Enhancement of dimensional accuracy of dieless tube-drawing process with vision-based fuzzy control

Dieless tube drawing offers flexibility to reduce the cross-sectional area of a tube only by adopting a drawing speed higher than the feeding speed without requiring dies or a mandrel. However, this promising technique faces the technological challenge of enhancing the drawn part accuracy. The deformation behavior of the process depends on the processing conditions. Therefore, real-time monitoring and control are necessary. A vision sensor was employed to monitor the minimum outer radius, and a fuzzy controller was selected to control the deformation according to the desired geometries by adjusting the drawing speed. Shift reference was introduced to compensate the deformation delay of dieless tube drawing. The dimensional accuracy of the drawn part improved after implementing the proposed control system. However, an oscillation of the profile on the drawn part caused by improper adjustment of drawing speed. An adaptive fuzzy controller was introduced to perform appropriate adjustment of drawing speed on the basis of the deformation characteristics. The result shows that a drawn outer radius accuracy of up to 99.6% was achieved with a 1.6% variation on the axial direction at a 44% reduction ratio. This performance was verified under various reduction ratios and feeding speeds.