Date of Award

Spring 2023

Document Type


Degree Name

Doctor of Philosophy (PhD)


Electrical and Computer Engineering

First Advisor

Demerdash, Nabeel A. O.

Second Advisor

Richie, James E.

Third Advisor

Povinelli, Richard J.


The goal of this dissertation is to establish computationally efficient large-scale design optimization procedures for induction motors that are not only computationally efficient, but also do not sacrifice the accuracy of the results. To achieve this goal, the topic of the lengthy numerical transient response of Time-Stepping Finite Element Analysis (TS-FEA) of induction motors was investigated. This is because this lengthy transient affects the computational efficiency aspect of the optimization process. The effect of different parameters on the numerical transient response phenomenon were studied, and two different techniques for mitigation of this numerical transient were evaluated. The superior one amongst the two techniques was highlighted and implemented on the case-study induction motor. This technique was also extended to induction motors under faults. Applying Maximum Torque Per Amp (MTPA) or constant Volts/Hz control on uncharacterized design candidates of a large-scale evolutionary optimization process that are assumed to be operating with a sine-wave current-regulated drive is a major challenge and process accuracy may suffer significantly. In this dissertation, a novel and computationally efficient technique was introduced that can apply the MTPA or constant Volts/Hz control on uncharacterized design candidates without any prior knowledge of the motor equivalent circuit parameters. A new computationally efficient VBR-based routine for calibration of the motor temperatures estimation was introduced in this dissertation. This hybrid approach is a fusion between the VBR-based TS-FE and the thermal network models of an induction motor and is aimed to enhance the accuracy of the induction motor optimization process. In addition, the sensitivity analysis for both purely magnetic and Multiphysics cases were performed to identify and deactivate the less impactful design variables to enhance the computational efficiency of the optimization process. Eventually, the techniques established throughout this dissertation are utilized to perform computationally efficient large-scale design optimization of induction motor that does not sacrifice on accuracy. The 5-Hp case-study induction motor was electromagnetically optimized under three different scenarios DE-based algorithm. In addition, a multi-physics design optimization routine was carried out with its own objective functions and design constraints that evaluated the motor in thermal, mechanical, and magnetic environments in real-time.

Available for download on Thursday, April 03, 2025

Included in

Engineering Commons