Estimation of the Optimum Wind Turbine Size for Two Different Offshore Sites and Wind Farm Rated Powers MSc. Candidate: Shajid Kairuz Fonseca Supervisor: Dr. ir. Michiel B. Zaaijer Faculty of Aerospace Engineering Wind Energy Department August 30th 2017

Outline Introduction Approach Models Results Sensitivity analysis Conclusions and recommendations

Outline Introduction Approach Models Results Sensitivity analysis Conclusions and recommendations

Problem Overview: Large turbines more energy yield, but high costs Which is the optimum turbine size? Difficulties: The industry has not reached to this optimum size Lack of data regarding turbine components dimensions and costs The optimum turbine size is not the same for all situations

Objectives: Main Objective Gain insight into the optimum size of wind turbines for offshore applications. Specific Activities Determine scaling approach for main components of an offshore wind farm. Determine how the cost of the components changes for different turbine scales and case characteristics Establish a cost of energy function to identify the optimum turbine size in each case. Determine components that have largest impact on optimum turbine scale with a sensitivity analysis. 

Outline Introduction Approach Models Results Sensitivity analysis Conclusions and recommendations

How to define ‘size’ for a wind turbine? The turbine size is defined in terms of the turbine power and the rotor diameter Although a correlation may be expected between these two parameters they are independent from each other

What is optimum? The Levelized Production Cost (LPC) is the objective function Composed by: Cinvestment  Capital investment costs CO&M  Operation and Maintenance costs Cdecommissioning  Decommissioning costs r  project discount rate T  project lifetime A  Annuity factor Ey  Energy yield

Which LPC components to model? Costs: Turbines Support Structure Electrical Infrastructure O&M Energy yield Individual turbine energy yield Wake efficiency Electrical losses Availability

How to approach upscaling? Square cube law Empirical Analysis

Case Studies Since the optimum turbine size is not the same for all situations Far North Sea Baltic Sea Farm rated power  500MW Distance to grid  100km Water Depth  40m Farm rated power  100MW Distance to grid  20km Water Depth  15m

The MZ tool Outputs: Inputs: System Farms Support structure Sites Layout Electrical system Maintenance Cost Details Inputs: Farms Sites Turbines

Outline Introduction Approach Models Results Sensitivity analysis Conclusions and recommendations

Turbines cost model Blades costs ≈ 32% of turbine cost excluding tower and transformer Other components as: Direct drive generator Power converter Pitch system Cost models taken from (L. Fingersh et al., 2006) NREL

Turbines cost model - Blades Driver of blades cost  materials  carbon fiber  hybrid carbon-glass fiber blade

Support structure cost model Monopile foundation  most used for offshore wind farms & suitable up to ≈ 40m Main drivers for the cost: Water Depth  Site location Wave conditions  Site location Thrust Force  Turbine characteristics (Pnew & Dnew) Assuming that the new turbine achieves the same maximum CP & CT @ rated wind speed as the reference turbine (V80-2MW) Scaling law: Specify the concept: Monopile Components: Monopile, transition piece, tower, scour protection etc

Support structure cost model MZ inputs Maximum operational thrust  scaling law Rotor radius  empirical approach Yaw diameter  square cube law Front area nacelle  square cube law Tower top mass  square cube law MZ outputs Technical data & costs of: Monopile Transition piece Tower Scour protection Installation foundation

Support structure cost model Power Thrust Diameter

Electrical Infrastructure cost model HVAC system  cost effective up to 100-120km Main components: Cables Transformers Shunt reactor Switch gear Largest weight in the total cost of the system Dependent on: Distance to shore  Case characteristic Farm rated power  Case characteristic Turbine rated power  Turbine characteristic Specify the concept: AC transmission, also the components that make the whole: cables, transformers, switch gear, shunt reactor

Electrical Infrastructure cost model MZ tool inputs: Distance to shore  (100km for Far North Sea & 20km for the Baltic Sea) Voltage at grid coupling point  400kV for both cases Voltage of the turbine generator  empirical approach for the upscaled turbines A square layout is established

Electrical Infrastructure cost model

O&M cost model The MZ tool requires of the following inputs: Preventive maintenance cost & frequency Consumables repairs costs The MZ tool gives the following outputs: Their cost increase linearly with the ratio of the turbine powers Operation costs Administration Grid charge Bottom lease Maintenance costs Consumables repairs & service Personnel Access vessels Lifting equipment Subsea inspections

O&M cost model After using the MZ tool the cost models developed for the O&M components: Administration – fixed cost for both cases Grid charge – proportional to farm energy yield Consumables repairs Consumables service Bottom lease – proportional to the farm area: Fixed cost per case  dependent on farm rated power

O&M cost model After using the MZ tool the cost models developed for the O&M components: Personnel Access vessel Lifting equipment Subsea inspections Good correlation with the number of turbines in the farm, and therefore to the turbines rated power

Energy yield model Energy yield of a wind turbine  Thr  Number of hours in a year Vin  cut-in wind speed Vout  cut-out wind speed P(U)  power @ wind speed U  power curve f(U) du  Weibull distribution Difference of 0.5Wh/y between 1.95MW to 20MW Specify the height for the weibull parameters: 100m, mention it in the final report

Outline Introduction Approach Models Results Sensitivity analysis Conclusions and recommendations

Results After introducing the components costs and the energy yield model in the LPC function:

Results LPC as a function of the turbine power for the optimum turbine sizes in the range from 2 to 20MW in steps of 1MW Optimum turbine size  11MW-176m Optimum turbine size  6MW-147m

Results Behavior in the contribution to the LPC for different components plus the energy yield trend for the optimum turbine sizes in the range from 2 to 20MW in steps of 1MW for the Far North Sea case

Outline Introduction Approach Models Results Sensitivity analysis Conclusions and recommendations

Far North Sea case Sensitivity of the model when varying some components costs Also: The lower the price ratio of carbon fiber to glass fiber, the larger the optimum size (13MW @7.5, 14MW @5.0, hypothetical case of just glass fiber  15MW )

Baltic Sea case Sensitivity of the model when varying some components costs Also: The lower the price ratio of carbon fiber to glass fiber, the larger the optimum size (8MW @5.0, hypothetical case of just glass fiber  10MW )

Outline Introduction Approach Models Results Sensitivity analysis Conclusions and recommendations

Conclusions The optimum turbine size is result of trade-off between components that their cost reduce and the components that their cost rise as turbine scale increases. The optimum turbine rated power for a far offshore, large wind farm is in the range of 10 to 13MW while for a small, near shore wind farm is in the range of 5 to 7MW. The blades cost are a limiting factor for the increment in rotor diameter for large turbines scales. The main objective of the project was achieved which was getting insight into the upscaling of wind turbines for offshore applications.

Recommendations Incorporate support structure cost models of other foundations types and an electrical infrastructure cost model for HVDC transmission, so the model is more versatile. Obtain actual costs of the largest blades manufactured nowadays so that these real costs are compared to the ones given by the blade cost model.

Thank you for your attention! Questions?

Approach Annuity factor Real interest or discount rate

Approach Full list of OWF components: Capital Investment:   Support Structure Electrical Infrastructure Turbines Boat landing structure Offshore Platform Measuring tower Onshore premises Harbor use Warranty Transportation of the RNA onshore Installation of the RNA offshore Dune crossing Grid connection Transmission cable installation Infield cables installation Engineering Site assessment Management O&M: Administration Grid charge Bottom lease Insurance Personnel Access Vessels Lifting Equipment Subsea inspections Consumables repair Consumables service Management Decommissioning: Removal of the RNA’s Removal of the foundations and scour protection Removal of the offshore platform and meteo tower Removal of the transmission cable Removal of the infield cables Site clearance Disposal of the RNA’s Management

Cost models Turbine - Blades Spare

Cost models Turbine - Blades weight ratio material cost ratio total cost ratio

Cost models Turbine - Blades Spare

Cost models Support structure cost model  (5) (6)  (7) (8)

Cost models Support structure cost model

Cost models Electrical Infrastructure - Layouts

Cost models Electrical Infrastructure – Generator Voltage

Cost models Electrical Infrastructure

Cost models O&M

Cost models O&M

Cost models O&M

Energy yield model Wind speed scaling for different heights:

Energy yield for the Baltic Sea case computed with the MZ tool Energy yield model Energy yield for the Baltic Sea case computed with the MZ tool

Energy yield model To determine the power curve for the upscaled turbines: Assume that upscaled turbines achieve the same maximum CP at rated power as the reference turbine (V80-2MW) Adjust the wind speed axis of the power curve for the upscaled turbines with Ec.6 Vnew is introduced in Ec. 5 and using the same power coefficient values that the reference turbine has  power of the upscaled turbine @ certain wind speed Repeating step 3 for the velocity range of Vin to Vout  power curve dependent on the rotor diameter and the turbine power  (5) (6)

Increment when going from 400 to 300W/m2 Results After introducing the components costs and the energy yield model in the LPC function: Increment when going from 400 to 300W/m2 LPC as a function of the power and the rotor diameter for the Baltic Sea case

Results Far North Sea

Results Far North Sea

Results Far North Sea

Results Baltic Sea

Results Baltic Sea

Results Baltic Sea Behavior in the contribution to the LPC for different components plus the energy yield trend for the optimum turbine sizes in the range from 2 to 20MW in steps of 1MW for the Baltic Sea case

Results Baltic Sea

Results Gradients comparison of the dominant components in the model results for the Far North Sea case Gradients comparison of the dominant components in the model results for the Baltic Sea case

Sensitivity study Far North Sea Baltic Sea

Sensitivity study Far North Sea

Sensitivity study Baltic Sea

Sensitivity study When the costs of the consumables repairs scales with the ratio of the upscaled turbine rated power to the reference turbine rated power to a power larger than 1

Appendix Characteristics of the case studies:

Appendix V80-2MW technical data:

Retrieved from: https://www. technologyreview