Bestemmelse af ekstrem respons og svigtsandsynligheder for vindmøller under normal drift ved brug af kontrolleret Monte Carlo simulering

Sørensen, Søren ; Nissen, Jonas ; Jørgensen, Jacob

4. term

Structural and Civil Engineering, Master

2009

2009-06-15

192 pages

The purpose of the thesis is to test alternative methods for estimating the extreme response for wind turbines under normal operation. Currently, different extrapolation methods are used for determining the extreme responses by extracting maxima from a few simulations of normally 10 minutes. These extrapolation methods are associated with unknown statistical uncertainties. In the first part of the thesis, the extrapolation method suggested by the IEC-61400 standard is described and the associated uncertainties and problems are clarified and the extrapolation method is tested for determining the extreme response for a wind turbine model. Afterwards two simple structural systems are presented, a simple shear frame and a simple wind turbine model, which will be the basis for evaluation of the alternative methods presented. The alternative methods Russian Roulette and Splitting with Distance Control (RR&S), Importance Sampling (IS) and RESTART are presented as alternatives to the present extrapolation methods. The latter of the three does not provide promising results, whereas RR&S and IS provide promising results on both of the simple structural systems reducing the computational time significantly. Finally, the RR&S algorithm is implemented on the aero-elastic code FAST. The algorithm provides promising results for estimating the extreme responses if disabling the wind turbine control system and no significant improvements if the control system is enabled.

Formålet med projektet er at afprøve alternative metoder til bestemmelse af ekstremlaster for vindmøller under normal drift. I øjeblikket anvendes ekstrapoleringsmetoder til bestemmelse af ekstremlaster ud fra få maxima fra et par simuleringer af normalt 10 minutter. Disse ekstrapoleringsmetoder er forbundet med ukendte statistiske usikkerheder. I den første del af projektet, beskrives ekstrapoleringsmetoden foreslået af IEC-61400-standarden og de dermed forbundne usikkerheder og problemer er belyst. Metoden er benyttet til bestemmelse af ekstremlaster for en reference vindmølle. Derefter beskrives to simple strukturelle systemer; en simpel forskydningsramme og en simpel vindmøllemodel, som vil være grundlaget for evalueringen af de alternative metoder. De alternative metoder Russisk Roulette and Splitting with Distance Control (RR&S), Importance Sampling (IS) og RESTART der fremlægges som alternativer til den nuværende ekstrapoleringsmetode. Sidstnævnte af de tre giver ikke lovende resultater, hvorimod RR&S og IS giver lovende resultater på begge simple strukturelle systemer, hvor beregningstiden reduceres betydeligt. Endelig er RR&S algoritmen implementeret på den aeroelastiske kode FAST. Algoritmen giver lovende resultater ved bestemmelse af ekstremlaster, hvis vindmøllens kontrolsystemer deaktiveret, hvorimod algoritmen ikke forbedrer beregningstiden hvis kontrolsystemet er aktiveret.

wind turbine ; monte carlo ; importance sampling ; controlled monte carlo ; extreme response / load ; russian roulette and splitting ; FAST ; oscillating systems ; Sample control function ; FORM ; variance reduction ; vindmølle ; kontrolleret monte carlo ; ekstrem respons / laster ; russisk roulette og splitting ; oscillerende systemer ; kontrolfunktion ; variansreduktion

Download
View record in AAU Student Projects