Research On The Evaluation Approach And Application Of The Ability Of Dimensional Reduction Multivariate Measurement System

Measurement system analysis(MSA)plays an important role in helping production organizations improve product quality and is an important tool in quality decision systems.The traditional MSA only considers a single quality feature.However,with the development of production technology,the data dimension that characterizes product quality is not limited to a single variable.It often has multiple quality characteristics and has a high correlation between each other.MSA may lead to misjudgment of quality decisions and bring inestimable potential losses.If the capability evaluation method of a multidimensional measurement system with dimension reduction is adopted,such hidden dangers can be avoided.At present,most manufacturing enterprises are driven by user needs as their core production philosophy.Different user groups have different quality tendencies for certain quality characteristics.In addition,the existing capability analysis methods of dimension-reducing multivariate measurement systems are mainly based on principal component analysis(PCA),which has certain limitations,and there is insufficient explanation between the quality characteristics of the measurement system and the contribution rate of error fluctuations.In this paper,by improving the application research of multi-component measurement system capability evaluation methods in complex production environments,this paper expands the exploration ideas of dimension-reduction type measurement system evaluation methods in terms of the inconsistency of the importance of quality characteristics and the insufficient explanation of variation fluctuations.(1)This article takes the multi-value measurement system as the research object,comprehensively extracts the measurement information,comprehensively considers the correlation between the importance of the quality characteristics subjective preference and the quality characteristics,uses the projection pursuit dimension reduction idea and adopts the IPSO algorithm to extract the best From the perspective of subjective and objective comprehensive weighting,the projection direction proposes a multi-measurement system capability evaluation method that considers subjective preferences of quality characteristics.By fitting the projection score,the acceptability of the multivariate measurement system is discriminated against.Compared with other dimensionality reduction methods,the acceptability of the measurement system has been further improved.The impact of subjective preference on the quality characteristics of the acceptability of the measurement system is initially discriminated against.And the relationship with the measurement system error source.(2)Most of the dimensionality reduction methods can only judge whether the multivariate measurement system meets the requirements of the production environment from a quantitative point of view,but the interpretation of the variation fluctuation of the measurement system is poor,it is difficult to quantify the specific correlation between the multivariate quality characteristics and the variation fluctuation.Under the premise of considering the inherent correlation of multi-quality characteristics,the relationship between the convenience of comprehensive dimensionality reduction method and the fluctuation of system fluctuations,based on the principal component analysis,the varimax factor rotation method is adopted,and the expansion method of PCA is used-The factor analysis method evaluates the capabilities of the multivariate measurement system,identifies the relationship between individual factors and the multivariate quality characteristics,and evaluates the relationship between sources of fluctuation variation.(3)Select the welding workshop of Dongfeng Yueda Kia No.2 Factory for empirical research.Through the different angles of the inconsistency between the multiple quality characteristics and the interpretation of the quality characteristics to the error source,the methods of projection pursuit and factor analysis are used to deal with some problems in the spot welding process of the dimensionality reduction multi-dimensional measurement system capability evaluation.Empirical research shows that the two types of methods can be matched with each other,and cross-complementary use strategies are adopted according to different production conditions,which is more suitable for the actual application environment under the actual production environment,and is more conducive to the promotion and application of multiple measurement system analysis methods in the enterprise.