Model Based Identification of Hydrodynamic Loads and System Parameters for Offshore Wind Turbine Monopile Support Structures: A measurement driven, model based, identification approach for the joint estimation of wave loads and system paramaters

Model Based Identification of Hydrodynamic Loads and System Parameters for Offshore Wind Turbine Monopile Support Structures: A measurement driven, model based, identification approach for the joint estimation of wave loads and system paramaters

Fallais, D.J.M.

Lourens, E. (mentor)

Mechanical, Maritime and Materials Engineering

Marine & Transport Technology

MT - Ship and Offshore Structures

2015-10-23

Identifying unknown forces and system parameters belongs to the field of force and system identification. A specific area of this field is concerned with performing this identification based on physical models and measurable response data. The physical model which is used for this so called model-based identification has to resemble the structure on which the identification will be performed. In this thesis a model based and measurement driven identification approach is used to identify a hydrodynamic load, a system parameter and the response behavior of an offshore wind turbine mono-pile support structure. Having the opportunity to improve knowledge on uncertainties such as hydrodynamic loads or e.g. stiffness and damping properties, may allow to generate more accurate input for response calculations which in return can lead to a more cost efficient design. This work is based on the principle of adaptive filtering. Adaptive filtering is a concept used in state estimation which allows to augment the state vector by additional variables which are to be estimated. For problems in structural dynamics the state is typically defined by the set of displacements and velocities which are required to describe the response of the structure. However, augmenting the physical model and the state vector by a parameter to be estimated allows for the joint estimation of the states and a parameter. Extending this concept such that response governing forces are included in the "augmented state vector" and the "augmented system" allows to perform a joint estimation of response governing forces, parameters and the systems state. As a consequence of estimating a parameter in conjunction with forces and states, the system will inherently become nonlinear. Therefore the extended kalman filter, which is a well known non-linear state estimation technique, is selected as method in order to identify equivalent wave loads, parameters, and the systems state. In this thesis a realistic finite element model of a 6 megawatt offshore wind turbine in 40 meter water depth is used. based on this model artificial measurement data is generated using hydrodynamic loading associated with normal and extreme sea states. The finite element model is subsequently augmented, modally reduced, and coupled to the extended Kalman filter. In order to investigate the properties and the applicability to a practical problem three different types of estimation are performed: First the concept is applied to jointly estimate an equivalent hydrodynamic load and the states of a offshore wind turbine support structure using a limited number of response measurement data. For this case the model is assumed to be known perfectly. The results show that the select method is capable of identifying equivalent forces which resemble the exact forces up to a high degree. Secondly the method is used to perform a non-linear joint force-parameter-state estimation. This technique is used to perform a free decay test to identify one modal damping parameter in conjunction with the states and a equivalent force. Further the same technique is used for the normal sea state conditions to demonstrate the potential in estimating one equivalent wave load in conjunction with one global stiffness related parameter and the states of the system. The results show that the proposed method is capable of tracking both the hydrodynamic loading and a response governing parameter of the support structure. Third. The concept will be used in order to perform a joint parameter-state estimation which aims at identifying the modal damping properties under the assumption that parts of the system and the hydrodynamic load are known. The results for this case show that the method is capable of accurately identifying the selected parameter.

Force identification

http://resolver.tudelft.nl/uuid:8448d07e-c96e-4811-ab7b-f5a236ff897a

Student theses

master thesis

(c) 2015 Fallais, D.J.M.