Applications of the Autocorrelation Function and Fourier Analysis to Quality Control.

摘要

This research investigates the feasibility of using Fourier Transform Theory and the autocorrelation function for the purpose of detecting when a process is out-of-control. Two methods are proposed, one using the spectral density of process data and the other using the autocorrelation of the process data. Density functions of the test statistic in each case are derived with the assumption that the process data is normally distributed. The results of simulations ran to verify the derived distributions are given, and comparisons are made between the effectiveness of the two proposed methods and a concentional control chart. The author feels that these methods are of interest to persons concerned with quality control, but should not be applied until more research is performed.