Hydro-Elastic Contributions to Fatigue Damage on a Large Monopile

Transcription

1 Hydro-Elastic Contributions to Fatigue Damage on a Large Monopile Jan-Tore H. Horn a,b, Jørgen R. Krokstad b,c and Jørgen Amdahl a,b a Centre for Autonomous Marine Operations and Systems - AMOS b Department of Marine Technology, NTNU c Statkraft January 21, EERA DeepWind 216

2 Motivation EERA DeepWind 216 2

3 Motivation EERA DeepWind 216 2

4 Motivation EERA DeepWind 216 2

5 Overview Introduction Structural model Hydrodynamic models Challenges Results Evaluation by time-frequency analysis Conclusions EERA DeepWind 216 3

6 Hydrodynamic Modeling Motivation: Need accurate and realistic load models to evaluate control strategies Upscaling of monopiles will give new response characteristics Load theory validation and relative impact on lifetime estimation EERA DeepWind 216 4

7 Hydrodynamic Modeling Motivation: Need accurate and realistic load models to evaluate control strategies Upscaling of monopiles will give new response characteristics Load theory validation and relative impact on lifetime estimation EERA DeepWind 216 4

8 Hydrodynamic Modeling Motivation: Need accurate and realistic load models to evaluate control strategies Upscaling of monopiles will give new response characteristics Load theory validation and relative impact on lifetime estimation EERA DeepWind 216 4

9 Hydrodynamic Modeling Motivation: Need accurate and realistic load models to evaluate control strategies Upscaling of monopiles will give new response characteristics Load theory validation and relative impact on lifetime estimation Findings: Significant higher order contributions to fatigue damage in sea-states with H S > D/2 Necessary to include diffraction effects on second order inertia forces Higher order loads more prominent at low damping levels EERA DeepWind 216 4

10 Hydrodynamic Modeling Motivation: Need accurate and realistic load models to evaluate control strategies Upscaling of monopiles will give new response characteristics Load theory validation and relative impact on lifetime estimation Findings: Significant higher order contributions to fatigue damage in sea-states with H S > D/2 Necessary to include diffraction effects on second order inertia forces Higher order loads more prominent at low damping levels EERA DeepWind 216 4

11 Hydrodynamic Modeling Motivation: Need accurate and realistic load models to evaluate control strategies Upscaling of monopiles will give new response characteristics Load theory validation and relative impact on lifetime estimation Findings: Significant higher order contributions to fatigue damage in sea-states with H S > D/2 Necessary to include diffraction effects on second order inertia forces Higher order loads more prominent at low damping levels EERA DeepWind 216 4

12 Model Hub height = 115m Mean depth = 3m Soil = 42m Damage calculation EERA DeepWind 216 5

13 Model Hub height = 115m Mean depth = 3m Soil = 42m Damage calculation EERA DeepWind 216 5

14 Model Parameters: Diameter 9m Depth 3m Structural damping 3% of critical damping (Rayleigh) Aerodynamic damping Constant Rayleigh included in structural Soil Non-linear springs for sand and clay Natural periods Mode 1: 4.2s, Mode 2: 1.s Sea-states FLS EERA DeepWind 216 6

15 Modal analysis Mode 1: 4.1 seconds Mode 2: 1. seconds Mode 1 Mode EERA DeepWind 216 7

16 Wave load models O1 O1D O2 O3 FNV3 O1P O2P Linear waves Linear waves with diffraction (MacCamy and Fuchs) Second order contribution from kinematics stretching Third+fourth order contribution from kinematics stretching Third order FNV - direct implementation First order distributed pressure from panel code Second order total force from panel code EERA DeepWind 216 8

17 Diffraction - MacCamy and Fuchs Correction of wave load due to interaction with large-volume structure. a eq = equivalent water particle acceleration. aeq/a Exact Fitted function πd/λ EERA DeepWind 216 9

18 Diffraction - MacCamy and Fuchs Correction of wave load due to interaction with large-volume structure. a eq = equivalent water particle acceleration. Total wave load [N] Morison MacCamy & Fuchs correction Time [s] EERA DeepWind 216 9

19 Second order wave elevation Elevation [m] First order Second order sum-frequency Time [s] EERA DeepWind 216 1

20 Second order sum-frequency from panel code HydroD - Wadam EERA DeepWind

21 Second order sum-frequency from panel code Non-dimensional resulting pressure in x-direction over the column as a function of ω 1, ω 2 and z. EERA DeepWind

22 Second order sum-frequency from panel code The second-order pressures are lumped to z = and act as a point force. { N } F = R M n=1 m=1 ζ a,n ζ a,m QTF e i(φm+φn) EERA DeepWind

23 Load application Distrubuted, point force or moment? F e 2kz M z/h = (a) kd=.4 (a) kd=1.1 (a) kd=3.6 (b) (c) z/h = 1 z/h [-] (a) (b) (c) M/M max [-] EERA DeepWind

24 Load application Distrubuted, point force or moment? F e 2kz M z/h = (a) kd=.4 (a) kd=1.1 (a) kd=3.6 (b) (c) z/h = 1 z/h [-] (a) (b) (c) M/M max [-] EERA DeepWind

25 Third Order FNV Third order horizontal force from linear elevation and diffraction potential assuming deep water: [ ( FNV (3) Fx = ρπr 2 ζ 1 ζ 1 u tz + 2ww x + uu x 2 ) g u tw t ( ) ] ut (u 2 + w 2 ) + βg g u2 u t z= EERA DeepWind

26 Kinematics models Assuming kζ a = O(ɛ) kd = O(δ) where ɛ 1 and ɛ δ EERA DeepWind

27 Kinematics models Assuming kζ a = O(ɛ) kd = O(δ) Order of horizontal inertia forces: A B A: ɛ 1 δ 2 z F G H C D E Φ 1 Φ 2 ζ 1 ζ 1 + ζ 2 x EERA DeepWind

28 Kinematics models Assuming kζ a = O(ɛ) kd = O(δ) Order of horizontal inertia forces: A: ɛ 1 δ 2 B+C: ɛ 2 δ 2 z F G H C D E A B Φ 1 Φ 2 ζ 1 ζ 1 + ζ 2 x EERA DeepWind

29 Kinematics models Assuming kζ a = O(ɛ) kd = O(δ) Order of horizontal inertia forces: A: ɛ 1 δ 2 B+C: ɛ 2 δ 2 D+E+F: ɛ 3 δ 2 z F G H C D E A B Φ 1 Φ 2 ζ 1 ζ 1 + ζ 2 x EERA DeepWind

30 Kinematics models Assuming kζ a = O(ɛ) kd = O(δ) Order of horizontal inertia forces: A: ɛ 1 δ 2 B+C: ɛ 2 δ 2 D+E+F: ɛ 3 δ 2 G+H: ɛ 4 δ 2 z F G H C D E A B Φ 1 Φ 2 ζ 1 ζ 1 + ζ 2 x EERA DeepWind

31 Kinematics models A B C D E F G H F x /(.5πρD 2 ) O(F x ) F x Frequency h u 1,t(z)dz ɛδ 2 ζ a 1ω h u 2,t(z)dz ɛ 2 δ 2 ζ 2 a 2ω max(,ζ1 ) u 1,t ()dz ɛ 2 δ 2 ζ 2 a 2ω max(,ζ1 ) zu 1,tz dz ɛ 3 δ 2 ζ 3 a 1ω + 3ω max(,ζ1 ) u 2,t ()dz ɛ 3 δ 2 ζ 3 a 1ω + 3ω max(,ζ1 +ζ 2 ) max(,ζ 1 ) u 1,t ()dz ɛ 3 δ 2 ζ 3 a 1ω + 3ω max(,ζ1 +ζ 2 ) max(,ζ 1 ) zu 1,tz ()dz ɛ 4 δ 2 ζa 4 4ω max(,ζ1 +ζ 2 ) max(,ζ 1 ) u 2,t ()dz ɛ 4 δ 2 ζa 4 4ω EERA DeepWind

32 Third order forces Total F x FNV3 [N] F x (D) [N] FNV3 First term in F [N] -1.5 x ζ [m] Time [s] FNV (3) Fx = ρπr [ζ u tz + 2ζ 1 ww x + ζ 1 uu x 2 g ζ 1u t w t ( ) ] ut (u 2 + w 2 ) + βg g u2 u t z= EERA DeepWind

33 Third order forces Total F x FNV3 [N] F x (D) [N] FNV3 First term in F [N] -1.5 x ζ [m] Time [s] [ FNV (3) Fx = ρπr 2 ζ1 2 u tz + 2ζ 1 ww x + ζ 1 uu x 2 g ζ 1u t w t ( ) ] ut (u 2 + w 2 ) + βg g u2 u t z= EERA DeepWind

34 Kinematics models Notation Fields Description O1 A First order incident wave potential O1D A First order incident wave potential w/diffraction O2 B+C Second order incident wave potential and stretched first order potential O3 D+E+F+G+H Third and fourth order force from stretched first and second order potential O1P A First order diffraction pressure from panel code modeled as acceleration O2P B+C Total second order diffraction force from panel code FNV3 N/A Third order FNV ringing force based on first order incident potential EERA DeepWind

35 Wave kinematics grid Logarithmically distributed in z-direction to increase accuracy in wave-zone 4.3 million points for 3 minute simulation with dt =.1 and N Z = 4, large files! 1 z [m] -1 Gridpoint -2 E[ζ m ] -E[ζ m ] Mean surface N Z EERA DeepWind

36 Sea-states Chosen sea-states for Dogger Bank conditions. JONSWAP spectrum with peak parameter 3.3 is used KC max=3 Extra H S [m] KC max=2 KC max=1 KC max= T P [s] Figure: Sea-states with finite depth KC number for h=3m. EERA DeepWind 216 2

37 Sea-states Chosen sea-states for Dogger Bank conditions. JONSWAP spectrum with peak parameter 3.3 is used. No. H S [m] T P [s] f HS,T P [-] KC max [-] πd/λ [-] * EERA DeepWind 216 2

38 Results For each sea-state and hydrodynamic model, 3 3 minute simulations have been run without wind Average findings presented Small variances between the seeds EERA DeepWind

39 Results Baseline maximum moment D RFC D /D M/M RFC,O1 /D f RFC,Cd= RFC,LD RFC,O1 /D /D HS,T P O1D O1D+O2 O1D+O2+O3 O1D+O2+FNV3 O1P O1P+O2P O1D O1P+O2P+FNV3 O1D+O2 O1D+O2+O3 O1D+O2+FNV3 O1P O1P+O2P O1P+O2P+FNV3 O1D O1D+O2 O1D+O2+O3 O1D+O2+O3 O1D+O2+FNV3 O1D+O2+FNV3 O1P O1P+O2P O1P+O2P+FNV3 O1P+O2P+FNV Sea-state EERA DeepWind

40 Results Fatigue damage relative to first order incident wave D RFC D /D M/M RFC,O1 /D f RFC,Cd= RFC,LD RFC,O1 /D /D HS,T P O1D O1D+O2 O1D+O2+O3 O1D+O2+FNV3 O1P O1P+O2P O1D O1P+O2P+FNV3 O1D+O2 O1D+O2+O3 O1D+O2+FNV3 O1P O1P+O2P O1P+O2P+FNV3 O1D O1D+O2 O1D+O2+O3 O1D+O2+O3 O1D+O2+FNV3 O1D+O2+FNV3 O1P O1P+O2P O1P+O2P+FNV3 O1P+O2P+FNV Sea-state EERA DeepWind

41 Results Relative fatigue damage accounting for probability of occurrence D RFC D /D M/M RFC,O1 /D f RFC,Cd= RFC,LD RFC,O1 /D /D HS,T P O1D O1D+O2 O1D+O2+O3 O1D+O2+FNV3 O1P O1P+O2P O1D O1P+O2P+FNV3 O1D+O2 O1D+O2+O3 O1D+O2+FNV3 O1P O1P+O2P O1P+O2P+FNV3 O1D O1D+O2 O1D+O2+O3 O1D+O2+O3 O1D+O2+FNV3 O1D+O2+FNV3 O1P O1P+O2P O1P+O2P+FNV3 O1P+O2P+FNV Sea-state EERA DeepWind

42 Results Relative fatigue damage accounting for probability of occurrence D RFC D /D M/M RFC,O1 /D f RFC,Cd= RFC,LD RFC,O1 /D /D HS,T P O1D O1D+O2 O1D+O2+O3 O1D+O2+FNV3 O1P O1P+O2P O1D O1P+O2P+FNV3 O1D+O2 O1D+O2+O3 O1D+O2+FNV3 O1P O1P+O2P O1P+O2P+FNV3 O1D O1D+O2 O1D+O2+O3 O1D+O2+O3 O1D+O2+FNV3 O1D+O2+FNV3 O1P O1P+O2P O1P+O2P+FNV3 O1P+O2P+FNV Sea-state EERA DeepWind

43 Results Increased fatigue damage for lightly damped system: 3% 1% D RFC D /D M/M RFC,O1 /D f RFC,Cd= RFC,LD RFC,O1 /D /D HS,T P O1D O1D+O2 O1D+O2+O3 O1D+O2+FNV3 O1P O1P+O2P O1D O1P+O2P+FNV3 O1D+O2 O1D+O2+O3 O1D+O2+FNV3 O1P O1P+O2P O1P+O2P+FNV3 O1D O1D+O2 O1D+O2+O3 O1D+O2+O3 O1D+O2+FNV3 O1D+O2+FNV3 O1P O1P+O2P O1P+O2P+FNV3 O1P+O2P+FNV Sea-state EERA DeepWind

44 Results Drag force contribution due to wave elevation and increasing KC-number D RFC D /D M/M RFC,O1 /D f RFC,Cd= RFC,LD RFC,O1 /D /D HS,T P O1D O1D+O2 O1D+O2+O3 O1D+O2+FNV3 O1P O1P+O2P O1D O1P+O2P+FNV3 O1D+O2 O1D+O2+O3 O1D+O2+FNV3 O1P O1P+O2P O1P+O2P+FNV3 O1D O1D+O2 O1D+O2+O3 O1D+O2+O3 O1D+O2+FNV3 O1D+O2+FNV3 O1P O1P+O2P O1P+O2P+FNV3 O1P+O2P+FNV Sea-state EERA DeepWind

45 Time-frequency analysis Wavelet analysis revealing most dominating oscillation periods. O1D, Sea-state /2 1/4 1/8 1/16 1/32 1/64 Period [s] T(IMF1) 8 T(IMF2) O1D+O2+FNV3, Sea-state 5 12 Time [s] /2 1/4 1/8 1/16 1/32 1/64 Period [s] T(IMF1) 8 T(IMF2) Time [s] EERA DeepWind

46 Time-frequency analysis Wavelet analysis revealing most dominating oscillation periods. O1D, Sea-state /2 1/4 1/8 1/16 1/32 1/64 Period [s] T(IMF1) 8 T(IMF2) O1D+O2+FNV3, Sea-state 5 12 Time [s] /2 1/4 1/8 1/16 1/32 1/64 Period [s] T(IMF1) 8 T(IMF2) Time [s] EERA DeepWind

47 Time-frequency analysis Wavelet analysis revealing most dominating oscillation periods. O1D, Sea-state /2 1/4 1/8 1/16 1/32 1/64 Period [s] T(IMF1) 8 T(IMF2) O1D+O2+FNV3, Sea-state 5 12 Time [s] /2 1/4 1/8 1/16 1/32 1/64 Period [s] T(IMF1) 8 T(IMF2) Time [s] EERA DeepWind

48 Conclusions When H S > D/2, significant contributions to fatigue damage from higher order loads are observed Higher order effects not important for smaller sea-states - overall small contributions when frequency of occurrence is accounted for Lower damping level results in more prominent contributions from higher order forces Drag forces still important when wave elevation is accounted for - need sensitivity study of C D A Morison type loading for second order load seems to be predicting very large responses and fatigue damage for large sea-states - elevation important Important to include diffraction effects - both first and second order EERA DeepWind

49 Conclusions When H S > D/2, significant contributions to fatigue damage from higher order loads are observed Higher order effects not important for smaller sea-states - overall small contributions when frequency of occurrence is accounted for Lower damping level results in more prominent contributions from higher order forces Drag forces still important when wave elevation is accounted for - need sensitivity study of C D A Morison type loading for second order load seems to be predicting very large responses and fatigue damage for large sea-states - elevation important Important to include diffraction effects - both first and second order EERA DeepWind

50 Conclusions When H S > D/2, significant contributions to fatigue damage from higher order loads are observed Higher order effects not important for smaller sea-states - overall small contributions when frequency of occurrence is accounted for Lower damping level results in more prominent contributions from higher order forces Drag forces still important when wave elevation is accounted for - need sensitivity study of C D A Morison type loading for second order load seems to be predicting very large responses and fatigue damage for large sea-states - elevation important Important to include diffraction effects - both first and second order EERA DeepWind

51 Conclusions When H S > D/2, significant contributions to fatigue damage from higher order loads are observed Higher order effects not important for smaller sea-states - overall small contributions when frequency of occurrence is accounted for Lower damping level results in more prominent contributions from higher order forces Drag forces still important when wave elevation is accounted for - need sensitivity study of C D A Morison type loading for second order load seems to be predicting very large responses and fatigue damage for large sea-states - elevation important Important to include diffraction effects - both first and second order EERA DeepWind

52 Conclusions When H S > D/2, significant contributions to fatigue damage from higher order loads are observed Higher order effects not important for smaller sea-states - overall small contributions when frequency of occurrence is accounted for Lower damping level results in more prominent contributions from higher order forces Drag forces still important when wave elevation is accounted for - need sensitivity study of C D A Morison type loading for second order load seems to be predicting very large responses and fatigue damage for large sea-states - elevation important Important to include diffraction effects - both first and second order EERA DeepWind

53 Conclusions When H S > D/2, significant contributions to fatigue damage from higher order loads are observed Higher order effects not important for smaller sea-states - overall small contributions when frequency of occurrence is accounted for Lower damping level results in more prominent contributions from higher order forces Drag forces still important when wave elevation is accounted for - need sensitivity study of C D A Morison type loading for second order load seems to be predicting very large responses and fatigue damage for large sea-states - elevation important Important to include diffraction effects - both first and second order EERA DeepWind

54 Thank you for your attention. - Questions? EERA DeepWind