Resumen de Determining Statistical Process Control Baseline Periods in Long Historical Data Streams
Hang Zang, Susan L. Albin, Steven R. Wagner, Daniel A. Nolet, Suhail Gupta
Baseline data used to create statistical process control models are often extracted from a long stream of data that contain observations from both successful and unsuccessful productions. Baseline periods must then be identified to ensure the on-line monitoring based on these SPC models is effective. A method is proposed that identifies baseline periods where the quality is good, the quality variable has a stable distribution, and the time intervals are sufficiently long. The method is shown to be robust to the distribution of the quality variable in tests on a real dataset from a melting process and in simulation experiments.
