Professional Documents
Culture Documents
Abstract: The subject of deciding on the most appropriate of Reference [4]. As a result, the Audible Sound & Vibrations
method and conditions for accurately measuring the noise level (ASV) Subcommittee (SC) of the IEEE Transformers
of a transformer has been discussed through the transformer Standards Committee recently approved implementing the
industry for decades. It is the objective of this Paper to Sound Intensity Method in C57.12.90 [1], again, provided that
contribute to this effort by identifying more appropriate appropriate conditions are specified.
conditions for both the Sound Pressure Method and the Sound Section II of this paper presents results of a comprehensive
Intensity Method. The Paper presents results of comprehensive study performed by the authors of this paper to determine the
indoor measurements made of the frequency spectrum and most appropriate conditions for more accurate measurement of
total noise of a large number of power transformers using the transformer noise using the Sound Intensity Method. These
two measuring methods. First, the paper presents data used to
conditions are proposed to be used by both IEEE and IEC
develop appropriate conditions for more accurate measurement
Standards.
of transformer noise using the Sound Intensity Method. The
Also, the IEC Standards, Ref. [5] and [6] allow using a
paper then presents data used to develop a more accurate and
reliable Sound Wall – Reflection Correction to be used with Sound Wall – Reflections correction to be used along with the
the Sound Pressure Method. Lastly, the Paper presents Sound Pressure method. Recently the ASV SC of the IEEE
appropriate method of measuring Load noise of power Transformers Standards Committee approved correcting for
transformers. Sound measuring conditions developed and Sound Wall – Reflections in C57.12.90 [1].
presented in this paper are being proposed to replace existing Section III of this paper presents a proposal for a correction
conditions in the IEC and IEEE Industry Standards of to be used by IEC and IEEE Standards to use along with the
transformer noise testing. Sound Pressure Method to properly account for Sound wall
Keywords: Power Transformers, Transformer Noise, Sound Power reflections when performing indoor sound measurements of
Measurement, Transformers Industry Standards transformer noise.
I. INTRODUCTION Finally, the IEC standards allow measuring Load noise
under some conditions. The Standards also allow measuring
Indoor measurement of the true noise level of a transformer
Load noise at less than full – load and correcting up for full –
is critical, especially when requirements for low and ultra –
load. In Reference [8] of this paper, the authors demonstrated
low noise levels are to be met. It is even more critical when the
the importance of measuring Load Noise in addition to No –
levels of individual frequency components of transformer
Load noise, as the total noise of the transformer, including
noise are to be guaranteed.
Load noise, can be several dB > No – Load noise, especially in
Presently, the IEEE Standards [1] specify using only the
low no – load noise transformers. Again, the ASV SC of the
Sound Pressure Method and allowing only a correction to be
IEEE Transformers Standards Committee recently approved
made for the ambient noise under specific ambient noise
implementing Load noise measurements in C57.12.90 [1].
conditions. The IEEE Standards do not allow using the Sound
Section IV of this paper discusses the IEC conditions and
Intensity Method and also do not allow a correction for the
corrections of Load noise measurements. It offers a proposal
Sound Wall – Reflections to be used with the Sound Pressure
for improved accuracy to be incorporated in both the IEC and
Method. In addition, the ambient noise conditions, as specified
IEEE Standards.
by both the IEC and IEEE Standards, are often not feasible to
meet when measuring Low & Ultra – Low noise power and II. DEVELOPMENT OF CONDITIONS FOR APPROPRIATE NOISE
distribution transformers. MEASUREMENT USING THE SOUND INTENSITY METHOD
References [2] & [3] have demonstrated that the Sound IEC allows using the Sound Intensity method, provided the
Intensity Method is the more accurate Method of measuring difference between the measured values of the total dB (A)
transformer noise since it measures the true noise radiated by using the Sound Pressure and the Sound Intensity Methods is <
the transformer, avoiding errors caused by ambient noise, near 8 dB. The selection of this condition was based on limited data
– field effects, and Sound Wall – Reflections. This is provided collected and presented in a CIGRE paper, Ref. [7]. However,
that appropriate conditions are met. This is also the conclusion over the years, it was realized that this condition is not
__________________________________________________ accurate enough. As a result, less and less IEC customers now
Dr. Ramsis Girgis is presently the R&D manager at the ABB approve using the Sound Intensity Method. In Reference [3],
Power Transformer plant, Saint Louis, MO. the authors of this paper presented data that demonstrated the
Mr. Mats Bernesjö is a Development Engineer at the ABB inaccuracy of this 8 dB IEC condition. The data showed that
Power Transformer plant, Saint Louis, MO. this limit should be in the 4 – 5 dB range, not 8 dB.
2
(Im - Ic)
-6
[Im] and the corresponding theoretical estimates of the -7
accurate Sound Intensity levels [Ic] have been plotted versus -8
-9
the corresponding differences between measured Sound -10
Pressure [Pm] and Sound Intensity levels [Im]. The quantity -11
[Im - Ic] represents the measuring error in the Sound Intensity -12
-13
Measurement. The estimated Sound Intensity levels [Ic] have -14
been calculated using the measured sound pressure levels [Pm] -15
-16
corrected for the actual ambient noise level at the 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
corresponding frequency component, the IEC Sound Wall – (Pm - Im)
Reflection (Environmental) factor K, Ref. [5], and the Near –
Field effect. A value of 0.2 was used for the average acoustic (b) Total dB (A)
absorption coefficient “ ” in the calculation of the
Environmental factor K in the factory test environment. No Figure 1 – Error in Intensity Measurement of Load noise
sound wall reflection correction was made for measurements vs. [Pm - Im], measurements made at factory environment
in sound rooms. Finally, an average value of 1.5 dB was used 4
for correcting for Near – Field effects as previously concluded 3
2
by the data presented in References [2] & [3]. 1
0
Figure 1 below presents such plots for Load Noise measured in -1
-2
a typical factory test area. Figure 2 presents the same but -3
measured in a sound – room with very low ambient noise level -4
(Im - Ic)
-5
as well as walls covered with highly sound absorbing material. -6
The figures show a clear correlation between the quantity [Pm -7
-8
- Im] and the quantity [Im - Ic]. These results show an error in -9
the Intensity measurement of < 1 dB for [Pm - Im] < 4 dB. The -10
-11
error is 1 – 2 dB for [Pm - Im] between 4 & 5 dB, 2 – 3 dB for -12
[Pm - Im] between 5 & 6 dB, etc. -13
-14
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
4
(Pm - Im)
3
2
1
Figure 2 – Error in Intensity Measurement of Load noise
0 vs. [Pm - Im], measurements made at a large sound–room
-1
-2 The plots also demonstrate that a [Pm - Im] value of 8 dB, as
-3 allowed in the IEC Standards, represents an error of about 4 –
-4
-5 5 dB; the noise level measured using the Sound Intensity
(Im - Ic )
(Im - Ic)
-5
Measured Sound Intensity level, Im: 63.7 dB -6
[Pm - Im] = 72.1 - 63.7 => 8.4 dB -7
-8
Measured ambient level: 65.5 dB -9
IEC wall correction factor K: 1.8 dB -10
-11
Near – field effect: 1.5 dB -12
Ic = [72.1 - 65.5] - 1.5 - 1.8 => 67.7 dB -13
-14
(Im - Ic) = 63.7 - 67.7 => - 4.0 dB 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
The measured Sound Intensity level in this case is 4 dB lower (Pm - Im)
than the theoretical value. (c) – 400 Hz component
Figure 3 presents data, of frequency components and total 4
Core noise of the same 36 power transformers, presented in the 3
2
same manner as that of Figure 1 for Load Noise. Again, these 1
transformers were tested in a factory environment. 0
-1
4 -2
3 -3
-4
(Im - Ic)
2
-5
1
-6
0
-7
-1
-8
-2
-9
-3
-10
-4
(Im - Ic)
-11
-5
-12
-6
-13
-7
-14
-8
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
-9
-10 (Pm - Im)
-11
-12 (d) – Total dB (A)
-13
-14 Figure 3 – Error in Intensity Measurement of Core noise
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 vs. [Pm - Im], measurements made at factory environment
(Pm - Im)
(a) – 125 Hz component The figures show that, except for the data points in
yellow, the same correlation exists between the error
4 quantity [Im - Ic] of core noise and the quantity [Pm - Im]
3
2 as observed in Figures 1 & 2 for Load noise. Again the
1 figures show an Intensity measurement error of < +1 dB
0
-1 for [Pm - Im] < 4 dB. The error is 1 – 2 dB for [Pm - Im]
-2
-3
between 4 & 5 dB, 2 – 3 dB for [Pm - Im] between 5 & 6
-4 dB, etc.
(Im - Ic)
-5
-6 The data points in yellow in Figures 3 (b) and 3 (d)
-7
-8 correspond to Ultra – Low core noise transformers, where
-9 the 250 Hz component and the total Core noise are in the
-10
-11 range of, or below, the ambient noise. In such cases, the Ic
-12 calculation is not valid and, hence, the quantity [Im - Ic] is
-13
-14 not valid. However, in the majority of these cases, the
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 quantity [Pm - Im] is > 6 dB.
(Pm - Im)
-5
-6 considered invalid.
-7
-8 Table I – Values of Acoustic absorption coefficient per IEC
-9
-10
-11 Average
-12 acoustic
-13
-14 Description of room absorption
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 coefficient,
(Pm - Im)
Figure 4 – Error in Intensity Measurement of Core noise vs. Nearly empty room with smooth hard walls
0.05
[Pm - Im], measurements made at a large Sound Room made of concrete, brick, plaster or tile.
Partly empty room, room with smooth walls. 0.10
Based on above data, the authors propose using the following
corrections when measuring transformer noise using the Sound Room with furniture, rectangular machinery
0.15
Intensity Method: room, rectangular industrial room.
No correction for (Pm - Im) 4 dB Irregularly shaped room with furniture,
0.20
Add 1 dB to measured value of Im for 4 < [Pm - Im] 5dB irregularly shaped machinery or industrial room
Add 2 dB to measured value of Im for 5 < (Pm - Im) 6dB Room with upholstered furniture, machinery or
industrial room with a small amount of acoustic
If (Pm - Im) > 6 dB, measurements using the Intensity Method 0.25
material (for example, partially absorptive
should not be allowed. Instead, the Sound Pressure method ceiling) on ceiling or walls
would have to be used; correcting for Sound Wall –
Reflections and ambient noise. Obviously, this is provided that Room with acoustic materials on both ceilings
0.35
conditions of ambient noise, as specified in the IEC and IEEE and walls
Standards, are met. Room with large amounts of acoustic materials
0.50
on ceilings and walls
III. THE SOUND WALL – REFLECTION CORRECTION
[ENVIRONMENTAL FACTOR] Typical Values of Sound Walls – Reflection Correction
Correction per present IEC Standards
In order to have an appreciation for typical magnitudes of
When a transformer is tested indoors for noise level, sound this correction and the impact of the different parameters on
reflections from walls, and other large objects, adds to the the magnitude of this correction, this factor was calculated for
measured noise level when using the Sound Pressure Method. a transformer with a measuring surface area of 100 m2. The
Transformer noise contains discrete tones that are affected by calculations were made for values of the Wall Sound –
room acoustics, room geometry, and reflecting objects. In Absorption Coefficient of 0.10, 0.15, 0.20, and 0.35 which
order to account for the effect of these sound reflections, the corresponds to a large variety of test room types. Also, the
IEC Standards, Ref. [5], provides the following calculation calculations were made for three different test room sizes from
formula for what is termed as the “Environmental factor K”. 2000 m2 to 8000 m2. Table II below presents the calculated
The calculated value is to be subtracted from the measured values of the wall sound reflection correction using the IEC
value of the noise level: formula.
5
Table II – Example values of sound reflection correction measure Load noise of a transformer if the transformer is to be
loaded to any significant levels during operation.
Wall Sound – Absorption
Transformer Area Coefficient -
100
100 m2 90
40
8000 1.8 1.2 1.0 0.6
30
Reiplinger's Equation
Proposed changes to IEC Sound – Reflection Correction 20
Measured
10
First, and for most, the table above shows that typical values of 0
this correction for a reasonable ratio of Area of Room / Area 30 60 90 120 150 180 210 240 270 300 330 360
Rated Power, MVA
of Transformer are in the 1 – 3 dB range. In fact, a wall
reflection of 3 dB means that the wall reflections are of a Figure 5 – Measured values of Load noise versus calculated
magnitude equal to the total noise of the transformer. using Reiplinger’s equation of Ref. [4]
Therefore, it is proposed that a limit of 3 dB is applied to this
correction (Environmental factor “K”). Measuring load noise at reduced load current
Secondly, Table – II shows that it would be sufficient to use According to the IEC Standards, if the measurements can only
values of 0.10, 0.20, and 0.35 for the coefficient “ ” to be performed at a reduced current (Current lower than the
account for three different types of rooms. The error involved rated full – load current of the transformer), the sound power
in using these 3 values, versus the original 7 values in the IEC level at the rated current shall be calculated by the following
Table – I, is within a fraction of a dB. It is, therefore, proposed equation:
LW = 39 + 18 log(MVA) , dB (A)
95
90
85
the guaranteed no – load Sound Power level of the
transformer, load current sound measurements are not 80
appropriate.” 75
The above formula presents magnitudes of load noise as a
70
function of the transformer MVA rating regardless of the
Measured
design of the transformer or the physics of the electromagnetic 65
40 Log (I)
forces which produce this noise. It is, therefore, expected to
60
not provide accurate values of load noise. This is demonstrated 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3
in Figure 5 below where measured load noise levels are I, P.U.
compared with those calculated using the above equation for Figure 6 – Measured Load noise levels at different loads
power transformers of a wide range of power ratings. It is
evident that the formula mostly provides an average value of Note that measuring Load noise at lower currents can be
Load noise level but at the same time, for some designs, it can difficult as load noise levels at such currents can be in the
be as much as 5 – 10 dB in error, Ref. [8]. Hence, the above range of, or even lower than, ambient noise.
equation can not be used to judge whether Load noise is high
enough to warrant measuring it. It is, therefore, proposed to
6