ML19343A173
| ML19343A173 | |
| Person / Time | |
|---|---|
| Site: | 05000142 |
| Issue date: | 12/31/1976 |
| From: | Rubin M CALIFORNIA, UNIV. OF, LOS ANGELES, CA |
| To: | |
| Shared Package | |
| ML19343A172 | List: |
| References | |
| NUDOCS 8009160322 | |
| Download: ML19343A173 (64) | |
Text
{{#Wiki_filter:-- - - - . _ - _ _ _ _ _ _ - _ _ . O.
@ Copyright by Mark Phillip Rubin 1976 8 0 0 916 0.31G-
.a e
i The thesis of Mark Phillip Rubin is approved. 4 s 4A
; 6 W' Ivan Catton Niudr*kI -a% & , "
lJ.'es'G.E(inger
/
/ l
{dr__- . Willia:a E. Kas nberg, j:or=ittee Ch .a n 4 University of California, Los Angeles 1976 f t 11
. - _ . . ~ . - _ . ~ - . .
TABLE OF CONTENTS Page LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . . iv LIST OF FIGURES . . . . . . . . .. . . . . . . . . . . . . . . . v LIST OF SYMBOLS . . . . . . . . . . . . . . . . . . . .. . . . . vi ACKNOWLEDGMENTS . . . . . . . . . . . . . . .. . . . . . . . . . vil ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii CHAPTER 1. INTRODUCTION . . . . . . . . . . . . . . . . . . . . 1 CHAPTER 2. EXPERIMENTAL PROCEDURE . . . . . . . . . . . . . . . 8 CHAPTER 3. RESULTS . . . . . . . . . . . . . . . . . . . . . . . 18 CHAPTER 4 COMPARISON OF EXPERIMEhTAL RESULTS TO MODEL PRE-DICTIONS . . . . . .. . . . . . . . . . . . . . . . 44 CHAPTER 5. CONCLUSIONS . . . . . . . . . . . . . . . . . . . . 51 REFERENCES . . . . . . . . . . .. . . . . . . . . . .. . . . . 55 iii
LIST OF TABLtd Table Page 1 SF C ncentrations for June 22 Experiment . . . . . . . . 23 6 2 Argon Concentrations for June 22 Experiment . . . . . . . 29 Concentrations for July 8 Experiment . . . . . . . . . 33 6 4 Argon Concentrations for July 8 Experiment . . . . . . . . 34 5 SF Concentrations for July 27 Experiment . . . . . . . . 38 6 6 Argon Concentrations for July 27 Experiment . . . . . . . 39 7 Composite of Argon Concentrations . . . . . . . . . . . . 40 3 Relation of Turbulence Types to k'eather Conditions . . . . 47 9 Pasquill Stability Categories . . . . . . . . . . . . . . 48 10 Parameters i;r Model Calculations . . . . . . . . . . . . 49 11 Dispersion Model Predictions . . . . . . . . . . . . . . 50 12 Percentage of Federal limits . . . . . . . . . . . . . . . 52 i 1 1 iv i i i
)
LIST OF FIGURES Figure Page 1 Gas Chromatograph Output from SF 6 Analysis . . . . . . . 12 2 SF Calibration Curve . . . . . .. . . . . . . . . . . . 14 6 3 Meteorology Data for August 11 .. . . . . . . . . . . . 16 4 Path of Plume with Southwesterly Wind . . . . . . . . . . 19 5 Math Science Roof Sampling Locations . . . . . . . . . . 20 6 Campus Sample Locations . . . . . . . . . . . . . . . . . 22 7 Plume Path for June 22 Experiment . . . . . . . . . . . . 26 8 Plume Path for June 22 Experiment . . . . . . . . . . . . 27 9 Plume Path for July 8 Experiment . . . . . . . . . . . . 31 10 Plume Path for July 8 Experiment . . . . . . . . . . . . 32 11 Plume Path for July 27 Experiment . . . . . . . . . . . . 36 12 Plume Path for July 27 Ekperiment . . . . . . . . . . . . 37 13 Smoke Plume from June 22 Experiment . . . . . . . . . . . 41 14 Smoke P!ume for July 8 Experiment . . . . . . . . . . . . 42 15 The product of c a as a Function of Downwind Distance from the Source ? i . . . . . . . . . . . . . . . . . . . 46 1 V
LIST OF SYMBOLS cc cubic centimeters m meter al c111111ter ppm parts per million ppt parts per trillion SF sulfur hexafluoride 6
-0 uc 10 curies l
1 1 I vi
ACKNOWLEDGMENTS The analysis equipment necessary for SF6 tracer studies was made available by Professor Frederick Shair of the California Institute of Technology Department of Chemical Engineering. His help, and that of Brian Lamb, was invaluable in completing this study. I would like to thank the staff of the Nuclear Energy Laboratory; Jack Horner, Neil Ostrander, Charles Ashbaugh, Ron Bolek, and Donald Browne for all the help they gave on this project; along with Phyllis Gilbert who typed the final copy of the thesis. My grateful appreciation is extended to all the students who spent long hours walking around campus taking air samples for these experi-ments. I also would like to thank Leslie Cave who suggested this study, and my cocmittee members, Williac ' . Kastenberg, Ivan Catton, and James G. Edinger. vii
ABSTRACT OF THE THESIS Atmospheric Dispersion of Argon 41 from the UCLA Nuclear Reactor by Mack Phillip Rubin Master of Science in Engineering University of California, Los Angeles, 1976 Professor William E. Kastenberg, Chairman An experimental program was developed to determine to what degree the general public has been exposed to radioactive Argon 41 released in the exhaust gas from the UCLA Nuclear Reactor. Actual Argon 41 concentrations were too low for direct measurements, so the chemical tracer sulfur hexafluoride was used to predict Argon concen-trations. The SF was injected directly into the reactor exhaust 6 stack. Downwind air samples were taken with hypodermic syringes. The air samples were analyzed using gas chromatography with electron cap-ture detectors. The data was used to determine averaged Argon expo-sures and was compared to the predictions of the wake cavity mixing model for atmospheric dispersion. The results presented here show that Argon releases from the UCLA reactor result in concentrations to the general public that are well within federal limits. Additionally, the results showed that the wake cavity mixing model is conservative in its prediction of atmos-pherie dispersion. viii
4 CHAPTER 1 INTRODUCTION The work reported in this thesis was undertaken in support of 1 an Amendment [1] to tne operating license for the UCLA nuclear reac-i tor. In 1974 the annual facility review by the Nucluar Regulatory Commission (NRC) showed that previous estimates of the amounts of activated argon gas (Argon 41) being released to the environment had been severely underestimated. Concerned with possible radiological exposures to the general populace in excess of that allowed n Ne Code of Federal Regulations 1 (10-CFR-20) (2], the Nuclear Regulatary Commission restricted the oper-ating time on th( UCLA reactor and ordered a review by the Nuclear Energy Laboratory (NEL) staff as to what steps must be taken to assure that the UCLA reactor was in compliance with federal regulations. The federal regulations which apply to this situation stipulate the concentrations of radioactive species which can be released to the environment. The problem confronted by the NEL staff was to ensure to the satisfaction of the Nuclear Regulatory Commission that at no location beyond the controlled release point of the reactor stack, did Argon 41 concentrations exceed 4x10" ue/ml, the limit set forth in 10-CFR-20. In this work an improved monitoring system utilizing a new, = ore sensitive ioni:ation chamber and associated electronics, yielded accurate values of the quantities of Argon 41 leaving the reactor 1
stack. The stack is located on the eighth floor of the Math Science Complex. However data was lacking on the atmospheric dispersion which takes place ir. the vicinity near the stack. Argon 41 concen-trations in the stack gas flow itself were definitely beyond allow-able limits, so it was necessary to determine the dillution that took place in this stream before it impacted on locations where possible exposures to the general populace could occur. Te abtain data on the atmospheric dispersion of Argon 41, prior to this project, the NEL staff embarked on the experimental program to capture and measure air samples in the vicinity of the reactor stack during reactor operations. The program entaileo taking 100 ml air sa=ples in plastic squeeze bottles by pouring water out of the bottles and latting air displace the wat.r. The bottles were then placed on a 3 inch Sodium Iodide scintillation crystal connected to a multi-channel analyzer. The goal of this program was to determine how much radioactive Argon 41 was contained is this sample of 100 ml of air, allowing calculations to be made on Argon concentrations in the sampled areas. Unfortunately poor detector geometry, noise electronics, and a low Arb m concentration =ade it impossible to distinguish the Argon being released by the reactor from natural background. Ulti=ately (for the purpose of supporting Argon 41 concentration claims to the Nuclear Regulatory Commission for the license amendment), atmospheric dispersion estimates were made by the consulting firm Applied Nucleonics, using a modified Gaussian dispersion model based on the work done by Pasqual [3]. This is the most videly accepted 2
dispersion model. However Caussian models were originally developed to describe atmospheric dispersion over relatively long distances (cens of kilometers) and over level ground with no perturbations, such as buildings. It would be expected that the effect of large buildings and roof structures, such as exist around the UCLA reactor 4 stack. would tend to cause turbulence in the wind scream and increase local dispersion, lowering Argon concentrations. However direct evidence did not exist to support claims of actual dispersion coeffi-cients. Therefore the research described in this thesis was under-taken to provide experimental verification of dispersion estimates provided by the mo m . The primary goal of this research was to determine experimentally the Argon 41 concentrations which would actually be present when the UCLA reactor was in operation. Many methods of detection were origi-nally investigated. To claim validity, any detection system should be able to detect concentrations of Argon below the permitted NRC release limits. This mandates the. detection system to a sensitivity
-8 j of better than 4x10 uc/ml.
From the onset of the experimental progra= this criterion posed a problem. Concentrations of this level equate to a mass concentra-19 tion of 14 atoms /ce; a mass ratio of over one part in 10 of M gon to air. This is many orders of =agnitude, the detectability of mass spectroscopy or atomic absorption. This suggests that some =ethod of radioactive decay analysis is the only way to achieve the sensitivity necessary. Several schemes utilizing this method were investigated. Detection systems based on either large volume ion chambers or large 3
crystal (5 to 8 inch), good geometry, scintillation systems appeared , to have the ability to measure the required concentrations. However the required systems were not available at UCLA, and building them would run into thousands of dollars for the ion chamber, and tens of thousands of dollars for a scintillation system. Since virtually no funding existed for this research, finances seemed to preclude the development of the required radioactive decay detection system. The results of a preliminary study, up to this point in the research, indicated the inability or impracticality of available methods for actually detecting the Argon 41 being released from the UCLA reactor. The one remaining alternative, use of chemical tracers at first appeared unusable for the situation being studied. A method which had been used successfully in power plants, fluorescent par-ticles into the exhaust gas, is a technically messy operation which would have exceeded the financing of the research. Use of chemical tracers such as carbon =onoxide, or various hydrocarbons, were attractive in that they could be detected by smog research analysis equipment which was available locally, and at a cost that was within the ability of the project to finance. However the high and varying background levels of these compounds in the Southern California basin made it impossible to achieve the detection sensitivity needed. A relatively new chemical tracer, sulfur hexafluoride, just recently introduced for atmospheric studies, was investigated and seemed very promising for the UCLA study. The tracer, known as SF6 ' appears in only very minute quantities in the environment and is 12 The upper limit of detectability easily detectable to one part in 10 . l 4 , i I L J
is one part in 10 0. Therefore, by releasing one part per million of SF t air in the reactor exhaust stack, it would be possible to 6 measure atmospheric dispersion factors of one million. This turned out to be more than sufficient. What made this technique especially attractive was the fact that this tracer was being used by the Chemical Engineering Depart =ent at the California Institute of Tech-nology in Pasadena. An arrangement was worked out whereb; the analysis equipment needed for SF tracer studies was made available for a very 6 nominal fee. While use of a chemical tracer precluded the actual measurement of the Argon being released from the UCLA reactor, this technique appeared quite satisfactory. Use of the tracer gave accurate dilution factors from the reactor exhaust to any point on campus where gas samples were taken and analyzed. Since very good instrumentation exists for determining the Argon 41 concentration within the exhaust stack itself (the source concentration), it was a relatively straight-forward matter to calculate the actual Argon 41 concentrations at any desired spot during a reactor run. Hewever, for this technique to be valid, it is necessary ta deter-mine that SF w uld behave the same as Argon in the exhaust gases. 6 Both SF and Argon are relatively inert, so it would not be expected 6 that chemical reactions would deplete either material during their dispersion. The only major difference between the two materials is molecular weight. SF , with a molecular weight of 146, is over three 6 times heavier than Argon. This u uld suggest the possibility of SF 6 ecting differently than Argon due to buoyancy effects. However 5
- . .__ . . - - _ _ . __ _ _ _ . _._ ~_
4 meteorological studies have shown that diffusive separation, as the ; separation of heavy and light gases by gravity is termed, proceeds I extremely slowly except at great altitudes. Winds have an opposing , } mixing action, so that below 70 km light and heavy gases act uniformly and do not separate. Argon 41 is radioactive and will decay while i-SF will n t. 6 However since the Argon transit times from stack to i
- sample points are under 4 minutes, essentially no decay would take place in the isotopt which has a 1.83 hr halflife. Therefore it is expected that the use of SF 6as a tracer for Argon would yield good predictions.
Federal regulations limit the release of radioactive nuclides t
-8 to the environment. In the case of Argon 41, this is 4x10 c/ml. r However there is no stipulation on the maxi =um instantaneous concen-i tration that can be released to the environment. Radioactive releases can be averaged both over the time the release is =ade, and the occupancy factor for the location where the release is made. If run-ning the UCLA reactor caused an Argon 41 concentration of 4x10" uc/ml f
at some specified location accessible to the public, this would be in excess of legal limits. However if the reactor only ran 5% of the time, the actual instantaneous concentration can be averaged over time to give an actual' exposure to the public of 2x10~ uc/ml, which is within federal li=its. Similarly if this specified site is only occupied 10% of the time the reactor is operating, the calculated . exposure to the public can be reduced by another order of magnitude. , There were two separate areas on cSmpus that appeared to offer
- the greatest possibility of excessive radiological exposure to the i
6
public. The first area was on the reLf of the Math Science Complex, near the reactor exhaust stack. Occupancy factors here are low (10*.), but this area is near the high source concentration of the reactor stack. The second potential hazard area was within the Math Science buildings themselves. Prevailing southwesterly winds blow the gas plume from the reactor stack directly towards one of the =ain air conditioning inlets for the building. Since there are other air inlets to the buildings Argon concentration within the building would not be expected to be as high as on the roof itself. However the much higher occupancy factor in the building (100*.) could possibly give high averaged exposures. The investigation described in this thesis centered on determin-ing exposures to the public for these areas, as well as trying to locate other areas which might experience high Argon concentrations. Additionally, the adequacy of the previously mentioned mathematical model for atmospheric dispersion, as it related to the situation at UCLA, was determined. In chapter 2 the experimental procedure,is outlined. Chapter 3 contains the results of the tracer experiments. Chapter 4 gives comparisons between the mathematical dispersion model and the actual experimental results. Chapter 5 contains the conclusions drawn from the work done in this thesis. ; i 1 i 7 \ l 1 l
CHAPTER 2 EXPERIMENTAL PROCEDURE For the chemical tracer studies described in this thesis, a tracer gas, sulfur hexafluoride (Sf ), 6 was injected directly into the ventillator intake of the reactor high bay. This is the same route which Argon 41 would follow after it is produced by the reactor. The SF 6 was contained in a pressure cylinder and delivered through a single stage regulator. Since a constant flow of the gas was extremely important, some thought was given to the use of a two-stage precision regulator. However SF6 is a liquified gas and, at t 70*F, is a two-phase mixture at 340 psi in its pressure cylinder. Since a two-phase mixture existed in the pressure cylinder and delivery rates would be low, it was determined that the pressure vari-ations in the SF cylinder would be minimal, and that a significant 6 improvement in constant delivery rates would not be obtained with an expensive two-stage regulator. To assure that the flow of the tracer gas into the exhaust stream was constant, SF6 g s flow was metered through a precision Gilmount flow meter with vernier controls. It was desirable to release less than one part per million,by volume, of SF int the reactor exhaust stack gas. This would allow 6 the additional check of direct sampling of the exhaust gases for SF g concentration. If tracer gas flow calculations proved inaccurate, this sampling would provide a correct source concentration for SF *6 l At concentrations in excess of one part per million, the analysis 1 8
4 system available f 5F detection saturated and would not provide 6 useful data. The UC* actor facility exhausts 14,000 cubic feet per minute of air from the stack. This equates to 3.96x10 ml/ minute. The flow meter being used to deliver the SF 6has a maximum flow rate for air of 260 ml/ minute. Being corrected for a gas of the density and viscosity of SF , it would indicate at full scale, a flow rate of 6 147.6 ml/ minute. This gave a volume ratio of SF e air o 6
. x10
(.37 parts per million). The first experimental run used this flow rate of 147.6 ml/ minute of SF6 . Subsequent runs had the SF 6 1"" reduced to 118.1 ml/ minute so that the flow meter reading would not be at full scale. A flow of 118.1 ml/ minute corresponded to a scale reading of 80%, and made it easier to monitor the flow accurately. Each of the. Uree experimental ru..; was three hours in length. The experiment began with the start of the SF fl w. The gas flow 6 was continuously monitored and adjusted to remain at the desired discharge rate. Thirty minutes af ter the tracer gas release begcn the first set of atmospheric air samples were taken at desired loca-tions. Samples were taken by a group of engineering students and staff using 35 cc hypodermic syringes. At each sample location approximately 30 cc of air was taken. Sample takers had their watches synchronized and all took their air samples within seconds of each other. Samples were taken at 30-minute intervals until the completion of the experiment (i.e. when there were six samples for each location). After being used to gather a sample, each hypodermic syringe was hermetically sealed to await analysis of its air sample. 9
Tracer gas flow was ceased 15 minutes after the last sample was taken. Air samples were also taken before the SF6 release began so that background concentrations of SF 6
' if present, could be deter-mined. Original background ceasurements showed quite high concentra-tions of SF6 , up to 800 parts per trillion. This caused some initial concern because the literature predicted a background below 4 or 5 parts per trillion. SF is used industrially, most co==only as an 6
insulating material under pressure, in high voltage transformers. A leaking transformer in the vicinity of the campus could have been respo.'sible for the high backgrounds. This could have severely ef fected the accuracy of the experiment. However further investiga-tion showed that the high backgrounds were due to the SF6 pressure bottle used in the experiment. Even with all valves closed, the bottle leaked trace amounts of SF . The high sensitivity of the 6 SF6 detection system being used, indicated significant amounts of SF6 ' even when no noticeable leak could be discovered in the pressure bottle. To overcome this problem, the SF6 pressure bottle was kept off campus and was only brought in just prior to the start of an experiment. Background air samples were taken before the pressure cylinder arrived on campus to arcertain whether or not any other sources of the gas existed, and to give the necessary information on what adjustments must be made in the results of the air analysis due to background presence of SF6. This background monitoring proved useful, as it showed large quantities of SF t be on campus at the 6 start of the August 11 experiment. It was believed that nearby 10 1
construction workers had broken open high voltage equipment insulated with SF . The data from this experiment was discarded. Background 6 measurements for the other experiments showed negligible amounts of SF in the campus environment. 6 Air samples were analyzed the morning after they were taken, utilizing equipment at the California Institute of Technology, Depart-ment of Chemical Engineering. The equipment was under *.nc direction of Professor Frederick Shair. The SF e ntent of ea.:h of the air 6 samples taken during the tracer release at UCLA wr.s determined through the use of gas chromatographs using electron capture detectors. Each chromatograph contained a detector of concentric cylindrical design with a 200-millicurie triciated titanium foil. The detector was pulsed at 38 volts for a 1-usec duration, and the pulse period was adjusted to 200 usec for optimum SF sensitivity. 6 Each chromato-graph contained a 3.5 f t by 0.25 inch stainless steel column ;acked with 80-100 mesh SA molecular sieve. The carrier gas was purified nitrogen. The column and detector operated at ambient temperature [4). The output from the gas chromatograph is fed to a strip chart recorder, giving a trace cuch as shown in Figure 1. The first peak on the chart is the SF 6 pe . e coc entration of SF in the sample 6 being analyzed is a function of the area under the peak. His signal comes through the chromatograph, 18 seconds after an air sample is injected into the column. Thirty-five seconds after the sample is introduced into the system, the oxygen in the sample reaches the detector and causes a very large signal which saturates the detector. The oxygen peak incapacitates the system. =aking the detector unable 11
. . Er . _ . ._4 __ , .... ,_ . _ ._. ._
r L _. _ __..,,_w.
. _ _ .. . _ . _,._..w.....
.__4.......,__._4._.__._ _ _._.
{ . _ .+ . . _ ..
~ ._.._ _ . . _ . ._._. .....p _
._ . _ . _. _ . + -
.._4..._. . . _ _
. . _ _ $,_. _ . _ . .._p . . _. .p. _.
....._9
...._. _. , _ . ...._.4...._.
- 1. _. . . ._.
.. . t_.__.
._. t, ._ _ _
- t. . . _ .. . _ . _ . _ . ._.- __. _ _ _ . . .
T. ,
.t. t ,. _._ _ .
6
) .. t ,_4 _.. .
1 -. . t . . . .. . .
. 1.. .
- x. _
. Oxygen ._ -. .
.._x _
*E. .
.,___ of _ . _ _ .__
- . .__ Je .
.a
+._..
o . . O ~ . . _ . . _.
. __ i x
.__ 4_ _ _
. .i
- r. _._ __ _. . ._ ,
.. . _ . . . . _ _ . . _ . . . .. . _...a . .._. ._ _.._-.. .
_ t, . _ _. _ . ._ . _ _ ._. _...
. _ . ... y
- v. .__ _ . . . .
_.. .., ___ 4. . . . _ _ . ____ _....___ _ . . _ _ _ . __ _ _ . _ _ . .__-_.._r _- _. . . . _ . _ _ . .. .._ _ . _-.__ _ _ . _ . . __.
$ TIME _ ;
u ... Gas Chromatograph Output from Sf6 Analysis i 1 Figure 1 pa
to respond to any further samples fer approximately five minutes. For this reason the chromatographs are usually ganged in sets of four, allowing up to 80 samples to be analyzed each hour. A sample calibration curve giving SF concentrations as a func-6 tion of area is shown in Figure 2. The chromatographs had a linear to 10 '. At SF6con-
~
range of 10 , from SF6 concentrations of 10-centrations above 10 -9 the signal becomes non-linear and increases very rapidly. For this reason, and to greatly improve ease of analysis, the outputs from the gas chromatographs were also fed into a digital integrator. Even with the fast rise time that would be seen with heavily concentrated SF samples, the digiti.1 integrators gave accu-6 rate areas of the electron capture detector output curves. The use of the digital integrators gave a useful range for the analysis system of 10 -l to 10-6 with no recalibration of the system being needed.
~
This ability to measure, from one part per million to one part per trillion, made this analysis system extremely useful for the tracer study described in this thesis. Accurate meteorology data for the period during an experiment was also needed, both for later analysis of mathe=atical dispersion models, and to ascertain that the required meteorological conditions were met during a tracer release. For this purpose the automatic wer.cher station located on the roof at the northeast corner of the Math Science Complex was used. Every thirty seconds this station would sample wind direction and speed, as well as dew point, solar radiatirn and temperature. The da:a was coded in colors and displayed on a multichannel recorder. The weather data for the August 11 experi-13
10-8 i i s jiiii i i yjiin
- i i l l ::: i i:l sin : : lisii 10-7 = - -:
=
b i _ 10-8 = _ _
=
1 CSf6 ) ! ! 10-8 = - _
=
10-10 = _ _
=
~
i _ 10-11 _ _
=
= ~
b 1 0-12 iII" i iilli" iiilloi i i ilini i i'liin 102 108 104 105 108 107 INTEGRATOR AREA SF6 Calibration Curve Figure 2 14
mental run is shown in Figure 3. That data was in a very awkward form to use. Every green dot corresponded to instantaneous wind direction, while each brown dot corresponded to wind speed. Each of these readings was taken every thirty seconds, giving approximately 360 data points for each variable per rur.. Both wind direction and speed scales on the charts were non-linear, so reducing the data for average speed and direction was a tedious proce.ss. A decision was made to conduct the tracer studies only when winds of a generally southwesterly nature were blowing. The reason; for this were two-fold. First, thic is the prevailing wind condi-i tion during daylight hours and appearing approximately 70% of the a time [3]. Secondly, and most importantly, it was determined that winds of this direction would cause the highest radiation exposure due to Argon 41 being released from the UCLA reactor stack. South-westerly winds blow the plume from the exhr:st stack directly towards a ledge 30 feet away, which overlooks the stack, and then across the Math Science Complex roof. This roof contains many astronomy installa-tions and is accessible to the general public. This is the area which was identified as the location of potentially highest exposure. Addi-tionally a southwesterly wind would blow the exhaust plu=e directly towards a majot air conditioning inlet plenum for the building. This had been identified as another potential danger area. From a radio-logical' safety standpoint, it is unfortunate that the prevailing wind conditions are those that would cause the highest radiation ex-posures to the public. However, for tha purposes of this research, it it a benefit. Determining Argon concentrations, under this most 15 l 1
E E. C C: C O eq eq
'7 r4
~
- l. , . .
.L r-1
- - s, . f*
**p e a u os -
O f.> h~
.'/A -
C
. .s f .
9 '. a Mi f,
,A-
*r' 3 C . 4 A,
,,r.
!s*.O ,c. .s- <
3 o *t. * *
;- . #- *. .,,4 m . . , ,. ****** ,. , . . . .' ' > j' ,. . g
. <:+. . i.&. - *
. o
=
a et e .,.,t s,>.. .. . 8.s*......v: 6
,n - r e.
a .< a u y r,= U .v' . 3 u 9 ,- -
.,- ,, = = =
a u g . .,
- s. . . A T
- 3. -
"3
". r . %...
e s
,. .i .. .
3 e
. . v. - = . .,.
3
. s. . .
u 2 s y
.,: :c,' . -
...t., , . . .
C
.t .
- itto ,-
- j.
C %
~. ...j . . . , .
u ...w
, s f,,*
. 1***,' .
a
- C '. . g. ',.f . s' s .'
-* . * , s' t *% % y C
8 4*.*.' y e N O a e 4
+ .,.
u r-M e e,=e
~s C . .a 1
l g - ool e L 20!)DNbb 16 l 1 l
hazardous wind condition, would give values which can be considered conse rva tive. The concentrations determined are the highest whic* would be encountered. As a further tool to investigate the behavior of the exhaust plume from the reactor stack, smoke bombs were set off on top of the stack, and photography employing special filters to accentuate the smoke trail was utilized. This was done in an attempt to identify roof-top turbulence and, possibly, building downwash and wake effects. Additionally it was hoped that the s=oke would help provide information on the velocity head of the exiting air from the stack. i 17
4 4 d U 1 iL CHAPTER 3 j- RESULTS The first tracer release was made on June 22, 1976. A large area of the UCLA campus was sampled during this run. The intent of this first experiment was to identify the high SF concentration loca-6 4. , tions. The highest expected concentration location, the roof of the i Math Science Complex, was sampled. However, additional ground level and elevated air samples were taken all along the path of the exSaust plume to determine if any unexpected hot spots existed. Figure 4 shows the path the plume would take across campus with a prevailing southwesterly wind. Figure 5 shows the orientation of the reactor exhaust stack and. surrounding roof structures. The figure also shews the locations of the various rooftop sampling sites used in the experi-ments. Not all locations were used for each experiment. Location A was the exhaust stack itself. The sampling syringe was inserted down into the stack to obtain source-strength concentrations of the tracer gas. Locations B, C, D, E, and F were on the ninth floor roof-ledge overlooking the reactor exhaust stack. Location E was in the direct path.of.a southwest wind. Location G was next to the air-conditioning inlet plenum. Location H was the Meteorology Depart =ent weather sta-tion on the northeast corner of the roof. Locations I, J, and K were inside the Math Science building, directly below the air-conditioning inlet plenum. Location I was on the seventh floor, J on the third floor, and K on the first floor. Location L was in the south wing 18
i A l B l C i s 1,i : ..: . P: t q.s _i,D i, s3 t
- g .' ; =. q,
- a. \ g g.
5
\-
T .\
.s t-
! 4 q 1..:. R _.. . f,'_-..'h
,. .,..- .. ~
.- 9, a . .
- 1. .c.
n . .i . . O, t ... y- \ "
. =: -
o '4*
- 2 -
2 N gLi
. , n "@ t e eLyA.. . .
tecwo p
.- . p .,- - r w a .. -
w.. ~. - . .- e 3 .,,... f: -- p. g
. g. . .. m x..3 _
l ? {' u. ?E a:&l' -
- y.
Q
;
- t 4 =
==.
.. n 4
. g .yg --
. .c .
,o e -
,e... '
. = . .- .
, .r.
5 5::.
- r _
b" ' 5
. UC.LA eh e
.,^~ . ..
g.s .
* :~ * ~ . '
=; -- Isin,rF fluu t ai . utTsai siti L1 *:w A l 8 l C O
D D 19 oa w .1.l - a
l espected p1wne direction N l_ . . . . l.oC H
.-~;. [- *
. . .- m .
'r N'
' Ventilatio intake l'lHl!
g" !I
-.- l
[ ,[
, Loc G ( 3 I
/. s ,-( '
l l
.'.,_, l ' SAstrottomy l m- . - . - - ---m - ~ -
i I Observotories l 9'h Floor Rool s- - V Elev 48T MSL s
~-
Loc 0' Loc D Loc F~ /%
~~~~
E.- s e ', 8 -- -- - Loc C 't cE 8 thfloor ROOI , Mot hematical Sciences Elev 475'MSL .
. . ~ . - . . . - . t jI unu - -
10 20 F EET 1~- 10 O l
~~ ~
y Reactor _ l 'f Reactor - - ~ St o c k j ., f Elev 4875 ! '*3 .. h .. 1 MSL { Doeller Hall }
,[ * = Sampling Locations II tl :=1 r; .- 1r kOOf SAP' FLING LOCAllONS ON MTM SCl[ME COMPLE A FICLIRE 5 w w - w x:, ... _ w . == r
of the b'uilding complex on the fourth floor near room 4711. Since a precise southwesterly wind could r.ot be expected, the attempt was made to bracket the path of plume travel so that, even with the normally expected wind fluctuations, the center of the plume would be encountered. Figure 6 shows the locations of the rest of the sampling points M-Z; the line drawn from the reactor stack shows the expected direc-tion of the wind and exhaust plume. Location M was at ground level next to the campus placement center. This location was selected in an attempt to identify any downwash effect from the Math Science building. Location N was the rocf of Kundsen Hall, and location 0 the roof of Kinsey Hall. These samples were taken in an attempt to identify the height of the plume. If concentrations at these heights were greater than those measured at ground level, it would suggest that the plume had not yet dispersed to ground level. Location ? was the flagpole at Royce Quad, location Q the center of Schoenberg Quad, and location R the southeast corner of Schoenberg Quad. These three samples were all taken at ground level and were meant to tra-verse a perpendicular line to the direction of plume travel. Location S was the roof of the Law School at the edge of campus. Locations T, U, and V were another traverse across the plume taken at ground level on Hilgard Avenue just off campus. Location W was in front of Royce Hall at ground level. Location X was the foot of Janss steps. Loca-tion Y was in front of the Student Union. And location Z was in front of the Chemistry Building. Locatiens V-Z were not expected to be in the plume path, but were sampled to determine that no high concentra-l l l 21 l
I e h e't t y. j if I f 4 ". b .
#0
- g ?.,Lu; ~ ' ' "f . ,1 5
\ h*
, O ,
I
#"^Yr'u'aI5No$
sCu '" k _ . .fu ' "j. (@ p'x w. M c--g 5
= u'~f il
; J 3
h +,; y. j' hl : ]' ['l8 ' Q- /* ~
-kthj 'b .....
HuwahiriES.
- C***'~ '
..$- [ l/ _
jh! {@ j a i _. -
.. . =. i . . _ _ . = .
"" I
, .$ 1 DicmSON PLAZA k s.<.57/:: . - c.ect
'$' r
! . A M E s t RHOF F ,q}
PAR E '4 c , h .
. ,/ eas statiow v'
a ,, "ir -?= c q m, l ( 1$a 'g '- J. / {I I, ., ub.
- CouAf y j
)
j $ indiesen samle location 3 o, - ., ....................
^
- ~
_ i '>00 9 *E57*000 N[,$.i, , Q, h n CA"WS 5" E '#*8
~
' '~~ '
CIRCLE drive 84 R a i N ", l U"0 g_ ,.j p, SCit NC E S i r;7 ] ,, } Q 9' 9 ' {} ' j 39 j se Ju 6 bd "
Table 1 SF6 Concentrations (ppt) from June 22 Experiment Time Location 1:30 2:00 2: 30 3:00 3:30 4:00 Averace 5 5 A 5.9x10 5.9x10 5 5.8x10 5.9x10 5 5.8x10 5 5.8x10 5 5.85x10 5 E 5292 2102 3412 1893 1232 10563 4082 F 3339 501 2041 1080 1843 3219 2003 G 626 1620 3211 1319 2079 2200 1842 H 24 15 1002 265 793 617 452 M 28 1417 40 30 14 64 265 N 5 6 4 19 4 88 21 0 14 69 9 14 14 26 24 P 9 9 12 10 5 3 8 Q 10 27 8 20 10 130 34 R 11 9 7 19 13 111 28 S 845 5 8 10 16 11 149 T 15 25 16 2 3 6 11 U 22 2 9 10 2 3 8 V 2 1277 7 14 31 3 222 23 m
tion areas were resulting from local building wake effects. For the first experiment on June 22, samples were taken at locations A, E, F, G, H, M, N, 0, P, Q, R, S, T, U, and V. The first air samples were taken at 1:30 pm. Samples were taken every 30 min-utes until the conclusion of the experiment at 4:00 pm. SF e ncentra-6 tions far the sampled locations are shown in Table 1. As indicated by the values for location A, the SF6 concentration in the exhaust stack was slightly above calculated values. Measurements yielded a
-7 while the predicted concentration was SF6 concentration of 5.85x10 ,
-7 is well within the target "ange of below 3.7x10" . However 5.85x10
-6 1.0x10 . Measurements also showed that the SF 6 fl w int the exhaust stack remained quite constant, keeping the stack concentration to within 1.57. of its mean value throughout the gas release.
Of particular interest waa the variation with time of the SF 6 concentrations at various sample locations. Location B had an average SF e ncentration of 4082 ppt, but samples at this location 6 ranged from 1,200-10,000 ppt. This large variation was due to the short sampling time needed to fill the 35 ce syringes. This effect was predicted by Slade [5). Local wind variations caused significant wavering of the exhaust plume. Since it only took seconds to fill the syringes, they could have been filled when the plume had shifted. Alternately if the sampling location was not generally in the plume centerline, a gust could have cisplaced the plu=e towards it. This problem is sometimes surmounted in atmospheric dispersion studies by taking what is known as hour-averaged samples. For these, a =echanical device slowly pulls back the plunger of a syringe, filling the syringe 24
in one hour. Samplers such as these were not available for this study. However, whil_e data in this form is useful in the developnent of mathematical dispersion models, it was not necessary for this work. The techniques used here can identify high concentration areas. Then, by increased sampling in these areas during later experiments, average values can be obtained for the concentrations. Data from the automatic weather station showed that meteorologi-cal conditions were close to ideal for the duration of the experiment. Average wind direction was from 231*, which is just 6* of f from due southwest; vind speed was 7.7 knots. Figure 7 shows the direction of the plume across the roof. I Figure 8 indicates the path of the plume across the campus. l The actual wind direction is shown. This is the plume centerline direction of travel. Also shown is a 30' subtended angle around the baseline wind direction. Slade [5] indicates that within approxi-mately 1000 meters of the source, this is the angle through which detection of the plume would be expected. The sampling locations are well within 1000 meters, so this estimate is useful. From Figures 7 and 8 it can be seen that the plume was well-bracketed by sampling locations, with many sites well within the plume itself. To calculate the actual Argon concentrations at the locations where air samples were taken, it was necessary to determine the I dilution that took place in the exhaust gases from where they were released (stack concentration), to cne locations where they were i sampled. Thi.= was obtained by dividing :he exhaust stack concentra- ! i tion of SF 6by the concentration measured at the sample locations. 25 1
/
plure centerline
\
is' ort centerline il *
.p - . i i Loc H l -l l ___m - a is arv tentertiae .[{
J, x, s y niilotion intak s C'
; 'r '
l Loc ild l yy
. . _ _ - __= ! / -- i
}ii 9th Floor Roof
' Astronomy lt--M l
b l Elev 489' MSL
~
- Observatorses h_ __
^ V
~
. l. II Lo,c 8
'l
, L o,c D , L o,c F j- ,
--A , _ ___ .
______f.n_l[ Loc C J
, j t.oc l 8 floor hoof I
Elev 4 75' MSL Mothematical Sciences i l t - ~ ~ ~ -
; urn ss - -_an n :a g ;
==_;
10 0 10 20 rEET j -) I Reac tor Rene f or < g_L n 1_ Stock l i E le v 4"'
, I_.
O Mst D .
+
- n[ Boelt r Hall
,) e = Somphng Locotions t
2:_ - a-=
-:~:' w PLlet PATH F04 JUNE ?? (RPERIM hi
- .?--:. = . - =
s n -1: i!GUkl ?
-T'__
A _ _um -
. , , , _ w u_. - - - , a . . , _ a , , ,
1 e 4
. fgn ,4 -
7
,, .,,..... 11 .
, 2
,, o
,, j- l t,-
s
- E ,
- m
~\
1
- V' s s
i i s c m$wnimouunv I j isn SCULPTumE GARCE4
'[I t ,
i i
,, ) ,
l
,y 3 f :~.f~ amer% g o, %g p ;.3 %- ..
2ii-glci,i.1
.. , .. ~Jd a i n
, y n -
covar f m,.c , N3 I l ' l , / . S r ' V l,o o,c gso,, ;e n k. - h- ;7
'f
,, oicmsome e zA $ o ,.o .
/ c,.ct, i Loc !
L. - , m
. .. =
L / fgefg, Loc
.z .
f
" =
. ME Y E RHOF F / ?
PARx .Cc 0 . LOC R . *. u,-) , k
. r. , ,,., c'
,. 3 :- -
./ a, . ! ,-
no a -
,1 6
couY
~
^
"'" [
n
,. f... *,,v ,
} ,
g et!'a06 stt e- ,
.~
COUEt P g . 3 Im
,/ ADeltoM o,
2 1 a ,,
-r
/
i
$.......inese.t..
....u......, .n1. i x.e t on i
;::.u. s y s, m,<,,,,c..m u F,<s1-=ciacLE opivt Pannia G v c
k ' ', ; N f @ NN-T nr:rpt -
%-.- , ,j (. scit =ct s .,
) i
_ s _ .. l
.r,,r - ,-----,,_.--,-..__,-_.-...-,,,.-,,,._-._.,m_,., __.._._.__.,_,.mm.,-_..,,-. . - - _
From Table 1 the average SF6 concentration at location E was 4082 ppt. 5 Since the stack concentration of SF was 5.85x10 ppt this gives a 6 dilution factor of
-3 4082/5.85x105 = 6.98x10 .
To obtain the Argon concentration from this, it is only necessary to know the Argon concentration of the stack exhaust gases when the reactor is operating. A review of this past year's reactor operating log showed that a steady-stage Argon concentration was reached in the reactor exhaust stack af ter a period of approx 1=ately three hours
-5 at full power. This concentration was 1.1x10 pc/ml. The Argon concentration for sampling location E is then 1.1x10 -5 pc/=1 x 6.98x10 = 7.68x10 -8 ucjmy ,
-3 This is the actual average Argon concentration at location E when the reactor is operating. However since the NRC has limited the UCLA reactor to operating 18.87. of the time, and has determined the occupancy factor for the roof of the Math Science Complex to be 10*.' [1], the actual exposure to the public at location E is
-8
.1 x .188 x 7.68x10 = 1.44x10 -9 pc/=1 .
-0 This is well within the federal limit of 4x10 pc/ml. Table 2 gives the dilution factors, Argon concentrations during a reactor run, and.
Argon concentrations corrected to reflect reactor utilization and occupancy factors. It is this last column of the table which shows the true radiological exposures to the pub *ite, 28
Table 2 Argon Concentrations (pc/ml) for June 22 Experiment Location Dilution Factor Argon Concentration Argon Concentration During Reactor Run Corrected to Show Public Exposures
-3 -8 1.44x10
-9 E 6.98x10 7.68x10 F 3.42x10
-3 3.77x10
-8 7.08x10
-10
-8 -10 G 3.15x10 3.46x10 6.51x10
-0 8.50x10
-9 1.60x10
-10 H 7.73x10 M
4.48x10 ' 4.93x10
-9 9.26x10
-11
-5 3.95x10
-10 7.42x10-12 N 3.59x10 0 4.10x10
-3 4.51x10
-10 8.48x10 -12 P 1.37x10
-5 1.50x10
-10 2.83x10
-1 Q 5.81x10 -5 6.39x10
-10 1.20x10
-11 R 4.74x10 -5 5.26x10
-10 9.90x10 -12 2.55x10
-0 2.80x10
-9 -11 5 5.27x10
-5 -10 -12 T 1.88x10 2.07x10 3.89x10
-5 -10 -*
U 1.37x10 1.50x10 2.83x10 2
-9 -11 V 3.79x10 ' 4.17x10 7.85x10 29
The results from the first experiment substantiated the belief that the highest instantaneous Argon concentrations would be four.d on the Math Science Complex roof. There was a slight building downwash ef fect as indicated by the concentration levels at semple location M. This was the furthest location from the exhaust stack where any significant levels of Argon were found. The high Argon concentration at location G, the air inlet to the building, also indicated that measurements of Argon concentrations within the building would be necessary. The second tracer experiment was conducted on July 8,1976. Sampling locations were concentrated on the roof-top, with the first series of internal building samples also being taken at location J. The wind direction was from 226* for this experiment, l' off due southwest. The average wind speed was 5.8 knots. Figures 9 and 10 show the plume path across the roof-top ar.d ca= pus respectively. The results of air sample analysis are given in Table 3, while dilution factois and Argon concentrations are given in Table 4. For this experiment samples were taken at locations A, B,C,D,E,F,G, H, J, M, N, 0, Q, and U. The SF rele se was begun at 1:30 pm, and air 6 samples were taken from 2:00 until 4:30 pm. It should be noted that for location J, the corrected Argon concentration cannot be averaged over both operating time and occupancy factors to give true exposure esti=ates. Since concentrations at location J were within the Math building which was always occupied, occupancy factors were assumed to be 100% and the corrected Argon concentrations could only be averaged 30
plume centerll,me x _ _. ___ _ 15 off center 11:4 - It* of f centeritee
\ Loc H
=
' Ventitation niche '( ~ N @
~
l [h__: l t cG \
/ .x g
~ Astronomy c
-- -, Observatories l .
x s v 9'h F loor Roof .
~ ~
E lev 4 89' MSL L o,c ._ __ Loc ~ -._ _ _ . . _ _
/%
7 __ __ -.l~~'~ -- __ _ . _ _..L. o c B. . , , Loc C t E 8 oor Roof , Mat hematical Sciences t Elev 47$'MSI u
- sn u s s a e , = k
" 10 O 10 20 FEET - , l
._ Reactor -
. . _ .[ Rec;; . gl,
}-
Stock f t ie v 48 f 3- - *h Boeller 11011 i J M s t. l
,j e Somphng Locations PLUML PATH IL3 stulY 8 IIP (lllMthi
. = ,: == z : = = -
IIGURL 3 _tttlr 12 m x wwers-~r=.3 _=:-: d CC.
15 off centerline 3- e [I[ e o f 1 9 2[ydM(
/.
0
- c. , m .. g .-y .
I , 7 ' (RA IN D MUR L H f O SCULPfumE GA
'u ? ' % . y N
- k, J as N E *
-7r* , - n, plume
< cerrterline c, r .,
~)
8 7 r
']e 9, . ,, foe f(7,f c - -
, j '
E , , , Muwape TIES i j,ygg k '5 b ," 150 ft l : ,'- - - 't-- Q ,Q ' s DECK $0N * :p a,tesos %v c actt / mterline L s -, ., r
, , 1
- (T .
! ,e ai T- -
W if -- 2_ a i
,. Loc Q METERHCFF { . n "
~~f PARK i
, L '
g, -. [ - Q-' Q $ Q u . s..
--c , ',
,[ '
f~ \
I
- g .,U'="s b .m ;
qW. . I *en :: i
/ /
. n c ie
_ _- u mt.e y - , u. ( o gsf aci ut
~
P n COUR5 ~ y l o I
$ indicates earvle locatim ADDITION o, L w z a . . . n.. .i, u n . en,
- - ~ ~ ~
h v
, 7 %0 wESTWOOD .Nc',j,jj
~
g h PLtNE PATH FOR J11Y I 5 CIACLE ORivE p a e Ki m.G .f M TICLE 10 gr p q f, ,, SCIE NC E S -'^"rgr") g'g! l@hI I' 32 2O 3A2 a a n $s
Table 3 SF Concentrations (ppt) for July 8 Experiment) 6 Time Location 2:00 2:30 3:00 3:30 4:00 4: 30 Average 5 5 5 5 A 3.9x10 5 3.9x10 3.9x10 5 3.8x10 3.9x10 3.9x10 3.88x10 5 B 594 155 1040 135 616 1248 631 C 486 395 205 221 4625 1928 1310 D 880 60 1141 256 5127 2957 903 E 724 126 4574 3888 96 247 1609 F 574 2114 1816 4070 1649 1255 1913 G 2158 2658 641 731 4653 562 1900 H 61 172 40 129 32 56 82 J 727 767 762 1061 1067 944 883 M 39 26 91 15 185 38 65 N 21 121 24 24 62 19 45 0 20 28 52 46 106 24 46 R 13 49 23 56 54 19 36 U 10 20 29 39 6 24 21 33
Table 4 Argon Concentrations (uc/ml) for July 8 Experiment Location Dilution Factor Argon Concentration Argon Concentration During Reactor Run Corrected to Show Public Exposures
~ -8 -10 B 1.63x10 1.79x10 3.36x10
-8 -10 C 3.38x10~ 3.71x10 6.98x10 3 -8 -0 D 2.33x10 2.56x10 4.81x10
-3 -10 E 4.15x10 4.56x10~ 8.58x10
-8 -9 F 4.93x10 -3 5.42x10 1.02x10 G 4.90x10
-3 5.39x10
-8 1.0lx10 -9
~
~9 -11 H 2.11x10 ' 2.32x10 4.37x10
-3 -0 4.73x10
~9 J 2.29x10 2.52x10
-11 1.84x10 ~9
~
M 1.68x10 ' 3.46x10
-9 2.40x10
-11 N 1.16x10~ 1.28x10
-0 -9 2.45x10
-11 0 1.19x10 1.30x10
-5 -9 -11 R 9.28x10 1.02x10 1.92x10
-5 -10 1.12x10
~U U 5.41x10 5.95x10 34
over the time throughout the year when the rew tor was not running.
-8 Therefore the actual Argon concentration in the building of 2.52x10 pc/ml could ,only be reduced by the utili::ation time of the reactor
(.188)togiveavalueof1.67x10]8 c/ml. This was the highest '
.. ,e corrected Argon exposure that had been detected and indicated that [
further sampling within the building was necessary. #' The third SF release was conducted on July 27, 1976. The pluce 6 direction and sampling locations are as shown in Figures 11 and 12. Samples were taken at locations A, E, F, G, H, I, K, L, P, W, X, Y, and Z. The wind direction was from 235*,10' of f due southwest. The average wind speed was 6.1 knots. Table 5 gives the results of the air sample analysis, and Table 6 giver dilution factors and Argon concentrations for this experiment. Locations K, L, Q, W, Y, and Z showed no SF 6 above background and appeared to be completely out of the plume. Table 7 is a composite of all the sampling locations. When a particular location was sampled during more than one experiment, its Argon concentration was averaged. Locations E and F were sampled during each of the three experiments, while locations such as Z were sampled only during a single experi=ent. Figures 13 and 14 show the behavior of smoke released directly into the exhaust stack air stream. In Figure 13 the exhaust stack top is level with the top edge of the surrounding enclosure. Both figures show considerable mixing and dispersion of the exhaust plume shortly after leaving the exhaust stack. The velocity of the air leaving the stack was 41 ft/sec. The velocity of the southwesterly 35
plure centerline
-- ,-- . . . _ _ .___._N _ _ _ ,
r- , _ . 1---- o it off centerline i,
, N __
Loc H r----- ._ ~ Vent ~ ation intoke
' y' N
~~~
[,> [ f
! Loc G 15 oH (enter use jg I
x 1 l VN - g Ng
- - - - .-- m. _l ,
nomy i
,, , ^,
Obser votaries j 9'h Floor Roof )
~
;l Elev 489'MSL ll t uc r _ jf fr- - m n r _- - --
~ ' ~~ ~ j Lu.c D
Q D .
, ~ ~ ~~ ~
Loc C oc E 8'" Floor Roof Mo t hema t ical Sciences Elev 4 7 5' M SL t uuu-- 10 O
.m.
10 20 FEET [
,/
l n _~- J
' Reactor ....~ ,
~~
Re OC1or # - l Stock Elev ~ + 4 ti / 5' q ust Boelter Hall k l i
,- * = Sompting Locations
. - . - - - . . . ,,__,,,,g_,
ocoun na y =1., _ _ . x _ . + ., _ J- :_
. an fhlh
~ $
- I g ' f
. (2 l
, , . . 3 R) .- tc M..
g 8 ; , p 15' of f centerline
- n N + g'
,s 7
. . . ,, I - #
$ \ j
_ l ,
,P S
*j61RAkF11NoWUMPHY t - scut,TURE GARCEN I
* > , 1 l@
N i 't - -- Jin l 1, as - - , ,. f P - / 1
~
..o.
cL
#'[q y (.. . -couar
{u . ir,Es.I~ ~ ,~~
~ * '
.,.T.
~ @f ' [.: -^p Ci a
- m. 4 - - # s
- f. q',) m@IQ' \
't 2, e, ~~ #
.4
/j ..
2 r e L i , r ,,_ pq f ./ 1r - E i - toe ; . . ___/- - I
-i .);.\'jy' T .Cxso~ o
.t Y.* " (./ 4.u. :e .[ f c , .c a l __
- / i. ,
i :- ([V . = -{__ [, -. nts.
' II '
h'
= -*
METERHOFF '
. - M .' 15 off ,. F.AM centtrline j -
j
., , w c. .c -
- m. ~
< /, ,..
,a e<
f ,,
~
l' m ., i G
~ 38 Cou i - - s
~-
2
- @ . .$ , /#
- 1r.41c.t e. .. ,le 1 x. t oe womes o, - -1 '
...........u.......,
N - Oq - [ hh : ^'a
@ as, -o e ^
;;:n, g,
- ~ gg, par'"
CiRCtE opivt ,A9meNG w g= .% { ~ j !' SCIE NC E S
- -- 3 ] [
Table 5 SF ncentrati ns (ppt) for July 27 Experiment 6 Location Time 2:00 2:30 3:00 3:30 4:00 4:30 Average 5 5 5 5 5 5 5 A 3.8x10 3.9x10 3.9x10 3.9x10 3.8x10 3.9x10 3.S7x10 ~ E 2094 2597 773 272 1629 1932 1550 F 778 394 3104 770 2342 583 1328 G 1350 1590 1336 2118 136 1129 1276 H 51 31 22 1578 11 206 316 I 576 718 629 573 549 513 593 K 473 684 832 1414 1038 876 886 L 2 2 0 2 4 2 2 P 6 0 0 0 5 2 2 W 0 0 3 2 2 3 2 X 2 3 0 0 0 1 1 Y 2 0 2 2 3 2 2 Z 0 4 2 4 5 0 3 l l l 1 38 i
Table 6 Argon Concentrations (pc/ml) for July 27 Experiment Location Dilution Factor Argon Concentration Argon Concentration During Reactor Run Corrected to Show , Public Exposure
-3 ~8 8.28x10
-10 E 4.01x10 4.41x10
-3 -8 7.10x10 -10 F 3.43x10 3.77x10
~3 ~8 G 3.30x10 3.63x10 6. 82x10-10
-9 -10 4
H 8.17x10~ 8.98x10 1.69x10
~9 1.53x10 ~3
-8 1 1.69x10 3.17x10
~8 -9 K 2.29x10~ 2.52x10 4.73x10 L 0 0 0 P 0 0 0 W 0 0 0 X 0 0 0 Y 0 0 0 Z 0 0 0 39
Table 7 Composite of Argon Concentrations (pc/ml) Location Argon Concentration Corrected to Show Public Exposure
-10 B 3.36x10
-10 C 6.98x10
-1 D 4.81x10 E 1.04x10-
-10 F 8.13x10 G 7.81x10
-10
-10 H 1.24x10
-9 I 3.17x10 J 4.73x10
-9
-9 K 4.73x10 L 0
-11 M 6.36x10 N 1.57x10
-11
-11 0 1.65x10
-12 P 1.42x10 Q 1.20x10 -11 R 1.46x10 -11
-11 S 5.27x10
-11 T 2.83x10 U 7.02x10
-12
-11 V 7.85x10
'J 0 X 0 Y 0 2 0 40
0
=
pu s.sr;,Q,r2+ Pv~=: <waqp w g ;m.w. r . - 3p, a '. '_ . - > ^ >y. ; K' m
.K.* *fia.h.,-.4%@'!fw-sh.
w: -!?Y - t
. a. v.A #e**. t*= h7.e,n,-u-> ;.,. e w we,, . s.w.
. h'bl
- v..m.% s.50;1
..e f- . ' a r* . -
'->-p& Wo~s ti w
%f x
qt! d ts-~ rpMN,,;;s.e.yl+:ew ry%.'i,~
-*yy"9y A f ,%%.~c.0 G. %t
;% ss S er,* -
ia.y- \gn4%,4 '- e * ., ys-.-
. 19A.,.yf4Q,v.g*..T ,; '\ Lwk' ~. -*~ . ~
G.%' <.1;h .himigy *c- 4.,%,7
$f.tD ! g [~ } & ;iM:* ' '.'
% . 4,.Tr,
.,- ;y c- x.
#$.. }. . ,f' tr- .- Ms.m. -
' ap 25 l.
.'r4'e < c=h
~: "'" ,,, .,
S"
- w(> .
s y 4 . o
% C
- 4.r
. .'- c
..E.
q;.
..p-V i
i~q * ~ . , ,, .Jf t; Ji 4 o 5-L , , .. v . o . > r- c g
~
* ^ '/ %pj::p%;',
~ ,
ra = 5,-
,..- ' i.e > "W5: htimeh m
.. e.gs.w, n Ml i@in C M 4.,: ye
,' , u
,n di: ' .1 -
E
- s
) ., . l. O
- z. :,;r;.".
~.a.. 7 M :g.:.M~% i.ng. - - g h.- L o
Jd -4 4 ; . .a.
~-w.,:
Q,.,ap'I.n ,4 (q%. 3:. ?,il[".. '.k' be u a
!y d;,,ELGgM*g' C N-Ty g3 O
; . ;.y =~ t ., >
- 5
~ E!'fd)+NM....p .MM* g g$ . f.;p.. %. . , . $.'g .[.
-- ~
g+
. v.
g4 (.4 m% A' S 2 @T t .gS.%-'C 3W7 M.p?..Ary . P'
- gn.+4 M-f 1./ %,,
l TE'.h TY: fo,iN Y . .
"J
,pr .,4 .1'4.h* ip. jp rym%'Y :..:
. ~
~
yyg.pq?!.1;,~*9:-!, .. 11 y fje
-th %q)e. al m Q N /,/,.--, s- c ; -:,w we t - '
al. .
, J. S.w.N,T8.N.. .*..N
.: .. ~e I'",N.
y-
' .l*!Pm' . . N yd m .%f'*:.'_*~
.v . Q..'5 1* '!.- ..l[,G '
((. .q L .' ig
. , -c.o O
..% ~-y e _ ty b-y.'.'ii
~g w.
s 'M.a . ^:.z
.- M &-J*XZ d;e_1 %.w ,
~. .
jll[,,iha S O AO - Q h ["k. - c
ww ,- . . . . . ,. P. . . ,q - C' f :i .w%. . . FJ. 1:' $ ?T l* f ..M? . I 4% - (t .e C- ' e.N% i 5 p 1%,., ss - [.- . l w s s y. : . _l . 8 e r .r . . , .., E. '
.. .- I r s . u
)
, ; e v.-~.-
, j u s, .
u .w. n, .- . 7 j z$fdE tY. a 55! $ = [ S4t
- 3. p
!Di'
;a
- w a e
l'j', It' i6 s 8
.i' t . . .\
i
} . :
L
- a. :
. =
r . o 4 e
- $ :: ; m 8
i
.. e
[ lt.
. In
) ;- ,
t- g?L - i i c. t g
)y:r.' v!! '
pb. t..y) i
- r. 3.iLf %c. inr :.'
mm o ,. ,, 3. D D D 'j 42 _c o f o . )) J o
wind blowing across the roof was 13 f t/sec for Figure 13 and 10 f t/see for Figure 14. Results of Halitsky [5] for building wake and wind tunnel studies indicate that a smoother rising plume would be expected for the plume exit velocity to wind-speed ratio of our experiment. This suggested that the effect of local turbulence caused by building structures could significantly increase plume dispersion. This phenomenon was further investigated by the comparison of the mathe-matical dispersion model to the experimental data. 43 c __
CHAPTER 4 COMPARISON OF EXPERIMENTAL RESULTS TO MODEL PREDICTIONS The atmospheric dispersion model which the Nuclear Regulatory Cocaission deems appropriate to the UCLA situation is the wake cavity mixing model. It is applied where the height of the exhaust stack is not twice the height of the surrounding structures. The model assumes a cavity area near the point of release where uniform mixing takes place. It was derived from the earlier Pasquill/ Gaussian dis-persion model which assumes Gaussian distribution of the plume in both vertical and horizontal directions. The wake cavity model states that concentration of radioactive gases in the area near a short exhaust stack can be - rressed by [6] 6 Q 10 X(x,u) = 2: 0ou yz where X is the downwind concentration in pc/ml, x is the downwind distance at which the concentration is being calculated in =eters, Q is the discharge rate from stack in uc/sec, cy.c are the standard deviations of the concentration distribu-tions'in the y (horizontal) and : (vertical) directions. These parameters are functions of atmospheric stability and downwind distance, , i u is wind speed in m/sec. ] l l l
The atmospheric stability is described by classes A-E, and was origin-ally defined as indicated in Table 8. More recently, atmospheric stability classes have also been determined by the standard deviation of the wind direction over a specific ti=e period. These definitions are shown in Table 9. The product of a c , as a function of stability yz class and downwind distance, is given in Figure 15. The wake cavity mixing model was used to predict what the Argon concentrations would be at the roof-top sampling locations of the July 8 experiment. The wind speed was 3.12 m/sec and the daytime insolation was strong. From Table 8 this would suggest a stability class of B, moderately unstable. However 0 f r the period of the experiment 6 was determined from the automatic weather station chart to be 29.5*, which fits stability class A, extremely unstable. The parameters needed in the wake cavity mixing model are given in Table 10. The predictions of the model versus the actual measure =ents from the July 3 experi=ent are given in Table 11. Both stability classes A and B are used to ascertain which fits experimental results better. The data in Table 11 shows that the model predictions, based on stability class A, do fit the experimental results better, but are still conservative. The best model prediction occurred at location F, but even there the prediction was 15 times the measured value. At other locations, model predictions were as much as 105 times the actual measured concentration. 45 i l
r-g ,( f f pp q p p.-, 4 ~g.g M. L 6 . o{r .{ ~ p .7' g a n L7.cy @i y.m g ik.g;. ;. p. r y~pr ;~:~-4.t .m . : , - _ _ . - 4_iw , b.ym.- .g-#p.i e,. . i ," t .R P :;U FM
! r a-
'_ F G ,I b, - M.:
a=r
- + - + - . - 9r -
;. 3 -- - r q . ; j. . 1 r 4.
, y --- !t ;7: n .r p: - ; .: .yq -- .
7M,$I% .444 y.i: .-F- .M E jb'TU , t@d }j}I@ M f [.-] .f v.
-j i. p i.j.g.f I d 4 w e
>!W- %p Np Nh = Yi e .
itthj N-O . ,$i SM.N O -4
-M'd h iii tE $ "g.M '7
].[ }j. ~* i - ..J l E3
. E.t .dYY ! _u. .. .Ub
.. ....t
- d. J.~ ; .-5:-M-h
..... ' a.. .n F.
5.Y Mb.'
- a. :.,3 ... . ::.. r.:._; .....o.:.- c_. =
~
=_ '.'r' n
_:a_:.=_2 _r_d
- W. ;Ny
- . i:a..2.=...-3 5 :;
h_m..__:_._. [ ,I_ . .. ... : - f3 %i-f Q - f -- ,f~'; h m...
/. ...
. + . - ....a..
. , . .. . . ....v .
. . . s. . ... . --
U im. "
; Lip;44. . 77 [N M T @liII' F_ .U.S
@111[D
'TFfi %.*fr .,y.g E q z.Ii,, ' l *+
' ' -j '-N-E W ; . @h[ jM
-!- J i M 4:
. _f :/ /:- 4 _i._a. 2M- . HFfi 4 9-d F'
~~
-F 'A a/- :-fi: - 1; . . o.. Jp =M - i-R F Hlf-i i-+C -H 4 , I le- ! iF -W 7 JT'r F F B F M G
+4 ..
*; i 1
T, F + y*,-w,p 3 g M
._+; g .--. n-4 --. . .-- ,!--".m T-P-" 7_. A
^f. - f tt*j + em.3 g
-.d., :. .-_. . { _.=---+-iw mu2g _q
- 4: .
-ep.n.. .
t 4 1 :_,- 43
.,in.
1 1 -J . i . . . :Li s..!r i- F 3
;I7.m .' i-a El-E:-'
j- dy - i -+ : --+9~:.--- =- M .- "..d_: =EE 'EE' Z : . ".-~
,.:: .:: a :! : .: :r- :r. - - - ~ _ _ . _ - _
..__.._=..._. , f. _ =
_. f _= ...f. ._= .-:_:=.f.._.._ . - - --
. . . ': _.1--- 2:: .; : : f.:;;; :x.r-fz :: /---~ :/. :-
..........- .; _q...._.,
. . . . . . . ..,..._._z z..._ . , . .
M,
- f. . f. . . ..
r i -
- i. - 7 1 w , , j.2..p g ~Iup.f_.t yg g y pcp.gp.,y gg i
m . f i f N: [ ! [J ' 2-i f Q j't - n m.
*a x n u.
.: 77, ~-M J n
if 1
- q_o, ,! '.. -.!-SII,'I Nj '.b@)
*M ...
. . 4 ., p ,.- _p.
4 . . . . . . . . . . ' [ NI[ .
.- .E,rs,.,7 r.a N, $.,~9,
.p= 4 r y yg. r qp- .
- I !i-
*g,7 = M -E y t ..
.+4-,, -f- . _-
-_ ff & '--f.r-l: [~. l ..a.k :$
- f - -- . - i ,
II l
..._.2._.__d_.",,,i.......i.. : :E ; E-at
.-.2 . . - . .-
; . ,' , ,-h
. . , _ _ ,+ , , , ._ ,-,. _55.___ . - iii
, , , , _~i '.=Js-5
.t. - c e: : :c: : _.
I :----- : .=-?.._---=- ~:: _ ;.. _ - :2 . _ .__ e 2 : --/ - .n " - - - - - - - - - - : : :.: : :.: . :.-
"' ;. f - ~.I. - *: u ._ ? * * * ~ '
.- j
..-n .
. .+. . . - . . . - .. - ....
C ._, ! 4 _ gig 1.2 '
"% c71r 1[lE i;r g j ca ._ _ _
. dy _ .:h I; M' [ R $' .u
~
-t---
". t-.- _ . - -g-
; m. - . r:-
?:
_,:--. -: i. .. ..;. . ; : ;-. g . .s :' ;
- _: 7
; ' i * . e ; - .
?^Q ...:. @ 5_ ;,, .Gi.d.11=.5_= .d E f-h - E
.g .L7 7 17. +;4 ...;.4 e." 4...J y g_T. g...... g j .1) ]p . ..'
y 4 {
. .. _ _ . . . t c M
t y.- ,:u%;.4.P_4 .- .,n .g _M.:.
.n . ' _ a 1 1.1 ...
y - ; ..L. . . . ,.4. nw :.7: _ - - - _ = - -=;
.r .. = ; . . g -q . s - w .
, i
; . . r. .4 , .
gg;, 3- =.; .-
..q_y \
.i~ ; -
- s. 3 - Ti.Pi : : . - . :i= =E = .?
1.+ f
. a.4 a. -_ -
. . - .- k _ -- 3.; ,___,3_ ;wi.1 - - - - _= - _-
- - , - - /---t . .: ii- u Eicia::_ ~~~;":::: :;;i.*E l 2-- iE 3;_" Z
. .. .=/._.::-
- - - - -~~
L.
. .__..~.. . . - ......- .
(. . . . . . n. _ . , . . -- ..p.. pzypm;
,_ . q~.m '
x' q4ppy m.m '
.. t . >;--
p
+-.---Ni-L _,, ,. F #f . 4 -- t T- "~. T4 .r: r -- 2 7 69 E i ,.
- : np.-
. : m'.--- is 2 i di.c--- wr- r.tw ---+-*r
-q
, . M'Mi$ , ": N .*4 _
I* m - N b .1---3 '.--U',"Q 6--*+- 4 ,- -- -,-c+ h. n W . ..tg.ep p; . , c , , ... p.-. y. n h,.; -.y ..y; N;,ky , n].e ,,. + @ y 2-w:.-.t. y, .
,, p.
F- ,- , r .,r.- 66 .r -.t-- h ,,,1,*. .$..;- M k Q . J.- d 4_Q I-.N.ip344 '.W.I e.g . J._ I y, l [ t $ .M *% @?
". " ^-^-2
- 2 ,' _ - HI '- - :
} y 5 F.
- 4. : : i lh:. ;..$T&{ _ ' { - . ~ ~ I 7j . ,
.- - . . ; ! i _" 1 ' . . IE
. _i _-F.J: 9. _-E_?
t ..
.ir -
*r. . .. = ~+++m m--- . . 7. '.i.~f3 ' M_ ~~~ t+ : r __:-- -
.....'37~
. . : : ' . _. ; .._ m.=. .:. =. . r. .r__ . . .~
-. . . . . 1. :. :. :.N. .._:. r: == =_. ;, ; _- _: .
. . . _ .: 1. : .::- . c _ _, - ". . . - --- t .
!! .1"-~"' - * - --- . . : n.:;
-- - :;:: .~ - ;- := . . . . . . -- - __._ .. - -
'****T:.-- .
l T.. ......
~...
l 1 0t i 10 100 o . . . .. . o . . . .a i . Figure is The product of <,<. as a function of downwind distance from the source. D N O b 46 a u)h_[_h m
s s 8 e
/ n n 3i o I
d F E P, D z s l E D D D o l e I u i eb h T nr o tl a t I i ol ib t D h C d as e N T nt v O osy o C c l b ye a R l l t E at a y l i rh r k T t ge s A uid E el o e W h t NSM h t O g - - - T i B C C D D f l DEF o 8 S S E n e P o l Y i b T t a n c T E o a C i r N t f E a L l e . t U ut t a B sa h h R nr B C D g t U I e - B - - C i T d A B C n s eo a F mM r O i o d t e N y s y n O a sn a i I D nos d f T oin e A i t o , d L t ii t E g id t s s . R d ni a i s n B nod c ! t o A - 8 C C ocn r s u r A c o e so t ec v el S el o nc l b e i b al y d y at b v ub t sa a o s nt e l d nus h c e _ u n r _ yu o f e yl t ov _ d c l ey o _ ne et l e ec i s mat l e _ W/ erh b r s _ m r t leg a gi _ e 2 2 4 6 6 t i c e _ c , x ol i dh ai t < > EMS l c f c p ei r e - - - p hh up
*A +T w SS A1 C5
~ .
zN -
Table 9 Pcsquill Stability Categories (o g) A, extremely unstable 25.0* B, moderately unstable 20.0* C, slightly unstable 15.0* D, neutral 10.0* E, slightly stable 5.0* F, moderately stable 2.5' i 48 ! 1
Table 10 Parameters for Model Calculations Locations C D E F G H x (meters) 5.4 8.23 10.29 10.82 20.88 55.02 u (m/sec) 3.12 3.12 3.12 3.12 3.12 3.12 00, for (M ) .95 2.6 3.8 4.5 15.5 110 Sla5111tyA o c, for (M ) .7 1.95 2.6 3.2 10.0 62 SlaBility B Q (uc/sec) 72.68 72.68 72.68 72.68 72.68 72.68 ' r .,
~h t
t 49
Table 11 DISPERSION MODEL PREDICTIONS (uc/ml) Location Experimental Results Model Prediction Model Prediction Stability Class A Stability Class B 3.71x10
-8 3.91x10
-6 -6 C 5.30x10 2.56x10
-8 -6 -0 D 1.43x10 1.90x10
~0 -6 E 4.56x10 9.28x10~ 1.43x10 F 5.42x10
-0 8.23x10
-7 1.16x10
-6
~0 G 5.39x10 2.39x10~ 3.71x10~
-9 ~7 ~0 H 2.32x10 3.37x10 6.00x10 50
CHAPTER 5 CONCLUSIONS The experimental results obtained in this research showed that radiclogical exposures to the public from Argon 41 released by the UCLA reactor are well within legal limits. Table 12 gives a summary of Argon concentrations corrected for reactor utilization and occu-pancy factors at the various locations where samples were taken. The table also includes what percentage of the legal limit (4x10' uc/ml) these concentrations are. The highest exposures to the public were discovered to be within the Math Science building. This occurs because a =ain ventilator intake for the building was found to be directly in the path of the reactor's exhaust plune. However, while this concentration was the highest found in the s:udy, it was still only 11.8% of the allcwable federal limit. These results seem to indicate that there is no danger to the public from current reactor Argon releases provided that the legal limits do not yield a danger to the public. It appears that no modifications to the reactor oper-ating procedures are necessary and that the recent operating license amendment should remain in force (providing the reactor use time is 0.188). The results also indicate that the current NRC model for atmos-pheric dispersion yield results that were conservative for the specific situation at UCLA. The model predicted levels of Argon which were 5com 15 to 105 times higher than the experimentally determined values. 51
Table 12 Percentage Location Corrected Argon of Federal Concentration Limits
-10 0.99 B 3.36x10
-10 1.75 C 6.98x10
-1 D 4.81x10 1.20
-9 2.60 E 1.04x10
-10 2.03 F 8.13x10
-10 1.95 G 7.81x10
-10 0.31 H 1.24x10
-9 7.93 I 3.17x10 4.73x10
-9 11.83 J
K 4.73x10- 11.83 L 0 0 M 6.36x10
-11 0.16
-11 0.04 N 1.57x10
-11 0.04 0 1.65x10
-12 0.004 P 1.42x10
-ll Q l.20x13 g,g,
-A R 1.46x10 0.04
-11 0.13 S 5.27x10
-11 T 2.83x10 0.07
-12 0.18 U 7.02x10
-11 0.2 V 7.85x10 W 0 0 X 0 0 Y 0 0 Z 0 0 .
1 1 52
This indicates that roof-top turbulcnce resulted in greater mixing of the exhaus*. gases than was predicted by the model. This is sup-porr:eil by snoke plume pictures which displayed an earlier plume dispersion than was forecast by wind tunnel tests. It appears that both the dispersion model and the wind tunnel tests are based on a clean cubical roof-top with no structures to perturb the air stream as it flows across the roof-top. The complicated roof geometry of the Math Science buildir;, coupled with the many structures on it, has resulted in a more turbulent wind condition on the roof than allowed for in the model and wind tunnel tests. Subsequently the dispersion of the exhaust plume from the UCLA reactor is considerably greater than predicted in these studies. The use of SF6 as a tracer to determine atmospheric dispersion showed itself to be a valuable tool for situations such as exist at UCLA. Wherever it is necessary to determine low le el environmental concentrations of gases such as Argon, released from a reactor, use of SF6 circumvents the requirement of elaborate and expensive radio-active detection equipment. An added advantage of this technique is that dispersion estimates can be obtained without the necessity of releasing radioactive species to the environment during the testing procedure. The normal radioactive gas release rate must be known, but once determined, no further releases are required. The use of this easily handled tracer makes it possible to determine actual local concentrations resulting from environmental releases of radio-active gases. Use of this data, to supplement use of conservative 53
model predictions, should aid in documenting environmental release claims to the Nuclear Regulatory Coc: mission. S t.
rf REFERENCES P
- 1. U.S. Nuclear. Regulatory Commission. " Amendment to Facility Operat-ing License," Docket 50-142, Washington, D.C., February 1976.
- 2. U.S. Nuclear Regulatory Commission. " Rules and Regulations,"
Title 10, Code of Federal Regulations, pt. 20, Washington, D.C., April 1975.
- 3. Applied Nucleonics Company. " Atmospheric Dispersien Analysis'of Argon-41 Discharges from the UCLA-NEL Nuclear Reactor," Santa Monica, California, February 1975.
, _ 7 9 $ e- .
f .
, G-, r*
- 4. Drivas, Peter J. and Fredrick H. Shair. "The Chemistry, Dispersion, and Transport of Air Pollutants 6 d from Fossil Fuel Power Plants in California," Division of Chemistry and Chemical Engineer-ing, California Institute of Technology, Pasadena, California, June 1975.
- 5. Slade, David H. Meteorology and Atomic Energy 1968, U.S. Ato=ic Energy Cocnission, Washington, D.C. , April 1968.
- 6. U.S. Atomic Energy Commission. " Draft Regulatory Guides for Imple-mentation," Docket RM-50-2, Washington, D.C., February 1974.
l l l l I 9 f f 55
- - . - _ _ _ _ . _ . - . - . . - . . _ _ . , - , - , _ - . -}}