Free fatty acid reduction in waste cooking oil using biomass derived heterogeneous catalyst and machine learning-driven optimization

Abstract

BACKGROUND: The depletion of fossil fuel reserves underscores the urgent need for sustainable alternatives, and biodiesel derived from waste cooking oil (WCO) emerges as a promising solution. This approach not only addresses the challenges of disposal of waste cooking oil from eateries but also contributes to a circular economy by converting waste into a valuable resource. RESULTS: The present work focuses on the synthesis of a heterogeneous biomass-derived mixed-metal oxide (WO3/SiO2) catalyst for targeting Free Fatty Acid (FFA) reduction in WCO. The properties of the catalyst were analysed using XRD, FTIR, BET, and SEM with EDS analysis. The reaction parameters such as catalyst loading (2 to 4 w/w%), methanol to oil molar ratio (15:1 to 25:1), and the reaction time (2–10 h) were varied while keeping the reaction temperature constant at 65 °C for the FFA reduction reaction. A Machine Learning technique such as Random Forest was used to optimize the reaction parameters. A maximum FFA conversion of 62.16% was predicted by the model at optimum reaction parameters of methanol to oil molar ratio as 19.7368:1, amount of catalyst as 3.6 w/w%, and time of 9.2 h, which agreed well with the experimentally obtained optimal reaction parameters of methanol to oil molar ratio as 20:1, amount of catalyst as 4 w/w%, and duration of reaction as 10 h. Although the FFA conversion is modest, the as-prepared catalyst effectively reduces FFA in used cooking oil below 2% without further modifications. CONCLUSION: This work highlights the effective utilisation of machine learning for optimization of reaction parameters that targeted the reduction of acid value in WCO, thereby facilitating its efficient conversion into biodiesel. © 2025 Society of Chemical Industry (SCI).