Design optimisation of a superconducting AC generator

The feasibility of applying rational methods of optimisation to a superconducting generator design is presented. A new design method is developed which permits the optimisation of a superconducting generator to take place. The method is used in a computer program software package which systematically determines the optimum design parameters of the machine components. The optimisation techniques and the logical structure of the computer package are described. The optimisation problem is solved by using the interior penalty function method in which the direct search methods of Hook-Jeeves (pattern) and Nelder-Mead are used. A numerical example demonstrating the effectiveness of the program is given. A new analytical method is also presented for the prediction of the inner rotor bursting speed by taking into account the geometry, the applied stresses and material properties. The result of this method is used as mechanical performance constraints. Moreover, analytical presentation of other machine performances such as the critical characteristic of the superconductor are presented