Robust Design Optimization Methods for High Quality Producible Electrical Drive Systems




Electrical drive systems are key components in many modern appliances, as well as industry equipment and systems. In order to achieve the best design objectives, such as high performance and low cost, various optimization methods have been developed for design optimization of electrical machines and drive systems. The traditional design optimization is at the component level, e.g. optimization of a motor design or the parameters of a control algorithm. However, modern appliances or systems demand that the drive systems be specifically designed and optimized to provide full support to their best functionalities with multiple performance indicators. For such applications, an application-oriented multi-objective system-level design optimization method is developed. Because of the complexity of drive system design that involves many disciplines, such as electromagnetics, materials, mechanical dynamics including structural, thermal, and vibrational analyses, power electronic convertors, and control algorithms, a multi-level optimization method was developed by the authors to improve the effectiveness of the optimization of electrical machines as well as drive systems.

On the other hand, the real quality of motors and drives in mass production highly depends on the available machinery technology and those unavoidable variations or uncertainties in the manufacturing process, assembly process and operation environment. The manufacturing precision and tolerances are two main issues in the manufacturing process, including mainly the variations of material characteristics, and dimensional variations of parts of drive systems. The assembly process variations mainly include the lamination of silicon steel sheets, and misalignments of stator, rotor and PMs. The operating uncertainties mainly include the load variations, changes of electrical and mechanical parameters, such as the changes of resistance and inductance due to the operational temperature rise, and fluctuations of drive voltage.

Limited by these variations in the practical machinery technology, an aggressively optimized design may be difficult for high quality batch production and end up with high rejection rates. Similarly, variations in system parameters and operational conditions may also lead to sub-optimal performance, and in a severe case, even instability. To solve this type of problems, the methodology of Six-Sigma quality control can be adopted to develop a robust design optimization method to guarantee the high quality batch production of drive systems.

Based on many years of research experience, this talk aims to present a brief introduction of efficient application-oriented, multi-disciplinary, multi-objective, and multi-level design optimization methods for advanced high quality electrical drive systems. The multi-disciplinary analysis includes materials, electromagnetics, thermotics, mechanics, power electronics, applied mathematics, machinery technology, and quality control and management.



Prof. Jian Guo Zhu

Distinguished Professor of Electrical Engineering

Director, Centre for Green Energy and Vehicle Innovations

Head, School of Electrical, Mechanical and Mechatronic Systems

University of Technology Sydney,Australia


Brief Biography:

Prof. J. G. Zhu received his BE in 1982 from Jiangsu Institute of Technology, China, ME in 1987 from Shanghai University of Technology, China, and Ph.D in 1995 from University of Technology Sydney (UTS), Australia, all in electrical engineering. He currently holds the positions of Distinguished Professor of Electrical Engineering and Head for School of Electrical, Mechanical and Mechatronic Systems at UTS, Australia. His research interests include electromagnetics, magnetic properties of materials, electrical machines and drives, power electronics, green energy systems and smart micro grids.

He has been a team leader and chief investigator for over 50 government and industry funded research projects, and based on his research findings, published 2 books, 10 book chapters, 248 journal articles, and 554 conference papers.