info@presidion.com +44 (0)208 757 8820 (UK) +353 (0)1 415 0234 (Ireland)