Multi-objective design of mechanically-commutated DC machines

In the research presented herein, population-based design of mechanically-commutated DC machines is considered. To set the stage for design, a model of the machine is derived to predict the electromagnetic torque, commutation interval, conduction loss, and core loss based upon material property, geometry, and excitation. A key component of the model is an analytical expression for the flux density within the machine from which the armature winding inductance and voltage constant are obtained. Using the model, multi-objective optimization is performed to establish Pareto-optimal front between mass and power loss for a host of design constraints. Finally, a design from the Pareto-optimal front is validated using Finite Element Analysis (FEA).