Biohydrogen production in a fixed-bed reactor using biochar and automatic control via machine learning

Process: 447489/2024-0 (CNPq Call 21/2024)

Funding Modality: Bolsa Conhecimento Brasil

Start Date: 2025

End Date: 2029

Total Funding: R$ 688,999.84

Project Summary

Arrakis Bioenergia is leading an innovative project to develop a bench-scale anaerobic reactor for biohydrogen production through dark fermentation, with a focus on low-carbon solutions, automation, and the use of biochar as microbial support.

By integrating machine learning techniques and real-time sensors (pH, temperature, pressure, and gas composition), the system will be capable of self-adjusting operational conditions to mitigate instabilities and optimize hydrogen generation from organic waste—particularly highlighting vinasse from the sugarcane industry.

Objectives

  • Design and build the automated reactor

    Develop a bench-scale fixed-bed reactor packed with biochar, suitable for dark fermentation of organic

  • Automate real-time control

    Develop a system that automatically adjusts operational variables based on real-time data.

  • Apply machine learning algorithms

    Use supervised and unsupervised learning to predict instabilities and optimize the process.

  • Validate the model

    Test the system using vinasse from ethanol production as a real substrate, assessing stability and yield.