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.


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

http://trackingthesun.lbl.gov/

http://utilityscalesolar.lbl.gov/


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


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.


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


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


PV projects

7

Photo Credit: sPower SEPV Palmdale East


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


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%


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


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.


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)


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


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:


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.


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.


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 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


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


NREL 201525 MW-DC

National AverageUnion Labor

BNEF 2015Unknown Size

California

GTM 201510 MW-DCCaliforniaEPC Only

NREL 2015100 MW-DC


NREL 201525 MW-DC


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


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


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


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


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


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


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

https://emp.lbl.gov/publications/maximizing-mwh-statistical-analysis


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


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


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


Concentrating Solar Power (CSP) Projects

29

Photo Credit: Solar Reserve: Crescent Dunes


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


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


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)


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


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


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


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.

http://utilityscalesolar.lbl.gov/




Utility-Scale Solar 2015 - Lawrence Berkeley National ... · Presentation Outline . Strong growth...

Documents

Transcript of Utility-Scale Solar 2015 - Lawrence Berkeley National ... · Presentation Outline . Strong growth...