STATISTICAL OPTIMIZATION AND EXPERIMENTAL ASSESSMENT FOR BIODIESEL PRODUCTION FROM WASTE COOKING OIL

Abstract

In this study, statistical validation methods using model fitting and analysis of variance have been developed for biodiesel production from waste cooking oil. Box–Behnken design for full factorial design has been created to verify the experimental runs for maximum free fatty acids conversion. A catalytic approach for biodiesel production from waste cooking oil using potassium hydroxide has been employed. This work has focused on the transesterification of the free fatty acids in the feedstock to produce fatty acid methyl esters. The influence of three independent factors; methanol to oil molar ratio, potassium hydroxide concentration, and the reaction time on fatty acid methyl esters yield has been investigated. Multivariate approaches have been efficiently applied to develop a statistical model and optimize the independent process variables. Response Surface Methodology and contour plots have been implemented for different parameters’ combinations for maximum FAME conversion. The response has been analyzed following the P and F values of all variables as well as the adjusted and predicted R-squared. A Quadratic model has been developed representing the empirical relationship between the process variables and the response. Different statistical validation techniques have been used to check and validate the adequacy of the predicted models including the analysis of variance at a 95% confidence level as well as model fitting with a p-value equal to 0.0002 and R-squared equal to 0.9960. The factor coefficients have been compared to identify the relative impact of the factors. The properties of the produced biodiesel have been measured and compared with the standard values. The results in optimum conditions have been reported at 5.7:1, 0.61%, and 121 minutes for molar ratio, catalyst loading, and reaction time, respectively. The reported optimum process conditions have been simulated using Aspen HYSYS.