Optimal design of multicharacteristic inspection plan udner inspection errors and statistical dependency.

Abstract

Complete inspection plans have become increasingly important in the area of quality control due to the growth in modern manufacturing systems that makes complete inspection inexpensive and reliable. In the process of inspection, an inspector is likely to commit Type I and Type II errors in his judgement about the product quality. In this thesis the effect of Type I and Type II errors are investigated on multicharacteristic repeat inspection plans for critical components. The practicality of the inspection models are enhanced by modifying them for the case where the defective rates of the characteristics are statistically dependent. The results indicate that the effect of errors and statistical dependency are significant and should be incorporated in the design of the inspection plans. Since the error probabilities are a function of incoming quality, the suggested procedure for estimating Type I and Type II errors for a given incoming quality is utilized to incorporate the dynamic behavior of inspection errors into the repeat inspection models and their effect is studied. It is noticed that varying inspection errors have a significant effect on the plans in terms of expected total cost. Then a factorial experiment is conducted to find the factors and interactions that have a significant effect on the performance measures of the inspection plan.