Progress in process operation by goal oriented advanced control

Abstract

Larger competition, increasing demand for quality, and the necessity for lower inculpation of resources set the scene for production, storage, and processing of agricultural products and derivatives in the future. There is no doubt that automatic control will play a significant role in achieving these objectives in the chain from producer to customer. Advances in sensor technology, microelectronics, and control theory presently offer solutions far beyond the classical PID-controller. Yet, major gains will not be obtained if the ultimate goal is lost out of sight. This means that control solutions, how ever advanced they may be, will have to be placed in the frame of overall system operation in order to be profitable. An analysis of the situation in post-harvest processing shows that the bottle-neck in achieving economically attractive solutions lies in the necessity to have suitable models describing the product's behaviour in function of the environmental variables. Once such models are available the concepts of optimal predictive control provides an attractive framework for profitable process operation and control. Optimal solutions can be computed either off-line, and implemented with - possibly advanced - low-level feedback compensators, or implemented directly on-line using a receding horizon approach. Successful implementation of such goal oriented operation requires close cooperation between process engineers and control engineers. The philosophy is illustrated by examples from ongoing research projects on advanced control methods and dynamic optimal operation in the field of production, post-harvest technology and food processing.