Lead Session Chair:
Stephan Barth, Managing Director, ForWind - Center for Wind Energy Research, Germany
Benjamin Rocher (1) F P Anne-Laure Caouissin (1) Arnaud Salou (1) Franck Schoefs (2) Marc François (2) Matthieu Gélébart (1)
(1) STX France Solutions, Saint Nazaire, France (2) Laboratory of civil and mechnical engineering, Nantes, France
Printer friendly version: Print
Download poster(1.41 MB)
Tool for decision support for maintenance of offshore support structures of wind turbine
With the development of offshore renewable energy market and especially number of offshore wind farms, maintenance activities are increasing significantly. To limit the cost involved by the latter, it becomes crucial to improve fatigue design process in welded joints of steel offshore structures by mixing robust and probabilistic numerical analysis and structural health monitoring.
Up to now, current fatigue design criteria for offshore structures are based on S-N curves, rainflow cycle counting and Palmgren-Miner approach. It is well-known that this approach could be dangerous or over-conservative. In addition, most current fatigue analysis is based on spectral method. However, the latter does not allow engineers to take into account time-dependent parameters.
A method based on a two-scale probabilistic model for high-cycle fatigue life predictions has been developed. This method may be divided in three steps: The first two steps are commonly used. The third one is a new approach. The computation process is described below:
Main body of abstract
- A realistic dynamic method integrating aerodynamic loads, hydrodynamic loads and structural responses of the entire wind turbine is developed to generate time series of wind loads applied on the midpoint between the transition piece and the mast.
- According to the wind load time series and associated sea-state, time series of stress are then obtained by using a finite element code modelling a jacket beams model. Mentioned stresses are combined with the stress concentration factors to estimate stress values at weld toe.
- Sheet metal fabrication and assembly by welding with conventional methods generate microcracks which can cause to macroscopic cracks. A post-processing of stress results has been performed by using the two scales damage model (TSDM) developed by Lemaitre & Desmorat . To take into account the micro defects in the material, especially at the weld toes, elasto-plastic damageable constitutive law is used for material behaviour. Material parameters are previously identified on fatigue test results or S-N curves. Time history and mean stresses are taken into account because the TSDM is an incremental process of cumulative plastic strain.
Reliability approach is also under investigation presently because fatigue design parameters have significant uncertainty. This method allows taking into account:
- Uncertainty of material parameters included in fatigue analysis. Indeed, material parameters of TSDM are now considered as random variables.
- Environmental variability e.g. wave height, period and direction, wind speed and direction.
- Uncertainty of measurements issued of structural health monitoring
Thanks to the structural health monitoring, it becomes possible to have real-time state information concerning offshore structures and then anticipate their maintenance requirements. Real-time fatigue damage can directly be evaluated with TSDM using real-time signal of stress. It permits to identify the welded joints which need to be inspected during next operation on site. Moreover, if crack initiation is detected, material parameters can be updated using Bayesian updating method and used for future evaluation of reliability index.
With this new fatigue design method of offshore support structures, by mixing numerical analysis and structural health monitoring, reducing future maintenance cost of offshore wind turbines seems to be realistic.