Springback Prediction And Process Optimization Of Flexible 3D Stretch-bending Technology For Aluminum Profile

Along with the development of lightweight vehicles and the industry, the traditional 2D stretch-bending forming for aluminum components can no longer meet the industrial demand, and 3D stretch-bending forming components have become increasingly popular. In this paper, a new flexible 3D stretch-bending forming process is presented to rapidly reconstruct the die surface and process a variety of cross-sectional profiles. Given the new forming process,3D stretch-bending can be performed, and the forming surface can be reconstructed, thus effectively shortening the time spent in designing, debugging and manufacturing die, and also increasing productivity. The 3D stretch-bending forming of aluminum profiles is a complicated mechanics process. It is difficult to precisely control the shape of formed profile, and forming defects such as springback, wrinkle and fracture can easily take place in the process of forming process. In order to acquire higher-quality aluminum components after stretch-bending forming, the technological parameters in the forming process should be strictly controlled and optimized. Herein, flexible 3D stretch-bending forming process is studied systematically. In addition, springback prediction and process optimization methods are presented.A new flexible 3D stretch-bending forming process is presented in this article, which combined the stretch-bending process and multi-point forming. The 3D stretch-bending forming is decomposed into horizontal bending and vertical bending in the new process.The principle and forming steps of 3D flexible stretch-bending forming are introduced, as well as the structure of key part unit-body. Besides, a finite element model of flexible 3D stretch-bending forming is put forward. Research contents include model simplification, simulation algorithm selection, material constitutive model, element selection, treatment of contact and friction, boundary conditions. This research provides theoretical basis for the finite element simulation of flexible 3D stretch-bending forming.Springback is the most difficult problem in stretch-bending forming. According to the characteristic of springback in 3D stretch-bending forming process, a springback prediction model using support vector regression algorithm is proposed in this article. The artificial neural network algorithm is commonly used in sprinback prediction, support vector regression algorithm is a machine learning method similar to artificial neural network algorithm. However, the generalization ability of support vector regression algorithm is stronger than that of artificial neural network algorithm. The rule of elasticity modulus, yield stress, pre-stretching elongation, post-stretching elongation, horizontal bending angle and vertical bending angle on springback are analyzed. Further, these parameters are taken as input parameters; the horizontal springback angle and the vertical springback angle are taken as output parameters. On this basis, the support vector regression method is employed for construction of a springback prediction model of flexible 3D stretch-bending forming. The accuracy of sprinaback prediction using support vector regression algorithm is higher than using s artificial neural network algorithm.The stretch force remains unchanged in the traditional stretch-bending forming. The method of using variable force trajectory is proposed in this article, and the optimization model of variable force trajectory is established with the target of springback reduction. A complex T profile case is applied for analysis of the stress-strain states at various stages of 3D stretch-bending forming process. Further, the model for mechanical analysis is established. Tension is divided into four stages, encompassing the pre-stretching stage, the horizontal stretch-bending stage, the vertical stretch-bending stage and the post-stretching stage. Stretch forces in four stages, pre-stretching elongation and post-stretching elongation are used as design variables, based on which a variable stretch force trajectory optimization method is presented in order to control the springback. The constraints in the optimization model are the minimum thinning rate and maximum thickening rate of profile. The experimental data are used to set up a surrogate model of objection and constraints using response surface method. Afterwards, the particle swarm optimization is utilized to optimize the variable stretch force trajectory. After optimized variable stretch force trajectory can effectively minimized the springback.Flexible 3D stretch-bending forming for rectangular profiles is analyzed, and the multi-objective optimization model of rectangular profiles. The forming defect of sagging is liable to produced in rectangular profile flexible 3D stretch-bending forming because of its hollow section. The method that can fill steel block to suppress sagging is put forward in this paper. Moreover, the influences of pre-stretching elongation and post-stretching elongation are analyzed on sagging. It has been found that increasing pre-stretching elongation and post-stretching elongation could lead to increase the sagging; otherwise it can effectively reduce springback. With the purpose of controlling springback and sagging simultaneously, a multi-objective optimization model is built of rectangular profiles 3D stretch-bending forming. In the optimization model, the objective function is to minimize the springback and sagging. The constraints in the optimization model are the minimum thinning rate and maximum thickening rate of profile. The experimental data are used to set up a surrogate model of objection and constraints using response surface method. A non dominated genetic algorithm is used to optimize the sets of solutions, and the optimal solution is obtained after analysis. With the optimized combinations of pre-stretching elongation and post-stretching elongation, springback and cross-sectional sagging can decline at the same time.A die surface optimization method of flexible 3D stretch-bending forming is presented based on iterative springback compensation. Aiming to control springback, the die surface is iteratively compensated by springback amount until the shape errors meet the forming requirements. The concept of segmental compensation factor is proposed, the profile is divided into two segments. And the springback compensation factor is adopted according to the springback amount. Using the optimized die surface for flexible stretch-bending forming, shape errors of the target shape can greately reduced.