ABSTRACT

Switched Reluctance Motor (SRM) is becoming increasingly popular in industrial applications such as wind energy systems and electric vehicles due to its simple and rugged design, capacity for high-speed operation, resistance to high temperatures, and fault-tolerant features. Despite these features, SRM has some drawbacks such as torque ripple and acoustic noise. Hence, there is need for detailed design and analysis of the motor. Finite element analysis of a 3-phase, 6/4, 1.5 kW SRM has been performed for enhanced performance in this dissertation. An intelligent optimization program based on multi-objective genetic algorithm (NSGA-II) was implemented to study the effects of saturation, stator yoke, pole embrace of the stator and rotor on the average torque, efficiency, average total loss, and average torque ripples. It was observed that the stator pole embrace, and yoke thickness, are the key parameters to the optimization objectives. The comprehensive performance of the optimized design in terms of average torque and efficiency were 5.59 Nm and 96.20% respectively, which were 15 % and 13.2 % better than the base model. Also, there were 30.1 % and 1.55 % reduction in average total loss and torque ripple in the optimized model when compared with the base design. This proves the success of the NSGAII intelligent optimization program as a framework to optimize the specified objective functions. Furthermore, 2D thermal FEM model was established in Maxwell 2D and was exported to lumped parameter thermal network of MotorCAD. The lumped parameter method was used to analyze the temperature of the motor in which all parts of the heat path were combined to form thermal circuit system for the entire motor. It was observed that flow rate II of 15 l/min gave the lowest temperature levels of 251.5 0C and 158.3 0C in the winding and 150.2 0C and 198.1 0C in the lamination at both regions respectively; hence, it was adopted for further thermal analysis of motor over wide range of speed. Maxwell’s stress tensor method was used in the 2D electromagnetic FEA analysis to compute radial and tangential electromagnetic forces that were applied on the stator tips in Maxwell 2D/3D. It was observed that tangential force increased by 48.6% in the optimized model which indicates higher torque generation than the initial model whereas the radial force decreased by 19.2% in the optimized model which shows that the effect of vibration was minimized in the optimized model. The full acoustic fingerprints of the normal and optimized models were also given by the waterfall diagrams which displayed the equivalent radiated power level that estimated the radiated structure-borne sound power from the vibrating structural surface which is the stator surface in this dissertation. For laboratory testing of the studied motor, the electromagnetic torque equivalent of the electronic load box has maximum limit torque of 6.6 Nm, which was the adopted torque limit for experimental verification. The experimental results sufficiently validate the accuracy of the proposed SRM model.