Evaluation of exergy performance and onion drying properties in a multi-stage semi-industrial continuous dryer: Artificial Neural Networks (ANNs) and ANFIS models

Abstract In this research work, two intelligent methods, namely artificial neural networks (ANNs) and adaptive neuro-fuzzy inference system (ANFIS) are applied to predict the moisture ratio (MR), energy utilization (EU), energy utilization ratio (EUR), exergy loss and exergy efficiency of onion slices drying process by a Multi-Stage Semi-Industrial Continuous Belt (MSSICB) dryer. The experiments are carried out for various temperature levels (40, 55, 70 °C), air velocities (0.5, 1, 1.5 m/s) and belt linear speeds (2.5, 6.5 and 10.5 mm/s). The results demonstrate that by using the high levels of air temperature and velocity along with a low belt linear speed, the values of effective moisture diffusivity ( D e f f ), color change (ΔE), EU, EUR, and exergy loss increase whereas the drying time, specific energy consumption (SEC) and exergy efficiency decrease. The highest values of D e f f , SEC and ΔE are found to be 3.74 × 10-11  m2/s, 484.54 MJ/kg and 23.65, respectively. EU and EUR vary between 0.0641 to 0.2451 kJ/s and 0.0848 to 0.5028, respectively. Moreover, exergy loss and efficiency vary in the ranges [0.0182-0.0555] kJ/s and [62.33- 89.25] %, respectively. Moreover, the results showed that the prediction accuracies of MR, EU, EUR, exergy loss and exergy efficiency by the use of ANNs model were 0.9995, 0.9957, 0.9984, 0.9960 and 0.9979, respectively so that the values of 0.9998, 0.9972, 0.9991, 0.9962 and 0.9985, obtained by application of ANFIS model, respectively. Therefore, the selected ANN and ANFIS models can be used confidently to estimate the exergy efficient drying conditions for a sustainable drying process by MSSICB.