Practical Challenges of
Data Analytics
Leigh Lapworth
Chief Design Systems Architect
Coping with Big Data – an Analytics and Computational Perspective
7th Jan 2015, UCL, London
© 2015 Rolls-Royce plc
The information in this document is the property of Rolls-Royce plc and may not be copied or communicated to a third party, or used for any purpose other
than that for which it is supplied without the express written consent of Rolls-Royce plc.
While the information is given in good faith based upon the latest information available to Rolls-Royce plc, no warranty or representation is given
concerning such information, which must not be taken as establishing any contractual or other commitment binding upon Rolls-Royce plc or any of its
subsidiary or associated companies.
Trusted to deliver excellence
© 2015 Rolls-Royce plc
Contents
Company overview
Product challenges
Data challenges in:
• Simulation
• Testing
• Manufacture
• Service
© 2015 Rolls-Royce plc
Rolls-Royce – 2013 Highlights
© 2015 Rolls-Royce plc
© 2015 Rolls-Royce plc
Meeting the challenges of modern technology
Steel starts to glow red at 700°C
© 2015 Rolls-Royce plc
Data in simulation
Increasing size
and complexity of
simulation data
10x increase in FEA
analyses per year
since 2005
Courtesy NAFEMS
© 2015 Rolls-Royce plc
Analytics for simulation
Directed data
interrogation and
discovery
Data
sentencing
© 2015 Rolls-Royce plc
In-situ solver
analysis
Simulation challenges
Variety and volume
• Multiple codes and formats (in-house and COTS)
Veracity
• Validation represents a major business investment
• Increased physical fidelity will drive orders of magnitude
increase in model size and HPC requirements
Velocity
• Large models stress IT networks
© 2015 Rolls-Royce plc
Data in testing
Under cowl
temperatures
up to 350oC
10km of cabling and 15km of tubing
Centrifugal loads
on rotating
instruments up to
100,000g
© 2015 Rolls-Royce plc
Measurement uncertainty
Guarantees
• We sell engine performance with possible high financial
penalties if we miss guarantees
Accuracy
• Performance critical measurements need very high
accuracy
• Thrust accuracy is equivalent to weighing a jelly baby on
bathroom scales in a force 7 storm
• All of these require secondary measurements and stable,
accurate references and well controlled calibration
• Instructing such measurements need c.50 pieces of coordinated data fed through a business system
© 2015 Rolls-Royce plc
Analytics for testing
Variety and volume
• Steady state, transient, dynamic, high speed photography
• File sizes range from 1Mb to > 1Tb
Veracity
• Accurate to ±0.1%
Velocity
• Dynamic data needs reduced and presented in frequency
format for safety monitoring in <0.5secs from acquisition
• 10,000fps for digital photography
© 2015 Rolls-Royce plc
Data in manufacture
Coordinate measurement, optical scanning, etc.
© 2015 Rolls-Royce plc
Analytics for manufacture
Variety and volume
• Statistical process control  RAG
• Coordinate measurement and scanning
• Inspection: manual  digital x-ray
Veracity
• Target 99.9% right first time
• Modelling to improve processes and reduce cost
Courtesy Advanced Forming Research Centre,
University of Strathclyde
© 2015 Rolls-Royce plc
Data in services
12 billion miles
flown every
year
A Rolls-Royce
powered aircraft
takes off or lands
every 2.5 seconds
22 million data
reports a year
© 2015 Rolls-Royce plc
Analytics for services
Sense
Acquire
ACMS Reports
via ACARS
Engine Monitoring
Unit
Global Network
eg: SITA
Transfer
QAR, DFDR
Flight
Log
Sheets
Ground-based
information,
e.g. oil uplift
Ground
Station
Analyse
Condition monitoring,
Data processing &
storage,
Data access & reports,
Forecasting services
Act
© 2015 Rolls-Royce plc
Closing the loop - feedback to design intent
Full Closed loop, high detail
Service verification
Rapid collation of full
product life cycle data in the
event of an in-service issue
or a re-design
Evidence based, auto risk
assessments
Rules / Models
Risk Assessment
Dimensions
As
Designed
© 2015 Rolls-Royce plc
As
Manufactured
Events
Duty Cycle
As
Operated
Condition
As
Maintained
Data curation
Products remain in service for many years
• T56 (C-130 Hercules) – 18,000 engines since 1954
• 2014 T56 upgrade – extends C-130H fleet to 2040
Geometry Data
• Physical drawings
• Legacy Databases and codes
Functional Data
• Legacy codes, some >30 years old
• Bespoke formats – own and 3rd party
• Rate of change of codes, databases, formats etc. is
increasing
© 2015 Rolls-Royce plc
Not just aerospace
© 2015 Rolls-Royce plc
Closing remarks
Holistic view
• Huge variation in data types, volume, velocity and veracity
• Closed loop systems offer product and service innovation
Functional view
• Data analytics is essentially functional evaluation
• Black swans – need to keep positive and negative results
IT view
• Fully connected systems
• Operational effectiveness relies on data models & standards
• Plan for legacy data, codes, media etc.
© 2015 Rolls-Royce plc