| Sheet metal stamping is a very important processing method in modern industry,especially in the automobile field.Most parts of the automobile body are made up of the covers,and the springback defect in the cover stamping process will seriously affect the assembly accuracy of the automobile.The springback can be controlled by optimizing the process parameters to obtain the automobile covers that meet the production requirements.In sheet metal stamping,the material parameters and process parameters have different degrees of fluctuation,and this parameter uncertainty will reduce the product stamping quality,so a robust design is required.Meanwhile,the metamodels are widely used to replace the real models in robust design,so the metamodel uncertainty is inevitable.If the metamodel uncertainty is neglected,the optimal solution may be invalid.Therefore,a new robust design method with the consideration of parameter uncertainty and metamodel uncertainty is proposed,and the proposed method is applied to the springback control for the complex stamping part.The main work is as follows:In order to analyze the factors affecting the production quality,a Likelihood Function Factor Screening(LFFS)method is adopted.The feasibility of the proposed LFFS method is demonstrated by a case from the petrochemical industry.Considering the influence of the metamodel uncertainty on the robust design,a Bayesian Estimation based on the Genetic Algorithm(BEGA)method is applied according to the Genetic Algorithm and the Bayesian Estimation theory,which estimates the metamodel coefficients.The effectiveness of the proposed BEGA method is demonstrated by four mathematical examples.Considering the influence of the parameter uncertainty on the robustness of stamping part product quality,a tolerance robust model with the inner and outer loop optimization structure is improved based on the tolerance robust model with the single loop optimization structure and the theory of interval analysis.The proposed model is applied to the robust design for the NUMISHEET93 square box,and the deterministic optimization design method is used for the comparison research.The results show that the tolerance robust design method with the inner and outer loop optimization structure can effectively improve the robustness of the stamping part product quality.The proposed robust design method is used to control the springback for the NUMISHEET 2016 Land Rover part,and the maximum springback amounts of the three specific sections are selected as the optimization objectives.Firstly,the LFFS method is used to determine the key factors that have significant influence on the optimization objectives.Secondly,the metamodels between the key factors and the optimization objectives are established by the BEGA method.Then,the tolerance robust model with the inner and outer loop optimization structure is established.Finally,the established tolerance robust model is optimized by the Genetic Algorithm to obtain the optimal process parameters.The tolerance robust design method with the single loop optimization structure is used for the comparison research,and the results show that the proposed robust design method improves the robustness of the three optimization objectives by 28 percent,29.2 percent and 35.3 percent respectively.Therefore,using the proposed robust design method considering the two uncertainties to optimize the process parameters can effectively control the springback,and the robustness is better. |
