Controlling first-part quality in small-radius press brake bending operatio
ns is motivated by the need to produce small-lot parts of high quality. A n
ew technique for angle control has been developed based on an experimentall
y measured model of material properties, tooling geometry variables, and on
-line measurements of part deformation. This technique requires the off-lin
e measurement of part thickness and hardness, plus the on-line measurement
of loaded angle and press ram position. Linear regression and neural networ
k techniques are compared to develop a robust model of springback. A contro
l scheme is introduced that provides accurate first-part bend angles for sm
all-lot sheet metal manufacture.