The experimental design and statistical analysis were created and performed using Design-Expert 8.0 (Stat-Ease, Inc, Minneapolis, USA) . Response surface methodology with a five-factor, three-level Box-Behnken design (BBD) was utilized to study the effect of the hydrolysis parameters on the FFA contents in the products. Five independent factors, namely temperature (X 1 ,℃) , water amount (X 2 , ％) , DESs amount (X 3 , ％) , lipase dosage (X 4 , ％) , and reaction time (X 5 , h) were used, and three levels were applied to each factor (Table 2) . The ranges of the variables investigated were selected according to the preliminary experiments. The response was FFA content in the product mixture. The five replicates of the center points assay the repeatability of the designed method and used to examine the experimental error. The relationships between the response and the variables were expressed as a quadratic polynomial regression model. The mathematical model proposed for the response of Y was:
where X i and X j represent the independent variables; β 0 , β i , β ii , β ij are the model intercept, linear term coefficient, quadratic term coefficient, and the interaction regression coefficient, respectively. The fit of the model was analyzed by coefficients of determination (R 2 ) and the analysis of variance (ANOVA) using Design-Expert 8.0 (Stat-Ease Inc., Minneapolis, USA) .
where X i and X j represent the independent variables; β 0 , β i , β ii , β ij are the model intercept, linear term coefficient, quadratic term coefficient, and the interaction regression coefficient, respectively. The fit of the model was analyzed by coefficients of determination (R 2 ) and the analysis of variance (ANOVA) using Design-Expert 8.0 (Stat-Ease Inc., Minneapolis, USA) .