09:30 - 11:00 Innovative design and validation tools
Guidelines to optimize design processes in terms of cost and time-to-market while ensuring necessary standards when variants are introduced, during early stages of new concepts and as the industry moves to larger turbines.
- Progress on smart rotor control using sensing strategies for optimal design
- Probabilistic design framework to quantify uncertainties and reduce risk in design
- New approach to validation of simulation models for grid studies to deliver variants earlier and economically
- Time benefits through novel dual axis fatigue testing of blades
- Delegates will take away guidelines for the optimal design of smart rotor systems
- Delegates will be able to implement a probabilistic design method enabling assessment of which uncertainties have most influence on overall COE
- Delegates will be able to propose a feasible alternative to traditional validation with field measurement data
- Delegates will be able to advocate optimized dual axis fatigue testing of blades
Lead Session Chair:
Michaela O'Donohoe, Adwen, Spain
Peter Jamieson, University of Strathclyde, United Kingdom
Ricard Buils Urbano (1) F Graeme McCann (1) David Langston (1) Ed Norton (1) Lindert Blonk (2)
(1) DNV GL - Energy Advisory, Bristol, United Kingdom (2) DNV GL - Energy Advisory, Groningen, The Netherlands
Printer friendly version: Print
Presenter's biographyBiographies are supplied directly by presenters at EWEA 2015 and are published here unedited
Ricard Buils has been working in the wind industry for 9 years. He is currently the Head of the Loads Analysis Section at DNV GL - Energy Advisory, within Turbine Engineering department in Bristol. He obtained an MSc in Mechanical Engineering (aeronautics specialisation) from the UPC in Barcelona (ETSEIB) and the ULB in Brussels in 2006. After gaining experience in wind energy resource assessment at ERSA in Barcelona, he joined Garrad Hassan in 2007 (now DNV GL). Since then he has participated in several research projects of different nature, including topics as blade aerodynamics optimisation, floating wind modelling and probabilistic design.
A probabilistic design framework applied to a 7MW wind turbine for robust cost of energy optimisation
Current wind turbine design methods generally follow a deterministic methodology based on current design standards in which design uncertainties are modelled primarily by safety factors and characteristic design parameters. A number of authors have considered probabilistic design approaches, where the uncertainties in the process are quantified and propagated explicitly through the design methodology. Most studies to date have implemented this using Structural Reliability Analysis (SRA) techniques in order to estimate the annual probability of failure for a given component and the work presented here extends that approach with the addition of an engineering Cost of Energy (CoE) model to provide a quantitative methodology for wind turbine concept design.
In this paper, a probabilistic design framework will be presented and applied to a 7MW wind turbine previously designed according to state-of-the-art deterministic methods following IEC 61400-1 edition 3. In particular, the work will present the application of SRA to the analysis of different turbine components and failure modes with design uncertainties propagated through to CoE and the resulting turbine design compared to the original deterministic design.
A probabilistic design framework has been implemented, which makes use of “classical” SRA techniques such as First Order Reliability Method (FORM). This provides a more robust design approach, which aims at achieving a consistent level of safety across the overall wind turbine system. A limit state function is defined for each limit state of interest, and the reliability level or annual probability of failure of the component and failure mode may be calculated. Typical wind turbine applications have a target annual probability of failure in the order of 1e-4 to 1e-5.
Main body of abstract
An inventory of uncertainties in the design process has been built and categorized in different uncertainty distributions for both the load and resistance side. These are used as inputs to the framework and applied to the 7MW turbine designed using a standard deterministic approach, with the aero-elastic simulations performed using Bladed. Limit state functions have been implemented for different components each with different levels of complexity:
- Tower: fatigue and buckling limit states
- Mainframe: fatigue and extreme (yield) limit states
- Blade tip to tower clearance limit state
- Aerodynamic instability (flutter) limit state
For each, a FORM analysis is performed and an annual probability of failure calculated and compared to the original deterministic design, to identify over/under-conservatively designed components. The method also allows identification of which uncertainties significantly govern the design for each limit state.
Furthermore, the framework has been coupled to an internal CoE calculation tool, which makes use of engineering models to predict CoE of the entire wind plant system at conceptual stage. The results show which uncertainties in certain components of the turbine have most influence on the overall CoE of the system.
Current wind turbine design standards follow a deterministic approach by which uncertainties in the design process are modelled through fixed partial safety factors. Probabilistic design methods aim at explicitly quantifying the various uncertainties in the design process, and thus have the potential to provide a more robust approach by minimizing the risks of both over and under-conservativism.
A probabilistic design framework has been implemented and applied to an existing 7MW wind turbine design. Several limit state functions have been modelled and the reliability levels calculated and compared to the original deterministic design. The framework has also been used to assess which design uncertainties have most influence on the overall CoE.
Readers of this paper will learn about the current state of the art in probabilistic design methods applied to wind turbine design. Specifically, readers will learn about when and how to use such methods to reduce design risks and achieve a more cost effective design. A range of recommendations for further investigation in this domain will also be presented.