SLIDE 1
The jour he journey i ney in r n railw ailway ay ana analytics - - PowerPoint PPT Presentation
The jour he journey i ney in r n railw ailway ay ana analytics - - PowerPoint PPT Presentation
The jour he journey i ney in r n railw ailway ay ana analytics ytics po power ered by ed by AI: AI: Towar ards ds railw ailway ay 4.0 4.0 Professor Diego Galar Lulea University of technology Head of Maintenance &
SLIDE 2
SLIDE 3
But here be the dragons!!, approaches fail to scale
SLIDE 4
What analytics can be performed on railway?
SLIDE 5
Analytics and expectations also change
SLIDE 6
Types of data analytics
SLIDE 7
Descriptive analytics
SLIDE 8
Types of data analytics
SLIDE 9 PROGNOSTICS DATA MINING BLOCK
Data Reduction
Historical and live data
RELEVANT FEATURES
UNIFIED DATA FORMAT DATA PREPARATION
Feature selection
Optimal thresholds For features
ADVANCED PREDICTION PREDICTIVE MAINTENANCE ESTIMATION
UPDATE MAINTENANCE PLAN PROCESS
Diagnostic analytics
SLIDE 10
Types of data analytics
SLIDE 11
Feature of item n crosses boundary in time t RUL considering two features Feature of item n crosses boundary in time t+dt
Bearing mounted by Contractor 1 Bearing mounted by Contractor 2
Predictive analytics:RUL prediction
SLIDE 12
Types of data analytics
SLIDE 13
Types of data analytics
SLIDE 14
The way forward
SLIDE 15
16
Where analytics should be performed?
SLIDE 16
Edge agents versus cloud centralized
SLIDE 17
SLIDE 18
SLIDE 19
SLIDE 20
SLIDE 21
SLIDE 22
AI workflow @edge
SLIDE 23
Huge gap between data science and O&M
SLIDE 24
Now casting
1) What has happened 2) What is happening
Forecasting
3) What will happen in the future 4) When will it happen
What can I see in my data?
SLIDE 25
Domain knowledge and physics sometimes is not in the data
SLIDE 26
The method, let us twin reality
SLIDE 27
The twin as a service provider
29
SLIDE 28
The picture of Dorian Gray
SLIDE 29
Digital Twin: A virtual instance for services
SLIDE 30
Digital Twin Solution Architecture
SLIDE 31
Di Digital gital twin twin bas based ed on O
- n OT
SLIDE 32
Di Digital gital twin twin bas based ed on O
- n OT
On board Wireless System
Machine Maintenance Analytics
Data Information Knowledge
Internet
eMaintenanc e Cloud Server
SLIDE 33
Di Digital gital twin twin bas based ed on O
- n OT
SLIDE 34
Wha hat a t abou bout IT t IT sys systems? tems?
SLIDE 35
Rule-1 FailureMode(?x) ^ hasHappened(?x, true) ^ Device(?y) ^ happenedAt(?x, ?y) ^ FailureMode(?z) ^ theEndFffectIs(?z, ?x) ^ FailureMode(?a) ^ theHighEffectIs(?z, ?a)?theDirectFailureCauseIs(?x, ?a) ^ hasHappened(?a, true)
1 1 2 2
Tax axono
- nomies
mies and and
- ntologies
- ntologies
SLIDE 36
TRANSFORMA TRANSFORMATIVE TIVE MAI MAINTE NTENAN ANCE CE SOL SOLUTIO UTIONS NS Inte Integration & A tion & Applica pplication of tion of Tec echno hnologies logies
IT
OT
SLIDE 37
IT OT
Digital Digital twin twin OT/IT T/IT inte integration tion
SLIDE 38
Digital twin Digital twin OT/IT inte T/IT integration tion
Technical services Truck scheduling On board Wireless System
Machine Maintenance Analytics
Data Information Knowledge
Internet
eMaintenanc e Cloud Server
SLIDE 39
The he Way ay Forw
- rwar
ard
Sensemaking Algorithms All Digital Data
Context Engines
Time Computing Power Growth
SLIDE 40
Conte Context xt-aw awar are e Mainten Maintenanc ance e Decisi Decision
- n Suppo
Support t Solution Solution
Digital twin based on context
Data Fusion & Integration Big Data Modelling & Analysis Context sensing & adaptation
Information models Knowledge models Context models
Maintenanc e Data
SLIDE 41
Let us b Let us be car e careful big eful bigger = smar ger = smarter? ter?
- tolerate errors?
- discover the long tail and corner
cases?
- more data, more error (e.g.,
semantic heterogeneity)
- still need humans to ask right
questions, lack of analytics
SLIDE 42
Bl Blac ack Sw k Swan Lo an Losse sses
- Loss Distribution
- Tail events are rare
– very little data
- Typically strong
model assumptions
SLIDE 43
Da Data d ta driv riven en or mode
- r model based?
l based?
SLIDE 44
Ev Evolution
- lution of
- f the
the Pr Proce
- cess
ss
90s
Integration of Product Design and O&M
80s 2000 2016…..
3D
Digital Mockup
Digital twin
Knowledge Capture
2D
Technological Advance
Design & Validation of products
SLIDE 45
Hybrid Hybrid & C & Conte
- ntext Dri
xt Driven en Se Services vices
Physics
- f failure
based Data driven Hybrid models Context Awareness
Context Driven Services
SLIDE 46
Di Digital gital twin twin hyb hybrid rid
IT
ET
OT
SLIDE 47
Di Digital gital twin twin hyb hybrid rid
SLIDE 48
SLIDE 49
Hybrid Digital T Hybrid Digital Twin win
51
Physical asset
Historical records Pre- processing Feature extraction Physics- based model Diagnosis Prognosis Synthetic data Cloud computing Pre- processing Feature extraction Hybrid engine Failure detection Maintenance planning Maintenance action
OFFLINE PROCESS: VIRTUAL COMMISSIONING ONLINE PROCESS: OPERATIONAL LEVEL ANALYSIS CONTEXT
CM data Risk mitigation / actuator Trend analysis Risk assessment
SLIDE 50
Application of railway twins
Virtual commissioning services Virtual assets O&M information Data loop Service/Repair Shop Design feedback loop
SLIDE 51
Some hints
SLIDE 52
Concluding remarks
- Digital twins and Hybrid models are needed
for virtual commisioning to deliver O&M services
- O&M based on Data driven solutions can lead
to catastrophic failures
- Life extension is not possible with big data
analytics
- Manufacturers must provide the integration of
systems and data
- Digital twin 4.0 will consider evoltionary
models and normality dynamics
SLIDE 53