Modeling of palm kernel oil from SC-CO₂ extraction / Muhammad Zulfirdaus A’adnan

A mathematical model is developed in order to study the effect temperature and pressure toward the extraction of palm kernel oil using a supercritical carbon dioxide. A sets of secondary data were obtained from previous studies which consist of a sets of temperature, pressure and amount of oil yield. In this study, the artificial neural network is used to simulate the effect of temperature and pressure towards the oil extraction. In order to obtain the desired output, the suitable number of neuron must be selected. To choose the suitable number of neuron, the sets of data were inserted into the neural network model and let to train by manipulating the variable from 1 neuron to 25 neuron. The result for mean squared error for each variable were then tabulated and compared. Based on the result, the least MSE was 18 neuron number. Then, by using the chosen neuron number the simulation were let to run and trained until the best regression is obtained. Regression R Values measure the correlation between outputs and targets. R value of 1 means a close relationship, 0 a random relationship. The regression obtained from the simulation is 0.958. This indicate that there are a good relation between the experimental and theoretical data. The data for experimental and theoretical were then compared and the error obtained is within the acceptable range which is 3.356%