Text-based algorithms for automating life cycle inventory analysis in building sector life cycle assessment studies

Abstract

Life Cycle Assessment (LCA) is essential for evaluating the environmental impact of sustainable activities in industry. Despite its importance, there exist challenges negatively impacting its deployment, particularly the time-consuming process of gathering inventory data. This research introduces a novel framework that leverages advanced text-based algorithms from Natural Language Processing (NLP), significantly enhancing the efficiency of data collection in LCA studies. Focusing on the inventory phase, the novelty of this research lies in its ability to reduce data collection time by an estimated 80%–90% compared to conventional methods and improve accuracy by directly extracting materials from bills of quantities (BoQs), which usually list all the construction materials. While our methodology shows promise, it faces challenges due to project complexity, particularly the need for consistent terminology between BoQ and reference databases, though future advancements in matching algorithms may enhance our approach’s efficiency. Real-world case studies demonstrate the framework’s effectiveness, offering flexibility across industries and system complexities.