The Application Research Of Data Mining And Spc In Manufacturing Quality Control

This paper compares the characteristics between traditional control chart and CUSUM control chart. CUSUM control chart is provided in the prediction application of the process running stability by subparagraph. And it's reliability is proved by the result of experiment and provides a new way of thinking for fault diagnosis .Statistics analysis is used to find root cause of staking pressing abnormal on accumulator staking station. With 5M1E used in this case, narrows the range of possible causes and finds the root cause finally.ID3,NB decision tree and ANN are used in data of hi-pressure performance tester and compared their results. The results indicate it can classify two new and troublesome failure modes in assembled ABS unit.In this paper, Apriori algorithm is studied and introduced it's main algorithms of association analysis. Based on the manufacture and maintenance log files, Apriori is used in high FTQ issue of hi-pressure performance tester. Compared the result with different min. support values and combined the spastics chart to explain the association rules , some useful rules are obtained successfully in this case.