Main content area

A model for streamlining and automating path exchange hybrid life cycle assessment

Stephan, André, Crawford, Robert H., Bontinck, Paul-Antoine
The international journal of life cycle assessment 2019 v.24 no.2 pp. 237-252
automation, case studies, data collection, decision making, greenhouse gas emissions, information processing, input output analysis, life cycle inventory, models, path analysis, supply chain
PURPOSE: Life cycle assessment (LCA) is inherently complex and time consuming. The compilation of life cycle inventories (LCI) using a traditional process analysis typically involves the collection of data for dozens to hundreds of individual processes. More comprehensive LCI methods, such as input-output analysis and hybrid analysis can include data for billions of individual transactions or transactions/processes, respectively. While these two methods are known to provide a much more comprehensive overview of a product’s supply chain and related environmental flows, they further compound the complex and time-consuming nature of an LCA. This has limited the uptake of more comprehensive LCI methods, potentially leading to ill-informed environmental decision-making. A more accessible approach for compiling a hybrid LCI is needed to facilitate its wider use. METHODS: This study develops a model for streamlining a hybrid LCI by automating various components of the approach. The model is based on the path exchange hybrid analysis method and includes a series of inter-related modules developed using object-oriented programming in Python. Individual modules have been developed for each task involved in compiling a hybrid LCI, including data processing, structural path analysis and path exchange or hybridisation. RESULTS AND DISCUSSION: The production of plasterboard is used as a case study to demonstrate the application of the automated hybrid model. Australian process and input-output data are used to determine a hybrid embodied greenhouse gas emissions value. Full automation of the node correspondence process, where nodes relating to identical processes across process and input-output data are identified, remains a challenge. This is due to varied dataset coverage, different levels of disaggregation between data sources and lack of detail of activities and coverage for specific processes. However, by automating other aspects of the compilation of a hybrid LCI, the comprehensive supply chain coverage afforded by hybrid analysis is able to be made more accessible to the broader LCA community. CONCLUSIONS: This study shows that it is possible to automate various aspects of a hybrid LCI in order to address traditional barriers to its uptake. The object-oriented approach used enables the data or other aspects of the model to be easily updated to contextualise an analysis in order to calculate hybrid values for any environmental flow for any variety of products in any region of the world. This will improve environmental decision-making, critical for addressing the pressing global environmental issues of our time.