Aeroelastic stability analysis and certification of wind turbine blades

Σχοινάς, Παναγιώτης; Schoinas, Panagiotis

dc.contributor.author	Σχοινάς, Παναγιώτης	el
dc.contributor.author	Schoinas, Panagiotis	en
dc.date.accessioned	2022-10-13T08:12:26Z
dc.date.available	2022-10-13T08:12:26Z
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/55908
dc.identifier.uri	http://dx.doi.org/10.26240/heal.ntua.23606
dc.rights	Default License
dc.subject	Aeroelastic	el
dc.subject	Stability	el
dc.subject	Certification	el
dc.subject	Wind	el
dc.subject	Turbine	el
dc.title	Aeroelastic stability analysis and certification of wind turbine blades	en
dc.contributor.department	Τομέας Ρευστών, Εργαστήριο αεροδυναμικής	el
heal.type	doctoralThesis
heal.secondaryTitle	Αεροελαστική ανάλυση ανεμογεννητριών και πιστοποίηση	el
heal.classification	Δομική Μηχανική	el
heal.language	el
heal.access	free
heal.recordProvider	ntua	el
heal.publicationDate	2022-04-15
heal.abstract	The present thesis is related with the key issue of wind turbines aeroelastic stability and the modern research developments of the certification processes analyzed to IEC certification code. The first chapters deal with certification issues of wind turbines. The certification process is essential for the designers, especially the evaluation of fifty years design load base target. W/T rotor blades are large composite structures operating in a completely stochastic environment. Hence, the applied wind loads and further the developed stress resultants in the rotor blade sections are stochastic themselves. Moreover, stochastic behaviour is also exhibited by composite materials showing great scatter both in their fatigue and static mechanical properties. A rational way to quantify the variability in the basic variables and take into account these uncertainties in the final design of the structure is provided by probabilistic methods. Towards this, it was used as data input for the stochastic methodology an already known database of experimental data for the evaluation of composite mechanical properties. The stochastic models of composite material properties reproduce the statistical uncertainty of the blade beam properties, which resulted from the heterogeneity of composite materials. In terms of wind inflow, its stochasticness is reproduced with 10 minute aeroelastic simulations. The simulations represent the loading over the whole structure with Kaimal wind spectrum. This spectrum is calculated with the relevant simulation software INWIND which was developed in the laboratory of aerodynamics. Concerning the extreme loading, the long-term probability distribution for the extreme load is evaluated using load extrapolation technics according to IEC 61400-1 certification code. After the introduction of the 3rd edition of the IEC Standard 61400-1, designers of wind turbines are now required, in one of the prescribed load cases, to use statistical extrapolation techniques to determine nominal design loads. For the present thesis, a series of data simulation made for the NREL 5MW turbine, in order to compare the performance of several alternative techniques for statistical extrapolation of rotor loads. The methods are the GM and the POT method. Using each one of those, fifty-year return loads are estimated for the selected wind turbine. Two methods for extracting maximum values from time series and three cumulative distribution functions (CDFs) to these maxima data are analysed and compared between each other, in order to find out which is the correct choice for collecting data and which CDF is the appropriate one to extrapolate data gathered. Also, a convergence analysis has been made for the evaluation of the extrapolation method and the necessary number of aeroelastic time series in particular. This study was made directly to the long term distribution of the extreme values. To the final CDF of the extreme values, the statistical uncertainty due to the limited number of aeroelastic simulations was accounted for. Also, the different alternative techniques of data collection and statistical extrapolation methods for the rotor loads prediction were compared. These methods are: the method of collecting one maximum value from the whole time simulation and the method of selecting all values above a threshold. Finally at the end of the process, the 50 year design load value is estimated. Also, the selection of parametric distribution used for fitting is analyzed. Firstly, the prediction of extreme loads under turbulent wind input is presented. Then the uncertainty for the composite 26 material properties used in the blade construction is introduced. So, initially the extreme loads are calculated for the case of fixed composite material properties, and then similar estimates are obtained for lognormally distributed material properties. The properties considered are: the E1, the E2, the G12 and the ν12. The E1 is the tensile modulus of elasticity along the fibers of the composite. The E2 is the tensile modulus of elasticity vertical to the fibers. The G12 is the shear modulus and finally the ν12 is the Poisson ratio. In both sets of estimated extreme loading the wind is turbulent. From these separate estimates, conclusions are made regarding what is the effect of the material properties on the design load estimations and the stress analysis of a blade section with statistical extrapolations. A statistical code has been also developed with the commercial software Matlab, for the IEC code of W/T certification and especially for Annex F. Annex F refers to the extreme design load forecast. While this study makes use of aero-elastic simulations data in addressing statistical load extrapolation issues, the findings should also be useful in other ways. For example, the results are useful in similar questions regarding extrapolation techniques, distribution choices, and the amount of data that are needed. In any case, for the needs of the certification process, it is important to have a fast and precise code in order to have as many as possible aeroelastic calculations under realistic computation cost. For these needs a reduced order model has been developed for the simulation of the dynamic response of a W/T with twenty-two (22) DOFs in total. The formulation of the dynamic equations of the problem is based on the Hamilton’s theorem. A simulation code was also programmed for the dynamic response and the analysis of a floating wind turbine, based on the aforementioned model, which was verified with the results from the finite element analysis code hGAST. Similarly, a code was developed to simulate the aeroelastic behavior and taking into account the effect of aeroelastic deflections to the aerodynamic loads of the rotor. For the system of equations, aeroelastic stability analysis was made with Coleman’s transformation, in order to eliminate the periodic terms. Coleman’s transformation is used to enable extraction of modal frequencies, damping, and periodic mode shapes of a rotating W/T by describing the rotor DOFs in the inertial frame. The Coleman approach is valid only for a homogeneous system. Disparate systems, e.g. an unbalanced rotor, are treated with the general approach of Floquet analysis. Floquet does not provide a unique reference frame for observing the modal frequencies, to which any multiple of the rotor speed can be added. This indeterminacy is resolved by requiring the periodic mode shape to be as constant as possible in the inertial frame. The modal frequency is thus identified as the dominant frequency in the response of a pure excitation of the mode is observed in the inertial frame. The corresponding code and the Floquet method were developed for the stability analysis of balanced and unbalanced W/T systems. A separate routine was programmed for the eigenvalue identification of the system. The effect of the periodic terms on the stability of the wind turbine was examined assuming mass difference for the blades, wind yaw etc. The tool was validated against system identification with results from the hGAST FEM tool for the NREL 5MW wind turbine. Concerning the eigenvalue identification process, in homogeneous conditions the periodic mode shape contains up to three harmonic components, but in disparate conditions it can contain an infinite number of harmonic components with frequencies that are multiples of the rotor speed. These harmonics appear in calculated frequency responses of the turbine. In order to identify the 27 right eigenvalues from all multiples of the rotor speed, the appropriate identification method has been implemented. So, the specific ROM can be used for fast aeroelastic calculations in order the design – certification process to be as fast as possible for the cases that the model is accurate and covers important part of the aeroelastic calculations.	en
heal.sponsor	ΙΚΥ	el
heal.advisorName	Βουτσινάς, Σπυρίδων	el
heal.committeeMemberName	Ριζιώτης, Βασίλης
heal.committeeMemberName	Φιλιππίδης, Θεόδωρος
heal.committeeMemberName	Μπούρης, Δημήτριος
heal.committeeMemberName	Παπαδάκης, Γεώργιος
heal.committeeMemberName	Σαραβάνος, Δημήτριος
heal.committeeMemberName	Σπηλιόπουλος, Κωνσταντίνος
heal.academicPublisher	Εθνικό Μετσόβιο Πολυτεχνείο. Σχολή Μηχανολόγων Μηχανικών	el
heal.academicPublisherID	ntua
heal.numberOfPages	176
heal.fullTextAvailability	false