Page 205 - Materials Chemistry, Second Edition

L1644_C05.fm Page 178 Monday, October 20, 2003 12:02 PM

crucial for decision-makers in order to understand the variables to be acted upon.
Also it would be very useful to know the parameters that might be neglected,
especially if it is difficult to get detailed information about them. Sensitivity can be
analyzed by an approach that displays sensitivity as a percentage of contribution
from each parameter to the variance of the final result. Crystal Ball Version 4.0, the
software package selected to perform an example of application for the use of MC
simulation, approximates this approach by lifting to square the correlation coeffi-
cients of ranks and normalizing them to 100%.

5.5 UNCERTANTY ASSESSMENT IN DIFFERENT
ENVIRONMENTAL IMPACT ANALYSIS TOOLS

In this section, methods of uncertainty assessment in different environmental impact
analysis tools are presented.

5.5.1 UNCERTAINTY ASSESSMENT IN LIFE-CYCLE INVENTORY

Figure 5.2 presents an adaptation of the procedure for uncertainty analysis in LCI
as it is reported in the literature (Meier, 1997; Maurice et al., 2000). The first step
refers to the compilation of LCI data. If all the parameters that might have reper-
cussions on the final result were considered, an exhaustive study would need to be
carried out; however, not all these data are relevant. Hence, only the most relevant
factors must be selected and, for some parameters, can be assumed to have fixed
values. Once the essential factors have been selected, a characterization of the
probability distributions is carried out. Therefore, the data are classified into two
groups: extensively available data, for which average and standard deviation can be
calculated, and data based on little information, for which literature and expert

Compilation of Life-cycle
inventory data

Selection of essential factors Neglected parameters

Characterization of
probability distribution
Distribution for data based
on little information
Determination of distribution by literature and experts
for extensively available data

Monte Carlo Simulation Sensitivity Analysis
Analysis and
discussion of results

FIGURE 5.2 Procedure for the uncertainty and variability assessment in the life-cycle inven-
tory. (Reprinted from J. Cleaner Prod., 11, Sonnemann, G.W. et al., pp. 279–292, ©2002
with permission from Elsevier.)

© 2004 CRC Press LLC