The tutorial session is scheduled for Monday, October 11, 2010. All tutorials are free to all registered PHM 2010 attendees.
Tutorial Title: Introduction to Diagnostics
Tutorial Presenter: Greg Kacprzynski, Impact Technologies
This tutorial focuses on the design, development, implementation, and evaluation of Diagnostics and Condition Based Maintenance (CBM) technologies. We will describe and demonstrate some of the leading PHM/CBM technologies currently being implemented in various health monitoring systems applications. In addition, the important concepts associated with verifying and validating these technologies using relevant performance metrics specifically tied to failure mode detection, fault isolation, and prediction requirements will be covered. We will provide practical insight into how organizations are currently deploying PHM/ CBM systems. These experiences may range from newly designed or legacy experiences in aircraft, other mobility applications, and land-based equipment.
Greg Kacprzynski is Director of Engineering Services and co-founder of Impact Technologies, LLC. Greg directs engineering teams that develop software and provide consulting support associated with system health management including diagnostic, prognostic and reasoning technologies, physics of failure modeling, intelligent control, and maintenance optimization. In his 15 years of experience in this domain, Greg has been technical lead on a wide variety of PHM/CBM related programs for both DoD and commercial applications and published more than 25 papers and journal articles. His education was at Rochester Institute of Technology (BS/MS) in Mechanical Engineering.
Tutorial Title: Introduction to Prognostics
Tutorial Presenter: Abhinav Saxena, Research Scientist, SGT at NASA Ames Research Center
This session will focus on the concepts and basics of prognostics from condition-based systems health management viewpoint. Participants will be introduced to a prognostic framework that will contrast the differences with other techniques and philosophies of prognostics used in other domains. Examples will be used to illustrate various types of prediction scenarios and what does it take to set up a desired prognostic system. This will include discussions on significance of run-to-failure data, requirements and specifications generation for prognostics, prediction algorithms, post prognostic reasoning, etc. The session, then, will go into the details of setting up a prognostics problem and algorithm development using data-driven and model based approaches including data preprocessing and feature extraction steps. Discussion on prognostic performance evaluation and performance metrics will conclude the technical discussion followed by a general discussion on open research problems and challenges in prognostics.
Abhinav Saxena is a Research Scientist with SGT Inc. at the Prognostics Center of Excellence NASA Ames Research Center, Moffett Field, CA. His research involves developing prognostic algorithms and methodologies to standardize prognostics that include performance evaluation and requirement specification for prognostics of engineering systems. He has been involved in PHM research for the last seven years and has published several papers on these topics. He is a PhD in Electrical and Computer Engineering from Georgia Institute of Technology, Atlanta. He earned his B.Tech. in 2001 from Indian Institute of Technology (IIT) Delhi, and a Masters Degree from Georgia Tech in 2003.
Tutorial Prepared by: Piero P. Bonissone, Chief Scientist, GE Global Research (Tutorial Presented by Neil Eklund)
After a brief introduction to model fusion (or ensemble learning), we show how its effectiveness depends on the diversity of the individual models. We describe Random Forest (derived from CART models) as a typical example of ensemble learning. Then, we focus on three key PHM functions, Anomaly Detection (AD), Diagnostics (D), and Prognostics (P). We analyze three real-world case studies to illustrate each PHM functions. We develop: (1) an AD model for simulated aircraft engines, based on the fuzzy selection of complementary models, locally trained auto associative neural networks; (2) a diagnostic model for simulated aircraft engines, based on local dynamic fusion of competing classifiers; (3) an ensemble of predictive models, based on local dynamic fusion of competing predictors of efficiency and emissions of a coal-fired boiler.
A Chief Scientist at GE Global Research, Dr. Bonissone has been a pioneer in the field of fuzzy logic, AI, soft computing, and approximate reasoning systems applications since 1979. He is a Fellow of IEEE, AAAI, IFSA, and a Coolidge Fellow at GE Global Research. He served as Editor in Chief of the International Journal of Approximate Reasoning for 13 years. He has co-edited six books and has over 150 publications. He received 50 patents issued from the USPTO (plus 51 pending). He has (co-)chaired 12 scientific conferences and symposia focused on Multi-Criteria Decision-Making, Fuzzy sets, Diagnostics, Prognostics, and Uncertainty Management in AI. In 2002, he was President of the IEEE Neural Networks Society (now Computational Intelligence Society). He has been an Executive Committee member of NNC/NNS/CIS society since the past 16 years and an IEEE CIS Distinguished Lecturer since 2004.
Tutorial Title: Uncertainty Management in Prognostics
Tutorial Presenter: Yongming Liu, Assistant Professor, Department of Civil and Environmental Engineering, Clarkson University
This tutorial focuses on the uncertainty management for physics/model-based prognosis. The discussed uncertainty management framework is applicable to many different prognostics problems and is demonstrated for fatigue damage prognosis of metallic materials/components. Four major components are covered in this presentation. First, accurate physics-based mechanism modeling is discussed which aims for reliable median/mean remaining useful life prediction. Following this, the uncertainty quantification and propagation methodology is discussed with special focus on efficient uncertainty propagation analysis. Next, a general Bayesian updating framework is discussed for model and uncertainty updating using diagnostics results. Finally, rigorous verification and validation is discussed. Prognostics metrics and Bayes factor are demonstrated for mode performance evaluation.
Dr. Yongming Liu is an assistant Professor in the department of civil and environmental engineering at Clarkson University. His research interests include fatigue and fracture analysis of metals and composite materials, probabilistic methods, computational mechanics, and risk management. He completed his PhD at Vanderbilt University, and obtained his Bachelors’ and Masters’ degrees from Tongji University in China. Dr. Liu is a member of ASCE and AIAA and serves on several technical committees on probabilistic methods and advanced materials. His group is current working on several projects for probabilistic prognosis sponsored by NASA, NSF, and FAA.
Tutorial Title: Accelerated Life Testing (ALT) in Electronics
Tutorial Presenter: Ephraim Suhir, Professor, Dept. of Electrical Engineering, Santa Cruz,
We address the role, objectives and attributes of, and the challenges and pitfalls in, the accelerated life testing (ALT) and highly accelerated life testing (HALT) in electronics and photonics engineering, including the ALT/HALT interaction with the product development testing and especially with qualification testing (QT). We illustrate the major concepts using, as suitable examples, assemblies and packages experiencing thermal loading. We elaborate on a novel approach of the probabilistic design for reliability (PDfR). This approach, based on the ALT/HALT conducted for the most vulnerable (weakest link) elements of an assembly, package, product or a subsystem of interest, complemented by extensive predictive modeling (PM) and sensitivity analysis efforts, enables one to predict, minimize and, if necessary, even control a low probability of failure in the field. We show also how such an ALT/HALT-PDfR-PM effort could be successfully employed in a new approach to qualification specification and testing of electronics and photonics products. The major concepts are illustrated by examples based primarily on the author’s published work, mostly in application to assemblies and structures experiencing thermal loading.
Dr. Suhir is Distinguished Member of Technical Staff (ret), Basic Research Area, Bell Labs, Murray Hill, NJ. He is currently on the faculty of the Electrical Engineering Department, University of California, Santa Cruz, CA, is Visiting Professor, Department of Mechanical Engineering, University of Maryland, College Park, MD and CEO of the ERS Reliability Engineering Co. Dr. Suhir is Foreign Full Member (Academician) of the National Academy of Engineering, Ukraine, and Fellow of the Institute of Electrical and Electronics Engineers (IEEE), the American Physical Society (APS), the American Society of Mechanical Engineers (ASME), the Institute of Physics (IoP), UK, and the Society of Plastics Engineers (SPE). He has authored about 300 technical publications (patents, papers, book chapters, books). He received many professional awards from ASME, IMAPS, IEEE and Bell Laboratories.
Tutorial Title: Verification & Validation for PHM
Tutorial Presenter Guillaume Brat, NASA Ames Research Center
This tutorial will discuss V&V techniques for prognostics and health monitoring applications, and especially applications relying on non-deterministic algorithms. The goal is to describe how current and advanced V&V techniques can help retire risks associated with the use of non-deterministic algorithms so that the PHM applications they support can be certified and deployed in the field. The tutorial will cover V&V techniques addressing the whole development lifecycle. For example, model checking helps in doing early V&V and mitigating design decisions in early design, static analysis can verify code implementing a PHM system, and, advanced testing techniques can validate a PHM system and gain insights in off-nominal cases.
Dr. Brat received his M.Sc. and Ph.D. in Electrical & Computer Engineering from The University of Texas at Austin. He is the deputy area lead for the Robust Software Engineering group at the NASA Ames Research Center ; he focuses on research and application of sound and complete static analysis (based on abstract interpretation) to the verification of large software systems. He also serves as the Principal Scientist on the V&V of Flight Critical Systems effort (funded under the Aviation Safety program in NASA ARMD), which conducts research in V&V techniques for Aerospace systems, including PHM applications.