Few-sample multi-objective optimisation of a double-sided tubular machine with hybrid segmented permanent magnet

Double-sided tubular machine (DSTM) is very suitable for wave energy conversion but easily suffers from high thrust ripple. In order to get the minimum cogging force with the maximum thrust force, a new DSTM with hybrid segmented permanent magnet array is proposed and optimised by a novel iterative few-sample multi-objective optimisation method. The novel optimisation method is based on an iterative Taguchi method framework to obtain optimal design with only few samples. To solve the low precision problem of the iterative Taguchi method, a surrogate-model based multi-objective optimisation algorithm that uses a general regression neural network, a speed-constrained multi-objective particle swarm optimisation and an exponentially weighted moving average are embedded into this framework. The optimisation result is compared with other alternative topologies and methods, and a prototype is manufactured for testing experiment.