Adaptive heuristic temperature control of a batch polymerisation reactor

Abstract

The purpose of this research is to improve and apply the temperature control of a free radical solution polymerisation of styrene and to examine its performance on the basis of adaptive heuristic criticism (AHC) control. This algorithm consists of a three-layer feed forward artificial neural network (ANN) which uses supervised learning with reinforcement in a unique topology. This algorithm has two neurone-like adaptive elements and a difficult learning control problem which can be solved by means of a learning system with a single associative search element (ASE) and a single adaptive critic element (ACE). AHC uses a type of control system whose output value is either maximum or minimum. The controller will take in process data on-line and update the weights to proper ones in the control of the process. The performance results of the AHC controller are compared with the results obtained by using conventional Deadbeat algorithm. AHC control system shows satisfactory behaviour to track the reactor temperature. (C) 2003 Elsevier B.V. All rights reserved.