TPT September 2009

new bend die. A third model was simulated without an applied boost load, using the original bend die profile. The same friction and other conditions were applied to the three models except as mentioned above. The bend OD wall thinning ratio, ID wall thickening ratio, and ovality predicted by the three models are plotted in terms of the bend angle. As shown in Figure 5, the clamp die, or the head of the tube is considered the start of the bending angle (0°). The tail of the tube is considered the end of the bending (180°). With no boost load, the OD wall thinning ratio is about 16%, as shown in Figure 6. The distribution of the OD wall around the bend was predicted to be fairly uniform. With the optimised boost load schedule, the OD wall thinning was effectively controlled for bending angles less than 150°. The OD wall thickness is directly affected by the boost load. The higher boost load at the beginning of the bending process is predicted to result in a low amount of OD wall thinning, around 3% to 4%. However, as bending progresses past the 90° position, the maximum boost load is limited due to the possibility of the tube detaching from the die, and the amount of OD wall thinning is predicted to increase. The OD wall thinning ratio predicted from the FEA model is about 10% close to the tube tail. The two bend die geometries are predicted to result in similar levels of wall thinning at the OD of the bend. Without boost load, the wall thickening at the bend ID is about 20%, as shown in Figure 7. With boost load, the ID wall thickening is much greater, with values of 35 to over 40%. The bend die profile is also predicted to affect the ID wall thickness and the ovality significantly, as shown in both Figures 7 and 8. It is interesting that for this specific case, the bend die profile did not affect the OD wall

Angle (Degree)

Ovality Ratio

 Figure 8 : Ovality distribution

thickness significantly while it did affect the other two geometric characteristics of the bend. With the new bend die, the ID wall thickening is predicted to be about 40%, versus 35% from the original die. The ovality ratio is shown in Figure 8; a positive ratio means the vertical OD is greater than the horizontal OD. With no boost load, the ovality is greater than with a boost load and the tube slightly collapses, as signified by the positive sign for the ovality ratio. When a boost load is used, the ovality is reduced and the sign is reversed. Besides the effect of the bend wall thickness, the bend die geometry has an impact on ovality. The effect of the bend die on the ovality ratio should be a combined effect of boost load and the pressure die profile as well. The results from FEA simulations can be effectively used to set-up a tube bending process and improve the quality of the bend geometry from the first bend trial. However, to be successful as a design tool, the model must be validated against experimental results to be certain that the model variables are equivalent to the actual process variables, namely friction, tube strength, tube dimensions and the variations of these parameters. The following example compares the results of the FE model presented above and tube bending experiments. Machine settings A Pines CNC 150 HD tube bender, shown in Photo 2, was available for the practical tests. The CNC 150 HD has programmable booster pressures and 180 pressure zones. Using the research data to program the tube bender

ID Wall Thickening Ratio OD Wall Thinning Ratio

Angle (Degree)

 Figure 6 : OD wall thinning ratio distribution

 Figure 7 : ID wall thickening ratio distribution

 Photo 2 : Pines CNC 150 HD tube bender

Angle (Degree)

133

www.read-tpt.com

S eptember 2009

›