rolls-royce data – strictly private 5 steps to heaven – a rolls-royce perspective ibm almaden...
TRANSCRIPT
Rolls-Royce data – strictly private
5 Steps to Heaven – A Rolls-Royce Perspective
IBM Almaden Institute, San Jose7th May 2008
Dr Ian K Jennions
EHM Global Capability Owner
©2008 Rolls-Royce plcThe 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.
This 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.
2
© 2008 Rolls-Royce plc
5 Steps to Heaven – A Rolls-Royce
Perspective
Introduction
Our Service Business
Asset Management
Business Landscape
5 Steps to Heaven
Simulation of Total Care
Value in Data
Technology Successes
Marine EHM (Equipment Health Management)
AE3007 HP Vane
QUICK
Future
Current Issues
Summary
3
© 2008 Rolls-Royce plc
Rolls-Royce’s four markets
4
© 2008 Rolls-Royce plc
Rolls-Royce Group plc - 2007
5
© 2008 Rolls-Royce plc
Video: ‘New Thinking, Joined Up Thinking'
6
© 2008 Rolls-Royce plc
5 Steps to Heaven – A Rolls-Royce
Perspective
Introduction
Our Service Business
Asset Management
Business Landscape
5 Steps to Heaven
Simulation of Total Care
Value in Data
Technology Successes
Marine EHM (Equipment Health Management)
AE3007 HP Vane
QUICK
Future
Current Issues
Summary
7
© 2008 Rolls-Royce plc
Changing Customer Requirements
Traditional Support Spare parts Repair & overhaul
Enhanced Support Data & forecasting
services Technical & logistics
support Customer training
Advanced Support Comprehensive package integrating
elements of basic and enhanced support Spares inclusive R&O contracts
Total Support Complete, availability- based service. Can cover
all off-aircraft, and some on-aircraft activity Configuration management & reliability
enhancements covered
Extended Support Partnered capability to cover complete weapon systems Turn-key services Non-propulsion related support solutions
Increased Scope
Cu
sto
mer
Val
ue
Rea
ctiv
e
Ali
gn
ed
Inte
gra
ted
8
© 2008 Rolls-Royce plc
The EHM Transformation Staircase - Defence
Increasing Scope of Service Provision
Increasing Scope of Service Provision
Basic EHM
- Lifing – cyclic counts- Digital borescope
Advanced EHM- Lifing- Performance trending- Oil trending- Vibration trending- Improved life counting- Basic diagnostics
Total Care EHM- Data fusion –
oil/performance/vibration- Integration with LCC &
WSG tools- Advanced diagnostic- Dynamic performance
model (COMPASS)
Customer Exploits Customer Exploits Info. to Reduce Cost Info. to Reduce Cost
Extended EHM- Piece part diagnostics- Prognostic capability- Networking of systems
(BROADEN)
Future EHM
- Unknown at this time, but could include Powerplant and/or platform health monitoring
Commodity Commodity ServiceService
RR Exploits Info. to Increase RR Exploits Info. to Increase Cost Effectiveness of Product Cost Effectiveness of Product
Stage 1
Stage 2
Stage 3
Stage 4
Stage 5
Basic Support
Traditional
Enhanced Support
Forecasting
Total Support
Engine Availability
Advanced Support
Asset Management
Extended Support
Total Management
Customer Risk Customer Risk Service Provider Risk Service Provider Risk
9
© 2008 Rolls-Royce plc
Total Care Simulation - Example
Product Operation aircraft, engine, airport,
operations control
Product Behaviour engine, module, part, virtual
characteristic
Asset Management parts, modules, dedicated &
lease pool spare engines, operator interaction
Manufacturing & Supply aircraft, engines, modules,
parts
Transportation & Logistics engines, modules, parts
Engine Repair, Overhaul & Test capacity, capability, resourcing, availability
Component & Module Repair
10
© 2008 Rolls-Royce plc
Total Care Simulation- video
11
© 2008 Rolls-Royce plc
–– –
Integrated Products and Services (IPAS)
–– –
Service Engineers/Overhaul trained and motivated to share knowledge with design. Tools, methods and processes available to support data and knowledge capture
Product designed to support the service offering –existing service knowledge base fully used, continuous improvement via optimized IPAS Processes
Designer able to search easily for knowledge via intuitive interface, ontology models and social solutions. Able to do LCC trade-offs, understands service needs
Designer Knowledge
Desktop (WP1)
LCC Modelling Toolkit (WP3)
Service Data
Product Designer
Service Engineer/ Overhaul Shop
Social Issues and Solutions (WP1)
–
Customer Customer
Customer
Optimised IPAS Processes
Key:
Service Engineers/Overhaul trained and motivated to share knowledge with design. Tools, methods and processes available to support data and knowledge capture
Product designed to support the service offering –existing service knowledge base fully used, continuous improvement via optimized IPAS Processes
Designer able to search easily for knowledge via intuitive interface, ontology models and social solutions. Able to do LCC trade-offs, understands service needs
Designer Knowledge
Desktop (WP1)
LCC Modelling Toolkit (WP3)
Service Data
Designer Knowledge Desktop
LCC Modelling Toolkit
Service Data
Product DesignerProduct Designer
Service Engineer/ Overhaul Shop
Service ProviderService Provider
Social Issues and Solutions
New (IPAS)
Existing
New (IPAS)
Existing
–
CustomerCustomer Customer
New (IPAS)New (IPAS)
Knowledge flow
12
© 2008 Rolls-Royce plc
The IPAS System Architecture
KnowledgeDesktop
SDMSAP RCP Reports Other
LegacyDoc
LegacyData
Source2
NewDoc
NewData
Source4
ETC Data
Information
Knowledge
K-Search
K-Forms
Mechanisms
Features
Parts
Materials
Data Types
Ontologies
K-Forms
Web Services
13
© 2008 Rolls-Royce plc
Legacy data is largely unstructured and is captured and stored in a variety of ways.
Development of ontologies for hardware description, failure mechanism etc. allow us to make sense of unstructured legacy data.
K-Search provides a means of interrogating legacy data and presenting structured search results back to the user
K-Search is a learning system that builds an understanding of document content such that after the ‘learning process’ >95% rate of correct automated document content classification can be achieved.
IPAS Results – the ability to search legacy data
Automatically extracting knowledge from legacy unstructured data
14
© 2008 Rolls-Royce plc
5 Steps to Heaven – A Rolls-Royce
Perspective
Introduction
Our Service Business
Asset Management
Business Landscape
5 Steps to Heaven
Simulation of Total Care
Value in Data
Technology Successes
Marine EHM (Equipment Health Management)
AE3007 HP Vane
QUICK
Future
Current Issues
Summary
15
© 2008 Rolls-Royce plc
Marine Off-shore Supply and Service Vessel
Propulsion units monitored: 4 Bergen 6 cylinder C
engines. 2 Azipull thrusters. 2 Tunnel thrusters. 1 swing up thruster.
16
© 2008 Rolls-Royce plc
Ethernet
Propulsion
EHM unitEngines
EHM unit
Thrusters
EHM unit
Engines
Data Coll. unit
RRM
Ship network
EHM
Central unit
RRM
Service Centre
Vessel 49 EHM Network Installation.
17
© 2008 Rolls-Royce plc
AE3007 HPT Vane Distress
Leading edge burns away over time Causes: cooling blockage, oxidation Condition that typically occurs on high time engines
(primarily on-condition engines) Causes increased stress on the 1st stage blade leading to
high cycle fatigue (secondary damage) Turbine blade fracture leads to in flight shut down and a 2x
shop visit cost
18
© 2008 Rolls-Royce plc
AE3007 HPT Vane Distress
Vane distress showed up as performance changes. Data fusion allows clear signature to be set
Fuel flow, turbine temp, pressure & speed 50 hour window after detection with 96% confidence of
preventing HPT blade failures Allows time to schedule removal Thresholds set to balance false positives and false negatives Relatively small number of false alerts and false negatives
19
© 2008 Rolls-Royce plc
AE3007 HPT Vane Distress - Savings
The alternative of not having EHM for this issue Allow HPT-1 blade failures Pull engines in mod campaign Implement a 50 hr boroscope inspection interval
8 yr EHM Savings estimated at over $60M Based on mod campaign or inspections both ($$$$) Savings continue as engines age and experience vane
distress
20
© 2008 Rolls-Royce plc
So What’s the Problem ?
With high-integrity systems such as jet engines, conventional fault detection methods have limited capability:
The most important examples (i.e. the abnormalities) are very rare Some of the fault conditions may not have been seen before Subtle pre-cursor events often exist prior to failure but these are rarely detected by employing conventional techniques.
Advances being made at RR via advanced computational and signal processing methods, specifically:
QUICK, learn a description of normality and test against it (OBS) Crack detection and propagation (Cranfield University) Knowledge Based Diagnostics (Sheffield University) DAME (Universities of York, Leeds, Sheffield and Oxford)
21
© 2008 Rolls-Royce plc
Quantitative User friendly Intelligent Collection of Knowledge (QUICK)
QUICK is a system used for real time acquisition and intelligent analysis of vibration and performance data from gas turbine engines.
QUICK has been developed over a number of years between Oxford University, Oxford BioSignals and Rolls-Royce
The innovation in the product lies in:
The use of Computational Intelligence to quantify features Building models of normality Fusion of data from disparate sources for engine diagnostics. The system is exploitable in other fields (e.g. medical)
22
© 2008 Rolls-Royce plc
QUICK video: Innovation Award
23
© 2008 Rolls-Royce plc
Novelty Thresholds – Bayesian EVTNovelty Thresholds – Bayesian EVT
2323
24
© 2008 Rolls-Royce plc 2424
Example results – 826 testsExample results – 826 tests
Method 1: Bayesian EVT Method 2: Classical EVT Method 3: Expert Heuristic
Data Set Sensitivity Specificity Sensitivity Specificity Sensitivity Specificity
A1 0.75 1.00 0.75 0.97 0.75 0.93
A2 1.00 1.00 1.00 0.90 1.00 1.00
B1 - 1.00 - 0.99 - 0.99
B2 - 1.00 - 0.85 - 0.45
B3 - 0.99 - 0.70 - 0.68
B4 - 1.00 - 0.70 - 0.65
Overall 0.78 0.99 0.78 0.83 0.78 0.73
25
© 2008 Rolls-Royce plc
5 Steps to Heaven – A Rolls-Royce
Perspective
Introduction
Our Service Business
Asset Management
Business Landscape
5 Steps to Heaven
Simulation of Total Care
Value in Data
Technology Successes
Marine EHM (Equipment Health Management)
AE3007 HP Vane
QUICK
Future
Current Issues
Summary
26
© 2008 Rolls-Royce plc
Different perspectives – system boundaries
27
© 2008 Rolls-Royce plc
EHM Intelligence Dichotomy Fleet centred view - the realisation of full value by having an accurate, real time fleet view of the
state of all assets Platform centric view - Devolving EHM intelligence to the local platform to compensate for:
– limitations in communicating with a central information hub (a lack of infrastructure and capacity)
– Requirements for assets to operate and be maintained autonomously– Concerns for data privacy and IPR
LOCAL PLATFORM
Often has to operatedeployed and autonomously,
and often has an ability toself maintain. If the platform
is mobile, datacommunications may be
limited
Data Communications infrastructure
CENTRAL HUB
Processes EHM datacentrally and issuesalerts. Maintains an
overview of the conditionof the full fleet. Is able tomitigate the operational
consequences offailure
LOCAL PLATFORM
Often has to operatedeployed and autonomously,
and often has an ability toself maintain. If the platform
is mobile, datacommunications may be
limited
LOCAL PLATFORM
Often has to operatedeployed and autonomously,
and often has an ability toself maintain ‘on mission’. Ifthe platform is mobile, datacommunications channelsmay change and be limited
Communicationsbandwidth may belimited and may be
expensive
Real time monitoring offers advantages ofbetter resolution, earlier detection and a
wider scope of failure mode detection - atthe price of having to process much
higher data volumes
Some organizationsdo not want to export
their data beyondtheir firewalls
THE RESULT IS PRESSURE TO PUSH EHM INTELLIGENCE TOWARD THE PLATFORM
28
© 2008 Rolls-Royce plc
Vision: HUMS Information Fused with Operations for Optimised Fleet Management
SAP
CoreControlCoreAlert
Service Data Manager Aero Exec Integrator
Gold (T45 Programme)
29
© 2008 Rolls-Royce plc
Vision: The Operations Room of the Future?
Oil uplift and utilizationdata transferred from
Squadron and pilot reports
EHUMS data download from A/C using data retrieval unit (DRU)
routed via Ground Support System (GSS) to Operations Centre
Customers alerted by pager, e-mail or cell phone
Customers can view status on DefenceManager
Operations Room - processes information
Service team monitors output and liases with customer
Repair and Overhaul
Agent-based aftermarket modelling and decision
support
All Operations linked through global GRID system
AI fault recognition and diagnosis
30
© 2008 Rolls-Royce plc
5 Steps to Heaven - Summary
In the OEM to Service business transformation data is key
Operating on the data to create information and knowledge has been demonstrated, including value, in a number of areas.
With risk shift from Customer to OEM in a Service model EHM is risk mitigation, informing the Aftermarket business and helping to minimise unplanned events
Basic EHM has been yielding good results for a number of years, the next steps are much more complicated
31
© 2008 Rolls-Royce plc http://www.cranfield.ac.uk/cua
Vision A world-class research and demonstration Centre of Excellence, seeking to establish industry standards and yielding competitive advantages in Integrated Vehicle Health Management (IVHM) for its industry members.
Strategic Aims • Attract Investment – supplement launch funding by attracting additional investment from new core partners/members and research funding bodies• World Recognition – working with industry and academic leaders• Technology Transfer – implementation of concepts and technologies within the industry and develop the associated skills base• Self Sustaining – after 5 years• Delivery of Competitive Advantage – to the key stakeholders
Cranfield Integrated Vehicle Health Management Centre of Excellence
Generic Architecture
CoE Concept Funding/ Governance