Advanced modelling and evolutionary multiobjective optimization of motor drive for aerospace applications

Cheong, Benjamin

Advanced modelling and evolutionary multiobjective optimization of motor drive for aerospace applications

Publication Date

July 24, 2020

Creator

Cheong, Benjamin

Abstract

The More-Electric Aircraft (MEA) and hybrid/all-electric propulsion concepts have been introduced to improve the efficiency and reduce the environmental impact of air travel. Within the MEA concept, it is proposed to replace heavy pneumatic, mechanical or hydraulic driven elements with electrical systems, and to this end, an increasing number of motor drives is required on-board. Likewise, within hybrid/all-electric propulsion concepts, motor drives play a pivotal role in handling the drivetrain electromechanical power conversions. However, the greatest challenge concerning the wide employment of motor drives is in achieving high power density and high efficiency in their design. The main goal of this PhD thesis is to develop advanced models and multiobjective optimisation techniques for motor drive system-level weight and efficiency optimisation. This holistic approach considers all components within a typical motor drive, i.e. the electrical machine, and the power converter with its associated cooling system and passive filter components. The novel contribution of this work is in combining a comprehensive knowledge of machine and converter modelling, with modern optimisation algorithms to form a practical system-level motor drive design and optimisation tool. The first part of the thesis presents a computationally-efficient analytical model developed for a surface permanent magnet machine, to evaluate its multiphysics operation and performance. Detailed finite-element analysis and experimental tests are carried out to validate the model's accuracy. A similar analytical model is also developed for the power converter, to predict its performance and size the associated cooling system and passive components. The model's accuracy is also verified using time-domain simulations and experimental results. Next, the machine and converter models (component-level) are integrated into one system-level model, to comprehensively evaluate the motor drive's performance and weight. To model extra losses in the electrical machine due to power converter pulse width modulations, a physics-based harmonic loss model is also proposed and incorporated within the system-level model. Owing to the availability and capability of modern computing, researchers and engineers are increasingly turning to evolutionary algorithms to solve complex optimisation problems. The second part of the thesis discusses the application of evolutionary multiobjective optimisation (EMO) algorithms for an optimisation-based motor drive design, with the system-level model at its core. In particular, three popular EMO algorithms, namely the NSGAII, SMPSO, and GDE3, are considered, and practical guidelines for their implementations are provided. For illustration, a motor drive optimisation case study is presented for a benchmark prototype motor drive, while visualisation plots are employed to analyse the optimisation results. Finally, the three EMO algorithms are compared against each other, and the SMPSO algorithm is found to be most competitive for a multiobjective motor drive optimisation problem.

Item Type

ethesis

Thesis Type

PhD

Supervisors