Cost Minimization of a Permanent Magnet Eddy Current Brake by Multiobjective Particle Swarm Optimization Based on Nonlinear Reluctance Network Modeling

Abstract

Design optimization of a permanent magnet eddy current brake (PM-ECB) is performed by applying multiobjective particle swarm optimization (MO-PSO) for cost minimization. A previously designed and patented PM-ECB is used as a reference model in the study. A quasi-3-dimensional (3D) analytical modeling approach based on a reluctance network considering the actual structure of the reference PM-ECB is proposed and verified. The Gauss–Seidel method is used as a nonlinear solver for the reluctance network modeling, and the braking torque is calculated considering both the skin effect and the armature reaction. Multiobjective optimization is developed by applying a particle swarm algorithm, and a 3D Pareto front is provided to demonstrate all non-dominating design points. Three cost functions, viz. rated braking torque, magnet mass, and magnetic flux density of the yoke, are selected as the objectives for the optimization problem, and the optimum design point is addressed in detail. The optimized design is validated by 3D-FEA and experiments. The results indicate that a 40% reduction in the magnet volume could be brought about by the optimized PM-ECB design with practically the same braking torque. Further, a 40% cost reduction in the optimized brake could be achieved compared with the reference one.