how does risk aversion shape overplanting in the design of ......motivation modelling case study and...
TRANSCRIPT
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
Motivation Modelling Case Study and Results Conclusions and Future Work References
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 Modelling Case Study and Results Conclusions and Future Work References
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
Motivation Modelling Case Study and Results Conclusions and Future Work References
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]
Motivation Modelling Case Study and Results Conclusions and Future Work References
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
Motivation Modelling Case Study and Results Conclusions and Future Work References
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
Motivation Modelling Case Study and Results Conclusions and Future Work References
Modelling of Overplanting
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
Motivation Modelling Case Study and Results Conclusions and Future Work References
Modelling of Overplanting
Modelling of Overplanting
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
Motivation Modelling Case Study and Results Conclusions and Future Work References
Modelling Risk Aversion
Modelling Risk Aversion
[Rockafellar and Uryasev 2002]
Risk metric
ρα[λ, LCOE ] = λCVaRα[LCOE ] + (1 − λ)Median[LCOE ]
Motivation Modelling Case Study and Results Conclusions and Future Work References
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
Motivation Modelling Case Study and Results Conclusions and Future Work References
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 %
Motivation Modelling Case Study and Results Conclusions and Future Work References
Local Sensitivity Analysis
Local Sensitivity Analysis - wind speed
Motivation Modelling Case Study and Results Conclusions and Future Work References
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
Motivation Modelling Case Study and Results Conclusions and Future Work References
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?
Motivation Modelling Case Study and Results Conclusions and Future Work References
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.
Motivation Modelling Case Study and Results Conclusions and Future Work References
Questions
How does risk aversion shape overplanting in the design of offshore wind farms?
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).