One of the unique features of the PHM conferences is the free technical tutorials on various topics with comprehensive introduction to the state-of-the-art. The proposed tutorials address the interests of varied audience: beginners, developers, designers, researchers, practitioners, and decision makers who wish to learn a given aspect of PHM.
“Introduction to Diagnostics”
Prof. Hyunseok Oh
School of Mechanical Engineering Gwangju Institute of Science and Technology Korea
Bio: Hyunseok Oh received the B.S. degree from Korea University, Seoul, South Korea, in 2004, the M.S. degree from the Korea Advanced Institute of Science and Technology (KAIST), Daejeon, South Korea, in 2006, and the Ph.D. degree from the University of Maryland, College Park, MD, USA, in 2012. He is currently an Assistant Professor with the School of the Mechanical Engineering, Gwangju Institute of Science and Technology, Gwangju, South Korea. Dr. Oh received the A. James Clark Fellowship in 2007, the IEEE PHM Data Challenge Competition Winner in 2012, the PHM Society Data Challenge Competition Winner in 2014 and 2015, and the ACSMO Young Scientist Award in 2016.
“Prognostics 101: Concept, Methods, Issues and Applications”
Dr. Dawn An
Daegyeong Division/Aircraft System Technology Group Korea Institute of Industrial Technology Korea
Bio: Dawn An received the B.S. and M.S. degrees of mechanical engineering from Korea Aerospace University in 2008 and 2010, respectively. She started a joint Ph.D. at Korea Aerospace University and the University of Florida in 2011, and received her Ph.D. in 2015 as a jointly conferred degree. After she got her degree, she worked as a postdoctoral associate at the University of Florida for one year. She is now a senior researcher at Korea Institute of Industrial Technology. Her current research is focused on prognostics and health management not only for a machine but also the human body.
“Deep Learning for PHM”
Prof. Seungchul Lee
Mechanical Engineering Ulsan National Institute of Science and Technology Korea
Bio: Seungchul Lee is an assistant professor at the school of Mechanical and Nuclear Engineering at UNIST, Korea. His research focuses on machine learning for an anomaly detection, stochastic modeling for manufacturing processes and system decision support tools with mechatronics applications. He extends his research work to developing self-sustainable systems via an intelligent, informatics, and IoT system design at UNIST. He received the BS (2001) and PhD (2010) in Mechanical Engineering from Seoul National University, Korea, and from the University of Michigan, USA, respectively.
“Degradation Modeling and Analysis for Prognostics and Health Management”
Prof. Tao Yuan
Department of Industrial and Systems Engineering Ohio University USA
Bio: Dr. Tao Yuan is an assistant professor in the Industrial and Systems Engineering Department at Ohio University. He received the Bachelor of Engineering degree in Thermal Engineering from Tsinghua University with honorsin 2000. He obtained the Master of Science degree in Aerospace Engineering in 2003 and the Master of Engineering degree in Industrial Engineering in 2004 from Texas A&M University, College Station, and the Ph.D. degree in Industrial Engineering from The University of Tennessee, Knoxville in 2007. He joined the Industrial and Systems Engineering Department of Ohio University in September 2008. His research interests are in yield, quality, and reliability issues related to nanoelectronics manufacturing, fatigue behavior of advanced engineering materials, applied statistics and stochastic processes, and Bayesian statistics.
“Uncertainty Management in Prognostics”
NASA Ames Research Center
Bio: Shankar Sankararaman received his B.S. degree in Civil Engineering from the Indian Institute of Technology, Madras in India in 2007 and recently obtained his Ph.D. in Civil Engineering from Vanderbilt University, Nashville, Tennessee, U.S.A. in 2012. His dissertation titled "Uncertainty Quantification and Integration in Engineering Systems" focuses on the various aspects of uncertainty quantification in civil, mechanical, and aerospace systems. His research interests include probabilistic methods, risk and reliability analysis, Bayesian networks, system health monitoring, diagnosis and prognosis, decision-making under uncertainty, treatment of epistemic uncertainty, and multidisciplinary analysis. He is a member of Non-Deterministic Approaches (NDA) technical committee at AIAA and a member of Probabilistic Methods Technical Committee (PMC) at ASCE. Currently, Shankar is a researcher with Stinger and Ghaffarian Technologies (SGT Inc.) at NASA Ames Research Center, Moffett Field, CA, where he develops algorithms for uncertainty assessment and management in the context of system health monitoring, prognostics, and decision-making.
Prof. Tao Yuan
Department of Industrial and Systems Engineering
Trends and Recent Advances of
Industrial Big Data Analytics and Cyber Physical Systems
for PHM Applications
This tutorial session will discuss commonly used degradation modeling and analysis methods
such as stochastic process models, general degradation path models, maximum likelihood
methods, and Bayesian methods. Random-effect modeling and hierarchical Bayesian modeling
approaches will be discussed. Methods for deriving failure-time distributions and remaining useful
life distributions will be introduced. Various examples and case studies will be used to illustrate the
models and methods.
Prof. Seungchul Lee
Ulsan National Institute of Science and Technology,
Deep Learning for PHM
Deep learning, considered as one of the breakthrough technologies in machine learning in recent years, has attracted tremendous research attention in both academia and industrial communities. It involves learning good representations of data through multiple levels of abstraction, and can discover complicated underlying structure and features, thus achieving an improved predictive performance. As a result, PHM community also starts to apply deep learning technologies to PHM applications. In this tutorial, I will first provide an overview of deep learning technology including NN, CNN, RNN, and Autoencoder. In particular, the basic python codes with the TensorFlow library will be shared and explained during the tutorial. This tutorial can help more researchers in the PHM community to understand the philosophy of deep learning and to utilize the provided codes in their practices. Ultimately I hope this tutorial can stimulate more research interests towards deep learning technology within our PHM community.
Uncertainty plays an important role in prognostics and health management. This is because, prognostics deals with the future prediction of engineering systems and it is almost impossible to make precise predictions regarding the future. In fact, it will be meaningless to make predictions without incorporating any uncertainty regarding the future behavior of the system. A good prognostic algorithm needs to be able to identify and assess the various sources of uncertainty that affect prognostics, systematically compute the joint effect of these sources of uncertainty and quantify the resultant overall uncertainty in the remaining useful life prediction. This tutorial will discuss various activities such as uncertainty quantification, propagation, and management in the context of prognostics and remaining useful life prediction. Walkthroughs of several algorithms that assist in these activities will be provided, and these algorithms will be illustrated with numerical examples pertaining to industrial applications.
Dr. Dawn An
Daegyeong Division/Aircraft System Technology Group,
Korea Institute of Industrial Technology,
Prognostics 101: Concept, Methods, Issues and Applications
Prognostics is to predict future degradation behavior and remaining useful life of in-service systems based on condition monitoring data. In general, prognostics methods can be categorized into physics-based and data-driven approaches. There are several issues to be considered for practical prognostics. In physics-based approaches, prediction results can be accurate with a small number of data by utilizing a physical degradation model. Since, however, physical models are rare in practice, data-driven approaches are commonly employed. Usually many sets of data are required for the approaches, but it is very expensive to obtain the data in time and cost. Before the prognostics approaches are considered, there is a fundamental problem that degradation feature should be extracted first from indirectly measured signals. In this tutorial, these issues are discussed in addition to introducing the concept, methods and applications of prognostics.
Prof. Hyunseok Oh
School of Mechanical Engineering,
Gwangju Institute of Science and Technology,
Introduction to Diagnostics
Diagnosis for engineered systems can be defined to detect, locate, and identify faults. For safety-related and mission-critical systems, it is extremely critical to detect and identify any types of potential faults as early as possible and avoid dangerous situations. This tutorial provides an overview of fault diagnosis including fault detection, fault isolation, and fault identification. Basics and illustrative examples will be explained. Some case studies will be also presented such as power plants, urban railway, and manufacturing equipment. In this tutorial, participants can expect to learn the fundamentals of fault diagnosis for engineered systems. This tutorial may be a prerequisite lecture to understand other advanced topics such as prognostics, deep learning, degradation modeling, and uncertainty management.