Résumé
The complexity of some food-packaging systems requires mathematical models of mass transfers to accurately predict the evolution of headspace gas composition during storage and better anticipate food shelf-life. In this study, a numerical 2D model of O 2 /CO 2 transfer in a cheese packaging system was proposed. The results showed significant involvement of gas sorption/desorption in food and permeation through the tray in evolution of gases in headspace. In this context, only a 2D model provided a good fit to experimental results, as gas permeation at the food/tray interface was non-negligible, limiting 1D approaches. This validated 2D model was then used in a global sensitivity analysis employing the Morris method to quantitatively identify the most influential mass transfer parameters affecting prediction accuracy. The results revealed a significant impact of O 2 and CO 2 permeability of the tray and their activation energies, as well as O 2 and CO 2 solubility and diffusivities in food. Therefore, to obtain a robust model of a food packaging system, it is recommended to integrate gas transfers between food and tray when the food-tray interface area is significant (2D model) and to precisely evaluate the gas permeability of the tray. These transfers can be neglected when this interface area is minimal (1D model).