Drying plant-based materials—such as agricultural waste and forestry residues—can account for more than 70% of the total energy required before those materials can be converted into fuels or bioproducts. That front-loaded energy demand has been a persistent barrier to scaling biomass-based energy systems.
The project brings together advanced AI developed at South Dakota Mines with applied energy systems research at INL to explore ways to make the process more efficient, lower-cost, and more viable at commercial scale.
Unlike combustion or conversion technologies, biomass drying is often treated as a preprocessing step rather than a core efficiency challenge. Yet moisture content directly affects downstream performance, transport costs, and overall energy balance.
According to Kazi Khoda, assistant professor of chemical and biological engineering at South Dakota Mines, the inefficiency of current drying methods limits biomass’s competitiveness as a clean energy input.
The research team is examining new approaches that pair recovered waste heat—including heat from nuclear power facilities—with AI-based predictive models to better control how moisture moves through plant materials.
Khoda’s research group, PRocess Optimization, Design InteGration, and Informatics (PRODIGI), is developing physics-informed machine learning models that simulate moisture transport inside biomass. Rather than relying on trial-and-error experimentation, the models allow researchers to test drying strategies virtually before physical systems are built.
This approach is intended to reduce both energy consumption and development risk, particularly for industrial-scale deployment where inefficiencies are magnified.
At INL, the work is led by Nepu Saha, a research scientist and principal investigator on the project, who brings expertise in energy systems integration and applied modeling.
The project reflects a broader shift in clean energy research: efficiency gains are increasingly coming from process optimization and data-driven control, not just new generation technologies.
By addressing energy demand at the preprocessing stage, the research could improve the overall economics of biomass-based fuels and materials—particularly in regions where waste biomass is abundant but energy costs are high.
The work also aligns with national efforts to integrate AI into energy system design while maintaining transparency and physical realism, an area where purely data-driven models have faced skepticism.
In addition to its technical goals, the project includes a workforce development component. Students at South Dakota Mines will have opportunities to collaborate with INL researchers and participate in federally funded work tied to national energy priorities.
Participants will gain hands-on experience in clean energy innovation, data science, and advanced manufacturing applications—skills increasingly in demand as energy systems become more data-intensive.
