Artificial Neural Network Modeling of Water Activity: a Low Energy Approach to Freeze Drying

A method for reducing the energy consumption during freeze drying has been proposed. Water activity variation with time has been explored for button mushroom (Agaricus bisporus L.). The effect of primary and secondary drying temperatures on water activity was found significant (p < 0.05) as compared to sample thickness and pressure. The economics of the process showed that an energy reduction up to 34.9% could be achieved if the final water activity was constrained at 0.6. Artificial neural network tool has been used to develop a model for predicting the water activity precisely for a given combination of time, initial moisture content, vacuum pressure, sample thickness, and primary and secondary drying temperatures. The model-predicted values were found to be in good agreement (R = 0.97) with the experimental data. The model developed is expected to extend its aid in energy reduction for freeze drying of other food products.