intelligent asset management physics based digital twins...sap intelligent asset management asset...

PUBLIC

Intelligent Asset Management

Physics based Digital Twins

Marit Reiso, PhDProject Manager, Product DeliverySAP Norway Engineering Centre of Excellence, PEI

3PUBLIC© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ

Strong Momentum for Intelligent Asset Management

76%of organizations consider it important to predict

potential failures by leveraging data models

Source: Business Performance Benchmarking by SAP 2017

Only 21% of organizations employ predictive

and preventive maintenance effectively

21%

of organizations consider it important to run real-time

asset management processes

83%Only 13% of organizations are able to drive asset

performance based on analysis of real-time

sensor data, along with historical maintenance

data

13%

Digital readiness surveys show that companies clearly see

the need to leverage the digital capabilities to optimize their asset management:

4PUBLIC© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ

Motivation – Intelligent Asset Management

As

se

t C

on

dit

ion

TimeTotal Failure

Functional Failure

Human

T

F

Equipment

Predictive Maintenance

PPotential Failure

P

P

P

More time to respond enables

greater flexibility to dynamically

plan maintenance events

5PUBLIC© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ

Maintenance and

Service Management

SAP S/4HANA

SAP Leonardo IoT

Asset Strategy and

Performance

Asset

Network

and

Collaboration

Asset

Health

Prediction

and

Optimization

SAP Enterprise Asset ManagementIntelligent Asset Management

6PUBLIC© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ

SAP Digital Twin

Business View

Statistics View Physics ViewIntegration

SAP Intelligent Asset Management Asset Health Prediction and Optimization

9PUBLIC© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ

SAP Intelligent Asset Management Asset Health Prediction and Optimization

Sensor Data75 78 82 79 75 78 82 79Sensor Data

Simulation-based Digital Twins

Leverage IoT enabled engineering simulation models for asset health

prediction and optimization based on multi-physics simulations

Data Science

Use machine learning to provide advanced notice of a failure to

reduce the number of unplanned downtime maintenance

10PUBLIC© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ

Digital Twin for Structural Dynamics

11PUBLIC© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ

Asset Health Prediction and OptimizationExamples

Wind turbines 2 Bridges 3 Vibrating equipment1

12PUBLIC© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ

ExamplesExtreme loads from rapid changes in stateProduction > stop > production

1

13PUBLIC© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ

ExamplesOne radical happening “eats” 4 days normal life

1

14PUBLIC© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ

Examples – Produced power vs fatigue1

15PUBLIC© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ

Asset Health Prediction and OptimizationExamples

1 Wind turbines 2 Bridges 3 Vibrating equipment

16PUBLIC© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ

Photo: Rambøll på oppdrag for Statens vegvesen

Photo: Rambøll på oppdrag for Statens vegvesen

Photo: Rambøll på oppdrag for Statens vegvesen

Idealized model

Photo: Morgan Frelsøy/OPP

Example bridgesGlobal structural deterioration

2

17PUBLIC© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ

Sensor

3975

Sensor

3972

Virtual

sensor

3975

Virtual

sensor

3972

Photo: Morgan Frelsøy/OPP

Example bridgesGlobal structural deterioration

2

Recapture

physical

sensor

behaviour

18PUBLIC© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ

14 Bridge Crossings

Sensor 3972, az

S0 S5

Sensor

3972

Example bridgesGlobal structural deterioration

2

19PUBLIC© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ

Asset Health Prediction and OptimizationExamples

1 Wind turbines 2 Bridges 3 Vibrating equipment

20PUBLIC© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ*Additional license - **Additional installation and license – C(Partially) Compatibility scope

Example Vibrating Equipment3

21PUBLIC© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ

SAP Intelligent Asset Management Asset Health Prediction and Optimization

Sensor Data75 78 82 79 75 78 82 79Sensor Data

Simulation-based Digital Twins

Leverage IoT enabled engineering simulation models for asset health

prediction and optimization based on multi-physics simulations

Data Science

Use machine learning to provide advanced notice of a failure to

reduce the number of unplanned downtime maintenance

22PUBLIC© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ

Machine Learning/Teaching

http://adilmoujahid.com/posts/2016/06/introduction-deep-learning-python-caffe/

Starting point: 50/50

23PUBLIC© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ

Machine Learning in industrial application is

used to determine «normal vs irregular»

WHAT IS NORMAL?

• The laws of physics are constant (=Normal?!)

SAP Digital Twin can be used to train the

Machine Learning Algorithms

SAP Digital Twin and Machine Learning

24PUBLIC© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ

• Continuous versus campagne

• Integrated platform in one core SAP system. No stand-alone measuring system

• Real time asset behaviour

• Integrate maintainance, inspection, risk matrix and geographical overlay

• Based on the need for prediction; simulations using sensor data or the Digital Twin models can predict global behaviour

Intelligent Asset ManagementAdvantages

Marit Reiso, PhD

SAP Norway Engineering Center of Excellence

M: +47 90 14 15 08

E: m.reiso@sap.com

Thank you.

AI/ML | IoT | Analytics

Intelligent Technologies

Digital Platform

DataManagement

CloudPlatform

CustomerExperience

Manufacturing& Supply Chain

Digital Core PeopleEngagement

Network & SpendManagement

Intelligent Suite