Beruflich Dokumente
Kultur Dokumente
support structure modelling on simulation results for a 5-MW turbine on a tripod substructure
0.7
0.6
0.5
Difference [%]
0.4
0.3
0.2
Figure 4: Tripod (left) with load output position (dotted circle) and
0.1 shell central joint (right).
0
0 10 20 30 40 50 The model described herein has a number of DOF of approximately
Number beam elements [-] nshell = 50, 000
Figure 3: Natural frequency of mud brace for different discretiza- Tripod modeled with beam- and superelements
tions.
The superelement model is based on the shell model where all
the joints -the shell substructures- are replaced by superelements
As expected, the differences decrease with increasing number of el- using a Guyan reduction procedure. The rest of the beam structure is
ements. However, even for the simplest model the difference to the unchanged compared to the other models.
most detailed model - which is assumed to be "correct" - is clearly un- The number of DOF of this structure is approximately nsuper = 400,
der 1 %. As the computational effort for the time domain simulations a value even significantly below its counterpart for the basic beam
should be kept in mind, such small errors are accepted and the model model. The reason for this is as follows: The DOF of each superele-
is not refined. The number of DOF of the resulting beam model is ment are only the DOF of its master nodes. As mentioned before,
approximately nbeam = 600. several supplementary members have to be included in the basic beam
To conclude, this model is as realistic as achievable with standard model to allow for more realistic load simulation ("overlapping mem-
beam elements and reasonable effort. Therefore, it is suitable to in- bers") at the joints. Furthermore, changing properties of the tubes
vestigate the differences between beam models and more sophisticated such as diameters Di or thicknesses ti are modeled with several beam
structural models. elements and at the intersection points of tubes a node has to be de-
fined which leads to more elements as well. The parts of the struc-
ture modeled with those supplementary beam elements become part
Tripod modeled with shell- and beam elements of the superelements and therefore those DOF are not in the superele-
ment model. Figure 5 shows the beam model of a pile sleeve joint
This model is based on the beam model, however the beam ele- with fourteen nodes (black dots) and a resulting number of DOF of
ments representing all joints are replaced by shell joints. A standard nbeam,1 = 84 (left) and the corresponding superelement with six mas-
four-noded shell element, well suited for the modeling of warped ter nodes (black dots) and therefore nsuper,1 = 36 DOF (right).
shell structures with moderate thickness and six DOF per node is
used.
Furthermore so-called "master nodes" are defined to connect the de-
tailed sub-models to the residual structure. Being located in the center-
lines of the tubular chord and brace members those nodes are rigidly
connected to the outgoing beam element and rigidly connected to the
detailed model via radially arranged rigid link or stiff beam connec-
tions. Figure 4 shows the tripod model with shell joints and the central
joint in detail.
In defining the shell structures, the appropriate size and shape of the el-
ements have to be guaranteed. This topic has not been investigated ex-
tensively, as very details of the joints are not subject of this study and
strong influences on the more global phenomena investigated herein
are not expected. However the static and dynamic plausibility of the
model is checked.
The stiff connections between the shell and beam members suppress
the ovalisation of the circular cross section of the joints in the proxim- Figure 5: Pile sleeve joint modeled with beams (left) and as superele-
ity of the connection. This mainly influences the joints if the distances ment (right). Nodes are shown as black dots in both models.
to the master nodes are too short and might be investigated using para-
metric studies. However, this is not expected to be of major impor-
tance for this study, as the influence of joint models on more global
phenomena is investigated and not the joints themselves at a very high
level of detail.
difference of a pair of natural frequencies is approximately 50 % from
the 3rd to the 6th natural frequency. Even the lowest natural frequen-
S IMULATION RESULTS cies, associated with the first global bending modes of the structure,
show a difference of approximately 6 %.
In the following exemplary results of simulations with the three The preliminary study leads to the following results:
models are presented. All relative values in comparisons are related
to the model expected to be the most accurate. For comparisons 1. The superelement model leads exactly to the same results as the
between shell models and other models, the values are related to most accurate shell model, therefore superelement models are
the results from the shell models and for comparisons between very suitable for modeling this type of structure.
superelement- and beam models all values are related to the results
from the superelement models respectively. 2. There are significant differences between the results obtained
with the superelement model and with the beam model, which
means that a significant increase of simulation accuracy is ex-
Preliminary study on support structure models pected using superelements compared to the state-of-the-art ap-
proach.
In an initial study, the dead weights and eigenstates of the beam-, the
shell- and the superelement support structure models are checked. Concerning point 1 it should be noticed, that the accuracy of the shell
A model of the rotor nacelle assembly (RNA) is not used for the model - with a number of DOF around two orders of magnitude higher
following reasons: For the mass comparison, the RNA is just not than in the other models - is generally not questioned for the simula-
relevant, as the mass of the RNA would be the same in all three tion of the investigated type of steel structure as long as the modeling
models. Also, the shell model of the support structure is part of the is realized carefully. Furthermore, there are significant differences,
investigation and it is currently not possible to use shell elements in especially concerning mass matrix description (cf. Vorpahl (2009)),
the aero-elastic code ADCoS-Offshore. Therefore, a model including but those differences have no influence on the relevant parameters for
a detailed RNA and a shell support structure is not available for a dynamic analysis of the degree of detail performed in aero-elastic
comparison of eigenstates. codes.
Table 1 shows the total steel masses of the structures and the mass In this section, support structure models are analyzed. The influence
difference between the beam and the shell model. of the significant differences between beam- and superelement support
structure models (cf. point 2) on the OWT models - including the RNA
Table 1: Dead weight of all three support structure models and differ- - is investigated below.
ence between shell- and beam model.
4
beam
3.5
shell
3 super
Frequency [Hz]
2.5
2
1.5
1
0.5
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Number eigenstate [-]
Figure 6: First 15 natural frequencies of beam-, shell- and superele- Figure 7: 10th eigenmode at 1.23 Hz (beam model, left) and at
ment support structure models. 1.08 Hz (superelement model, right).
2 1
beam
1.5 super 0.8
1P 3P 6P
1
0.6
0.5
0 0.4
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0.2
Number eigenstate [-]
0
25
0 0.2 0.4 0.6 0.8 1 1.2 1.4
Difference [%]
20 Frequency [Hz]
15
10 Figure 9: Frequency ranges suspect to significant dynamic excitations
5 (1p, 3p, 6p) due to turbine operation (boxes). Full system
0 natural frequencies for the beam model (dotted lines) and
the superelement model (continous lines).
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Number eigenstate [-]
Besides many details that are not of primary interest for this study, the
Figure 8: First 15 full system natural frequencies for both models following results are found: Firstly, the frequencies associated with
(top) and differences between those frequencies (bottom). the first global bending modes of the support structure are shifted by
approximately 5 % towards lower frequencies as mentioned before.
Those frequencies are among the most important design parameters
Compared to the results obtained with the support structure models for OWT support structures as it is common practice do design struc-
without RNA in the previous section, some changes are obvious in tures with a first natural frequency between the 1p and the 3p range
both figures. Mainly due to the influence of the RNA on global sup- (soft-stiff design) of the turbine. The allowable frequency gap is rel-
port structure frequencies and due to low blade bending frequencies, atively small, especially as a distance of about 10 % between natural
frequency and excitation range has usually to be provided for safety Figure 10 shows the fore-aft tower top displacement (displacement
reasons (not included in figure 9). For the NREL turbine, the more along the wind and wave direction) over the first 50 s for the beam-
detailed modeling would lead to higher theoretical safety margins, as and for the superelement model.
Series.Run = LC_1_2_1
the frequencies are shifted towards the dead center between the 1p and Output Channel = TTDspFA 15
the 3p range. 0.55
Secondly, the 3rd full system natural frequency is shifted into the upper Company Name = beam beam super
Displacement [m]
0.5 Name =
Simulator ######## 201209_2 ########
3p frequency range due to the more realistic support structure model- 0.45
1 Time TTDspFA #BEZUG! #BEZUG! Time TTDspFA #BEZUG!
ing. The difference between the 3rd natural frequency calculated with 2
0.4
3
(sec)
0.0000
(m) #BEZUG! #BEZUG! (sec)
0.3854 #BEZUG! #BEZUG! 0.0000
(m) #BEZUG!
0.4256 #BEZUG!
the beam model and the superelement model is about 7 % as men- 0.35
4 0.0500 0.3854 #BEZUG! #BEZUG! 0.0500 0.4256 #BEZUG!
tioned above. The excitations due to disturbed wind flow around the 0.3
5 0.1000 beam model
0.3855 #BEZUG! #BEZUG! 0.1000 0.4257 #BEZUG!
tower in the 3p range are considered to bring more energy into the 6
0.25 0.1500 superelement
0.3856 #BEZUG! #BEZUG! 0.1500model0.4259 #BEZUG!
7 0.2000 0.3859 #BEZUG! #BEZUG! 0.2000 0.4263 #BEZUG!
system than its harmonics and even more than the excitations in the 0.2
8 0.2500 0.3863 #BEZUG! #BEZUG! 0.2500 0.4268 #BEZUG!
1p range (cf. Fischer and Kuehn (2009)). Furthermore, the natural 0 9 0.3000
10 0.3871 #BEZUG! #BEZUG!
20 30 0.3000 0.4277 #BEZUG!
40 0.4289 50
frequency is found near the upper bound of the 3p range which is as- 10 0.3500 0.3882 #BEZUG! #BEZUG!
Time [s] 0.3500 #BEZUG!
11 0.4000 0.3897 #BEZUG! #BEZUG! 0.4000 0.4305 #BEZUG!
sociated with rated rotor speed. It is obvious, that the turbine operates 12 0.4500 0.3916 #BEZUG! #BEZUG! 0.4500 0.4325 #BEZUG!
much more often at rated rotor speed than at lower speeds, as this is Figure 10: Tower top displacement over time for baseline turbine with
13 0.5000 0.3939 #BEZUG! #BEZUG! 0.5000 0.4350 #BEZUG!
the operational speed at all wind speeds above rated wind speed. The 14 0.5500 0.3967 #BEZUG! #BEZUG! 0.5500 0.4378 #BEZUG!
beam support structure (dashed line) and superelement
15 0.6000 0.3998 #BEZUG! #BEZUG! 0.6000 0.4411 #BEZUG!
maximum rotor thrust occurring at rated wind speed makes a strong support structure (continuous line) respectively.
16 0.6500 0.4033 #BEZUG! #BEZUG! 0.6500 0.4447 #BEZUG!
dynamic excitation at this speed even more probable. 17 0.7000 0.4081 #BEZUG! #BEZUG! 0.7000 0.4495 #BEZUG!
Thirdly, the 9th natural frequency is shifted to the upper 6p range. This 18 0.7500 0.4125 #BEZUG! #BEZUG!
The significant difference of the two results is evident; the average
19 0.8000 0.4174 #BEZUG! #BEZUG!
0.7500
0.8000
0.4541 #BEZUG!
0.4591 #BEZUG!
is not too critical, other natural frequencies are found in the 6p range 20 0.8500 0.4227 #BEZUG! #BEZUG!
displacement difference is approximately 10 %. As expected, the su- 0.8500 0.4645 #BEZUG!
as well. 21 0.9000 0.4283 #BEZUG! #BEZUG!
perelement support structure is more compliant and therefore leads to
0.9000 0.4702 #BEZUG!
22 0.9500 0.4339 #BEZUG! #BEZUG! 0.9500 0.4760 #BEZUG!
All in all, the support structure modeling with superelements leads to larger displacements. The paths of the curves are similar, but not ex-
23 1.0000 0.4395 #BEZUG! #BEZUG! 1.0000 0.4818 #BEZUG!
several changes in comparison to the basic beam model concerning the actly the same. As this structural answer is a result of the external
24 1.0500 0.4448 #BEZUG! #BEZUG! 1.0500 0.4873 #BEZUG!
full system eigenstates. The number of order of the natural frequen- 25 1.1000 0.4499 #BEZUG! #BEZUG!
loads (that are the same in both cases) and the characteristics of the 1.1000 0.4926 #BEZUG!
26 1.1500 0.4546 #BEZUG! #BEZUG! 1.1500 0.4976 #BEZUG!
cies varies, as local modes are influenced differently. This leads to structure (that are not the same) this was expected as well. As it was
27 1.2000 0.4590 #BEZUG! #BEZUG! 1.2000 0.5022 #BEZUG!
coupled modes that do not occur with the basic model. Furthermore, shown in this paper before, the natural frequencies associated with the
28 1.2500 0.4630 #BEZUG! #BEZUG! 1.2500 0.5064 #BEZUG!
the relative portions of single modes in one coupled mode may vary. 29 1.3000 0.4667 #BEZUG! #BEZUG!
first tower bending modes, that are among the most important struc- 1.3000 0.5105 #BEZUG!
30 1.3500 0.4701 #BEZUG! #BEZUG! 1.3500 0.5143 #BEZUG!
As expected, the RNA - which is the same in both models - leads tural properties, differ about 5 %.
31 1.4000 0.4732 #BEZUG! #BEZUG! 1.4000 0.5178 #BEZUG!
to lower frequencies associated with global modes due to its signifi- The following figure (cf. figure 11) shows the bending moment around
32 1.4500 0.4760 #BEZUG! #BEZUG! 1.4500 0.5210 #BEZUG!
cant mass at the tower top. It leads to smaller differences between 33 1.5000 0.4785 #BEZUG! #BEZUG!
the local z-axis, in the upwind leg near the pile sleeve joint over a 1.5000 0.5239 #BEZUG!
the two models, as the frequencies associated with global modes and 34 1.5500 0.4806 #BEZUG! #BEZUG! 1.5500 0.5265 #BEZUG!
fraction of 50 s of the load case for both models. The load output
35 1.6000 0.4824 #BEZUG! #BEZUG! 1.6000 0.5287 #BEZUG!
blade modes get closer together and are nearly the same respectively. position is visualized in figure 4 (dotted circle). The local z-axis lies
36 1.6500 0.4837 #BEZUG! #BEZUG! 1.6500 0.5305 #BEZUG!
However, there are differences of up to 23 % for the first 15 natural Series.Run = 37 1.7000
LC_1_2_1 0.4846 #BEZUG! #BEZUG!
perpendicular to the wind direction in the horizontal plane. 1.7000 0.5319 #BEZUG!
frequencies. The frequency differences of about 5 % associated with Output Channel = 38 1.7500
M152N148 0.4850
47 #BEZUG! #BEZUG! 1.7500 0.5329 #BEZUG!
39 1.8000 0.4850 #BEZUG! #BEZUG! 1.8000 0.5334 #BEZUG!
the first global bending modes are not negligible especially if the im-
2.5 Name = 40 1.8500 beam 0.4845 #BEZUG! beam
#BEZUG! 1.8500 super
0.5335 #BEZUG!
Bending moment [MNm]
Company
portance for support structure design is taken into account. The 3rd Simulator Name = 41 1.9000 ########
0.4836 #BEZUG! 201209_2
#BEZUG! 1.9000 ########
0.5331 #BEZUG!
42
1 1.9500 M152N148
Time 0.4822 #BEZUG!
#BEZUG! #BEZUG! 1.9500 M152N148
#BEZUG! Time 0.5323 #BEZUG!
#BEZUG!
natural frequency is shifted into the upper 3p range, which is assumed 2 43
2 2.0000
(sec) (kN)0.4804 #BEZUG! #BEZUG! 2.0000
#BEZUG! #BEZUG! (sec) (kN)0.5311 #BEZUG!
#BEZUG!
to be critical and the most important result. Furthermore, the 9th nat- 1.5
44
3 2.0500 0.4784 #BEZUG!
0.0000 ######## #BEZUG! #BEZUG!
#BEZUG! 2.0500 0.5297 #BEZUG!
0.0000 ######## #BEZUG!
45
4 2.1000 0.4762 #BEZUG!
0.0500 ######## #BEZUG! #BEZUG!
#BEZUG! 2.1000 0.5281 #BEZUG!
0.0500 ######## #BEZUG!
ural frequency is shifted into the upper 6p range, a fact that should be 46 2.1500 0.4744 #BEZUG!
5 0.1000 ######## #BEZUG! #BEZUG!
#BEZUG! 2.1500 0.5268 #BEZUG!
0.1000 ######## #BEZUG!
treated with caution as well. 1 47
6 2.2000 0.4721 #BEZUG! #BEZUG!
0.1500 ######## #BEZUG! #BEZUG! 2.2000 0.5250 #BEZUG!
0.1500 ######## #BEZUG!
48
7 2.2500 0.4696 #BEZUG!
0.2000 ######## #BEZUG! #BEZUG!
#BEZUG! 2.2500 0.5232 #BEZUG!
0.2000 ######## #BEZUG!
0.5 49
8 2.3000
0.2500 ########
beam
0.4671 model #BEZUG! 2.3000
#BEZUG!
#BEZUG! #BEZUG! 0.5212 #BEZUG!
0.2500 ######## #BEZUG!
50
9 2.3500 superelement
0.4645 #BEZUG! model
#BEZUG!
0.3000 ######## #BEZUG! #BEZUG! 2.3500 0.5191 #BEZUG!
0.3000 ######## #BEZUG!
Time domain simulation 0 10 0.3500 ######## #BEZUG! #BEZUG! 0.3500 ######## #BEZUG!
11 0.4000 ######## #BEZUG! #BEZUG! 0.4000 ######## #BEZUG!
50 60 70 80 90 100
This section presents the results from one exemplary time do- 12 0.4500 ######## #BEZUG! #BEZUG!
Time [s] 0.4500 ######## #BEZUG!
13 0.5000 ######## #BEZUG! #BEZUG! 0.5000 ######## #BEZUG!
main simulation load case that was run with the baseline turbine RNA 14 0.5500 ######## #BEZUG! #BEZUG! 0.5500 ######## #BEZUG!
including the beam support structure and the superelement support 15 0.6000 ######## #BEZUG! #BEZUG!
Figure 11: Bending moments in upwind leg over 50 s for beam 0.6000 ######## #BEZUG!
structure. 16 0.6500 ######## #BEZUG! #BEZUG!
(black) and superelement model (light grey).
0.6500 ######## #BEZUG!
17 0.7000 ######## #BEZUG! #BEZUG! 0.7000 ######## #BEZUG!
The selected load case is an exemplary power production design load 18 0.7500 ######## #BEZUG! #BEZUG! 0.7500 ######## #BEZUG!
case (DLC) as defined in IEC (2005), which is based on the descrip- 19 0.8000 ######## #BEZUG! #BEZUG! 0.8000 ######## #BEZUG!
tions in Fischer et al. (2009) for a Dutch North Sea Site. One could The average bending moment resulting from the simulation with the
20 0.8500 ######## #BEZUG! #BEZUG! 0.8500 ######## #BEZUG!
speak of a typical fatigue load case for North Sea conditions as the super element model is slightly higher than the moment calculated
21 0.9000 ######## #BEZUG! #BEZUG! 0.9000 ######## #BEZUG!
22 0.9500 ######## #BEZUG! #BEZUG! 0.9500 ######## #BEZUG!
occurrence of its metocean parameters, that are shown in table 2, is with the beam model. Furthermore, the amplitudes of the fluctuating
23 1.0000 ######## #BEZUG! #BEZUG! 1.0000 ######## #BEZUG!
highly significant. bending moments are much bigger for the beam model.
24 1.0500 ######## #BEZUG! #BEZUG! 1.0500 ######## #BEZUG!
In table 3 the damage equivalent bending moments (damage equiva-
25 1.1000 ######## #BEZUG! #BEZUG! 1.1000 ######## #BEZUG!
26 1.1500 ######## #BEZUG! #BEZUG! 1.1500 ######## #BEZUG! 8
Table 2: Wind (average wind speed, turbulence intensity, spectrum) lent load: DEL) derived from the ten-minute time series for N = 2·10
27 1.2000 ######## #BEZUG! #BEZUG! 1.2000 ######## #BEZUG!
and wave (significant wave height, peak period, spectrum) load cycles and different Wöhler material exponents (m) are shown to
28 1.2500 ######## #BEZUG! #BEZUG! 1.2500 ######## #BEZUG!
29 1.3000 ######## #BEZUG! #BEZUG!
point out the impact of the higher load amplitudes. 1.3000 ######## #BEZUG!
parameters for the exemplary load case. 30 1.3500 ######## #BEZUG! #BEZUG! 1.3500 ######## #BEZUG!
The differences in the damage equivalent loads for this load case and
31 1.4000 ######## #BEZUG! #BEZUG! 1.4000 ######## #BEZUG!
load output are very significant for all Wöhler material exponents. For
32 1.4500 ######## #BEZUG! #BEZUG! 1.4500 ######## #BEZUG!
C ONCLUSION AND OUTLOOK ANSYS (2007). Release 10 Documentation for ANSYS. SAS IP Inc.
In this paper, an aero-elastic model of a 5MW OWT on three Craig, R. R. and Bampton, M. C. C. (1968). Coupling of substructures
different types of tripod support structure models is investigated. A for dynamic analysis. In AIAA Journal, volume 6.
state-of-the-art beam model, a detailed shell model and a superele-
ment model derived from the shell model. De Vries, W. (2008). Soft-stiff bottom mounted support structures -
The beam model used here is already a good representation of the analysis of monopile and multi-member support structures for off-
structure, its details are based on findings from the Offshore Code shore wind turbines. Upwind deliverable D4.3.2 (WP4: Offshore
Comparison Collaboration. The shell model leads to a much more foundations and support structures), Delft University of Technol-
accurate representation of the critical joints combined with the draw- ogy. Restricted to project members.
back of several thousands of supplementary DOF that would lead to
unacceptable computational effort for OWT load calculations. Using Fischer, T., de Vries, W., and Schmidt, B. (2009). Design basis -
the superelement model, the stiffness representation (stiffness matrix) upwind deep water site. Upwind deliverable (WP4: Offshore foun-
of the joints is as accurate as for the shell model whereas the dynamic dations and support structures), Endowed Chair of Wind Energy
representation (mass matrix) is by far accurate enough to meet the re- (SWE) at the Institute of Aircraft Design Universität Stuttgart.
quirements of fully coupled OWT simulation. The number of DOF -
and therefore the computational effort - is reduced, even compared to Fischer, T. and Kuehn, M. (2009). Site sensitive support structure and
the beam model. Preliminary model checks show the expected results machine design for offshore wind farms. In European Wind Energy
concerning masses and eigenstates of the support structure models. Conference (EWEC). Endowed Chair of Wind Energy (SWE) at the
A full system eigenanalysis shows that the superelement model leads Institute of Aircraft Design, Universität Stuttgart.
to several changes compared to the beam model. This concerns
the number of order of eigenmodes, natural frequencies and coupled Guyan, R. J. (1965). Reduction of stiffness and mass matrices. In
AIAA Journal, volume 3.
eigenstates. Even the 1st global support structure bending modes are
affected. Furthermore, the 3rd and the 9th natural frequencies are IEC (2005). Wind turbines - Part 1: Design requirements. IEC 61400-
shifted into the frequency ranges that should be avoided in OWT de- 1, 3.0 edition.
sign due to excessive dynamic loads (3P, 6P). A design based on the
beam model would therefore probably not be conservative in terms of Jonkman, J., Butterfield, S., Musial, W., and Scott, G. (2009). Defini-
global dynamic excitations. tion of a 5-MW Reference Wind Turbine for Offshore System De-
An exemplary time domain fatigue load case shows, that there are sig- velopment. Technical Report NREL/TP-500-38060, National Re-
nificant differences in the results due to the different support structure newable Energy Laboratory (NREL).
modeling approaches. On the one hand, the structure as a whole be-
comes more compliant due to the superelements. This leads to higher Jonkman, J. and Matha, D. (2009). A quantitative comparison of
displacements of the turbine for the presented load case, a fact that the responses of three floating platforms. In European Offshore
must be kept in mind as larger displacements lead to larger inertia Wind Conference (EOW). National Renewable Energy Laboratory
loads and therefore one might underestimate loads in some cases with (NREL).
the beam model. On the other hand, a more compliant joint reduces
the load amplitudes and therefore the DEL in the member near the Kleinhansl, S., Mayer, M., and Mangold, A. (2004). ADCoS - A Non-
joint significantly. This is an important outcome as design driving linear Aeroelastic Code for the Complete Dynamic Simulation of
joints may be dimensioned based on the loads gained from aero-elastic Offshore- Structures and Lattice-Towers. In Deutsche Wind Energie
modeling at those positions following the relevant standards. In such Konferenz (DEWEK). Aero Dynamik Consult Ingenieurgesellschaft
cases, the more detailed superelement approach would allow for much m.b.H.
less conserative fatigue design of the joints.
Despite these results, showing the strong influence of the superelement Larsen, T. J., Madsen, H. A., Thomsen, K., and Rasmussen, F. (2007).
joint modeling compared to the state-of-the-art beam model, it must Reduction of teeter angle excursions for a two-bladed downwind
be noted, that only exemplary load cases have been calculated so far. rotor using cyclic pitch control. In European Wind Energy Con-
The next steps of work are the definition and simulation of a set of load ference (EWEC). Risø National Laboratory Technical University of
cases based on the relevant guidelines and standards, the identification Denmark.
of critical load outputs and the quantification of the effects described
in this paper on, for instance, lifetime DELs. NASTRAN (2004). MSC.NASTRAN. Advanced Dynamics Analysis
However, it is a main result of this study, that the superelement mod- User’s Guide. MacNeal-Schwendler Corporation.
Nichols, J., Camp, T., Jonkman, J., Butterfield, S., Larsen, T., Hansen,
A., Azcona, J., Martinez, A., Munduate, X., Vorpahl, F., Kleinhansl,
S., Kohlmeier, M., Kossel, T., Böker, C., and Kaufer, D. (2009).
Offshore Code Comparison Collaboration within IEA Wind Annex
XXIII: Phase III Results Regarding Tripod Support Structure Mod-
eling. In 47th AIAA Aerospace Sciences Meeting. Garrad Hassan &
Partners Limited (GH).
Tarp-Johansen, N. J., Andersen, L., Christensen, E. D., Mørch, C.,
Kallesøe, B., and Frandsen, S. (2009). Comparing sources of damp-
ing of cross-wind motion. In European Offshore Wind Conference
(EOW). DONG Energy.
Vemula, N. K. (2009). Jacket foundation support structure. Upwind
deliverable D4.2.6 (WP4: Offshore foundations and support struc-
tures), Rambøll Wind Energy. Restricted to project members.
Vorpahl, F. R. (2009). Implementation of a superelement approach in
a design tool for offshore wind turbines with braced support struc-
tures. Upwind deliverable D4.3.4 (WP4: Offshore foundations and
support structures), Fraunhofer Institute for Wind Energy and En-
ergy System Technology (IWES). Restricted to project members.
Vorpahl, F. R., Blunk, M., Wingerde, A. V., Busmann, H.-G., and
Kleinhansl, S. (2009a). Implementation of a superelement approach
in a design tool for offshore wind turbines with arbitrary support
structures. In European Offshore Wind Conference (EOW). Fraun-
hofer Institute for Wind Energy and Energy System Technology
(IWES).
Vorpahl, F. R., Huhn, H., Busmann, H.-G., and Kleinhansl, S. (2007).
A Flexible Aero-elastic Simulation Approach for Offshore Wind
Turbines. In European Offshore Wind Conference (EOW). Fraun-
hofer Center for Wind Energy and Maritime Engineering.
Vorpahl, F. R., Wingerde, A. V., Blunk, M., Busmann, H.-G., Klein-
hansl, S., Kossel, T., Kohlmeier, M., Böker, C., Kaufer, D., Azcona,
J., Martinez, A., and Munduate, X. (2009b). Validation of a fi-
nite element based simulation approach for offshore wind turbines
within iea wind annex XXIII - simulation challenges and results
for a 5-MW turbine on a tripod substructure. In Proceedings of
the Nineteenth International Offshore and Polar Engineering Con-
ference. Fraunhofer Institute for Wind Energy and Energy Systems
Technology (IWES).