Use of fuzzy neural-net model for rule generation of activated sludge process

Abstract

Activated sludge plant is usually difficult to operate and control because of its complex operational behaviour and usual significant process disturbances. To increase safety and improve operating performance of this biological wastewater treatment process, it is important to develop computer operational decision support systems. This intelligent computing system is able to assist ordinary operators to work at the level of a domain expert in daily operation. Neural network techniques and fuzzy logic methods have become important and effective tools to help build such intelligent system. The artificial neural network technique is powerful because it can learn to represent complicated data patterns or data relationships between input and output variables of the system being studied. Nevertheless, it has limitations in performing heuristic reasoning of the domain problem. On the other hand, expert systems are good at performing heuristic reasoning by making use of logic rules. It is, however, generally weak for knowledge acquisition. In this study, a fuzzy neural model is developed for addressing the operating problems of activated sludge processes, relating to prediction and heuristic understanding of the sludge age. Neural network techniques and fuzzy logic are used in model development. Simulation studies show that this fuzzy-neural network model obtained is able to extract fuzzy rules from a set of numerical data that can be used to carry out heuristic reasoning.