Analysis of crystallization phenomenon in Indian honey using molecular dynamics simulations and artificial neural network

Abstract

Molecular dynamics simulation was performed on sugar profile and moisture content-based mixture systems of six Indian honey samples. Comparative studies were performed to understand the interactive effects of fructose, glucose, sucrose, maltose and water on crystallization. All simulations led to formation of stable crystal but with different interaction energies. Post-simulation analysis showed that Fructose:Glucose of 1.18 formed the most stable crystal with highest van der Waals and electrostatic interactions. The stability of crystal was further validated with least gyration radius (209 ± 1.81 nm²), accessible surface area (4.09 ± 0.04 nm) and root mean square displacement (3.51 ± 0.00261 nm). Results indicated that not only Fructose:Glucose ratio but also sucrose, maltose and water had a significant effect on the overall crystallization process. The simulation data was used to train the artificial neural network which predicted the stability of honey crystallization depending on Fructose:Glucose and Glucose:Water ratios.

Keywords: Artificial neural network; Crystallization; FG ratio; Honey; Molecular dynamics.