Journal of Researches in Mechanics of Agricultural Machinery

Journal of Researches in Mechanics of Agricultural Machinery

Predicting and Modelling of Mulberry Moisture Content in Drying Process using Mathematical Models and Artificial Neural Network

Authors
mohammadreza asghari
rahim ebrahimi
davoud ghanbarian
bahram Hosseinzadeh
Abstract
Mulberry is one of the fruits with high levels of useful sugar which one of ways to maintain the product is drying. Evaluation of the moisture content reduction is essential to find the appropriate drying time and conditions. In this study, mulberry drying was carried out using a hot air dryer in three temperature levels (50, 60 and 70 °C) and three air speed levels (1.5, 2 and 2.5 m/s). Results of comparing the measured moisture loss with the predicted values by mathematical models indicated that the Page model, due to the high correlation coefficient and low root mean square error can be used for describing and predicting mulberry drying kinetics. Results of modelling by neural network indicated that the leading model with training algorithm of Trainlm, overall structure of (4-25-1) and threshold functions of Logsig and Tansig had the highest performance (R2=0/9999 and MSE=0/000009) and consequently, it offered tbetter results as compared with other topologies.
Keywords
Mathematical modelling-Mulberry.-Drying-Artificial neural network-Moisture content-