Shifts recognition in correlated process data using a neural network

Traditional statistical process control (SPC) techniques of control chartin g are not applicable in many process industries because data from these fac ilities are autocorrelated. Therefore the reduction in process variability obtained through the use of SPC techniques has not been realized in process industries. Techniques are needed to serve the same functions as SPC contr ol charts, which are to identify shifts in correlated parameters. Neural ne tworks are a potential tool for identifying shifts in correlated process pa rameters, as data independence is not an assumption of neural network theor y. In this research, a back-propagation neural network is utilized to ident ify shifts in process parameter values from AR( 1) time series models with varying values of the autocorrelation coefficient phi. To rnd the appropria te number of input nodes for use in a neural network model, the all-possibl e-regression selection procedure is applied. In addition, time series resid ual control charts are also developed for the data sets for comparison. As the results reveal, networks were successful at separating data that were s hifted one, two and three standard deviations from non-shifted data for gen erated process data. The SPC control charts were not able to identify the s ame process shifts. In the other words, the neural networks can be used to identify shifts in process parameters. Therefore, it is allowing improved c ontrol in manufacturing processes that generate correlated process data.