Process capability analysis for manufacturing processes based on the truncated data from supplier products

Abstract

Quality data fraud not only destroys the trust between suppliers and customers but also misleads the decision-making when choosing suppliers. Thus, it is preferred to use the quality data measured by customers to evaluate the manufacturing process capability indexes (PCIs). In practice, the suppliers always conduct a preliminary internal inspection to eliminate the nonconforming items before selling products, and quality data measured by the customers are truncated by the specification limits, which makes it difficult to measure the PCIs. This paper proposes a novel method to estimate the PCIs based on the truncated data. First, we propose a new data filling method called the QA-EM by integrating the EM and quantile-filling algorithms. Consequently, the truncated data can be converted into pseudo-complete data. A comparison study with other methods is further carried out to demonstrate the superiority of our proposed method. Then, various interval methods for estimating PCIs are applied to calculate the lower confidence limits of based on the pseudo-complete data. We investigate the performance of different methods in terms of coverage rate. The results indicate that the generalised confidence interval method performs better than the competitors. Finally, an industrial example is presented to illustrate the application of our method.