Moving Towards Large(r) Rotors Is that a good idea?

P. K. Chaviaropoulos, CRES

H.J.M. Beurskens, SET Analysis

S.G. Voutsinas, NTUA

EWEA 2013 ConferenceVienna, 4-7 February

V164-8.0 MW

Siemens 6 MW Offshore WT

Alstom Haliade 150-6MW

Sway 10 MW Turbine

Upwind/NREL/DOWEC UpScaled to 10 MW

SeaTitan 10 MW (HTS generator)

ROTORUPWIND [6] / Upscaled

SWAY [5]SeaTitan 10MW [4]

Siemens 6 MW [3] Scaled to 10 MW

Vestas 8 MW [1] Scaled to 10 MW

Alstom 6 MW [2] Scaled to 10 MW

Diameter (m) 178 164 190 199 183 194Blade Length (m) 87 97 89 95Design RPM 8,56 12 10 8,52 9,39 8,91Rated Wind Speed (m/s) 11,3 13 13 13Tip Speed (m/s) 80,0 103,0 99,5 88,7 90,2 90,3

Motivation

Motivation

Large Multi MW Offshore commercial turbines with• Higher Tip Speed (~90m/s)• Large Diameter

Concept 10 MW designs with• Even Higher Tip Speed (~100m/s)• Even Larger Diameter

High Tip Speed -> High TSR-> Lower Solidity -> Thicker airfoilsHigher Diameter-> Higher Loads for OPT Cpmax

HOW DOES THAT WORK ?

Quantification with BEM Theory

Ref: Peter Jamieson, Innovation in Wind Turbine Design, A John Wiley & Sons, Ltd., Publication, ISBN 978-0-470-69981-2, 2011

All Functions of a, λ, x

Low Induction Rotors (LIRs)

Conventional DesignFor given Ro and λ

Maximize CP(λ,a)

0 0.2 0.4 0.6 0.8 1

alfa

0

0.1

0.2

0.3

0.4

0.5

CP

0

0.05

0.1

0.15

0.2

0.25

CM

(0)

CP

CM(0)

RIGHT: Plots of non-dimensional coefficients, candidates for blade optimization, versus axial induction coefficient a. Application for Β=3, k=100 and λ=8.85 (providing the maximum CP value =0.4966 for the selected B and k combination).

New, R is a design variable

Maximize {[CP(λ,a).R2]/[CP0(λ0,a0).R0

2]}Constrained by[CM(0)(λ,a).R3]/[CM0(0)(λ0,a0).R0

3]=1

Or Maximize { CP(λ,a)/ CM(0)(λ,a)2/3}0 0.2 0.4 0.6 0.8 1

alfa

0

0.4

0.8

1.2

1.6

2

CP/C

M(0

)2/3

0

0.2

0.4

0.6

0.8

1

CP/C

M(0

)1/3

CP/CM(0)1/3

CP/CM(0)2/3

• As a result, the optimal rotor will have a larger radius: R/R0 = 1.136.

• It will capture more energy: [CP(λ,a).R2]/[CP0(λ0,a0).R02] = 1.087 and it will be less

loaded than the initial one (design CT and CM(r) will be less), operating at a lower axial induction value a=0.187 instead of a0=0.33.

• In other words, we sacrificed CP to increase energy capture with a larger rotor diameter, maintaining the aerodynamic bending moments at their initial level.

• The obvious question is “Is that cost effective, as long as we have a 13% longer blade now”? In a very primitive approach and taking only the aerodynamic moments into account, we can assume that the new blade can maintain the cross-sections of the initial blade as long as they have the structural strength to undertake the reference bending loads which are not altered.

• This means that the weight and “cost” of the new blade will increase by a factor of R/R0. This would result in an increase of the levelized cost of the blade component of 4.6% (13% additional cost, 8.7% more power). Nevertheless, and since the levelized cost of the rotor blades is a small fraction of the overall levelized cost of electricity, particularly offshore, the selection of a larger, less loaded rotor for offshore turbines seems cost effective.

Low Induction Rotors (LIRs)

0 0.2 0.4 0.6 0.8 1

x=r/R

0

0.04

0.08

0.12

0.16

c.c L

/R

Reference

New

0 0.2 0.4 0.6 0.8 1

x=r/R

0

20

40

60

φ(d

eg)

ReferenceNew

0 0.2 0.4 0.6 0.8 1

x=r/R

0

0.04

0.08

0.12

0.16

0.2

CM

ReferenceNew

0 10 20 30

Wind Speed (m/s)

0

4

8

12

16

N (

rpm

)

0 10 20 30

Wind Speed (m/s)

0

0.1

0.2

0.3

0.4

0.5

Cp

0 10 20 30

Wind Speed (m/s)

0

1

2

3

4

5

Pow

er (

MW

)

Low Induction Rotors (LIRs)

Characteristic properties of rotors with the same root bending moment designed for different values of the axial induction factor. Plots are presented for the rotor diameter (D), the power production at design wind speed P (Vdes), the levelized rotor cost (LCE) and the annual energy production AnEP. All properties are divided by their corresponding reference values (a=1/3)

Low Induction Rotors (LIRs)

LIRs in Large Wind Farms

0 20 40 60 80 100

WT Number

46

48

50

52

54

56

Ca

pac

ity F

act

or

(%

)

0 20 40 60 80 100

WT Number

2

4

6

8

10

12

14

Wak

e L

oss

es

(%)

• Capacity factor and wake losses per turbine in a 10X10 offshore wind farm for 8D spacing. Red dots address the initial turbines (highly loaded) and blue squares the less loaded turbines. The dashed red and the blue line correspond to the wind farm mean values.

• The less loaded turbines increase the wind farm capacity factor by nearly 6%. This comes partly (3%) from the increased annual production of the larger diameter turbine and partly (another 3% roughly) from the reduction of the wake losses due to the lower axial induction and, therefore, thrust coefficients of the larger rotors.

LIRs in Large Wind Farms

5 6 7 8

X times Diameter (125 m)

2050

2100

2150

2200

2250

2300

AE

P (

GW

h)

• Annual energy production of a 10X10 offshore wind farm as a function of turbines spacing (x times Diameter). Red dots address the initial turbines (highly loaded) and blue squares the less loaded turbines.

• The 6% gain in annual energy production is more or less flat and independent from the turbine spacing. Evidently, this is very much connected to the accuracy of wake calculation in multi-raw offshore wind farms with engineering models, like the one used.

Synthesis

• We investigated in some depth the impact of low induction rotor designs to the cost of offshore wind energy.

• We first demonstrated that a low induction high swept area rotor can capture more energy than a conventional design that aims at CP,MAX, without a burden on the aerodynamic loading of the blades.

• Then we estimated the additional cost of the larger rotor and concluded that this cost is smaller than the expected benefit in energy production, especially offshore where the capital cost of the rotor corresponds to a small fraction (around 5%) of the levelised cost of wind electricity.

• This analysis was first done in stand-alone operation. Moving to the wind farm level we anticipated additional benefits for the low induction rotor designs in terms of energy capture and wake losses (also related to extra fatigue loading).

• Closing, we want to make clear that this paper does not produce an actual design. It should be conceived as a first attempt for understanding possible “non-conventional” concepts that might be beneficial for the design of large offshore rotors. The methodology used and the assumptions made regarding the rotor and wind turbine loading are therefore pretty crude. For instance, in our analysis we have had no discussion on extreme loads, fatigue loads, blade-tower clearance issues etc.

THANK YOU