ANN Modelling for Prediction of Moisture Content and Drying Characteristics of Paddy in Fluidized Bed

Phyu Phyu Thant, P.S. Robi, P. Mahanta

Drying characteristics of paddy were studied in inclined bubbling fluidized bed dryer at the air temperatures of 55, 60 and 65°C, air velocities of 1.1, 1.6, and 2.1 m/s, dryer inclination angles of 0Ëš, 15Ëš and 30Ëš and inventories of 0.5 to 2.5 kg. By applying the artificial neural networks (ANNs), moisture content of paddy was predicted under the various input conditions of different drying air temperatures, superficial air velocities, inclination angles of dryer, inventories and drying time. The learning of ANN is accomplished by feed forward back propagation algorithm. The simulated results are compared with the experimental results. The effect of input parameters is significant on the moisture content and drying time. The optimized ANN was found 12 neurons in hidden layer. The 1st and 2nd functions are tansig and logsig, respectively at 84 iterations and error goal is 0.00006. The ANN model gives the average absolute relative error (AARE) of an acceptable level of 3.3% with a correction coefficient (Rcc) of 99.6% and it is found that moisture content predicted by the neural network model developed in this work is in a good agreement, which have a non-linear relationship with each other is believed to be an accurate prediction the moisture content of grain.