WIT Press

Fuzzy Genetic Algorithm Based Inductive Learning System (FGALS): A New Machine Learning Approach And Application For Chemical Process Fault Diagnosis

Price

Free (open access)

Paper DOI

10.2495/AI960121

Volume

16

Pages

9

Published

1996

Size

61 kb

Author(s)

I. Burak Özyurt & Aydin K. Sunol

Abstract

Fuzzy Genetic Algorithm Based Inductive Learning System (FGALS): A New Machine Learning Approach and Application For Chemical Process Fault Diagnosis I. Burak Özyurt, Aydin K. Sunol Chemical Engineering Department, University of South Florida, Tampa, FL, USA Email: ozyurt@sunflash.eng.usf.edu, sunol@sunset.eng.usf.edu. In today’s complex chemical processes, extracting of general knowledge from the noisy raw process data, coming continuously from the sensors, is an important issue. In this paper, an approach for symbolic knowledge extracting from noisy raw process data based on genetic algorithms (GAs), namely Fuzzy Genetic Algorithm based inductive Learning System (FGALS), is illustrated. The developed system is able to extract knowledge from the process data in the form of natural language like fuzzy rules. The system is also able to use available domain knowledge and it is robust to noise. The applicability of the developed system for fault diagnosis is shown on a hydrocarbon chlorination pla

Keywords