Intelligent pH control using fuzzy linear invariant clustering

Document Type

Conference Proceeding

Publication Date

1-1-1996

Abstract

This study explores the application of a fuzzy clustering algorithm in the field of Chemical Process Control. Process control is fundamental to Industrial Chemical Operations. The control problem considered in this study is a two level cascade control of the pH of a chemical stream. The pH is controlled by the addition of two chemicals - Sulfuric acid (to lower the pH) and Caustic (to increase the pH). The fuzzy clustering algorithm developed by J. Bezdek et al [1], and independently by Kundu and Chen [2] is used in this study to identify fuzzy rules from numerical I/O data points. The algorithm replaces the notion of a single representative point of a cluster with a more general notion of a hyperplane for each cluster. In this study, a simulation of the control problem has been generated and a menu driven GUI has been developed which enables the user to simulate different states of the control problem by modifying the tuning parameters. Preliminary experiments show that the rules learned by the fuzzy clustering perform well. These results provide support for the use of fuzzy clustering algorithms in process control.

Publication Source (Journal or Book title)

Proceedings of the Annual Southeastern Symposium on System Theory

First Page

514

Last Page

518

This document is currently not available here.

Share

COinS