Systematic disaggregation: a hybrid LCI computation algorithm enhancing interpretation phase in LCA

Before the advent of large databases, practitioners often lacked data for calculating life cycle inventories, but the actual computation was a straightforward task. Now that databases represent supply chains including feedback loops and several thousand unit processes and emissions, more formalized calculation methods are necessary. Two methods are widely used: sequential method and matrix inversion. They both exhibit different advantages and drawbacks. The present paper proposes a hybrid algorithm combining the advantages of both methods while minimizing their inconveniences. Sequential algorithm requires a form of cutoff criteria, as the supply chains are of infinite length in the presence of feedback loops. The proposed implementation allows the detailing of individual paths until their upstream contribution falls below a user-defined disaggregation criteria, while also allowing the total impact scores of all paths to be stored and considered. The output is then structured to facilitate consultation and re-aggregation, enhancing the work of practitioners in the interpretation phase of LCA. The algorithm is a variation on structural path analysis and accumulative structural path analysis. It is computationally efficient and uses a reporting threshold criterion based on multiple impact categories. Although the algorithm leads to a more voluminous inventory than matrix inversion, it produces detailed, useful information on the particular instances of processes responsible for the impacts. An average laptop can compute the results within seconds. This algorithm has the potential to improve the interpretation phase of LCA. More specifically, selective replacement of values (characterization factors, input from technosphere, or emission intensities) in parts of the process tree can be applied without affecting the rest of the system. LCA software would benefit from the inclusion of the algorithm presented in this paper. It produces additional information on the structure of the supply chain and the impacts of its constituents, which would be available for a more in-depth interpretation by practitioners. Its potential for understanding the propagation of uncertainty and acceleration of Monte Carlo assessment should also be investigated.