DiAGreen

Crop production in greenhouses supplies most agricultural products in industrialized countries while being one of the most energy-intensive operations in the agricultural sector. In the Mediterranean basin, it is estimated that the energy consumption for heating can achieve 440 kWh/m2/y. Most of the energy used for climate control and carbonic fertilization of these structures comes from fossil fuels. With the recent sharp increase in the cost of fossil energy sources due to the geopolitical threads, the impact on the price of the final product is increasingly important. In the light of the decarbonization (following Mission 2 – Green transition of Italian National PNRR) and to contrast the cost increase in greenhouse crop products, there is the need to disruptively innovate how greenhouses consume energy sources. Although strategies for reducing energy use were well studied in the past (e.g., thermal insulation, shading and HVAC control, renewables), there is still the need to integrate them into a framework able to dynamically respond to drivers such as outdoor (changing) climate, production needs, (variable) costs. At the same time, interpretation of data acquired from new and cost-effective digital devices for monitoring and control-ling the greenhouse environment is essential to feedback on the operation, cultivation, and environmental impact of the greenhouse itself. The DiAGreen project aims at creating and testing digital twins, a new generation of virtual representations of agricultural greenhouses that serves as the real-time digital counterpart of physical greenhouses. The proposed digital twins exploit the capabilities of multi-scale and multi-domain greenhouse models, historical measurements from sensors, optimization approaches, and future predictions for decreasing the energy use and thus the carbon impact of greenhouses, adapting the best strategies to current and future time steps. The digital twin model framework will be built on previous skills of the research consortium and developed based on two test sites to test their capabilities and potential results. In the first test site, the experimentation will be focused on the control of the climate and the energy source shifting. In the second one, the impact of CO2 fertilization will be tested by integrating it into the digital twin and assessing the CO2 capability of speeding up the crop growth (thus enhancing the overall energy efficiency) and, at the same time, the crop quality measured through HPLC analysis. This project, through the development of greenhouse digital twins, makes one step forward toward the Green-house 4.0 Industry. Digital twins can remarkably improve the design and operation of greenhouses, optimized in terms of energy efficiency, productivity, sustainability, and quality of the productions facing the future challenges of fossil fuels availability and costs, climate change, and future food demand.

Results achieved

During the implementation of the project, major challenges concerned the complexity of greenhouse microclimate modelling, the calibration of shallow geothermal energy models, the integration of monitoring data with numerical simulation frameworks, and the need to align advanced modelling tools with the practical and economic expectations of stakeholders. The CFD activities required careful definition of boundary conditions, turbulence modelling assumptions and mesh resolution in order to ensure both numerical stability and acceptable computational times; these issues were addressed through mesh-independence analyses and systematic experimental validation. In the geothermal modelling activities, residual uncertainties linked to soil heterogeneity and seasonal variability highlighted the need for longer-term monitoring, sensitivity analyses and techno-economic optimization. Additional improvements were identified in the direction of automated anomaly detection, redundancy of critical measurements, standardized data quality procedures, continued stakeholder feedback, reduced-order models for real-time applications, and possible use of cloud-based computational resources.

The research team of the University of Bologna contributed to the design, development, deployment, and validation of a low-cost, open-source control and monitoring platform for greenhouse CO2 enrichment, as well as to the implementation of a real-time, model-based CO2 dosing controller supporting the crop experiment. In the design phase preceding the experimental trials, the unit defined the overall system architecture, based on a Raspberry Pi gateway and supervisory controller, Modbus RTU over RS485 communications, Node-RED orchestration flows, InfluxDB time-series storage, and Grafana dashboards for real-time visualization. Prior to and during greenhouse setup for the trials, UniBO supported the selection, integration, and commissioning of low-cost sensing and actuation components, including the NDIR CO2 sensor, RS485 multi-drop communications, and the CO2 flow regulation device connected through Modbus RTU. During the experimental phase, the unit implemented and tested the closed-loop CO2 regulation logic using a PID-family controller in software, with control actions configured around daily CO2 enrichment windows and differentiated sampling frequencies during and outside active operation. Continuous activities during experimentation also included data acquisition, storage, integrity checks, standardized metadata management, dashboard-based real-time decision support, and retry mechanisms to handle communication failures. In October–November 2025, during the basil greenhouse experiment, UniBO implemented on the Raspberry Pi gateway a real-time model-based controller able to compute the optimal CO2 supply rate by coupling canopy assimilation modelling, ventilation-driven loss estimation, and supervisory logic aimed at reducing waste under high ventilation conditions. Throughout the trials, the unit also enforced software safeguards, including input validation, fail-safe zero-supply behaviour in the event of missing or invalid data, conservative indoor CO2 thresholds, and multi-level logging of controller states. During and immediately after the experimental trials, UniBO contributed to the validation of the implemented solution through the analysis of logged time series and performance indicators, with the objective of quantifying controller behaviour, set-point tracking, and resource efficiency.

An analysis procedure was developed for the design and optimization of systems harnessing shallow geothermal energy coupled with heat pumps to heat and cool interior spaces for controlled environment agriculture. The procedure developed led to the evaluation of the geothermal heat and cooling potential of a site and the analysis of energy flows achievable for a given technical configuration. The resulting tool can support the design of optimal shallow geothermal system configurations, considering space constraints, while integrating with indoor environmental control systems for agricultural productions. Additionally, the results are suitable to usefully inform the process of modeling and simulation of renewable energy utilization.

The DiAGreen project has been implemented in full coherence with the objectives and activities originally approved. All Research Units have ensured the timely execution of the planned WPs, maintaining alignment with the scientific, technical and organizational framework defined at the proposal stage.

The results confirm the effectiveness of the integrated multi-domain approach adopted by the project for modelling and analyzing the nexus between indoor environment, crop production and energy use in greenhouses. The development and validation of advanced modelling tools, the implementation of experimental activities at the two test sites, and the realization of a Digital Twin framework for greenhouse systems represent significant progress toward the digitalization, decarbonization and energy optimization of protected cultivation systems.

The collaboration among the research units proved to be effective and scientifically integrated, making it possible to effectively integrate modelling, experimentation, monitoring and control activities. The interdisciplinary nature of the project strengthened the robustness of the outcomes and enhanced their applicability to real greenhouse contexts. The scientific outputs generated during the project have contributed to the dissemination of knowledge and to the consolidation of a national research network on sustainable greenhouse systems.

Overall, the project has achieved its intermediate and final objectives and has established a methodological and technical basis for further development and potential transfer of the proposed solutions. The results obtained are consistent with the strategic goals of Mission 4 and contribute to increase research capacity, innovation and sustainability transition in both the controlled-environment agriculture and energy engineering sectors.

D.D. del MUR n. 104 del 02/02/2022

Codice progetto MUR: 2022FPHNXZ

CUP: J53D23002100006

Coordinatore di progetto: Politecnico di Torino

Ruolo UNIBO: RL

Research areas: Agricultural engineering

Scientific officer: Stefano Benni

Duration: 28/09/2023 - 27/09/2025

Research group: Stefano Benni, Alberto Barbaresi, Francesco Tinti (DICAM), Emanuele Bedeschi.