Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Utility-Scale Solar 2016
An Empirical Analysis of Project Cost, Performance, and Pricing Trends in the
United States
Mark Bolinger, Joachim Seel, Kristina Hamachi LaCommare
Lawrence Berkeley National Laboratory
September 2017
This research was supported by funding from the
U.S. Department of Energy’s SunShot Initiative.
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Presentation Outline
Strong growth of the utility-scale solar market provides increasing amounts of empirical project-level data that are ripe for analysis
1. Solar deployment trends (and utility-scale’s relative contribution)
7. Future outlook
2
Key findings from analysis of the data samples (first for PV, then for CSP):
2. Project design, technology, and location
3. Installed project prices
4. Operation and maintenance (O&M) costs
5. Performance (capacity factors)
6. Power purchase agreement (“PPA”) prices
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Utility-scale projects have the greatest capacity share in the U. S. solar market
The utility-scale sector accounted for 72% of all new solar capacity added in 2016 and 61% of cumulative solar capacity at the end of 2016
3
Sources: GTM/SEIA Solar Market Insight Reports, Berkeley Lab
We define “utility-scale” as any ground-mounted project that is larger than 5 MWAC Smaller systems are analyzed in LBNL’s “Tracking the Sun” series (trackingthesun.lbl.gov)
9 22 70 267784
1,8032,855
3,9224,149
10,636
8,6726,337
7,0848,735
10,44510,612
6975
250
877110
0
0
25
50
75
100
125
0
4,000
8,000
12,000
16,000
20,00020
07
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
E
2018
E
2019
E
2020
E
2021
E
2022
E
Cum
ulat
ive
Sola
r Cap
acity
(GW
)
Annu
al S
olar
Cap
acity
Add
ition
s (M
W) Utility-Scale CSP
Utility-Scale PV Commercial PV Residential PV
Columns show annual capacity additions,area shows cumulative capacity
PV is shown in WDC while CSP is in WAC
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Solar power was the largest source of U.S. electric-generating capacity additions in 2016
Led by the utility-scale sector, solar power has comprised >25% of all generating capacity additions in the United States in each of the past four years
In 2016, solar made up 38% of all U.S. capacity additions (with utility-scale accounting for 26%), and was the largest source of new capacity, ahead of both natural gas and wind
4
0%
5%
10%
15%
20%
25%
30%
35%
40%
0
5
10
15
20
25
30
35
40
2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
Sola
r Cap
acity
Add
ition
s (%
of T
otal
)
Tota
l Ann
ual C
apac
ity A
dditi
ons (
GW
AC)
Utility-Scale SolarDistributed SolarWindOther REGasCoalOther non-RE
Total Solar (right axis)
Distributed Solar (right axis)
Utility-Scale Solar(right axis)
Sources: ABB, AWEA, GTM/SEIA Solar Market Insight Reports, Berkeley Lab
Note: This graph follows GTM/SEIA’s split between distributed and utility-scale solar, rather than our 5 MWAC threshold
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Solar penetration rates approaching or exceeding 10% in several states
5
• Solar penetration rate varies considerably depending on whether calculated as a percentage of generation or load (e.g., see Vermont)
• Contribution of utility-scale also varies (a minority in northeast states and Hawaii, a majority in other states and overall)
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Utility-Scale PV
6
Photo Credit: Community Solar Amazon Solar Farm US East 1
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Historically heavy concentration in the Southwest and mid-Atlantic, but now spreading to Southeast and Northwest
7
Primarily fixed-tilt c-Si projects in the East
Tracking (c-Si and, increasingly, thin-film) is more common in the Southwest
State Cumulative Capacity MW-AC %
2016 2015 CA 54% 56% AZ 9% 12% NV 8% 7% GA 6% 3% NC 5% 6%
2015
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Historically heavy concentration in the Southwest and mid-Atlantic, but now spreading to Southeast and Northwest
8
Primarily fixed-tilt c-Si projects in the East
Tracking (c-Si and, increasingly, thin-film) is more common in the Southwest
State Cumulative Capacity MW-AC %
2016 2015 CA 54% 56% AZ 9% 12% NV 8% 7% GA 6% 3% NC 5% 6%
2016
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Utility-scale PV continues to expand beyond California and the Southwest
Strong percentage growth outside the established markets: 7 new states added their first utility-scale solar project: OR, ID, MN, VA, AL, KY, SC Georgia added 726 MWAC – the second-largest amount of new solar capacity among all states in 2016 Texas doubled its annual new capacity with 263 MWAC Florida started growth spree with 229 MWAC – with substantially more planned for coming years
9
69%
76%47%
40%
55%
20%
16%
24%
29%
22%
21%
8%
4% 7%
9%
0
2
4
6
8
10
12
14
16
0
1
2
3
4
5
6
7
8
<=2010 2011 2012 2013 2014 2015 2016
Annu
al P
V Ca
paci
ty A
dditi
ons (
GW
AC)
Installation Year
All Other States Southeast Southwest California
Cum
ulat
ive
PV C
apac
ity (G
WAC
)
Columns show annual capacity additions (left scale)
Areas show cumulative capacity (right scale)
PV project population: 427 projects totaling 16,439 MWAC
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
The eastward expansion is reflected in the buildout of lower-insolation sites
2016 was the 3rd year of declining median solar resource (measured in long-term global horizontal irradiance (GHI)) as the market expands to less-sunny states
Fixed-tilt PV is increasingly relegated to lower-insolation sites (note the decline in its 80th percentile), while tracking PV is increasingly pushing into those same areas (note the decline in its 20th percentile)
All else equal, the buildout of lower-GHI sites will dampen sample-wide capacity factors (reported later)
10
3.5
4.0
4.5
5.0
5.5
6.0
2010n=10
175 MW
2011n=34
478 MW
2012n=43
946 MW
2013n=38
1,344 MW
2014n=64
3,166 MW
2015n=87
2,870 MW
2016n=146
7,385 MW
Annu
al G
HI (k
Wh/
m2 /
day)
Installation Year
All PV Fixed-Tilt PV Tracking PV
Median values shown, with error bars indicating 20th and 80th percentiles
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
PV project population broken out by tracking vs. fixed-tilt, module type, and installation year
2016 Trends: Increasing dominance of tracking projects (79% of newly installed capacity) relative to fixed-tilt projects (21%)
Continued strong growth in c-Si capacity (77%) relative to thin-film capacity (23%). Largest c-Si manufacturers are Trina (22%), and Jinko (14%), Canadian Solar (14%) and SunPower (8%), while the thin-film market is dominated by First Solar (97% of the installed capacity).
11
PV project population: 427 projects totaling 16,439 MWAC
0
2
4
6
8
10
12
14
16
2007-2009 2010 2011 2012 2013 2014 2015 20160
1
2
3
4
5
6
7
8
Cum
ulat
ive
Capa
city
(GW
AC)
Installation Year
Annu
al C
apac
ity A
dditi
ons (
GW
AC)
Tracking Thin-Film Tracking c-Si Fixed-Tilt Thin-Film Fixed-Tilt c-Si
Columns show annual capacity additions (left scale)
Areas show cumulative capacity (right scale)
0.62
0.92
4.74
1.11
3.45
2.02
2.42
8.48
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
The median inverter loading ratio (ILR) has risen over time, though not much since 2013
12
As module prices have fallen (faster than inverter prices), developers have oversized the DC array capacity relative to the AC inverter capacity (i.e., the ILR) to enhance revenue
The ILR (DC:AC ratio) seems to have stabilized around 1.3 on average, though considerable variation remains
Fixed-tilt PV has more to gain from a higher ILR than does tracking PV; the highest ILR projects tend to be fixed-tilt
All else equal, a higher ILR should boost capacity factors (reported later)
1.00
1.05
1.10
1.15
1.20
1.25
1.30
1.35
1.40
1.45
2010n=10
175 MW
2011n=34
478 MW
2012n=43
946 MW
2013n=38
1,344 MW
2014n=64
3,166 MW
2015n=87
2,870 MW
2016n=144
7,325 MW
Inve
rter
Load
ing
Ratio
(ILR
)
Installation Year
All PV
Fixed-Tilt PV
Tracking PV
Median values shown, with error bars indicating 20th and 80th percentiles
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Median installed price of PV has fallen steadily, by over 65%, to around $2.2/WAC ($1.7/WDC) in 2016
13
Installed prices are shown here in both DC and AC terms, but because AC is more relevant to the utility sector, all metrics used in the rest of this slide deck are expressed solely in AC terms
The lowest 20th percentile fell from $2.2/WAC ($1.6/WDC) in 2015 to $2.0/WAC ($1.5/WDC) in 2016 Minimum price among our 88 projects in 2016 was $1.5/WAC ($1.1/WDC) This sample is backward-looking and may not reflect the price of projects built in 2017/2018
0
1
2
3
4
5
6
7
8
9
10
2007-2009n=5
75 MW
2010n=10
175 MW
2011n=29
428 MW
2012n=40
915 MW
2013n=38
1,344 MW
2014n=64
3,166 MW
2015n=87
2,870 MW
2016n=88
5,497 MW
Inst
alle
d Pr
ice
(201
6 $/
W)
Installation Year
Median (DC) Individual Projects (DC) Median (AC) Individual Projects (AC)
1.0
1.5
2.0
2.5
3.0
3.5
4.0
2016
Inst
alle
d Pr
ice
(201
6 $/
W)
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Pricing distributions have continuously moved towards lower prices over the last 5 years
14
Both medians and modes have continued to fall (moving towards the left) each year Share of relatively high-cost systems decreases steadily each year while share of low-cost systems
increases Price spread is the smallest in 2016, pointing to a reduction in underlying heterogeneity of prices across
all installed projects
0%
10%
20%
30%
40%
50%
60%
≥ $1.25< $1.75
≥ $1.75< $2.25
≥ $2.25< $2.75
≥ $2.75< $3.25
≥ $3.25< $3.75
≥ $3.75< $4.25
≥ $4.25< $4.75
≥ $4.75< $5.25
≥ $5.25< $5.75
≥ $5.75< $6.25
Installed Price Interval (2016 $/WAC)
2016n=885,497 MW
2015n=872,870 MW
2014n=643,166 MW
2013n=381,344 MW
2012n=40915 MW
Proj
ect S
hare
of A
nnua
l Pric
e Sa
mpl
e
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Tracking projects were $0.15/WAC more costly (at the median) than fixed-tilt projects in 2016
Tracking’s empirical cost premium has varied somewhat over time, but in general has declined considerably since 2010
Upfront cost premium usually compensated by higher annual generation
15
0
1
2
3
4
5
6
7
8
9
2007-2009n=5
75 MW
2010n=10
175 MW
2011n=29
428 MW
2012n=40
915 MW
2013n=38
1,344 MW
2014n=64
3,166 MW
2015n=87
2,870 MW
2016n=88
5,483 MW
Inst
alle
d Pr
ice
(201
6 $/
WAC
)
Installation Year
All PV
Fixed-Tilt PV
Tracking PV
Median values shown, with error bars indicating 20th and 80th percentiles
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
2016 project sample hints at possible economies of scale (at least up to 100 MW)
Modest economies of scale evident in the sample, from $2.3/WAC for projects smaller than 20MWAC to $2.1/WAC for projects between 50 and 100MWAC
But higher costs for the 100+ MW projects, several of which have been under construction for several years, possibly reflecting a higher-cost past. In addition, larger projects may face greater development, regulatory, and interconnection costs that could outweigh any economies of scale.
16
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
5-20 MWn=35
505 MW
20-50 MWn=13
455 MW
50-100 MWn=25
1,952 MW
>100 MWn=15
2,585 MW
Inst
alle
d Pr
ice
(201
6 $/
WAC
)
Project Size Range (MWAC)
All PV
Fixed-Tilt PV
Tracking PV
Median values shown, with error bars indicating 20th and 80th percentiles.
Figure only includes 2016-vintage projects.
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Project prices vary by region
Price differences driven in part by technology ubiquity (e.g., higher-priced tracking projects are more prevalent in the Southwest and California)
Other factors may include labor costs and share of union labor, land costs, soil conditions or snow load, and balance of supply and demand
17
Southwest: NV, UT, CO, AZ, NM
Southeast: AR, AL, FL, GA, KY, MD, NC, SC, VA
Northeast: NJ, NY
Midwest: IN, MN
Northwest: ID, OR
Not included: HI, TX
3.0
2.4 2.4
3.1
3.7
#N/A
2.42.1 2.1
#N/A
1.9
2.5
0
1
2
3
4
Californian=67
3,544 MW-AC
Southwestn=33
2,303 MW-AC
Southeastn=52
1,816 MW-AC
Northeastn=6
57 MW-AC
Midwestn=8
143 MW-AC
Northwestn=3
98 MW-AC
Inst
alle
d Pr
ice
(201
6 $/
WAC
)
Select Regions of the United States
2015 2016 U.S. national median 2016
Bars show median values, with 20th and 80th percentiles.
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Bottom-up models roughly consistent with LBNL’s top-down findings
LBNL’s top-down empirical prices are fairly close to modelled bottom-up prices GTM project represents only turn-key EPC costs and excludes permitting, interconnection, transmission,
developer overhead, fees, and profit margins Difficult to ensure consistency of scope in cost categories and time horizon (under construction vs. operation
date)
18
Prices are presented here in $/WDC for consistency with how they are presented by NREL, BNEF, and GTM 0.64 0.64 0.64
0.48 0.53 0.64 0.64 0.640.48 0.53
0.10 0.10 0.100.07 0.11
0.09 0.09 0.090.07 0.11
0.20 0.27 0.27
0.20 0.150.26 0.32 0.32
0.250.25
0.320.38
0.55
0.29 0.36
0.340.41
0.60
0.320.40
0.160.27
0.29
0.18
0.16
0.28
0.31
0.18
1.42
1.66
1.85
1.231.14
1.49
1.75
1.96
1.30 1.28
LBNL Fixed-Tilt: 1.55LBNL Tracking: 1.73
$0.0
$0.5
$1.0
$1.5
$2.0
NREL 2016100 MW-DC
NationalAverage
Non-UnionLabor
NREL 201625 MW-DC
NationalAverage
Non-UnionLabor
NREL 201625 MW-DC
NationalAverage
Union Labor
BNEF 2016NationalAverage
c-Si
GTM 201610 MW-DC
NationalAverageEPC Only
NREL 2016100 MW-DC
NationalAverage
Non-UnionLabor
NREL 201625 MW-DC
NationalAverage
Non-UnionLabor
NREL 201625 MW-DC
NationalAverage
Union Labor
BNEF 2016NationalAverage
c-Si
GTM 201610 MW-DC
NationalAverageEPC Only
Fixed-Tilt Tracking
Proj
ect C
ost o
r Pric
e (2
016
$/W
DC)
Other (Developer Overhead + Margin, Contingencies, Sales Tax) Design, EPC, Labor, Permitting, Interconnection, Transmission, Land Tracker / Racking, BOS Inverter Module
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
O&M cost data still very thin
Only a few utilities report solar O&M costs, slow emergence of project-specific O&M costs
O&M costs appear to be declining over time, to $17.8/kW-year and $8.2/MWh in 2016 (slight increase from 2015)
Cost declines may reflect economies of scale
Cost range among utilities continues to be large
19
Year PG&E PNM Nevada Power Georgia Power APS PSEG FP&L
MWAC project # MWAC project # MWAC project # MWAC project # MWAC project # MWAC project # MWAC project #
2011 #N/A #N/A #N/A #N/A #N/A #N/A #N/A #N/A 51 3 #N/A #N/A 110 3 2012 50 3 8 2 #N/A #N/A #N/A #N/A 96 4 #N/A #N/A 110 3 2013 100 6 30 4 #N/A #N/A #N/A #N/A 136 6 #N/A #N/A 110 3 2014 #N/A #N/A 55 7 #N/A #N/A #N/A #N/A 168 7 #N/A #N/A 110 3 2015 150 9 95 11 #N/A #N/A #N/A #N/A 191 9 #N/A #N/A 110 3
2016 150 9 95 11 16 1 36 2 237 10 44 3 110 3
predominant technology Fixed-Tilt c-Si 4 Fixed-Tilt,
7 Tracking Tracking c-Si Fixed-Tilt c-Si Tracking c-Si Fixed-Tilt c-Si mix of c-Si and CSP
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
25.8% average sample-wide PV net capacity factor, but with large project-level range (from 15.4%-35.5%)
Project-level variation in PV capacity factor driven by: Solar Resource (GHI): Highest resource quartile has ~8 percentage point higher capacity factor than lowest Tracking: Adds ~4 percentage points to capacity factor on average across all four resource quartiles Inverter Loading Ratio (ILR): Highest ILR quartiles have ~4 percentage point higher capacity factor than lowest
20
0%
5%
10%
15%
20%
25%
30%
35%
40%
1ILR
2ILR
3ILR
4ILR
1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4
Fixed-Tilt Tracking Fixed-Tilt Tracking Fixed-Tilt Tracking Fixed-Tilt Tracking
1st Quartile Solar Resource 2nd Quartile Solar Resource 3rd Quartile Solar Resource 4th Quartile Solar Resource
Cum
ulat
ive
Net
AC
Capa
city
Fac
tor Simple Mean
Individual Project
14 p
roje
cts,
165
MW
8 pr
ojec
ts, 1
20 M
W
15 p
roje
cts,
202
MW
3 pr
ojec
ts, 3
7 M
W
4 pr
ojec
ts, 5
4 M
W
7 pr
ojec
ts, 1
14 M
W
10 p
roje
cts,
170
MW
4 pr
ojec
ts, 8
9 M
W
11 p
roje
cts,
237
MW
8 pr
ojec
ts, 3
11 M
W
13 p
roje
cts,
332
MW
1 pr
ojec
t, 23
MW
4 pr
ojec
ts, 6
05 M
W
8 pr
ojec
ts, 1
06 M
W
16 p
roje
cts,
945
MW
12 p
roje
cts,
580
MW
6 pr
ojec
ts, 6
1 M
W
13 p
roje
cts,
146
MW
3 pr
ojec
ts, 5
3 M
W
9 pr
ojec
ts, 3
25 M
W
2 pr
ojec
ts, 2
74 M
W
10 p
roje
cts,
323
MW
15 p
roje
cts,
464
MW
14 p
roje
cts,
183
MW
6 pr
ojec
ts, 7
36 M
W
10 p
roje
cts,
1,1
73 M
W
4 pr
ojec
ts, 7
9 M
W
14 p
roje
cts,
350
MW
4 pr
ojec
ts, 1
58 M
W
4 pr
ojec
ts, 1
32 M
W
Sample includes 260 projects totaling 8,733 MWAC that came online from 2007-2015
ILR Quartile ILR QuartileILR QuartileILR QuartileILR QuartileILR QuartileILR QuartileILR Quartile
7 pr
ojec
ts, 1
76 M
W
1 pr
ojec
t, 10
MW
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
For those who prefer to think geographically rather than in terms of insolation quartiles…
Not surprisingly, capacity factors are highest in California and the Southwest, and lowest in the Northeast and Midwest
Although sample size is small in some regions, the greater benefit of tracking in the high-insolation regions is evident, as are the greater number of tracking projects in those regions
21
Regions are defined in the map on slide 8
18.0% 19.0%20.7% 21.0% 21.6%
25.9% 25.2%
18.8%20.5%
23.7% 24.0%
29.3% 30.2%
0%
5%
10%
15%
20%
25%
30%
35%
Northeast Midwest Southeast Hawaii Texas Southwest California
Aver
age
Cum
ulat
ive
Net
AC
Capa
city
Fac
tor Fixed-Tilt Tracking
21 p
roje
cts,
219
MW
1 pr
ojec
t, 6
MW
10 p
roje
cts,
96
MW
3 pr
ojec
ts, 2
6 M
W
27 p
roje
cts,
474
MW
15 p
roje
cts,
375
MW
3 pr
ojec
ts, 3
0 M
W
15 p
roje
cts,
1,0
69 M
W
2 pr
ojec
ts, 4
4 M
W
10 p
roje
cts,
262
MW
33 p
roje
cts,
2,2
36 M
W
70 p
roje
cts,
2,7
60 M
W
50 p
roje
cts,
1,1
36 M
W
No
trac
king
proj
ects
in H
I yet
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
More recent PV project vintages have higher capacity factors on average
Average capacity factors driven higher from 2010- to 2013-vintage projects by an increase in ILR (from 1.17 to 1.28), tracking (from 14% to 54%) and average site-level GHI (from 4.97 to 5.29).
But since 2013, average long-term site-level GHI has decreased while tracking has increased (with ILR roughly unchanged), leading to stagnation in capacity factors among 2014 and 2015 projects.
22
21.5%23.6% 24.6%
26.7% 26.4% 26.5%
22.0%24.1% 24.8%
26.9% 26.2% 26.5%
0%
5%
10%
15%
20%
25%
30%
2010 Vintage 2011 Vintage 2012 Vintage 2013 Vintage 2014 Vintage 2015 Vintage
7 Projects 30 Projects 37 Projects 48 Projects 53 Projects 78 Projects
144 MW-AC 440 MW-AC 892 MW-AC 1,720 MW-AC 2,785 MW-AC 2,660 MW-AC
2016 Cumulative
Mea
n N
et A
C Ca
paci
ty F
acto
r
ILR = 1.17
14%Tracking
GHI = 4.97
ILR = 1.23
49%Tracking
GHI = 5.13
ILR = 1.18
50%Tracking
GHI = 5.17
ILR = 1.28
54%Tracking
GHI = 5.29
ILR = 1.29
60%Tracking
GHI = 5.19
ILR = 1.30
67%Tracking
GHI = 5.11
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Performance degradation is evident, but difficult to assess and attribute at the project-level
Fleetwide degradation appears to exceed the 0.5%/year benchmark commonly assumed in PPAs and pro forma models
Contributing factors (other than actual degradation) could include inter-year resource variability (e.g., several bad solar years in a row), curtailment (which has become an issue in California – the largest market), and an inconsistent sample (which drops off quickly) in each successive year
23
Graph shows indexed capacity factors in each full calendar year following COD. No attempt has been made to correct for inter-year resource variation or other factors.
80%
85%
90%
95%
100%
105%
1 2 3 4 5 6 7 8 9
260 182 129 81 43 13 6 3 1
8,733 6,073 3,288 1,567 667 227 83 29 7
Median (with 20th/80th percentile error bars) Capacity-Weighted Average Simple Average Representative 0.5%/year degradation rate
Years post-COD:
Sample projects:
Sample MWAC:
Inde
xed
Capa
city
Fac
tor (
Year
1=1
00%
)
Sample includes projects with COD from 2007-2015
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Combination of falling installed prices and better project performance enables lower PPA prices
PPA prices are levelized over the full term of each contract, after accounting for any escalation rates and/or time-of-delivery factors, and are shown in real 2016 dollars
Top graph shows the full sample; bottom graph shows a sub-sample of PPAs signed post-2014
CA and the Southwest dominate the sample, but in recent years the market has expanded to other regions
Hawaii projects (included here for the first time) show a consistent and significant premium over the mainland
Three PPAs featuring PV plus long-duration battery storage do not seem to be priced at a prohibitive premium to their PV-only counterparts
Smaller projects (e.g., 20-50 MW) are seemingly no less competitive
>90% of the sample is currently operational
24
$0
$50
$100
$150
$200
$250
Jan-
06
Jan-
07
Jan-
08
Jan-
09
Jan-
10
Jan-
11
Jan-
12
Jan-
13
Jan-
14
Jan-
15
Jan-
16
Jan-
17
PPA Execution Date
California
Southwest
Texas
Southeast
Midwest
Hawaii
Leve
lized
PPA
Pric
e (R
eal 2
016
$/M
Wh)
550 MW
210MW
50 MW
Sample includes 189 contracts totaling 11.7 GWAC
12 MW
$0
$20
$40
$60
$80
$100
$120
$140
Jan-
15Fe
b-15
Mar
-15
Apr-
15M
ay-1
5Ju
n-15
Jul-1
5Au
g-15
Sep-
15O
ct-1
5N
ov-1
5D
ec-1
5Ja
n-16
Feb-
16M
ar-1
6Ap
r-16
May
-16
Jun-
16Ju
l-16
Aug-
16Se
p-16
Oct
-16
Nov
-16
Dec
-16
Jan-
17Fe
b-17
Mar
-17
Apr-
17M
ay-1
7Ju
n-17
Jul-1
7Au
g-17
Sep-
17
PPA Execution Date
California Southwest Texas Southeast Midwest Hawaii
Leve
lized
PPA
Pric
e (R
eal 2
016
$/M
Wh)
100MW
Sample includes 52 PPAs totaling 3,276 MWAC that were signed since the start of 2015
12 MW
includes battery storage
50MW
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
On average, levelized PPA prices fell by >75% from 2009 through 2016
Top figure presents the same data as previous slide, but in a different way: each circle is an individual contract, and the blue columns show the average levelized PPA price each year
Steady downward trend in the average PPA price over time has slowed in recent years as average prices approached and then fell below $50/MWh
Price decline over time is more erratic when viewed by COD (orange bars in bottom graph) rather than by PPA execution date (blue bars)
Though the average levelized price of PPAs signed in 2016 is ~$35/MWh, the average levelized PPA price among projects that came online in 2016 is significantly higher, at ~$60/MWh
2017 is provisional and currently reflects a very small sample and a high proportion of high-priced Hawaiian PPAs, plus several PPAs with long-duration battery storage
25
0
50
100
150
200
250
200617
200715
20083
770
200916
1,030
201030
1,746
201120
1,790
201217
1,073
201319
478
201430
1,503
201539
2,383
20166
556
20177
337
Generation-Weighted Average
Individual PPA (Hawaii PPAs shaded orange)
PPA Year:Contracts:
MW:
Leve
lized
PPA
Pric
e (R
eal 2
016
$/M
Wh)
$0
$50
$100
$150
$200
$250
2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017
Generation-weighted average based on the year in whichcommercial operation was fully achieved
Generation-weighted average based on the year in whichthe PPA was executed
Leve
lized
PPA
Pric
e (R
eal 2
016
$/M
Wh)
2017 isprovisional
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
The value of solar has declined in America’s largest solar market
With increasing solar penetration in California, solar curtailment has increased and solar’s wholesale energy value has declined
In 2012, when solar penetration was ~2%, solar earned 126% of the average wholesale power price
In 2016, with solar penetration at ~12%, solar earned just 83% of the average wholesale power price
Based on data for the first half of the year, this value decline is likely to continue in 2017 (1Q17 was particularly bad – bottom graph)
Most other markets are not yet facing this value decline, due to lower levels of solar penetration
26
0%
15%
30%
45%
60%
75%
90%
105%
120%
135%
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
2012 2013 2014 2015 2016 1H2017
Solar Penetration Rate (left axis) Solar Curtailment Rate (left axis) Solar Value Factor (right axis)
Ener
gy V
alue
of S
olar
Rel
ativ
e to
a 2
4x7
Flat
Blo
ck
CAIS
OSo
lar P
enet
ratio
n an
d Cu
rtai
lmen
t Rat
es
CAISO curtailment data not available prior to 2015
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
20%
2015 2016 2017 2015 2016 2017 2015 2016 2017 2015 2016 2017
1Q 2Q 3Q 4Q
Solar Penetration Rate (left scale) Solar Curtailment Rate (left scale) Solar Value Factor (right scale)
Sola
r Cur
tailm
enta
nd P
enet
ratio
n Ra
tes
Ener
gy V
alue
of S
olar
Rel
ativ
e to
24x
7 Fl
at B
lock
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Levelized PPA prices track the LCOE of utility-scale PV reasonably well
Using empirical data from elsewhere in the report, along with a number of assumptions (e.g., about financing), we calculated project-level LCOEs for the entire sample of projects for which we have CapEx data (14.3 GWAC)
Central estimates of LCOE track median PPA prices (levelized over 30 years in this case, and shown by COD rather than by execution date) reasonably well, suggesting a fairly competitive PPA market
PPAs are lower than LCOEs because they reflect receipt of the 30% ITC and perhaps also state-level incentives
27
NOTE: LCOE calculations do NOT include the 30% ITC (whereas PPA prices do reflect the ITC, and perhaps also state-level incentives)
0
50
100
150
200
250
300
350
201010
175
201128
423
201238
863
201338
1,344
201463
3,160
201586
2,883
201683
5,407
2017TBDTBD
2016
$/M
Wh
Capacity-Weighted Average LCOE Median LCOE Simple Average LCOE Individual Project LCOE Median Levelized PPA Price (by COD)
COD:Projects:MW-AC:
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
PV PPA prices generally decline over time in real dollar terms, in contrast to fuel cost projections
Two-thirds of PV sample has flat annual PPA pricing (in nominal dollars), while the rest escalate at low rates
Thus, average PPA prices tend to decline over time in real dollar terms (top graph)
Bottom graph compares recent PPA prices to range of gas price projections from AEO 2017, showing that…
…although PV is currently priced higher than the cost of burning fuel in a combined-cycle unit, over longer terms PV is perhaps likely to be more competitive, and can help protect against fuel price risk
28
$0
$50
$100
$150
$200
$250
2006
2008
2010
2012
2014
2016
2018
2020
2022
2024
2026
2028
2030
2032
2034
2036
2038
2040
2042
2044
2046
2048
2050G
en-W
eigh
ted
Aver
age
PPA
Pric
e (2
016
$/M
Wh)
2006 (7 MW, 1 PPA)
2008 (770 MW, 3 PPAs)
2010 (1,746 MW, 30 PPAs)
2014(1,503 MW, 30 PPAs)
2007(5 MW, 1 PPA)
2015 (2,383 MW, 39 PPAs)2016 (566 MW, 6 PPAs)
0
10
20
30
40
50
60
70
80
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2016
$/M
Wh
Overall range of AEO 2017 gas price projections (converted to $/MWh terms) AEO 2017 reference case gas price projection (converted to $/MWh terms) Generation-weighted average PV PPA price over time Median PV PPA price (and 20th/80th percentile bars) over time
PV PPA sample includes 29 PPAs signed 7/2015-8/2017 and totaling 2,184 MWAC(PPA sample excludes Hawaiian projects)
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Utility-Scale Concentrating Solar Thermal Power (CSP)
29
Photo Credit: Solar Reserve: Crescent Dunes
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Sample description of CSP projects
After nearly 400 MWAC built in the late-1980s (and early-1990s), no new CSP was built in the U.S. until 2007 (68 MWAC), 2010 (75 MWAC), and 2013-2015 (1,237 MWAC)
Prior to the large 2013-15 build-out, all utility-scale CSP projects in the U.S. used parabolic trough collectors
The five 2013-2015 projects include 3 parabolic troughs (one with 6 hours of storage) totaling 750 MWAC (net) and two “power tower” projects (one with 10 hours of storage) totaling 487 MWAC (net)
30
CSP project population: 16 projects totaling 1,781 MWAC
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Not much movement in the installed price of CSP
Small sample of 7 projects (5 built in 2013-15) using different technologies makes it hard to identify trends
That said, there does not appear to be much of a trend (in contrast to PV’s steady downward trend)
To be fair, newest projects are much larger, and include thermal storage and/or new technology (power tower) in some cases, making comparisons difficult
31
0
1
2
3
4
5
6
7
8
9
10
11
2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
Inst
alle
d Pr
ice
(201
6 $/
WAC
)
Installation Year
CSP Trough CSP Tower Median PV (for reference)
68 MWAC
250 MWAC with 6 hours of storage
75 MWAC 250 MWAC each
377 MWAC
110 MWAC with 10 hours of storage
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Several newer CSP projects continued to underperform relative to long-term expectations
32
The two “power tower” projects (Ivanpah and Crescent Dunes) were hit with closures in 2016 that negatively impacted capacity factors. The Crescent Dunes closure lasted into 2017.
Solana was at reduced capacity for part of 2016 due to micro-burst storm damage, and for part of 2017 due to a transformer fire.
Genesis and Mojave were both largely on target in 2016 Most newer CSP projects generally performing better than older CSP projects, but not necessarily
any better than (and in some cases worse than) local PV projects
0%
5%
10%
15%
20%
25%
30%
35%
2008 2009 2010 2011 2012 2013 2014 2015 2016
SEGS I & II
SEGS III-IX
GenesisSolana
Ivanpah
MojaveNevada Solar One (dashed)
For reference: average PV in CA, NV, AZ (red diamonds)
CrescentDunes
Net
Cap
acity
Fac
tor(
sola
r por
tion
only
)
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Though once competitive, CSP PPA prices have failed to keep pace with PV’s price decline
When PPAs for the most recent batch of CSP projects (with CODs of 2013-15) were signed back in 2009-2011, they were still mostly competitive with PV
But CSP has not been able to keep pace with PV’s price decline Partly as a result, no new PPAs for CSP projects have been signed in the U.S. since 2011
– though the technology continues to advance overseas
33
$0
$50
$100
$150
$200
$250
Jan-
07
Jan-
08
Jan-
09
Jan-
10
Jan-
11
Jan-
12
Jan-
13
Jan-
14
Jan-
15
Jan-
16
Jan-
17
PPA Execution Date
PV in CA, NV, AZ (for comparison)
CSP trough w/o storage
CSP trough w/ storage
CSP tower w/o storage
CSP tower w/ storage
Leve
lized
PPA
Pric
e (R
eal 2
016
$/M
Wh)
250 MW
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
0
5
10
15
20
25
30
35
40
California Southeast Southwest Northeast Texas Central Northwest
2013
2014
2015
2016So
lar i
n Q
ueue
s at Y
ear-
End
(GW
)
020406080
100120140
Wind Solar Gas Other Storage Nuclear Coal
Entered queues in that year
2016
GW
in Q
ueue
s at Y
ear-
End
Entered queues in prior years
Looking ahead: long-term ITC extension should support continued growth in the utility-scale solar pipeline
121.4 GW of solar was in the queues at the end of 2016—up from 56.8 GW at end of 2015, and more than six times the amount of installed capacity at the end of 2016
83.3 GW of the 121.4 GW total first entered the queues in 2016 Very strong solar growth in all regions, with the possible exception of the Northwest The Southeast moved ahead of the Southwest for the number two position behind California
34
Graphs show solar and other capacity in 35 interconnection queues across the US:
• Inset compares solar to other resources (2016 only)
• Main graph shows location of solar (2013-2016)
• Not all of these projects will ultimately be built!
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Questions?
Download the full report, a data file, and this slide deck at:
http://utilityscalesolar.lbl.gov
Download all of our other solar and wind work at:
http://emp.lbl.gov/reports/re
Follow the Electricity markets & Policy Group on Twitter:
@BerkeleyLabEMP
35
Contact:
Mark Bolinger: [email protected]
Joachim Seel: [email protected]
This research was supported by funding from the U.S.
Department of Energy’s SunShot Initiative.