Multi-objective optimisation of a magnetic gear for powertrain applications

This paper proposes the study of a magnetic gear by means of multi-objective optimisation. Magnetic gears are structures able to transfer motion between two or more mechanical axles with a given speed ratio. In this aspect they are analogous to a classical mechanical gear but they allow to transfer motion without contact between the moving parts. To replace a standard mechanical gear in an automotive powertrain, the design of a magnetic gear able to transfer the torque needed by the application is required. At the same time, beside the torque, the replacement of the mechanical gear should not alter the dynamic of the powertrain so that also mechanical parameters, as for instance the moment of inertia, must be considered in the design. To cope with these requirements, a multi-objective optimisation approach is proposed that maximises the transmitted torque per unit mass of the magnetic gear while minimising the moment of inertia of the moving parts. Due to the computational intensive evaluation of the magnetic performance, the optimisation is carried out by a deterministic optimisation algorithm coupled to a weighted sum approach of objectives. Results are discussed and compared with those of a commercial application.