They are seeking to maximize equipment uptime while minimizing costs. This course will focus upon an integrated hardware/software approach to machine health maintenance by introducing a systematic framework to failure mode and effects criticality analysis, database management and means to diagnose machine/component impending failure conditions and to prognose their remaining useful lifetime.

Attendees will be exposed to novel software and hardware architectures for failure mode and effects analysis, data processing and diagnostic /prognostic algorithms that will enable them to implement innovative and cost-effective health maintenance programs on their own critical processes; they will become familiar with flexible and open interface platforms and witness demonstrations of simulated and actual diagnostic systems; R&D agendas and future commercial and military applications of these technologies will be reviewed.

• Systems Approaches to PHM/CBM
• Failure mechanisms of critical systems
• Hardware and software requirements for data processing, Diagnostics and Prognostics
• Open system architectures and human system interface platforms
• Tools to represent and manage uncertainty
• Methods to evaluate performance and assess cost-effectiveness

• Manufacturing Research Center - PHM Laboratories
• Aerospace UAV Laboratories
• Mechanical Engineering Laboratories
• Ultrasonic Sensors Laboratory

• Georgia Tech Background in CBM/PHM
• Motivation and a Work Plan

The Systems Approach to PHM

• Trade Studies
• Failure Modes and Effects Criticality Analysis (FMECA)
• System Test Plan Design
• Comparison of Data Distributions/Statistical Measures
• Performance Metrics
• Verification and Validation of PHM Systems
• PHM Design Analysis (Impact Technologies, LLC)

Tools and Toolboxes for PHM Design

• Fuzzy Logic and Neural Networks
• Genetic Algorithms and Dempster-Shafer Theory

Sensors and Sensing Strategies

• Transducing Principles/Sensors
• Dynamic Metrology
• Data Base Management
• Data Pre-Processing/Data Mining, Data Fusion
• Feature Selection and Extraction

Fault Diagnostics -- Algorithms and Examples

• Introduction
• Physics Based Fatigue Models
• Model Based Techniques/Model Based Reasoning
• Fuzzy Logic and Neural Network Approaches
• Application Examples and Demonstrations
• Real-Time Diagnostics of a Mechanical Face Seal

Prognostics

• ARMA/ARIMA Methods
• Weibull Distribution and other Probabilistic Algorithms
• Dynamic Wavelet Neural Networks and Virtual Sensors
• Uncertainty Representation and Management
• Prognostics Methods and Systems -- Application Examples (Impact Technologies, LLC)

Prognostics Health Maintenance (PHM) and Condition Based Maintenance (CBM)

• An Intelligent Agent Based PHM Architecture
• Case Based Reasoning in PHM
• How Condition-Based Maintenance Changes Logistics

Performance Assessment

• Performance of Software Algorithms
• Economic Justification Methods

Open Systems Interface and Human System Interface Platforms

• Conventional and New Communication Platforms
• Open Systems Interface Platforms
• Human System Interface

Hardware Platforms

• Data acquisition, Communications and Computing Requirements

Test Cases

Fault-Tolerant Control

Open Discussion

• Issues and Concerns Raised by Participants

For additional information, contact the course coordinator: Dr. George Vachtsevanos School of Electrical and Computer Engineering Georgia Institute of Technology Atlanta GA 30332-0250 (404) 894-6252 E-mail: gjv@ece.gatech.edu

Mr. Gary O'Neill is a former naval aviator with over 15 years of experience in the Navy's aviation industrial facilities. He is currently developing research in data centered maintenance and logistics for the Georgia Tech Research Institute's Logistics and Maintenance Applied Research Center.

Dr. Kai Goebel received his M.S. and Ph.D. degrees from the University of California at Berkeley in 1993 and 1996, respectively. He is currently with GE Corporate Research and Development in Schenectady, NY. His research interests include monitoring, diagnostics, and prognostics, diagnostic information fusion, adaptive diagnostic systems, and soft computing. He has designed and fielded diagnostic systems for a variety of applications, e.g., aircraft engines, medical equipment, paper mills, etc.

Dr. George Hadden is a Senior Research Fellow at the Honeywell Laboratories and a recipient of the H. W. Sweatt award - Honeywell's highest technical honor - for a preventive maintenance expert system. He is the technical lead for two Condition Based Maintenance programs to detect and predict failures in jet engine components. In addition, he leads programs for the destructive testing of large mechanical equipment (Office of Naval Research) and the design of a failure prediction system for shipboard systems (Newport News Shipbuilding). Recently, he served as Principal Investigator of a project to perform Condition Based Maintenance on Navy shipboard equipment for the Office of Naval Research.

Dr. Michael J. Roemer serves as the Director of Engineering at Impact Technologies. He served previously as the Vice President of Engineering at STI Technologies. Dr. Roemer is the Principal Investigator for several programs related to advanced Machinery Health Monitoring and Decision Support Technologies funded by USAF, Navy, DARPA, Army and EPRI. He serves at the lead for the Boeing/Sonic Cruiser Team and FCS developing PHM for complex platform systems with logistics integration.

Carl S. Byington is the Manager for R&D at Impact Technologies where he directs research activities in PHM related programs funded by government agencies and private industry. HE was previously the Head of the Condition-Based Maintenance Department at Penn State's Applied Research Laboratory. Mr. Byington was the 2000 Technical Program Chairman of the Society for Machinery Failure Prevention Technologies Division of the Vibration Institute. He is currently serving as the Vice-Chairman of the CBM Committee of the ASME's Tribology Division.