| In the optimization design process of modern manufacturing industrial products/processes,robust parameter design has gradually become one of the widely used and effective methods.Through the research on the horizontal combination of controllable factors,the optimal combination of parameters is selected to reduce the sensitivity of products/processes to noise factors,so as to reduce and control the influence of fluctuations in the manufacturing process from the root.To a certain extent,this method effectively improves the robustness of product/process quality and plays an increasingly important role in many fields.With the complexity of products/processes in manufacturing industry,the robust parameter design problem with multi-quality characteristics has become increasingly prominent.Because different responses have different quality characteristics and may have certain correlation with each other,as well as model uncertainty between responses and controllable factors,it is difficult to find a set of optimal parameter combinations of controllable factors in the optimization process so as to optimize all quality characteristics.In order to solve the above problems,this paper proposes to combine principal component analysis and grey relational analysis to obtain a new way to build a robust parameter design model.The main contents of this paper include:(1)In this paper,principal component analysis is used to reduce the dimension of multiquality characteristics in robust parameter design.The problem of incomplete and uncertain information is dealt with by using grey relational analysis method to solve the problem of model uncertainty in robust parameter design.Through the combination of principal component analysis and grey relational analysis,a robust parameter design model is constructed.Its essence is to transform multiple quality characteristics of different types into a comprehensive performance index with promising characteristics,that is,comprehensive grey relational degree.The multi-quality characteristic is converted into a single quality characteristic to reduce the error in the robust parameter design process.(2)In this paper,the upstream and downstream products of rare earth process flow,namely high purity neodymium oxide and neodymium iron boron magnetic materials,are selected for example analysis to verify the advantages and disadvantages of the robust parameter design model proposed in this paper.The method proposed in this paper is combined with the actual production process data of two rare earth products,and the multi-quality characteristics are converted into comprehensive performance indexes,and the comprehensive grey correlation degree is obtained.Then,the MINITAB software is used for factor effect analysis,and finally the optimal parameter combination in the two examples of preparing high-purity neodymium oxide and NdFeB magnetic materials is obtained.(3)In this paper,the signal-to-noise ratio prediction method is used to predict the signal-tonoise ratio value under the optimal parameter combination obtained by the method proposed in this paper,and the signal-to-noise ratio value obtained by other methods is predicted,and the comparison between the two is made to verify the effectiveness of the method in this paper.After obtaining the optimal parameter combination,by predicting the signal-to-noise ratio of each response of the actual process data and comparing the signal-to-noise ratio prediction values of the methods proposed in this paper with those of the methods proposed in reference documents,the signal-to-noise ratio prediction value obtained by the methods proposed in this paper is larger,therefore,the method proposed in this paper is better and can realize the quality robustness of rare earth products.Finally,on the basis of summarizing the above research results,the paper puts forward the problems to be solved urgently in the future robust design research of rare earth production process,and determines the focus and direction of future research. |
PDF Full Text Request