The performance of wind farms Evidence from the UK, Denmark and the - - PowerPoint PPT Presentation
The performance of wind farms Evidence from the UK, Denmark and the US Professor Gordon Hughes University of Edinburgh 28 th January 2013 Background Wind power is a capital intensive technology Projections of load factors are critical
2
Projections of load factors are critical for any
Systematic variations in wind conditions Variations across sites & operational regimes Changes in composition of turbine types, ages, etc
62 62 62
50 100 150 50 100 150 50 100 150 50 100 150
55578 55579 55581 55747 55795 55796 55968 55992 56111 56211 56212 56225
Load factor (%) Months since Jan 2000
Graphs by plant_id
Consistent estimates of coefficients under quite
Standard errors consistent if errors are
Cross-check using bootstrap standard errors No assumptions required about the distribution of
7
Analysis of ROC monthly output for 2002-12
Strong incentive to report reliable data
Standard panel fixed effects models allowing for age,
Multiplicative (log-linear) specification for load factor Unit fixed effects capture site & location effects Total of 296 separate reporting units analysed with total
Extensive testing to identify additional factors which
8
9
10
11
12
Data reported to Danish Energy Agency
Nominally for each turbine but aggregated for sites linked to
Data summed to wind farms defined by location, turbine
Onshore wind farms in Denmark are smaller and
823 wind farms commissioned in 1992 or later with total
Limited sample of offshore wind farms
all in shallow water & most constructed in 2002 or later 30 wind farms with total capacity of 860 MW
13
14
15
Data from Energy Information Agency (EIA)
Annual figures for plant characteristics Monthly data for output – reported either monthly (largest
Data cleaning
Inconsistencies over time in reported capacity, etc Sample of 673 plants over US Initial analysis focuses on Texas [92 plants, 11.2 GW] and
Extended to WSC (TX & OK – 13.6 GW), WNC (MN, IA, KS,
16
17
18
19
20
5 10 15 20 Wind farm age 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
21
22
23
24
25
*** p<0.01, ** p<0.05, * p<0.1 Robust standard errors in parentheses 291 291 291 291 Number of groups 0.460 0.560 0.579 0.658 R-squared (0.067) (0.065) (1.552) (1.320) 1.140*** 3.413***
29.70*** Constant (0.001) (0.026) 0.00529*** 0.113*** Month (0.003) (0.075) 0.204*** 5.191*** Wind (0.002) (0.002) (0.038) (0.035)
Age^2 * large (0.001) (0.001) (0.021) (0.023)
Age^2 * medium (0.001) (0.001) (0.008) (0.007) 0.000307 0.00033 0.00417 0.00542 Age^2 (0.016) (0.016) (0.269) (0.252) 0.0735*** 0.0689*** 1.515*** 1.288*** Age * large (0.015) (0.015) (0.238) (0.227) 0.0422*** 0.0412*** 0.659*** 0.513** Age * medium (0.018) (0.015) (0.339) (0.257)
Age Wind & Time Full Wind & Time Full Multiplicative model Additive model
31
Average
2011
2010
2009
2008
2007
2006 0.17 2.56
2005
2004 0.56 10.31 0.14 3.44 2003 0.69 11.46 0.21 3.83 2002 0.68 10.40 0.18 2.07 2001 0.71 9.55 0.06
2000 Multiplicative Additive Multiplicative Additive Capacity weights Equal weights
32
Bootstrap standard errors in parentheses; *** p<0.01, ** p<0.05, * p<0.1 0.529 0.527 0.425 0.421 R-squared 291 291 291 291 Observations (0.034) (0.030) (0.588) (0.564) 0.0723** 0.0803*** 0.552 0.817 Constant (0.006) (0.182)
(Year - 2000) * Scotland (0.003) (0.003) (0.069) (0.064)
Year - 2000 (0.001) (0.001) (0.016) (0.017)
Capacity in MW (0.044) (0.041) (0.955) (0.924) 0.137*** 0.133*** 2.865*** 2.725*** Wales (0.054) (0.038) (1.587) (0.910) 0.241*** 0.202*** 6.554*** 5.259*** Scotland (0.047) (0.047) (1.050) (0.999) 0.195*** 0.200*** 4.520*** 4.683*** Northern Ireland (4) (3) (2) (1) Multiplicative model Additive model
33
34
35
Economic life of wind farms is no more than 15 years
Many onshore wind farms re-powered after 10-12 years After 10 years the residual value of turbines is low, but there
Costs of meeting renewable targets much higher than
Higher capacity required due to lower average load factors Shorter operating lives implies higher replacement costs
Implications for financing investments
Not attractive to many infrastructure investors Higher cost of capital due to uncertainty about length of
36
Impact of load factors on levelised costs
Increase from £86 to £183 per MWh for onshore wind and
Market prices and/or subsidies required to fund
DECC/CCC analysis based upon erroneous assumptions Allowance of £8-10 per MWh for price differential Total subsidy required in range £115-145 per MWh without
Extra cost of offshore wind not as large as often thought
37
38
39
Wake effects & maintenance regimes Community ownership: capital vs operating costs Resistance to larger turbines
Planning restrictions: upland sites and greater
Treating wind turbines as short rotation forests Link to cost of capital
40
More detailed analysis of existing data Longer term monitoring of installations Implications for maintenance regimes
Who should take performance risk?
Structuring subsidies Impact on financing new projects