| Summary: | 博士 === 國立成功大學 === 統計學系碩博士班 === 97 === Process capability indices have been widely used in industry today since they provide quantitative measures for evaluating process performance. Instead of using process capability index, the process incapability index Cpp has an advantage of separating the information of process accuracy and precision for a symmetric tolerance. Using Cpp process incapability index, the process performance and its risk level can easily be understood by quality practitioners. Another process incapability index C"pp is to measure the process performance for an asymmetric tolerance. However, it can’t correctly reflect the true manufacturing risk. In order to accurately measure the process performance for both symmetric and asymmetric tolerances, a new process incapability index using desirability function is proposed. The relationship between various new process incapability indices and their associated non-conforming rates has also been explored. In the first part of this dissertation, two numerical examples are illustrated to show that the manufacturing risk can be correctly reflected by our proposed univariate process incapability index.
Furthermore, due to the fact that the quality of an industrial product is determined by two or more interrelated quality characteristics. A new multivariate incapability process index is developed by taking the correlation among multiple quality characteristics into account. Analogous to univariate process incapability index, the multivariate process incapability indices can be viewed as an extension of the univariate process incapability indices and thus the manufacturing risk for a multivariate manufacturing process can be evaluated accordingly. Finally, the correlated risk assessment using the new multivariate incapability indices and their associated interval estimates is further demonstrated by two numerical examples.
|
|---|
