BELNA, Maëllis; NDIAYE, Amadou; TAILLANDIER, Franck; FERNANDEZ, Christophe; AGABRIEL, Louis; GÉSAN-GUIZIOU, Geneviève

doi:10.1016/j.eswa.2022.117624

Métadonnées

Afficher la notice complète

Licence d’utilisation du document

BELNA, Maëllis
Institut de Mécanique et d'Ingénierie [I2M]

NDIAYE, Amadou
Institut de Mécanique et d'Ingénierie [I2M]

TAILLANDIER, Franck

Langue

Article de revue

Ce document a été publié dans

Expert Systems with Applications. 2022-11-01, vol. 205

Résumé en anglais

Optimizing food processes is a complex task made harder by gaps in knowledge on process performance mechanisms, by the complexity inherent to the food product itself, and by the many heterogeneous variables involved in prediction models. Microfiltration of skim milk with a 0.1-µm pore-size membrane is a typical example (MF). MF is commonly used as a unit operation to separate the two major milk protein into valuable fractions for cheesemaking and food formulations. However, despite its importance for the dairy industry, the MF process has never been optimized to integrate conflicting stakeholder objectives such as maximizing quality of product outputs while minimizing cost inputs and addressing environmental impacts. This work addressed the multiobjective optimization of 0.1-µm skim milk MF by considering conflicting stakeholder-defined objectives and integrating expert and scientific knowledge into the formulation of the multiobjective problem. The multiobjective MF problem was modelled by considering the quality of product outputs, operating variables, process design and cost inputs, and using both scientific data and expert knowledge. Over a thousand Pareto-optimal solutions were found, including solutions close to current industry practice but also innovative new solutions. This work opens new perspectives for using multiobjective optimization techniques to design and optimize food processes.< Réduire

Mots clés en anglais

Dairy sector

Expert knowledge

Microfiltration

Modelling

Multiobjective optimization

Real-world problem

URI

https://oskar-bordeaux.fr/handle/20.500.12278/172376

DOI

http://dx.doi.org/10.1016/j.eswa.2022.117624

Projet Européen

FEDER contract no. EU000171, INRAE funding agreement 3000129

Unités de recherche

Institut de Mécanique et d’Ingénierie de Bordeaux (I2M) - UMR 5295