Internal Normalization Procedures in the Context of LCA: A Simulation-Based Comparative Analysis

摘要

Normalization is a procedure used to convert absolute values of a system, generally expressed in different measurement scales, into normalized values, thus enabling comparison, ranking, and aggregation of attribute values. In the context of the Life Cycle Assessment (LCA), normalized results can be obtained using internal and external approaches. The latter requires normalization factors gathered within a precise spatial context (e.g., a country), and this data usually originates from environmentally aware nations. However, several countries, such as Brazil, lack this sort of data; therefore, it is more difficult to apply representative external normalization factors. Alternatively, one may apply an internal normalization approach since the analysis of the data is specific to individual assessments, thus simplifying LCA in "non-normalized" countries. Since there are many internal procedures and the literature lacks discussions on how they perform in LCA contexts, it might be challenging for decision-makers to select and apply them as Multiple Attribute Decision Making (MADM) methods. In order to fill this research gap, we performed exploratory research aiming to compare eight procedures of internal normalization through a Monte Carlo Simulation using artificial data. Results indicate that procedures of internal normalization generally present a good performance since they influence the choice of the preferable alternative in 30% of the simulations. Additionally, only two internal normalization approaches have reduced ranking performance. On the other hand, the least influential procedures on the final ranking of alternatives were Vector Normalization and Simple Normalization using the maximum value as a reference.