ML20138D220
| ML20138D220 | |
| Person / Time | |
|---|---|
| Site: | Harris  |
| Issue date: | 10/18/1985 |
| From: | Lee V ANALYSIS & COMPUTING |
| To: | |
| Shared Package | |
| ML20138D190 | List: |
| References | |
| OL, NUDOCS 8510230214 | |
| Download: ML20138D220 (30) | |
Text
truirnexm m w w
"\' ,,
W /
UNITED STATES OF AMERICA bJ "
Od NUCLEAR REGULATORY COMMISSION t>
pocoMgqe' ,
p,a ,-
3 BEFORE THE ATOMIC SAFETY AND LICENSING BOARD ?
, ')Q x, ,F J Mf'e , * ,
'~
In the Matter of CAROLINA POWER AND LIGHT COMPANY AND NORTH CAROLINA EASTERN MUNICIPAL Docket Nos. 50-400 OL POWER AGENCY 50-401 OL (Shearon Harris Nuclear Power Plant, Units 1 and 2) )
TESTIMONY OF VAN M. LEE REGARDING EDDLEMAN CONTENTION 57-C-3 Q.1. Dr. Lee, where do you live and what is your occupation?
A.I. My name is Van M. Lee. I reside at 26 Laura Drive, Westbury, New York 11590. I am the president and principal consultant of Analysis
& Computing, Inc. of New York. My areas of specialization are noise assessment and acoustical design, communication and warning system design and evaluation, and computer modeling.
Q.2. Dr. Lee, what is your education and experience which pertains to this contention?
A.2. I have a Ph.D. degree from New York University majoring in Noise and Acoustics with a minor in Applied Mathematics. Over the last twelve years, I have worked as a consulting engineer on noise and acoustics aspects of power plant, steel plant, oil refinery and platform, highway, subway, and tunnel blasting and ventilation hok T
D
[
- 2 projects, and other specialized computer modeling studies. I have also taught as ar. assistant professor in Computer Science at New York State University, Queens College, and New York Institute of Technology.
I am a member of the American Association for the Advancement of Sciences, New York Academy of Science, the Acoustical Society of
- America among others, and a eember of the standards committee of the American Society of Testing and Materials on environmental acoustics. I have conducted research on and developed procedures and models for the assessment of sound propagation outdoors. I
- have been retained by. International Energy Associates Limited (IEAL) since June 1983 as an acoustics consultant dealing specifically with the sound level coverage evaluation of nuclear power plant fixed siren systems.
ACOUSTICAL TERMIN0 LOGY Q.3. Would you please present a brief overview of the acoustics of fixed sirens?
A.3. The human ear is able to distinguish among the millions of sounds around us. Each sound, such as trumpet, fog horn, or siren, has distinctive pitch, loudness, and quality. The source of every sound is a vibrating body. In the case of a drum, the vibrating drumhead pushes against the air everytime it moves outward. It I
shoves the air molecules against other air molecules, compressing a
the air. This compression moves away, as the drumhead moves inward, leaving a region where the air is slightly thinner than normal. On the next outward push, another compression is formed and started on its way outward. We call these pulses of compression and rarefactions " pressure waves". When waves of sufficient strength reach your ears, they cause your eardrums to vibrate. It's these vibrations which your brain interprets as sound.
Despite the human ear's sensitivity to minute changes in air pressure, it is only when the changes are repeated in rapid successions, at least twenty times a second, or as we say in scientific terms, a frequency of 20 cycles per second, or 20 hertz, that the human brain perceives them as sound. On the other hand, vibrations that occur more than about 20,000 hertz cannot be heard by the average human ear.
This range of frequencies, 20 to 20,000 hz, is called the audible range, and it varies with different people and with different ages.
The siren, invented by Cagniard de la Tour, is in fact the historical instrument for the furdamental experiments on frequencies. The Cagniard siren consists essentially of a stiff disc, capable of revolving about its center, and pierced with one or more sets of holes at concentric circles on the disc. A windpipe in connection with bellows supply a stream of air perpendicular to the disc. As
( the disc turns, a succession of puffs of air escape through it, and l
l when the velocity is sufficient, they blend into a note. If two circles of holes are pierced on the disc with the outer containing as many as twice the number of holes in the inner circle, the two i
notes are found to stand to each other in the relation of octaves.
Thus, in passing from any note to its octave, the frequency of vibration is doubled. The audible frequency range is divided in the acoustics profession into octave bands of frequencies, and sometimes into finer bands of one-third octave bands.
Human ears are most sensitive to vibrations in the frequency range between 1,000 and 4,000 hertz. Changes in the atmospheric pressure of about one part in ten billion, if repeated 3,500 times a second, will send an audible sound to the brain. The threshold of hearing is equivalent to ~the turbulent noise of the flow of blood in the vessels of the ear, or 0.0002 micro-bar in pressure compared to the atmospheric pressure of 1 bar, whereas the threshold of pain is about 1,000 micro-bar, or a ratio of one million to one. This wide
, range is inconvenient to express in written form, thus, a logarithmic unit, called the Decibel is introduced to compress the scale of expression as is done in electronic units. When applied to sound, the decibel or decibel level or the sound pressure level is 10 times the logarithm of the ratio of acoustic pressure to a refence pressure at the threshold of hearing of the average person for sound at a frequency of 1,000 hertz standardized at 0.0002 micro-bar.
i I
l l
However, the human ear does not have uniform sensitivity to all frequencies, and is less efficient at low and high frequencies than it is in medium or speech-range frequencies. This characteristic, called pitch, depends not only on-frequency but also upon the loudness of a sound. For example, lower frequency tones seem lower pitched when played very loudly, whereas high frequency tones seem higher pitched when played very loudly. In attempting to relate the physical measurements of sound in terms of decibels and frequency to the subjective evaluation of. loudness by the human ear, various "weightings" or weighted networks of filters were developed to obtain a single number representing the sound level of a sound containing a wide range of frequencies in a manner representative of the ear's response. A popular method which has been found to give reasonably good. agreement for most sounds in our environment is the A-weighted Sound Pressure Level, or dBA. Other weightings are "B",
"C", "PNL", "EPNL", etc. for different types of sound. The C-weighting is commonly used to rate industrial machinery and equipment sound output and is almost flat or unweighted in that it accounts for sounds in all frequencies almost equally.
Sound pressure levels from a source often vary from time to time, such as a fire siren. To quantify a time varying signal, many sound pressure level readings are usually taken over a period of time. Statistics can then be compiled from these readings, the level exceeded ten percen_t of the time is the ten-percentile level, and so forth, These are called Percentile Sound Levels. To
4
\
l
. . l quantiff a time varying signal which lasts over a short period of time, such as a 3-minute siren sounding, a Single Event Level (SEL) is used. It is the equivalent steady sound level which would contain the same sound energy as the time varying signal during the same time period.
More formal definitions of these and other additional terms are ,
provided in the Glossary attached to this testimony.
Q.4 Did you conduct a FEMA-43 review of the Shearon Harris siren system?
A.4 I conducted a technical review of the fixed siren system design with respect to FEMA-43, Section E.6.2.1, as submitted in Carolina Power and Light Company's "Shearon Harris Nuclear Power Plant Alert / Notification System Report" dated January, 1985. FEMA-43, Section E.6.2.1, details the criteria for siren system evaluations.
It requires topographical maps identifying siren locations, areas
, with population of 2,000 persons per square mile or greater, and 70
. and 60 dBC siren sound level contours or alternatively the 10 dB l
above ambient sound level contours. It also requires documentation of siren make and model, mounting heights, and verification of siren acoustic output ratings, and description of how resultant l
siren sound level contours were computed; in particular, with respect to site-specific weather and topographical conditions. For those designs with the stated objective of achieving +10 dB above ambient, detailed requirements on ambient background surveys are enumerated. A system is considered acceptable if it provides a
__ __ _ _ _ ~ _ _ _ _.__.._- _ _
i minimum of 60 dBC for areas with less than 2,000 persons per square mile and a minimum of 70 dBC for areas with more than 2,000 persons per square mile, or alternatively, if it provides a minimum of +10 dB above the daytime ambient background noise. levels, throughout the populated areas of the EPZ.
There were three aspects to my review: (1) review of input data.
(2) review of calculation process, and (3) review of the system coverage. Input data review consisted of verifying siren sound output levels against anechoic chamber or field test measurements, reasonableness of meteorological conditions used for computation of site prevailing summer daytime conditions, site terrain character-istics necessary for special treatment, and completeness of ambient t'ackground noise data if applicable. There are at least three analytical avenues to evaluate a licensee's procedure for calcu-lating siren sound levels and resulting sound level contours away from the sirens.
The first is to use the accepted siren output data and employ the 10 dB loss per distance doubling attenuation rate. This 10 dB rule-of-thumb should provide conservative range estimates within 10,000 feet of the siren in the absence of special conditions. The procedure can be extended to hilly or built-up areas by further accounting for the simple barrier diffraction loss. Should the licensee's procedure be more conservative than the 10 dB rule with
further adjustments for other terrain features when needed, the licensee's procedure can be judged as acceptable.
The second approach is to perform an independent sound propagation calculation using the same input information. This is an accepted engineering approach and often the only viable alternative in those cases in which the event to be gauged has not occurred or cannot be economically or reasonably conducted to provide satisfactory and realistic verification. The central focus in such an approach is the method or computer model employed to perform the independent calculation. The computer model chosen for this task of siren ,
evaluation must first be based on known principles of sound propagation in the atm'osphere and over the ground; second, and more important, its application to similar situations must have demonstrated reasonable success in quantifiable statistical terms.
One such computer model, the Outdoor Sound Propagation Model (05PM), is employed for this purpose and is discussed in further detail later. Irrespective of the methods or computer models used, the 10 dB rule-of-thumb, as postulated in CPG 1-17 from empirical experience in the acoustical profession fcr mid-range distances of ,
2,000 to 10,000 feet'in the absence of intervening terrain or man-made structures under usual weather conditions, offers a ready reference for other, more sophisticated methods of computing siren coverages.
i l
l
The last and most desirable approach is to have sufficient good quality field data to directly verify the siren effective ranges and the coverage claimed. Such, however, is seldom the case.
Frequently, the licensee, in its siren system design and evaluation effort, solicits the aid of an acoustical consultant, and some field measurements of siren sound levels are performed. Such data, when available, are invaluable in the review process. Statistics of measured siren sound. levels versus distance can provide a quick check of the siren range calculation procedure, and can be used to infer the validity of the system coverage claimed.
All three approaches were applied in the Quality Assurance Verification (QAV) review of the Shearon Harris siren system design.
Since the licensee's consultant, Acoustics Technology, Inc. (ATI),
used a proprietary computer model, comparisons of its results with the 10 dB-rule and OSPM can only be made through statistics.
Log-linear regression analyses of the logarithms of distances and <
corresponding siren sound levels indicated that the 60 dBC design range (no area within the EPZ is identified to have population density of 2,000 persons per square mile or greater) for the FS-1000 siren as estimated and employed by the licensee is conserva-tive (approximately 7,100 feet range versus 9,000 feet using the 10 dB rule and an effective range of 8,400 feet as evaluated through OSPM). Because ATI included some field measurements from other siren system designs in the Shearon Harris design review, those data were not judged to be site-specific and acceptable for review. !
t I
l
The QAV review consisted of 11 sirens in and around the fuquay-Varina quadrangle out of the 62 siren system. The Design Report indicated that there were populated areas within the EPZ with less than 60 dBC siren sound levels. Ambient noise surveys were performed within those areas for consideration of the +10 dB above background criteria. However, upon reviewing the submitted ambient data, the ambient survey procedure was judged to be inadequate. The Shearon Harris siren system was found to be conditionally acceptable per ding
< resolution of those populated areas shown on Map 3 of the Design Report with less than 60 dBC coverage.
Q.5. Dr. Lee, please describe how you utilized the Outdoor Sound Propagation model?
A.S. To evaluate the effectiveness of a siren system within the 10-mile EPZ against the stated criteria for 60 and 70 dBC or 10 dB above the background levels, it is necessary to be able to predict what siren sound levels will exist, under a given set of weather conditions, at numerous points through the area due to the entire array of sirens. This statement contains three qualifiers of importance: the first, the set of weather conditions, involves a time scale in effect; the second, at numerous points throughout the area, involves a spatial effect; and the last, due to the entire system of sirens, involves a selective or superpositioning effect depending on the types of siren (stationary or rotational). A realistic problem of this magnitude can be treated by first solving the problem of predicting the siren sound level at one receiver l
I
location or receptor due to one siren under one set of specified weather conditions, and then being able to repetitively apply the method for all desired siren-receptor pairs, accounting for the logarithmic summation or maximal selection effect appropriate for the type of siren, and over a desired sequence of site weather conditions. The following deals specifically with the methodology employed in the OSPM dealing with prediction of siren sound levels for a single siren-receptor pair under specified weather conditions.
To predict the resulting siren sound level at a receptor due to a given siren, one must have knowledge of the sound levels that exist close in to the siren, and then be able to calculate sound levels at large distances from the siren, or at the far-field. The siren output level and its directivity are assumed to be known through measurement. The investigation concerned with sound propagation in the atmosphere and over the ground is the subject area of atmospheric sound propagation. The subject is under constant review with the most comprehensive and in-depth review to date being the report by Nyborg and Mintzer (1955); although the subject was more recently reviewed by Embleton (1982), Piercy et al (1975), Putnam (1975),
and others. Only Nyborg and Mintzer offered a practical approach to consider the problem as a whole, decoupling the various factors affecting sound propagation and weighting their contributions in the overall problem. Recent major contributions were made in the areas of atmospheric absorption of sound, and the treatment of the ground as an acoustic ~ impedance layer; these refinements are yet to I
be incorporated. With some exceptions, the OSPM follows closely the computational procedures in Nyborg and Mintzer, and its essen-tial elements are also in line with the procedure outlined in NASA's Technical Memorandum X-56033 for computing aircraft noise propagation (Putnam 1975). The application of OSPM, along with other surface interpolation and grid manipulation routines, to an entire siren system of fixed types (stationary and rotating) of sirens is discussed later.
The method - OSPM - to be described is meant to be an additional alternate means of verifying the validity of a licensee's calcula-tions, particularly in situations in which the 10 dB rule may be too gross or deficient.
The approach taken is~ to consider the dominant factors independently and separately and to account for their interaction effects through a weighting of field experience as recommended in Nyborg and Mintzer. Following this approach, as adopted in Beranek et al (1971) and Piercy et al (1977), the most significant influence on sound level at any position is the distance between the receptor and the source. In an atmosphere with no losses and far from any boundaries, the change in sound pressure level, dl, due to a change in distance to r from a reference distance r is, g
i dL= -20 log (r/rg ).
I l
This represents the well-understood inverse square law of spherical divergence for a point source in decibel form. For each doubling of distance, the sound pressure level decreases by 6 decibels.
A second dominant and ever present factor is the additional loss, or excess attenuation, sound waves suffer along their propagation paths due to air absorption processes. Part of the acoustic energy is changed into heat, and part changed into internal energy within the air molecules themselves. The atmospheric absorption losses are quantified in terms of ambient temperature and humidity conditions for distinct frequencies, and is linearly proportional to the distance traveled. The Society of Automotive Engineers issued ARP-866 for determining this absorption coefficient in 1964.
In 1969, Piercy of the Canadian Research Council proposed a nitrogen relaxation role in addition to the classical absorption and the oxygen relaxation rule considered. This resulted in a new calculation procedure adopted by the American National Standards Institute (ANSI) in 1978. It was felt (Smith, 1973) that ARP-866 overestimated the attenuation.at frequencies from 1,000 to 4,000 hertz and underestimated the attenuation at frequencies above 4,000 hertz. The ARP-866 was programmed into the OSPM, and since the siren frequencies are typically below 1,000 hertz, the computations were not updated to reflect the new ANSI procedure. For the FS-1000 sirens deployed in the Shearon Harris siren system (with dominant frequencies at 550 hertz), air absorption amounts to 0.5 to 1.0 dB per 100 meters.
i l
I l
1
When a sound wave propagates nearly parallel to the earth's surface, the amplitude and phase of the sound waves are greatly dependent upon the acoustical properties of the earth's surface.
The ground surface characteristics vary from acoustically hard to acoustically soft, where an acoustically hard surface is one that acts as a perfect reflector of sound and an a:oustically soft surface is one that acts as an absorber of sound. The presence of the ground plane creates complex reinforcement and destructive interference patterns, and can significantly distort the sound field, particularly for pure-tone sounds in aquiescent atmosphere.
When a sound source and the receiver are both close to the ground and the separation distance is large compared to the wavelength, we have what is called propagation at grazing incidence as formulated by Ingard (1951). The sound field consists of two major regions.
In the first the sound pressure level fluctuates about an average curve sloping approximately 6 decibels per distance doubling.
Beyond a certain distence from the source, depending upon the heights of the source and the receiver and the acoustical property of the ground, the level ray decrease by about 12 decibels per distance doubling. It is to be noted that excess attenuation due to ground impedance is not linearly proportional to distance, and that maximum ground attenuation typically occurs in low frequencies.
2 The ground attenuation in OSPM is determined empirically based upon the extensive field experimental data of Parkin and Scholes (1965).
Thus in the absence of other factors, such as obstruction of line-of-sight by terrain, or shadows caused by wind and temperature gradients, or presence of heavily forested areas, it can be seen that the 10 dB per distance doubled rule is a simplified procedure in which the spherical divergence over distance (6 dB/DD), the atmospheric absorption (typically 0.5 dB per 100 meters), and the ground attenuation which may approach 6 dB/DD and is not propor-tional to distance, are accounted for and is indeed conservative.
4 Figure 1 from Oleson and Ingard's 1959 year-long field measurement program illustrates the behavior of sound pressure level over distances in the range of up to 5,000 feet, and as can be seen the sound level drops initially at the rate of 6 dB per distance doubling, and at some point between 500 to 1,000 feet the decay rate starts to increase and may eventually approach a drop-off rate of 12 dB per distance doubling.
Wind and temperature gradients or differences in wind speed and temperatures at various heights from the earth's surface produce excess attenuation effects on sound propagation, since they affect the speed of sound at different levels thereby affecting the direction to which the sound waves travel. The velocity of sound propagation in the atmosphere va' ries as the square root of the absolute air temperature, and wind velocity adds to or subtracts from the sound velocity, depending on whether the propagation is upwind or downwind. During the day, which is normally adiabatic, temperature decreases with increasing height above the ground, and I
I i i l I
- ,; I I I I I I I I i Sw _*
I
\ 's
\ N g s -
\ \sq
- N g 's WIND 20-25 MPH -
N (NO SHOWFALL) 1
% g
\ N &
N ss .... . . . .
J _
12 DB/DDd 608/DD A '
, . 's, _
e N >s. s N
- \ / s s
s
\ / 's s l
El _
g
' 's _
E \ \ \ 's N
\ N---\ S s
~}
~
\ \ \ \ % .
~
y 1 _
\ NO WIND ;\ \, /
~ FREQUENCY:600-120d CPS MICROPliONE HEIGHT:17 FT.
\(N SNOWAW s\
\
g W$ s
~
l ~ GROUND COVER: 3 - 4 FT. .
\ N[ s FRESH SNOW . \
DURING MODERATE l _. SOUND WIND . _
i > > SNOWSTORM WIND 25 .30 MPH ,, _
- MICROPHONE ~ POSITIONS
- l 1 2 3 4 5 6 7 8 9 10 I I I I 'l i I I I I I I 30 - 50 100 300 500 - 1000 3000 -5000 DISTANCE FROM . SOURCE (FEET)
FIGURE 1
l
. l the decreasing vertical sound speed profile causes the sound to bend upward (principle of least action). Since the wind almost always increases with increasing height above the ground, the sound speed profile decreases more in the upwind directions, thus bending the sound rays further upward, reinforcing the temperature effect.
For downwind propagation, the sound speed profile increases with increasing height above the ground and the sound bends toward the ground, opposite to the direction of temperature induced curvature.
In this case the net curvature of the sound path depends on the relative magnitude of the wind and temperature gradients. One of the most important effects of this refraction phenomenon is the creation of acoustic shadow zones. Nyborg and Mintzer's formulation of shadow zone formation and distance to shadow boundary were adopted in OSPM.
Barriers are simplified geometrical idealizations of natural and man-made obstructions to sound propagation. As such, the limitations in application are obvious as in cases of undulating
! terrain.' and rows of single-family houses. However, these limitations are related strictly to approximatiens in geometry rather than the acoustics. The subject of sound wave diffraction Ly an obstacle has been investigated in great length with many recent studies conducted by the U.S. Federal highway Administration, Office of Research. The results indicate that the simple procedure proposed by Maekawa (1966) based on simple geometrical l theory of diffraction is quite satisfactory for most practical l
t l
outdoor situations. The method used in OSPM is based upon"the work of Maekawa, with a slight modification to take into account the effect of the sound velocity profile brought about by the wind and temperature gradients.
Two other attenuation processes are treated by OSPM. One is foliage attenuati.on which is applicable only when penetration is significant, more than 300 feet of the sound path, using from Beranek et al (1971):
AF=0.01 r * (f)1/3 The other is scattering by turbulence. The forward scattering theory presented by Brown and Clifford (1976) for sound propagation in the surface and free-convection layers of the lower atmosphere for a ground-based transmitter-receiver is adapted for use in OSPM.
Some qualifications must be stated with regard to the adjective "special" as used in NUREG-0654; the special weather conditions of concern here are not " extreme" weather conditions such as hail, storm, or the like. The reasons for not consioering such extreme but nevertheless possible scenarios for evaluation are basic: the first reason is that the acoustic knowledge does not exist, and the second reason is that the myriads of combinations of weather parameters that could possibly constitute extreme conditions would render such a task impossible to evaluate. The qualifier, l
l l
t "special", is interpreted to mean identifiable and quantifiable prevalent weather patterns that could seriously affect the siren f coverage under usual conditions. These effects include land-sea breezes along coastal areas, valley flows, desert inversions, etc.
Such site-specific characteristics with prominent seasonal or diurnal patterns can be identified and should be incorporated as {
f input weather conditions for the evaluation of siren systems. l Q.6. What conditions did you use in evaluating the Shearon Harris siren !
system, and why?
A.6. I used average summer daytime conditions. NUREG-0654/ FEMA-REP-1, Appendix 3, Section C-3, provides the design criteria against which siren alert and notification systems are evaluated. The design i rationale is to achieve a target level of 10 dB above average daytime ambient background, and the siren system's acoustical design may use average daytime conditions. The choice of summer versus other seasons is intended to provide an added margin of I conservatism in the design since (1) the average ambient background which is dominated by human activities, primarily traffic, is :
usually higher in summer than in uther seasons (Safeer 1973), and daytime ambient background noise levels are typically 6 to 10 db higher than nighttime levels (U.S. EPA, 1974), and (2) the siren sound may be additionally attenuated by the fuller summer vegetative covering.
.There was a popular belief up to 1894 (Batemant 1914) that fog, rain, dust, and snow cause significant sound attenuation. However, studies of these conditions by Tyndall and Henry in England (Lord Rayleigh, 1945), Livermore and King in North America, and Sieg in Germany (Sieg, 1940) indicated that the attenuations due to such ,
effects are insignificant for sound with frequencies below 1,000 hertz, and siren frequencies are typically below 1,000 hertz. This is further confirmed by more recent studies conducted by NASA and the U.S. Army (Putnam 1975, DeLoach 1975, Delany 1971, Henley and Hoidale 1971, Oleson and Ingard 1959, Ingard 1955 and 1953).
Oleson and Ingard's study carried out in 1956 at the Army Ordnance ;
Test Depot in Maynard, Massachusetts for entire years through fog, rains, and snowstorms is particularly relevant since few acoustic field measurement programs were ever planned for such weather conditions. A major finding was that at frequencies between 150 and 2,400 cycles per second, or hertz, the sound attenuation due to rain or fog is almost negligible; and at frequencies above 1,000 '
hertz the sound attenuation due to falling snow becomes significant.
Thus, winter snowfall or autumn fogging or spring raining conditions are not expected to be less favorable to siren sound propagation i than sunmer conditions.
As concluded by Oleson and Ingard and all other outdoor sound propagation studies, the single most important factor is the effect of wind and temperature gradients. The acoustics associated with
! this so-called refraction effect is quite well understood (Piercy a
et al 1977). It is known that temperature gradients under 10 meters above the ground in the summer are greater than those in the winter, and that summer daytime lapses (temperatures at higher elevations are less than the ground level temperature) predominate, cs do nighttime inversions in the winter, thus making sumer daytime conditions conservative (tending to bend the sound rays skyward) for sound propagation relative to nighttime and winter conditions (Nyborg and Mintzer 1955).
The average wind shows a definite diurnal variation close to the ground. It reaches a maximum around noon and a minimum at midnight, thus maximizing the potential of acoustic shadow formation during the day and minimizing the same during the night. Though little information is available concerning the wind structure close to the ground, it is however strongly dependent on the temperature gradient. In the daytime when the ground is hotter than the air above, there is a tendency for-the air layer to move upwards and thus stir up the air, encouraging turbulence in the atmosphere. In the nighttime, on the other hand, when the temperature gradient is reversed, the atmosphere is more stable, and the air flow has a tendency to be laminar. In other words, the strength of the l
turbulence at night is ordinarily less than in the daytime which may be one of the reasons that sound propagation at night is "better" than in the daytime.
Thus, in the context of ambient background noise level and outdoor sound propagation, the specification of summer daytime conditions in FEMA-43 as siren system design conditions is a prudent and proper choice as practiced in most engineering designs.
Q.7. Dr. Lee, what work have you done in connection with this contention?
A.7. To assess the nighttime alerting capability of the Shearon Harris siren system, the OSPM was applied repeatedly to each of the 68 sirens in the final system ( 62 in the original system and 6 additional sirens in the supplementary submittal) with nighttime weather conditions from on-site meteorological tower data. Siren data consisting of siren power levels in octave bands from 31.5 to 8,000 hertz, siren locations and mounting heights above ground elevation were input. Topographical profiles for each siren along sixteen compass bearings out to 12,000 feet were extracted from U.S.
Geological Survey maps; input information consisted of receptor coordinates, dominant ground type between the siren-receptor pair, any line-of-sight obstructions and distance from siren to, and relative height of, the obstruction, significant foliage penetration
! and distance thereof, if any.
The output of each siren run, consisting of estimated siren sound pressure levels in dBC at each one of the 112 polar grid points is fed individually into a surface interpolation and mapping program to generate a finer mesh (at grid interval of 150 meters) over the
, entire EPZ. Since the sirens are of the rotating type, the
individual siren sound levels at a given receptor cannot be reliably synchronized to achieve the logarithmic additive effe:t; instead, the maximum of multiple siren sound level contributions is selected. This maximal selection process was performed at each of the grid mesh points (215 x 215 points). And final contours were drawn through this mesh at intervals of 5 dB increments using a standard least-square weighted contouring routine. The resulting contour map was scaled and overlaid on a household location map supplied by Carolina Power & Light. The number of houses located within 10 dB increment contours was then counted. The distribution of houses in the Shearon Harris EPZ versus siren sound levels was tabulated as follows:
TABLE 1 DISTRIBUTION OF HOUSEHOLDS BY SIREN SOUND LEVELS IN DBC SIREN LEVEL AVERAGE DBC AVERAGE DBA PERCENT l
100-110 105 102.4 7.42 90-100 95 92.2 15.24 l 80-90 .85 i 81.9 33.61 70-80 75 71.8 39.83 :
60-70 65 61.8 3.90 '
The siren contours were generated in dBC since NUREG-0654 f specifies dBC, and as discussed previously, the A-weighted sound pressure levels are commonly used for human hearing purposes. ;
To properly convert the siren sound levels from dBC to dBA for
[
awakening assessment, the following process was performed in lieu of i
. - , -. . . - . ye .
the difference of 3.2 dB from A-weighting to C-weighting for the 500 hertz octave band containing the fundamental siren frequency. This additional cautionary step was taken to account for the possible non-uniform frequency dependent attenuations over distances. Four sirens (#63, #65, #66, and #67) were re-run in OSPM, and this time the output was of dBA at each of the same polar grid points as previously done. The paired dBC and dBA values were stratified, !
and statistics computed, with the results as follows:
TOTAL OBSERVATIONS IN FILE: 448 BASIC STATISTICS FOR FOLLOWING GROUP:
GROUP = 1.000 (= > 100 dBC)
DBC DBA DIFF N OF CASES 76 76 76 MINIMUM 104.800 101.900 2.500 MAXIMUM 111.400 108.700 2.900 MEAN 109.318 106.728 2.591 STANDARD DEV 2.071 2.147 0.129 BASIC STATISTICS FOR FOLLOWING GROUP:
GROUP = 2.000 (= >90 dBC and <100 dBC) ,
DBC DBA DIFF N OF CASES 57 57 57 4 MINIMUM 97.600 94.700 2.700 MAXIMUM 97.900 95.100 2.900 MEAN 97.872 95.065 2.807 STANDARD DEV 0.086 0.114 0.032
BASIC STATISTICS FOR FOLLOWING GROUP:
- GROUP =
3.000 (=>80dBCand < 90dBC)
DBC DBA DIFF N OF CASES 99 99 99 MINIMUM 81.100 77.900 3.000 MAXIMUM 89.400 86.400 3.200 MEAN 85.139 81.998 3.141 2
STANDARD DEV 3.617 3.665 0.052 BASIC STATISTICS FOR FOLLOWING GROUP:
GROUP = 4.000 (= > 70dBC and <80dBC)
DBC DBA DIFF N OF CASES 103 IL3 103 MINIMUM 70.300 67.100 3.100 MAXIMUM 80.000 76.800 3.200 MEAN 74.737 71.540 3.197 STANDARD DEV 2.356 2.360 0.017 BASIC STATISTICS FOR FOLLOWING GROUP:
, GROUP = 5.000 (= > 60 dBC and < 70dBC)
DBC DBA DIFF N OF CASES 99 99 99 MINIMUM 60.200 57.000 3.000 MAXIMUM 69.900 66.700 3.200 MEAN 65.380 62.189 3.191 STANDARD DEV 2.941 2.943 0.032 The mean differences from the above analyses were used in Table 1 to convert the C-weighted siren sound levels to corresponding A-weighted levels.
i e
-- _ ,--,--w- - - - . - + -e+- %~ - -- , - - y- ----.rw y p - - 4 --e e
I From Table 1, it is noted that the Shearon Harris siren system provides full coverage of all populated areas in the EPZ under summer nighttime conditions in that it provides at least 60 dBC to 100 percent of the population (the minimum of 70 dBC does not apply since no area within the EPZ is identified as having population density of more than 2,000 persons per square mile). It will be ,
s'hown through Dr. Kryter's and Dr. Nehnevajsa's testimony that such l a physical acoustic coverage meets the NRC 10 CFR 50 Appendix E design objective for prompt public notification systems.
~
k
. . - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ , _ _ _ _ _ , _ _ _ __ _ _ .--_, _ _._. ~ _
GLOSSARY A-weighted Sound Level (dBA):
The ear does not respond equally to sounds of all frequencies, and is less efficient at low and high frequencies than it is at medium or speech-range frequencies. Thus, to obtain a single number representing the sound level of a noise containing a wide range of frequencies in a manner repre:;entative of the ear's response it is necessary to reduce, or weight, the effects of the low and high frequencies with respect to mediun frequencies. The resultant sound level is said to be A-weighted, and the units are dB. A popular method of indicating the units, dBA, is frequently used. The A-weighted sound level is also called the ..otse level.
C-weighted Sound Level (dBC):
C-weighting is an essentially flat weighting network in frequency ranges from 10 to 10,000 hertz with slight suppression of very low and very high frequency components of a sound spectrum. It is used primarily for acoustic output rating of industrial machines and equipment.
Decibel:
The decibel, abbreviated "dB", is a measure, on a logarithmic scale, of the magnitude of a particular quantity, e.g., sound pressure with respect to a standard reference value.
Hertz (Hz):
Frequency expressed in cycles per second.
Level:
The level of an acoustical quantity in decibels is 10 times the logarithm of the ratio of the quantity to a reference quantity of the same physical kind.
Sound Precure:
The minute fluctuations in atmospheric pressure which accompany the passage of a sound wave. These pressure fluctuations on the tympanic membrane (ear drum) are transmitted to the inner ear and give rise to the sensation of audible sound.
Sound Pressure Level (SPL):
The level of sound pressure squared and averaged over a pericd of time, the reference quantity being the square of 2 x 10-5 newions per square meter.
Single Event Level (SEL):
4 The equivalent steady noise level which in a given period of time would contain the same noise energy as the time varying noise during the same period.
Single event' noise is specified by the sound exposure level measured during a single event. . It can be closely approximated by:
SEL =L + 10 log T/2 (dB) max 10 where L = maximum sound level as observed on the A scale of a
. max standard scund level meter on the slow time characteristic.
arid T = duration measured between the points of (L -10) in seconds. max l
1 8
I i
l i
l l
l
BIBLIOGRAPHY IEAL. 1983. Analysis of siren system pilot test. Argonne National Laboratory.
FEMA. 1983. Standard guide for the evaluation of alert and notification systems for nuclear power plants.
FEMA-43. ;
NRC/ FEMA. 1980. Criteria for prer'iration and evaluation of radiological emergency response plans and preparedness in support of nuclear power plants.
FEMA. 1980. Outdoor warning systems guide. CPG-1-17.
Society of Automctive Engineers.1964. Standard values.
of atmospheric absorption as a function of temperature and humidity for use in evaluating aircraft flyover noise. ARP-866.
IEAL. 1985. Shearon Harris Nuclear Power Plant site-specific offsite radiological emergency preparedness alert and notification system quality assurance verification. IEAL-R/85-39. FEMA.
Embleton, T.F.W. 1982. Sound propagation outdoors -
improved prediction schemes for the 80's. Noise Control Engineering: 18(1).
Piercy, J.E. et al. 1977. Review of noise propagation in the atmosphere. JASA 61(6).
Nyborg, W.L. and D. Mintzer. 1955. Review of sound propagation in the lower atmosphere. WADC 54-602.
Putnam, T.W. 1975. Review of aircraft noise propagation.
NASA Flight Research Center. TM X-56033.
Piercy, J.E. 1969. Role of the vibrational relaxation of nitro JASA 46(3) gen in the absorption cr' sound in air.
American Ne tional Standards Institute.1978. Metnod for the calculation of the absorption of sound by the atmosphere. S1.26.
Smith, C.M. 1973. Atmospheric attenuation of aircraft noise - experimental values measured in a range of climatic conditions. Hawker Siddeley Aviation Ltd.
HAD-R-GEN-214.
l i
l L
c Ingard, U. 1951. On the reflection of a spherical sound wave from an infinite plane, JASA 23(3).
Parkin, P.H. and W.E. Scholes. 1965. The horizontal propagation of sound from a jet engine close to the ground at Hartfield. J. of Sound and Vibration 2:353-374.
Maekawa, Z. 1966. Noise reduction by screens of finite size. Memoirs of Engineering Faculty. Kobe University, Japan. 12:1-12.
Beranek, L.L. (ed) 1971. Noise and vibration control.
McGraw-Hill.
Brown, E.H. and S.F. Clifford. 1976. On the attenuation of sound by turbulence. JASA 60:688-794.
U.S. Environmental Protection Agency (EPA). 1974.
Information on levels of environmental noise requisite to protect public health and welfare with an adequate margin of safety. 550/9-74-004.
Safeer, H.B. 1973. Community noise levels - a statistical phenomenon. JASA 26:489-502.
s Batemant, H. 1914. Monthly weather review. 42:258.
Delany, M.E. 1971. A review of sound propagation in loser atmosphere. British Acoustical Society.
DeLoach, R. 1975. On the excess attenuation of sound in the atmosphere. NASA TN 0-7823.
Ingard, U. 1955. The physics of outdoor sound. Proceedings of the 4th Annual Noise Abatement Symposium: 11-25.
Ingard, U. 1953. Review of the influence of meteorological conditions on sound propagation. JASA 25:405-411.
Oleson, S.K. and U. Ingard. 1959. Field measurements of sound propagation. MIT.
Rayleigh, Lord.1945. Theory of sound. Dover Publication.
l l
1 l
- N.