Application of Fuzzy - NSGA-II for achieving maximum biodiesel yield from waste cooking oil

Abstract

Abstract The main objective of this study was to optimize the parameters necessary for biodiesel production from waste cooking oil (WCO) using fuzzy modeling and Non-Dominated Sorting Genetic Algorithm-II (NSGA-II). To establish the relationship between input parameters (methanol, temperature, reaction time, and catalyst concentration) and the output (biodiesel yield), a fuzzy inference system was utilized. The input variables were normalized and represented by linguistic variables such as "small," "moderate," and "large," while the biodiesel yield was expressed in linguistic variables such as "small," "moderate," "moderately large," and "large." The defuzzification process was employed to obtain crisp values of the predicted output using the Centre of Gravity method, resulting in a mathematical model. This mathematical model was then used as an input function in NSGA-II to determine the optimal input process variables. NSGA-II was selected as it is a versatile and efficient metaheuristic algorithm that incorporates non-dominated sorting, accelerated strategies, and does not require sharing parameters. The developed model exhibited a high R square value of 96.34%, indicating a strong relationship between the input and output variables. The optimal transesterification process for biodiesel production from WCO was determined to be a methanol quantity of 150ml, reaction temperature of 62°C, reaction time of 63 minutes, and catalyst concentration of 7.5 grams. This optimized process yielded a conversion efficiency of 97.36%. Additionally, further experimentation was conducted using the Box-Behnken experimental design method to explore a wider range of input combinations. The maximum combination of input variables resulted in a biodiesel yield of 96.88%. Both the predicted and experimental biodiesel yields met the ASTM standard, with a conversion efficiency exceeding 96.5%. Overall, this study successfully optimized the biodiesel production process from waste cooking oil by employing fuzzy modeling and NSGA-II, leading to high conversion efficiencies and meeting industry standards.