Research On Springback Prediction In Sheet Metal Forming Based On Radial-based Agent Model

As an important method of metal plastic forming, sheet metal forming has been widely used in several industrial fields, such as Aerospace, Auto and Household Appliances. And bending is one of the important processes of the cold stamping forming of sheet metal. As a result of that sheet metal forming needs to go through rather complex plastic deformation process, hence, determining the process of sheet metal forming reasonably can improve the quality of products and the stability of production. At present, to shorten the cycle of products’ development and manufacture and to reduce the development costs of the products are treated as the surviving goals of processing manufacturing industry. So the process of sheet metal forming and its optimization have attracted many researchers’eyes around the world wide.With the development of the industry, some new materials appear, during the practical production, the properties of the materials, the fluctuation of processing parameters and the reduction of the Die’s precision can lead the decline of the products’ quality. And the spring back of sheet metal is the most obvious phenomenon during the sheet metal forming, meanwhile, sheet metal’s spring back also is a common phenomenon which is caused by the unload of the external bending force. The spring back value greatly affects the geometric accuracy and subsequently the assembly performance of the shaped component. Because of many factors which can affects the spring back value, however, to calculate the spring back value accurately at theory is rather difficult. So the process parameters’optimization of the sheet metal forming and the forecasting of the spring back value have important practical significance for improving the qualities of products and the efficiency of the production.Using the Aluminum alloy as the research object, this paper has done the following works:(1) The deformation mechanism and the characteristics of the sheet metal at different bending deformation stages and the main factors that influence the spring back are analyzed. The structure of approximate model and the evaluating of approximate model’s precision are studied in the paper.(2)In order to improve the efficiency of the production of the sheet metal forming, this paper presents a method of approximate model which is based on radial basis function. The nonlinear function is fitted by using the radial basis function that is at the circumstances of different kernel functions. The research shows the approximation model which makes the Gaussian function as the kernel function has the highest precision when the numbers of the samples are determined(3) When the U-shaped bending forming part of the NUMISHEET ’93is used the research object, the numerical simulation is done by utilizing the Dynaform, and the parameter setting of finite element simulation of the U-shaped part’s bending spring back and other related issues are also analyzed. It is verified that finite element model in this paper is correctness. Make spring back angle as, select(4) When the material parameters are certain, spring back angle is defined as the evaluation index, and punch velocity, blank holder force, friction coefficient and mold clearance are selected as the process parameters which affect the spring back value. When the experiment level of the parameters are determined, the orthogonal experiment method is applied to design the experiment scheme, then the spring back results of all the experiment schemes which are numerically simulated by the Dynaform are obtained. Based on the results, the effect of different process parameters to the spring back angle is analyzed, and the process scheme is preliminarily analyzed, and the primary and secondary orders of factors which affect spring back value and the most excellent process program are obtained lastly.(5) The radial basis function is utilized to construct the surrogate model of prediction of the U-shaped spring back, and the nonlinear mapping relation of the stamping process parameters and stamping spring back value is established. Through analyzing the surrogate model, the toolbox of MATLAB is used to optimize the process parameters and predict spring back results.