Predictive Maintenance for Fuel Cell Systems FCS

1
Predictive Maintenance forFuel Cell Systems (FCS)
Dr. Haixia Wang Research Associate NSF
Industry/University Cooperative Research Center
on Intelligent Maintenance Systems
(IMS) University of Cincinnati Prof. Jay
Lee Ohio Eminent Scholar L.W. Scott Alter Chair
Professor Director of the NSF Industry/University
Cooperative Research Center on Intelligent
Maintenance Systems (IMS) University of Cincinnati
2
Unmet needs in FCS maintenance

• FCS is a complex system including the
  interactions of mechanical, chemical, and
  electrochemical subsystems
• no mature, well-developed diagnostic tools in
  existence for trouble shooting
• no trained business to support field service of
  fuel cells

Fuel Cell Systems
Mechanical components
Chemical components
Sensors/ detectors
Troubleshooting
Trained personnel in field service businesses
3
Unmet needs in FCS maintenance

• Lack of historical data since FCS technology is
  in its infancy
• No data collection plan for the purpose of
  predictive maintenance
• Membrane temperature?
• air flow speed ?
• Hydrogen pressure stack inlet ?
• Air pressure stack inlet ?
• .?

FC designer
FC Manufacturer
FC service
FC consumer
4
Gaps and unmet needs in the area of FCS
maintenance
Lack of instruments and standard tools for
diagnostics and health monitoring
Fuel Cell System
On Board Diagnostic and Prognostic
Standard maintenance and troubleshooting
instruments, adaptable to all vehicles
5
IMS Proposed Approach
Select fuel cell systems and measurement tools
from company members
Gather requirements from different players
(manufacturers, instrumentation companies, users,
etc)
Develop maintainability analysis/failure mode
analysis targeted to diagnostic and prognostic of
FCV using FMECA
Identify failure modes, detection technique for
FCV from company members
Diagnostic
Prognostic
Define measurement requirements for diagnostic
tools
Modify one of the IMS tools for monitoring fuel
cell health degradation
6
Provide requirements to instrumentation members
for modifying/developing measurement tools for
diagnostic and troubleshooting
Modify one of the IMS watchdog systems for FCV
monitoring
Demonstrate the developed instruments and
diagnostic tools in real environment of car
manufacturing company members and field service
industry
Define standards for all stages of data
collection/input/output for FCV
Approve standard by consortium members
Test products for suppliers and working with Ohio
state government for standard development
Develop patents to protect the intellectual
property of consortium
7
Introduction Fuel Cell Failure Modes

• Membrane cracks and tears
• Membrane pinhole blisters
• High membrane degradation rate
• Reduced fuel efficiency and thermodynamic
  efficiency
• Anode catalyst poisoning by carbon monoxide
• Other MEA failure modes can be identified through
  accelerated failure tests and analysis

8
Introduction Available Monitoring Techniques

• http//www.calce.umd.edu/general/Facilities/eds.ht
  m
• scanning electron microscopy (SEM),
• energy dispersive spectrometry (EDS),
• cyclic voltammetry (CV),
• various contact angle techniques

9
Value Shift Prognostics at IMS Center
Information
Data
Degradation
Failure
10
IMS Prognostics - An Analogy
Data Streamlining
Food
Nutrition
11
Watchdog Agent Technique
Working Process
Signal Collection Feature
Extraction
Provide warning
Most Recent Behavior
Normal Behavior

Performance/Health Prediction
Health Feature Radar Chart
Low humidity
Performance/Health Assessment
12
IMS Methodology 5S Approach
13
Data Streamlining Summary
Maintenance Database
No Maintenance Record
Expert Knowledge
Data Cleaning, Filtering and Sorting
Smart Processing
Identify Critical Components
Synchronize
Expert Knowledge
Data Collection
Standardize
Data Cleaning and Reduction
Determine a Clean and Reduced Data Set
Sustain
14
Watchdog Agent Toolbox
14
15
Automatic Algorithm Selection
QFD Quality Function Deployment AHP
Analytical Hierarchy Process
15
16
Brief Introduction to the IMS Center
17
Introduction to the IMS Center Smart Informatics
to Excel Productivity
NSF I/UCRC since 2000
Jay Lee, Jun Ni, Jag Sarangapani Univ. of
Cincinnati, Univ. of Michigan, Univ. of
Missouri-Rolla NSF Industry/University
Cooperative Research Center on Intelligent
Maintenance Systems www.imscenter.net
18
The IMS Consortium

 
19
IMS Mission Statement

• To enable products and systems to achieve and
  sustain near-zero breakdown performance, and
  ultimately transform machine condition data to
  useful information for improved productivity and
  asset utilization.

20
IMS System-of-System Approach
Near 0 Downtime
Design for
Just-in-Time Service
Health Monitoring
Product or System
Reliability and
Serviceability
Sensors Embedded
In Use
Intelligence
Product
Product
Center
Redesign
Smart
Condition-based Maintenance (CBM)
Design
Communications
Self
-
Maintenance

Tether
-
Free

Redundancy
(Bluetooth)

Active

Internet

Passive

TCP/IP
Enhanced
Six
-
Sigma
Design
Watchdog Agent is Trademarks of IMS Center
21
Focus Areas of Each Site
22
Reconfigurable Platform
IMS Infotronics Platform
23
Readiness of Commercial Platform
CompactRIO reconfigurable embedded system
UNO-2160 Watchdog Agent System

• Robust Embedded OS
• Windows CE.NET
• Windows XP embedded

24
Reconfigurable Prognostics
Reconfigurable prognostics systems are mandatory
for cost-effective implementation
24
25
Value Flow
Risk
Watchdog Agent
IMS Platform
Decision Support Tools
Company Designated Projects and Implementations
Validation
(Shared Tools Knowledge)
Research Tools
Testbeds
IMS Projects
Contracts
Value
Membership
26
IMS Impacts to Companies Rapid Prognostics
Development and Deployment
RISK COST
Today
TIME
RISK COST
TIME
27
Questions?
28
For More InformationPlease visit
www.imscenter.net

• Thank
• You !

29

• Extra Slide

30
Challenges

• What to monitor?

• Need to analyze the characterization of the
  component microstructures, and their physical,
  chemical, electrochemical, and interfacial
  properties
• Membrane temperature?
• air flow speed ?
• Hydrogen pressure stack inlet ?
• Air pressure stack inlet ?
• .?