The Research On Dynamic Multivariate Quality Control For Multi-variety And Small-batch Production

Nowadays,the manufacturing industry is facing a great change not seen in a century.With the improvement of people's living standard and the diversification of product demand,the traditional manufacturing mode of mass production turns to the mode of multi-variety and small-batch production.Because the multi-variety small-batch production mode has the characteristics of multi-index,multi-category,multi-evaluation index and small sample data,the traditional quality control method can no longer satisfy the quality control under the multi-variety small-batch production mode.In recent years,although many achievements have been made in the quality control method under the multi-variety small-batch production mode,there is still a lot of work to be done to effectively carry out the multi-quality control under the multi-variety small-batch production mode.Therefore,it is of great theoretical and practical significance to study dynamic multivariate quality control in the mode of multi-variety and small-batch production.In this thesis,automobile crankshaft is taken as the research object,and the dynamic multivariate quality control method under the multi-variety small-batch production mode is studied.The specific research contents are as follows:(1)Using fuzzy matter element method to determine similar processes.For solving the problem of data shortage of multi-variety and small-batch real-time data collection,combining the real-time data of RFID technology,the original process is determined by fuzzy matter element theory method and the similar process is found,and the similar parts family is constructed to achieve sample expansion.(2)The data are standardized by the data conversion method which combines raw data with real-time production data,and then the standardized data are tested for normal distribution and variance mean consistency.(3)Multiple process capability analysis using combined weights.The entropy weight method combined with expert experience gives the weight of each key quality characteristic of crankshaft.Based on several unit process capability indices,the method of calculating the combined weight multivariate process capability index is established,and the production process is analyzed.(4)Using Bayesian approach to build a multivariate quality control model.The prior information of Bayesian method is composed of raw data combined with originalproduction information,and the mean parameter estimation of Bayesian variance is carried out in combination with real-time data.The Bayesian dynamic single-valued moving range control diagram of real-time data is established to identify the causes of production variation and propose improvement measures.(5)Adopt multivariable Bayesian update and cost function.The whole optimization model of production process is established with the known parameters in the production process,which provides an economical and feasible research method for the implementation of dynamic multivariate quality control method under the multi-variety small batch production mode.