APPLICATION OF FUZZY RULES-BASED MODELS TO PREDICTION OF QUALITY DEGRADATION OF RICE AND MAIZE DURING HOT AIR-DRYING

The concept of modelling using a rules -based approach is introduced a nd developed for the prediction of the quality degradation of grains d uring hot air drying. It is then validated on two typical examples of quality degradation during drying: the head kernel yield of paddy rice and the wet-milling quality of maize. The linguistic approach gives a representation of the experts understanding of phenomena and leads to a good kinetic model after identification of the parameters using gen etic algorithms. This study also points out the robustness of the mode l prediction, even if noisy data are used, and its ability to predict transient behaviours under certain conditions.