A three-phase mathematical model to monitor alcoholic fermentation and phenolic extraction of Vitis Vinifera L. red wines
The main goal of this study was to explore the development and validation of a dynamic mathematical model to describe the oenological and phenolic development during the alcoholic fermentation process of red wines. A three-phase model including the cap, must and lees phases was attempted. The model uses grape compositional data to forecast the expected progression of phenolic extraction and production of fermentative products. The likelihood of the calculated parameters was tested on a new set of fermentations to assess model accuracy and robustness. In addition, the model parameters were explored to determine the range of applicability of the proposed mathematical model. For the must phase, the phenolic and some of the most relevant oenological parameters have shown R 2 often higher than 0.8 and mean average predictive error (MAPE) below 20 % indicating comparable levels of accuracy to those presented in literature while also expanding on the number of components of interest. Moreover, model validation has shown that model performance can be affected by fermentation temperature, especially when fermentations are conducted at higher temperatures (28 °C). The mathematical model presented in this study was able to accurately forecast the fermentation kinetics of both phenolic and oenological parameters. The model incorporated the polymerisation of anthocyanins and the transfer of compounds to the lees phase representing a step forward in understanding the kinetics of red wine fermentation and contributes to the development of future methods for monitoring and controlling red wine fermentations. • A three-phase mathematical model to describe red wine fermentations is proposed. • Grape compositional data is used to forecast phenolics and alcoholic fermentation. • Phenolics and relevant oenological parameters in the must showed R2>0.8 and MAPE<20 %. • Anthocyanin polymerisation and transfer of compounds to the lees was modelled.
Authors
- Tobias Louw
- Jose Luis Aleixandre-Tudo
- Wessel du Toit
- Kiera Lambrecht
Institutions
- Stellenbosch University (ZA)
- Universitat Politècnica de València (ES)
Publication Details
- Journal
- Journal of Food Engineering
- Published
- 2025-12-20
- DOI
- https://doi.org/10.1016/j.jfoodeng.2025.112932
- Citations
- 1
- Primary Topic
- Fermentation and Sensory Analysis
- Type
- article
- Field-Weighted Citation Impact
- 1.16