The Design And Application Of The Prediction System For Bending Springback Based On GABP Network

Bending is one of the most important forming processes in sheet metal forming, and springback is a universal phenomenon during bending. The existence of springback would bring out lots of problems, such as decreasing the parts' geometric precision, having a bad influence on the subsequent assembling, increasing the workload in adjusting modules, or extending the product development cycle, so it will be significant to study springback during bending process. In this paper, the springback of U-shaped beam parts is mainly researched. First, the actuality of three aspects of springback research is introduced, which includes the theory research, Finite Element Method(FEM) simulation and experiment controlling of springback. Then the applications of Artificial Neural Networks(ANN) and Genetic Algorithm(GA) in the field of material processing are generally introduced. After that, the mechanism of springback and the insufficiency of analytics in calculating springback are exposed. A simulation example of springback is demonstrated with the FEM software of Dynaform after describing how to set parameters in FEM simulation for U-shaped parts. A GABP prediction model is built up and realized by the computer language of MATLAB after exposing the principle of ANN and GA. Finally, the sensibility of nine factors in springback is researched with the method of range analysis, using the sample data of GABP, which would guide the springback controlling during the course of real production. The feasibility of system is verified by error analysis and result of comparing the system result with the traditional conclusions on the single factor affection law of springback.In this paper, ANN and GA technology are introduced to the field of sheet metal springback research, and a prediction system, which is realized by MATLAB GUI(Graphical User Interface), is developed. The efficiency and accuracy would be greatly promoted by using this system.