ML20205J684

Testimony of Kd Kryter Re 860116 Memorandum & Order Concerning Limited Reopening of Record on Eddleman Contention 57-C-3.Comments Provided Re Issues 2 & 3.Related Correspondence

ML20205J684
Person / Time
Site:	Harris
Issue date:	02/25/1986
From:	Kryter K AFFILIATION NOT ASSIGNED
To:
Shared Package
ML20205J680	List: ML20205J676 ML20205J684
References
OL, NUDOCS 8602260311
Download: ML20205J684 (22)

Text

. -. . . .

i

&TED CORRESPONDENCL, DOLMETED UNITED STATES OF AMERICA USNRC.

NUCLEAR REGULATORY COMMISSION BEFORE TIiE ATOMIC SAFETY AND LICENSING BOARID6 EB 25 - A9 :41 '

0FFILE Gr wt :rd r In the Matter of ) 00CMEflNG & SE; Vin 3

) BRANCH CAROLINA POWER AMD LIGHT )

COMPANY AND NORTH CAROLINA -) Docket Nos. 50-400 OL EASTERN f.!UNICIPAL POWER ) 50-401 OL -

AGENCY )

)

- (Shearon Harris Nuclear Power Plant, )

i Units 1 and 2) -)

i TESTIMONY OF I:ARL D. KRYTER CONCERNING MEMORANDUM AND ORDER (LIMITED REOPENING OF THE RECORD ON EDDLEMAN CONTENTION 57-C-3)-

j My name is Karl D. Kryter. I reside at 01357 Heron Drive, Bodega Eay, California. My professional qualifications are discussed in my ,

previous testimony following Tr. 9690.

i-Reference is made to Memorandum and Order (Limited Reopening of the Record on Eddleman Contention 57-C-3) dated January 16, 198G.

Comments are given below with respect to Iscues #2 and #3, page 9 of the

! subject Memorandum and Order.

1 Issue #2, Review of Krallmann study and discussion of possible reasons for differences in probabilities of arousal as fo~ recast by j Kryter (based on Horonjeff, et al. data), and observed probabilities

) in I:rallmann study.

Figure 1 illustrates the large differences found between SEL and t

f percentage sleep arousal according to the Horonjeff, et_ al. and Krallmann-

8602260311 860221 PDK ADOCA 05000400 T PDR -

4

, 3 00 - __ - - - - _ .

7

_ ~ ,  ;,

e Kra11mann, Siren- - o' -

90 # ' e 4:00-5:00AMg' s

! 80 o 1 30-2:30 % M sd ' ~

to) s'f,

~

d ~& Aver. 0-5:00A2

_. 7 0 / .sc/ y

/ '

E e0-1:15 AM # O * ~~ pb r ,C (<*c' 5 60 . _ _ g . _ _ _ _ _ _ _

50 -

/ 9 # .e' 40 p, - . - - ~

- i e,s s ,.- .

}LO p[

f Horonjeff et al, Test

-5 20 l o* . - . Trans. Line Noise -

E NUREG/CR-2654,PNL-4227 x -

// .~..-_.

7 . Horonjeff et al, Low Freq. Noises 35

--[ 45 ;5

/ (Trnnsformer Airconditioner.Treffic) 65 75

~

ld'5 85 95 ~ 115 SEL IN BEDROOM Fig.l. Probabilityof sleep arousal persons j age 45 ye irs. Adjustment of Krallmann data, dashed Jine, based on analysis of Griefahn and Jansen of habituation to sleep arousal in experimental situations. The curves for the Horonjeffe et al data are taken from Figure 7A KryterjTestimony Regarding Eddleman Contention 57-C-3.

The curves above about SEL 80 are extrapolations based on other research findings. The " pure-tone" correction for saliency is based on Fig. 3 from Kryter and Pearsons.

s research investigations and the NUREG/CR 2654 PNL 4227 guide. Figure 2 shows one-third octave band spectra of the Horonjeff, et al. and Krallmann studies at an equal overall dBA level.

As will be discussed below, some of the differences in sleep arousal for scr;c of the noises in Figure 1 when at the same SEL can be explained as being due to inadequacies in dBA-weighting and one-third octave spectra measures for predicting loudness , and possible unassessed ecoustical masking conditions present in some of the tests. However, strictly basic acoustical-perceptual transform relations would not be Fufficient Oven if known exactly to explain the extent of these differences.

Interpretation of the Horonjeff, et al. and Krallmann findings, along rith the results of some generalizeri analyses of a number of other research studies, provides scientifically reasonable conclusions with respect to the problems at hand. Before presenting this analysis, a brief description of procedures and subjects ured in the IIoronjeff, et al. and I

Erallmenn studies are outlined. (The Krallmann study end its findings were unknown to me when I previously presented testimony on Eddleman Contention 57-C-3. )

Krallmann . Although certain details of the study are not fully described in his report, the results of the Krallmann study must be censidered, in my opinion, as applicable for purposes of estimating arousal from sleep by the alren alerting system of the Shearon Harris 4 EPZ. The subjects in the Krallmann study came from a v'ide variety of work backgrounds, and were attendees at a ra,e week course at an Air Raid Protection School in Germany. The subjecte slept in a dormitory i

4

... _4-SIMULATED TRANSFORMER ,

APP \. LINE SPECTRUM SIRENS e'

• ,.,,,'s (Krallman & FS 1000) 90 -

N 80 i z

A i

i N/ 'd- n s

, i

, 1 i f. 5, .

n f 80 ,=- f 70 - -

.--- w s o

,3 1 I / , \ ,\

i 70 -

60 j -' =,

'g

, / ,

\

TE M TRANSMISSION LINE .

(i

]60 -

50-- '* . N

]

(Horonjeff etal) 5,

,~ ng j .

! 4 650 -

40--

1 APPROX. AVERAGE SPECTRA " LOW"' I S '

FREQ. NOISES l  !

E Horonjeff et al, Simulated Transformer, 40 -

30-

-(Air Conditioner and Distant T'ra ' '

l I  : ,

30 I I I I I 20 I I I ' I II

.63  ;

125 250 500 1K 2K 4K 8K 463 125 250 500 1K 2K 4K 8K

t. ta FREGUENCY (H2) FREQUENCY (H2) 100 -

100 -

AIR CONDITIONER DISTANT TRAFFIC 90 90 i

2

~

80 ~

80*-

s 1

70 -

70 -

1 1

• 60 -

60 -

1 1

0 50 -

50 -

8 3

40 -

40 -

I I I I I I II I I I I I I I Il 30 463 125 253 500 1K 2K 4K 8K 463 125 psa, Maa- 1x 2K 4K SK t: FREQUENCY (HZ) t=8 FREQUENCY (HZ)

Figure 2, Showing that when all the noises are a level of about 81.5 dBA, the " low" frequency noises tested by Horonjeff et al (Simulated Transformer, AirConditioner and Distant Traffic) exceeded the spectrum of their Test Transmission Line noise in the region of about 400 Hz to 4000Hz by an average of 15 dB or so. Also, it it seen that the siren signal exceeds the spectrum of the Test Transmission Line 3 Noise by a 12 dB or so in the region of about 400 Hz and that, indicatedbythedotsandline(f)thesiren i signal consisted primarily of "line" or " pure-tone" components  !

rather than the more continuous spectrun components present  !

in that frequency region in other noises.

. -. - ~. .~ - .

room with provisions for the simultaneous testing of up to twenty-four persons. Each person had at their bed a switch which was to be pushed

. whenever awakened by the sound of a siren. The siren signal was presented at five different levels, in 5 dB steps from 40-60 dBC. The levels were presented in randomized order, over 98 individual test rights.

Krallmann presented the recorded siren signal once per night at a precolected level (40, 45, 50, 55 or F0 dBC) and time (0-5 a.m. , divided i

into 15 minute segments). The signals were presented at a steady level for 45 seconds. The number (617) and range (16-70 years, average age 43.88 years) of the subjects involved is much larger in the Krallmann f

I study than in other research studies on sleep.

f Iloronjeff et al. Fourteen adults, averaFe age forty-three years,

were exposed over twenty-one consecutive nights to four different noises at different levels of intensity and two, for each noise and level, temporal 1 patterns. One pattern was a steady-state level for forty-five minutes and l the other was a rising and falling level (two dB per second), somewhat i

similar to the variation in level.of the signal of a rotating siren.

The noises were presented via loudspeaker under remote telephone-line control placed in the bedrooms of the homes of the subjects. The subjects were 4

instructed to press a switch next to their bed when awakened.

i Psychoacoustic Factors i

i The principal psychoacoustic perceptual and physiological factors to be considered in assessing sleep arousal from the EPZ siren alerting 4

signal are discussed below, i

i

> . , --ry% w-y-- -

t-p9 e w<- +M=P e =- =p- t-FNMe*- -Pe --- =w -T -

w fg !+ =--gwew'-r4+-4yd-*7-'- pr-yw- -4' t-PW' * *F -'Y-W t--

'O

. 1. Audibility and Masking. It is taken that to cause arousal a sound must perforce be audible. A limiting factor is then, of course, t hether a given sound or noise exceeds the threshold of hearing of a person in the quiet , or the masked threshold of hearing due to other sounds or noises in the environment. -

b'ith regard to . assessing the arousability of a siren signal it is important to note that the masked threshold for the siren is determined primarily by the sound energy in a narrow band, say 1/3 octave wide, around 500 liz, and not necessarily by the overall level in dBA or dBC.

For a tonal signal, such as from a siren, the tone will be readily audible to the awake person when its sound pressure level exceeds by 2 dD or so the 1/3 octave level of a broad band background noise having the same center frequency.

Iloronjeff, et al. suggested that some 10 dB greater arousability of th Test Transmission Line noise than the transformer, air conditioner and traf'ic noises was due to the presence in their studies of a background mashing n aise. When at an overall level of about 35 dBA the spectre of the latter :: noises fell somewhat on that of the background noise over the entire spectrum, indicating some masking, and, therefore, i

reduced arousability of the test noises. On the other hand, Test Transmission Line noise at about 35 dBA exceeded the background noise in the frequency region above 500 Hz by 5-15 dB, and was below the i background noisc by 5-15 dB in the frequency region below about 500

[ llz . This would indicate that there would be 10 dB or so less masking of the higher frequency components of the Test Transmission Line ncise than ,

1 of the ~other noises, about the difference in general arousability found j among the noises (see Figure 1).

licrever , this explanation is probably not applicable for the noises when presented at levels above 50 dBA or so. At these higher levels, masking of any of the test noises by the background noise should be negligible, and cannot fully explain arousability difference between the l Test Transmission Line and the other noises.

I l

2. Loudness and Duration. The moment-to-moment subjective impressier of the intensity, or loudness, of sounds and noises have been found to be correlated with overall dBA and PNdD levels when wideband noises are judged relative to each other, or tones vs. tones. To the degree this is the esse, the differences between the I rallmann and Horonjeff, et al. , as well as other similar research findings, are probably not significantly due to the use of the A-weighting, or PNdB, as the i

baric unit of sound pressure IcVel measurement.

In my previous testinony on the Eddleman contention, it was suggested thct the most probable explanation for the 10 dB or so greater j arousability, at a given SEL, of the Test Transmission Line noise compered to the other !!oronjeff, et al. noises was the 10-15 dB greater i intensity of the Test Transmission Line noise, as shown in Figure ? in the mid-to-high frequency region, where the ear is most sensitive. It a

was pointed out that the critical bands of the cor for loudness in the low frequency region where the "other" noises had most of their energy, are I

wider than the 1/3 octave bands used in the calculation of the overall 4 dBA level. This would lend to cause over-estimation by dE A measures of the loudness of these predominately low frequency noises.

Also, as noted above, 'foronfeff, et al. suggested that there was possibly some masking in the lower frequencies by ambient noise, making

I

! then less audib!c. The fact that ' the Test Transmission Line noisc spectra was the one most similar to the Federal Signal siren dictated, at that time, the use of the Horonjeff, et al. data for the Test Transmission Line noise as the most appropriate available for estimating the arousability of the siren.

3. Duration-Presenec. Obviously the longer a sound is present during sleep, the greater is the probability of arousal given the cyclic variability of the sensitivity of the auditory systen during sleep. In addition, it is established from psychoacoustic research that although the r.icrent-to-moment loudness remains essentially constant, a sound or noise continuing in time becomen more unwanted and noticeable. Its presence over time is perceived as having, usually, an objectionable and arousing characteristic.

Judgement tests of noiseness or unwantedness when people are i awake, and numbers of awakenings when asleep , are reasonably well correlated with the sunmed A- or PNdB-weighted energy in either steady or interrupted broadband sounds presented over relatively short periods (minutes to hours) of time. It is believed that the conversion of j I:rallmann's data to SEL's (adding 16.5 dB to peak dDC for 45 second I durations - 3 dB to convert dBC to dDA) provides sound exposure units j

directly comparable, in this regard, to those used for the Horonjeff, et i

al noises and the FST model 2000 siren.

4. Saliency. Figure 1 shows that curve 3 for the deepest sleep period, 0-1:15 a.m., of the Krallmann data indicate about 17 dB greater effective arousability, for a given SEL, than the floronjeff et al Test Transmission Line noise. According to conventional loudness data and  ;

l i

I concepts, the noise of the Test Transmission Line should be equal in loudness to thet of the siren because they are equal SEL, and both have significant amounts of unmasked energy in the mid-to-high frequencies.

The most apparent explanation for some of this difference is that the siren signal is more salient of noticeable (and sleep arousing) thm the j

broader spectrum Test Transmission Line noise even when equally loud.

Eryter and Pearsons found that the presence of some pure-tone frequency components in a broad-band noise caused the noise to be i judged as being significantly more objectionable or noisy than loud. The I

] degree to which this was the case depended upon the tone-to-background j noise ratio and the frequency of the tone, as is shown in Figure 3. Note that according to Figure 3, a correction of up to 9 dB would be added to I the dDA level of the siren signal in erder to properly assess its judged noisiness as compared to that of a broad band noise of the same dBA or J

PrdB level. As seen in Figure 1, a 9 dB correction would be adequate i

to make the Urallmann data, when also adjusted for equal 1

exposure-habituation (to be discussed below) , consistent with the

!!aronjeff, et al. data for the Test Transmission Line data.

Whether this " extra" unwantedness, beyond that due to loudness, 9

makes pure-tones, or very narrow bands of sound at about 400 Pz or so relatively more sleep arousing hac not , to my knowledge, been r:ystematically studied. However, the results of the Krallmann study s

strongly indicate that this might be the case. (It is noted that

, Zimmerman exposed subjects to 1 second pulses - 1 second on, I second

]

off of an 800 Hz tone and found after 8 seconds, at a level of 63 dDA, i

l SEL 60, 100% awakenings. The temporal nature of these impulses are perhaps too unusual to be considered in the present discussion.)

i a

. ._ - - =. . _ . . - - .- ..

10 -

r

.. I oe ..i.ae I uson i. eel ue.n.. 1 I I !

• • *-~'~'*

=iTAitw e taitw=itaitw. w ...

i 1 Mi M l 25 ! 25 Imi 2 ====r=***w

!j E

g IC 2 1 40 i 10 , 15 t ,$ 13l 3 ,,,,,,, ,,,,,,,,

w -.*=~.*=.

l' s

3 I C1 4 l 40 I 3 I5 t o l .5 i e i to l ic t i oi 3

~ w.

.w 2 8.

l

~ -

l,*3...,-- ;,it

,g ". , s t ,

2". _e . e"_,*..*.-

. . .r . * *

1 l

', , 7 i

i a a e n , , , , , , ,

,, 000 rm f ase center fregancy - na F i g u re 3 . Cometion ia decibels, to be applied to the $PL of a broadband sound containing a pure tone or very narrow band of energy. From Kry t er and Pea rsons .

I r

l i

f I

}

4 i

4 l

1 l

1

Another possible interpretation of the auditory attribute of saliency and sleep arousal therefrom, as distinct from overall loudness, is that the saliency and sleep arousal from hearing but one frequency component of a noise is just as arousing as hearing its total loudness.

5. liabituation. In the Krallmann experiment each subject apparently partleipt.ted for up to 4 consecutive nights of sleep, whereas in the !!cronjeff, et al. study the subjects were tested for 21 consecutive nights. As seen in Figure 4 there is some habituation-accomodation, Icr. ding to reduced sleep nrousal, to noises according to the number of sleep nights involved. If the data for the Krallmann study represent the averogc arousability over trio nights (the average for the first, second, third end fourth sIcep nights), according to Figure 4, had his study.

continued for seven or more nights the arousability percentages would have dropped by about 10 percentage points, i.e. from 45% to 35%,

equivalent to about an 8 dB difference in SEL.

For purposes of making comparisons with the IIoronjeff, et al.

study, it is perhaps appropriate to adjust the Krallmer.n data downwards (or the lloronjeff, et al. data upward), by 10 percentage points to allow for equal habituation to the experimental conditions. This adjustment is shown in Figure 1. liowever, the actual Krallmann data without any adjur,tment may be proper for purposes of estimating arousal from the -

stren in the Shearon liarris EPZ in that to a large extent the people will not be habituated to sleep arousal by the stren stFnn!, although they will obviously be more " habituated" to the sleeping arrangements, being in their own beds.

theoretical habituation to scoustical stimuli during sleep; number of swekenings in the first test night: e.g. 50%

75

- 1 I I i l 8 0-reac tuans

" I ' "

f65 / .... . ..... ..s. .

(

.5 55

& - for example -

45 ' ' '

x '....4...e.,.,..,...,

4 C l E 35

.k .

~,,, ,, . t e on ,,,s 25 0 1 23 4 5 6 7 8 9 101112 test night (averagel 0-reactrons = reactions less then a enonge in steep store esiculated from: 8 pubirentrons wrth 12 subrects on 802 noents after 8130 noose stomuti.

oortreft nesse, cann noose, truch noose (4046 do(Al1 Fi gure 4. si.ep disturbances due to noiw From Griefahn and Jansen.

l

.i e

J

-- . - , - - .- -. , . . - , , ,, , - - - - - , . - , - - .-, c- -. --

(One of the uncertainties about the Krallmann data is the effect on 4 sleep patterns of dormitory sleeping for a period of one week as compared i to sleeping in one's own bedroom. It is conjectured that s!cep arousal mcy be enhanced in the dormitory situation, at least in the periods of light sleep. This is one reason that it will be proposed below that the Krallmann data for the " deep sleep" period of 0-1:15 a.m. be used for purposes of comparison with the lloronjeff, et al. data and for estimating

! arousals by the siren in an EP2. Attempts are being made to contact Dr.

F,rnilmann to obtain possible additional information on this and other details of his sturly.)

6. Depth of sleep. It is known that the case with which people are aroused from sleep is dependent upon the stage of sleep and, although these stages are somewhat cyclic throughout the sIcep night, most of the deertmt sleep periods occur v:! thin the first hours of sleep. However, there variables have not generally been controlled or systematically investigated in most sleep studies. For example, Horonjeff, et al.

presented arousing signals scattered throughout the night but averaged the responce data over the entire night.

A rigrnificant feature of I:rallmann's study is that the percentages of sleep arousal from the stren are reported for different clock hours during the night, from midnight to 5 a.m. (0-5:00 a.m.). His data are shown in Fig.1, curvec 1, 2 and 3, and in the following table.

t 1

. _ . __ - . -. - . . - - . - . - .~ -_ . - - . . - - . -

l t

i i

0-1:15 1:30-2:30 4:00-5:00 Av.0-5a.m.  :

a.m.* a.m.** a.m.***

i dBC dBA SEL Aver. Krall. ,p .17 o

60 57 74.5' 6d' 82 97 80 79 l 55 52 60.5 62 77 92 77 72 t

50 47 64.5 52 67 88 60 66 45 42 59.5 44 47 64 52 59 40 37 54.5 39 58 82 60 52 j t

• Krallmann, Figs. 12, 14, 16, 18, 20.

" Krallmann, Figs. 21-25.  !

"* Krallmann, Figs. 26-31.

C!corly, from the data in the above table, and as shown by Curve 3 for r the Urallmann data on Fig. I., the percentages of arousal for the 0-1:15 ,

a.m. period are by far more conservative than the average for the night l overall, fren 0-5:00 a.m. The Horonjeff, et al. data (averaged over the  ;

entiro nirht minus I hour before normal waking time, when their experi-ments were always stopped) could also be expected to overestimate arousals from nelse relative to the 0-1:15 a.m. data of Krallmarn. I Hnwever, it is the writer's opinion, based on personal experience and  !

discuscions with other researchers of sleep, that the data in most sleep l stud!cs tend to be collected during the earlier (deeper stages) rather than the later hours of s!cep. All things considered, it is estimated that the prohnble somewhat underestimation of arousals from the Krallmann '

0-1:15 a.m. data relative to the lloronjeff, et al. approximately-all-night  ;

data is compensated, for purposes of this comparison, by the overestima-l tion of arousal in the Krallmann data because of his dormitory sleeping ,

l arrangement for the subjectr as compared to own-bedroom in the  !

Iforonjeff, et al. study.

7. Age of subjects. As described in the original Kryter testimony, j arousability from sleep is somewhat a function of a person's age and a 1  ;

l 1

precedure for adjusting experimental data, or estimates, of sleep arousal for different age groups was furnished. No adjustments are required between the Krallmann and IIoronjeff, et al. data because the average ages of both groups of subjects are cbout the same, 43 years.

Conclusions. Issue #2

1. Differences in arousability due to differences in sleeping conditions (dormitory vs. private bedrooms) are estimated to be such as to approximately cerrpensate for differences in arousability to be expected when ecmpuring the deep-sleep period (0-1:15 a.m.) data of Krallmann with the overall sleep night data of lloronjeff, et al. Accordingly, the greater arousal (equivalent to about a 17 dB difference in SEL) found between the respective Krallmann stren and the !!oronjeff, et al. test i

transmission line noise data can, it appears from present psychoacoustle 1:nowledge, be explained and reconc!!cd by:

(a) A "Falloney", or tonal-detectability factor equivalent to about 9 dP for the 400-500 I!r frequency region of the siren signal;

! (b) A "llabituation" factor equivalent to about 8 d!1 for the 4 nights (average of 7) of exposure in the Krallmann siren study as compared to the 21 nights (average of 10,5) in the Itoronjeff, et al. study.

2. The greater sleep arousability of the test transmission line noise relative to the other !!oronjeff, et al. nolces (transformer, air conditioner and traffic) and to the fiUREG/CR-2654, PflL-4227 guide (which appr:ars to be based on an averaging of rJ1 the lloronjeff, et al. data) is probably due to the predominately low-frequency, below 400 ffz or so, content of

i i

the other noises, and possibly to some extent masking of these noises by

typical ambient bedroom noise.

l

3. The 1 rallmann data for the 0-1:15 a.m. sleep period is taken to be a conservative means for estimating the percentage of one person, age i 45 years, arousals to be expected from exposure to the FST Elodel 1000 siren signal at a dven SEL indoors. [

i l

I

} L 1 .

e i

1 1

(

l t

I i

}

4 I

1 a

i i l

t i

i i

1 I

! [

t i

. References Iforonfeff, R.D. ; J4nnet t , - R.L.g., and Teffeteller, S.R.: Sleep Interference, EA-1240, Vol. 2. Project 852, Electric Power Research Institute, Palo Alto, CA 94304 (1979).

Iloronjeff, R.D.; Fidell, S.; Teffeteller, S.R.; and Green, D.Bl. :

Behrvioral Awakening as Functions of Duration and Detectability of Noise Intrusions in the !!ome. J. Sound & Vib., Vol. 84 No. 3, 1982, pp.

327-336.

Griefahn, B. : and .Tansen, G. : EEG-Responnes Caused by Environmental l'oise Duri:,g Sleep -

Their Relationships to Exogenic and Endogenic In fluences. Sci. Total Environ. , Vol. 10,1978, pp.187-199.

Krytc r, l{ . D . and Pearsons, K.S.: Judged Noisiness of a Dand of Random Ncise Containing an Audible Pure Tone, J. Acoust. Soc. Amer.,

Vol. 38, No.1. July 1965, pp.106-112.

NUREC/CR-2654, PNL-4227, Procedures for Analyzing the Effectiveness of Siren Systems for Alerting the Public , Prepared for Nuclear Reg.

Commission, Washington, D.C. 20555, Feb.1982.

Zimraernan, W.B. : Sleep hientation and Auditory Awakening Thresholds, Psychophysiology, Vol. 6,1970, pp. 540-549.

Issue #3. Explanation of method for converting dBA levels to SEL.

The Memorandum and Order (Limited Reopening of the Record on Edt!!cman Contention 57-C-3) page 7 cites the following formula for a 3-minute exposure to a siren at 4 RPM:

Indoor SEL = dB A + 20 - attenuation The comparable calculation in the Kryter Testimony (p. 26) was based on data, turnished by Dr. Van Lee, from the Federal Signal Company.

There data show a duration to the 10 dB downpoints from the peak level of about 4 ? seconds for the Model 1000 siren operating at 4 PPPl. The rise and decay times around the peak are not perfectly linear, and include about 0.4 dB more energy than would a linear rise-decay time pattern. This was considered no negligible for reasons to be discussed later. Accordingly, the calculation in the Kryter Testimony was:

frdoor SEL = dBA + 10 - attenuotion (who rt. SG stron bursts = 75.6 seconds = 19 dD; or 4 HPfl x 3 min. x 3 activationa x 4.2/2 seconds l "4.2/2 seconds" assumes a linear rise-decay pcttern.)

The 1 dB differcree between the two calculations (+20 vs. +10) rests, I believe, on the finding in the Nemorandum and Order that the rise-decay pattern is significantly non-linear, as shown in Fig. 5. The inercased sound energy contributed by the area above the linear. straight line, pattern amounto to about 1.1 du, or dDA + 20 (.1) = SEI.,

t,'hich of these directivity patterna prevalle is a complex acoustical-atmospheric function depending on the distances of the measuring points from the alren. At closo distances (100 ft.) the pattern

4

- 19 -

s i

.I *'  !

l l' l

II r1 % p$l.qp $f d(A W LI' s d? VC i

< l' Y i I p Y

(V .

I ,b i

lh- i, i

f m -

. 4

,g'l b's ,

i, b h i

A i

'I t w

I i 4

• I  ; wx ,- .. ,

v c 1

, m

, Q r ,. 1L&-

t ., ,,. ,

u,,

l i?/.a d i

h *

,s

,i i

e a

5 f -

[  ;

h i

s l , 'i I 5 j',l' s I

[i t

i s e

f v

(

l g

, ....+.

M '

,< sll l' ,.lif '

'sa TY

.sl a A 6 M ,..

li sifs Ti hlt- ll 9

n 3

,c l *,1

,i y .y'

-l " l l q l' TJg ' pl[

1:

l

...m,. 3 .n..... . . . . . .. . , , . . . . . . .. , . . , . . , . , .., ,

i

{ '_ p - n . . . i r ri ,in r,iti l .m o n c. .n , , n + 3. .,

.. ~ ,

j  !

is more non-linear (" rounded" at its peak) than at farther distances (4500  ;

i i ft.), as seen in Fig. 6 compared to Fig. 7. As the circumference of the circle is enlarged around the siren (points more distant) the changes observed in Fig. 7 are to be expected because of radiation from lateral i points phenoraena and atmospheric perturbations due to greater path

{ lengths involved. In so far as Fig. 6 is representative of the FS Model 1000 siren level-temporal-pattern, (estimated to be about 5 secondo, rcther than 4.2 seconds, to the 10 dD downpoints) 20 dB should be 1

) ad&d tc peak dBA to calculate SI:L; if a significant non-linear riso-decay 1

pattern is assumed, the number to be added should be 21 dB rather than 4

> 20 dB.

i

! It is concluded that the conversion used in the original 1;ryter

)

.cstimony for peak dBA to SEf, of the FST Model 1000 stren may be j conservative by I to ? dD.

t l l

1 .

1 I 1

t 4

l l

i i_ _ _ . - - - - - , . - - - - . . - _ . . . , - , _ _ . - - - _ . . _ . _ , . _ _ . - - . - _ _ , _ _ _ _ - _ - - - - . . - . _ _ _ _ _ - - - - _ , _ _ , , --,,-_..- - _ __--._

JG .. . . . . . . _ . - .

First A cti v atio n _ . .. .-

.,n

..- - ....,......t%..-..-.===

~r . - - - - - -.

.. o

---a-- ---.

9 ..

100 ft, from

.. .p . i.

1

...._._,.1_-..._ . __. . 1.. .. _ . .

-t ttc

- . . ;- - - .:.: 4 :_ n ._ .c_.-J. 4 .. siren

- ~ -

c.

H A

10 0 h.4MC jWM.m  :--.__p - " ~ +. :Uiw .ma:O_a~ -- WW u.

Me-y

,m .

. . .a _,. i. >, ;y T I 90 t ,

i si...> <

0 30 60 90 120 i -

150 o TIME (SEC'6NDS) m D

Se con d A ctiv a tion

• . ' * ~

_- . A. - 1.~ .. ./  ; i.p-. . -- - , .. .~.-4 . .:

m

\- .

> s

..... .. y _, -4

~

.g y 99g 4 _ . - . .

g g _i < -. - ..

.m _

w ~

Z 10 0 '

2 ,

- r, -

-- [ . ;4,,-

• 90 e i i i -

y 0 30 x 60 90 c 120 150 c TIME (SECOND'S) g Thirc Activation 2 . 30 _ _ . .

3 .. _ . . .

O uJ 10 0

.  %. i.

........4..

. . 7 . .. ~20 .

~

- . :4 .. _._q _ .. . 1.. . _ .s. u. - .. a_.

110 t

.z^ . -r. w_ n 5 -f--=.se FAIW w m _&. t ', .

i toc m1 Dy89 g.t.,=.e

_4Mf .  %}l!Y.F- y f,ti.j--+ f i . _ _ ... _ .

.._.a ,

~

go i

i i -

0 30 60 90 i

120 150 TIME (SECONDS)

SOUND PRESSURE LEVEL VARIATION VERSUS TIME OUTPUT AT 100 FEET FROM ROTATING SIREN SIREN NUMBER: 4 hAccusric TECHNCL.0GY INC Tigurela, from letter of January 8,191I0 to Administrative Judges f Xelley, Dright and Carpen,ter in the matter of Caroline Powe r Lt.tsht Company etc , f ropThoms a An __ D0000f2o_Jhorn9 M Aco" =* ' - --

__ _ - - ~ - ~ - ~ ~ ~ ~ ~

. t_- - -

- ? -- Mossutomont Location: 10

,0 i' f""T" ,

_p p, 1 _.

1 4. _ .

T ,

< i't, 60 > u 't3 r. e1%1, . r .

1 3Y--

o g.,

p , ,. 4 g_

. ..g .,4 _ g. g _'

4 , , _

-- i.  :. q 1

,a. .f.

50 M - -

'_' - [ #~ f 4800 ft, from Siren p3  :,

9 k j

~

40 z- ll -

t:-

p.I .

s- >> 1-s i e m C 30 60 90 s

120 o 150

= TIME (SECONOS) 0 Se conc Activ ation

~

2 80 _~ .

Measurement Location:6

.- ~ -

,,3 , -

W '

3 --

re- - -+ - -

4800 it,

'- rom w _ ,-- c4 , 4

a 60 e r . .. q

.L- g, ,  ;' s Siren i .

g 1 1 1 3

W 1,

_ _ 4-* r -

m R c 50 , ... n - **.

z

.,- I 3

1

, g_ ..,t. .,4 ,, -  ; . ,

-- i.:,

_3; g . .

_ 43 a -

, :1 , o ,. .

e 'O s ig- -

1 i i i y 0 30 60 c: 90 120 s

150

c. TIME (SECONDS)

=,

Thire Activation 3 80 - Measurement Location:2 o ~~ ~'

M 70' . 4700 tt, t ror-L

.t f_t '

I n - Sirer y X* *.

da 2

, ;y --+-

y

, . ;_ ,__- , ,- g ,, _ , _ --

s e a J, 9 + i- l- . i N*+--- ' -

50 - 4+4 4th

-3

. ..r .p g

ng i 40 i 3.-

.' :i' .-*M

, .i' 1 ,,

6 a - - -

0 30 60 a . . -

90 100 150 TIME (SECONOS)

SOUND PRESSURE LEVEL VERSUS TIME OUTPUT AT FAR-FIELD MEASUREMENT LOCATION 500 Hz OCTAVE BAND SIREN NUMBER: 4 Acourne TaCHNOLOGY INC.

Figuro7. from letter of January 8,1 o to Administrative Judges i

Kelley, Bright and Carpen,ter in the matter of Carolina Power Li,tght I

Company etc, fromThomas A. Dexter, Counsel for Applicants.

i