Adaptation d'une base de donnees pour l'integration d'un algorithme de resolution d'inventaire en ACV dynamique

Life cycle assessment (LCA) is an environmental tool, which takes into account the potential environmental impacts of a product over its entire life cycle, from raw material extraction to endof- life. This approach allows the identification of the major contributors to the environmental impact of a product, and may support decision-making between two technologies or two products based on environmental criteria. Dynamic LCA is an LCA that takes into consideration the temporal profile of emissions and extractions during a product life cycle. Then, using the dynamic characterization factors, impacts are assessed over a fixed and finite time horizon in a coherent temporal framework. To represent complex systems with infinite supply chains, LCA practitioners often use life cycle inventory (LCI) databases. However, the temporal information available in LCI databases isn't sufficient for dynamic LCA. Therefore, it is not possible in dynamic LCA, for now, to take into account the temporal profile of emissions and extractions occurring in background processes of product systems.;This thesis aims to evaluate how sensitive dynamic LCA results are to the addition of temporal information to background processes in order to elaborate recommendations to LCI database developers about the integration of new temporal parameters for dynamic LCA. The research objective is to develop a methodology to show the importance or not to introduce temporal information in the background system for dynamic LCA results and to apply it to study cases for the global warming impact category.;The selected unit processes of ecoinvent v2.2 database for temporal differentiation were: infrastructure processes (14% of the database), crop and forestry processes (3% of the database), stored products (4% of the database), and disposal to landfills (2% of the database). In the end, temporal parameters were added to 22% of the database unit processes. These parameters were: replacement material proportion, rotation time, storage time, construction time, infrastructure occupation, and demolition time.;The temporal distribution of elementary and intermediate flows was estimated with phases defined by temporal parameters (e.g. construction, use and end-of-life phases). These distributions were considered constant over time, with the exception of carbon uptake and emissions to air from landfills. For reasons of computing time, only the elementary and intermediate flows in the first thirteen-levels of the supply chain were temporally differentiated for the implementation of dynamic LCA.;The research hypothesis was verified; for 8.6% of the 4034 product systems of the ecoinvent v2.2 database, the temporal differentiation of background processes changed global warming results, compared to dynamic LCA that considers the temporal profile of foreground processes only, by at least 10%. The LCA results that were the most sensitive to temporal differentiation of background unit processes were those from product systems in the categories " Energy, wood ", " Material, wood ", " Material, paper ", " Infrastructure " and " Electricity ".;The main limitations of the method were 1) the research objective was verified only for the global warming category, 2) the uncertainty of the temporal data was not evaluated, 3) the results obtained with dynamic LCA with only foreground processes temporally differentiated were sensitive to the positioning of allocation unit processes, and 4) 2.3% of the results in dynamic LCA with foreground and background processes temporally differentiated had 90% and less of the impact due to the impact of temporally differentiated elementary flows.;In conclusion, it is recommended that LCI database developers integrate new temporal parameters, as the ones proposed, to allow the implementation of dynamic LCA with the consideration of the temporal profile of emissions and extractions of background processes for global warming impacts. Moreover, LCI database developers should pay attention to the coherence in the temporal considerations of the processes during modelling. Furthermore, it is recommended to LCA practitioners to temporally differentiate for dynamic LCA the background of the product systems from the sensible sectors identified, primarily for the temporal profiles from the crop and forestry phase.