Robust Parameter Design Of Sheet Metal Forming Based On Kriging Model

In modern quality engineering theory,"fluctuation" is the fundamental cause of product quality problems.Robust parameter design is an efficient and low-cost method to control the quality fluctuation in a minimum range.The implementation of robust parameter design is to reduce and control the fluctuation of quality response index.During the implementation of robust parameter design,however,because the cost and time,could not have a lot of physical experiment was carried out,this inevitably there is a small amount of data,resulting in the researchers can’t well recognition model,and build the model caused a great deal of uncertainty,severely reduces the fitting precision of the model.In order to reduce the uncertainty of proxy model in the optimization process of automobile sheet metal forming and improve the quality of sheet metal forming,robust parameter design method based on Kriging model was proposed in this paper.The optimal parameter combination that can generate the desired quality response was determined according to the few available experimental data.The specific research contents are as follows:(1)Impact factor selection based on likelihood function method.Evaluation in order to measure the result of the sheet metal forming,the quality of the need to make sure that correspond to various aspects demand response index,and the measure standard setting,determine the quality of sheet metal forming response index,in view of the impact factor too much of the problem of high cost,large amount of calculation and test using the method of likelihood function of each influence factor screening,The factors affecting the quality index are selected to improve the optimization efficiency and reduce the calculation cost.(2)Construction of Kriging model.After the quality response index and its influence factors are determined,the Kriging method can be used to establish the proxy model between the quality index and each factor.First,the ordinary Kriging method is used to construct the initial Kriging proxy model,and then the Kriging model adding point criterion considering the prediction accuracy of update point is adopted.The initial Kriging model is updated until the set stop update conditions are met and the final Kriging model is output.(3)Robust parameter design based on satisfaction function method and dual response surface method.In the process of multi-response parameter design,if only the optimality of optimization results is considered,the resulting design may have great fluctuation,which makes the optimization design particularly sensitive to the changes of relevant parameters.To solve this problem,this paper combines the dual response surface method and satisfaction function method.A robust parameter design method based on the Kriging model was designed instead of the polynomial model in the traditional double response surface method.The functional relationship between the impact factors and the quality response was constructed based on the existing few experimental data.At the same time will build good Kriging model instead of dual response surface method(DRSM)of the polynomial model,quality the mean and standard deviation of the response.model is.set up respectively,and adopts improved.satisfaction.function.method,the response mean and standard deviation are taken into account at the same time,both to ensure the final result of optimality,and combining the forming process of robustness.Compared with other methods,the effectiveness of the proposed method is proved.