Outrunner generator with optimized cogging torque pattern for an electromechanical energy harvester

In contrast to widely applied permanent magnet (PM) machine optimization for minimization of cogging torque, this paper describes a method for shaping the cogging torque of a PM outrunner machine towards a desired sinusoidal torque pattern. The underlying goal of this approach is to compensate the cogging torque of a kinetic energy harvester (KEH) with the optimized counter cogging torque of a generator connected to the same shaft. Therefore, the total cogging torque is highly reduced, and a self-starting electromechanical energy harvester, comprising KEH and outrunner generator is formed. The key degree of freedom for shaping the cogging torque is the sinusoidal modulation of the machine's air gap. An algorithm based on the multi-dimensional Secant method, which is related to the multi-dimensional Newton-Raphson method, first evaluates the cogging torque of a given generator geometry with two-dimensional finite element method (2-D FEM) simulations and then iterates the geometry of the outrunner machine until the cogging torque target is achieved. Using the presented optimization approach, a generator design with the desired sinusoidal cogging torque pattern is obtained, achieving a total cogging torque reduction of the overall electromechanical energy harvesting system of 90%.