Elsevier

Engineering Structures

Volume 31, Issue 10, October 2009, Pages 2236-2246
Engineering Structures

Simulation of offshore wind turbine response for long-term extreme load prediction

https://doi.org/10.1016/j.engstruct.2009.04.002Get rights and content

Abstract

When there is interest in estimating long-term extreme loads for an offshore wind turbine using simulation, statistical extrapolation is the method of choice. While the method itself is rather well-established, simulation effort can be intractable if uncertainty in predicted extreme loads and efficiency in the selected extrapolation procedure are not specifically addressed. Our aim in this study is to address these questions in predicting blade and tower extreme loads based on stochastic response simulations of a 5 MW offshore turbine. We illustrate the use of the peak-over-threshold method to predict long-term extreme loads. To derive these long-term loads, we employ an efficient inverse reliability approach which is shown to predict reasonably accurate long-term loads when compared to the more expensive direct integration of conditional load distributions for different environmental (wind and wave) conditions. Fundamental to the inverse reliability approach is the issue of whether turbine response variability conditional on environmental conditions is modeled in detail or whether only gross conditional statistics of this conditional response are included. We derive long-term loads for both these cases, and demonstrate that careful inclusion of response variability not only greatly influences such long-term load predictions but it also identifies different environmental conditions that bring about these long-term loads compared with when response variability is only approximately modeled. As we shall see, for this turbine, a major source of response variability for both the blade and tower arises from blade pitch control actions due to which a large number of simulations are required to obtain stable distribution tails for the turbine loads studied.

Introduction

Statistical extrapolation of extreme loads is being increasingly used in the design of offshore wind turbines against ultimate limit states, and a recent draft [1] of design guidelines from the International Electrotechnical Commission (IEC) also recommends its use. Statistical extrapolation involves integration of the distribution of turbine loads given specified environmental states with the likelihood of occurrence of the different environmental states; the (conditional) load distributions are obtained by means of turbine response simulations.

While extrapolation methods are relatively well understood for onshore wind turbines [e.g., [2], [3], [4]], they present several challenges for offshore turbines. For one, the offshore environment involves, as a minimum, the consideration of waves in addition to wind; hence, the number of random variables describing the environment increases. As a result, the domain of integration expands and it can often become impractical to perform computationally expensive simulations over this entire domain if one uses the basic extrapolation approach that involves direct integration. It is thus of interest to explore efficient alternative extrapolation techniques for offshore wind turbine design. A second challenge is that extrapolation of turbine loads needs to recognize the dependence on two (or more) random processes representing the environment—wind and waves, say—each of which influence turbine loads in distinct ways. Several studies in recent years have focused on the complexity of these issues in the offshore environment and have addressed comparisons of alternative methods to extract turbine load extremes [5], possible reduction in simulation effort by careful selection of critical environmental states [6], use of the environmental contour method [7], and use of a suitable percentile of the wave-related random variable (conditional on wind speed) in lieu of the full joint wind-wave distribution [8].

On related matters to those highlighted in these previous efforts, we attempt here to answer several open questions regarding how the peak-over-threshold method should be used with the environmental contour method; whether or not the environmental contour method, which requires considerably less simulation effort is as accurate as direct integration in statistical load extrapolation; and whether or not variability in turbine loads must be carefully accounted for in order to yield accurate long-term loads. To address these issues, we derive long-term loads using a model of a utility-scale 5MW offshore wind turbine that was developed at the National Renewable Energy Laboratory (NREL), and which is assumed to be sited in 20 m of water. Stochastic time-domain simulations of turbine response form the basis for this study. While the inflow turbulence describing the wind field is simulated using similar procedures to those for onshore turbines, excitation from waves is simulated assuming simplified linear irregular wave kinematics that may not be suitable for this shallow-water site. In shallow waters, irregular waves are more appropriately modeled using a second-order wave theory such as that developed by Sharma and Dean [9]. While such nonlinear wave modeling capabilities were not fully integrated with wind turbine aeroelastic simulation software such as FAST [10] at the time of this writing, such enhancements are planned, given the preponderance of shallow-water sites for offshore wind turbines.

The outline of this work is as follows: after describing the extrapolation methods and the simulation model, we examine turbine response statistics for several representative environmental conditions. We then discuss application of the peak-over-threshold (POT) method to derive probability distributions of turbine loads. We illustrate how long-term loads can be derived using the Inverse First-Order Reliability Method, first by omitting turbine load variability as in the environmental contour (EC) method, and then by explicitly accounting for this variability (given environmental state). Comparison of EC-based long-term load predictions with those obtained by direct integration is discussed. We also discuss how turbine control actions influence variability in long-term loads. Finally, we compare predictions of rare (long-term) load fractiles based on the POT and global maxima methods.

Section snippets

Load extrapolation methods

Design Load Case (DLC) 1.1b of the IEC 61400-3 draft design guidelines [1] for offshore wind turbines, which is based on DLC 1.1 of the IEC 61400-1 guidelines [11] for onshore wind turbines, recommends the use of statistical extrapolation methods to predict turbine characteristic loads for an ultimate limit state. In direct integration, which is most often employed in statistical extrapolation for wind turbine extreme loads, one seeks to estimate the turbine long-term load, lT, associated with

Simulation model

A 5MW wind turbine model [15] developed at NREL and closely representing utility-scale offshore wind turbines being manufactured today is used in our simulation studies. The turbine is assumed to have a hub height of 90 m above the mean sea level, and a rotor diameter of 126 m. The turbine is a variable-speed and collective pitch-controlled machine, with a maximum rotor speed of 12.1 rpm. The rated wind speed is 11.5 m/s. The turbine is assumed to be sited in 20 m of water; it has a monopile

Turbine response

We are interested in the response of the turbine only while it is in an operating state. Accordingly, we perform response simulations for mean wind speeds ranging from cut-in to cut-out wind speeds (i.e., 3 to 25 m/s, here). As a function of the mean wind speed in each simulation, the turbulence intensity is assumed per IEC Class I-B site conditions using the Normal Turbulence Model (NTM) [11]. The peak spectral period for the waves is modeled as a function of significant wave height based on

Short-term extreme load distributions

The short-term distribution of turbine extreme loads, FL|X(l), which enables prediction of long-term loads according to Eq. (1), requires data on load extremes. The global maximum and peak-over-threshold methods are commonly used to extract load extremes from time series data. We use the peak-over-threshold method here, as it can provide a large number of load extremes from a given number of simulations, resulting in better definition of distribution tails which is important when extrapolating

Comparison of POT and global maxima

In the preceding discussions, we used the peak-over-threshold (POT) data to extract load extremes. An alternative approach is to use global (or epochal) maxima in which only statistics of the single largest load value from each simulation are used. It is of interest to examine how long-term load predictions differ from the two methods. We fit two-parameter Weibull distributions to the tails of global maxima data for the design environmental states, and estimate load fractiles required with the

Conclusions

Our objective in this study was to derive long-term loads for a utility-scale 5MW offshore wind turbine sited in 20 m of water. The focus was on the out-of-plane blade bending moment at a blade root and the fore-aft tower base moment at the mudline. Load extremes data needed to establish short-term load distributions were extracted from time series of turbine response simulations using the peak-over-threshold method. Long-term loads were estimated using 2-D and 3-D inverse first-order

Acknowledgments

The authors gratefully acknowledge the financial support provided by two grants from the National Science Foundation—CMMI-0449128 (CAREER) and CMMI-0727989. They also acknowledge assistance from Dr. Jason Jonkman at the National Renewable Energy Laboratory with the wind turbine simulation model used in this study.

References (18)

  • IEC 61400-3. Wind turbines—Part 3: Design requirements for offshore wind turbines. Intl. Electrotechnical Commission,...
  • P.J. Moriarty et al.

    Extrapolation of extreme and fatigue loads using probabilistic methods

    (2004)
  • K. Saranyasoontorn et al.

    Design loads for wind turbines using the environmental contour method

    J Solar Energy Eng, Trans ASME

    (2006)
  • Ragan P, Manuel L. Statistical extrapolation methods for estimating wind turbine extreme loads. In: Proc. ASME wind...
  • Cheng PW. A reliability based design methodology for extreme response of offshore wind turbines. Ph.D. dissertation....
  • E. Norton

    Investigation into IEC offshore draft standard design load case 1.1

  • K. Saranyasoontorn et al.

    On assessing the accuracy of offshore wind turbine reliability-based design loads from the environmental contour method

    Int J Offshore Polar Engrg

    (2005)
  • N.J. Tarp-Johansen

    Extrapolation including wave loads—Replacing the distribution of Hs by a suitable percentile

  • J.N. Sharma et al.

    Second-order directional seas and associated wave forces

    Soc Petroleum Eng J

    (1981)
There are more references available in the full text version of this article.

Cited by (0)

1

Currently at Stress Engineering Services, Houston, TX, USA.

View full text