Page 170 - Biofuels for a More Sustainable Future

154 Biofuels for a More Sustainable Future

3.2.3 Inventory analysis
The environmental, economic, and social data for each technology are the
same as those discussed in the previous case study (see Section 3.1) with two
major amendments: firstly, coal with carbon capture and storage (CCS) is
added due to its potential for future deployment; secondly, the impacts of
coal CCS, wind, and solar PV are altered through time using learning rates
and parameter amendment due to their relative immaturity and rapid devel-
opment. The assumptions and data are outlined later, with more information
available in Stamford and Azapagic (2014).
The use of learning curves to estimate future cost reduction for energy
technologies is well established as a useful tool for scenario analysis
(McDonald and Schrattenholzer, 2001; Rubin et al., 2007; Ferioli et al.,
2009; van den Broek et al., 2009; IEA, 2013). It is defined as the reduction
in costs that are achieved for every doubling of installed capacity, and this
principle can be in LCSA as well as pure costing exercises (Stamford and
Azapagic, 2018).
Prior work has suggested a learning rate for PV systems of approximately
18% (IEA, 2010) and 12% for wind (IEA, 2013). For coal CCS, the rate is
less certain due to the lack of commercial deployment, but rates of 3.5%–
4.9% have been estimated (Rubin et al., 2007).
As described in Ferioli et al. (2009), the relationship between cost reduc-
tion and cumulative output can be expressed as:

b

x t
ðÞ
ðÞ
Cx t ¼ Cx 0 ð monetary unitÞ
x 0
where
x t ¼Cumulative installed capacity at a point in time t (GW)
x 0 ¼Arbitrary starting point in time in cumulative capacity (GW)
C¼Installed system cost at either x 0 or x t (monetary unit)
b¼A learning parameter (dimensionless)
The learning parameter b represents the slope of the power-curve fitting
the cumulative installed capacity x t against the installed costs C over the time
period of interest. The learning rate LR can then be determined using b as
follows:
LR ¼ 1 2 b x100 %ðÞ

In this case, learning curves are used to estimate future reductions in costs
and employment levels (which are closely related to costs). Future change to