Offshore wind turbine design Addressing uncertainty drivers Sten Frandsen Niels Jacob Tarp-Johansen Erik Asp Hansen Michael Høgedal Lars Bo Ibsen Leo Jensen Risø National Laboratory DHI – Water and Environment Vestas Wind Systems Aalborg University ELSAM Engineering

Utilising experience from demonstration projects Objectives of work presented: To get the most out of the two Danish demonstration projects in terms of structural design

North Sea: Horns Rev - monopile

….and Nysted: Demonstration project in the Baltics Concrete, gravity foundations ( km ) Nysted

Design components Design in general has many important sub-components like: Environmental protection Cabling Preparation and construction etc. – and: Structural loads, response and design loads For this project the target is design loads, extremes and fatigue: Measurements Wind and wakes Waves and current Soil conditions Response Synthesis of load cases

The design process Response integration

Measurements COMPONENTS: FOCAL POINTS: Model verification measurements Models Design calculations External conditions’ recording Atmospheric measurements, incl. wake measurements Wave measurements Geotechnical measurements FOCAL POINTS: Model verification measurements Interpretation of specific loads based on response measurements. Load response measurements Extremes from measurement. Metocean measurements Measurement of 2D/3D wave kinematics. Measurements of scour level. Wake measurements Conceptual – which characteristics of the wake affect the wake loading and how? Relationship between the wind turbine characteristics and flow characteristics Reference mean wind speed. Measurement of turbulence, wake width etc. to which the rotor is actually exposed. measurements wind and wakes waves and current soil conditions aero/hydro-response synthesis

Measurements on offshore wind farms measurements wind and wakes waves and current soil conditions aero/hydro-response synthesis

Instrumentation, Horns Rev Wind turbine Wind Waves measurements wind and wakes waves and current soil conditions aero/hydro-response synthesis

Wind and wake-induced loads COMPONENTS: External conditions Gusts, mean wind climate, turbulence, 50y extreme wind, shear, spectra, air density Wakes CT curves for stall and pitch-regulated wind turbines Turbulence as function of separation and wind speed Separation of effect of turbulence and mean deficit Discrete flow structures in wakes Rotor aerodynamics Adequacy of contemporary aero-codes FOCAL POINTS: Extreme gusts during normal operation CT curves vs. turbulence and velocity deficit Models for aerodynamic rotor loading Updating of models for extreme and fatigue loading measurements wind and wakes waves and current soil conditions aero/hydro-response synthesis

Extreme gust during normal operation measurements wind and wakes waves and current soil conditions aero/hydro-response synthesis

Waves and current COMPONENTS Determination of the sea states that shall be considered in design (wave height, period and direction, current, current direction, water level, etc.). Evaluation of combined undisturbed wave and current kinematics at the OWT position. Evaluation of the time varying loads from the kinematics. FOCAL POINTS Cases where the engineering models are sufficiently accurate and cases where more precise modelling is required. Determination of kinematics and time varying loads.. ..from steep waves superimposed on current, by use of a 3 dimensional Navier-Stokes solver, to be applied where simpler methods are not applicable. This action is parallel to attempts to derive approximate methods for implementation in aero-elastic codes. Contributions from loads on appurtenances. Determination of maximum run-up. measurements wind and wakes waves and current soil conditions aero/hydro-response synthesis

Soil conditions COMPONENTS Design drivers in general: FOCAL POINTS: Water depth Possible erosion Size and type of wind turbine Environmental conditions (wave height, current, ice, etc.) Economics and politics FOCAL POINTS: Lateral pile resistance The ULS calculations for the piles are to be compared with an elasto-plastic finite element Analysis with the same soil conditions. Load-deflection curves (p-y curves): Performance of the p-y curves for pile diameters larger than 4 meters should be investigated. Curvature of the p-y curves in the pre-plastic portion is typically approximated by parabolic expressions. These approximations are not useful in the small-strain area. Behaviour of piles under cyclic loading. measurements wind and wakes waves and current soil conditions aero/hydro-response synthesis

Aero/hydro-elastic modelling COMPONENTS Effect of choice of structural modelling scheme on the response modelling, viz. e.g. finite-element method or modal formulation used for structural elements Aerodynamic and hydrodynamic load models’ ability to determine correctly the loading, based on proper input parameters representing the external conditions Foundation models ability to represent correctly the soil- structure interaction for both simple and more complex foundation types and the effect of the selected implementation scheme on the response modelling FOCAL POINTS Compare the response from simulations from different structural models on a generic turbine on a number of identical artificial test conditions Compare the response from different structural models with measurements of response and external parameters exist Feasibility of improving elements of the structural models Implement improvements and quantify reduced uncertainties by performing verification modelling measurements wind and wakes waves and current soil conditions aero/hydro-response synthesis

Handling and synthesis of response calculations 1st edition of standard for design of offshore wind turbines IEC61400-3 is currently being issued in its final draft Through a lengthy process the draft was created as an extension of the onshore standard This implies, that relative to the onshore standard the number of load cases has increased substantially due to the addition of wave loads In turn, also the composition of relevant load cases has become a more difficult task measurements wind and wakes waves and current soil conditions aero/hydro-response synthesis

Load cases – who many do we need? measurements wind and wakes waves and current soil conditions aero/hydro-response synthesis

Computational needs with present draft standard A complete set of simulations: 1000-1500 runs of each 10min duration measurements wind and wakes waves and current soil conditions aero/hydro-response synthesis

Extreme load cases Normal operation Special load cases measurements wind and wakes waves and current soil conditions aero/hydro-response synthesis

What – if anything – is less than perfect in IEC61400-3? Not enough specific in terms of definition of loads and how to combine load cases The extend of load cases may signal comprehensiveness, also in terms of accuracy The many load cases in reality reflect uncertainty – “some extra load cases don’t hurt” The instructions regarding how to perform site assessment are not sufficiently specific Several components – viz. the descriptions given previously in this talk – may be improved First and last – there are too many load cases. measurements wind and wakes waves and current soil conditions aero/hydro-response synthesis

Suggested structure of design process Aim is to optimize structural design by reducing uncertainty of component models reducing uncertainty in load systhesis rationalising load cases tune load case synthesis Thus, main design components: Site assessment Dynamic analysis Load cases and synthesis measurements wind and wakes waves and current soil conditions aero/hydro-response synthesis

Site assessment Measurements and hindcast Joint probability density function (JPDF) ways of estimating its uncertainty measurements wind and wakes waves and current soil conditions aero/hydro-response synthesis

Dynamic analysis Investigate the dynamic properties of the wind turbine adequacy of aero/hydro elastic model Identify possible peculiarities of design This action should encompass special load cases Investigate whether structure can be subdivided into components Sensitivity analysis, including effect of joint-action of climate variables; theresponse function is to be mapped for all relevant combinations of x measurements wind and wakes waves and current soil conditions aero/hydro-response synthesis

Load cases and synthesis Formal load cases should cover wt in operation Derive conditional distributions on basis on prescribed random variability of response, conditioned on external conditions Simulate without aero/hydro-elastic code to determine unconditional distribution Derive equivalent load conditioned on external conditions Extreme response from Fatigue from measurements wind and wakes waves and current soil conditions aero/hydro-response synthesis

Conclusion Optimisation and decrease of uncertainty of the design of offshore wind turbines are sought Therefore, a number of design-process components that contribute the most to the aggregated uncertainty has been identified Considerations regarding handling of load cases and synthesis these have been presented