AN INTEGRATED NEURAL-NETWORK AND EXPERT-SYSTEM APPROACH TO THE SUPERVISION OF REACTOR OPERATING STATES IN POLYETHYLENE TEREPHTHALATE PRODUCTION

With the emphasis on quality indices in polyethylene terephthalate (PE T) production, it is highly desirable to assess, and subsequently to m aintain, individual reactors at satisfactory operating states. To fulf il these aims, it is proposed to use an approach that integrates artif icial neural networks (ANNs) with an expert system (ES); the purpose o f the former is to estimate the quality indices from reactor process v ariables, whilst that of the latter is to assess current operation, an d thence to advise on, for instance, the application of optimisation p rocedures to any of the reactor controllers. In addition, an expert sy stem is used to filter plant measurements before they are input to the ANNs; the aim is to suppress gross errors that can cause the ANNs to output incorrect conclusions. The ANN training algorithm is based on a Quasi-Newton method with a self-scaling variable metric (SSVM), becau se simulation results show that the algorithm has high performance esp ecially in terms of its speed of convergence. The work was implemented on a large-scale PET plant, with the software installed as an Applica tion Module of a Honeywell TDC-3000 distributed control system. (C) 19 98 Published by Elsevier Science Ltd. All rights reserved.