Environmental Engineering Reference
In-Depth Information
10
Large-Scale Fabrication, Learning Curves, and Economics
Including Storage
10.1
Fabrication Methods Vary but Exhibit Similar Learning Curves
We are interested in the techniques needed to greatly increase the production and use
of renewable energy. In terms of fabrication, it is well known, and makes common
sense, that the cost of making a given device falls as methods of fabrication are
improved with increased output. For photovoltaic modules, the cost per watt Wp over
30 years is summarized in Figure 10.1, plotted versus accumulated production
(volume) in megawatts. The value Wp is the power the panel will produce under full
illumination, 1000W/m
2
. The average cost is shown, including various cell types, but
this curve primarily re
ects silicon solar cells, which still represent 90% of
production.
This nearly linear log
-
log plot of cost
C
per peak watt, Wp, versus manufactured
volume
V
, is consistent with
V
0
Þ
L
C
=
C
0
¼ð
V
=
;
ð
10
:
1
Þ
so that
ln
ðC=C
0
Þ¼L
ln
ðV=V
0
Þ:
ð
10
:
2
Þ
The learning parameter
L
in this case is about 0.33. Note that one decade of cost,
from$10 down to $1, occurs in about three decades of volume, from100MWto about
10
5
MW
100 GW (which is extrapolated). As we saw at the end of Chapter 8, the
volume in 2010 is reported as 15.9GW (http://pvinsights.com/Report/
ReportPMM04A.php), so the extrapolation seems reasonable.
On the Internet today, one can
find many vendors (http://Wpww.wholesalesolar.
com/solar-panels.html) of solar cell modules, with offered prices generally in the
range $3/Wp to $5/Wp.
The price of silicon solar cells has recently fallen, reflecting increased production
capacity of silicon and of silicon cells. Aconsequence is that smallermanufacturers of
innovative products have faced losses. The September 2011 bankruptcy of the CIGS
producer Solyndra [124] has been described as a cash
flow problem arising from
¼
