Knowledge-based rules for control of the sake (Ginjoshu) making process and their application in fuzzy control

Abstract

For temperature control of excellent sake (Ginjoshu) making process, the construction of the knowledge-based rules was studied From the interview with process operators (Toji), 4 knowledge based rules were listed up. Especially, it was found that Alcohol-Baumé plot and BMD curve were important as a reference for making Ginjoshu with good quality. Fermentation period was separated into 4 control regions. Fuzzy production rules and membership functions for temperature control in each region were constructed by using above references. Tuning of membership function was done using the simulator as proposed by us and the brushed-up rule was applied to 25L and 250L-Ginjoshu fermentation. Time course of temperature, acidity, Baumé, glucose, pyruvic acid, some alcohols and some esters were judged to be similar to those of traditional fermentation process by Toji.