Prediction of thin layer drying of edible mushroom moisture content by feed forward artificial neural networks method

Nowadays the edible mushroom has got a great consideration as a high protein and low calorie food source and also its medicinal consumption. As the production of edible mushrooms increases, it is necessary to notice on its storability, waste decreasing and usage of dried mushrooms. For this reason drying of this crop as a practical way has been considered. Nowadays advantages of artificial neural network are noticeable for simulation and prediction of necessary parameters in drying processes. The objective of this study is moisture content prediction of edible mushroom with arithmetic neural network during drying process. In this study the thin layers of edible mushroom were dried by hot air dryer in three temperature levels: 40, 50, 60°C and three air flow speed levels: 0.5, 0.7, 1 m/s until the moisture content reached to 10% (d.b.). Regression models and feed forward neural network with Levenberg–Marquardt learning algorithms were used to estimate and predict the moisture content of mushroom thin layers. The results showed that the quadratic degree regression model with determination coefficient of 97.0 and the error sum of the squares of 0.071 and also the neural network with 3-20-18-1 structure with sigmoid threshold and logarithmic function in comparison with other topologies had better results to predict the moisture content of mushroom thin layer during drying.