Computational framework for multi-objective optimization of activated biochar properties using machine learning and evolutionary algorithms

Description

Biochar made from farm residues can store carbon and also serve in energy storage devices, but tuning it for both jobs at once means balancing rival goals. Mohammad Fazle Rabbi builds a simulation framework that searches for the best trade-offs. A random forest model stands in for lab experiments, and an evolutionary algorithm explores process settings.

The search covered 800 combinations of temperature, time, heating rate, activation chemical, and feedstock type. The stand-in model predicted surface area and capacitance with high accuracy, though carbon stability proved harder to model. The optimization found settings reaching a surface area near 1,094 square meters per gram alongside strong carbon dioxide uptake. Rabbi presents the framework as a way to design dual-purpose biochars on a computer first.

Catalog record

Source: https://europepmc.org/articles/PMC13376612