Knowledge Graph-Based Mapping and Recommendation to Automate Life Cycle Assessment
Tao Peng,Lu Gao,Reuben S. K. Agbozo,Yuming Xu,Kateryna Svynarenko,Qi Wu,Changpeng Li,Renzhong Tang
DOI: https://doi.org/10.1016/j.aei.2024.102752
IF: 8.8
2024-01-01
Advanced Engineering Informatics
Abstract:The increasing global attention on environmental issues has heightened the demand for a quantitative evaluation of environmental impacts, which primarily relies on using Life Cycle Assessment (LCA). Despite the availability of foreground data from enterprises and the use of existing LCA software to streamline the executing LCA process, LCA practitioners still grapple with the time-consuming task of querying domain knowledge, selecting background data, and entering numerous parameters into the LCA software. To enhance the efficiency and overall performance of LCA-based evaluations, this paper introduces a knowledge graph-based method towards automated LCA. This method aims to recommend background datasets, encompassing flows and processes, and to automate life cycle modeling and LCA calculations. The proposed approach demonstrates significant improvements, with a flow recommendation Precision@10 of 79.52%, surpassing current search engines by 4.26 times, and a corresponding 2.45 times reduction in response time within the Top 10 results. Furthermore, processes are ranked based on the knowledge graph and geographical inclusion relationships. This aids in extracting system boundaries and functional units, facilitating process recommendations and sup-porting decision-making. After obtaining the life cycle inventory (LCI), an open-source LCA software, OpenLCA, is utilized to extend and refine the automated life cycle modeling and calculations. The proposed method is validated through a case study on an electrical product, and a prototype system is designed to ensure straightforward result interpretation. In conclusion, this method can efficiently select background dataset, automate life cycle modeling and LCA calculation, and improve the readability of LCA results.