Digital twins in food processing: A conceptual approach to developing multi-layer digital models

The modern food processing industry faces complex and multi-faceted decisions from the day a new food product/process is formulated to its commercial-scale production. Key challenges that must be considered during the lifecycle include shifting consumer behaviour, circular production goals, food safety and sensory quality demands. Reviewing the state-of-art in digitalization of food processing shows that modelling, sensors and big data analytic methods are available to address these needs individually. In this work, we propose a multi-scale digital model concept to support complex decisions that must be made through the product lifecycle, unlocking the full benefit of the digitalization drive. The proposed multi-scale concept consists of 5 separate modelling layers that captures both consumer behaviour (at sensory scale and consumer sales scale) and production plant behaviour (at phenomena, unit operation, and plant level) which is then connected to make a single multi-scale model. Two conceptual case studies of cream cheese fermentation and meat freezing are used to describe the information flow between the layers. Technical challenges related to acquisition and curation of data at multiple scales and the models economic value proposition are also discussed.