Quality in Every Gram: AI-Driven Raw Material Analysis for Digital Twin-Based Food Production
| Reference number | |
| Coordinator | Proteinish AB |
| Funding from Vinnova | SEK 5 000 000 |
| Project duration | April 2026 - March 2028 |
| Status | Ongoing |
| Venture | Advanced digitalization - Industrial needs-driven innovation |
| Call | Industrial applied AI by advanced digitalization 2026 |
Purpose and goal
The project develops an AI-based real-time system for consistent quality in manufacturing textured plant proteins. Swedish raw materials vary significantly between batches — beyond current process control capabilities. A digital twin analyses raw material properties via industrial sensors and generates control recommendations for the extruder. Validated in pilot environment, the system targets 10–30% waste reduction and a scalable platform for AI-driven process control in the food industry.
Expected effects and result
The project is expected to reduce raw material waste by 10–30% and improve product consistency in the manufacturing of textured plant proteins. Reduced waste translates directly to lower climate impact and increased resource efficiency. The technical platform is scalable and broadly applicable across the food industry, strengthening Swedish competitiveness in sustainable production. Scientific results will be published openly to enable dissemination across the sector.
Planned approach and implementation
The project is implemented in five work packages over 24 months. Initially, raw materials are characterised through spectroscopy and laboratory analysis to establish a data foundation. AI models are then developed, trained on collected data, and integrated in real time with the production system via industrial system integration. The solution is validated in an industrial pilot environment. The project is managed by the coordinator with continuous reporting to Vinnova.