Invited Speeches

Ke Feng
Xi'an Jiaotong University, China

Title: Digital Twin-driven Health Management and Remaining Useful Life Prediction of the Gearbox Transmission System
Abstract: The gearbox transmission system plays a vital role in advanced manufacturing, aerospace, renewable energy, vehicle, and mining system. Its degradation and failure would cause unexpected economic loss and even serious accidents. For example, the degradation and failure of the gearbox will impair the performance of the machine tool, affecting the production quality and quantity significantly and resulting in enormous economic loss. Therefore, monitoring the health condition of the gearbox transmission system is of great significance. However, the gearbox transmission system usually operates in harsh working environments, and it is difficult to conduct regular manual inspections and maintenance. Thus, the use of advanced online algorithms to monitor the degradation status of the gearbox transmission system and predict its remaining useful life (RUL) can bring significant benefits to industry practices. Digital twin (DT) is a virtual representation (mirror) of a physical structure or a system in real space along its lifecycles. Through real-time interaction between the virtual model and physical structure, the degradation status of the system and its RUL can be reflected and evaluated effectively. Thanks to its unique specialty, DT has recently received considerable attention from the research community. However, due to the complex structures and harsh operation conditions, research on DT-based gearbox transmission system RUL prediction is limited. Moreover, existing conceptual approaches have limitations in indicating the specific contact status and providing insights into the degradation stages of gearbox transmission systems, which greatly benefit RUL prediction.

Biodata: Ke Feng is a Full Professor at Xi’an Jiaotong University, China. He is a Marie Curie Fellow (Imperial College London & Brunel University London). He received a Ph.D. degree from the University of New South Wales, Australia, in 2021. He worked at the University of British Columbia and the National University of Singapore in 2022 and 2023, respectively. His main research interests include digital twins, vibration analysis, structural health monitoring, dynamics, tribology, signal processing, and machine learning. He is recognized as the Emerging Leader (2023) by the Measurement Science and Technology journal. He is the Associate Editor and Guest Editor of several journals, including Mechanical Systems and Signal Processing, IEEE Transactions on Industrial Cyber-Physical Systems, Journal of Intelligent Manufacturing, Structural Health Monitoring, Engineering Applications of Artificial Intelligence, IEEE Transactions on Instrumentation and Measurement, Neurocomputing, Measurement, IEEE Sensors Journal, IET Collaborative Intelligent Manufacturing, Measurement Science and Technology, Advances in Manufacturing, Proc Inst Mech Eng B J Eng Manuf, Journal of Central South University, etc.

Weiwen Peng
Sun Yat-sen University, China

Title: Uncertainty-aware Deep Learning Based Health Prognostics: A Bibliography Review and Some Demonstrations
Abstract: Deep learning based health prognostics is receiving ever-increasing attention. Most existing methods leverage advanced neural networks for prognostics performance improvement, providing mainly point estimates as prognostics results without addressing prognostics uncertainty. However, uncertainty is critical for both health prognostics and subsequent decision making, especially for safety-critical applications. In this talk, we are trying to deliver a comprehensive introduction of uncertainty-aware deep learning based health prognostics, including major motivation of uncertainty quantification in deep learning prognostics, main research line with bibliographical analysis, and some demonstrations with recent works in our research group. The demonstrations mainly focus on Bayesian deep learning realized through Monte Carlo Dropout and variational inference. In particular, Bayesian deep learning with multi-scale convolution neural network, bi-directional neural network, multibranch neural network, Bayesian transfer learning and Bayesian deep active learning are presented through applications to health prognostics of bearings, lithium-ion batteries and turbofan engines. Strengths and shortcomings of uncertainty-aware deep learning prognostics are summarized and highlights for future research are presented as well.

Biodata: Weiwen Peng is an Associate Professor in the School of Intelligent Systems Engineering, Sun Yat-sen University, China. Dr. Peng received the B.E., M.E. and Ph.D. degrees in Mechanical Engineering from the University of Electronic Science and Technology of China, Chengdu, China, in 2009, 2012 and 2015, respectively. From 2014 to 2015, he worked as a Visiting Graduate Researcher in University of California, Los Angeles. From 2016 to 2019, he worked as a Research Fellow in National University of Singapore, Singapore. Dr. Peng’s research interests include system reliability, Bayesian machine learning, prognostics and health management.

Jiawen Hu
University of Electronic Science and Technology of China, China

Title: Optimal Replacement Policy for K-out-of-N Systems with Periodic Imperfect Inspections
Abstract: The K-out-of-N system structure is widely used in industrial and military applications for its high reliability, and condition-based maintenance is a promising way to enhance maintenance efficiency. While existing studies on inspection and replacement policies often assume perfect inspections, real-world inspections can be imperfect due to noise and disturbances. This study investigates an optimal replacement policy for K-out-of-N systems with periodic imperfect inspections. We model component degradation with a three-state continuous-time Markov chain and use a state-observation matrix for imperfect inspections. We first analyze a 1-out-of-2: G system using a partially observable Markov decision process to derive the optimal policy that minimizes long-run discounted maintenance costs. The resulting maintenance policy is a twodimensional control-limit policy, determined using a value iteration algorithm. We then extend this model to a general K-out-of-N: G system and address the curse of dimensionality with a point-based value iteration method. A numerical study and sensitivity analysis demonstrate the effectiveness of our proposed policy.

Biodata: Jiawen Hu is an associate professor in School of Astronautics and Aeronautic, University of Electronic Science and Technology of China, Chengdu, China. He received the B.S. degree in mechanical engineering from Shanghai Jiao Tong University, Shanghai, China, in 2009, the M.S. degree in mechanical engineering from Chinese Academy of Sciences, Beijing, China, in 2012, and the Ph.D. degree in industrial engineering from Shanghai Jiao Tong University, Shanghai, China, in 2017. He was a research fellow with the Department of Industrial Systems Engineering and Management, National University of Singapore from 2017 to 2020. His research interests include maintenance optimization, degradation modeling. His work has appeared in journals including IEEE Transactions on Industrial Informatics, IISE Transactions, IEEE Transactions on Reliability, Reliability Engineering & System Safety, Journal of Manufacturing Systems, International Journal of Production Research.

Jiateng Yin
Beijing Jiaotong University, China

Title: Disruption Management of Urban Rail Transit Systems - A Resilience Optimization Approach from the Passengers’ Perspective
Abstract: Urban rail transit (URT) systems are exposed to risks of natural disasters and man-made disruptions, which not only cause train delays and overcrowding of platforms, but can even lead to serious safety incidents. Therefore, it is crucial to design a resilient URT system that provides commuters with as reliable a service as possible. This study first proposes to assess the resilience of URT systems with the emerging virtual coupling technologies, which allow trains to be coupled/decoupled dynamically. In contrast to the existing studies, where the system resilience is usually quantified through a resilience triangle that reflects the recovery of system performance, our study proposes a novel resilience assessment metric of URT from the perspective of passengers, which depicts the quick evacuation of delayed crowds in case of disruptions. To further enhance the system resilience, we focus on utilizing train platoons to transport the delayed passengers and we construct a mixed integer programming (MIP) model to optimize the rescheduling decisions of train platoons. Considering different types of disruptions in practice (e.g., different occurrence time of disruptions, disrupted sections), we develop a two-step simulation-optimization approach to solve the MIP model, which simulates the influences of random disruptions on passengers and then optimizes the rescheduling decision of train platoons in an iterative scheme. Finally, the proposed approach is validated using real-world data from the Beijing Metro. Through quantifying for a large set of disruptions, the simulation results demonstrate that the involvement of train platoons can effectively enhance the system resilience of URT against unexpected disruptions.

Biodata: Jiateng Yin is a professor at the School of Systems Science, Beijing Jiaotong University. He is a recipient of the National Science Fund for Excellent Young Scholars and was selected for the Youth Talent Support Program by the China Association for Science and Technology. He received his Bachelor's and Ph.D. degrees from Beijing Jiaotong University in 2012 and 2018, respectively. He was a visiting scholar with WISC-Madison, MIT and NUS. His primary research areas include train operation control, operational optimization management in rail transit, and artificial intelligence. He has published over 50 papers in journals such as Transportation Research Part B/C/E, Journal of Operations Research, European Journal of Operational Research, Networks and Spatial Economics, and Reliability Engineering & System Safety, with more than 2,800 citations. Five of his papers have been selected as ESI Highly Cited or Hot Papers, and he has been recognized as one of the world’s top 2% scientists.

Guangxin Jiang
Harbin Institute of Technology, China

Title: Adaptive Importance Sampling for Efficient Stochastic Root Finding and Quantile Estimation
Abstract: In solving simulation-based stochastic root-finding or optimization problems that involve rare events, such as in extreme quantile estimation, running crude Monte Carlo can be prohibitively inefficient. To address this issue, importance sampling can be employed to drive down the sampling error to a desirable level. However, selecting a good importance sampler requires knowledge of the solution to the problem at hand, which is the goal to begin with and thus forms a circular challenge. We investigate the use of adaptive importance sampling to untie this circularity. Our procedure sequentially updates the importance sampler to reach the optimal sampler and the optimal solution simultaneously, and can be embedded in both sample average approximation and stochastic approximation-type algorithms. Our theoretical analysis establishes strong consistency and asymptotic normality of the resulting estimators. We also demonstrate, via a minimax perspective, the key role of using adaptivity in controlling asymptotic errors. Finally, we illustrate the effectiveness of our approach via numerical experiments.

Biodata: Guangxin Jiang received his Ph.D. in Applied Mathematics from Tongji University, China, in 2015. He is currently a Professor at Harbin Institute of Technology (HIT) and an Adjunct Professor at the University of Science and Technology of China. Before joining HIT, he was an Assistant Professor at Shanghai University. His research focuses on simulation analytics and machine learning, with applications in financial engineering and risk management, FinTech, manufacturing, and supply chain management. He has published on academic journals such as Operations Research, INFORMS Journal on Computing, IEEE Transactions on Automatic Control, IISE Transactions. He currently serves as an Editorial Board Member for the Asia-Pacific Journal of Operational Research and as a Youth Editorial Board Member for Fundamental Research.

Qingqing Zhai
Shanghai University, China

Title: A Wiener Process Model for Degradation Modeling of Heterogeneous Populations
Abstract: Degradation is an important phenomenon for industrial products, which manifests as the gradually deterioration of some performance characteristics. The recent decades have witnessed a fast growth in the number of studies in statistical degradation modeling, due to the increasing accessibility of degradation data. Stochastic process serves as an important tool for degradation modeling, because it naturally captures the inherent randomness in the degradation phenomenon and can be easily extended to characterize other aspects, such as heterogeneity of the degradation data. I would like to introduce a Wiener process model for degradation modeling of heterogeneous populations and the related estimation and applications.

Biodata: Qingqing Zhai received the Ph.D. degree in systems engineering from Beihang University, Beijing, China, in 2015. He is currently an Associate Professor with the School of Management, Shanghai University, Shanghai, China. He has authored or coauthored more than 40 articles including JASA, Technometrics, IISE Transactions and IEEE Transactions. His main research interests include statistical degradation modeling, probabilistic reliability modeling, and game theories in reliability.

Tangfan Xiahou
University of Electronic Science and Technology of China, China

Title: Physics-Informed Neural Network for Health Status Assessment of Control Rod Drive Mechanisms of Nuclear Reactors
Abstract: Control rod drive mechanism (CRDM) is a pivotal component in nuclear reactors, responsible for regulating the reactor’s power output and ensuring its safe shutdown. Its reliability is paramount for the economic and safe operation of the reactors. This talk introduces a novel approach to CRDM health status assessment using physics-informed neural networks (PINN). The proposed CRDM-PINN method leverages simulation experiments and current data, combining prior physics knowledge to construct a health status assessment model. It not only provides accurate predictions of CRDM motion indicators but also reduces the need for extensive data collection. The CRDM-PINN model’s accuracy is demonstrated through case studies based on simulation data and real condition monitoring data, showing that it outperforms surrogate models that do not incorporate the physics information.

Biodata: He is currently an Associate Professor with the School of Mechanical and Electrical Engineering, University of Electronic Science and Technology of China. He received his Ph.D. degrees in mechanical engineering from UESTC in 2022. He has published more than 30 peer-reviewed papers in international journals. His research interests include reliability modeling under uncertainty, Dempster-Shafer evidence theory, and prognostics and health management (PHM). He has received several awards, including Excellent Doctoral Dissertation in Management Science, IEEE reliability society Student Achievement Award, IISE annual conference 2022 QCRE Best Track Paper Award Finalist. He serves as the guest editor of Reliability Engineering and System Safety, and editorial boards for International Journal of Reliability Safety and Industrial Engineering in Chinese. He is also a member of IISE and IEEE.

Lu Jin
University of Electro-Communications, Japan

Title: Condition-Based Maintenance for Systems Under External Shocks
Abstract: System failures and malfunctions arise from both internal deterioration and external shocks. This talk focuses on optimizing condition-based maintenance policies for systems subjected to these external shocks. We first examine a single-unit system that deteriorates according to a discrete stochastic process, considering the interactions between internal deterioration and external shocks. The analysis is then extended to a multi-unit system, where each unit’s deterioration follows a continuous stochastic process. Using a Markov decision process, we optimize the maintenance policies for both systems with the goal of minimizing the total expected discounted cost. Under reasonable assumptions, we derive properties of the optimal maintenance policies, which can facilitate the determination of the optimal solution. Finally, numerical examples are provided to illustrate these policies.

Biodata: Lu Jin is currently an Associate Professor in the Department of Informatics, University of Electro-Communications, Tokyo. Her current research interest includes system reliability, maintenance planning and optimization, and decision-making with uncertainties. She received the Outstanding Young Scientist and the Best Paper Award of IEEE Reliability Society (Japan Chapter) in 2006, the Nikkei QC Literature Prize in 2014, and several paper awards as a co-author. She is a member of the Japanese Society for Quality Control (JSQC) and Reliability Engineering Association of Japan (REAJ), the Institute of Electronics, Information and Communication Engineers (IEICE), Japan Society of Maintenology, and IEEE.

Qiuzhuang Sun
The University of Sydney, Australia

Title: Optimal Abort under Imperfect Condition Monitoring
Abstract: While most on-demand mission-critical systems are engineered to be reliable to support critical tasks, occasional failures may still occur during missions. To increase system survivability, a common practice is to abort the mission before an imminent failure. We consider optimal mission abort for a system whose deterioration follows a general three-state (normal, defective, failed) semi-Markov chain. The failure is assumed self-revealed, while the healthy and defective states have to be predicted from imperfect condition monitoring data.
Due to the non-Markovian process dynamics, optimal mission abort for this partially observable system is an intractable stopping problem. For a tractable solution, we introduce a novel tool of Erlang mixtures to approximate non-exponential sojourn times in the semi-Markov chain. This allows us to approximate the original process by a surrogate continuous-time Markov chain whose optimal control policy can be solved through a partially observable Markov decision process (POMDP). We show that the POMDP optimal policies converge almost surely to the optimal abort decision rules when the Erlang rate parameter diverges.
This implies that the expected cost by adopting the POMDP solution converges to the optimal expected cost. Next, we provide comprehensive structural results on the optimal policy of the surrogate POMDP. Based on the results, we develop a modified point-based value iteration algorithm to numerically solve the surrogate POMDP. We further consider mission abort in a multi-task setting where a system executes several tasks consecutively before a thorough inspection. Through a case study on an unmanned aerial vehicle, we demonstrate the capability of real-time implementation of our model, even when the condition-monitoring signals are generated with high frequency.

Biodata: Qiuzhuang Sun received the joint B.E. degree in industrial engineering and in computer science from Shanghai Jiao Tong University in 2015 and the Ph.D. degree in industrial and systems engineering from the National University of Singapore in 2019. He is currently a Lecturer (Assistant Professor) in the School of Mathematics and Statistics, the University of Sydney. His research interests include maintenance modeling, degradation analysis, and data-driven optimization.

Yifan Cui
Zhejiang University, China

Title: Fiducial Inference in Survival Analysis
Abstract: In this talk, we introduce novel nonparametric and semiparametric fiducial approaches to censored survival data. We propose Gibbs samplers and establish Bernstein-von Mises theorems. We also demonstrate our estimators by extensive simulations and real data applications.

Biodata: Yifan Cui is a Faculty Member of Data Science at Zhejiang University. Before joining ZJU, he was an Assistant Professor at Department of Statistics and Data Science, National University of Singapore. He has worked at Wharton School of the University of Pennsylvania as a Postdoctoral Researcher and obtained his Ph.D. from University of North Carolina at Chapel Hill. He is an elected member of the International Statistical Institute.