Künstliche Intelligenz in Instandhaltungsprozessen

Willkommen zum Gedankenaustausch über Künstliche Intelligenz (KI) in Instandhaltungsprozessen! Am 10. September 2024 haben wir die Gelegenheit, von Experten wie Thomas Schäfer von RES Consulting und Markus Seidl von SAP Deutschland zu lernen. In dieser Veranstaltung werden wir die Grundlagen der KI, die Algorithmen und Modelle, die SAP in seinen intelligenten Funktionen verwendet, sowie spannende Anwendungsfälle in der Instandhaltung erkunden. Lassen Sie uns gemeinsam entdecken, wie KI als Motor der vierten industriellen Revolution fungiert und welche neuen Berufe in der Zukunft entstehen werden!
Hier sind drei Beispiel-Fragen, die Sie über die Datei stellen können:
1. Was sind die Hauptthemen, die während des Gedankenaustauschs behandelt werden?
2. Wer sind die Referenten und welche Unternehmen vertreten sie?
3. Warum ist Künstliche Intelligenz heute so wichtig für die Instandhaltungsprozesse?

1. Management Circle
Gedankenaustausch:
Kl in Instandhaltungsprozessen

10. September 2024

2. Vorstellung der Referenten

Gedankenaustausch: KI in Instandhaltungsprozessen

a

Thomas Schäfer Markus Seidl

Geschäftsführender Gesellschafter Principal Solutions Architect
RES Consulting GmbH, Rheinberg SAP Deutschland SE & Co. KG, München

Ansprechpartner
der SAP

Stellv. Sprecher

RES >

3. Agenda

Gedankenaus

Was heißt eigentlich KI - Eine kurze Einführung

Welche Algorithmen/Modelle verwendet SAP in neuen
Intelligenten Funktionen?

Welche Anwendungsfálle KI gibt es in der Instandhaltung?

4. Warum hat Kl heute eine so groe
Bedeutung gewonnen?

Je - Gegäikenaustausch Klin
dhaltungsprozessen

5. Kl als Motor der 4. industriellen Revolution

Kúnstliche Intelligenz

RES >

6. BF es Prozent unser ie
»

Kinder werden 2035

»
»

7. In welchen Bereichen entstehen gerade
KI Anwendungen?

Je - Gegäikenaustausch Klin
dhaltungsprozessen

8. Landwirtschaft Y e

Sport Justiz

Verwaltung \ N

»
Handel OZ

Bildung So i a

9. A(BNC) = 22
n(B) =68
n(C) =84

%) = logix- l0guy
d arb =b+e
albre)=abrac 10000a+ 100b- 58

126 =6y
2x + 2y=20

10. Definition KI

Kúnstliche Intelligenz
= keine eindeutige Definition für den Begriff »künstliche Intelligenz«

= KI steht für alle Methoden, die es einer Maschine erlauben,
kognitive Fahigkeiten einer menschlichen Intelligenz
nachzuahmen

= insbesondere Fähigkeiten wie Lernen und Problemlósen, aber

auch Handeln, Wahrnehmen, Planen, Wiedererkennen,
Kommunizieren und sogar Neues erschaffen

= Unterscheidung zwischen schwacher KI und starker KI

11. Unterscheidung starke / schwache KI

Kúnstliche Intelligenz

Starke KI Schwache KI

+ Ein starke Kl kann selbstständig « Die schwache KI (auch als methodische KI
Aufgabenstellung erkennen und definieren bezeichnet) besitzt keine Kreativität und
und sich hierfür selbständig Wissen der keine expliziten Fähigkeiten selbstständig im
entsprechenden Anwendungsdömäne universellen Sinne zu lernen.
erarbeiten und aufbauen.

+ Sie untersucht und analysiert Probleme, um + Mit ihr kann man klar definierte Aufgaben
zu einer adäquaten Lösung zu finden - die mit einer festgelegten Methodik bewältigt,
auch neu bzw. kreativ sein kann. um komplexere, aber wiederkehrende und

genau spezifizierte Probleme zu lösen.

Quelle: hitps,/Khws.de/thematik/starke:vs-schwache:ki

13. X-4X+555

\bl _ a
i a E X-4X <0' pe ~
N a n(BNC)=22 ON
n(B) =68
ge: bi = =) n(C) = 84 Luis
2364028 = $ n(BUC) =n(B)+n(C}n(BnC)
46498012090,
Be IS
x
EM
log,a
rlog,x
xy) = log,x + logey
$) =log.x- logsy
A

+ RES

ai > consis

15. Konnen Computer wirklich Denken?

o - Gedankenaustausch Kl in
altungsprozessen

18. Algorithmus

Kúnstliche Intelligenz

= Algorithmen sind eine definierte Abfolge einzelner eindeutiger Handlungsvorschriften zur
Lösung einer bestimmten Aufgabe

= Programmiersprachen übersetzen Algorithmen in Boolscher Algebra, dabei bedeutet Null,
»Strom aus«, oder Eins, »Strom an«

. Programmier- 1010
Algorithmus = sprache = 1010

RES >

19. Unterschied Computerprogramm und KI

Kúnstliche Intelligenz

Lin a
Em 0

Du kaufst ‘| Teil.

20. Vielen-Dank für Ihre Aufmerksamkeit!

Kontaktdaten
0 [phone] /+4941 70 22,52 397
[email protected]
9 Orsoyer Str. 8, 47495 Rheinberg,
L_J] vwwresconsuiting.de
Management Cifcle - Gedäfkenaustausch Klin =

32 Instandhaltungsprozessen =

21. Klin der
Instandhaltung

Markus Seidl, SAP
Principal Solution Architect

EXTERNAL 10. September 2024
Management Circle

22. Disclaimer

‘The information in this presentation is confidential and proprietary to SAP and may not be disclosed without the permission of SAP.

Except for your obligation to protect confidential information, this presentation is not subject to your license agreement or any other service

or subscription agreement with SAP. SAP has no obligation to pursue any course of business outlined in this presentation or any related document, or
to develop or release any functionality mentioned therein.

This presentation, or any related document and SAP's strategy and possible future developments, products and or platforms directions and
functionality are all subject to change and may be changed by SAP at any time for any reason without notice. The information in this presentation is not
a commitment, promise or legal obligation to deliver any material, code or functionality. This presentation is provided without a warranty of any kind,
either express or implied, including but not limited to, the implied warranties of merchantability, fitness for a particular purpose, or non-infringement.
This presentation is for informational purposes and may not be incorporated into a contract. SAP assumes no responsibility for errors or omissions in
this presentation, except if such damages were caused by SAP's intentional or gross negligence.

All forward-looking statements are subject to various risks and uncertainties that could cause actual results to differ materially from expectations.
Readers are cautioned not to place undue reliance on these forward-looking statements, which speak only as of their dates,
and they should not be relied upon in making purchasing decisions.

23. Klin der Instandhaltung
Was ist SAP Joule und was bringt es?

Fazit, Ideen und Ausblick

EXTERNAL

24. Intelligent Asset Management with SAP

Solution areas
to be powered
Business Process Flow: Maintenance Strategy to Execution by Al

| a 5:
— Maintenance Planner
‘Subject Matter Expert

Business Application: Intelligent Asset Mat

any Employee

Business Technology Platform

25. SAP Asset Performance Management

Failure Mode Analytics with PoF Curve
Monitor Asset Health and Maintenance Demand

Key Features: ‘Raw water Pump ADO NOT CHANGE ANY DATA) a

Use out of the box model based on Weibull u
algorithm to process raw data de a E om

Derive probability of failure

Calculate an indicator forecast to show trends of
indicator values

Predict Failure Date and Time to Failure in days Va

Store derived data against the asset indicators to gain
and provide deeper insights into the asset health [

Integrated into standard SAP processes a a A i e sd = a

26. SAP Asset Performance Management

Visual Inspection for Assets

$8 Identify defects, assembly errors and measurements by alerting reliability engineers

VI closes the gap between technology and processes and enables
everyone to integrate computer vision in their inspection processes.

Key Features:

ONO)
- Orchestration of VI scenarios by managing computer
vision models fully integrated in business processes I \
+ Out-of-the-box integration with various SAP solutions to = +

plan and execute Visual Inspection steps

» Assist and automate visual inspections through
simple and intuitive apps including feedback loop

» Integrated into standard SAP processes

27. Asset Performance Management Visual Inspection with Drones * La €
Detect failures on insulators as part of overhead lines

Expert Feedback optimizing Al

Monitor Indicator View Alerts & Confirm cr
Values and Execute Failure Mode Ne
ules

Maintenance
tion or Order

Define Asset Master Define Visual inspection
ta, Measurem
Failure Modes

Insulator Imae
Drone insp

28. Inspection Process — Conceptual Design

Continuous optimization of Al Model

Al Analyzing and

processing Image Y

Drone shots of Checking Create Alert or Perform
insulators on ct Notification Maintenance
overhead lines Measure

Pre-processing of Object Predict on
images Detection image

Customer Computer Vision Pipeline SAP APM

29. Similar use cases in Manufacturing and Quality Management

Assisted Mode:
Al-Assist the Expert
(without defects)

Manual Mode: Assisted Mode:
Expert trains Al Al-Assist the Expert
(with defects)

30. SAP Asset Performance Management

Detect anomalies and generate maintenance backlog

2 Increase asset uptime and lower maintenance costs by alerting reliab engineers

A «9
Equipment anomalies and failures often go unseen,
exposing companies to increased risk of undetected

malfunctions or failures, leading to potential downtime
and higher maintenance costs.
Key Features:

An anomaly detection algorithm works in the background
to automatically identify potential machine failures
based on loT sensor data

Reliability Engineers are alerted about potential
machine failures, assess the detected anomalies and
can trigger suitable follow up actions

Integrated into standard SAP processes

31. SAP Business Network with ChatGPT PoC

es

Tons of documents with valuable information are kept in pdf =
formats. The challenge is to get fast access to retrieve desired
information in a smart way.

Model gets trained by documents like pdf
via SAP Business Network Asset Collaboration
using a partner solution*

Technicians get easy access via Chatbot prompt

Integrated into standard SAP processes
* Nihilent Partner Solution

32. SAP Business Al approach

Relevant. Reliable. Responsible.
Finance | Supply Chain | HR | Procurement | Marketing and Commerce | Sales and Services | IT

Joule
A copilot that truly understands your business.

Cloud ERP Human capital ‘Spend management Customer relationship. Business Technology
management and business network management Platform
X ANTHROP\C aws = cohere © databricks

@DataRobot ;oogle Cloud HE Microsoft

33. Artificial Intelligence in DSC Operate — Current Ideas and Plannings cr

SAP Field Service Management

Al-based scheduling
Predictive routing

Intelligent fitering and natural language
search*

Job prediction duration‘

Equipment summary report & activity,
summary report*

Generative Al for Admin*

Simulate and compare change:
policies"

scheduling

agement

Estimate time to failure and probability of
failure based on Weibull analysis

Detect anomalies and create alerts"

Al enabled visual inspection for condition
monitoring"

Custom Al models and algorithms*

Intelligent support for assessments
and smart recommendations

Cloud Asset Mobile Execution &
nagement / RSH =! Dynamic Form:

Smart work order completion + Aldriven maintenance execution: *

using text mining to propose © Voice recognition and voice to text

object part and damage codes functionality

Incorporating Joule Assistant" ‘© Completion work screen

Intelligent maintenance order .

m Incorporating Joule Assistant” 0%
recommendation”

+ Arasssted browsing of knowledge bases*
based automatic scheduling and expected
‘chedulé attalimient + Voice recognition and voice to text
functionatty* "™"

+ Equipment & activity summary report® "9

* Al generated report indicating state of
equipment ran

34. SAP Mobile Execution and Dynamic Form:

e recognition and voice to text function:

Improve technician productivity with voice to text
functionality to assist with task execution

Technicians face productivity challenges due the amount of data they have to
manually fill in when executing job and confirming operations.

SOLUTION
SAP Business Al

+ Use your voice to give commands to the solution to support in data entry when Subject to change
confirming operations and work orders.

OUTCOMES

Increased productivity and wrench time
showcasing the direct impact ofthe speed of execution while reducing errors through the elimination
of manual data entry

35. oule is not Prompt
Engineering

36. Joule, The « ONE » Digital Assistant for the SAP Cloud Business Software

tr

28918 mem x com u y

The same Informati the same Interaction — always, everywhere

37. Joule

Distinguish the Different Levels of GenAl Usage

Joule Ux

+ Natvely integrated,
out-ofthe-Box CoPilot

+ Consistent Ux across.
applications.

Joule Context

+ Semantic ftering of
available scenarios &
knowledge based on
user input

+ Context is dynamically
assembled upon user,
application & history

Joule LLM Access.

+ Bestof-breed partner
ums

+ Access via GenAl Hub to
ensure secure, private &
safe interactions

Joule Functions

+ Seamless & secure
integration to SAP apps
Via business connector

* Nested dependencies of
functions to model
complex dialog flows.

Grounding
nomade Retevant

ser trance Ervine query como
MT) 4 Grounded spon
How can help you?

‘content Building

Va Gens ab

cour? Supply Cain

Business Applications

38. nial j When can | use
À Joule?

One digital assistant throughout the
SAP cloud Portfolio!

39. view © a

WDF_50_Navigation

41. Loading your story
..o

my pumps”
outlet

‘ver the past 3 months including à

SACStrys ere

42. Markus,
can | assist you?

indicator Trend et ie tin 22028
Tit (lis 113
== ans ‘ u. => Cesteadadibosror"ny unos

». D 1 halo meto compare the eth
Sore he inet and lt pres wth

» Sempere

nn, over the at 3monthsindrg trend

hat

is generating

43. Was kann KI für die Instandhaltung leisten?

a — 3 2.8 E
.. E .- e Es ist a I Bereich, um KI-Fa jen sas 2 "

HET s ist'der Schlüssel zum Erfolg * +

a

Pr y über Produktesun ion HOOD rechatiun bis zur AlBerbetriebnahme* Qs
. = SAP Joule wird superleis: sfáhig, integriert und allgegenwártig = '
? . . = e

* 4

44. Thank yo

Contact information:
Markus Seidl, SAP
Principal Solution Architect

[email protected]

48. Wir freuen uns auf Fragen und eine
angeregte Diskussion.