A statistical screening technique to derive simplified characterization models between life cycle inventory (LCI) flows and life cycle impact assessment (LCIA) indicators: An application for the ADP indicator of construction materials and products - SF&S Accéder directement au contenu
Communication Dans Un Congrès Année : 2010

A statistical screening technique to derive simplified characterization models between life cycle inventory (LCI) flows and life cycle impact assessment (LCIA) indicators: An application for the ADP indicator of construction materials and products

Résumé

The Life Cycle Assessment (LCA) methodology is known to be a powerful tool for the analysis of the main flows of energy, matter and resulting emitted pollutants within a given system boundaries of any products or processes. Yet, to date, it is often seen as a complex task to carry out especially because of the huge amount of data - at the inventory step - to deal with. In this paper, we investigate some statistical methods to produce simplified characterization models (SCM) for the usual life cycle impact assessment (LCIA) indicators. The methodology aims at detecting the most significant and robust LCI flows that predominantly contribute in the calculation of each LCIA indicator. The problem cannot be solved using classical stepwise regression techniques due to the type of data. We explored a French LCI database for the building sector gathering 108 generic materials data and 376 EPDs data based on a same LCI model and expressed as kg or m² of functional unit [1]. First, different indicators are calculated for each material and product in the database using full additive models and second, the level of significance of each LCI flow is assessed using its statistical p.value and transforming it into a rank score. This second step is repeated 10000 times on resampled subsets of the database. This technique proves to be quite efficient to detect the most robust contributors for a given LCIA indicator. An application for the ADP indicator limited to the non energetic resources (ADP-CRNE) shows the feasibility of deriving a simplified characterization model (SCM) with the help of only 8 resources namely the boron, the cadmium, the chromium, the copper, the tin, the molybdenum, the lead and the zinc. We then attempt to validate the SCM on two building case studies in order to assess whether the weighting of the functional units (kg, m²) of materials and products results in deviation of the SCM. It was finally found that the relative deviation between the detailed ADP-CRNE and the simplified one is only 1,86% for the masonry constructive system (62.37 vs. 61.21 kg eq-Sb) and 1,47% for the wooden constructive system (2.77 vs. 2.68 kg eq-Sb).
Fichier non déposé

Dates et versions

hal-00952137 , version 1 (26-02-2014)

Identifiants

  • HAL Id : hal-00952137 , version 1

Citer

Sébastien Lasvaux, Bruno Peuportier, Jacques Chevalier, Philippe Garat. A statistical screening technique to derive simplified characterization models between life cycle inventory (LCI) flows and life cycle impact assessment (LCIA) indicators: An application for the ADP indicator of construction materials and products. LCA X - Bridging Science, Policy, and the Public, Nov 2010, Portland, Oregon, United States. ⟨hal-00952137⟩
394 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More