Abstract
We have been developing optimal design methods for high-field and highly homogeneous superconducting magnets for NMR study using mathematical programing method. It is hoped that the error of produced magnetic field is very small, however, the error field is caused by winding errors, warp of coils, ferromagnetic objects around a magnet and so on. It is needed to reduce the error field in the design stage. The presented design technique is combination of the simulated annealing, which is one of most effective methods for nonlinear optimization problems, and the sensitivity analysis. In the design stage of the high-field and highly 4emoge-neous superconducting magnets, this method reduces the error coefficient of the magnetic field homogeneity around the center region, which is caused by winding errors. The detail of algorithm and the examples of its application to the high-field and highly homogeneous superconducting magnets are shown.
