Design and modeling of hot compressed water extraction process for palm oil

Abstract

The steady state HCWE process for palm oil and the minor component have been modeled using the multivariable regression method. The MISO strategy model with two independent variables; pressure and temperature and four process outputs; overall oil yield, ß-carotene, a-tocopherol and a-tocotrienol concentration have been employing to getting four multivariable equations in this process. At the first step, three different models including linear and nonlinear, have been tested and analyzed to getting the best fit model based on the adjusted r2, residue, p value and RMSE value. Three process outputs that is overall oil yield, ß-carotene and a-tocotrienol concentration was successfully being predicted using the polynomial equation. Meanwhile, it's found that all of these simple models are not enough to model the concentration of a-tocopherol. The selected model will be utilized for the process optimization in the development of the palm oil HCWE control system.