The model for verifying software should be decided considering many viewpoints, for example, ease of experiment. Then, the general requirements of the most suitable model for 3-D nonlinear eddy current analysis are examined in this paper. Some factors affecting the analysis and experiments are also investigated, in order to obtain accurate and reproducible results. For example, various types of elements are compared with each other. A transient nonlinear model is proposed, and the flux and eddy current distributions obtained by using the finite element method are compared with experimental results. © 1990, IEEE. All rights reserved.