Development of a Smart Supervisory Control System in a Sugar Mill Crystallisation Stage
This paper discusses the development of an expert advisory system designed to provide expert knowledge in the control and management of a sugar mill crystallization stage. The smart supervisory control system is fundamentally a hybrid fuzzy expert system that incorporates fuzzy logic, relational databases and process models of the cystallisation stage. The primary topic of this paper will be a description of the framework of the smart supervisory control system with focus on: (1) system architecture, and (2) process models. The models presented are a core component in the expert system framework to predict of future pan stage operating conditions.
The meta consequent was used for "integrating" data which may be in different data platforms, eg linguistic or quantititive. It is rather an approach for dealing with different data and models, allowing wider flexibility than the conventional defuzzifier which can only work in a single data format.
Not much neural nets have been used so far, most of them are conventional data modelling tools which are good enough. The concept appeals to industry partner and further investigation into a commercially viable product is being done.
Who's online
There are currently 0 users and 172 guests online.
What are the advantages of using meta-conseqent layer to the traditional defuzzifier?
What was the motivation behind using fuzzy logic to neural networks? Have you encountered any difficulties in applying the AI techniques? How long has the project been running for and when do you estimate to finish?
Regards