| Failure in the drawing of sheet metal parts is caused predominantly by wrinkling or tearing, with both modes of failure influenced by local sheet tension. This research determined the suitability of drawbeads for the active control of sheet tension during the drawing process.; A multiple action computer controlled hydraulic strip drawing test apparatus, designed and constructed for this research, simulates the flow of sheet metal over the die shoulder and through the controllable drawbead of a sheet metal forming die. A bidirectional force transducer with a lockable roller was designed and constructed to measure strip tension before and after the die shoulder.; The strip tension data was used to simultaneously measure the drawbead restraining force and to calculate two different coefficients of friction; a flat plate coefficient of friction applicable to the flat blankholder surfaces, and a die shoulder coefficient of friction applicable to the drawbead and the curved blankholder surfaces.; The coefficients of friction were shown to be time varying functions of the lubricant, lubricant viscosity, lubricant additives, application rate, and die surface preparation procedure. The die shoulder coefficient of friction decreased under increasing strip tension, due the effect of flattening of surface asperities causing lubricant escape channels to close up, trapping the lubricant.; The drawbead restraining force was shown, both experimentally and analytically, to be a complex function of drawbead position and drawbead velocity history. The response (drawbead restraining force) of the controllable drawbead as a function of drawbead position and velocity was simulated using ABAQUS. The simulation results closely matched experimental results from the strip test apparatus. The simulation showed that the drawbead kinematic effects were due to the drawbead losing contact with the strip on fast retraction and the effects of residual bends being pulled out of the strip.; A control model was constructed using drawbead restraining force as the desired output and drawbead position as the control variable. The control model was used to design an optimal PI controller for the drawbead. The response was shown to be dominated by accumulated transport lag in the drawbead hydraulic system. |
