Abstract
Abstract
This study focuses upon process design and forming analysis of a permalloy shielding can (PSC) supporting automobile multidisplay parts to indicate accurate information about the car. This study is particularly important, since the shape fixability of PSC is known to affect quite sensitively magnetic properties such as coercivity and permeability. In this study, an optimizing technique based on a neural network (NN) using an orthogonal array (OA) has been developed for deep drawing processes to obtain PSC parts that have good magnetic properties. The punch radius, die radius, and blank holding force are considered as design parameters in the deep drawing process. Also, in order to attain optimum magnetic properties, the roundness of the PSC part must be maintained within tolerances after the annealing process. The objective functions are hemming defects such as creepage, warp, tightness, and roundness in the cylinder prehemming and hemming process. The NN has been implemented to minimize the objective function and to investigate the influence of the prehemming angle on the hemming process. In the present study, the optimum prehemming angle selected firstly to obtain roundness and then to keep warp and tightness relatively insignificant was 124°. The results of analysis to validate the proposed design method are presented.