Prediction of Pistachio Nuts Moisture Ratio Using a Proposed Recurrent Neural Network and Genetic Algorithm

Predictive models of recurrent neural networks were proposed in order to obtain on-line prediction of moisture kinetics during drying of pistachio nuts The experiments were conducted at four air temperatures (25, 40, 55 and 70 degrees C), three air velocities (0 5, 1 0 and 1 5 m/s) and two relative humidities (5 and 20%) The best topology of neural network for each state of drying conditions to predict moisture ratios was found For each drying condition two variables were used to find the best predictor topology of proposed recurrent neural networks number of delays in the input layer and the number of nerons in the hidden layer In order to reach this goal two methods were used trial and error method and genetic algorithm The derived models can be used for on-line state estimation and control of drying processes