Research On Causes Of Abnormal Quality Of Manufacturing Process Based On Multi-factor Fusion

Under the background of China's manufacturing transformation and upgrading,market dynamics and increasingly fierce competition,reliable product quality has become the key to a company's development and market competition.However,due to the complexity and variability of factors affecting product quality,it is difficult to locate and trace the quality impact factors of products.Among the many factors affecting product quality,some may affect the quality of the product alone,and some may have a combined effect,which together affects the quality of the product.In order to find out the many factors that affect quality factors,eliminate the causes of quality anomalies,and control product quality,we must thoroughly explore the causes of abnormal product quality.Since the fluctuation of product quality obeys certain statistical laws,the paper aims to use statistical process control(SPC)to detect abnormal fluctuations in quality.When abnormality in quality is detected,system diagram method is used as an analysis tool.The system map lists all the factors that may affect the quality indicator.By combining the evaluation indicators with the field survey comprehensive scoring method,the influencing factors are ranked,and the key influencing factors with higher priority are selected.This method reduces the complexity of the original information and facilitates further optimization of the influencing factors.For the key influencing factors selected,using the multivariate statistical principal component analysis method,the key factors of the screening were further reduced and simplified by SPSS software.In the case of the least loss of original information,it is integrated into several major anomalies.These main reasons represent the information of a certain aspect,which are independent of each other and have no overlapping of information.On the basis of principal component analysis,the response surface analysis method is used to analyze the fusion effects of each main cause.Firstly,the main effects and interaction effects of the main factors are obtained by the variance analysis method of the response surface,and the saliency is evaluated.Then,trace and optimize the cause of the anomaly: The main reasons for the absence of interaction effect are optimized by singlefactor.The main reason for the significant interaction effect,using regression and variance analysis to obtain the degree of significance of the interaction effect and effective quality index regression prediction model.According to the order of significance,the abnormal factors can be traced,and the regression model can be used to predict the quality indicators.With the help of the interaction effect diagram of the factors,the adjustment and optimization of various factors can be realized.Finally,the multi-factor fusion quality anomaly analysis method is used to study the X-cylinder low-pressure casting process of an auto parts manufacturing enterprise,and the feasibility and effectiveness of the method are verified.