A fuzzy state estimator for fed-batch fermentation

One of the main problems in controlling fermentation is that the important variables cannot be directly measured. One way round this problem is to use secondary measurements in conjunction with an estimator which calculates the values of the control variables as the fermentation proceeds. Conventional estimators are based on mathematical models and have not, to date, been greatly used by industry. This paper describes the early stages of the development of an estimator based on fuzzy logic. The model the estimator uses is defined by qualitative statements and estimates specific growth rate in a fed-batch bakers yeast fermentation. The estimator's performance is compared with simulation results