Applying Artificial Neural Network for Shrinkage Prediction of Green Pea in a Microwave Assisted Fluidized Bed Dryer

In this search drying characteristics of green pea (Pisum satium) with an initial moisture content of 76% (db) was studied in a fluidized bed dryer assisted by microwave heating. Four drying air temperatures (30, 40, 50 and 60ºC) and five microwave powers (180, 360, 540, 720 and 900W) were adopted. At each drying operating conditions the volume of green pea was calculated by measuring the three principal characteristic dimensions and the variation of the ratio of mean diameter of the kernel to its initial mean diameter was investigated for different operating conditions. It has been shown that power of microwave and temperature of drying could reduce the shrinkage of particles considerably. Furthermore, in this study, the application of Artificial Neural Network (ANN) for predicting the drying time (output parameter) was investigated. Microwave power; drying air temperature and green pea moisture content were considered as input parameters for the model. An ANN model with 4 neurons was selected for studying the influence of transfer functions and training algorithms. The results revealed that a network with the logsig transfer function and trainrp back propagation algorithm made the most accurate predictions for the green pea drying system. In order to test the ANN model, (RMSE), (MAE) (SE) were calculated and found that the random errors were within and acceptable range of ±5% with correlation coefficient (R2) of 98%.