Modelling and first order optimisation of the air-cored tubular PM machine using polynomial approximation

Abstract

Finite element analysis remains the most reliable method of simulating electromagnetic systems. When attempting to optimise a system, FEA can be computationally expensive and time consuming. This paper demonstrates the use of polynomial approximation to FEA results to minimise the necessary number of FEA simulations. Using this method, a simulation is developed, capable of accurately reproducing the magnetic field above one pole in the air-cored tubular permanent magnet machine for any combination of machine parameters in a pre-chosen design space. The same method is also employed to optimise the machine parameters to produce the maximum reaction force for a fixed current. A second optimisation is also presented based on the capital cost per average power produced per unit length of translator in a typical wave motion.