Accuracy of calculation procedures for offshore wind turbine support structures

Pauline de Valk – 27th of August 2013

Content Introduction Approach Modeling Results Conclusions and recommendations

1

Content Introduction Approach Modeling Results Conclusions and recommendations

1

Offshore wind energy shows potential to become one of the main energy suppliers

Demand for energy continues to increase

Offshore wind energy More steady wind flow

and average wind speed is higher than onshore

Introduction Approach Modeling Results Conclusions and recommendations

2

2009 2020 20350

5000

10000

15000

20000

25000

30000

35000

40000

Re-new-ables

Nuclear

Oil

Gas

Coal

Energy demand

TW

h

Offshore wind energy shows potential to become one of the main energy suppliers

Energy demand

TW

h

Demand for energy continues to increase

Offshore wind energy More steady wind flow

and average wind speed is higher than onshore

Cost of energy (€/kWh) should be decreased Structural optimization

design

2Introduction Approach Modeling Results Conclusions and recommendations

2009 2020 20350

5000

10000

15000

20000

25000

30000

35000

40000

Re-new-ables

Nuclear

Oil

Gas

Coal

Optimize structural design of the support structure

Support structure one of the main cost items

In order to optimize one should have confidence in the outcome of calculation procedures

3Introduction Approach Modeling Results Conclusions and recommendations

Thesis objective

‘‘Investigate the validity and conservatism

4Introduction Approach Modeling Results Conclusions and recommendations

Thesis objective

‘‘Investigate the validity and conservatism

of the current calculation procedures

4Introduction Approach Modeling Results Conclusions and recommendations

Thesis objective

‘‘Investigate the validity and conservatism

of the current calculation procedures for offshore wind turbine support

structures

4Introduction Approach Modeling Results Conclusions and recommendations

Thesis objective

‘‘Investigate the validity and conservatism

of the current calculation procedures for offshore wind turbine support

structures and propose improved procedures

4Introduction Approach Modeling Results Conclusions and recommendations

Thesis objective

‘‘Investigate the validity and conservatism

of the current calculation procedures for offshore wind turbine support

structures and propose improved procedures

based on these findings.”4Introduction Approach Modeling Results Conclusions and

recommendations

Content Introduction Approach Modeling Results Conclusions and recommendations

5

Offshore wind turbine support structure is custom engineered for every wind farm

Introduction Approach Modeling Results Conclusions and recommendations

6

Foundation designer (FD) Turbine designer (TD)

Offshore wind turbine support structure is custom engineered for every wind farm

Introduction Approach Modeling Results Conclusions and recommendations

6

Foundation designer (FD) Turbine designer (TD)

(Adjust) design foundation1

Offshore wind turbine support structure is custom engineered for every wind farm

Introduction Approach Modeling Results Conclusions and recommendations

6

Foundation designer (FD) Turbine designer (TD)

Integrate foundation modelin aero-elastic model

(Adjust) design foundation1 2

Offshore wind turbine support structure is custom engineered for every wind farm

Introduction Approach Modeling Results Conclusions and recommendations

6

Foundation designer (FD) Turbine designer (TD)

Integrate foundation modelin aero-elastic model

Run aero-elastic simulation(and adjust tower design)

(Adjust) design foundation1 2

3

Offshore wind turbine support structure is custom engineered for every wind farm

Introduction Approach Modeling Results Conclusions and recommendations

6

Foundation designer (FD) Turbine designer (TD)

Integrate foundation modelin aero-elastic model

Run aero-elastic simulation(and adjust tower design)

Extract interface loads/displacements between

tower and foundation

(Adjust) design foundation1 2

4

3

Offshore wind turbine support structure is custom engineered for every wind farm

Introduction Approach Modeling Results Conclusions and recommendations

6

Foundation designer (FD) Turbine designer (TD)

Integrate foundation modelin aero-elastic model

Run aero-elastic simulation(and adjust tower design)

Extract interface loads/displacements between

tower and foundation

(Adjust) design foundation

Apply interface loads/displacements on

detailed foundation model

5

1 2

3

4

Offshore wind turbine support structure is custom engineered for every wind farm

Introduction Approach Modeling Results Conclusions and recommendations

6

Foundation designer (FD) Turbine designer (TD)

Integrate foundation modelin aero-elastic model

Run aero-elastic simulation(and adjust tower design)

Extract interface loads/displacements between

tower and foundation

(Adjust) design foundation

Run simulation

Apply interface loads/displacements on

detailed foundation model

6

5

1 2

3

4

Offshore wind turbine support structure is custom engineered for every wind farm

Introduction Approach Modeling Results Conclusions and recommendations

6

Foundation designer (FD) Turbine designer (TD)

Integrate foundation modelin aero-elastic model

Run aero-elastic simulation(and adjust tower design)

Extract interface loads/displacements between

tower and foundation

(Adjust) design foundation

Run simulation

Apply interface loads/displacements on

detailed foundation model

6

5

1 2

3

4

Calculation post-processing analyses

Introduction Approach Modeling Results Conclusions and recommendations

7

fwind

fwave

Dynamic analysis Force

controlled

g

fwind

fwav

e

fwave g

Calculation post-processing analyses

Introduction Approach Modeling Results Conclusions and recommendations

7

fwind

fwave

Dynamic analysis Force

controlled

g

fwind

fwav

e

Dynamic or Quasi-static fwave g

Calculation post-processing analyses

Introduction Approach Modeling Results Conclusions and recommendations

7

fwind

fwave

Force controlled

g

fwind

fwind

fwav

e

fwav

e

ub

Dynamic or Quasi-static

Displacement controlled

Dynamic analysis

fwave g

fwaveub

Calculation post-processing analyses

Introduction Approach Modeling Results Conclusions and recommendations

7

fwind

fwave

Force controlled

g

fwind

fwind

fwav

e

fwav

e

ub

Dynamic or Quasi-static

Displacement controlledDynamic or Quasi-

static

Dynamic analysis

fwave g

fwaveub

Calculation post-processing analyses

Introduction Approach Modeling Results Conclusions and recommendations

7

fwind

fwave

Force controlled

g

fwind

fwind

fwav

e

fwav

e

ub

Dynamic or Quasi-static

Displacement controlledDynamic or Quasi-

static

Dynamic analysis

fwave g

fwaveub

Dynamic versus quasi-static analysis

Introduction Approach Modeling Results Conclusions and recommendations

Dynamic analysis

Quasi-static analysis

Only accurate if structure is excited below first eigenfrequency

8

Dynamic versus quasi-static analysis

Introduction Approach Modeling Results Conclusions and recommendations

Dynamic analysis

Quasi-static analysis

Only accurate if structure is excited below first eigenfrequency

8

Dynamic versus quasi-static analysis

Introduction Approach Modeling Results Conclusions and recommendations

Dynamic analysis

Quasi-static analysis

Only accurate if structure is excited below first eigenfrequency

8

Frequency

Am

plit

ude

≈1K__

≈1

ω2M__

Dynamic versus quasi-static analysis

Introduction Approach Modeling Results Conclusions and recommendations

Dynamic analysis

Quasi-static analysis

Only accurate if structure is excited below first eigenfrequency

8

Frequency

Am

plit

ude

≈1K__

≈1

ω2M__

Dynamic versus quasi-static analysis

Introduction Approach Modeling Results Conclusions and recommendations

Dynamic analysis

Quasi-static analysis

Only accurate if structure is excited below first eigenfrequency

8

Frequency

Am

plit

ude

≈1K__

≈1

ω2M__

Design cycle for offshore wind turbine support structure

Introduction Approach Modeling Results Conclusions and recommendations

9

Foundation designer (FD) Turbine designer (TD)

Integrate foundation modelin aero-elastic model

Run aero-elastic simulation(and adjust tower design)

Extract interface loads/displacements between

tower and foundation

(Adjust) design foundation

Run simulation

Apply interface loads/displacements on

detailed foundation model

6

5

1 2

3

4

Design cycle for offshore wind turbine support structure

Introduction Approach Modeling Results Conclusions and recommendations

9

Foundation designer (FD) Turbine designer (TD)

Integrate foundation modelin aero-elastic model

Run aero-elastic simulation(and adjust tower design)

Extract interface loads/displacements between

tower and foundation

(Adjust) design foundation

Run simulation

Apply interface loads/displacements on

detailed foundation model

6

5

1 2

3

4

Reduction of foundation to lower computation costs

Reduce large number of DoF into smaller set of generalized DoF Size(ũ) << size(u) Lower computation costs Approximation of exact

solution Reduction basis contains

limited number of deformation shapes

Only accurate if Spectral convergence Spatial convergence

Introduction Approach Modeling Results Conclusions and recommendations

10

Reduction of foundation to lower computation costs

Reduce large number of DoF into smaller set of generalized DoF Size(ũ) << size(u) Lower computation costs Approximation of exact

solution Reduction basis contains

limited number of deformation shapes

Only accurate if Spectral convergence Spatial convergence

Introduction Approach Modeling Results Conclusions and recommendations

10

= ++

Reduction of foundation to lower computation costs

Reduce large number of DoF into smaller set of generalized DoF Size(ũ) << size(u) Lower computation costs Approximation of exact

solution Reduction basis contains

limited number of deformation shapes

Only accurate if Spectral convergence Spatial convergence

Introduction Approach Modeling Results Conclusions and recommendations

10

= ++

Reduction methods

Guyan reduction

Introduction Approach Modeling Results Conclusions and recommendations

11

+ …. +

Static constraint modes

Reduction methodsCraig-Bampton reduction

Introduction Approach Modeling Results Conclusions and recommendations

12

Static constraint modes

+ + +

Fixed interface vibration modes

Reduction methods

Augmented Craig-Bampton reduction

Introduction Approach Modeling Results Conclusions and recommendations

13

+ + +

Fixed interface vibration modes

Modal Truncation vectors

Static constraint modes

Impact on fatigue damage results

Offshore wind turbine exposed to cyclic loading

Fatigue is one of the main design drivers

Impact of error in the reponse on the accuracy of the fatigue damage results

Introduction Approach Modeling Results Conclusions and recommendations

14

Content Introduction Approach Modeling Results Conclusions and recommendations

15

Monopile versus Jacket

Introduction Approach Modeling Results Conclusions and recommendations

16

Eigenfrequency

OWT model[Hz]

Foundationωfree [Hz]

Foundation ωfixed [Hz]

1st 0.30 6.73 42.8

Eigenfrequency

OWT model[Hz]

Foundationωfree [Hz]

Foundation ωfixed [Hz]

1st 0.27 1.06 4.09

Wind, wave and operational loads

Wind loads Random load, wide

frequency spectrum Excite frequencies up to 7

Hz

Introduction Approach Modeling Results Conclusions and recommendations

17

ω

Energy

Wind, wave and operational loads

Wind loads Random load, wide

frequency spectrum Excite frequencies up to 7

Hz Wave loads

Wave frequencies are generally lower

Excite frequencies up to 0.5 Hz

Introduction Approach Modeling Results Conclusions and recommendations

17

ω

ω

Energy

Energy

Wind, wave and operational loads

Wind loads Random load, wide

frequency spectrum Excite frequencies up to 7

Hz Wave loads

Wave frequencies are generally lower

Excite frequencies up to 0.5 Hz

Operational loads Rotation frequency of the

rotor (1P) Blade passing frequency

(3P)

Introduction Approach Modeling Results Conclusions and recommendations

17

ω

Energy

Energy

ω

ω

Energy

Content Introduction Approach Modeling Results Conclusions and recommendations

18

Introduction Approach Modeling Results Conclusions and recommendations

19

fwind

fwav

e

Dynamic analysis

fwav

e

Quasi-static post-processing analyses

Quasi-static post-processing analyses

Introduction Approach Modeling Results Conclusions and recommendations

19

fwind

fwav

e

fwave g

Quasi-staticForce controlledDynamic

analysis

g

fwind

fwav

e

fwav

e

Quasi-static post-processing analyses

Introduction Approach Modeling Results Conclusions and recommendations

19

fwind

fwav

e

fwave g

Quasi-static

fwaveub

Force controlled

Quasi-static

Dynamic analysis

g

fwind

fwind

fwav

e

fwav

e

ub

Displacement controlled

fwav

e

Accuracy of quasi-static post-processing

Introduction Approach Modeling Results Conclusions and recommendations

20

ωfree

Elastic energy in the foundation structure

En

erg

y [ J

]

Excitation frequency [Hz]

En

erg

y [ J

]

Accuracy of quasi-static post-processing

Introduction Approach Modeling Results Conclusions and recommendations

20

ωfree

ωfree ωfixed

Elastic energy in the foundation structure

En

erg

y [ J

]E

nerg

y [ J

]

Excitation frequency [Hz]

Expansion of reduced response

Response detailed foundation model obtained by expanding the reduced response of the foundation

Only accurate if model converges spectrally and spatially

Introduction Approach Modeling Results Conclusions and recommendations

21

fwav

e

fwav

e

~

fwin

d

fwav

e

~Expansion

Dynamic analysis

Spectral convergence

Introduction Approach Modeling Results Conclusions and recommendations

22

Relative difference eigenfrequencies of reduced OWT model

Frequency [Hz]

Spectral convergence

Introduction Approach Modeling Results Conclusions and recommendations

22

Relative difference eigenfrequencies of reduced OWT model

Frequency [Hz]

Introduction Approach Modeling Results Conclusions and recommendations

23

Relative energy difference of expanded response

Expansion of reduced response

Excitation frequency [Hz]

Introduction Approach Modeling Results Conclusions and recommendations

23

Residual correction

Expansion of reduced response

Excitation frequency [Hz]

Relative energy difference of expanded response

Expansion of reduced response

Introduction Approach Modeling Results Conclusions and recommendations

23

Residual correction

Excitation frequency [Hz]

Relative energy difference of expanded response

Post-processing analysis with reduced foundation in complete OWT model

Introduction Approach Modeling Results Conclusions and recommendations

24

fwave

fwave

fwind

fwave~~

Dynamic analysis

Post-processing analysis with reduced foundation in complete OWT model

Introduction Approach Modeling Results Conclusions and recommendations

24

fwave g

Force controlledDynamic and Quasi-

staticg

fwind

~fwave

fwave

fwave

fwind

fwave~~

Dynamic analysis

Post-processing analysis with reduced foundation in complete OWT model

Introduction Approach Modeling Results Conclusions and recommendations

24

fwave g

Force controlled

fwaveub

Dynamic and Quasi-static

Displacement controlledDynamic and Quasi-

static

g

fwind

fwind

ub

~fwave

~fwave

Dynamic analysis

fwave

fwave

fwind

fwave~~

Post-processing analysis with reduced foundation in complete OWT model

Introduction Approach Modeling Results Conclusions and recommendations

24

fwave g

fwaveub

g

fwind

fwind

ub

~fwave

~fwave

Force controlledDynamic and Quasi-

static

Displacement controlledDynamic and Quasi-

static

Guyan reduction Craig-Bampton reduction Augmented Craig-Bampton

reduction

fwave

fwave

fwind

fwave~~

Dynamic analysis

Post-processing analysis with reduced foundation in complete OWT model

Introduction Approach Modeling Results Conclusions and recommendations

24

Guyan reduction Craig-Bampton reduction Augmented Craig-Bampton

reduction

fwave g

fwaveub

g

fwind

fwind

ub

~fwave

~fwave

Force controlledDynamic and Quasi-

static

Displacement controlledDynamic and Quasi-

static

fwave

fwave

fwind

fwave~~

Dynamic analysis

Post-processing analysis with Craig-Bampton reduced foundation in OWT model

Introduction Approach Modeling Results Conclusions and recommendations

25

ωfree ωfixed

Relative energy difference with respect to exact solution

Excitation frequency [Hz]

Post-processing analysis with Craig-Bampton reduced foundation in OWT model

Introduction Approach Modeling Results Conclusions and recommendations

25

ωfree ωfixed

Quasi-static post-processing inaccurate ωfree and ωfixed within excitation spectrum

Dynamic post-processing accurate CB reduced model spectrally converged Internal dynamics included

Relative energy difference with respect to exact solution

Excitation frequency [Hz]

Post-processing analysis with Craig-Bampton reduced foundation in OWT model

Introduction Approach Modeling Results Conclusions and recommendations

25

ωfree ωfixed

Quasi-static post-processing inaccurate ωfree and ωfixed within excitation spectrum

Dynamic post-processing accurate CB reduced model converges spectrally Internal dynamics included

Relative energy difference with respect to exact solution

Excitation frequency [Hz]

Fatigue damage - Jacket

Introduction Approach Modeling Results Conclusions and recommendations

26

Relative damage difference with respect to exact damage

Expansion

Quasi-static Force controlled

Dynamic Force controlled

Quasi-static Displacement controlled

Dynamic Displacement controlled

Fatigue damage - Jacket

Introduction Approach Modeling Results Conclusions and recommendations

26

Relative damage difference with respect to exact damage

Content Introduction Approach Modeling Results Conclusions and recommendations

27

Conclusions Following aspects tend to influence the accuracy of

the calculation procedures: The characteristics of the structure

First fixed and free interface eigenfrequency Qs FC significantly underestimates fatigue damage for jacket

Use of a reduced foundation model in complete OWT model Spectral and spatial convergence Residual correction improves accuracy fatigue damage results

Post-processing method Dynamic post-processing provides accurate fatigue damage

results despite errors in interface loads/displacements

Introduction Approach Modeling Results Conclusions and recommendations

28

Conclusions Following aspects tend to influence the accuracy of

the calculation procedures: The characteristics of the structure

First fixed and free interface eigenfrequency Qs FC significantly underestimates fatigue damage for jacket

Use of a reduced foundation model in complete OWT model Spectral and spatial convergence Residual correction improves accuracy fatigue damage results

Post-processing method Dynamic post-processing provides accurate fatigue damage

results despite errors in interface loads/displacements

Introduction Approach Modeling Results Conclusions and recommendations

28

Conclusions Following aspects tend to influence the accuracy of

the calculation procedures: The characteristics of the structure

First fixed and free interface eigenfrequency Qs FC significantly underestimates fatigue damage for jacket

Use of a reduced foundation model in complete OWT model Spectral and spatial convergence Residual correction improves accuracy fatigue damage results

Post-processing method Dynamic post-processing provides accurate fatigue damage

results despite errors in interface loads/displacements

Introduction Approach Modeling Results Conclusions and recommendations

28

Conclusions Following aspects tend to influence the accuracy of

the calculation procedures: The characteristics of the structure

First fixed and free interface eigenfrequency Qs FC significantly underestimates fatigue damage for jacket

Use of a reduced foundation model in complete OWT model Spectral and spatial convergence Residual correction improves accuracy fatigue damage results

Post-processing method Dynamic post-processing provides accurate fatigue damage

results despite use of reduced foundation

Introduction Approach Modeling Results Conclusions and recommendations

28

Recommendations Apply the different calculation procedures in

BHawC with different load cases

Introduction Approach Modeling Results Conclusions and recommendations

29

Recommendations Apply the different calculation procedures in

BHawC with different load cases Set up clear guidelines for spatial

convergence Error estimation methods

Introduction Approach Modeling Results Conclusions and recommendations

29

Recommendations Apply the different calculation procedures in

BHawC with different load cases Set up clear guidelines for spatial

convergence Error estimation methods

Determine an efficient and accurate calculation procedure for more complex models

Introduction Approach Modeling Results Conclusions and recommendations

29

Recommendations Apply the different calculation procedures in

BHawC with different load cases Set up clear guidelines for spatial

convergence Error estimation methods

Determine an efficient and accurate calculation procedure for more complex models

Validate results with real OWTs and loads

Introduction Approach Modeling Results Conclusions and recommendations

29

Thank you for your attention

Levelized Cost of Electricity

Onshore LCoE

Offshore LCoE

Cost of electricityEEX Leipzig

100

10

1

4

1980 1990 2000 2010 2020 2030

€/kWh

Fatigue damage computation

Response Stresses SN-curveFatigue damage

Force versus displacement controlled

Force controlled approach

Displacement controlled approach

fwave g

fwaveub

Relative energy difference quasi-static analysis

Interface loads - Monopile

Interface loads - Jacket

Guyan reduced jacket in complete OWT model

Augmented Craig-Bampton reduction1. External load represented by a spatial and temporal part

2. Quasi-static response and orthogonalize w.r.t. fixed interface vibration modes

3. Orthonormalize w.r.t. each other

4. Construct reduction basis

Augmented Craig-Bampton reduced jacket in complete OWT model

Facts wind energy

Wind turbine Household

Power capacity 3 MW

Energy production 6 – 7,5 GWh per year Serves ± 2000

households

Average household 2,2 persons

Energy usage 3500 kWh per year

Requirement for calculation procedures

Detailed foundation in OWT model

Reduced foundation in OWT model

Expansion If spectrally and spatially converged

Force controlled

Dynamic If spectrally and spatially converged

Quasi-static If ωfree >> max(ωext) If spectrally and spatially converged If ωfree >> max(ωext)

Displacement controlled

Dynamic If spectrally and spatially converged

Quasi-static If ωfree >> max(ωext) If spectrally and spatially converged If ωfree >> max(ωext)

✔

✗

✔

✔✔✔

✔✔

✔

✔