Normal operating record data of a continuous digester pulping process are used to study the possibilities of improving process regulator control with stochastic discrete time series techniques. It is illustrated that inclusion of a secondary input that is uncorrelated with the main input in a univariate process model enables more efficient forecasting of disturbances to be made, resulting in better control for minimum mean squared output. The resulting control scheme is a form of feedforward control which calls for adjustments in only one main process input.

This content is only available via PDF.
You do not currently have access to this content.