Prediction of Drying Time and Moisture Content of Wild Sage Seed Mucilage during Drying by Infrared System Using GA-ANN and ANFIS Approaches

This study investigated the use of an adaptive neuro-fuzzy inference system (ANFIS) and genetic algorithm–artificial neural network (GA-ANN) for the prediction of drying time and moisture content of wild sage seed mucilage (WSSM) in an infrared (IR) dryer. These models (ANFIS and GA-ANN) were fed with three inputs of IR radiation intensity (150, 250, and 375 W), the distance of mucilage from the lamp surface (4, 8, and 12 cm), mucilage thickness (0.5, 1, and 1.5 cm) for prediction of average drying time. Also, to predict the moisture content, these models were fed with 4 inputs IR power, lamp distance, mucilage thickness, and treatment time. The GA–ANN model structure that used 4 hidden neurons, and modeled the drying time of WSSM with a correlation coefficient (r) of 0.984. Also, the GA–ANN model with 9 neurons in one hidden layer, predicts the moisture content with a high r-value (r=0.999). The calculated r-values for the prediction of drying time and moisture content using the ANFIS-based subtractive clustering algorithm were 0.925 and 0.998, respectively, that shows a higher correlation among predicted data and experimental data. Sensitivity analysis results demonstrated that IR intensity and mucilage distance were the main factors for the prediction of drying time and moisture content of WSSM drying, respectively. In summary, the GA–ANN approach performs better than the ANFIS approach and this method can be applied to relevant IR drying process with satisfactory results.