Optimization of Caper Drying Using Response Surface Methodology and Artificial Neural Networks for Energy Efficiency Characteristics

One of the essential factors for the selection of the drying process is energy consumption. This study intended to optimize the drying treatment of capers using convection (CD), refractive window (RWD), and vacuum drying (VD) combined with ultrasonic pretreatment by a comparative approach among artificial neural networks (ANN) and response surface methodology (RSM) focusing on the specific energy consumption (SEC). For this purpose, the effects of drying temperature (50, 60, 70 °C), ultrasonication time (0, 20, 40 min), and drying method (RWD, CD, VD) on the SEC value (MJ/g) were tested using a face-centered central composite design (FCCD). RSM (R2: 0.938) determined the optimum drying-temperature–ultrasonication-time values that minimize SEC as; 50 °C-35.5 min, 70 °C-40 min and 70 °C-24 min for RWD, CD and VD, respectively. The conduct of the ANN model is evidenced by the correlation coefficient for training (0.976), testing (0.971) and validation (0.972), which shows the high suitability of the model for optimising specific energy consumption (SEC).