Sway Measurement Overhead Power Line by an ITV-based Camera
Member Ryuichi Ishino (CRIEPI,ishino@criepi.denken.or.jp)
Non-member Tomomi Ishikawa (CRIEPI,iskw@criepi.denken.or.jp)
Keywords: image processing, overhead power lines, frequency response analysis for wind, wind engineering, ITV-based camera
In S eptember 1991 in Japa n, a typhoon destroye d many which is at 85,170,255m away from the position of ITV-based
powe r transmission lines and tower s. S inc e then, we have camera, to expe rimental over head powe r line and we
studied the ef fe cts of loca l winds on such fa cilitie s. The se compa red the powe r spec trum of ca ble displa ce ment betwe en
studies include wind-tunnel tests, numerical analysis, monitors our method and a theor etical ana lysis that is base d on
for var ious t opogra phie s, and in orde r to improve the design Da venport’s gust re sponse theor y and conside red the
method for overhead transmission lines. geometric stiffness for cable displacement by an average wind
When ana lyzing cable re sponse to wind, we use data fr om load. As shown i n F ig. 2, the pea k fr eque ncy obtained by the
wind veloc ity meter s. To ver ify whe the r the simula tion model image mea sureme nt method matche s the simula tion at ea ch
is valid, we compa re the ana lysis for ce to the data fr om a point i n the low-f reque ncy domain whe re a ca ble eige n-
tension s ensor and eige n-fr eque ncy to the data fr om an frequency exists.
accelerometer. However, we cannot obtain data on the average
cable di spla ce ment. B ec ause its initia l point is unknown. To Image measurement
check the simula tion model in deta il, a method for mea suring Theoretical analysis
cable sway displacement is needed. sec)
In orde r to obtain the ca ble swa y displa ce ment, We 10 5
2
developed a remote sensing method for measuring the position
1000
of a swa ying ca ble by ITV -video proc essing. Our method is
power spectrum(cm
that some targets on the cable are tracked in videotaped images. 10
Its initial position is defined as the cable position in no wind. In
order t o che ck the diffe re nce betwe en the re al displa ce ment 0.1
and measured displa ce ment, we compa red the displa ce ment 0.001
betwe en our method and lase r mea sureme nt of the ca ble swa y 0.001 0.01 0.1 1 10
frequency(Hz)
at 85, 170,255m aw ay for m ITV -base d ca mera . As shown in
F ig.1, the diffe re nce betwe en them is about 10cm at eve n
furthest point 255m. F ig.2 C ompa rison of powe r spec trums for displa ce ment
betwe en i mage mea sure ment method and theor etical ana lysis
12 at 85m in the horizontal direction
Horizontal direction
Vertical direction
10
S ince some of dominant pea ks at thre e positions ar e
Difference of measurement (cm)
8
obtained, we could ana lyze the ca ble swa y mode by
6
considering difference of phase of each positions. As shown in
Fig. 3, we can also know how the cable sway.
4
2 The first mode (0.115Hz)
The second mode (0.295Hz)
0
1
50 100 150 200 250 300
Arbitary unit
Target position(m)
0
Fig.1. Difference between the image measurement and laser
measurement methods
-1
Our m ethod gives not only the displa ce ment but also
0 50 100 150 200 250 300 350
fr equency response of t he ca ble for wind. Whe n we ana lyze Distance from tower 1(m)
how t he ca ble swa ys, the informa tion is ver y usef ul. S o we
analyzed the ca ble swa y by videotape d data in a typhoon and F ig 3. S wa y m od es i n h or iz o nt al d ir ec t io n f ro m
verified our method. For verification, we attached three targets, m ea su r em en t r es ul t