Design and optimization of a fuzzyneuro expert system for improving productivity of a chemical recovery boiler

Abstract

Information Technology application to industrial process control is a challenging task Model based control schemes require the existence of a suitable process model For non inear plants and processes the most suitable modeling based on the physics of the process and chemical reactions of the system is in place There are a number of data driven methods as well Data driven methods employ different approaches such as neural networks and fuzzy systems or a combination of both In the present work the structure and main functions of the supervisory level control system that has been developed for High Pressure Chemical Recovery Boiler is presented The input and output performance newlinedata are used in the model for discussions and analysis High energy performance along with continuous availability is an important criterion in high pressure Chemical Recovery Boilers The Chemical Recovery Cogeneration plant employed in paper and pulp Industries deals with a wide range of processes accepting information from a variety of inputs It has been observed that there is a significant variation in the production of steam on a continuous basis Although few basic relations connecting fuel quantity and its concentration with the steam output are known on a one to one basis the combined effect coupled with the effect of hidden parameters remains unknown as of today Hence the process of design of a fuzzy neuro expert system plays a significant role in predicting the set of actions in advance which need to be taken by the operator for newlineincreased productivity A brief report of the work carried is presented in the following paragraphs newline