The present investigation deals with the control of springback phenomena in the bending process of aluminium sheets by hybrid forming process. Metal substrates were pre-bent to nominal shapes on a built-ad-hoc mould after being constrained on it. Then, they were post-treated by high power diode laser to prevent the deformation of the pre-bent sheets after the release of the constraints. The extent of springback phenomena were estimated by measuring the difference between the nominal bending angles and those achieved on the unconstrained substrates after laser post-treatments. Analytical models, aimed at predicting the springback by varying the setting of the operational parameters of the forming process, were developed. Neural network solutions were also proposed to improve the matching between experimental and numerical data, with the Multi-Layer Perceptrons trained by Back-Propagation algorithm being the fittest one. On this basis, a control modulus very useful to practitioners for automation and simulation purposes was built-on. © 2011 Elsevier Ltd. All rights reserved.
Gisario, A., Barletta, M., Conti, C., Guarino, S. (2011). Springback control in sheet metal bending by laser-assisted bending: Experimental analysis, empirical and neural network modelling. OPTICS AND LASERS IN ENGINEERING, 49(12), 1372-1383 [10.1016/j.optlaseng.2011.07.010].
Springback control in sheet metal bending by laser-assisted bending: Experimental analysis, empirical and neural network modelling
BARLETTA, MASSIMILIANO;
2011-01-01
Abstract
The present investigation deals with the control of springback phenomena in the bending process of aluminium sheets by hybrid forming process. Metal substrates were pre-bent to nominal shapes on a built-ad-hoc mould after being constrained on it. Then, they were post-treated by high power diode laser to prevent the deformation of the pre-bent sheets after the release of the constraints. The extent of springback phenomena were estimated by measuring the difference between the nominal bending angles and those achieved on the unconstrained substrates after laser post-treatments. Analytical models, aimed at predicting the springback by varying the setting of the operational parameters of the forming process, were developed. Neural network solutions were also proposed to improve the matching between experimental and numerical data, with the Multi-Layer Perceptrons trained by Back-Propagation algorithm being the fittest one. On this basis, a control modulus very useful to practitioners for automation and simulation purposes was built-on. © 2011 Elsevier Ltd. All rights reserved.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
