Transcript Uncertainty Anal..

Approaches to Data Acquisition
• The LCA depends upon data acquisition
• Qualitative vs. Quantitative
– While some quantitative analysis is appropriate,
inappropriate quantification can obscure issues
• LCA Information Systems
– Non-prescriptive – global solution may depend on
locally “non-green” choices
– Must describe both data and uncertainty
Treatment of uncertainty
• The nature of uncertain parameters
– Random
– Unknowable
• We can use descriptive statistics to
characterize uncertain parameters
• Or we can simulate uncertain systems
– Monte Carlo sampling methods
Monte Carlo Simulation
• Uncertain independent variables are
described by probability density functions
• A cumulative
distribution function
is calculated
Monte Carlo Simulation
• The y axis is sampled randomly, and
corresponding x axis values are
selected
1.2
1
0.8
0.6
Series1
0.4
0.2
0
1
2
3
4
5
6
7
8
9 10 11 12 13 14 15 16 17 18 19 20
Monte Carlo Simulation
• Multiple independent variables can be
sampled for each model run
• N such calculations produce N model
results - itself a distribution, from
which descriptive statistics and
probabilities can be determined
Using MC in Data Acquisition
• Since many of the data values in our industrial
systems are uncertain, we can use Monte Carlo
sampling to gain estimates of these values
– e.g. we can estimate the average amount of lead
used in an iPad by combining the individual
components’ probability distributions
– Then running the calculation many times results in
a composite distribution of lead in the iPad