NutriTwin: a digital twin for more sustainable pig production

Why this project?

The most promising way to reduce the carbon footprint of pig meat and to increase the competitiveness of our production system is by improving feed efficiency and making rational choices regarding feed ingredients. For example, the feed sector aims to source half of its raw materials from by‑products of the food and biofuel industries by 2030. However, the selection of feed ingredients with a low ecological footprint must be carefully balanced against potential effects on production efficiency and feed costs, to ensure that overall improvements in carbon footprint remain achievable and economically viable. Further efficiency gains can also be achieved by capturing changes in nutrient requirements over time (e.g., farm‑specific seasonal trends), allowing feed composition to be tailored to those needs. The industry therefore urgently needs more knowledge and tools to accurately calculate and predict the effect of practical interventions on pig performance and the carbon footprint of pig meat.

Research approach and expected results

This cSBO project investigates how nutrient flows on pig farms can be modelled in a dynamic digital representation of the production process, i.e., a 'digital twin'. By continuously supplying up‑to‑date, farm‑specific training data to a digital‑twin architecture, the generic 'twin' evolves into multiple digital twin versions, each tailored to the specific pig farm from which the data originate. This digital twin captures the animals’ nutrient requirements and utilization efficiency according to the latest insights on that farm, enabling us to learn how each farm evolves over time and which feed strategies are most successful.

By additionally integrating dynamic and consequential Life Cycle Assessment (LCA) models into the digital twin, NutriTwin enables farm‑specific scenario simulations aimed at reducing the carbon footprint of pork production. To maximise efficiency and sustainability, the NutriTwin architecture will be built on readily available data streams throughout the across chain, and will therefore focus on removing key barriers to data interoperability and data privacy. For example, a federated‑learning framework will be developed that allows model training among chain stakeholders without sharing raw data.

Finally, an optimization algorithm will be developed that balances feed costs, animal performance, and ecological footprint. The result is a tool that demonstrates that cost efficiency can go hand in hand with a lower carbon footprint in pig production.

Target group

The project primarily targets companies in the pig production value chain: pig farmers, compound feed manufacturers, meat processors, and integrators.

Project partners

• UGent_LANUPRO (Project manager: Prof. Dr. Jeroen Degroote) 
• UGent_Biovism Lab (Prof. Dr. Jan Verwaeren)
• ILVO_T&V (Dr. Stephanie Van Weyenberg, Dr. Veerle Van linden)
• ILVO_Dier (Prof. Dr. Sam Millet and Dr. Sophie Goethals)

Flanders’ FOOD manages the project.

Participation

The project started on 1 December 2025 and runs until 30 November 2029. To ensure the highest possible relevance of the results, the project includes an Industrial Advisory Board (IAB). Participation in this board is possible until the end of the project.

What does participation in the advisory group involve?

• The IAB meets in person twice a year to receive updates on the latest project results and to provide input for the ongoing research. In total, eight in‑person meetings are planned over the four‑year period.
• During these meetings, you can actively contribute to shaping the direction of the research.

Participation fee and conditions

Participation requires a project contribution, which depends on company size:

• SMEs: €250 per year 
• Large enterprises: €1000 per year

For organisations joining later in the project, the participation fee will be charged retroactively.