Adaptive Estimation in Inferential Process Control

Two techniques for inferring process outputs which can only be measured at infrequent intervals of time, from other more rapidly sampled secondary outputs, are proposed. Both estimators have quite general structures, the parameters of which can be updated on-line enabling the tracking of slow variations in both process and disturbance characteristics. The proposed methods require minimal design effort, reduce the possibilty of static estimation error and also contribute towards reducing the problem of secondary measurement selection. A significant improvement in overall process control performance is demonstrated by studies using a highly nonlinear distillation column simulation package.