Research On Relevant Questions Of Springback In Three-dimensional Multi-point Stretch Bending Forming Of Aluminum Profile

At this stage, the improvement of people’s material life and the development of socialand economic have been restricted by the energy shortage and environmental pollution,especially the automotive industry that closely related to those two issues. Automotivelight-weighting is one of the important way to reduce energy consumption and pollution.Aluminum alloy has many advantages, such as low density, high strength, easy to recoveredand so on, and widely used in automotive safety parts and car body frame. As people’srequirements of the vehicle performance and appearance are continued to increasing, theexterior design of vehicle has becoming more and more complex, spatial structure pieceinevitable direction toward the three-dimensional surface, which makes the traditional wayof forming difficult to meet the requirements. By combining the three-dimensionalmulti-point flexible forming technology and traditional stretch bending process, it can notonly have advantages of multi-point flexible forming technologies, but also to ensureaccurate profiles forming the three-dimensional space, which provides a new moldingmethods to the rapid development of the automotive industry.Rebound phenomenon is a inevitable phenomenon in the process of metal plasticdeformation. It is caused by the historical accumulation occurred. In this paper, ABAQUSwas used to simulate the three-dimensional multi-point bending forming process of thecross-section aluminum which closed or semi-closed. Studied the effect of pre-stretched,pulled up the amount, horizontal bending angle, vertical bending angle, elastic modulus andyield stress on the springback. The results showed that:For semi-closed and closed cross-section aluminum, both horizontal and verticaldirection, the greater the elastic modulus are, the smaller the springback value, the greaterthe yield stress are, the greater the springback value; pretension increases, the springbackvalue decreases; while pulling up increases, the springback value fist decreases andgradually flatten out; vertical direction of the bending angle increases, the springback valuein horizontal direction reduces; at the same time, the vertical direction rebound increases,horizontal direction can get similar to the variation law of the bending angle; for two cross-section, the rebound resilience of the closed cross section is greater than that of thesemi-closed section.In addition, this article also take closed section aluminum as an example, based on theabove six parameters set up a six factors and five levels orthogonal test, using ABAQUSsoftware to numerical simulation the orthogonal test levels, based on the statistical analysisof test results data, get the optimal parameter combination:1.4%of the amount of pretension,1%filling amount,10°horizontal bending angle, vertical bending angle of10°,250000Mpaelastic modulus, and300Mpa yield stress. Finally, based on support vector machineregression to predict the springback, compared the predicted results with neural networkprediction results, proved that the SVR algorithm for springback prediction accuracy ishigher than the neural network algorithm; and validated using test method, test proved thatthe SVR algorithm can effectively predict the springback.