Integrating CFD-based digital twins into an IoT platform for enhanced greenhouse temperature spatial distribution monitoring

Greenhouse microclimates are highly heterogeneous, yet capturing this spatial variability typically requires dense sensor networks that are impractical due to cost and maintenance. Consequently, climate management often relies on incomplete spatial data. To address this, a synchronized Digital Twin (DT) was developed to characterize greenhouse microclimate behavior by generating air temperature estimates at 39 distributed locations. A key contribution is the integration of a CFD-based DT into a standardized, scalable IoT ecosystem, enabling continuous synchronization and operational deployment beyond offline simulation. The DT utilizes a CFD model in ANSYS FLUENT coupled with a nodal energy balance model to represent thermal dynamics under real operating conditions. This virtual model is deployed within a modular IoT architecture based on the FIWARE platform and OMA NGSI standards. Through the iVeg platform, physical sensors, actuators, and CFD-derived virtual sensors are exposed via Representational State Transfer Application Programming Interface (REST API) and an interactive dashboard, ensuring seamless interaction between the physical greenhouse and its virtual counterpart. Experimental validation demonstrates satisfactory agreement between simulated and measured temperatures, with offline validation achieving Mean Absolute Errors (MAE) values between 0.74 ∘ C and 1.04 ∘ C (RMSE: 0.94–1.35 ∘ C; NMAE: 3.55–5.28 %), and online validation achieving MAE values between 1.13 ∘ C and 2.12 ∘ C (RMSE: 1.45–2.35 ∘ C; NMAE: 5.3–9.3 %) across all validated sensor locations and contrasting meteorological conditions. By extending spatial temperature monitoring without additional hardware, this DT generates spatially distributed temperature estimates that complement physical measurements and establishes a foundation for predictive monitoring and decision support in greenhouse climate management.