AgenAI-ProcessAg: Intelligent AI models for real-time spectral data analysis and optimization in agro-industrial process monitoring

• Grant number: 25/20596-2
• Support Opportunities: Research Grants - Innovative Research in Small Business - PIPE
• Start date: February 01, 2026
• End date: January 31, 2028
• Field of knowledge: Interdisciplinary Subjects
• Principal Investigator: Wokimar Teixeira Garcia
• Grantee: Wokimar Teixeira Garcia
• Principal investigators: Felipe de Lima Rosa ; Isabela Ordine Pires da Silva Simões ; Jaime Finguerut

Abstract

The AgenAI-ProcessAg project proposes the development and validation of an intelligent system for real-time monitoring and optimization of agro-industrial processes, with initial application in the sugar-energy sector. The initiative will be carried out by the Brazilian company ITC - Instituto de Tecnologia Canavieira, responsible for the selection, installation, and calibration of NIR sensors at critical process points (syrup at the cooker inlet and molasses at the crystallization outlet), and by the Canadian company FactR Limited, which will adapt its DataPeak digital platform for continuous ingestion and analysis of spectral and operational data.Calibration campaigns will be conducted under varied operating conditions, ensuring robust predictive models across different cane qualities, processing strategies, and sugar types produced (VHP, VVHP, and domestic market). The spectral and laboratory data obtained will be integrated into the DataPeak digital platform, enabling the development of machine learning algorithms to estimate critical indicators such as stream purity, sucrose balance, and molasses exhaustion.Based on these models, it will be possible to generate process performance indices and automatic operating recommendations, establishing an "Intelligent Virtual Operator" capable of continuously learning from data, supporting real-time decisions, and reducing sucrose losses. The integration of NIR sensing, artificial intelligence, and a no-code digital platform will enable the creation of a predictive control environment, with direct impact on industrial efficiency, sustainability, and competitiveness of sugar mills.The project is positioned between TRL 4 and TRL 6, encompassing sensor installation and calibration, development of predictive models, integration into the digital platform, validation in an operational environment, and performance assessment. The resulting solution will have high replication potential in other continuous agro-industrial processes, consolidating itself as a groundbreaking innovation in online spectral monitoring and intelligent control in the sector. (AU)