Multi-Response Robust Parameter Design For Rare Earth Products

Robust design is an important technical method in quality improvement activities,which is usually used in the initial development and design stage of a product or production process,effectively improving the quality of the product by reducing quality variation.With the increasing complexity of products and production processes,robust parameter design not only measures and balances the robustness and importance of multiquality characteristics,but also needs to consider the correlation between responses and the conflict between multiple objectives.However,most studies on robust parameter design often ignore the correlation between responses.Therefore,the problem about related multiresponse robust parameter design is the focus of current research.In this paper,the quality design problem of NdFeB hot deformation process is taken as the research object.And how to construct a multi-response optimization model with related multiple responses and small samples is solved by combining partial least squares regression analysis with traditional mass loss function.The main research contents of this paper are summarized as follows:(1)The method used in this paper considers the correlation between multiple responses,and is especially suitable for quality design problems in small batches of complex products.It is difficult and limited to obtain relevant data of production process,and there is inevitably correlation between data,which can't be solved such problems effectively by ordinary optimization methods.(2)By calculating the mass loss function value of each mass characteristic as the response variable,the model between response variables and influence factors is based on partial least squares regression analysis to estimate the parameters,which improves the traditional quality loss function.Therefore,the multi-response with correlation is transformed into standardized loss deviation that is a new comprehensive variable with better explanatory,which effectively reduces the conflict and contradiction of multiresponse in the optimization process.(3)In this paper,the process of hot deformation of NdFeB is taken as the research background.Through a variety of auxiliary analysis techniques,the influence of controllable influence factors such as hot deformation temperature,deformation rate and percent deformation on magnetic properties is visually expressed in a two-dimensional plane.The ability of each controllable factor to interpret the set of response variables is analyzed,and the conclusion that the controllable factors such as particle size of magnetic powder MQPA,temperature and percent deformation have an important influence on magnetic properties is obtained.At the same time,the conclusions obtained in the paper are verified and explained from the perspective of the principle of hot deformation.At the end of the thesis,the robust design on the NdFeB hot deformation process needs to be further solved in future research,which is proposed on the basis of summarizing the above research results.