Experimental Investigation on Geometric Accuracy …
217
Fig. 9 Main effects plots for geometric accuracy
The normality plot of the residuals in Fig. 8 shows that many residual points
fall on a straight line which implies that the error distribution is normal and the
residuals follow the normality plot. Figure 9 shows the main effect plot for geometric
accuracy. As shown in the graph, feed rate has a small influence on geometric accuracy
compared to other parameters. As feed rate increases, the RMSE value decreases but
the variation is not significant.
It is observed from the main effect plot that with an increase in pitch size, RMSE
value increases. This is because the forming profiles generated are more intense when
the pitch value is set at a low value. And also the action of the forming tool is less
between two contiguous profiles which provides greater uniform material distribution and plastic deformation along forming depth. Therefore, for better geometric
accuracy, a small pitch should be used. It is observed from Fig. 9 that the RMSE value
decreases with an increase in the number of forming stages. In the first stage, since
undeformed sheet is deformed at wall angle of 40°, the springback is considerable.
Therefore, accuracy after first stage is low because of global and local springback.
Further, in second stage, 40° wall angle part is deformed (to 55° wall angle) rather
than deforming raw sheet. Thus, spring back in second stage is less relative to spring
back in first stage and similar is the case with third stage. So, as the number of
stages increases, spring back decreases and geometric accuracy increases. It can
be observed from main effects plots that maximum geometric accuracy is achieved
using three stages, 0.3 mm pitch and 3600 mm/min feed rate. Table 5 shows the
resemblance between the RMSE values of experimental and prediction (using the
developed mathematical model).
217
Fig. 9 Main effects plots for geometric accuracy
The normality plot of the residuals in Fig. 8 shows that many residual points
fall on a straight line which implies that the error distribution is normal and the
residuals follow the normality plot. Figure 9 shows the main effect plot for geometric
accuracy. As shown in the graph, feed rate has a small influence on geometric accuracy
compared to other parameters. As feed rate increases, the RMSE value decreases but
the variation is not significant.
It is observed from the main effect plot that with an increase in pitch size, RMSE
value increases. This is because the forming profiles generated are more intense when
the pitch value is set at a low value. And also the action of the forming tool is less
between two contiguous profiles which provides greater uniform material distribution and plastic deformation along forming depth. Therefore, for better geometric
accuracy, a small pitch should be used. It is observed from Fig. 9 that the RMSE value
decreases with an increase in the number of forming stages. In the first stage, since
undeformed sheet is deformed at wall angle of 40°, the springback is considerable.
Therefore, accuracy after first stage is low because of global and local springback.
Further, in second stage, 40° wall angle part is deformed (to 55° wall angle) rather
than deforming raw sheet. Thus, spring back in second stage is less relative to spring
back in first stage and similar is the case with third stage. So, as the number of
stages increases, spring back decreases and geometric accuracy increases. It can
be observed from main effects plots that maximum geometric accuracy is achieved
using three stages, 0.3 mm pitch and 3600 mm/min feed rate. Table 5 shows the
resemblance between the RMSE values of experimental and prediction (using the
developed mathematical model).