From solar heat to digital drying: integrating computational fluid dynamics, process optimisation, and digital twins for sustainable tuber-chip drying

Root and tuber crops such as cassava, yam, and sweet potato are central to tropical food systems but remain highly vulnerable to postharvest deterioration because of their high moisture content. Solar drying offers a renewable route to stabilisation, yet practical performance is constrained by fluctuating weather, non-uniform airflow, spatial temperature and humidity gradients, product loading, and strong coupling between external convection and internal moisture transport. This structured narrative mini-review, informed by SANRA and PRISMA reporting principles, synthesises evidence on active solar drying, computational fluid dynamics (CFD), experimental validation, response-surface and multi-objective optimisation, artificial intelligence (AI), sensor systems, and digital twins for tuber-chip processing. Six databases were searched, sources were screened against explicit eligibility criteria, and each of the 45 cited sources was classified by its relationship to tuber-chip drying so that tuber-specific evidence is separated from evidence transferred from other foods and dryer types. Only 12 of the 45 sources were tuber-specific; they concentrate on yam drying kinetics, cassava dryer performance, and one integrated machine-learning and 4E study of sweet-potato solar drying. CFD validation, robust optimisation, surrogate modelling, and digital-twin architectures are, in the evidence assembled, largely transferred from other products or from unloaded or non-tuber systems. Reported effective moisture diffusivities for yam span roughly 6.4 × 10 −9 to 1.6 × 10 −7 m 2 s -1 (up to 4.2 × 10 −8 m 2 s -1 for water yam); an indirect solar dryer for sweet potato achieved 15% overall thermal efficiency, 4.5 kWh per kilogram of moisture removed, and a 29-month payback; and a CFD–ANN–genetic-algorithm framework reproduced no-load airflow and temperature fields to within 7.3% error. Because these values arise from different dryers, products, and metric definitions, they are compared critically rather than ranked. The evidence suggests that CFD is most valuable when it resolves spatial transport phenomena that cannot be inferred from inlet and outlet measurements, and that AI is most defensible when it accelerates validated physics rather than replacing it. A prioritised research agenda is proposed, beginning with tuber-specific property data and loaded-dryer CFD validated against product moisture, and progressing conditionally to robust optimisation, validated surrogates, digital shadows, and multidimensional sustainability assessment for cassava, yam, and sweet-potato chips.