Transcript Integrated Vehicle Health Management in Network Centric

Smiths Aerospace
Integrated Vehicle Health Management in Network Centric Operations
International Helicopter Safety Symposium, Montreal
September, 2005
Piet Ephraim
www.smiths-aerospace.com
© 2005 by Smiths Aerospace: Proprietary Data
Outline
Network Centric Operation & its implications
Vehicle Health Management objectives and challenges
Background and Current developments
•
Comprehensive health management
•
On-board common computing platforms & networks
•
Ground system networks
•
New tools and architectures
Integrated Vehicle Health Management in the Net centric environment
Conclusions
© 2005 by Smiths Aerospace: Proprietary Data
Network Centric Operation (NCO)
NCO is a philosophy that aims to provide dispersed operations with:
• Greater speed, more precision, Fewer forces
• Information & Decision Superiority
• Shared Situational Awareness
• Interoperability
NCO includes ‘C4ISRS2’
• Command, Control, Computing, Communications
• Intelligence
• Surveillance
• Reconnaissance
• Support and Sustainment
© 2005 by Smiths Aerospace: Proprietary Data
NCO Implications
NCO implies:
• Greater reliance on maximised vehicle availability and reduced logistics
footprint – benefits afforded by Health Management
NCO requires:
• Information from data
• Timely delivery of accurate, coherent and comprehensive intelligence,
operational and logistics information
• Integration of sensors, decision makers, operational and support systems
through networked and integrated open systems
• Adaptability and extensibility
• Increased levels of autonomy
Health Management is an integral part of Net Centric Operations
© 2005 by Smiths Aerospace: Proprietary Data
Vehicle Health Management Objectives
Increased mission readiness, effectiveness and sortie rate
Reduced downtime (advise maintenance prior to return)
Improved safety
Reduced redundancy requirements
Reduced sustainment burden & logistics footprint
Address need for autonomous & integrated on-board health
management (e.g. for UAVs)
To provide the right information to the right people at the right time so that
decisions can be made and actions taken
© 2005 by Smiths Aerospace: Proprietary Data
Vehicle Health Management Challenges
Flexible, open Architectures
Improved Diagnostics & Prognostics - Decision Support tools
Optimised roles of, & interaction between, on-board and off-board
functions
Integration and Interoperability (sharing of monitored information)
Distribution of Data / Functionality - on-board & off-board
Autonomous (self-supporting) vehicle capability
Provide a demonstrated payback
© 2005 by Smiths Aerospace: Proprietary Data
Smiths Aerospace
Background and Current Development
www.smiths-aerospace.com
© 2005 by Smiths Aerospace: Proprietary Data
HUMS - 20 Aircraft types, 2 million flight hours
Bell-Agusta BA609
Agusta-Bell AB139
US Army
UH-60L &
MH-47E
Japan SH-60K
UK MoD
Chinook Lynx
Sea King
Apache
© 2005 by Smiths Aerospace: Proprietary Data
Example HUMS System
Rotor
Sensors
RT & B
Acceleromet
ers
On-board system
Engine
Acceleromet
ers
RT &B
Acceleromet
ers
Hanger
Bearing
Rotor Acceleromete
rs
Azimuth
Optical
Blade
Tracker
Altitude,
Airspeed &
Air
Temperatur
e Sensors
Area
Mic
Control
Position
Sensors
CG
Acceleromet
er
Pitch Roll
Heading
Sensors
Optical
Blade
Tracker
Rotor
Sensors
At aircraft
maintenance
Ground System Software
Depot Level
Fleetwide support
In-depth analysis &
Diagnostics
© 2005 by Smiths Aerospace: Proprietary Data
HUMS: Proven Benefits
HUMS: Proven Benefits
Increased safety
• Reduced fatal accident statistics
Significant annual savings:
Aircraft Usage
Monitoring – £600k
• Rotor track & Balance
Engine Health
Monitoring – £200k
Transmission Health
Monitoring – £1.0M
• Transmission Health
• Aircraft Usage
Rotor Track & Balance
– £1.5M
• Engine Health
Notable diagnostic successes:
• Minimised screening process
• Prevention of fleet grounding
© 2005 by Smiths Aerospace: Proprietary Data
Comprehensive Aircraft Health Systems
Doors and door actuators
STRUCTURAL HEALTH
ACTUATOR HEALTH
Engine Components
EDMS/IDMS
OIL CONDITION
VIBRATION USAGE
IGNITOR HEALTH
ROTOR HEALTH
LOD
Hi-Lift systems
STRUCTURAL HEALTH
Fuel Systems
FUEL QUALITY
LEAKAGE
PUMP HEALTH
Fuel & hydraulic tubes/hoses
SMART VALVES CORROSION
LEAKAGE
OBSTRUCTION DETECTION
Environmental Control
SUBSYSTEM HEALTH
Power Generation
GENERATOR HEALTH
Weapon Control & Release
SUBSYSTEM HEALTH
Integrated Avionics,
Flight Management, Data,
Displays
SUBSYSTEM HEALTH
LEAST DAMAGE NAV
Power Distribution
ARC FAULT DETECT
Current
Growth
Cable Harnesses &
Connectors
ARC FAULT PROTECTION
WIRE FAULT DETECT
Utilities Management
SUBSYTEM HEALTH
Fly-by-wire flight
control actuators
ACTUATOR HEALTH
Airframe components
STRUCTURAL HEALTH
© 2005 by Smiths Aerospace: Proprietary Data
On-board common core computing
Common Computing Platform
• Single computing resource runs
multiple applications
• Vehicle Management System for
X-47 J-UCAS
• Flight Management
• Flight Control
• Fuel, Power, Engine
Management
• C-130 AMP, KC-767 Tanker,
MMA, X-45 J-UCAS
• Boeing 787 Dreamliner
© 2005 by Smiths Aerospace: Proprietary Data
Smiths on-board networked systems on Next-generation
airliners: The Boeing 787 Dreamliner
Common data network
Common
computing
resource
Common core system
remote data concentrators
Enhanced airborne
flight recorder
Common data network
Common core system
remote data concentrators
The Smiths Common Core System (CCS)
is the central nervous system of the aircraft
© 2005 by Smiths Aerospace: Proprietary Data
Integrated Web-enabled HUMS Ground Support
Generic capability for aircraft and land vehicles
Meets deployment / non fixed base requirement for IVHM
Full range of IVHM functions & services
Remote
Download
Remote
Access
Windows
Groundstation
Data
Warehouse
Smiths Fault
Database
Smiths On-line Support Site
© 2005 by Smiths Aerospace: Proprietary Data
© 2005 by Smiths Aerospace: Proprietary Data
Lessons learned
Health & Usage Management has proven benefits in safety and
maintenance
New computing and communications provide the processing power
and data for comprehensive integrated vehicle health management
Existing health management functions are still heavily reliant on
people to provide prognostics, decision support and learning
Further development is required to improve:
• Prognostics
• Autonomous decision making
• Extraction of information from historic data
• Automatic capture of experiential data
© 2005 by Smiths Aerospace: Proprietary Data
New tools for data fusion, data mining and reasoning
Intelligent Management of HUMS
data
• CAA sponsored
• Effectiveness of AI techniques as a
method of improving fault detection in
helicopters
ProDAPS
• USAF sponsored
• Development of tools for PHM
• Application of tools to F-15 engine
Internal Development Activity
• Development of AI tools and techniques
• Application to
• Electrostatic engine data
• Flight Operational Quality
Assurance (FOQA)
© 2005 by Smiths Aerospace: Proprietary Data
ProDAPS component configuration for PHM
Ground-based
Reasoning
Diagnostics
Prognostics
Embedded
Reasoning
On-board components
applicable to in- dev. a/c
Diagnostics
Input to
Autonomous
Controls
Decision
Support
Recommended
actions
Autonomous
control
Data
Mining
Ground-based components
applicable to:
Legacy a/c
In-development a/c
Future a/c
Fleet
New knowledge
Anomaly models
On-board components
applicable to future a/c
© 2005 by Smiths Aerospace: Proprietary Data
ProDAPS
Positioned within the OSA-CBM
evolving Open System Architecture
standard
• ProDAPS provides high level intelligent
functions and capabilities to push
Health Monitoring to true IVHM/PHM.
Current capability gap, and key
target area for ProDAPS intelligent
systems tools, e.g.
•
Data fusion
•
Automated reasoning
•
Data mining (for empirical models)
Existing Smiths HUM systems
provide considerable functionality
in these areas.
7. Presentation Layer
6. Decision Reasoning
5. Prognostics
4. Health Assessment
3. Condition Monitor
2. Data Manipulation
1. Data Acquisition
© 2005 by Smiths Aerospace: Proprietary Data
Demonstration of ProDAPS data mining tool on
helicopter MRGB bevel pinion fault
1. Initial cluster model based on ‘healthy’ data
80% of all data (first 80% of flights for each
Gearbox A - 80% of all Data
gearbox)
4
20500
3
Score
20000
2
19500
19000
1
18500
0
0
2
4
6
8
10
No. of Clusters
Flight
Cluster
Gearbox B - 80% of all Data
MRGB Bevel Pinion
Gearbox C - 80% of all Data
4
4
3
3
2
2
1
469
433
397
361
325
289
253
217
181
145
73
109
1
37
1
0
0
Flight
2. Trend of faulty gearbox relative to
initial ‘anomaly’ cluster
Flight
3. Adaptive modelling to characterise ‘trending’ data
All data used
193
209
177
flight
Gearbox B
Gearbox B - All data used
157
145
133
121
97
109
85
73
61
25
13
6
5
4
3
2
1
0
1
491
456
421
386
351
Cluster
Cluster
Flight
316
281
246
211
176
141
106
71
36
37
34
31
28
25
22
19
16
13
7
10
4
1
0
1
6 per. Mov. Avg.
(Gearbox B)
100
6
5
4
3
2
1
0
37
Gearbox C - All data used
Gearbox C
200
49
300
-100
161
10
145
8
129
6
97
4
No. of Clusters
113
2
81
0
Gearbox A
65
21000
400
49
23000
22000
500
6
5
4
3
2
1
0
17
24000
Cluster
Score
600
33
Gearbox A - All data used
25000
1
Movement relative to Cluster 4 - Learnt on 80%
Flight
© 2005 by Smiths Aerospace: Proprietary Data
Smiths Aerospace
Future Integrated Information Systems Architecture
www.smiths-aerospace.com
© 2005 by Smiths Aerospace: Proprietary Data
Concept of On-board IVHM Operation
Vehicle Sensor Information
State Detection Data
Assess
Act
Adaptive Flight
Control System
IVHM
Control
Algorithms
High Level
Reasoning Engine
Surface
Control
Plan
Health
Assessment
Vehicle
Capabilities
Health Data
(Vehicle Subsystems
Health Data)
On-board Real-Time Replanning
Flight Management System
Mission
Planning
Flight Planning
© 2005 by Smiths Aerospace: Proprietary Data
Networked on-board and off-board IVHM System
Real Time Data
Acquisition
Anomaly
Detection
Data Fusion
Diagnostics
and
Prognostics
Mission
Information
Reasoning
and
Decision Component
On-board
Operation
Decision
Support
Components
Reasoning
Components
Data Mining,
Data Fusion
&
Analysis
Components
Data Warehouse
Off-board
Operation
© 2005 by Smiths Aerospace: Proprietary Data
Conclusions
Network Centric Operation requires vehicle health information in
order to achieve mission readiness goals whilst reducing logistic
support.
New architectures and network centric technologies will provide a
powerful framework for the exploitation, integration and distribution
of vehicle health information.
The use of AI techniques has shown considerable potential for
information extraction to meet the challenges of:
• Improved fault detection, diagnostics and prognostics
• Decision support, reasoning, data mining
• Improved payback through Optimal use of deployed assets
© 2005 by Smiths Aerospace: Proprietary Data