Multi-Objective Optimisation of Glycerol Carbonate Production Using Response Surface Methodology and Genetic Algorithms

Abstract

The industrial scale-up of glycerol carbonate (GC) synthesis via transesterification requires balancing competing process objectives, notably maximising product yield while minimising energy consumption and catalyst usage. Traditional single-objective optimisation fails to capture the complex, nonlinear interactions and inherent trade-offs between these competing variables. This study presents a comprehensive multi-objective optimisation framework for GC production over a synthesised BaCaO mixed oxide catalyst, integrating Response Surface Methodology (RSM) with the Non-Dominated Sorting Genetic Algorithm II (NSGA-II). A Central Composite Design (CCD) was employed to evaluate the effects of reaction temperature, catalyst loading, dimethyl carbonate (DMC)-to-glycerol molar ratio, and reaction time. Second-order polynomial RSM models accurately predicted glycerol conversion (R² = 0.985) and GC yield (R² = 0.988). To resolve the inherent trade-offs, the validated RSM models were embedded as objective functions within an NSGA-II framework. The algorithm successfully generated a well-distributed Pareto front, identifying a "Balanced" optimal operating condition (74.8 °C, 3.8 wt% catalyst, 2.8:1 DMC ratio) that maximises GC yield (91.2%) while minimising estimated thermal energy consumption. Furthermore, global sensitivity analysis using Sobol indices revealed that the DMC-to-glycerol ratio exerted the most significant influence on reaction efficiency (Total-effect index = 0.52). Experimental validation of the predicted optimum yielded a relative error of less than 1%. This integrated chemico-mathematical approach provides a robust, scalable decision-making tool for the sustainable and economically viable industrial production of GC.