| This dissertation presents thermal design methodologies for switched reluctance (SR) and brushless permanent magnet (BPM) machines incorporating the effects of machine geometry, configuration, material properties and parameters on the overall performance of the machines. The developed methodologies are verified with system level simulations and finite element analysis (FEA).; Analytical models based on machine geometry and material properties have been developed for calculating the machine heat losses. On the same basis, lumped parameter thermal networks have been developed for predicting the transient and steady state thermal characteristics of the two motors. Radiation and natural convection heat transfer effects are also incorporated into the analytical networks. Using the lumped parameter models, the effects of various design parameters and internal machine dimensions on the temperature rise have been analyzed in depth. Some of the parameters and dimensional ratios are seen to have significant thermal effects, while others have minimal. An optimization technique using Lagrange Multipliers has been applied to optimize the design of SR and BPM machines, with the most thermally significant parameters of the respective motors as the design variables. Torque density, temperature rise and envelope dimensions are used as the parameters of the cost function. System level simulations and finite element analysis are done to verify the electromagnetic design and the thermal performance of the optimized machines. Also a design methodology has been developed for an SR machine that operates with low duty cycle. The state space representation of the lumped parameter thermal model for SR machine is used to predict the thermal time constant, which is later utilized in resizing the motor for low duty cycle operation.; A new design methodology has also been developed for BPM machines, which extracts all the machine dimensions, both external and internal, from the required and given specifications.; Experiments are conducted to record temperature rise data at different locations of an SR motor in order to validate the lumped parameter thermal network. The experimental results have successfully correlated with the analytical predictions. |
