Utility-Scale Solar 2015 - Lawrence Berkeley National ... · Presentation Outline . Strong growth...
Transcript of Utility-Scale Solar 2015 - Lawrence Berkeley National ... · Presentation Outline . Strong growth...
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Utility-Scale Solar 2015
An Empirical Analysis of Project Cost, Performance, and Pricing Trends in the
United States
Mark Bolinger & Joachim Seel Lawrence Berkeley National Laboratory
August 2016
This research was supported by funding from the
U.S. Department of Energy’s SunShot Initiative.
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Part of broader solar work in the Electricity Markets and Policy Group
Solar Cost-Related Work: Annual solar “state of the market” reports
• Residential/commercial systems: http://trackingthesun.lbl.gov • Large ground-mounted systems: http://utilityscalesolar.lbl.gov
Derivative analyses (e.g., Academic Partnership Program)
Renewable Energy Valuation and Grid Integration
Rate-Design Impacts on the Economics and Deployment of DPV
Impact of DPV on Traditional Utility Business Models
Impact of Utility-Scale and Distributed PV on Real Estate Assets
Technical Assistance and Policy Evaluations (e.g., RPS analyses)
2
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Presentation Outline
Strong growth of the utility-scale solar market offers increasing amounts of project-level data that are ripe for analysis.
1. Introduction and description of broader technology trends
6. Future outlook
3
Key findings from analysis of the data samples
(We discuss PV projects first, then focus on CSP projects):
2. Installed project prices
3. Operation and maintenance (O&M) costs
4. Performance (capacity factors)
5. 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
Utility-scale solar had a 57% capacity share of 2015 installations and a 54% share of cumulative installations at the end of 2015
4
Sources: GTM / SEIA Solar Market Insight Reports, LBNL Database
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.
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Drivers of the utility-scale market: RPS
RPS has historically been a significant driver of utility-scale solar, particularly in the Southwest and Northeast
Recent RPS expansions (e.g., in CA, OR) will ensure future RPS relevance for utility-scale solar markets
Increasingly, utility-scale solar expansion into non-RPS states (Southeast) or continued deployment where RPS goals have been reached (e.g., Texas)
5
WI: 10% by 2015
NV: 25% by 2025
TX: 5,880 MW by 2015
PA: 8.5% by 2020
NJ: 22.5% by 2020CT: 23% by 2020
MA: 11.1% by 2009 +1%/yr
ME: 40% by 2017
NM: 20% by 2020 (IOUs)10% by 2020 (co-ops)
CA: 50% by 2030
MN: 26.5% by 2025Xcel: 31.5% by 2020
IA: 105 MW by 1999
MD: 20% by 2022
RI: 16% by 2019
HI: 100% by 2045
AZ: 15% by 2025
NY: 30% by 2015
CO: 30% by 2020 (IOUs)20% by 2020 (co-ops)10% by 2020 (munis)
MT: 15% by 2015
DE: 25% by 2025
DC: 20% by 2020
WA: 15% by 2020
NH: 24.8% by 2025
OR: 50% by 2040 (large IOUs)5-25% by 2025 (other utilities)
NC: 12.5% by 2021 (IOUs)10% by 2018 (co-ops and munis)
IL: 25% by 2025
VT: 75% by 2032
MO: 15% by 2021
OH: 12.5% by 2026
MI: 10% by 2015
Source: LBNL RPS Report 2016
Source: GTM 2016: The Next Wave of Utility-Solar
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Non-RPS drivers of utility-scale solar Voluntary Procurement:
3rd party-ownership with competitive PPA deals (+ Hedge Value) Utility-Owned Generation (Florida Power & Light, Georgia Power, Dominion,
Duke, PNM) PURPA (“avoided cost” rates for “Qualifying Facilities”)
e.g., North Carolina (~2700 MW), Utah (~700 MW), Idaho (~500 MW), Oregon Potential for boom and bust cycles
Retail Procurement Green tariffs / community solar Direct access (e.g., Apple 130 + 150 MW California Flats PPA) Community Choice Aggregation
Merchant Solar e.g., Barilla Solar in Texas
Clean Power Plan???
6
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
PV projects
7
Photo Credit: sPower SEPV Palmdale East
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
PV project population broken out by tracking vs. fixed-tilt, module type, and installation year
2015 Trends: Strong growth in c-Si capacity (81%) relative to thin-film capacity (19%), driven in part by the completion of the
very large Solar Star project (594 MWAC). Largest c-Si manufacturers are SunPower (33% of c-Si market), Trina (20%), and Jinko (16%), while the thin-film market is dominated by First Solar (93% of the installed capacity).
Increasing dominance of tracking projects (70% of newly installed capacity) relative to fixed-tilt projects (30%)
8
PV project population: 278 projects totaling 9,016 MWAC
1,502
2,823
3,709
915
640
201
1,631
334
0
1,000
2,000
3,000
4,000
5,000
6,000
7,000
8,000
9,000
2007-2009 2010 2011 2012 2013 2014 20150
500
1,000
1,500
2,000
2,500
3,000
3,500
4,000
4,500
Cum
ulat
ive
Capa
city
Add
ition
s (M
WAC
)
Installation Year
Annu
al C
apac
ity A
dditi
ons (
MW
AC) Tracking Thin-Film
Tracking c-Si Fixed-Tilt Thin-Film Fixed-Tilt c-Si
Cumulative Tracking Capacity is 4,684 MW (52%) (incl. hybrid projects with both thin-film and c-Si modules)
Cumulative Fixed-Tilt Capacity is 4,325 MW
Columns represent annual capacity additions
Areas represent cumulative capacity additions
Cumulative c-Si Capacity is 5,211 MW (58%)
Cumulative Thin-Film Capacity is 3,738 MW
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Historically heavy concentration in the Southwest and mid-Atlantic, but now spreading to Southeast
9
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 %
2015 2014 CA 56% 59% AZ 13% 17% NV 7% 5% NC 6% 2% TX 3% 3%
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:
North Carolina (quadrupling previous capacity with 15 new projects)
Georgia (nearly tripling previous capacity with 6 new projects totaling 177 MWAC)
Nevada (more than doubling previous capacity with 4 new projects totaling 349 MWAC)
10
25% 25% 69%
76%
47%
51% 46%
55%
20%
16%
25%
41% 29%
19% 11%
8%
28%
0
1,000
2,000
3,000
4,000
5,000
6,000
7,000
8,000
9,000
0
500
1,000
1,500
2,000
2,500
3,000
3,500
4,000
4,500
<=2010 2011 2012 2013 2014 2015
Cum
ulat
ive
PV C
apac
ity A
dditi
ons (
MW
AC)
Annu
al P
V Ca
paci
ty A
dditi
ons
(MW
AC)
Installation Year
All Other States
Southwest (NV, UT, AZ, NM, CO)
California
Columns represent annual capacity additions
Areas represent cumulative capacity additions
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
The eastward expansion is reflected in the buildout of lower-insolation sites
Historical trend of increasing solar resource quality for the average project site did not continue in 2015 – the first year where the Global Horizontal Irradiance declined
The wide 80/20 distribution of fixed-tilt PV reflects deployment throughout the US, whereas tracking PV is concentrated more in the high-GHI Southwest. However, 2015 shows an expansion of tracking into less-sunny areas (note the decline in the 20% percentile)
All else equal, higher GHI should boost sample-wide capacity factors (reported later). The effects of the lower GHI for new 2015 projects will be evaluated in next year’s report once they have been operational for a full year
11
5.01
5.39
5.28
3
4
5
6
2010n=10
(175 MW-AC)
2011n=35
(485 MW-AC)
2012n=43
(946 MW-AC)
2013n=38
(1,344 MW-AC)
2014n=64
(3,166 MW-AC)
2015n=83
(2,824 MW-AC)
Annu
al G
HI (
kWh/
m2 /
day)
Installation Year
Fixed-Tilt PV Tracking PVAll PVMarkers represent capacity-weighted averages, with 20th and 80th percentiles.
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
The average inverter loading ratio (ILR) has increased over time, to 1.31 in 2015
12
1.33
1.31
1.31
1.0
1.1
1.2
1.3
1.4
2010n=10
(175 MW-AC)
2011n=35
(485 MW-AC)
2012n=43
(946 MW-AC)
2013n=38
(1,344 MW-AC)
2014n=64
(3,166 MW-AC)
2015n=83
(2,824 MW-AC)
Inve
rter
Load
ing
Ratio
(ILR
)
Installation Year
Fixed-Tilt PV
Tracking PV
All PVMarkers represent capacity-weighted averages, with 20th and 80th percentiles.
As module prices have fallen (faster than inverter prices), developers have oversized the DC array capacity relative to the AC inverter capacity to enhance revenue
Fixed-tilt PV generally has a higher average ILR than tracking PV (fixed-tilt has more to gain from boosting ILR), dip in 2014 is skewed by several very large projects
The apparent decline in the capacity-weighted average ILR from 2013 to 2014 is related to several large projects – the median ILR held nearly constant in 2014
All else equal, a higher ILR should boost sample-wide capacity factors (reported later)
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Median installed price of PV has fallen steadily, by nearly 60%, to around $2.7/WAC ($2.1/WDC) in 2015
13
0
1
2
3
4
5
6
7
8
9
10
2007-2009n=5
(75 MW-AC)
2010n=10
(175 MW-AC)
2011n=29
(428 MW-AC)
2012n=39
(895 MW-AC)
2013n=34
(1,287 MW-AC)
2014n=59
(3,050 MW-AC)
2015n=64
(2,135 MW-AC)
Inst
alle
d Pr
ice
(201
5 $/
W)
Installation Year
Capacity-Weighted Average (DC) Median (DC) Individual Projects (DC) Capacity-Weighted Average (AC) Median (AC) Individual Projects (AC)
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 $2.3/WAC ($1.8/WDC) in 2014 to $2.2/WAC ($1.6/WDC) in 2015
Capacity-weighted average prices were pushed higher in 2014 and 2015 by several very large projects that had been under construction for several years (but only entered our sample once complete)
This sample is backward-looking and may not reflect the price of projects built in 2016/2017
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Pricing distributions have continuously moved towards lower prices over the last 4 years
14
Not only pricing medians but also pricing 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
Interquartile price spread is the smallest in 2015, pointing to a reduction in underlying heterogeneity of prices across all installed projects
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
>= $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 ($2015/WAC)
2015n=64(2,135 MW-AC)
2014n=59(3,050 MW-AC)
2013n=34(1,287 MW-AC)
2012n=39(895 MW-AC)
Proj
ect S
hare
of A
nnua
l Pric
e Sa
mpl
e
Sample Statistics for each Installation Year:
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Installed price decline led primarily by c-Si
Pricing has converged among the various mounting/module configurations over time
The two CPV projects built in 2011 and 2012 were priced similar to PV at the time, while the 2014 CPV project was at the very low end of price distributions (unfortunately, no price data was available for the new 2015 C7 project)
15
0
1
2
3
4
5
6
7
8
9
2007-2009n=5
(75 MW-AC)
2010n=10
(175 MW-AC)
2011n=29
(428 MW-AC)
2012n=39
(895 MW-AC)
2013n=34
(1,287 MW-AC)
2014n=59
(3,050 MW-AC)
2015n=64
(2,135 MW-AC)
Inst
alle
d Pr
ice
(201
5 $/
WAC
)
Installation Year
Fixed-Tilt c-Si Tracking c-Si Fixed-Tilt Thin-Film Tracking Thin-Film CPV All PV
Markers represent medians, with 20th and 80th percentiles.
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Tracking projects command a premium of $0.3/WAC
Not surprisingly, tracking appears to be slightly more expensive than fixed-tilt
New EIA summary statistics from project-level data for 2013 consistent with LBNL medians
16
2.5
2.8
0
1
2
3
4
5
6
7
8
9
2007-2009n=5
(75 MW-AC)
2010n=10
(175 MW-AC)
2011n=29
(428 MW-AC)
2012n=39
(895 MW-AC)
2013n=34
(1,287 MW-AC)
2014n=59
(3,050 MW-AC)
2015n=64
(2,135 MW-AC)
Inst
alle
d Pr
ice
(201
5 $/
WAC
)
Installation Year
Fixed-Tilt, all modulesTracking, all modulesEIA Fixed-Tilt Estimate 2013EIA Tracking Estimate 2013All PV
Markers represent medians, with 20th and 80th percentiles.
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
2015 project sample does not reflect economies of scale
Modular/scalable “power block” solutions from manufacturers like SunPower and First Solar may have already wrung out most of the cost savings otherwise available to larger projects.
Potential savings may not fully be passed through from EPCs to developers, or procurement savings may occur at the portfolio rather than project level
Several of the 100+ MW projects have been under construction for several years, possibly reflecting a higher-cost past. We find a correlation between increased COD-PPA lag and higher project prices (additional year of lag time results in premium of $0.5/WAC). Larger projects may face greater development, regulatory, interconnection costs that outweigh any economies of scale.
17
0
1
2
3
4
5
5-10 MWn=12
(108 MW-AC)
10-20 MWn=30
(475 MW-AC)
20-100 MWn=19
(853 MW-AC)
100-1,000 MWn=3
(699 MW-AC)
Inst
alle
d Pr
ice
(201
5 $/
WAC
)
Project Size (MWAC)
All PV Fixed-Tilt c-Si Tracking c-Si Fixed-Tilt Thin-Film Tracking Thin-Film
Markers represent medians, with 20th and 80th percentiles. Figure only includes 2015-vintage projects.
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Project prices vary by region
Price differences driven in part by technology ubiquity (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
18
Southwest NV, UT, CO, AZ,
NM
Southeast AR, TN, MD,
NC, GA
Northeast NJ, MA
Not included HI, IN, TX
3.2 3.0
2.3 2.5
2.9 2.7
2.3 2.3
0
1
2
3
4
Californian=53
(3,419 MW-AC)
Northeastn=12
(105 MW-AC)
Southeastn=28
(700 MW-AC)
Southwestn=21
(736 MW-AC)
Inst
alle
d Pr
ice
(201
5 $/
WAC
)
Select Regions of the United States
2014 median prices 2015 median prices U.S. national median 2015
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Bottom-up models roughly consistent with LBNL’s top-down findings
Prices presented here in DC terms, to be consistent with how presented by NREL, BNEF, GTM
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)
19
0.65 0.65 0.65 0.66 0.67 0.65 0.65 0.65 0.67
0.11 0.11 0.11 0.11 0.14 0.11 0.11 0.11 0.14 0.32 0.33 0.33 0.27 0.21 0.38 0.39 0.39 0.33
0.39 0.41 0.60 0.75
0.18
0.44 0.47 0.67
0.21
0.31 0.36 0.40 0.20
0.24
0.33 0.39
0.43
0.31
1.78 1.86
2.09 1.99
1.44
1.91 2.00
2.25
1.66
LBNL Fixed-Tilt 2.01
LBNL Tracking 2.10
$0.0
$0.5
$1.0
$1.5
$2.0
$2.5
NREL 2015100 MW-DC
National AverageNon-Union Labor
NREL 201525 MW-DC
National AverageNon-Union Labor
NREL 201525 MW-DC
National AverageUnion Labor
BNEF 2015Unknown Size
California
GTM 201510 MW-DCCaliforniaEPC Only
NREL 2015100 MW-DC
National AverageNon-Union Labor
NREL 201525 MW-DC
National AverageNon-Union Labor
NREL 201525 MW-DC
National AverageUnion Labor
GTM 201510 MW-DCCaliforniaEPC Only
Fixed-Tilt Tracking
Proj
ect C
ost o
r Pric
e (2
015
$/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, but largely consistent with early years of cost projections
Only a few utilities report solar O&M costs (see table), slow emergence of project-specific O&M costs
O&M costs appear to be declining over time (as fleet size increases), to $15.6/kW-yr and $7.3/MWh
Cost range among utilities continues to be large
20
30.6 29.3 28.1
18.0 15.6
19.0
14.1 13.7
8.3 7.3
0
4
8
12
16
20
24
28
0
10
20
30
40
50
60
70
2011 2012 2013 2014 2015
2015
$/M
Wh
2015
$/k
WAC
-yr
FPL CSP ($/kW-yr) Mean of all PV projects ($/kW-yr) Mean of all PV projects ($/MWh)
Whiskers represent least and most expensive utility
Year PG&E PNM APS * FP&L
MW-AC project # MW-AC project # MW-AC project # MW-AC project # 2011 #N/A #N/A #N/A #N/A 51 3 110 3 2012 50 3 8 2 96 4 110 3 2013 100 6 30 4 136 6 110 3 2014 150 7 55 7 168 7 110 3 2015 150 7 95 11 191 9 110 3
predominant technology Fixed-Tilt c-Si
4 fixed-tilt / 3 tracking thin-film,
4 tracking c-Si primarily tracking c-Si mix of c-Si and CSP
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
25.7% average sample-wide PV net capacity factor, but with large project-level range (from 15.1%-35.7%)
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
21
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
9 pr
ojec
ts, 1
06 M
W
6 pr
ojec
ts, 8
7 M
W
8 pr
ojec
ts, 8
3 M
W
4 pr
ojec
ts, 6
7 M
W
3 pr
ojec
ts, 3
6 M
W
3 pr
ojec
ts, 4
5 M
W
7 pr
ojec
ts, 1
01 M
W
2 pr
ojec
ts, 4
0 M
W
8 pr
ojec
ts, 1
16 M
W
4 pr
ojec
ts, 1
13 M
W
10 p
roje
cts,
269
MW
1 pr
ojec
t, 23
MW
6 pr
ojec
ts, 7
25 M
W
6 pr
ojec
ts, 8
7 M
W
7 pr
ojec
ts, 8
9 M
W
13 p
roje
cts,
606
MW
2 pr
ojec
ts, 1
9 M
W
12 p
roje
cts,
135
MW
5 pr
ojec
ts, 9
8 M
W
2 pr
ojec
ts, 3
8 M
W
1 pr
ojec
t, 24
2 M
W
7 pr
ojec
ts, 5
13 M
W
5 pr
ojec
ts, 8
4 M
W
10 p
roje
cts,
115
MW
4 pr
ojec
ts, 5
96 M
W
7 pr
ojec
ts, 9
84 M
W
5 pr
ojec
ts, 9
8 M
W
7 pr
ojec
ts, 2
83 M
W
1 pr
ojec
t, 18
MW
4 pr
ojec
ts, 7
2 M
W
Sample includes 170 projects totaling 5,907 MWAC that came online from 2007-2014
ILR Quartile ILR QuartileILR QuartileILR QuartileILR QuartileILR QuartileILR QuartileILR Quartile
1 pr
ojec
t, 20
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 (with the Southeast and Texas in between)
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
22
17.5% 18.3%20.2% 20.1%
25.6% 25.6%
18.7%19.9%
21.9% 22.7%
29.5%30.6%
0%
5%
10%
15%
20%
25%
30%
35%
Northeast Midwest Southeast Texas Southwest California
Aver
age
Cum
ulat
ive
Net
AC
Capa
city
Fac
tor Fixed-Tilt Tracking
16 p
roje
cts,
173
MW
1 pr
ojec
t, 6
MW
8 pr
ojec
ts, 8
6 M
W
2 pr
ojec
ts, 1
6 M
W
13 p
roje
cts,
179
MW
6 pr
ojec
ts, 1
01 M
W
1 pr
ojec
t, 14
MW
7 pr
ojec
ts, 1
51 M
W
14 p
roje
cts,
819
MW
38 p
roje
cts,
773
MW
27 p
roje
cts,
2,1
28 M
W
37 p
roje
cts,
1,4
62 M
W
Regions are defined in the map on slide 9
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
More recent PV project vintages have higher capacity factors on average
Higher capacity factors by vintage driven by an increase in tracking (most notably in 2011 and 2014), average inverter loading ratio (in every year), and long-term global horizontal irradiance at project sites (in 2011 and 2013)
The fact that single-year 2015 capacity factors (blue columns) show same trend as cumulative capacity factors (orange columns) suggests that inter-year resource variation is not much of a driver
23
21.0%23.4% 24.1%
26.8% 26.7%
22.1%24.2% 24.6%
27.1% 26.7%
0%
5%
10%
15%
20%
25%
30%
2010 Vintage 2011 Vintage 2012 Vintage 2013 Vintage 2014 Vintage
7 Projects 32 Projects 36 Projects 48 Projects 41 Projects
144 MW-AC 464 MW-AC 891 MW-AC 1,728 MW-AC 2,597 MW-AC
2015 Cumulative
Mea
nN
et A
C Ca
paci
ty F
acto
r
Mean ILR = 1.17
14% Tracking
Mean GHI = 4.97
Mean ILR = 1.23
50% Tracking
Mean GHI = 5.07
Mean ILR = 1.18
53% Tracking
Mean GHI = 5.15
Mean ILR = 1.28
54% Tracking
Mean GHI = 5.30
Mean ILR = 1.30
61% Tracking
Mean GHI = 5.28
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Performance degradation is difficult to assess at the project-level, due to modest impact
Over the 8-year period shown in the graph, a degradation rate of 0.5%/year would reduce a 30% capacity factor to 29%—a modest decline that could easily be swamped by inter-year variation in the strength of the solar resource
Though degradation is no doubt present in the graph above, the 2013-2015 decline (evident among Western projects in particular) is more likely attributable to below-normal summer insolation
24
15%16%17%18%19%20%21%22%23%24%25%26%27%28%29%30%31%
2008 2009 2010 2011 2012 2013 2014 2015
Net
AC
Capa
city
Fac
tor
Calendar Year
NV CO CO CA NV NM NV FL FL IL FL TX OH
Graph shows a time series of capacity factors by calendar year for the 13 projects in our sample that have been operating for at least 5 (and for as many as 8) years
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Earlier regression analysis offers additional insights into sources of net capacity factor gains
Graph progresses from a 2007 fixed-tilt project with average GHI and ILR (on the far left) to a 2013 fixed-tilt project with a higher GHI and ILR (in the middle) to a 2013 tracking project with the same higher GHI and ILR (on the far right).
25
More information at: https://emp.lbl.gov/publications/maximizing-mwh-statistical-analysis
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 2015 dollars
Top graph shows the full sample; bottom graph shows a sub-sample of PPAs signed in 2014 or 2015
CA and the Southwest dominate the sample, but 2014 and 2015 saw a broadening of the market to TX, AR, AL, FL—and even MN and MI
Strong/steady downward price trend since 2006 to <50$/MWh in 2015
Smaller projects (e.g., 20-50 MW) seemingly no less competitive
>75% of the sample is currently operational
26
$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
PPA Execution Date
California
Southwest
Texas
Southeast
Midwest
Leve
lized
PPA
Pric
e (R
eal 2
015
$/M
Wh)
550 MW
210MW
50 MW
4 of 5 regions now have PPA prices<$50/MWh (Midwest <$60/MWh)
Sample includes 136 contracts totaling 9.1 GWAC
$0
$10
$20
$30
$40
$50
$60
$70
$80
Jan-
14
Feb-
14
Mar
-14
Apr-
14
May
-14
Jun-
14
Jul-1
4
Aug-
14
Sep-
14
Oct
-14
Nov
-14
Dec
-14
Jan-
15
Feb-
15
Mar
-15
Apr-
15
May
-15
Jun-
15
Jul-1
5
Aug-
15
Sep-
15
Oct
-15
Nov
-15
Dec
-15
PPA Execution Date
California Southwest Texas Southeast Midwest
Leve
lized
PPA
Pric
e (R
eal 2
015
$/M
Wh)
150MW
45MW
Sample includes 40 contracts totaling 2,562 MWAC that were priced in 2014 or 2015
7 MW 100 MW
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
On average, levelized PPA prices have fallen by nearly 75% since 2009
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
Remarkably 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 commercial operation date (orange columns in bottom graph) rather than PPA execution date (blue columns)
Though the average levelized price of PPAs signed in 2015 is ~$40/MWh, the average levelized PPA price among projects that came online in 2015 is significantly higher, at ~$85/MWh
27
0
50
100
150
200
250
200617
200715
20083
770
200916
1,030
201026
1,640
201116
1,584
201214
931
201319
568
201416
918
201524
1,644
Generation-Weighted Average
Individual Contract
PPA Year:Contracts:
MW:
Leve
lized
PPA
Pric
e (R
eal 2
015
$/M
Wh)
$0
$50
$100
$150
$200
$250
2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
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
015
$/M
Wh)
2016 isprovisional
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
PV PPA prices generally decline over time in real dollar terms, in contrast to fuel cost projections
~70% of PV sample has flat annual PPA pricing (in nominal dollars), while the rest escalate at low rates
Thus, average PPA prices decline over time in real dollar terms (top graph)
Bottom graph compares 2015-vintage PPA prices to range of gas price projections from AEO 2016, showing that…
…although PV is currently priced higher than the cost of burning fuel in a combined-cycle unit, over longer terms PV is 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
015
$/M
Wh)
2006 (7 MW, 1 PPA)
2009 (1,030 MW, 16 PPAs)2008 (770 MW, 3 PPAs)
2010 (1,640 MW, 26 PPAs)
2011 (1,584 MW, 16 PPAs)2012 (931 MW, 14 PPAs)
2013(568 MW, 19 PPAs)
2014 (918 MW, 16 PPAs)
2007(5 MW, 1 PPA)
2015 (1,644 MW, 24 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
Overall range of AEO 2016 gas price projections (converted to $/MWh terms) AEO 2016 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 error bars) over time
2015
$/M
Wh
PV PPA sample includes 24 contracts priced in 2015, totaling 1,644 MWAC
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Concentrating Solar Power (CSP) Projects
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 storage and/or new technology (power tower) in some cases, making comparisons difficult
31
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Newer CSP projects have struggled with teething issues, but performance improved in 2015
32
Capacity factors at Solana and Ivanpah improved in 2015, but were still below long-term, steady-state expectations (the ramp-up is still in progress)
Genesis maintained its 2014 capacity factor (at expectations), but similar Mojave trough project fell a little short
Newer CSP projects generally performing better than older CSP projects, but not necessarily any better than PV projects
SEGS I & II have been decommissioned (and may be replaced with PV)
0%
5%
10%
15%
20%
25%
30%
35%
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
Net
Cap
acity
Fac
tor (
sola
r por
tion
only
)
SEGS I & II
SEGS III-IX
Genesis
Solana
Ivanpah
MojaveNevada Solar One (dashed)
For reference: average PV in CA, NV, AZ (red diamonds)
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 since 2011
33
$0
$50
$100
$150
$200
$250Ja
n-06
Jan-
07
Jan-
08
Jan-
09
Jan-
10
Jan-
11
Jan-
12
Jan-
13
Jan-
14
Jan-
15
Jan-
16
PPA Execution Date
PV (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
015
$/M
Wh)
250 MW
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
0
5
10
15
20
25
California Texas Southwest(NV, AZ, UT, CO, NM)
Southeast Central Northeast Northwest
Nam
epla
te S
olar
Cap
acity
(GW
)
Entered queue in 2015 Total in queue at end of 2015
0
40
80
120
160
Gas Wind Solar Nuclear Coal Other
Nam
epla
te C
apac
ity (G
W)
Looking ahead: long-term ITC extension should support continued growth in the utility-scale solar pipeline
December 2015’s extension of the 30% ITC through 2019 (along with the switch to a “start construction” rather than “placed in service” deadline), with a gradual phase down to 10% thereafter, should ensure continued momentum for the foreseeable future
56.8 GW of solar was in the queues at the end of 2015 (up from 44.6 GW at end of 2014): more than 5 times the installed solar capacity in our project population at the end of 2015
Solar was in third place in the queues, behind natural gas and wind
34
Graphs show solar and other capacity in 35 interconnection queues across the US:
• Inset compares solar to other resources
• Main graph shows location of solar
Project Site: http://utilityscalesolar.lbl.gov @BerkeleyLabEMP
Relative growth of solar pipeline in various regions suggests a broadening market
The utility-scale solar pipeline has been replenished and has even grown in recent years, despite the record buildout in 2014 and 2015
Although California and (to a lesser extent) the Southwest still dominate the interconnection queues, recent growth in the queues has come largely from outside of those two traditional markets—e.g., Texas and the Southeast, Central, and Northeastern regions
Not all of these projects will ultimately be built (some will undoubtedly fall by the wayside)
35
0
10
20
30
40
50
60
Total California Southwest Texas Southeast Central Northeast Northwest
purple shows new capacity that entered the queues in year shown20
1320
1420
15
2013
2014
2015
Sola
r in
Que
ues a
t Yea
r-En
d (G
W)
green shows legacy capacity that entered the queues in prior years
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
36
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.