Pressurized liquid extraction of Orthosiphon stamineus oil: experimental and modeling studies

Abstract

Extraction of Orthosiphon stamineus oil has been the subject of current study. In this case first based on Box-Behnken experimental design method, experimental work was carried out to find the effect of temperature, extraction time and the number of extraction cycles on extraction yield. Seventeen different experimental data were obtained and response surface modeling (RSM) was employed to find relation between extraction yield and process variables. A second order polynomial based on statistical analysis with 95% confidence limits was found as the best estimator of extraction yields. In the next step of the study, artificial neural network (ANN) as a soft computing method was applied to predict the oil yield. A multilayer perceptron (MLP) was used in this study. In order to implement an ANN, temperature, extraction time and the number of extraction cycles were selected as input variables and oil yield was considered as target variable. 70% of data were utilized for training and 30% of the remaining data were used for testing the best obtained network. The results illustrated that ANN method is more reliable than RSM method for extraction prediction and optimization. The optimum operating conditions were found at 100 °C, 10 min and 2 cycles.