In current paper, a modeling technique for time-series modeling is presented and applied for batch fluidized bed dryer. in this method, a neureo-fuzzy network is constructed based on dryer specifications and then is trained by recorded data. a neuro-fuzzy network can be translated to fuzzy inference system (FIS). trained neuro-fuzzy networks (FIS\'s) are checked by experimental data completely apart from training data. the achieved accuracy is very high, and as a significant result, the nature of the system can be expressed in fuzzy rules which are mainly formed by linguistic variables. since dryers are dynamic systems, recurrent (dynamic) neuro-fuzzy networks are used for modeling. steam batch dryer;s behavior is relatively complicated. initially, moisture content of drying material increases because of the absorption phenomenon, and then the moisture decreases with an approximately high and fixed slope, then at the third stage, the slopeof drying reduces significantly. in this research, despite classical methods, only one neuro-fuzzy network is constructed and trained for all of operating conditions.

In current paper, a modeling technique for time-series modeling is presented and applied for batch fluidized bed dryer. in this method, a neureo-fuzzy network is constructed based on dryer specifications and then is trained by recorded data. a neuro-fuzzy network can be translated to fuzzy inference system (FIS). trained neuro-fuzzy networks (FIS\'s) are checked by experimental data completely apart from training data. the achieved accuracy is very high, and as a significant result, the nature of the system can be expressed in fuzzy rules which are mainly formed by linguistic variables. since dryers are dynamic systems, recurrent (dynamic) neuro-fuzzy networks are used for modeling. steam batch dryer;s behavior is relatively complicated. initially, moisture content of drying material increases because of the absorption phenomenon, and then the moisture decreases with an approximately high and fixed slope, then at the third stage, the slopeof drying reduces significantly. in this research, despite classical methods, only one neuro-fuzzy network is constructed and trained for all of operating conditions.

Keywords