Drying of Apple Slices in Combined Heat and Power (CHP) Dryer: Comparison of Mathematical Models and Neural Networks

Abstract

Abstract Consuming a high amount of energy in industrial dryer makes the drying as an important and highly energy-intensive industrial process. In this study, the required energy for drying of apple slice was supplied from the outlet heat of an engine generation set exhaust. The drying behavior of apple slice was studied in a combined heat and power (CHP) dryer system at four engine load levels (25, 50, 75, and 100%) in order to provide different temperatures (50, 65, 80, and 95°C), and at three levels of drying product thickness (3, 5, and 7 mm) with the constant air flow velocity of 1 m/s. The empirical data from experiments with variants of semi-theoretical and empirical models were evaluated and finally, a suitable model proposed by Midilli et al. was selected to be the best model, as for as the RMSE, R2, and χ2 criterion is concerned. Comparing the results from implementing artificial neural network (ANN) and mathematical models, it was found that the dynamic ANNs is more powerful for modeling the drying process of apple slice in a CHP dryer system than static ANNs and mathematical models.