Generalized multi-scale kinetic model for data-driven modelling: Mangifera pajang antioxidant degradation in choline chloride/ascorbic acid natural deep eutectic solvent

Abstract

This paper formalizes a new multi-scale kinetic model for a process influenced by several limiting factors which are multi-scale in nature. The multi-scale kinetic model incorporates all of the dominant factors which can be expressed in the form of first-order kinetics where some of the factors could have delayed effects on the process. The applicability of this new model is demonstrated using a case study of Mangifera pajang antioxidant degradation modelling. The results show that the multi-scale kinetic model can provide improved modelling accuracy, with a Root Mean Square Error (RMSE) of 1.32%. Interestingly, the analysis of the multi-scale model suggests the presence of two groups of antioxidant compounds based on the degradation dynamics in the M. pajang extract. The dominant group of antioxidant compounds contributes about 87% to the total antioxidant concentration, and the degradation of this group is much slower than the non-dominant antioxidant compounds at pH 3. At a high temperature above 60 °C and pH above 7, the fast-degrading group of antioxidants becomes dominant in the extract. Significantly, the proposed multi-scale kinetic model is generic and flexible which can be applied to data-driven modelling of many different processes. • New generalized multi-scale kinetic model for improved modelling accuracy. • Antioxidant degradation modelling over broad range of experimental conditions. • Reveals two groups of antioxidants in the extract. • Provides more accurate half-life value of antioxidants.