Emission Factor Recommendation for Life Cycle Assessments with Generative AI

Area of Science:

• Environmental Science
• Computer Science
• Industrial Ecology

Background:

• Accurate greenhouse gas (GHG) quantification is vital for environmental impact assessment and mitigation strategies.
• Life Cycle Assessment (LCA) relies on emission factors (EFs) for estimating indirect emissions, a process currently manual, time-consuming, and prone to errors.
• Manual EF selection requires significant expertise and can hinder scalability in environmental reporting.

Purpose of the Study:

• To develop and validate an AI-assisted method for automated GHG emission factor (EF) recommendation.
• To improve the accuracy, efficiency, and scalability of EF selection in Life Cycle Assessment (LCA).
• To support organizations in their sustainability initiatives and progress toward net-zero emissions goals.

Main Methods:

• Utilized natural language processing (NLP) and machine learning (ML) to create an algorithm for automatic EF recommendation.
• Developed a system that provides human-interpretable justifications for recommended EFs.
• Implemented a tiered approach allowing expert assistance or fully automated EF selection.

Main Results:

• The AI-assisted method achieved an average precision of 86.9% for correct EF recommendation in fully automated mode.
• The method identified the correct EF within the top 10 recommendations with an average precision of 93.1%.
• Demonstrated effectiveness across multiple real-world datasets, confirming the method's robustness.

Conclusions:

• The AI-assisted approach significantly streamlines the EF selection process in LCA.
• This method enables more scalable and accurate GHG emissions quantification, facilitating corporate sustainability efforts.
• The tool supports organizations in achieving net-zero emissions targets by improving the reliability of environmental impact assessments.