Monitoring acoustic noise which is neither
corona nor Aeolian, a novel sensing method for stability in the power grid
Dr
Abstract
Acoustic noises are observed near 400KV circuits which
are not related to corona or Aeolian tones. The sensing method
adapted is to record the sound by using a laptop computer
with various microphones and an adapted FFT program with
output in a frequency -amplitude- time waterfall mode. Comparisons are
made with magnetic spectra and using different types of microphone. Mainly sub-
harmonic frequencies 8.33, 16.7,25, 33.3 and 37.5 Hz are recorded in the
acoustic spectra corresponding with only the usual higher harmonics in the
magnetic spectra. On a couple of occasions an inter-harmonic of 30 Hz
corresponding with 180 Hz in the magnetic spectrum is observed and parametric
conversion is postulated. The acoustic noise inter-harmonics generally
become more accentuated with an increased wind generation contribution to the
grid capacity. Possible theories of noise generation are advanced
including electro-seismic effect with and without
seismic feedback mechanisms. An alternative theory involving infra-structure
micro-cracks is briefly discussed. Possible space weather implications are
speculated upon.
Introduction
Further,
in this respect the present author a keen hill walker recently noticed a very
peculiar low pitched quasi -periodic humming sound whilst walking early on
morning in hills near Llanllechid Gwynedd. The pitch was estimated to
be much lower than 50 Hz and although the whole hillside seemed to
resonate with the noise its centre of activity seemed to lie in between two
high voltage power circuits on 400 KV National Grid single circuit and one 132
KV Manweb single circuit. The noise was definitely not corona as it
was not 100 Hz and there was no 'crackling' nor was it Aeolian tones as it was
not windy and the noise simply didn't have the right characteristic. The
author remembered a similar noise being reported to the
Synchronous
machines fed from harmonic sources can certainly generate unwanted and
additional acoustic noise13,14. In this respect this section of 400
KV circuit connects Deeside to the Dinorwig Pumped storage generation scheme
via the Pentir Inter-connector. There have been instabilities at Dinorwig in
the past15,16 . The author
wonders if increased power system harmonics and
inter-harmonics are once again provoking a problem at
Dinorwig. In such a respect ground borne vibrations may also be involved and
they can travel huge distances17,18
But
what possibilities are there for the overhead lines? It is well
known that high voltage circuits can emit corona noise. But this type of noise
has a frequency component which again is predominantly at 100 Hz and
sounds 'crackly'. On the other hand Aeolian tones sound just that, rather like
the whaling or howling wind or like blowing over a bottle of water and
again did not fit the bill for the noise being experienced.
The hypothesis is perhaps that the dirtier the content injected into the grid,
the more this new acoustic phenomenon will manifest. A proposed way of
testing is to try and correlate the appearance of the phenomenon with total or
else a significantly relevant portion of
Another
early morning visit to the same location where two
single circuit sections of the grid run paralleled one
400 KV owned by National Grid and one at 132 KV owned by Scottish
Power Manweb held the solution. The whole air, indeed the whole hillside
for several hundred yards was resonating with the noise. This time
however, a battery of equipment was available. Then and over the coming
days a series of experiments was performed. Some measurements were also made
under the 400 KV double circuit at Llanfair PG which links the Pentir
Inter-connector to Wylfa Nuclear Power Station.
Experimental ; Apparatus and methods
Low
frequency sound and infra sound is notoriously difficult to record using traditional
methods due at least in part to filtered frequency responses of
equipment. The approach here was to employ a laptop computer
running FFT spectral analysis software such that near real-time frequency-
amplitude -time plots could be obtained in the waterfall display mode. For
audio recording the computer electret microphone could be used or a modified
loudspeaker as a moving coil microphone and a small potted transformer was
available as an inductor for obtaining magnetic spectra. Electric field spectra
could be recoded simply by connecting an unscreened lead to the computer
microphone input. The acoustic recordings are not closely
calibrated. Traditional sound meters with dB(A) weightings under estimate low
frequencies but most of the readings are in the range 65- 90dB(A) based on a
separate acoustic spectrum and sound meter reading of moving car
infra-sound. The scales on each waterfall plot are merely relative as the
colour spectra were being adjusted to tease out the best signal to signal differentiations.
Experimental Recordings
Series of acoustic waterfall mode recordings were made under the 400 KV
single circuit at Llanllechid on three dates namely 22nd,24th
and 31st August 20011. These were supplemented by electric and magnetic recordings.
The mean magnetic field was 29 mG. A similar series of recordings
was made under the 400 KV double circuit at LlanfairPG on the same dates
allowing time to travel between the sites.
On 22nd and 24th August the noise described above could clearly be heard, on 31st August 2011 it could not. On 22nd August the weather locally was still and slightly damp. This does not of course imply that there was no wind offshore or anywhere else in the country for that matter. On 22nd August, corona noise could be heard at the Llanllechid site as a distinctly separate sound. . On 24th August the weather was cool, dry and breezy no corona noise could be heard. On 31st August the weather was still and dry, no corona noise could be heard at either site.
Results and Discussion
The
results are presented as simple printouts of the computer waterfall displays.
Figure 1 Llanllechid 400 KV single circuit 0722
The
acoustic spectrum shows signals at 16.7, 30 and 50 Hz. The
signal at 90 Hz is an
Clearly
the acoustic spectrum is showing up something which the magnetic spectrum is
not and certainly whereas the signals at 16.7 and 30 Hz appear to be quite
broad and continuous the signal at 50 Hz appears to a have a pulsating
amplitude. It is thought that this could well account for the phenomenon
heard by the author on the hillside. The observation of 30 Hz is
fascinating. It seems to coincide with an observation of 180 Hz in the magnetic
spectrum and one can speculate on a parametric conversion mechanism,
180EM -3x50EM = 30 AU. Where EM = electro magnetic and AU = audio.
In order to investigate the persistence and extent of this phenomenon
further measurements were made on subsequent days and at additionally at the LlanfairPG
site. Parametric behaviour of power systems has been discussed elsewhere19
Non- integer harmonics including 25 Hz and 180 Hz are referred to by
Jacobs et al (2002)20 and there is a specific reference to 180
Hz by Bollen.21
Figure 2 Llanllechid 400KV
Magnetic
signals present 0-50 Hz noise, 50 Hz, 100Hz, 180 Hz.
Acoustic
signals 16.7 Hz , 25 Hz, 30 Hz, 33.3 Hz, 38 Hz, 50 Hz, 90 Hz
On
this occasion the acoustic spectrum is slightly more complex than on the 22nd
August and the signals at 30 Hz and 33.3 Hz are intermittent. All the
signals appear to be pulsating more in amplitude. However the 180 Hz magnetic
signal is now continuous. The signal at 38 Hz could conceivable be the .75 sub
harmonic of 50 Hz. The other signals are as described previously except
now a signal at 33.3 Hz is also prevalent being the inter-harmonic related to
1/3 of 100 Hz which has showed up in the magnetic
spectrum.
Figure 3 Llanllechid 400KV
Magnetic
signals present 50Hz, 150Hz, 250 Hz. Acoustic signals present
16.7,25,33.3 and 50 Hz.
Surprisingly
some four minutes later the instability which gave rise to the 180Hz magnetic
signal and 30 Hz audio signal has disappeared. The magnetic signal at 100 Hz
has been replaced by the 3rd harmonic at 150 Hz but the other
acoustic signals remain stable and the 33.3. Hz signal is somewhat less
intermittent. Allowing for travelling time, measurements were next
made at Llanfair PG in an attempt to assess how widespread these effects are.
Figure 4, 24th
August LlanfairPG 400 KV Double circuit Top magnetic, bottom acoustic
At
Llanfair PG Magnetic signals present 50Hz, 150 Hz and some low frequency noise
mainly 0-70Hz.
Acoustic
signals present approx 8.5 Hz, 16.7 Hz, 25 Hz, 33.3 Hz, 50Hz, and some noise in
range 50-312Hz. The recordings here are strikingly similar to those
in figure 3 except for the absence of the 5th harmonic in the
magnetic spectrum and the presence of a signal centred on about 8.5 Hz in the
acoustic (infrasonic) spectrum which could potentially be the sixth sub
-harmonic at 8.33 Hz. The close similarity between the frequencies present now
completely rules out Aeolian tones as the mechanism of sound production because
the Llanfair PG circuit is a double circuit and would thus have completely
different wind loading and aerodynamic characteristics.
It
is noted that the acoustic sub-harmonics appear to be amplitude or pulse
modulated in some way.
This
could just be to load variations in the power system but it is interesting
also to note that power quality output form wind turbines has been shown
to contain amplitude modulation at the blade passing frequency typically
1.5 Hz for a three blade turbine and causes flicker in power syatems22
Figure 5
On
31st August, the grid is obviously behaving very differently and in
a far more stable manner than on the previous two dates. The only
acoustic signals of relevance are a very weak 25 Hz and a 50Hz signal.
The acousto-magnetic transducer recorded signals at 50 and 250 Hz and very
weak signals at 100 and 150 Hz. These signals were quite difficult to record on
account of a JCB working some 400 metres away. This did emphasise however what
a highly efficient acoustic radiator the hillside in question was.
Figure 6
The
only relevant frequencies are 25 and 50 Hz; there is considerable background
noise from other sources. There was no evidence of higher magnetic
harmonics or lower frequency acoustic inter-harmonics.
In
general the inter-harmonics are worst on the 22nd and 24th August
and virtually nil on the 31st August. Irrespective of
the mechanism for the generation of these acoustic inter harmonics which will
be discussed briefly later, what has been discovered is a very powerful and
sensitive way of monitoring the stability of power grids or sections thereof
and in that respect the author has already filed for patent protection. The
hypothesis advanced earlier is that wind generation and many other embedded low
carbon forms of generation contain inverter systems which produce
inherently dirty electricity. Ideally to test the hypothesis one
should seek out sources attached to to the nearest nodes in the grid.
Unfortunately small sources producing less than 90MW are not registered with
Exelon as separate generating units. For larger units, one can find out how
much wind power is being generated from 'bm reports' part of the Exelon
electricity trading website23. On the assumption that wind
input from more minor sources would have a similar geographical distribution
and in any event is not all accessible through data at Exelon, it was decided
to use the sum total of injected power from three major wind farms
geographically well spread across the
One of the largest inshore wind
farms in
DATE |
MW |
AUDIBLE
NOISE Y/N |
Inter-harmonics |
22ND
Aug |
88 |
Y |
Y
-large range |
24th
Aug |
119 |
Y |
Y
-large range |
31st
Aug |
34 |
N |
very
weak 25 Hz only |
According
to the above then, the hypothesis is very strongly supported by the present results.
It would seem there is a very low threshold in the section of grid examined for
disturbance by wind generation. As these farms are geographically well spread
it is fair to assume that a similar proportion of
The
mechanism of sound production itself could involve parametric conversion. At
least two separate and very different mechanisms are possible therefore further
investigation will be necessary to confirm.
Firstly,
it is known that geologists use electro-seismic prospecting and indeed the use
of seismic prospecting with power lines as a source has also been described
stating that harmonic and sub-harmonic frequencies of the power line
fundamental may be generated as seismic signals.26
Electro-seismic
conversion was first theorised on by Frenkle27 in 1944 and
later by Biot28 in 1962 and more recently by Pride29
2003 and in one sense can take place at electrical double layers at soil or
porous rock water interfaces and is therefore a commutative
process.
It
is interesting to note that one unstable mode of the power grid involving
30 Hz acoustic signals and 180 Hz magnetic signals was only noted adjacent to
the Llanllechid section of the single 400 KV circuit where the MANWEB 132KV
circuit runs nearby. It is possible that the instability
involves electro-seismic and seismo-electric electric coupling
between the two.
Another
variation on this theme might be that higher harmonics excite increased
non-linear behaviour of the synchronous motor-generator sets at Dinorwig
causing stronger earthbound seismic coupling of bearing vibration
which has been quoted by
Alternative
hypotheses could involve the cables, pylons and insulators themselves. Acoustic
waves at 50 Hz are observed in all the above cases. It is also know that
the electromagnetic component contains higher harmonics. Vibrations of these
structures can take place by Lorentz force, magnetostriction and electrostriction.
If micro-cracks and flaws are present at metallic contacts sub-harmonic
generation can take place31,32. Presumably as higher harmonics are
present there is more opportunity for sub and inter-harmonic generation by
theses mechanisms. This mechanism would be expected to give rise to
predominantly airborne vibrations. The hypothesis ought to be testable by
examining an unstable power grid at night when low frequency
atmospheric sound ought to propagate much further. Generation of the
frequencies observed by insulators or their partial breakdown can probably be
ruled out as such processes tend to give a flat frequency response from DC to
about 50 KHz and with emissions still measurable in the
radio frequency bands. Thermophone like noise generation by the cables
cannot be entirely rules out but would be expected to be strongest on a
working day when the load was greatest. This was not observed. There can
also be sound patterns similar to corona with loose connection arcing33
but these don't fit the bill of what was observed either.
Although it has been established
here that the instability in the sections of power grid monitored appear to be
due to wind generation, the technique ought not to be limited to that. In
one sense the 30 Hz phenomenon has been proven to be more of
a transient phenomenon than the rest. On one occasion, it came and went over a
four minute period. Kappenman, describes how space weather can cause
havoc on power grids. Solar storms can cause black outs on earth due to
powerful ground induced currents34.
However,
although a considerable body of literature exists by those working in the
field; it is generally far less well known that power grids can cause
space weather! So called PLHRS or power line harmonic radiation may affect
the earth's magnetosphere35-38
Thus
the author believes there is almost the possibility for the
The
observation of low frequency acoustic and infrasonic radiation from 400 KV
sections of the power grid not attributable to previously known causes and
revealed in this paper yields numerous potential opportunities. In
whatever case, it would seem an incredibly sensitive way of visualising sub-
harmonics and inter-harmonics in power systems. Sub and inter-harmonics with
period doubling have been observed which are known routes to chaos in a power
system42 ,
for a stable grid they ought then to be minimised at all costs. There can be no
such minimisation without monitoring and perhaps ultimately some kind of
feedback control. In order to expand
this method to cover more of the grid, there may be the opportunity to develop
long fibre optic based acoustic sensors for in-situ monitoring in the future. Alternatively,
several mobile phone companies are now making use of pylons for mounting of
their antennas and back haul dishes. There could be the opportunity for
co-operation in the development of wireless based monitoring
technologies along these lines.
Conclusions
· A new and sensitive method for monitoring potential instabilities on a
power grid is described.
· It is believed to be the first time that significant wide area acoustic
noise is reported and recorded from overhead 400KV circuits which is neither arc
breakdown nor corona nor Aeolian in origin.
· Sub-harmonic and inter-harmonic acoustic tones have been recorded which
seem to be involved in a parametric conversion process
with frequencies in the magnetic spectrum.
· Acoustic inter-harmonics have been
shown to become more severe when there is a
greater proportion
of wind generation on the grid.
·
Some possible hypotheses of the acoustic generation mechanism have been
advanced.
References
1. S.M. Wilkinson1 and N.A.Henderson, USE OF NETWORK ANALYSIS TO ASSESS THE VULNERABILITY OF AN ELECTRICAL SUPPLY NETWORK The 14th World Conference on Earthquake Engineering October 12-17, 2008, Beijing, China http://www.iitk.ac.in/nicee/wcee/article/14_06-0082.PDF
2.
ENSG ‘Our
Electricity Transmission Network a Vision for 2020’ URN 09/752
3. EWIS
European Wind Integration Study TREN/07/FP6EN/SO7. 70123/038509 http://www.wind-integration.eu/downloads/library/EWIS-Grid-Reinforcement-Projects.pdf
4.
Burak Ozpineci, Leon M. Tolbert Department of
Electrical and Computer Engineering University of Tennessee-Knoxville Knoxville,
TN 37996-2100 http://www.uv.es/emaset/iep00/IEP%2009-10/cycloconvertertutorial.pdf
5.
Gunther, E.W.; Electrotek Concepts Inc.,
6.
Wind Turbine Harmonics Caused by Unbalanced Grid
Currents By Hans
De Keulenaer / Published on Tue, 2008-05-06
7.
An Investigation of the Harmonic Emissions of Wind
Turbines, Sokratis T. Tentzerakis;
Issue Date: March 2007 Volume: 22 Issue:1 On page(s): 150 - 158 ISSN: 0885-8969 http://ieeexplore.ieee.org/xpl/freeabs_all.jsp?arnumber=4106019
8. Harmonic and interharmonic distortion analysis in the grid-connected wind electric generator, Kumar and Kannan, Intl.J.Energy Tecnol+Pol. 5(2) 2007 pp187-203
9. Guidance Notes Offshore Wind Farm Consents Process www.bis.gov.uk/files/file22990.pdf Youssef 2003- berr.gov.uk
10. Harmonic
Modeling of Solar Inverters and Their Interaction with the Distribution Grid Bosman,
A.J.A.; Cobben, J.F.G.; Myrzik, J.M.A.;
Kling, W.L.; Dept. of Electr.
11. Harmonic current emission of wind farms exceeding their limiting values, http://www.dewi.de/dewi/fileadmin/pdf/publications/Publikations/S14_5.pdf
12. Dr
13. Analysis of the Audible
Noise of Three-Phase Squirrel-Cage Induction Motors Supplied by Inverters
Belmans, Ronnie J. M.; D'Hondt, Lieven; Vandenput,
Andre J.; Geysen, Willy; Department of Electrical
Engineering, Katholieke Universiteit Leuven, Laboratory for Electrical Machines
and Drives, Kard. Mercierlaan 94, B3030
14. Noise
and vibration from induction machines fed from harmonic sources Yacamini, R.; Chang, S.C.; Energy
Conversion, IEEE Transactions on Issue Date: Jun 1995 Volume: 10 Issue:2
On page(s): 286 - 292
ISSN: 0885-8969 Dept. of Eng.,
15. Dewi
Jones and Sa'ad Mansoor, http://www.gwefr.co.uk/prevexp.htm
16. Reproducing
oscillatory behaviour of a hydroelectric power station by computer simulation S.P Mansoor ,
D.I Jones,
D.A Bradley, F.C Aris, G.R Jones Control
Engineering PracticeVolume
8, Issue 11, November 2000, Pages 1261-1272 http://www.sciencedirect.com/science/article/pii/S096706610000068X
17. Monochromatic
Components of the Seismic Noise in the NORSAR Area. E.Hjortenberg1, T.
Risbo2Geophysical Journal of the
18. Nature
of narrow-band signals at 2.083 Hz Götz H. R. Bokelmann* and
Stefan BaischBulletin of the Seismological Society of America; February 1999;
v. 89; no. 1; p. 156-164© 1999 Seismological
Society of America http://bssaonline.org/cgi/content/abstract/89/1/156
20. Shunt
hybrid power filter control in power systems with interharmonics Jacobs,
J.; Schroder, S.; De Doncker, R.W.; Inst.
for Power Electron. & Electr. Drives,
Print ISBN: 0-7803-7640-4
21. Bollen, Electric Power Systems Research 66(1), 2003
22. Models
for predicting flicker induced by large wind turbines Saad-Saoud,
Z.; Jenkins, N.; Centre for Electr. Energy, Univ. of
Manchester Inst. of Sci. & Technol. 790945 abstract Energy
Conversion, IEEE Transactions on Issue Date: Sep 1999 Volume: 14 Issue:3
On page(s): 743 - 748
ISSN: 0885-8969
23. http://www.bmreports.com/servlet/com.logica.neta.bwp_PanBMUData
24. On
the Effects of Subsynchronous Interharmonic Voltages on Power Transformers:
Single Phase Units Langella, R. Testa, A. Emanuel, A.E. Dipt.
di Ing. dell'Inf., Seconda Univ. di Napoli, Aversa Aversa : Power
Delivery, IEEE Transactions on
Issue Date: Oct. 2008 Volume: 23 Issue: 4
On page(s): 2480 - 2487 ISSN: 0885-8977 INSPEC Accession Number: 10235586
25. Improvement in the efficiency of
distributed power systems, G.Benysek, Bulletin of the
26. US patent 7227362
27. Frenkel
(1944) studied 'seismoelectric potentials' associated with the propagation of
longitudinal waves in moist soil. He used electrokinetic theory associated with
quasistatic Poiseuille flowin charged, cylindrical capillaries to model the
induced potentials.
28. Generalized theory of acoustic propagation in porous
dissipative media, MA Biot - J. Acoust. Soc. Am, 1962 -
pmi.ou.edu
29. The
electroseismic wave theory of FrenkelAuthor:Pride,
Steven R. Publication Date:11-10-2003Publication
30. An
investigation into the development of a condition monitoring/fault diagnostic
system for large reversible Francis type pump-turbinesS
31. Nonlinear
reflection of surface acoustic waves from contact lines B. A. Korshak and Yu. B. Korshak Technical Physics Letters Volume 34, Number 6, 486-488ov et al 2004 Abdel Salem et al 2001
32. John G. Kappenman, Storm
sudden commencement events and the associated geo-magnetically induced current
risks to ground-based systems at low-latitude and mid-latitude locations,
SPACE WEATHER,
33. A
theory of PLHR emissions to explain the weekly variation of ELF data observed
by a low-altitude satellite OA Molchanov, M Parrot, MM Mogilevsky… - Annales …,
1991 - cat.inist.fr http://cat.inist.fr/?aModele=afficheN&cpsidt=4986567
34. Daily
variations of ELF data observed by a low‐altitude
satellite, M Parrot, OA Molchanov, MM Mogilevski… - Geophysical Research …,
1991 - agu.org http://www.agu.org/journals/ABS/1991/91GL01352.shtml
35. Identifying
power line harmonic radiation from an electrical network 1, C.
J. Rodger, N.
R. Thomson http://hal-insu.archives-ouvertes.fr/hal-00317854/
36. Unintentional
man-made modification effects in the magnetosphereK Bullough,
T.R Kaiser, H.J Strangeways Journal of
Atmospheric and Terrestrial Physics
Volume
47, Issue 12, December 1985, Pages 1211-1223 http://www.sciencedirect.com/science/article/pii/0021916985900893
37. Magnetospheric
Effects of Power Line RadiationC.
G. Park and R.
A. HelliwellScience
38. Transmitter
simulation of power line radiation effects in the magnetosphere, C.G.Park and
DCD Chang, Geophysical Research Letters
5 (10), 1978 http://star.stanford.edu/~vlf/publications/214.pdf
39. DEMETER observations of ELF waves
injected with the HAARP HF transmitter, M.Platino, U.S. Inan, T.F. Bell,
M.Parrot, and E.J. Kennedy, Geophysical Research Letters, 33 (L16101), 2006.
40. Impact
of (almost forgotten?) infrasound from below on the upper atmosphere and
ionosphere, Lastovicka, J.;
Buresova, D.;
Chum, J.;
Drobzheva, Ya. V.;
Krasnov, V. M.
36th COSPAR Scientific Assembly. Held 16 - 23 July 2006, in
41. Influence
of infrasonic waves on the ionosphere. Krasnov , Ya.V. Drobzheva, Physics and
Chemistry of the Earth, Part C: Solar, Terrestrial & Planetary Science,
Volume
26, Issue 6, 2001, Pages 433-437 http://www.sciencedirect.com/science/article/pii/S1464191701000265
42.