IEEE Transactions on Power Apparatus and Systems, Vol. PAS-103, No.

2, February 1984 407

TELEVISION INTERFERENCE DUE TO ELECTROMAGNETIC SCATTERING BY THE

MOD-2 WIND TURBINE GENERATORS

K. H. Cavcey, Member IEEE L. Y. Lee, Member IEEE, M. A. Reynolds, Member IEEE

University of Missouri Bonneville Power Administration
Columbia, Missouri Portland, Oregon

Abstract - The characterization and identifi- demonstration of various wind turbine facilities to
cation of electromagnetic interference created by determine the feasibility of such systems. As part
the slowly rotating blade of a MOD-2 wind turbine of this continuing effort, the Bonneville Power
generator (WTG) is investigated. Field measurement Administration (BPA) in 1979 became the host agency
data supports the thesis that at ground level the for a three unit cluster of MOD-2 machines with a
near field scattered component is an amplitude total generation capability of 7.5 megawatts. These
modulated secondary signal. A prediction machines have a 91.5 m rotor blade on a 61 m tower.
methodology to define zones of interference to It is designed to operate between wind speeds of 14
television broadcast reception is presented. to 56 kilometers per hour.
The site selected for these wind turbines is
INTRODUCTION Goodnoe Hills, near Goldendale, Washington,
approximately 150 kilometers east north east of
Energy research over the past few years has Portland, Oregon, on the north side of the Columbia
turned to wind as a viable alternate energy source. River. This site was selected because of its
In the mid-70's the National Aeronautic and Space productive average annual wind speed. Figure 1
Administration (NASA) and the Department of Energy shows these wind machines at the site.
(DOE) actively funded the research, development, and

Figure 1 - DOE/BPA MOD-2 Wind Turbine Generators at Goodnoe Hills, Washington

In the vicinity of this site, there was a VHF
television transponder station owned and operated by
a common carrier. These transponders were used to
relay television broadcast channels 2, 6, 8, 10, and
12 from Portland, Oregon to rural communities in
eastern Washington and Oregon. The quad Yagi
83 SM 461-1 A paper recommended and approved antennas for these transponders were at a distance
by the IEEE Transmission and Distribution Committee of 520 meters from wind turbine No. 1. Figure 2
of the IEEE Power Engineering Society for presenta- shows an aerial map of this site.
tion at the IEEE/PES 1983 Summer Meeting,
Los Angeles, California, July 17-22, 1983. Manu-
script submitted February 4, 1983; made available
for printing May 2, 1983.

U.S. Government work not protected by U.S. copyright.

 Wind Turbine Generator Site
As reported in [1], [2], and [3], the metallic
rotating blades of these large machines generate
electromagnetic interference. Of immediate concern
Tdl -Jkr I 12J - 1 ___
to BPA was the possibility of interference to the
] sinO0
+ j2 r3
nearby VHF television transponders. In addition to
EO =

4, -7e
X +r- (4)
the transponders, there are farm homes nearby, most -g 41T I Xr 21]sin02 (5)
+
of which receive television = Id J 2,

broadcast with the e-jkr

assistance of outside antennas. The closest home
is more than 2 kilometers from this site. Hence, it
became important to predict and characterize the Where k = wave number = 2T/X
interference expected from these WTG units. This
may become an important siting criterium when X = wavelength of the incident radiation
commer- cialization of large wind
turbines becomes feasible. = characteristic impedance of the medium

THEORETICAL FORMULATION r = distance from the source

As the construction proceeded, efforts were made Idl = incremental dipole moment
to predict the scattering phenomena through the use
0 = angle between the z axis and the distance
of electromagnetic field theory and modeling
experiments. This problem may be analyzed by first vector using spherical coordinates
assuming that the incident far field radiation = angle between the x axis and the distance
(television broadcast signal) sets up in the
surface vector projected onto the x-y plane

of the wind turbine rotor a current

density JB. j =f
This current density leads one to view the total
blade surface facing the source as multiple sets All field components are assumed
of reradiating Hertzian dipole sources SIB. The to h.ave a
current for each source would be: carrier frequency of wo, so
61B = ds (1) the term e iWo is
~ tB implied and suppressed.
B= 2 (n x HB) (2) If the TB and EB result in creating
new
where HB incident magnetic component
= fields ER and HR received at a new point R,
n =unit vector normal to the surface the theory of reciprocity gives:

For a Hertzian dipole, the field

component:s in spherical coordinates, are: fjvIvEBBTRJR dv=v-=1gJRJBdER dv (6)
ER=Idi -jkr[2 +
X ] (3) To greatly simplify the formulation, the blade
cos
=-2 7
[r2 j 2:r3
geometry is assumed to be that of a flat
plate with a very high aspect ratio. Figure 3
shows the wind turbine blade as a
reradiating surface.
 409

Where 1 r
BLADE
= r12 j2 r1I

= wavelength of incident radiation

r1 = distance between phase center of blade
to received point R
r2 = distance between phase center of
blade to incremental source IR

n = characteristic impedance of the

RECEIVED medium n= unit vector normal to ds

POIN
T a= unit vector from the blade phase center
Fig. 3 - Wind Turbine Blade as a reradiating surface directed toward the received point R

Since the point R is located near the major If the f unction in (11) is integrated
source of interference, the near field components over its entire surface the following results:
are used to determine the electromagnetic
scattering. Substituting Idl with the dipole E Ti LW -jkr1 (-sin a)
moment Pe . Equations (3) and (5) become: Tr e x sinc
ER=PIR -
x sinc (L sinpcosa)
a12) (12)
-jkr12r
(p
( Pe x x
(7)
R 3 12
T L_ -j27 Where L = length of blade
+ 3J(Pex12)
1 jkr 12F (8) W = width of blade
Lr~~~1 1

Eo intensity of incident radiation at

the phase center of the blade
a = altitude angle between R, the
It is important to note that the term 1/i received point near the ground
the near field component. This is the and the blade phase center
major s of interference within a 1-
kilometer radius o P
= projected angle between the
machines. normal to the blade phase center
and that to the received point R
As seen from Figure 3, the contributio
sifl TTX
ER-bIB is by the use of the divergence theorem: slnc x =
TVX

The expression in (12) is that of two crossed

EB x HR - BR x H 3) dv dipoles but does not take into account incident
radiation which is not normal to the blade or the
fy (33B x xHR3RHxH (-in ds) rotation of the blade.
-JvRRJdvJR)vf EBJJRdvd Now consider the general case where incident
radiation scatters in an oblique manner of-f of the
v blade, as shown here in Figure 4:
Replacement of the source 6IB
from ds
results in zero for the first of two integrals BLADE
above for a very large radius. As a result: (PHASE
0 CENTER
8
f (3Bx 3ER x HB ) en ds)

fRdvB
BLADE
NORMAL
Solving f or an increment of ER at R yields:

bER
4, -jkr1 [ (-n x HB) (ni y
RECEIVED
POINT
r e 1a1 x [7al x
R-2Tr 9,i (11i ) Figure 4 - Electromagnetic Scattering by
ds Wind Turbine Blade.
x e-jkr2a
410 classes of signal levels. BPA has also e
stablished
If Ae is used to represent the effective three additional signal strength grades
scattering area of the blade seen by a receiver at [4]. The FCC and BPA signal grades in
point R, use of the double angle formula results in: dBuV/m are:

=ILW [1 + cos
2

Ae (13)

This would then describe a

region of uniform
interference level around the MOD-2 machine at hub
height.
Returning to equation (12), for simplification,
only the first sinc function is retained since it
dominates over the second. This will aid in numer-
ical computation and comparison to field data.
Assuming that a television receiver located at
point R cannot have an ideal omnidirectional
antenna, the final expression for the scattered
electric f ield component ER, using equations (12)
and (13)

E
is:
rAe1+ 'q +
-jkr
E[K0OL 71r2
2Trr3)e13
j-
x sinc ( L
sina cos - sin wct ) (14)
Where K = ratio of the signal received
directly at R to that which scatters
off of the machine hub or phase
center.
= rotational rate of the blade
which is 17.5 rpm, (35 'r ).

4/2 = half of angle between the

wind direction and the vector to the
received point R from the phase
center of the blade.
Thus, the interpretation of ER is that there
is, in the specular region, a near field component,
which is a sinc function pulse, that amplitude
modulates the video portion of the received
television signal. The pulse, which occurs twice for
each blade revolution, interfaces with the AGC of
television receivers and if strong enough, may cause
interferende to the video receptions.

FIELD MEASUREMENTS
Actual field strength measurements at the
wind turbine site were made with BPA Division of
laboratories personnel and its EMI van. The
vehicle is equipped with a pneumatically controlled
10 meters telescopic antenna mast and 120 volt 60 Hz
inverters for instrumentation power. Field
strength
measurement were taken using a Singer
Stoddart NM
37/57 Field Intensity Meter and the Singer 94455-1
biconical antenna. The output of the NM 37/57 was
then fed into an X-Y plotter. Actual received tele-
vision pictures were also recorded with the use of a
video recorder and television monitor.
The instrumentation described above conforms to
measurement requirement f or EMI testing as contained
in ANSI Standard C63.3. This standard
governs
measurements made in North America f or power-line
and electrical transmission facilities interference.
In addition, the United States FCC Rules and
Regulations quantifies specific field
strength
contours based upon average terrain and the absence
of external interference. The contours define
two
 2 3
TIME (SEC.)

Channels Channels
2-6 7-13
FCC Grade A 68 71
FCC Grade B 47 56
BPA Grade C 34-46 42
BPA Grade D (Fringe) 20-33 33-41
BPA Grade E (Far Fringe) Below 20 Below 33
It should be noted that due to the distance to
the signal sources (Portland) measurements at the
Goodnoe Hills wind turbine site were for the most
part low grade C through grade E.
The Eo in (12), the average incident field
strength at wind turbine rotor hub height, was
determined by attaching a calibrated biconical
antenna to the elevator of a nearby (190
m. away)
meterological tower. Field strength measurements in
dBuv/m were then taken on television channels 2, 6,
8, and 12 as a function of height. Af ter antenna
factors and cable loss were accounted for, an average
Eo at the hub (61 m.) was determined for a
normal reflection.
On the day that f ield measurements were taken,
the wind was blowing in such a manner as to cause
the turbines to point at 2700 magnetic. Since
Portland is approximately 2610, the 0/2
angle would
be 90. Measurements at different locations in front
of one machine did indicate a maximum f ield strength
at 2790 (2700 + /2) . Data taken and averaged for
various locations was solved f or Ae, the effective
scattering area of the blade. This Ae was calculated
to be 140.6 meters2, which compares favorably to
that of a model measured in an anechoic test chamber
at 3.034 GHz by the University of
Michigan Radiation Laboratory (140 meters2) [6].
Other observations of
interest were that the
interfering signals amplitude was frequency and
height dependent. With a fixed antenna height and
location, lower frequencies (longer wavelengths)
were always stronger. Secondly, at
a fixed fre-
quency and at close distances to the wind turbine,
lower antenna heights produced higher signal level
reading. The first observation indicated that as X
increased, ER increased. That, coupled with the
second observation, seemed to verify the thesis of
near field scattering effect, which in close to the
source, is composed of sky and ground waves which
are wavelength dependent. This is true only if the
reflector (wind turbine blade) is large compared to
the wavelength of incident radiation.
At this test site, channel 2 had the most
workable ambient signal level (30-35 dBuv/m) and was
used exclusively. Solution to Equation (14) for
Channel 2 video frequency (55.25 MHz) was programmed
into an Hewlett Packard-85 computer and the function
was plotted and compared to field data. Figure 5
shows both the measured and theoretical results.
37 channel 2 video
37 - 5 -_ E-j

33 _

32
 411

38
35
3S

ia THEORETICAL
a 120 LOCUS OF POINTS
WHERE:
a

27
/0 ERS- 0.5 dB

* I 2 3 4 5 0
00l 929.6 m

ThE
m:) SiW 5z/ L A + INDICATES WIND
Figure 5 Specular Region Video Carrier AM
- E TURBINE BLADE
Interference for Channel 2, 30 t
/
Measured (a) versus calculated (b) In

As reported in [1] and [3], the incoming signal

added to the scattered one produces the total
received signal in the specular region. The
signal received can be perceived as that being Fig. 6 - Calculated Region of Interference
modulated by With 0.5 dB Modulation as
the scatterer (wind turbine blade) at a Threshold of Interference
specific modulation index mo The
modulation, Srm, in dB, is given by: CONCLUSIONS
Electromagnetic interference prediction by means
1
2log1
+ m (1 5) of scattering phenomena has been examined and veri-
m m
fied both experimentally and theoretically. The
-

0
on
major observed effect to television reception is pulse
As reported in [3], this mo is based amplitude modulation (PAM) to the video por-
subjective judgement. It also changes with the tion of the received
ambient signal level. The observable signal. The PAM interferes
effect of a with the horizontal and vertical
marginally scattered signal is a slight blinking of synchronization,
the picture due to interruption or level change in producing a brightened image that can be annoying to
the television AGC circuits. the viewers.
With ambient levels in of 60 dBuv/m, the excess This interference is maximum on a line from the
MO was reported to vary from .15 to .35. In the machine out radially at an angle twice that of between
Goodnoe Hills area, the the incoming radiation and the blade normal line in
ambient broadcast signal the specular region. For MOD-2 machines at
varied between 20 to 40 dBuv/m depending on propa- Goodnoe Hills, it is desireable for TV antennas to
gation conditions and physical locations. Since the be at least 1000 meters f rom a machine. A direc-
ambient levels were significantly lower, it was tional antenna with an unidirectional pattern and
expected to result in a lower perception threshold high front to back ratio will greatly
for mo . This was found to be true as it was minimize the possibility of interference
observed that a 0.5 dB signal modulation was due to backward scattering.
perceptible. The modulation index calculated Statistical data taken from the site
based on this observation is 0.03. on wind
Using the 0.5 dB modulation level as television speed and wind direction and machine operation
interf erenc e threshold, it is interesting to calc u- indicated that the probability of a correct inter-
late the maximum extension of the cardioid or fering angle to the television transponder site
specular region for the worst case wind and radia- occurs about 200 hours/year. Even though
tion angles. Using equations (14) and (15), this this is a
calculates to approximately 930 meters. Sinc e the verysmall percentage of the time (2.3 percent), the
field conditions were not ideal to verify this number of viewers served by these transponders
number (wind, radiation angle, and terrain) only an justified relocation of the transponders to remote
approximation at the 40/2 = 90 was attempted. sites by utilizing microwave links.
Movement of the EMI vehicle (point R) to a distance
of 1050 meters produced the mo of .03. Figure 6 ACKNOWLEDGEMENT
shows a calcuated region of interference based on
field measurements of perceptible video interference The authors wish to acknowledge the guidance of
at 0.5 dB modulation. Mr. Donald J. Marihart, Chief of the Communications
Systems Branch, Division of Substation and Control
Engineering, and Mr. Philip E. Renner, Chief of the
Electrical Investigations Branch, Division of
Iaboratories, Bonneville Power Administration, to
which the success of this study was dependent.
Their experienced advice, counsel, and
encouragement are greatly appreciated.
412

REFERENCES

1. T. B. A. Senior, D. L. Sengupta and J. E.

Ferris, "TV and FM Interference by Windmills"
Final Report on Contract No. E-(11-1)-2846,
Energy Research and Development
Administration, Washington, D.C., 20001,
February 1977 (DOE/TIC-11348).

2. D. L. Sengupta, T.B.A. Senior and J. E. Ferris,

Television Interference Tests on Block Island,
R.I., "Technical Report on Contract No.
EY-76-S-02-2846, A004, Wind Systems
Branch, Department of Energy,
Washington, D.C. 20545, January 1980.
3. D. L. Sengupta and T. B. A. Senior, "Electro-
magnetic Interference to Television Reception Caused
by Horizontal Axis Windmills," Proc. IEEE, Vol. 67,
No. 8, pp. 1133-1142, August 1979.

4. IEEE Radio Noise and Corona Subcommitte Report,

"Review of Technical Considerations on Limits
to Interference from Power Lines and Stations,"
Paper F78 165-3, IEEE Winter Power Meeting,
January 1978, New York, New York.

5. T. B . A. Senio r an d D . L. Sengupta, " Larg e Wind

Turbine Siting Handbook: Television Interference
Assessment", Technical Report No. 5 on Contract
EY-76-5-02-2846, A004, Department of
Energy, Division of Solar Energy, Wind Systems
Branch, Washington, D.C. , April 1981.

6. K. H. Cavcey, "MOD-2 Television Interference

Study at Goodnoe Hills, Washington - Final
Report - on W.O. #846-034", Bonneville Power
Administration, Division of Substation and
Control Engineering, Communications Systems
Branch, Portland, Oregon 97208, July 30, 1982.

7. D. L. Sengupta and T.B.A. Senior, "Wind Turbine

Interference to Television Reception", Final
Report No. 4 on Contract No. EY-T6-S-02-2846,
A004, Department of Energy, Division of Solar
Energy, Wind Systems Branch,
Washington, D.C., September 1981.
8. L. Y. Lee, "Broadcast TV Signal Strength
Measurements and Wind Turbine Rotor Reflected
Signal Interference Investigations at Goodnoe
Hills. " Bonneville Power Administration,
Division of Laboratories, Report ERJ 82-57,
September 17, 1982.