(Note:Conference Tutorial of PHM-Harbin is free. )
Dr. Nagarajan Raghavan
Assistant professor in Division of Engineering Product Development, Singapore University of Technology and Design, Singapore
Dr. Nagarajan Raghavan is the assistant professor in Division of Engineering Product Development, Singapore University of Technology and Design. His work focuses on reliability modeling of nano-devices, physics of failure modeling, maintenance engineering, design for reliability and prognostics and system health management.
There are several electrical and mechanical components in various systems such as transistors, ball bearings, solder bond pads, batteries etc. that have existed for a long time and have sufficient failure history and supporting studies on the root cause of these failures. For such cases, the failure mechanism may already be quite well known. With prior knowledge of the physics of failure, a good prognostic study should not just be purely data-driven, but should also consider a mathematical formulation of the failure mechanism that is inserted into the data-driven algorithm as the state update equation. This approach ensures a more robust and accurate estimation of the remaining useful life (RUL) of the component/system. It also enables RUL prediction under variable loading conditions. In this tutorial, we shall look at the detailed step-by-step sequence towards implementing a proper PoF based prognostic study. This includes the smart design of canary test structures, stress and measurement methodology, failure analysis, mathematical model of the failure mechanism from the degradation pattern and subsequent estimation of the RUL. The complete cycle of a PoF based prognostic implementation will be illustrated using a couple of case studies. Challenges and issues with PoF based prognosis will also be discussed in the talk and scope for further research in this area will be presented.
Dr. Bin Zhang
Assistant professor in College of Engineering and Computing, University of South Carolina, USA
Dr. Bin Zhang is the assistant professor in college of engineering and computing, University of South Carolina. His research interests include Prognostics and health management, which covers fault detection and isolation, failure prognosis, and fault tolerance robotics, unmanned systems, electromechanics, and industrial electronics, intelligent systems and control dynamic systems, design, modeling, simulation and control.
Fault diagnosis and prognosis (FDP) plays an important role in the modern complex industrial systems. Diagnosis aims to monitor the fault state in real-time while prognosis predicts the evolution of fault state and remaining useful life (RUL). Traditional Riemann sampling-based FDP (RS-FDP) takes samples and executes algorithms periodically and, in most cases, requires signiﬁcant computational resources, which makes it diﬃcult to be implemented on hardware with very limited computational capabilities. To overcome this bottleneck, a Lebesgue sampling-based FDP (LS-FDP), in which FDP algorithms are implemented “as-needed”. In LS-FDP, a set of Lebesgue states are defined on the state axis. The computation of LS-based diagnosis is triggered only when the value of measurements changes from one Lebesgue state to another, or “event-triggered”. This method signiﬁcantly reduces the computation demands by eliminating unnecessary computation. This LS-FDP design is generic and able to accommodate diﬀerent FDP algorithms. In this presentation, the design of LS-FDP and its application to engineering systems will be discussed in details. The efficiency of LS-FDP is verified by comparison with those of its RS-FDP counterparts.
Prof. Haitao Liao
Hefley Professor of Logistics and Entrepreneurship, Industrial Engineering Department, University of Arkansas, USA
Dr. Haitao Liao is a Professor and Hefley Endowed Chair Professor in Logistics and Entrepreneurship in the Department of Industrial Engineering at University of Arkansas. His research is focused on both experimental and analytical studies related to Reliability Engineering and Service Logistics. His research interests include: (1) reliability models, (2) maintenance and service logistics, (3) prognostics, (4) probabilistic risk assessment, and (5) analytics of sensor data.
Accelerated testing has been widely used in reliability estimation for highly reliable products. Among different accelerated testing techniques, accelerated degradation testing is probability the most powerful one. In such tests, the degradation measurements of a product or material under different conditions are collected over time and modeled using various methods. In this tutorial, different methods for modeling accelerated testing data will be introduced and some applications of the resulting models in fault diagnostics and prognostics will be presented.
Keynote Topic: PHM in Railways : Big Data or Smart Data ?
RAM (Reliability Availability Maintainability) Director and PHM (Prognostics & Health Management) Director, Alstom Transport, France
Dr. Pierre Dersin obtained his Ph.D. in Electrical Engineering in 1980 from the Massachusetts Institute of Technology (MIT). He is now RAM (Reliability-Availability-Maintainability) Director and PHM Director in Alstom. He has contributed a number of communications and publications in IEEE conferences and journals in the fields of RAMS, automatic control and electric power systems. He was just elected Vice President, Technical Activities, of the IEEE Reliability Society and is a member of the IEEE Future Directions Committee.
Prognostics & Health Management (PHM) has undergone a very fast development since the beginning of the new century and holds a promise for making a number of industrial systems, such as railway systems, both more reliable and more cost-effective in terms of maintenance. The impressive developments in data science in recent years (the “Big Data”) provide powerful tools for extracting useful and actionable information from data acquired from the field or from test benches. Purely data-driven approaches require no physical understanding and are quite flexible but do require large volumes of data (pertaining to both healthy and degraded conditions) and their performance is highly dependent on the quality of those data. Computational load can be very high. But railway suppliers have accumulated decades of know-how on the physics of their systems, both in normal and degraded conditions. This knowledge can be exploited to the fullest by designing “virtual prototypes”, i.e. multi-physics models of the actual systems. The key challenge is taking uncertainty into account, for instance uncertainty in future operating conditions. Hybrid approaches, i.e. combining knowledge of physical processes and information from sensor readings to enhance diagnostics and prognostics capabilities, seem to combine the advantages of both methods. Model predictions can be adjusted using measured data (either off-line or on-line). The above considerations will be illustrated on railway subsystems, such as HVAC (heating, ventilation and air-conditioning).
Senior Researcher, General Motors R&D, MI, USA; Senior Scientist, United Technologies Research Center, CT, USA
Dr. Ming Cao received Ph. D. on Mechanical Engineering from Pennsylvania State University in 2003. He spent the last decade working on PHM solutions including human-moving vehicle systems such as car, elevator, helicopter, as well as fixed-wing aircraft & tubo engine system and industry systems. Dr. Cao has worked for Fortune 500 companies, such as General Motors, United Technologies, and Ingersoll Rand throughout his career. He is the member of AIAA Sensor Technical Committee, and ASME Transportation Technical Committee. He served as ASME IDETC AVTT Conference Program Chair for 2 years, and guest Editor of International Journal of Vehicle Dynamics.
Condition Based Maintenance (CBM), Prognostics and Health Management (PHM), Reliability Centered Maintenance (RCM), and the analytics supporting all these catchy and sounds-too-good-to-be-true after-market service solutions, such as Big Data, Predictive Service, have become so loud it is literally impossible to ignore them. The big driver, of course, is customer demands. These days, customers have a lot of options; so unless the OEM provides the best reliability, product availability, coupled with lowest overall cost-to-own, you are not going to win over your customers. Commercial engines are no exceptions to this trend. Actually, due to the catastrophic nature and traumatizing media impacts to the public (not only the victims and their families) after jet engine failures, the reliability/safety requirements on commercial jet engines far exceed those of ground transportation systems. The giants of the industry, with names such as GE, Pratt Whitney, Rolls Royce, are all rushing to develop, launch, and market their own CBM / PHM service solutions. American Airline, the largest airline in the world, is now demanding “contracts on wing”, i.e., paying for flight hours, not for leasing / purchasing the engines. This will be even more true when everyone wants “contracts on wing”, which in the speaker's view is not that far away from becoming reality. "Contracts on wing" solutions truly push OEM's to provide real-time, cost-effective engine PHM solutions; and there is no other way around it. This speech will discuss the burning PHM needs & technical challenges the commercial engine industry is facing, and touch base on some engine PHM R&D activities.
Prof. John W. Sheppard
College of Engineering Distinguished Professor and RightNow Technologies Fellow; Director of Numerical Intelligent Systems Laboratory, Montana State University, USA
Dr. John Sheppard serves as a member of the IEEE Computer Society Standards Activities Board and is the Computer Society liaison to IEEE Standards Coordinating Committee 20 on Test and Diagnosis for Electronic Systems. He is also the co-chair of the Diagnostic and Maintenance Control Subcommittee of SCC20 and has served as an official US delegate to the International Electro-technical Commission's Technical Committee 93 on Design Automation. He is also an Adjunct Professor in the Department of Computer Science at Johns Hopkins University. In 2007, he was elected as an IEEE Fellow.
To date, most practical implementations of Prognostics and Health Management (PHM) have focused on the health management aspects with only minor attention being given to fault prediction. Commercial and military aerospace systems often employ a wide variety of embedded and on-board sensors to track and estimate the health of systems, and this information is being employed in frameworks such as reliability centered maintenance and condition based maintenance. Recently, organizations such as NASA and the US Army have started to explore incorporating formal methods for online risk assessment to guide interpreting health assessments and making system maintenance decisions. In this talk, I will introduce work being performed at Montana State University where we are coupling traditional model-based diagnostic and health assessment methodologies with continuous-time probabilistic methods to track and predict the impact of emerging hazards in a system using real-time, condition-based assessments of system health and the emergence of likely faults. Specifically, I will address how we have integrated dependency-based methods for fault diagnosis with fault tree analysis to derive and use continuous-time Bayesian networks as an end-to-end approach in real-time monitoring, tracking, predicting, and mitigating risks associated with system failure.
Keynote Topic:FPGA-based Machine Learning for Prognostics and System Health Management
Prof. Philip Leong
Professor of Computer Systems in the School of Electrical and Information Engineering at the University of Sydney, Visiting Professor at Imperial College
Prof. Philip Leong received the B.Sc., B.E. and Ph.D. degrees from the University of Sydney. In 1993 he was a consultant to ST Microelectronics in Milan, Italy working on advanced flash memory-based integrated circuit design. From 1997-2009 he was with the Chinese University of Hong Kong. He is currently Professor of Computer Systems in the School of Electrical and Information Engineering at the University of Sydney, Visiting Professor at Imperial College, Visiting Professor at Harbin Institute of Technology, and Chief Technology Advisor to ClusterTech.
Machine learning has improved to a point where it can achieve near-human accuracy on difficult tasks such as image recognition, speech recognition, and machine translation. However, efficient implementation of machine learning algorithms, particularly for real-time applications remains a challenge. This presentation will first detail the improved Energy, Parallelism, Interface and Customisation (EPIC) opportunities offered by FPGAs over conventional technologies such as microprocessors and graphics processing units (GPUs). These enable systems with smaller footprint, operating at low power and achieving improved functionality. Next, our recent research on FPGA-based implementations of machine learning algorithms will be described. This work includes high-speed and low-latency implementations of kernel adaptive filters, random projections and binarized convolutional neural networks. The talk will conclude with a discussion of applications of this technology to prognostics and system health management.
Keynote Topic: Reliability Assessment for Fleets of Systems and Family of Products: from Practice to Theory
Dr. Loon Ching TANG
Director of Temasek Defence Systems Institute and a full professor of Department of Industrial & Systems Engineering at the National University of Singapore
Dr. Loon Ching TANG obtained his Ph.D degree from Cornell University in the field of Operations Research in 1992 and has published extensively in areas related to industrial engineering and operations research. He has been presented with a number of best paper awards including the IIE Transactions 2010 Best Application Paper Award and 2012 R.A. Evans/P.K. McElroy Award for the best paper at Annual RAMS. Prof Tang is the main author of the award-winning book:Six Sigma: Advanced Tools for Black Belts and Master Black Belts. Besides being active in the forefront of academic research, in the last 25 years, Prof Tang has served as consultant for many organizations, such as the Ministry of Home Affair, Singapore Power, Republic of Singapore Air Force, Seagate, HP, Phillips, etc, on a wide range of projects aiming at improving organizational and operations efficiency; especially through better management of engineering assets. He is currently a fellow of ISEAM, the Editor of Quality & Reliability Engineering International and a member of the advisory board of the Singapore Innovation and Productivity Institute.
In this talk, we shall present a few industrial cases in which reliability assessments are needed for making decisions relating to fleet management, product design and replacement strategies. These cases highlight the use of engineering knowledge with statistical modeling for better decision making. In particular, we look into approaches in analyzing large fleets of repairable systems, how to improve product design risk and mitigate warranty costs. These examples motivate some interesting theoretical development of new ideas in the analysis of repairable systems and the planning of reliability testing.
Keynote Topic: The ever changing paradigm of qualification of electronic products – Physics of Failure vs. Standards Based Qualification
Quality and Reliability Program Manager, Intel Corporation, USA
Dr. Preeti Chauhan obtained her Ph.D. in mechanical engineering 2012 from CALCE, University of Maryland College Park. She is now served as the reliability engineer and server Q&R program manager in Intel Corporation, Chandler, Arizona. She is quality and reliability expert in areas of epoxy underfill, mold underfill, mid-level interconnects, stiffener, thermal compression bonding, and chip attach for mobile, laptops, desktops, and server processors.
This topic is based on Physics of Failure. Physics of failure based qualification is also known as Knowledge based qualification (KBQ), which is a different paradigm as compared to standards based qualification (SBQ) of electronic products. Prognostics is often used with physics of failure methods to provide a robust product qualification. This topic will cover that as well.
Christian K. Hansen
Past President, IEEE Reliability Society; Professor and Associate Dean, College of Science, Technology, Engineering and Mathematics, Eastern Washington University, USA
Dr. Christian K. Hansen recently served as President of the IEEE Reliability Society (2014-2016) and is currently the Associate Dean and Professor of Statistics in the College of Science, Technology, Engineering and Mathematics (CSTEM), Eastern Washington University (EWU). Over the past two decades he has been active with the IEEE Reliability Society and has served in leadership positions that include vice-president of publications and treasurer before being elected to president in 2013.