Page 256 - Materials Chemistry, Second Edition

L1644_C06.fm Page 229 Monday, October 20, 2003 12:06 PM

It is generally accepted that damages to the manmade environment can be best
evaluated by external environmental costs; however, in the default selections, it has
been decided that this can also be done for damages to the AoP human health. The
natural environment will be assessed by REW (relative exceedance weighted) and
as a global indicator only the GWP is chosen. The default discount rate is 3%. An
intermediate aggregation is selected for damages to the manmade environment and
to human health.
Apart from the described selection of clear weighting schemes in order to obtain
meaningful indicators, it must be acknowledged that, in principle, determining which
dose–response and/or exposure–response functions to use and even which dispersion
model to apply implies indirect value choices that (especially in the case of the
dose– and exposure–response functions) can have very important influences on the
final result. Thus transparency on this point is recommended as well as checking
the preferences of the decision-maker; for instance, it can be said that, in general,
internationally accepted standard values have a high level of reliability. All the
presented criteria will be exemplified later in this chapter through a case study that
has demonstrating the functionality of the weighting and aggregation scheme as one
of its primary interests.

6.5.3 LIFE-CYCLE INVENTORY ANALYSIS
After the goal and scope definition, the life-cycle inventory analysis follows in the
same way as in an LCA according to ISO 14040. An overview of the LCI analysis
with its options is given in Figure 6.10.
If a situation of an existing process chain is assessed, the measured data (ELs)
from the core processes and those obtained from up- and downstream processes,
e.g., by questionnaires, can be used to feed the LCI spreadsheet model or software
tool that contains a more or less elaborated database with information for the
background processes. If a future scenario for process chain options will be assessed,
data can be generated by a model of core processes and linear adaptations of current
data obtained from up- and downstream processes. The model can be a modular
model, as described in Chapter 1, or a process simulator.
If the consideration of accidents was chosen in the goal and scope definition,
potential environmental loads through the accidents must be generated by simula-
tions with the corresponding analysis of the undesired events or accidents (AICHE,
1985; Aelion et al., 1995), as mentioned in Chapter 1. The proper LCI can be created
by a spreadsheet model, e.g., Castells et al. (1995), or by a commercial software
tool, e.g., TEAM, as presented in Chapter 2 and applied in Chapter 3. The incorpo-
rated database is important in the LCI model or software, especially for background
processes like electricity production (e.g., Frischknecht et al. 1996). Another optional
element is the uncertainty analysis, which can be carried out, for instance, by Monte
Carlo (MC) simulation, as described in Chapter 5. By using probability distributions
for the essential factors in an MC simulation (LaGrega, 1994), the inventory result
can be transformed from a concrete value into a probability distribution around a
mean value.

© 2004 CRC Press LLC