how does risk aversion shape overplanting in the design of ......motivation modelling case study and...

Motivation Modelling Case Study and Results Conclusions and Future Work References

How does risk aversion shape overplanting in the design ofoffshore wind farms?

Esteve Borras Mora1,2 James Spelling 2 Adriaan H. van der Weijde 3

1Industrial Doctoral Centre for Offshore Renewable Energy (IDCORE), University of EdinburghEdinburgh, EH9 3JL, UK

2EDF Energy R&D UK CentreInterchange, 81-85 Station Road, Croydon, CR0 2AJ, UK

3University of Edinburgh School of Engineering and the Alan Turing InstituteFaraday Building, The King’s Buildings, Mayfield Road, Edinburgh, EH9 3DW, UK

EERA DeepWind’19 Trondheim, 16-18 January 2019


Outline

1 MotivationOverplanting Studies

2 ModellingOffshore Wind Cost Modelling ToolFactors Affecting OverplantingModelling of OverplantingModelling Risk Aversion

3 Case Study and ResultsCase StudyDeterministic ResultsLocal Sensitivity AnalysisStochastic Results

4 Conclusions and Future Work


Motivation

� Farms subjected to a maximum exportcapacity agreed with the TSO

� Generators can export up to theircontracted maximum export capacity

� Majority of the time offshore wind farms arenot generating at full power

� Can overplanting result in better overalleconomics despite power output beingcurtailed at generations’ peaks?

[Wolter et al. 2016]

Overplanting

Over installing the offshore wind capacity to the fixed electrical infrastructure


Overplanting Studies

Overplanting Studies

Round 3 Offshore Wind

National Grid UK

[Grid 2008]

Dogger BankForewind UK

[Forewind 2012]

Connection OfferPolicy & Process

CER Ireland[Brid ODonovan 2011]

Academic Literature[Mcinerney and Bunn 2017]

Wind FarmZone Borssele

TenneT Netherlands[TenneT 2015]

Decision on InstalledCapacity CapCER Ireland[Morris 2014]


Offshore Wind Cost Modelling Tool

Offshore Wind Cost Modelling Tool

[Borras Mora 2017]

Characteristics

� Aim : rapidly evaluate the financialperformance of a farm

� Inputs : project specifications, technologychoices and market trends

� Outputs : financial metrics based on LCOE

� Structure : 4 main modules - Design, Cost,Financial and Stochastic

� Stochastic Framework: Quantitativeuncertainty management, Double loopMonte Carlo Simulation - inner loop withinAEP


Factors Affecting Overplanting

Factors Affecting Overplanting

Factors

� Ratio of wind turbine expenditure toelectrical infrastructure

� Wind speed distribution

� Wind turbine availability

� Inter-array cable availability

� Wake effects

� Electrical losses

� Degradation factor


Modelling of Overplanting


0 5 10 15 20 25 30 35 40

Wind Speed [m/s]

0

1

2

3

4

5

6

7

8

9

10

Win

d T

urbi

ne P

ower

[MW

]

Wind Turbine Power CurveConstrained Wind Turbine Power Curve

Modelling Type 1

0 10 20 30 40 50 60 70 80 90 100

Utilization per Year [%]

0

50

100

150

200

250

300

350

400

450

Win

d F

arm

Pow

er [M

W]

Aggregated Wind Farm Power CurveConstrained Wind Farm Power Curve

Modelling Type 2




Wind Speed Bin

Discretization / Monte Carlo

Wind Speed Sample

Calculate Wake

Effect Losses

Wake Effect Losses equal

to fwake?α

Accounting for Wake

Effects

Theoretical Power Curve

modified by α Yes

No

0 5 10 15 20 25 30 35 40

Wind Speed [m/s]

0

0.005

0.01

0.015

0.02

0.025

0.03

Fre

quen

cies

[-]

= 0.8 = 0.9 = 1

Cases

0

1

2

3

4

5

6

7

8

9

10

Loss

es [%

]

0 5 10 15 20 25 30 35 40

Wind Speed [m/s]

0

2

4

6

8

10

12

14

16

18

20

Pow

er [M

W]

= 1 = 0.9 = 0.8

36 38 40 42 44 46 48 50

Number of Wind Turbines Operating [#]

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Cum

ulat

ive

Dis

trib

utio

n F

unct

ion

[-]

90% Availability94% Availability97% Availability

0 50 100 150 200 250 300 350 400 450

Offshore Wind Farm Power [MW]

0

0.05

0.1

0.15

0.2

0.25

0.3

Fre

quen

cies

[-]

= 1 = 0.9 = 0.8

0 10 20 30 40 50 60 70 80 90 100

Utilization [%]

0

50

100

150

200

250

300

350

400

450

OW

F P

ower

[MW

]

Modelling Type 1Modelling Type 2


Modelling Risk Aversion

Modelling Risk Aversion

[Rockafellar and Uryasev 2002]

Risk metric

ρα[λ, LCOE ] = λCVaRα[LCOE ] + (1 − λ)Median[LCOE ]


Case Study

Case Study

Case Study

400MW commercial offshore wind farm400MW fixed maximum export capacity50-8MW WTGs0-14% overplanting

2% overplanting = 1 additional WTG

Characteristic Value UncertaintyWater Depth [m] 25 NoneDistance from shore [km] 25 None

Mean Wind Speed @ 100m [m/s] 9 N (9, 0.12)Wind Turbine Availability [%] 95 U(90, 97)Inter-Array Cable Availability [%] 99 U(97, 99)Foundation Type [-] Monopile NoneElectrical Infrastructure [-] HVAC NoneWind Turbine Type [-] 164-8 MW NoneWake effect [%] 10 NoneDegradation Factor [%] 0.5 None


Deterministic Results

Deterministic Results & Local Sensitivity Analysis

0 2 4 6 8 10 12 14

Overplanting [%]

100

105

110

115

Nor

mal

ised

Yie

ld [%

]

Unconstrained yieldConstrained yield

0 2 4 6 8 10 12 14

Overplanting [%]

98

99

100

101

102

103

104

Nor

mal

ised

LC

OE

[%]

8 m/s9 m/s10 m/s

0 2 4 6 8 10 12 14

Overplanting [%]

98

99

100

101

102

103

104

Nor

mal

ised

LC

OE

[%]

0 2 4 6 8 10 12 14

Overplanting [%]

98

99

100

101

102

103

104

Nor

mal

ised

LC

OE

[%]

90 %93 %95 %97 %


Local Sensitivity Analysis

Local Sensitivity Analysis - wind speed


Stochastic Results

Stochastic Results ρα[λ, overplanting ]

75 80 85 90 95 100 105 110

Normalised LCOE [%]

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08K

erne

l Den

sity

Fun

ctio

n [-

]

400 MW408 MW424 MW456 MW

90 100 110 120 130 140 150 160 170 180

Normalised LCOE [%]

0

0.01

0.02

0.03

0.04

0.05

0.06

Ker

nel D

ensi

ty F

unct

ion

[-]

400 MW408 MW424 MW456 MW

0 2 4 6 8 10 12 14

Overplanting [%]

99

99.5

100

100.5

101

101.5

102

Nor

mal

ised

Ris

k M

etric

[%]

Risk aversion = 0 = 0.25 = 0.5 = 0.75

Risk neutrality = 1


Conclusions and Future Work

Conclusions

� Development of a novel framework to evaluate overplanting

� Modelling Type 1 is easier to implement but may lead to an overestimation of the annualenergy production

� Modelling Type 2 is more accurate but requires higher computational costs

� Wind turbine availability is the most sensitive parameter to overplanting

� Previous studies based on low wind turbine availabilities rates or on Modelling Type 1, needto be revisited

� Optimal overplanting setup increased when considering the uncertainty quantificationframework regardless of risk appetite (from 2% to 4%)

� Overplanting the reference farm from 2% to 8% gives a better result than with nooverplanting for a risk neutral setting

Future Work

� How is overplanting influence by larger turbines and sites located futher from shore?

� How does risk aversion influence the decision for these new sites?


References

Borras Mora, Esteve (2017). “Transition from Deterministic to Stochastic Cost Models forOffshore Wind Farms”. In: Offshore Wind Energy Conference 44.June.

Brid ODonovan (2011). Connection Offer Policy & Process (COPP). Tech. rep. Commission forEnergy Regulation.

Forewind (2012). Environmental Statement Chapter 6 Appendix B Offshore Project BoundarySelection Report. Tech. rep. March.

Grid, National (2008). Round 3 Offshore Wind Farm Connection Study. Tech. rep.

Mcinerney, Celine and Derek W Bunn (2017). “Optimal over installation of wind generationfacilities”. In: Energy Economics 61, pp. 87–96. issn: 0140-9883. doi:10.1016/j.eneco.2016.10.022. url: http://dx.doi.org/10.1016/j.eneco.2016.10.022.

Morris, Nigel (2014). Decision on Installed Capacity Cap Installed Capacity Cap. Tech. rep.Commission for Energy Regulation.

Rockafellar, R Tyrrell and Stanislav Uryasev (2002). “Conditional value-at-risk for general lossdistributions”. In: Journal of Banking & Finance 26, pp. 1443–1471.

TenneT (2015). POSITION PAPER Overplanting. Tech. rep. TenneT, pp. 1–7.

Wolter, C et al. (2016). “Overplanting in Offshore Wind Power Plants in Different RegulatoryRegimes”. In: 15th Wind Integration Workshop 49.0, pp. 0–4. doi:10.1146/annurev.matsci.35.100303.110641.

https://doi.org/10.1016/j.eneco.2016.10.022

http://dx.doi.org/10.1016/j.eneco.2016.10.022

https://doi.org/10.1146/annurev.matsci.35.100303.110641


Questions

How does risk aversion shape overplanting in the design of offshore wind farms?

[email protected]

Acknowledgements

This work is sponsored by EDF Energy R&D UK and the Industrial Doctoral Centre for OffshoreRenewable Energy (IDCORE), a consortium of the University of Exeter, University of Edinburghand University of Strathclyde. IDCORE is funded by both the Energy Technologies Institute andthe Research Councils Energy Programme through grant number EP/J500847/1. Additionalsupport came from the UK Engineering and Physical Sciences Research Council through grantnumber EP/P001173/1 (CESI).

mailto:[email protected]

how does risk aversion shape overplanting in the design of ......motivation modelling case study and...

Documents