Prof. Zuomin Dong

Prof. Zuomin Dong

University of Victoria, British Columbia, Canada

Integrating Data-Driven Modeling, Design and Control Optimizations for Next Generation Clean Propulsion Systems of Vehicles and Marine Vessels

09:15-09:50 PDT, May 10, 2021

Abstract:
The rapid advance of numerical modelling, Model Based Design (MBD), metamodel-based global optimization, and intelligent system techniques provides new opportunities for design optimization and optimal control of complex mechatronic and embedded systems, such as electrified vehicles and marine vessels. Dedicated computer models are used to predict the performance of these systems through computation-intensive numerical simulations. In contrast to the subjects of the traditional engineering design of a standalone mechanical or electrical structure, these multiphysics systems consist of many sub-components of different types, including mechanical and electrical systems, electrochemical energy conversion process, and system controls. These systems present unlimited states of operation and rely on collaborative design and control optimization to achieve the best overall performance. The addition and reliance on embedded control strategies and algorithms constantly alter the behaviours of the whole system, making the solution of the globally optimal design and control much more challenging. A new and generic approach for addressing this complicated issue using integrated modeling, design and control optimization is presented. Research and developments on the modeling, design and control optimizations of the next-generation hybrid electric vehicles and marine vessels are overviewed.

Biography:
Zuomin Dong received his Ph.D. of Mechanical Engineering from the University at Buffalo, the State University of New York in 1989, started his academic career at the University of Victoria, Canada right after, and served as Chair of the Department of Mechanical Engineering at UVic for many years. He is a Fellow of CSME, Distinguished Professor of the Chinese Academy of Sciences (2016-18), and Associate Editors of several international journals. He recently received the 2019 Achievement Award from the Mechatronics and Embedded Systems Subdivision of the American Society of Mechanical Engineers.

Dr. Dong and his team carried out leading research on integrated modeling, simulation, design and control optimization of hybrid electric vehicles, marine vessels, hydrogen fuel cell systems, adaptive power distribution networks, and advanced manufacturing. He has established and led the UVic Green Vehicle Research and Testing Centre, the Clean Transportation research team, and the award-winning hybrid vehicle student competition team with significant funding.

Dr. Dong has provided excellent training to many students and researchers at all levels. Many of them became academic and industrial leaders in mechatronics, manufacturing automation, electrified propulsion of vehicles and marine vessels with advanced hybrid electric or fuel cell powertrain systems.

Dr. Dong also served as an advisor, a member of the Board of Directors, or committee Chair of several research funding programs, academic institutions and public companies. He has been closely collaborating with many industrial partners and transferred the developed technologies to industry.

Prof. Kui Wu

Prof. Kui Wu

University of Victoria, British Columbia, Canada

Reusing Backup Batteries as BESS for Power Demand Reshaping in 5G and Beyond

09:50-10:25 PDT, May 10, 2021

The mobile network operators are upgrading their network facilities and shifting to the 5G era at an unprecedented pace. The huge operating expense (OPEX), mainly the energy consumption cost, has become the major concern of the operators. In this talk, I will introduce our recent work on reducing energy cost by transforming the backup batteries of base stations (BSs) to a distributed battery energy storage system (BESS). Specifically, to minimize the total energy cost, we model the distributed BESS discharge/charge scheduling as an optimization problem by incorporating comprehensive practical considerations. Then, considering the dynamic BS power demands in practice, we propose a deep reinforcement learning (DRL) based approach to make BESS scheduling decisions in real-time. The experiments using real-world BS deployment and traffic load data demonstrate that with our DRL-based BESS scheduling, the peak power demand charge of BSs can be reduced by up to 27%, and the yearly OPEX saving for 2, 282 5G BSs could reach up to US$185, 000.

Biography:
Kui Wu entered the Special Class for the Gifted Young of Wuhan University, China, in 1985. He received the B.Sc. and the M.Sc. degrees in Computer Science from Wuhan University, China in 1990 and 1993, respectively, and the Ph.D. degree in Computing Science from the University of Alberta, Canada, in 2002. He joined the Department of Computer Science at the University of Victoria, Canada in 2002 as an Assistant Professor and is currently a Full Professor there. During his sabbatical years, he was a JSPS Fellow at the University of Tsukuba, Japan (2009), a Visiting Professor at City University of Hong Kong (2009, 2019), and a Visiting Professor at the Norwegian University of Science and Technology (2008, 2019). His research covers the theoretical foundation of computer networks (including stochastic modelling, network tomography, network calculus, dependence modelling), online social network analytics and e-business, computational sustainability, Quality of service (QoS) for cloud computing and content delivery networks (CDN), and computer/IoT/smartphone security. His research has been reported broadly by media such as MIT technology review, ACM TechNews, slashdot, Time Colonist, and Discovery News.

Dr. Yong Zhang

Dr. Yong Zhang

Huawei Technologies Canada

Personalized Federated Learning on Non-IID Data

10:25-11:00 PDT, May 10, 2021

Federated learning allows many clients to collaboratively learn a strong model without infringing their data privacy. One major challenge of federated learning is that the data distributions of different clients are non-IID. This makes it difficult to collaboratively train a single model to fit the data of all clients properly. As a result, it is important to collaboratively train personalized models such that each model fits the data distribution of a client well, and the data privacy of all clients remains intact. As a cloud provider, we design a novel method named federated attentive message passing (FedAMP) to tackle this challenging problem. The key idea is to let each client own a personalized model, and encourage clients with similar models to collaborate more in clusters during training. By passing model parameters between similar clients in a non-linear manner, FedAMP adaptively discovers the underlying collaboration clusters of clients without knowing the number of clusters a priori. This boosts the effectiveness of collaboration and leads to outstanding performance.

Yong Zhang currently is a Senior Principal Researcher at Huawei Technologies Canada and leading the big data and intelligence platform lab at Vancouver research center. Prior to that, he obtained his PhD in Mathematics from Simon Fraser University and was a postdoctoral research fellow at Stanford University, USA. His research interests include large scale numerical optimization and machine learning. His research works have been published in top-tier journals and conferences.

Dr. Hani Vahedi

Dr. Hani Vahedi

Power Electronics, Ossiaco Inc., Canada

PUC5 Converter, A Technology Transfer from Academia to the Industry

11:00-11:35 PDT, May 10, 2021

Power electronics converters are used in most of the power industries especially renewable energy conversion and electrified transportation. Due to unbelievable fast development of the semiconductor devices such as SiC and GaN, the size and cost of the market inverters are getting much smaller in recent years. On the other hand, the same story is happening to the digital signal processor (dsp) microcontrollers that facilitate implementing digital controllers for those power electronics inverters. Multilevel inverter technology is getting mature since last decade and replacing the conventional 2-level 6-switch 3-phase one in high power applications. As well, the 1-phase multilevel topologies could be seen in the market for medium power ratings. Although too many topologies are reported every day, a few of them could find their way to the industry. The main issue with most of the introduced topologies is the high number of components and especially using more than one isolated DC source. There are some certain differences between single-DC-source and multiple-DC-source multilevel inverter topologies which limit the application of multiple-DC-source ones in power systems. This presentation will focus on the advantages and applications of Single-DC-Source Multilevel Converters for EV and PV systems. It will describe a true story of transferring a multilevel converter topology called 5-Level Packed U-Cell (PUC5) from a research project in an academic lab of a university to the industry and market.

Hani Vahedi (S’10, M’20, SM’20) was born in Sari, Iran, in 1986. He received his B.Sc. and M.Sc. degrees in Power electrical engineering from K. N. Toosi University of Technology (KNTU), Tehran, IRAN in 2008 and Babol University of Technology, Babol, IRAN in 2011, respectively. He received his PhD with honor from École de Technologie Superieure (ÉTS), University of Quebec, in Montreal, Canada in 2016. He is the recipient of Best PhD Thesis Award for the academic year of 2016-2017 from ETS. He has published more than 60 technical papers in IEEE conferences and Transactions. He also published a book on Springer Nature and a book chapter in Elsevier. He has received best paper and presentation awards as well as travel assistance in numerous international conferences. He is an active member of IEEE Industrial Electronics Society (IES) and its Student & Young Professionals (S&YP) committee. He is a co-chair of special sessions, co-organizer of S&YP Forum and co-chair of 3M video session in IES conferences. He also serves as an Associate Editor for IEEE Transactions on Industrial Electronics. He is the inventor of PUC5 converter and holds multiple US patents and transferred that technology to the industry. Currently, he is a power electronics designer and chief scientist at Ossiaco Inc, Montreal, Canada, developing and commercializing the first bidirectional Electric Vehicle DC Charger based on PUC5. His research interests include power electronics multilevel converters topology, control and modulation techniques, power quality, active power filter, and their applications into smart grid, renewable energy conversion, UPS, battery chargers and electric vehicles.