THE DEVELOPMENT OF EXPERT SYSTEMS FOR THE AUTOMATIC MONITORING OF QUALITY PARAMETERS AND THEIR INTEGRATION INTO INDUSTRIAL AUTOMATION SYSTEMS

Thanks to the combined use of heuristic expert knowledge, mathematical models and information on previous fermentation procedures saved in data records, fermentation processes can be predicted in actual operations. With the aid of data measured at the beginning of a current fermentation process, the developed software system is able to adapt itself to the process and predict its course. It can also improve its prediction accuracy at a later stage and. after two days, the point in time at which the correct degree of fermentation for transferring the beer is reached can be reliably estimated. This estimate will thus enable the fermentation room foreman to predict the end of the fermentation process for each desired temperature course he can achieve with the temperature regulator installed on his fermenting tank. The on-line estimate of the extract decomposition rate is also a prerequisite for efficient model-aided temperature control during the fermentation process (Gee and Ramirez 1988, 1991). The reduction of temperature fluctuations enables tighter control over the fermentation process. In further work, this on-line estimate of the basic factors of the fermentation process is used to determine further significant factors in the process. An on-line estimate of the course of the diacetyl concentration is being prepared.