Utilization of change point detection based on information criterion for monitoring flatness data

Flatness, a critical quality metric in engineering, follows a complex normal range distribution that lacks an explicit analytical form. Currently, non parametric change point detection methods are used to monitor flatness data in quality control, but these methods suffer from low accuracy. Our contribution lies in solving the parameter estimation for the normal range distribution and in the numerical calculation of its density function, enabling the application of an information criterion-based change point detection method. This method offers higher accuracy and a lower false positive rate. Case studies are presented to demonstrate the effectiveness of the proposed method in change point detection.