In a major leap toward sustainable energy solutions for Africa, Nigerian researcher, Chiagoziem C. Ukwuoma, has unveiled a cutting-edge deep learning model that predicts hydrogen production from biomass gasification with unprecedented accuracy.
The study, titled “Sequential Gated Recurrent and Self Attention Explainable Deep Learning Model for Predicting Hydrogen Production: Implications and Applicability,” addresses critical challenges in Nigeria and across Africa, including heavy reliance on polluting fossil fuels, environmental degradation in oil-rich regions like the Niger Delta, and the urgent need for clean, renewable energy sources to combat climate change and health crises.
Ukwuoma’s innovative model combines gated recurrent units (GRU) with self-attention mechanisms to capture temporal patterns in biomass data, achieving near-perfect results: a Mean Absolute Error (MAE) of 0.102, Root Mean Squared Error (RMSE) of 0.160, and Coefficient of Determination (R²) of 0.999.
By leveraging explainable AI tools like SHAP and LIME, the model identifies key factors such as the percentage of plastics in mixtures and particle size, making predictions transparent and trustworthy for real-world applications.
This breakthrough uses abundant African biomass, such as agricultural waste from rubber seed shells and plastics, to produce hydrogen, a clean fuel that could reduce greenhouse gas emissions and replace dirty oil-based energy.
The research directly confronts Nigeria’s energy woes, where over 70 per cent of the population relies on fossil fuels amid frequent blackouts and pollution. In the Niger Delta, including Port Harcourt, black soot from oil flaring and illegal refining has caused a 30–50 per cent surge in respiratory diseases, contaminating air, water, and soil for millions.
Ukwuoma’s model enables optimised hydrogen production from local waste, potentially cutting CO2 and methane emissions by 15 per cent while creating jobs in green energy sectors. Across Africa, where biomass potential could reach 110 EJ by 2050, this AI tool supports the continent’s transition to net-zero, aligning with goals like the African Union’s Agenda 2063 and Nigeria’s Energy Transition Plan.
Simplified SHAP analysis from Ukwuoma’s model, showing how plastic percentage and particle size positively drive hydrogen predictions, while equivalence ratio has a negative impact. This transparency aids African industries in sustainable fuel production.
As a Nigerian, I’ve seen how oil pollution devastates communities in the Niger Delta, from black soot choking the air to health crises burdening families, said Ukwuoma, a Lecturer and Researcher at Chengdu University of Technology, Oxford Brookes College, China. My model isn’t just about accurate predictions; it’s about empowering Africa with homegrown solutions. By turning waste into clean hydrogen, we can reduce environmental harm, boost energy security, and foster economic growth without relying on fossil fuels.
Co-authored with experts including Prof. Dongsheng Cai and funded by the Natural Science Foundation of Sichuan and the National Natural Science Foundation of China, the study emphasises practical implications: enhancing biomass gasification efficiency, complying with environmental regulations, and lowering costs for hydrogen as a fuel for transport, power, and industry. In Nigeria alone, scaling this could address the black soot crisis by diverting plastic waste from landfills and reducing flaring, while tapping into Africa’s vast agricultural residues for renewable energy.
Advocates like the Society of Technology and Energy Professionals (STEP) hail the work as vital for holding oil companies accountable and promoting bioenergy. Ukwuoma urges partnerships with African governments, NGOs, and industries to deploy the model for pilot projects in the Niger Delta. The full paper is available https://doi.org/10.1016/j.apenergy.2024.124851 and Ukwuoma is available for interviews to discuss adaptations for African contexts.
About Dr Chiagoziem C. Ukwuoma
Chiagoziem C. Ukwuoma is a Nigerian lecturer and researcher in China, specialising in AI development and applications for energy and environmental challenges as well as medical image analysis. With affiliations at Chengdu University of Technology, Oxford Brookes College, his work bridges AI, machine learning, and Deep learning with sustainable development, focusing on issues relevant to Africa’s industrial growth. For more details or to schedule an interview, please contact: [email protected].
