Offshore Wind Turbine Support Structures

Transcription

1 1 Wind Power R&D Seminar Deep Sea Offshore Wind Royal Garden Hotel, Trondheim, Norway January 21, 2011 Effect of Foundation Modeling Methodology on the Dynamic Response of Offshore Wind Turbine Support Structures

2 2 Agenda Motivations o for research Research Questions Project details and methods Results Conclusions Further work

3 3 Motivations for Research To reduce the costs while increasing the performance and reliability of offshore wind energy through advancements in foundation modeling techniques and design methods Courtesy: GL-Garrad Hassan 2009

4 4 Higher costs largely due to offshore support structures Support structures make up a much higher percentage of the total costs offshore This trend is likely to continue as water depth increases at wind farm sites OFFSHORE Contribution to Total Cost Component Onshore Offshore Turbines (excluding 68-84% 49% works) Support Structure 1-9% 21% Grid Connection 2-10% 16% Consultancy 2-8% 9% Electric Installation 1-9% 5% Other 2-10% 1% Courtesy: EWEA 2010

5 5 Installation Difficulties Very large and expensive installation ti vessels are required Foundations and tower must be installed to very precise tolerances Many components must be installed in calm weather to avoid damage Bad weather can lead to large amounts of downtime, running up costs Foundation installation is the most time consuming part of the process Extremely large diameter piles or immensely heavy gravity based must be installed Preparation of the seabed and scour protection may be required Offshore foundations cost 2.5x more than for a similar land-based wind turbine

6 6 Reducing the costs... Efficiently designed support structures and foundations Specifically engineer foundations for loads and site conditions at each offshore wind turbine Develop computer software tools specifically produced for offshore wind turbine foundation design Mass production of offshore wind turbine support structures Towers and foundations must be designed in a way that is economical to mass-produce Efficient use of materials, manufacturing facilities, and manpower Purpose built offshore wind support structure manufacturing facilities will be needed Improved installation techniques and equipment New foundation technology which is easier and quicker to install Purpose-built installation vessels to install wind turbines in a cost effective manner

7 7 Pile Foundations Models Fully coupled finite element model simulation Most comprehensive modeling technique, includes many additional non linear effects Includes interactions between soil layers (vertical) and between adjacent piles (horizontal) C E Very time consuming and expensive, requires extensive soil lab testing R Sequential analysis with finite element simulations Combines the capabilities of the multiple non-linear spring model with finite element simulations Allows for dynamic FE simulations of the foundation without the need for a fully coupled model T A I N Multiple non-linear spring representation (p-y curves) Foundation modeled with springs distributed along length of pile T Y & C O S T Dependant on accurate soil profile and characteristic parameters Single non-linear spring representation Entire foundation modeled with single springs at mudline for each DOF Does not account for pile flexibility or soil profile non-homogeneity Model with an equivalent fixity it depth (Apparent Fixity it Length) Very simple and fast in computations, more representative than fixed condition Does not capture any soil-structure interaction Assume fixed boundary conditions Extremely simple, fast computations Gross misrepresentation of stiffness of the foundation S I M P L I C I T Y & S P E E D

8 8 Structural efficiency or design efficiency? Separately design each pile to give the minimum installation time and maximum structural efficiency Each foundation designed to only the minimum required length and diameter Less overall material use, reduced fabrication effort, less time and effort for installation More time, man-hours, and money spent during the testing and design phase Develop a single pile design that can be utilized for all structures in the entire wind park Foundations designed for worst case, many piles may be grossly overdesigned Higher overall material use, increased fabrication effort, more difficult installation Less time, effort and money spent in the testing and design phase VS

9 9 Research Questions 1. Do the most simple modeling techniques provide and accurate a enough description of the dynamic characteristics to be used for preliminary design and analysis? 2. Does the added accuracy and certainty in analysis and design of an offshore wind turbine foundation when using more advanced modeling techniques outweigh the additional costs of using such techniques?

10 10 Uncoupled Foundation Models Four different foundation modeling techniques are considered Fixed boundary conditions Apparent Fixity Length (AFL) Uncoupled Springs Distributed non-linear spring model using force-displacement (p-y) curves

11 11 Apparent Fixity Length Boundary effect of soil clamping is approximated by fixing the pile a certain depth (AFL) below the seabed AFL chosen to match the stiffness of the pile with distributed spring model Only matches at one given load due to non-linearity of p-y curves Can also be determined based on soil properties ~

12 12 Uncoupled Springs Static forces are applied in x, and z directions DOFs to determine the uncoupled spring stiffnesses Can be determined using two different approaches Appiled Force/Moment method Forced displacement/rotation method Can be modeled with linear or non-linear springs Borrowed from Zaaijer (2002)

13 13 Distributed NL spring models Force displacement (p-y) curves found in the design standard are used for horizontal and vertical displacements ISO 19902:2007(E) Petroleum and natural gas industries fixed steel offshore structures (Ch 17) Dependant on undrained shear strength profile, friction angle, unit weight of soil, and pile diameter Not really suitable to extremely large diameter piles (such as those used on monopile wind turbines) Hyperbolic force displacement relationship used for torsional stiffness Method developed by Randolph and Guo Torsional Piles in Non-homogenous Media (1996) Dependant on undrained shear strength profile, unit weight of soil, pile stiffness, pile diameter 1200 p-y curve for clay at depth z= t-z curve for clay at depth z= tion (kn/m 2 ) Subgrade React ction (kn/m 2 ) Subgrade Reac Horizontal Displacement (meters) Vertical Displacement (meters)

14 14 Support Structures Monotower Chose a generic design, representative of currently producing turbines 120m height, 35mm wall thickness, diameter tapering from 5.5m to 3m Full-height lattice tower Designed by former NTNU PhD student Haiyan Long 120m height, 4 legs, 10 sections. 21 meters wide at base, 4 meters at nacelle

15 15 Monotower Comparison

16 16 Lattice tower comparison

17 17 Conclusions Significant discrepancies noted between the different foundation models Not immediately clear which is most accurate, but worth investigating further The discrepancies are mostly due to dynamic amplifications Response is very sensitive to changes in the selected soil parameters More detailed soil descriptions and response models are needed Actual soil profile and soil properties from an offshore wind turbine site needed No interaction between soil layers or between adjacent piles Future models must include 3-D soil interaction effects Models must include time dependent effects such as drainage and dilatency

18 18 Research Questions 1. Do the most simple modeling techniques provide and accurate a enough description of the dynamic characteristics to be used for preliminary design and analysis? 2. Does the added accuracy and certainty in analysis and design of an offshore wind turbine foundation when using more advanced modeling techniques outweigh the additional costs of using such techniques?

19 19 Coupled Foundation Models Sequential Analysis FE Method Method used to investigate the response of a piled foundation to the loads experienced on an offshore wind turbine structure using the finite element method An iterative process of finite element simulations of the soil-pile structure and the wind turbine structure Does not allow data to feed into the aero-elastic code at each time step, only as initial iti conditions Static FEM NL Soil Springs HAWC2 Simulation Dynamic FEM Time Series Force Data

20 20 Coupled Foundation Models Fully-Coupled FE Model Foundation, or Geo module to be develop using open source FEM foundation code (such as OpenSees, Code Aster, etc.) Geo Module then fully coupled with an Aero-Servo-Hydro-Elastic code (FAST, FLEX5, ADAMS, etc.) Adding an analysis tool for the foundation system is the last piece needed d to provide a proper analysis of the entire wind turbine system

21 21 Further Work Develop a FEM code for foundation response which can be coupled to a Aero-Servo-Hydro-Elastic simulation (Aero-Servo-Hydro-Geo-Elastic) Can be implemented and coupled with FAST/ADAMS or other open source code Allows for a time domain analysis of the entire wind turbine system Investigate dynamic processes of scour and the impacts on soil stiffness and damping Changes in soil properties can have significant impacts on the fatigue life of the structure Impact will be more significant with shallow foundations such as suction caissons Extend investigations to suction caissons and other foundation solutions Potential foundation concepts can be used in conjunction with a number of different tower concepts Validate numerical models with field data (RAVEN)

22 22 Questions? Thank you for your attention ti Contact: Eric Van Buren, Ph.D. Candidate, NTNU