Food Industry 4.0: computational modeling trends for digital twins

In view of the Sustainable Development Goal (SDG) #2 'Zero Hunger', sufficient food production – while simultaneously reducing its environmental footprint – becomes a central challenge as the population expands worldwide. In order to manage ever-shrinking resources, Food Industry 4.0 has increasingly leveraged digital technologies to achieve food security. In this context, digital twins are virtual (i.e., in silico) replicas of real-world agroindustrial systems, designed to numerically simulate their performance and behavior using real-time data (e.g., from sensors). By enhancing food process control and predictive decision-making, digital twins can positively assist real-time optimization and drive innovation, but not without difficulty. Firstly, in view of a comprehensive 'digital-physical' interlink between agroindustrial assets, digital twins are prone to invoke complex mathematical models. Secondly, food production systems change with the prevailing scenario, thus bestowing a stochastic trait to influencing parameters as well as initial and/or boundary conditions for governing differential equations. Thirdly, commercial (off-the-shelf) software costs can be prohibitive for non-academic end-users, mainly small food producers. Also, the inherent complexities and specificities of agroindustrial systems require specialist-oriented simulation code often absent in general-purpose software. As both non-human and human actors are directly involved, an interdisciplinary and symbiotic ethos must be pursued in view of integrated, participatory and efficient digital technology transfer. As this work discusses, computational modeling trends for all-inclusive tailor-made digital twins to support Food Industry 4.0 include: (i) implementation of hybrid in-house simulators of food production systems by suitably merging mechanistic modeling with data-driven simulation; (ii) use of dimensionless modeling to expedite scale-up and optimization; and (iii) validating and transferring novel digital technologies and solutions (e.g., digital twins) toward strategic issues identified by the food producer. Last but not least, (iv) training (education) in using digital tools (including computational fluid dynamics and operations research) must be continuously provided via hands-on courses and workshops. The ultimate goal is to increase productivity, sustainability, and resilience of food production systems, thus upholding not only SDG-2 'Zero Hunger', but also SDG-3 'Good Health and Well-Being', SDG-6 'Clean Water and Sanitation', SDG 9 'Industry, Innovation, and Infrastructure', and SDG-15 'Life on Land'.