969

There appear to have been only two published investigations of power The curve for Farquharson and McHugh's investigation represents the
input to cylinders or cables vibrating in sine-shaped loops. One was reported average of results obtained with six different models, half of them having
by Carroll8 in 1936, the other by Adami & Batch29 in 1981. Estimates of stranded surfaces. The models were of several different diameters. We have
maximum 8r as a function of Ymax/d based on those sources are included reanalyzed the original data for the most energetic of the six models, a
in Fig. 7. smooth cylinder 4 inches in diameter identified as Model F, and those
The tests reported by Carroll were performed in a wind tunnel that results are represented by the circles in Fig 8.
apparently was built for the purpose. A 17m span of 35.6mm diameter Rigid model results such as those in Fig. 8 can be integrated to apply to
conductor under tension was exposed to a jet of air 15.2m wide and 0.5m sine-shaped loops, for comparison with the curves of Fig. 7, by the pro-
high. Vibration took place in 2 and in 3 loops in the tests leading to the cedure used by Farquharson and McHugh' 0. As noted in the section "Con-
curve of Fig. 7. Since the loop length-to-diameter ratios were quite large, ductor Motion During Aeolian Vibration", this procedure utilizes an assump-
wind tunnel end effects were probably very small. tion that is open to question, but it appears to be the only procedure
Carroll's curve should be compared with the upper sheet curve from the available. It entails expressing the rigid cylinder br (denoted brc) as a poly-
present tests. It is lower by a factor of about 2/3. nomial:
The tests reported by Adami and Batch were not described in significant
detail, but used untensioned cylindrical rods supported as pinned beams. 6rc = a0 + a, (Y/d) + a2 (Y/d)2 + a3 (Y/d)3 + . (9)
Two curves are shown in Fig. 7, one for single-loop vibration, and the other
for two-loop. The single-loop values of 6r are less than the present upper The integration involves multiplying the coefficients an by factors hn
sheet ridge by a factor that ranges from 0.85 to 1.0. The two-loop curve derived from the shape of the vibration mode, in this case a sinusoid.
is lower by a factor that ranges from about 0.55 to 1.0. Thus, it agrees
rather well with the curve based on Carroll's results.
6r for the loop (8rs) is given by:
These comparisons seem to suggest that the effective value of 6r is 6rs = bo + b (Ymax/d) + b2 (Ymax/d) + b3 (Ymax/d) + (10)
reduced when more than one loop is present. While such an effect cannot
be ruled out, there seems to be no aerodynamic justification for it, based where bn = anhn and the hn values are:
upon current understanding of the flow mechanisms involved in aeolian
vibration. A more likely explanation for the lower br values indicated by 1, 8/31r, 3/4, 32/15ir, 60/96, 192/1057r, 420/768 ---
Carroll's tests is that the flow in the jet of his tunnel contained more turbu-
lence than in the present tests. The description given of the tunnel is sug- Properly done, the integration must be applied to 6rc data obtained
gestive of this. with Vr and hence V, held constant, since V is nominally. constant over the
Lack of detailed information on the Adami and Batch tests precludes length of typical vibration loops. The integration may be carried out for
further comment on them. It must be said that, for evaluations using each of a number of values of Vr, to identify the Vr that yields the highest
models or conductors vibrating in loops, the present tests have, if anything, 6rs at each amplitude. The result of such a process should correspond to
increased the spread among results. the ridges shown in profile in Fig. 7.
In practice, however, it has been common to develop the profile of
COMPARISON WITH OTHER SOURCES - PURE TRANSLATION
the 6rc ridges for the rigid models, as illustrated in Fig. 8, and to apply
A search of published literature revealed reports of more than a dozen the integration of (10) to them. This procedure ignores the fact that differ-
investigations of power input to rigid cylinders undergoing purely transla- ent Vr values apply to different points along the ridge. Sample calculations
tory vibration. Various test arrangements and procedures were used. When have shown, however, that the errors that derive from this simplification are
the results were transformed to 5r to permit comparison, it was found that small (only a few percent).
a majority of the programs yielded roughly similar results for maximum The reverse of the process indicated in (10) was applied to the upper
br) over a significant range of Ymax/d. For the other programs, results sheet ridge from the present program, shown in Fig. 7. In other words,
differed from the above "core group" in manners suggesting that test the measured sine-loop curve was transformed to apply to rigid cylinders.
conditions were different in some respect. The similarity of many of the This was done to permit comparison of the present results with those from
results is the basis for summary curves presented by Iwan30 and by King31. the other sources shown in Fig. 8. The transformed results are represented
Curves representing maximum 6r as a function of Ymax/d are shown in in Fig. 8 by the square points. They fall about in the middle of the "core
Fig. 8 based upon eight of the core group investigations. Similarity in the group" band, correlating best with the results from Meier-Windhorst, Bate
slopes of the curves is evident, as is the dispersion in results of even this and Callow, Scruton, and Farquharson and McHugh's Model F. The de-
select group. parture from Meier-Windhorst and Bate and Callow at the highest ampli-
tudes results from the fact that only the lower sheet ridge was transformed.
In view of the correlation with those previous results, it is tempting to
conclude that there is confirmation as well. As noted earlier, however, there
is some question about the validity of the integration represented in (10)
and specifically about whether the elemental lengths of a vibration loop
are autonomous for vortex-locking and power-input purposes. Thus, the
confirmation is only as strong as the assumption of elemental autonomy.

CONCLUSIONS
B The purpose of the measurements was to narrow the uncertainty sur-
Fe> \rounding the power supplied by the wind to vibrating conductors, with
particular emphasis on steady-wind conditions, preparatory to developing
methods for dealing with turbulent winds. The degree to which this has
2
.04 06 0.1 0.2 0.4 06 1.0 2.0 been achieved must be judged in the context of the existing conflict between
YmoX/d the sine-mode tests of Fig. 7 and the rigid-body tests of Fig. 8. As can be
seen from Fig. 8, the present tests seem to correlate well with the central
Fig. 8 Comparison of present re suIts for maximum 6r with previous rigid- group of previous rigid-body tests. Thus that group would correlate about
body mode results. Prese*nt results are transformed by Eq. (10) to as poorly with the sine-mode tests of Fig. 7 as the present tests do. Sub-
apply to rigid body mo de. Key: B - Bate and Callow (7); D - ject to the questionable assumption, noted several times above concerning
Diana et al (14); Fa -F:arquharson and McHugh (10); Fe - Feng converting rigid-body data to apply to sine-mode vibration, the present
(12); K Griffin et al (116); M - Meier-Windhorst (9); Sc - Scruton
- tests support the central group of previous rigid-body tests, rather than
(11); Si Simmons and CDleary (19).
- the sine-mode tests.