Environmental Engineering Reference
In-Depth Information
Step-Relaxation
Step-relaxation relies of the quick release of strain energy in a pretensioned cable for
the excitation of the structure. A small device applying a force of approximately 50 lb was
used successfully on the 2-m VAWT [Carne
et al
. 1982]. A heated wire was used to burn
through a nylon cord to affect the quick release of the excitation force. For a modal test of
the 64-m diameter
Eolé
(Fig. 3-36) two separate force-application points were used, one on
the central column and one on a blade [Carne
et al.
1988]. Forces were applied through high-
strength steel cables loaded with a diesel-powered winch and restrained by a nylon strap.
Loads of 135 kN (30,000 lb) and 45 kN (10,000 lb) were applied to the column and blade,
respectively.
One difficulty in using step relaxation is related to the FFT algorithm. A quick release
basically applies a force which is a
Heaviside
function, for which there is no Fourier trans-
form. However, if the Heaviside function is passed through a high-pass filter, it can be con-
verted to a well-behaved function that is easily transformable with the FFT.
Natural Wind
Using the natural wind to excite the turbine has certain advantages, as discussed in
[Carne
et al.
1988]. Two of the most significant are the greatly reduced cost and complexity
of the test and the ability to test during windy conditions. Wind-excitation methods were
developed during test of the Eolé VAWT, and these produced modal frequencies in excellent
agreement with those obtained by step relaxation. However, damping information is not as
readily available from the power spectra obtained using wind excitation. Development of
the
natural excitation technique
(NExT) rectified this situation and will be discussed below
[James
et al
. 1995].
Measured Rotating System Modes of the 2-m VAWT
The results of the modal testing performed on the 2-m VAWT using step-relaxation
excitation are shown by the data points in Figure 11-9. The correlation here is typical of
wind turbine modal analysis. The average absolute deviation of the experimental frequencies
from the theoretical (tuned-model) frequencies is 0.5 Hz, or 2.2 percent. Deviations gener-
ally increase with increasing rotor speed, which is a characteristic of finite-element models
that do not have a sufficient number of degrees of freedom to accurately represent complex
mode shapes.
Search WWH ::
Custom Search