Real time Takagi-Sugeno fuzzy model based pattern recognition in the batch chemical industry

This contribution describes the real time pressure check pattern recognition of an industrial batch dryer. The goal is to identify the start of the drying process and to calculate the time elapsed between two consequent batch starts (batch time) right after the batch has completed. The presented pattern recognition method implements a supervised learning approach based on Takagi-Sugeno fuzzy (TS) models. The decision maker design is based on plant data compressed by the PI algorithm (OSI Software, Inc). It is concluded that the developed classifier is able to perform real time classification and the compressed PI data can be used in order to design data analysis tools which are useful for chemical batch plant operation investigations.