ALBERT

Description: With the current trend toward larger and larger horizontal axis wind turbines, classical flutter is becoming a more critical issue. Recent studies have indicated that for a single blade turning in still air the flutter speed for a modern 35 m blade occurs at approximately twice its operating speed (2 per rev), whereas for smaller blades (5–9 m), both modern and early designs, the flutter speeds are in the range of 3.5–6 per rev. Scaling studies demonstrate that the per rev flutter speed should not change with scale. Thus, design requirements that change with increasing blade size are producing the concurrent reduction in per rev flutter speeds. In comparison with an early small blade design (5 m blade), flutter computations indicate that the non rotating modes which combine to create the flutter mode change as the blade becomes larger (i.e., for the larger blade the second flapwise mode, as opposed to the first flapwise mode for the smaller blade, combines with the first torsional mode to produce the flutter mode). For the more modern smaller blade design (9 m blade), results show that the non rotating modes that couple are similar to those of the larger blade. For the wings of fixed-wing aircraft, it is common knowledge that judicious selection of certain design parameters can increase the airspeed associated with the onset of flutter. Two parameters, the chordwise location of the center of mass and the ratio of the flapwise natural frequency to the torsional natural frequency, are especially significant. In this paper studies are performed to determine the sensitivity of the per rev flutter speed to these parameters for a 35 m wind turbine blade. Additional studies are performed to determine which structural characteristics of the blade are most significant in explaining the previously mentioned per rev flutter speed differences. As a point of interest, flutter results are also reported for two recently designed 9 m twist/coupled blades.

Description: This paper describes a computer program developed for structural dynamic analysis of horizontal axis wind turbines (HAWT's). It is based on the finite element method through its reliance on NASTRAN for the development of mass, stiffness, and damping matrices of the tower end rotor, which are treated in NASTRAN as separate structures. The tower is modeled in a stationary frame and the rotor in one rotating at a constant angular velocity. The two structures are subsequently joined together (external to NASTRAN) using a time-dependent transformation consistent with the hub configuration. Aerodynamic loads are computed with an established flow model based on strip theory. Aeroelastic effects are included by incorporating the local velocity and twisting deformation of the blade in the load computation. The turbulent nature of the wind, both in space and time, is modeled by adding in stochastic wind increments. The resulting equations of motion are solved in the time domain using the implicit Newmark-Beta integrator. Preliminary comparisons with data from the Boeing/NASA MOD2 HAWT indicate that the code is capable of accurately and efficiently predicting the response of HAWT's driven by turbulent winds.