AI-powered platform for proactive grain quality monitoring for sustainable agro-industrial development
The project develops an AI-powered, multiparametric system for early detection of spoilage processes in grain silos. Unlike today’s reactive thermometry solutions that detect problems only after overheating occurs, our innovation integrates temperature, humidity, CO2, VOCs and biogases (H2S, NH₃, CH₄) into a unified physics-informed analytical model. This enables forecasting biological activity 24–72 hours before critical conditions arise. The system combines new sensor nodes, physics-informed AI models, a 3D digital twin of the grain mass, and an operator dashboard providing real-time analytics, early-warning alerts and energy-optimised aeration recommendations. The solution will be validated through laboratory tests and full-scale pilot installation at an operational grain elevator. INNOVINNPROM LLC, will develop sensor integration, data collection pipelines, ML forecasting models (LSTM/GRU), IoT dashboard, pilot deployment and a scalable business model. The project aims to reach TRL 7 by the end of the implementation period. Expected impact includes a 20–30% reduction of grain losses, 10–15% reduction of ventilation energy consumption, improved quality preservation and enhanced resilience of grain supply chains. The project will accelerate the digital transformation of agro-industrial storage infrastructure in Ukraine and the EU and create a commercially scalable solution with high patentability potential.
4/6 complete
Imported from OpenAIRE Ukraine | Grant 101312974.
Open sourceOverview
Grant agreement 101312974 | PQ-GRAIN
Open sourceThe project develops an AI-powered, multiparametric system for early detection of spoilage processes in grain silos. Unlike today’s reactive thermometry solutions that detect problems only after overheating occurs, our innovation integrates temperature, humidity, CO2, VOCs and biogases (H2S, NH₃, CH₄) into a unified physics-informed analytical model. This enables forecasting biological activity 24–72 hours before critical conditions arise. The system combines new sensor nodes, physics-informed AI models, a 3D digital twin of the grain mass, and an operator dashboard providing real-time analytics, early-warning alerts and energy-optimised aeration recommendations. The solution will be validated through laboratory tests and full-scale pilot installation at an operational grain elevator. INNOVINNPROM LLC, will develop sensor integration, data collection pipelines, ML forecasting models (LSTM/GRU), IoT dashboard, pilot deployment and a scalable business model. The project aims to reach TRL 7 by the end of the implementation period. Expected impact includes a 20–30% reduction of grain losses, 10–15% reduction of ventilation energy consumption, improved quality preservation and enhanced resilience of grain supply chains. The project will accelerate the digital transformation of agro-industrial storage infrastructure in Ukraine and the EU and create a commercially scalable solution with high patentability potential.