ML19324B755

Radioiodine & Particle Transmission Through Sampling Lines for Standby Gas Treatment Sys Effluents at Fermi 2, Final Rept

ML19324B755
Person / Time
Site:	Fermi
Issue date:	09/30/1989
From:	SCIENCE APPLICATIONS INTERNATIONAL CORP. (FORMERLY
To:
Shared Package
ML19324B751	List: ML19324B755 NRC-89-0221, Provides Results of Sample Line Loss Test & Evaluations Conducted to Fulfill License Condition 15 & Forwards Rept Entitled, Radioiodine & Particle Transmission Through Sample Lines for Standby Gas Treatment Sys Effluents....
References
NUDOCS 8911080146
Download: ML19324B755 (35)

Text

,;

'p4

(.

a f

, , _ . . . _ , .,r._,,

,. s.,,

j -

p, y, ,

'l U ,

{

t . ..e l g b,,, g i-t

.- ,e a

i 4

i

. .)

_ s.

_., . +, , >

':J! ,

s

i N' ' '

a a k:__ m .. . _ _ . . . . . .; .._. ;. . ,

. ..._.;_.,2.. .

(t An Employee-owned Company O

(

't I

\

i L

'Il

8911080i ADO g h10 3 4g ,

PDR PDC P

l i

RADI0 IODINE AND PARTICLE TRANSNISSION THROU(JI SAMPLING LINES FOR SGTS EFFLUENTS AT FERNI 2 4

Prepared by Utility Services Department Science Applications International Corporation Rockville, Maryland 20850 Prepared for

. Detroit Edison Company 6400 North Dixie Highway Newport, Michigan 48166 Final Report September 1989

~

s An Emyoyev0wned Company

r o .

) (

SUMMARY

r

) This report contains the results of measurements and calculations performed by Science Applications International Corporation (SAIC) to  !

' estimate radioiodine and particle transmission through selected sampling [

lines at Fermi 2. The sampling lines evaluated were those used to monitor l

) the Standby Gas Treatment System (SGTS) discharge. Two lines carry samples l to the SPING monitors and two lines carry samples to AXM Monitor grab (

sampling locations. The four lines are designated SPING-1, SPING II, AXM 1, I and AXM-II. The AXM lines are relied upon for the collection of post-

) accident grab samples of radiciodines and radioactive particles in gaseous l ef fluents. i

) RA01010 DINE TRANSMISSION The Radiciodine Line loss Test Facility in SAIC's Rockville Laboratory j has been used to measure the transmission of elemental iodine (12) through j

) replicas of air sampling lines. Elemental iodine is the most reactive of l the gaseous radiciodinn species expected in reactor effluents under normal l or post-accident conditions. It deposits on surfaces and may later be -

resuspended or tightly bound to the surface. Transmission of 131 12 through  !

) the lines was measured and the data were analyzed to determine average deposition velocities for 12 in the lines. Measurements of the subsequent  ;

resuspension of deposited radioiodine were also made. The data were analyzed to determine the average resuspension rate constant and the best estimate of ,

) the fixation rate constant. The experimental values of the deposition l velocity, and of the resuspension and fixation rate constants are used in SAIC's model of iodine transport to estimate the transmission factor for the j depositing species (TFd ) under a wide range of conditions. The TFd is the

) ratio of the 131 12 concentration at the line outlet to that at the inlet. l When a mixture of iodine species is present, the tr.snsmission for total gaseous iodine (TFg ) is similarly defined.

)

i

)

o o

)

Estimates of the deposition, fixation, and resuspension parameters obtained from measurements made as part of this and previous studies were used to estimate transmission factors for the SGTS lines. Expected iodine

) species distributions in the air being monitored were also considered in preparing recommendations for the most appropriate transmission factors.

Some' measurement data are available on the radioiodine species present during routine operation of the SGTS at another br,liing water reactor (see Section

) 2). During normal drywell purge, the expected transmission factors for total gaseous iodine (TFg ) are shown in the following table for the two SPING monitor lines. The two lines are comparable.

Expected TFo for Line Normal Ooerition SP!NG 1 0.97 SPING-Il 0.98

)

The recommended initial post accident transmission factors (TFs) for the SPING and AXM lines are shown'in the following table. Soon after the

) occurrence of an accident, the predominance of elemental iodine in the dischargs cannot be ruled out, so the initial values are for that species.

)

Initial (2-h) Post-Accident Transmission Factor for line Elemental Radioiodine (TFql

) SPING 1 0.68 SPING II 0.79 AXM-; 0.84 AXM II 0.86

)

t ii D

i

y?  ;

.As time passes, however, the mixture of radiciodine species in the closed ,:

containment building is expected to change. The transmission factors for {!

total gaseous iodine are expected to change with time. The following table  ;

k gives recommended values of TFg for selected times after the initial increase l in 131 1 concentration. There is no real difference between the expected l values of TFg for the two AXM lines and the results for the two SPING lines [

are also quite similar. f

)

Exoected Chances in Transmission Factors for Tqtal l

) Gaseous Iodine fTF g ) Followino an Ac@nt f tine- 2_h th 2.0_h 19..h 220_h  !

SPING 1 0.69 0.72 0.78 0.85 0.98 ,

) SPING II 0./9 0.82 0.86 0.91 0.99 l AXM 1 0.84 0.86 0.88 0.91 0.99 l AXM-II 0.86 0.88 0.89 0.92 0.99 l

) l l

l TRANSMISSION OF PARTICLES

), I Transmission factors for particles through the sampling lines were f computed using the OUCT code developed by SAIC. The code calculations are  !

based upon empirical measu.ements of particle deposit'on in vertical and horizontal tubes and of particle impaction in bends. Results, expressed as f.

particle transmission factors (TFp), were obtained for six particle daameters and five particlo densities.

There is unfortunately not a collection of data that defines the aerodynamic size distribution for radioactive particles in the containment building atmosphere following an accident. The following aerosol trans-mission factors are estimates. In inaking these estimates, it was assumed

that the radioactive particles will be in the normal density range cf

)

1.

iii

y_ -

4y- e- -o l

f..

1- 3 g/cm3 and that, because the SGTS discharge passes through high efficiency particulate air (HEPA) filters, only'a small fraction of the

). airborne activity will be associated with particles having diameters greater i than 1 pm.  !

<b i

) l Estimated Transmission I Line Factor (TFa I l

SPING-1 1 .,

SPING !! 1  :

" AXM ! 1  !

AXM II 1 i j

l 4

i The particle size most likely to penetrate HEPA filters is about [

) 0.3 pm. Transmission factors for this size and for a variety of othe-  ;

sizes and censities are presented in Section 4 of the report. .i

l

) .

r t I

s I

iv .

i

~

.r CONTENTS M  ;

SUMMARY

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . i  !

1. INTRODUCTION , ...................... 1

2. EVALUATION OF RAD 1010 DINE DEPOSITION AND RESUSPENSION , . . . 3 j 2.1 EXPERIMENTAL METHODS . . . . . . . . . . . . . . . . . . 5 ,

)' 2.2 MODEL OF RAD 1010 DINE BEHAVIOR 2.3 MEASUREMENT RESULTS 7

10 l

3. APPLICATION OF RAD 1010 DINE MEASUREMENT RESULTS ....... 17 [

1

4. CALCULAT.'ONS OF PARTICLE TRANSMISSION . . . . . . . . . . . . 23 l

r c

5. REFERENCES ......................... 28  !

).

i

);  !

)  ;

. i l' L t

i i

l

)- l f

v

]- l

m + .

)

1. INTRODUCTION

) This report contains the results of measurements and calculations performed to determine appropriate transmission factors for radioiodine and particles through selected sampling lines at Fermi 2. The lines evaluated were those transporting samples to the SPING and AXM monitors for the

) Standby Gas Treatment System (SGTS). There are two monitors of each type, termed Division I and Division II. Important characteristics of the four lines are given in Table 1. All of the lines are made of stainless steel.

)

Table 1. CHARACTERISTICS OF FERMI 2 SGTS SAMPLING LINES Inside

} SGTS Sampling Approximate Diameter Sampling Flow Rate line- Lenath fem) (cm) (cm3/ s)

SPING Monitor Division I 2680 1.166 1000 Division II 1620 1.166 1000

) AXM Monitor Division i 1560 0.704 100 Division II 1400 0.704 100

)

Fermi 2 staff have indicated to SAIC that the sampling lines operate at temperatures as high as 5100 with an airstream relative humidity as high as 97%. Under high humidity conditions, it is important to assure that the air stresm does not cool to the saturation temperature. If condensation of i water vapor did occur in the line, transmission of r:dioiodine would be poor and the samples would not be reliable. Variations in temperatures of as 1 l

)

r; a

e o

much as 10 0 C are not expected to affect the radiciodine transmission para-q meters, but could produce saturation conditions. The results presented in i ds report are predicated on the assumption that the heat tracing will  ;

h operate as designed and that :endensation of water vapor does not occur. .

Measurements of rauiciodine deposition on and resuspension from interior surfaces of replicas of sampling lines have been, performed by Science Applications International Corporation (SAIC) in the Rockville Line Loss Measurement Facility. The important physical and operating character-istics of the lines were replicated experimenta11f. The ' experimental testing methods are described in Section 2.1. The radiciodine depositien and resuspension model developed by SAIC for the evaluation of line loss l

data is presented in Section 2.2. The model was used to evaluate the experimental data and to compute appropriate deposition and resuspension parameters. Data evaluation is discuJsed in Section 2.3.

Section 3 contains the results of calculations of radioiodine trans-mission made using the SAIC model with pirameters derived from measurements.

) Both equilibrium and transient conditions are addressed. In Section 4, the results of calculations of the transmission of particles through all four lines are presented. References are listed in the last section.

2

^

3

I O. O j

)  !

2. EVALUATION OF RADIOIODINE DEPOSITION AND RESUSPENSION j I Transmission of radiciodine through sampling lines depends upon the )

characteristics of the line and the radioiodine species present. Ref-  !

erence 1 summarizes available measurement data on the distribution of l j radiciodine (1311 ) species in boi; aater reactor (BWR) effluents. Most )

) of the measure' rent results on which the mean values for BWRs are based are ]

given in References 2 and 3. Table 2 shows the average radioiodine species distributions that have been measured in BWRs. This average species distribution data applies to normal operation of a BdR with an augmented

) off gas (A00) system that virtually eliminates radioiodine releases from the steam jet air ejector. The species distribution is not constant but varies with time and operating conditions. The available measurements show that species distributions differ from source to source. Therefore, the  !

) mean values in Table 2 reflect a composite of many sources rather than a characteristic of all.

) i Table 2. AVERAGE RADIOI0 DINE (131 1 ) SPECIES DISTRIBUTION IN BWR EFFLUENT DURING NORMAL OPERATIONS Soecies Percent of Total ,

) Radiciodine Associated 7 with Particulates .

Elemental Iodine (12 ) 36 Hypoiodous Acid (HOI) 26

)

Organic lodides 31  :

y Elemental iodine gas (12 ) is the most reactive of the observed species and is the most likely to deposit in sampling lines. The deposi-tion velocity (V d cm/s) is the parameter often used to characterize the

) air-to surface transfer process. The deposition velocity of HOI is 3

J

c l (l ,

)  !

estimated to be <5% of *, hat for2 1 . Depositic. "eloci'ies of organic f iodides, such as methyl iodide (CH3 1), are even smaller, -0.17, cf the  ;

h deposition velocity of elemental iodine.4,5 Thus, most losses of radio-  ;

iodine in sampling lines will be due to deposition of 12 . lne amount of l

){

1 1. deposition of elemental iodine that occurs depends upon the design and l I

1 operation of the sampling line. Resuspension of deposited activity also depends upon the operating characteristics of the sampling system. -

)m a r

The iodine "racies distributions shown in Table 2 are not common to all i BWR sampling situations. Of particular relevance to the SGTS sampling lines are measurements of the radioioiline species in the SGTS exhaust st other [

)3 BWRs. Five measurements of the species of 131 1 at the inlet to the SGTS at f

~the Monticello Station between November 1974 and May 1975 were reported in j Reference 2 as was one measurement of the 131 1 species in the SGTS exhaust.

Like the results in Table 2, the Monticello data, shown in Table 3, I

)

{

TABLE 3. HEASUREMENTS OF RADIOI0 DINE (131 1 ) SPECIES DISTRIBUTION f IN SGTS AIR DURING NORMAL OPERATIONS AT MONTICELLO ,

Percent of Total 131 1  !

s May 1975 Simultaneous Measurementsb Meana Values Range of

)- for SGTS Values for SGTS SGTS .

Soecies Inlet SGTS Inlet Inlet Outlet  !

Radioiodine Associated 2.2 0.2- 8 2.3 2.8 with Particulates

) Elemental Iodine (12) 5.7 3--8 19 15  ;

Hypoiodous Acid (HOI) 21 15 28 50 18 Organic lodides 71 59- 82 28 64

) .

a. Average of four measurements,

b. One 23-h measurement period; ulue for SGTS Inlet not included in calculation of mean values for SGTS Inlet (second column).

)

6

o. .

.. represent normal operating conditions. Overall measurement uncertainties 3

for these results are estimated to be 10-20%. It can be seen from Table 3 that the SGTS inlet sample taken in May 1975 had a different distribution of

) gaseous species from the other four samples. The results for the three gaseous species (12 , H01, organic iodides) were all outside the ranges found for the other four measurements (shown in the third' column).

) The radioiodine species distributions in air in containment buildings of pressur G .. We reactors have been found to change with time.6 Similar changes would be expected in the BWR containment atmosphere. The data for the SGTS inlet, which indicate the containment (drywell) species

) distributicn at Monticello, are similar to equilibrium values for PWRs.

To evaluate radiciodine deposition and resuspension, replicas of sampling lines have been tested in SAIC's Line Loss Measurement Facility in ,

) Rockville. The following subuction describes the facility and the conduct  :

of tests. A model of radioiodine deposition and resuspension is discussed in Section 2.2 to provide the framework for data interpretation and i analysis. The results of the measurements are presented in Section 2.3. I

)

• 2.1 EXPERIMENTAL METHODS ,

Figure 1 is a schematic diagram of the line loss testing apparatus.

[ The line was coiled in an isothermal enclosure operated at the desired temperature and elemental iodine was injected into the line. During the l

injection period, radiciodine concentrations and species distributions f were measured using SAIC radiciodine species samplers. Sampler design and -

) operating characteristics are discussed in References 2 and 7. Following the injection period, the dilution chamber was isolated and the sampling line was purged with filtered laboratory air to measure resuspension of deposited radioicdine.

) ,

All sampling cartridges were counted for 13I I using a Ge(L1) spectro-I meter whose calOration is traceabit to the National Bureau of Standards.  :

Counting results were corrected for decay from the end of the sampling

) period to the time of analysis. Radioiodine air concentration resalts were 5

l 1

).

Inlet Sensors Air T (temperature)

) Q (flow rate)

H. (dew point) , ,

' ' Filters

1. ; C) todine ^ -

Dilution senerator V Chamber U elut, s a CD T, H, Q -+

y Colled G m

f Replico '

D y b Rgposs -- ~~

IMet g Samplers T, H, Q >

C & --

C D --

{ @ c) Docuum Pemp B

Outlet -- --

gup,,,

Samplers Filters l A B

-j 17 T

-9 Uecuum Pump-t l

V Figure 1. Schematic of Line loss Measurement Facility J.

6

)

i ,  ;

l

o. * -

n .

4 I

s also corrected for radioactive decay during sampling. *lhis latter decay  !

correction assumes that 'ne concentration of I3I I in the air strram was l constant during the sampling period. i l

For the laboratory tests, elemental iodine (1 )2 was generated using tne Dushman rear. tion:

l i

)  !

103 + SI' + 6H+ % 312 + 3H 2 O (1) f

,; 3 A reaction vessel (2 neck bolling flask) containing 20% sulfuric acid and  !

potassium iodate was simultaneously heated, stirred, and purged with a flow i of helium. The production rate of elemental radiciodine was controlled by l' using a peristaltic pump to inject radioactive sodium iodide solution into the reaction mixture. This method allows an almost constant production rate  ;

over the. time of the iodine injection, f

,o  ;

2.2 MODEL OF RADIOI0 DINE BCIAVIOR j

)- +

Prior to discussing the measurement results, it is useful to describe  !

the conceptual model'of radioiodine behavior in samp',ing lines and to define  !

the parsmeters being measured. Previous studies 2 of the behavior of air- f

) borne radioiodine in building ventilation exhaust and in discharge lines {'

have shown that' o decay of the short-lived isotoper exceeds the expected decay based 1 en air tranrit tires, implying a physical retention mechanism o chemical species changes occur that shift the activity balance from  !

reactive (depositing) to nonreactive forms.

A simpie two compartment model was developed 2 that incorporated the most probable mechanisms that account for the observed species c:1anges And  ;

, dep*;etiot, of short lived radioiodines. The mechtinisms are deposition of reactive species on surfaces, species transformations on the surfaces, and resuspension of deposited radioiodine. The deposit %n and resuspension I

7

)

• .. l

).

l phenoment, were also observed in laboratory studies of radioiodine behavior.7 l

'Although they have been observed frequently, the species transformation  ;

processes are not well understood. l The original mode 12 assumed that the total air volume and the interior )

surface of the line could each be treated as a single compartment. For long i sampling lines, that assumption is frequently not valid. A sampling line is now modeled as a sequence of segments in which the air concentration is '

approximately constant. The deposition and resuspension model is applied to j each segment of the line. A typical analysis using the model treats the ,

line as a sequence of twenty agments.  !

Figure 2 illustrates an imprcved model of radiciodine transmissien that has been found to be more representative of laboratory measurements of radioiodine behavier in sampling lines. Only two 'line segments are shown explicitly. The airborne activity leaving one line segment becomes [

the input for the next .egment. Radioiodine species transformations are ,

believed to take placa on the interior surfaces of the line, but the  !

) chemical transformations are not considered in detati in the model. Some i of the chemical changes lead to resuspensicn of various species; others are ,

involved in the fixation process. Although radioactive decay is not shown explicitly in Figure 2, it is included in the equations, as described below. '

)

The differential equations used to describe the transpor of activity

• in a line segment as a function of time (t, 2) are given below:

) dq

=

I+rqs ' (Av + 6 + A) q, (2) dq*

) =

Sq, - (r + A + () q, (3) dqr dt 44s

• A4 f (4)

)

8

)

l

. , o  :

r t s . , , j O i j

.O I

i

'O l r To Next  :

RADIOlODINE Segment input RADIOIODINE m m &

IN AIR IN AIR i lO i

O Hesuspension Resuspenslen Deposition Deposition  !

'g 1 P  !

1 P i

l RADIOlODINE RADIOlODINE  :

RESUSPENDABLE RESUSPENDABLE i FROM SURFACES FROM SURFACES j lO l _

i l

Fixation Fixation j lO 1 r I f  !

RADIOlODINE RADIOlODINE FIXED ON i FlXGD ON SURFACES SURFACES  ;

lO i i

l I b

i Figure 2. Model of Radiciodine Behavior in Sampling .

Lines (two line segments shown)

'0

, 9 O

(-

0, e

.[ where  ;

q, .is the activity (pCi) of the airborne radioiodine i I is the rate (pCi/s) at which activity enters the line I 43 is the activity (pCi) of radiciodine on line surfaces available l for resuspension  !

l r is the resuspension rate constnt (s*I)

Ay is the exhaust rate constant (s'I); it is the ratio of the 3

f. exhaust flow rate' (0, em /s) to the volume V s em ) of the line segment .

6 is the' deposition rate constant (s'I); it is the product of the deposition velocity (Vd, cm/s) and the surface to volume ratio

(

) (A/V, em'I) for the line ,

A is the radiological decay rate constant (s*I) f

$ is the rate constant (s-1) for the fixation process qr is the activity (pCi) of radiciodine fixed on the surface.  ;

) [

To apply the model given above the values of the parameters r. Ay , 6, l and d must be known. Using the sampling 15e descriptions data in Table 1,

, the values of (A/V) and Xy for the sampling 'iines are readi~iy computed. The l deposition, resuspension, and fixation parameters (V d , r, and () are best l

estimated from experimental data on radioiodine behavior in sampling lines.  !

The experimental measurements of deposition and tesuspensio. made f

(

using replicas of the sampling lines are discussed in the next section. '

/

The carameters needed to apply the 20 segment model to the four Fermi 2  :

sampling lines were derived from laboratory measurements. The derived  ;

parameters art only applicable to the 20-regment model discussed above.

[

Their use in a different kind of model is not appropriate and would yield invalid predictions of radiciodine transmission.

P 2.3 MEASUREMENT RESVITS ,

h l Measurements of the radiciodine deposition and resuspension in the two sampling lines are presanted below. Deposition measurements are presented i first. Subsection 2.3.2 contana the data from the measurements of ,

) resuspensicn of radiciodine.  ;

t S

L

D 2.3.1 Derosition Measurements I The results of radiciodine concentration measurements during radio-iodine injection are given in Table 4. Overall measurement uncertainties for these air samples are estimated to be about 5 percent. Test to test variability is discussed in Section 3. The radioiodine species distribu-I tions were measured during the injection period for the SPING 1 line. At least 97 percent of the injected 131 1 was identified as 12. Approximately 0.5 percent was estimated to be present as H01. About 2.6 percent of the activity was collected on the particulate filter. The less reactive

> organic iodides accounted for less than 0.3 percent of the total. During the injection period for the AXM 1 line, about 87 percent of the injected radiciodine was clearly identified as 1. 2 Nearly all the remaining activity was collected on the particulate filter. The organic iodides and H01 D accounted for less than 0.4% of the total. In both tests, mest nr all of the activity found on the particulate filter was prcbably 61emental iodine that had adhered to the filter. These species fractions are generally consistent with those measured previously in the test facility.

D Table 4. MEASURED RADIO 10 DINE CONCENTRATIONS DURING THE INJECTION PERIOD D

Average 131 1 Measured Line Sampling Concentra Transmission Desionation point __(gCi/cmgion ) Factor _

SPING-1 Inlet 9.60x10-7 0.91 Outlet 8.73x10'7 (4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />)

AXM 1 Inlet 1.04x10-6 0.78 Outlet 8.16x10-7 (4.8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br />)

D 11 B

1 I

1 y, , o l

)

The measured transmission factor during the injection period (IP) was about 0.91 for the SPING 1 line (IP = 4 h). Transmission of 12 through the I SPING-1 line was higher than expected. The expectation of a lower TFd (4 h)

) was based upon experience with lines of similar diameters. Although the I flow rates through the other lines were different, review of the current  !

experimental data base indicates that this result is at the extreme of the l range. The injection period for the AXM-1 line was slightly longer (4.8 h) l

) and the TFd was 0.78. Transmission of 12 through the AXM-! line was some- l what lower than expected, but was within the anticipated range of results.

The measured 131 1 concentrations were used to compute the average

) deposition velocity for 1 2 in each line section. It is assumed that the l fractional loss due to deposition is the scme in each of 20 segments and that resuspension can be ignored during the 4 hour4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> injection period. If C s I31 1 concentrations (pCi/cm3) at the inlet and outlet of the and C*, tre the

) segment, respectively, that assumption means (C*s/C g ) is the same for each j segment and thet (C*,/Cs )'

• I'd (IP). Under the assumed conditions, Equation (2) is simplified and its solution can be manipulated to yield an expression for the deposition velocity in the line. The average deposition

) velocity was computed using Equation (5) for a 20 segment line. .

I i

Q (1 - C,/Cs )  ;

V d" (5)

A

) s (C*/C s s)  :

where V

d is the average deposition velocity (cm/s) of 1 2 in any segment of ,

the line

) Q is the exhent .' low rat,e (cm /s) for the line (and for each j

segment)

A s

is the interior surface area (emI) of the line segment.

) This procedure was applied to the results for both the SPING-1 and AXM-I lines. The estimated deposition velocities were 0.010 cm/s and 0.0071 cm/s, respectively, for the two lines. A note of caution is appropriate here.

The procedure used to estimate values of Vd is approoricte only for the

20. segment model of the sampling lines.

J 12

)' l

, I

% g 2.3.2 Resuspension Measurements

. Following the period of deposition during radiciodine injection, filtered laboratory air, containing no 33II , was drawn through each line at the normal sampling rate. The line outlet concentration was monitored to determine the I3I l activity resuspended from the Interior surfaces.

Table 5 contains the results of measurements of 131 1 resuspended from the walls of the SPING 1 and AXM-! lines following deposition of I3I I in the two lines. The total measurement period was about nine days for the SPING I

) line and about ten days for the AXM ! line. The concentrations of resus-I3I l decreased monotonically with time for the SPING-1 line, but were pended more variable in the AXM 1 line. Overall mea.urement uncertainties for the tabulated concentrations are estimated to be between 5 and 15 percent.

).

) Table 5. MEASURED CONCENTRATIONS OF 131 1 DUE TO RESUSPENSION IN THE SPING ! AND AXM ! LINES SPING-1 Results AXM-1 Results Duration Average 131 1 Duration Average 131 1 Sampling of Period Concentragion of Period Concentration Period (minutes) fuci/cm ) (minutes) (uci/cmM I 1400 3.7x10-9 1060 3.8x10 9  ;

) 2 1440 7.8x10-10 1530 1.1x10*9 i 3 1360 3.8x10 10 1560 4.2x10-10 4 1515 2.9x10-10 1620 1.7x10*9 5 1470 1.4x10-10 1460 2.3x10-10

) 6 2820 1.1x10-10 1600 7.2x10-10 7 2820 6.3x10-Il 2610 8.4x10-ll i

8 -- -- 3050 1.8x10-10 13 >

h

r The concentration data shown in Table 5 were used to estimate the resuspension rate constant (r) for each sampling period and the fixation rate constant (3). The first step in this process is to determine s provisional values of the resuspension rate constant under the assumption that ( = 0. The procedure that was used is described below, f The surface activity at the start of the resuspension period (qso,pC1) was estimated using the measurements during radioiodine injection, using Equation (6):

q,, = ACQt g (6) where AC is the concentration difference between the inlet and outlet of the line and t g is the duration of the injection period. For a period of no radiciodine input and approximately constant air concentration, Equation (2) can be rearranged to yield an expression for the average resuspension rate constant r. The equation is C V (6 + ly + X) r = , (7) 9s 3

where C is the av6 rage concentration (pC1/cm ) of I3II . For the conditions stated above (I - O and q, constant), Equation (3) yields the following expression for the surface activity as a function of time:

)

q (t) =

s (1-e'*st ) + q so e s t (8) s

) where o - r + A + ( is the total removal rate constant describing loss of resu pendable I3I I from the surface. The average surface concentration during a sampling period (ts , s) under the stated conditions is given by y Equation (9):

14

)

o

)  !

i g 64, (3,,-as3t ) . q,L9__ (3,, ag t,) j s a, ,2 t ,, og ts 1

The mean surface activity during a period determines the computed  !

values of the resuspension rate constant and is, in turn, affected by it.  ;

j Equation (6) was used to calculate q3 , for the first measurement period. [

Equations (7) and (9) were used iteratively to compute self consistent i values of the resuspension rate constant and surface activity for the first j measurement period. Then Equation (8) was used to compute the surface activity that would be present at the beginning of the second measurement f

)

f period. The calculations using Equations (7), (9), and (8) were then j performed for each subsequent resuspension measurement period. The value  !

of 6 used in the calculations for each line section was based on the  !

deposition velocity estimated using the measurements discussed in Section f

)

2.3.1. I I

The calculated provisional resuspension rate constants decreased substantially with time. The rate of decrease was used to find an initial

) estimate of the fixation parameter p. Calculations of resuspension rate  ;

constants were then performed in the manner described above for a range of l values of $. The range ar.d variability of the computed resuspension rate constants was examined for each set of values corresponding to a particular j 4 The best-fit value of p was determined by finding that value which  !

minimized the variation in computed values of r.

The final resuspension rate conttants for the SPING I Line are shown in Table 6. It was found that 3.2x10-6 s'I was the best value for the ,

fixation rate constant p. The mean value of the resuspension rate constant

• I The range of the (1 standard deviation of the mean) was 2.410.5x10-6 s .

final values of r is about a factor of 3. This reflects a substantial improvement over the provisional values of r, which ranged over a factor of 20. Although substantial effort is required to determine the fixation parameter and corresponding resuspension rate constants, the results l indicate that it is worthwhile and that the revised model (Figure 2) is '

) an improvement over previous calculation schemes.

15

)

ht , ' o Table 6. BEST FIT RESUSPENSION RATE CONSTANTS DERIVED p FOR SPING 1 LINE AND AXM-1 LINES Best fita Best fitb Value of r (s-1) Value of r (s*l)

Period for SPING-1 Line for AXM-1 Line

'- 1 4.4x10 6 },4x10 6 2 1.7x10-6 5.3x10 7 3 1.4x10 6 2.9x10*7 4 1.7x10 6 1.7x10 6 5 1.4x10 6 3.5x10*7 6 2.3x10 6 1.6x10-6 7 4.2x10 6 3.0x10-7 8 -- 1.3x10 6 D

a. Best fit obtained for ( = 3.2x10 6 s-l.

b. Best fit obtained for p = 2.3x10 6 s*l.

Also shown in Table 6 are the final resuspension constants for the AXM 1 line. The mean value of the resuspension rate constant (1' standard deviation of the mean) was 8.812.7x10-7 s-l. The best fit was achieved for a fixation rate constant ( = 2.3x10-6 s*l. The greater variability, due to

' the observed fluctuations in the concentration of resuspended 1311 , was also D reflected in the factor of six range of values for the eight periods. The range of provisional values of r was more than a fattor of 20, so the revised model led to improvements in parameter evaluation for this line as

-well.

D-Application of the results of the measurements is discussed in the following section. It should be noted that the deposition, resuspension, and fixation rate parameters in this section are appropriate only for the I 20 segment model used to derive them from the measurement results, i

16 1

= ,

l

3. APPLICATION OF RADIO 10 DINE MEASUREMENT RESULTS The deposition and resuspension parameters derived from the experimental measurements are used in calculations of transmission of 1 2

through the sampling lines. Yne results are expressed in terms of the

) transmission factor (TFd ), defined to be the ratio of the concentration of 131 12 at the outlet of the line to the concentration entering the line.

, _ The lower the 1Fd , the greater the loss of 131 12 due to deposition in the line. The 20 segment model was used to calculate equilibrium transmission

(

J factors and to estimate radioiodine behavior during departures from equilibrium conditions. These calculations are discussed below, l The deposition velocities and resuspension rate constants for the lines were measured at the nominal line operating temperatures. There is evidence

that both parameters depend upon temperature; however, the data are limited i

in scope and, for deposition velocity, rather variable. Small changes (i 10 00) in the line operating temperature are not expected to have an important effect upon either the deposition velocity or resuspension rate i

constant. Temperature variations of this magnitude are not expected to affect the TF d, so long as condensation of water vapor does not occur.

Observed differences in multiple measurements of TFd at the same

temperature indicate the overall variability of the process. Duplicate l tests with a single line yielded transmission factors of 0.48 and 0.62.

l Tests of four lines of the same diameter and flow rate, but differing lengths, yielded transmission factors ranging from 0.45 to 0.72. When the h

l results were normalized to a common length, the variation in transmission l factors for the four lines was 0.55--0.61.

l h During a period of constant elemental iodine input to the sampling line, the surface activity in eae.h segment will gradually increase to an equilib'tur,value. the equilibrium is reached when deposition of elemental iodine onto surfaces in the segment is balanced by resuspension of 33I I from those surfaces. Calculations of the balance in each line segment and of 17

)

1

e

, o h

the net transmission of activity under those conditions were made using the 20 segment model and the values ofdV , r, and d derived from the laboratory

> measurements. The parameters used in the calculations are given in Table 7.

These are meant cf sets of results based upon experimental measurements using replicas of the Fermi 2 SPING 1 and AXM 1 lines and other lines comparable to them.

Table 7. PARAMETERS USED IN CALCULATIONS OF RAD 1010 DINE TRANSMISSION THROUGH FERMI 2 SGTS D EFFLUENT SAMPLING LINES Parameters 4 for Radiciodine Transmission Model Line Desionation yd-(SJDLil d IS'I I r (s*l l D

SPING-1 0.040 3.5x10 6 2.8x10 6 SPING !! 0.040 3.5x10-6 2.8x10 6 p AXP-1 0.0050 6.4x10-6 5.0x10-7 AXM-II 0.0048 6.4x10-6 5.0x10-7 D

a. Parameters determined from laboratory measurements of replicas and similar lines.

D The equilibrium transmission factor for the depositing species (TFde) is defined as the ratio of the 131 1 activity leaving the line to that D entering the line (I, pCi/s), when equilibrium has been reached. The equation is:

TF de "A9v ae(20)/I (10) 18 D

L

.. . 1i

)  !

I I3I I i

where q,,(20) is the equilibrium activity in the last (20th) line f

stgment and yl is the previously defined exhaust rate constant. The cal- ,l

) cu hted values of TF de are shown at the bottom of Table B. Although little is known about thn radiciodine species that would be expected following a  ;

" typical" accident, the equilibrium transmission factor for Ip (TFde) I

< provides a lower bound for the long term average transmission factor. l

)

Table 8. TIME DEPENDENCE OF TRANSHISSION FACTORS FOR 131 1 FOLLOWING A LARGE INCREASE IN EFFEUENT CONCENTRATION Transmission Factor for 131 1.2 Lif.dI Time (hours) SPING 1 SPING Il AXM-1 AXM II After Event Line line LiDL. Line 2 0.68 0.79 0.84 0.86 5 0.69 0.80 0.84 0.86 20 0.72 0.82 0.85 0.86 50 0.76 0.84 0.85 0.87 TOO 0.79 0.86 0.85 0.87 Equilibrium

) 0.79 0.86 0.85 0.87 Value (TFde)

The four lines are all estimated to have a TFd e of about 0.8 or greater for elemental iodine. As discussed below, expected values for a mixture of species are even higher.

If the input of depositing species to the sampling line increases from IA (pci/s) to IB (pci/s), it will take time for the line to reach a new equilibrium distribution of airborne and surface activity. The deviation would be greatest for releases during startup after an extended shutdown or 19

)

E-I

'toi fequilibrium.

'IB"I. The A 20-segment model was used to predict the rate of approach Table 8 also contains the computed values of the trans-mission factor for depositing species as a function of time for the four lines. Little change is seen with time for the AXM lines because the initial transmission factors are so high that a large change is not

' possible. For the two SPING lines, resuspension of previously deposited radiciodine causes the equilibrium transmission factors to be 10-16% higher

' than the initial values.

1 >

For the two SPING lines, the equilibrium transmission factor for total gaseous radimindine (TF g ) will be almost certainly higher than that

. quoted abovs for 12 . As Table 3 shows, elemental iodire accounted for only 15 percent of the total 1311 in the sample from the SGTS effluent. Because the deposition velocities of H01 and organic iodides are small, the trans-

{

mission factors for these species are approximately one. At any time, the frectional transmission of the gaseous radioiodine (TF g) is approximated by:

TF g

TFd x F, + Fh+F o (11) j

where F, is the fraction of the gaseous radioiadine that is in elemental form, F

h is the fraction of the gaseous radioiodine present as HOI, F

o is the fraction of the gnseous radiciodine that is ir oroanic form, I

a F and the other quantities were defined above. The average values of the  !

$ fractions F,, Fh , and Febased are calculable using the data in Table 3 and

. are 0.15, 0.19, and 0.66, respectively (based upon the single SGTS outlet '

measurement for normal purge operations).

_ If these average values apply to the Fermi 2 SGTS discharges, then

, .during normal purge operations, the values of TF ge would be expected to be

about 0.9) and 0.98 for the SPING-1 and SPING-II lines, respectively. Soon after a sharp increase in activity with no change in soecies, TF g is esti-mated to be 0.95 for the SPING-I line and 0.97 for the SPING-II line. Of

C course, the value of TFg , at any time depends upon the actual distribution

( 2n ql ll l

$ p ill In iiiiii < i lli i i iiis i iq i,g, p i

zpwr ,  ;

. .;

b. - l of radiciodine species. ' Better estimates can be made when more data on species distributions in SGTS effluents are available. I p )

The overall radiciodine transmission factor must also reflect the transmission factor for the pari.iculate fraction (TF,) which depends upon ,

the size distribution of the radioiodine activity. The overall transMssion T factor for radioiodine is: j

~

'F; -

Fp 'x TF, + TFg x (1 - Fp) _(12) ].

)

{a where pF is the. fraction.of the total radioiodine activity in particulate Previous measurements at the Monticello BWR indicate thatp F is less form.

than 0.03 for SGTS discharges (see Table 3). Values of transmission factors l

'for various particle sizes and densities are given in Section 4. l 4

j

'At times greater than a few hours after an accident, the transmission factors for total gaseous 131 1 (TFg ) in the AXM lines would be expected'to increase and exceed the corresponding values of TFd given in Table 8. As time passes, the fraction of the airborne radiciodine that is elemental would be reduced and the concentrations of less reactive forms would increase. There are a few post-accident measurements of radiciodine species -

in PWR containment buildings that indicate this change. The first '

measurements of radiciodine species in the THI-2 containment were not made

. u'nti . about' 3' months had elapsed.8 The initial measurements of 131 1 and 9

subsequent measurements of 1291 showed that the elemental iodine fraction was less than 0.1 at long times after the accident. -

[ ,

Measurements (,f species changes in the containments of PWRs during l normal operation may indicate the rate at which post-accident species  :

changes would occur. Table 9 contains estimates of Fe based upon measure-ments during normal operations in PWR containments.6 The:e estimates were f

used to project values of TFg for the two AXM lines. Although results are i

shown for both lines, there is no real difference betwoen them.

1 21 1 1

Y I

?

Y . --. - .- _ -

~

'o ,

, , J

):.

,, '7 2 1 Table 9. ESTIMATED CHANGES IN ELEMENTAL IODINE FRACTION AND yc PROJECTED TRANSM1SSION FACTOR (TFg ) WITH TIME Projected Estimated Time Elemental Iodine Transmission Factor >

(hours) Fraction (Fe ja .

LIEg) for AXM Lines  :

$YM AXM-II

2 1 0.84 0.86 ,

5 0.9 0.86 0.88 20 0.8 0.88 0.89 50 0.6 0.91 0.92 i s

200 0.1 0.99 0.99

) i, 7

! 4 3r ii Derived from best fit curve through data for PWR containments 't a.

- durino normal ooerations. The species distributions'there are '

presumably similar to those that would be found in the drywell. a.

b. Projection assumes that observed changes in Fe with time also apply to cost-accident conditions.

If the input of 1 2 into a line decreases sharply from I A to I C (pCi/s),withI A " I ,. then the measured outlet concentration can exceed  ;

C h the inlet concentration. This may occur due to resuspension of radioiodine previously deposited on surfaces of the line. The largest effect would be soserved if the line had reached equilibrium with the initial input rate before the decrease occerred.

The 20-segment model described above was used to estimate the magnitude

• of this effect. The resvits showed that even following uilibration of these sampling lines with re 1,2 the effect on a 24-hour air sample would ,

). be an overestimation of the 1311 release by less than 1% of the equilibrium release rate.

22

'c' .,., Il y l E i

/ 4.0 CALCULATIONS OF PARTICLE TRANSHISSION I b

The frectional transmission of particles through sampling lines depends i upon the characteristics of the particles and upon the characteristics of ,

the sampling line. The most important particle characteristics are the

> diameter and density. This section contains results of calculations of i transmission factors for the Fermi 2 lines described in Table 1 for a range l of particle sizes and densities.

[ The aerosol transmission calculations were performed using the DUCT computer code developed by SAIC. The computer code is based on empirical relationships for various depletion mechanisms. The calculations consider  !

Brownian and turbulent diffusion, sedimentation, and impaction of aerosol

) particles in the sampling line.

For each line segment, the DVCT code computes the fraction of particles of the'specified diameter and density that: (a) are deposited in a given section of line and (b) are lost by. impaction in bends in the line. The '

frecti:- deposited (F, dimensionless) is given 10 by F - 1 - exp (-4 V dL/hu) (13)

)

where Vd is the deposition velocity (cm/s) of the particles L is the length (cm) of the line segment h is the diameter (cm) of the line u is the mean fluid velocity (cm/s) in the line.

Separate calculations are performed for horizontal and vertical line segments. Slant sections are treated as horizontal, which generally over-j estimates the losses in the slanted line segments.

For horizontal lines, the deposition velocity is computed using the relationships of Matsui .qt. al.11 and Yoshioka et al.12 When gravitational  ;

[ settling contributes to the deposition velocity, that contribution is 23 Q .

r L

i + , '

h -

,- ' ' estimated using Stokes's Law. For. vertical line segments, the relationships g, developed by Beal l3 are employed to compute V ' '

d l N The overall transmission factor for a line reflects deposition in both

' the~ horizontal and vertical sections. The transmission factor for particles.  :

through the line, TF p, is computed using

[' TF p

=

(1 Fg ) (14y) (14)

, { where 'F H is the fraction of particles deposited in tt.e horizontal sections and bends i Fy is the fraction of particles deposited in the vertical  ;

. sections and bends. .

t The transmission factor for particulate material through a sampling +

J line can vary over a wide range, depending upon the size and density of the airborne particles. The. aerosol. being sampled will frequently consist of a R '

variety of partidles of differing sizes. For that reason, many calculations of transmission factors for specific particle sizes and densities were per-formed. The following paragraphs deal with the use of the computed values J' of transmission factors for specific pa<'icles (TF p) to determine the

transmission factor for the aerosol of interest (TF,). ,

j The transmission factor for a mixture of particles can be computed ,

using:

• b TF a

I(TF)g p x RF $ (15)

D' l where TF is the transmission factor for the aerosol, is the transmission factor for particles in the i th size (Th)$p O.. s range, End RF g is the friction of the aerosol radioactivity associated with particles in the i th size range (IRF4 = 1).

)

I 1

24 9:

".7 ,

i

s n ,. l 1

l I .The size ranges to be used in the calculations will be determined by the j

. impactor used for the measurements and its operating characteristics. )

)t Svne units divide the aerosol into six fractions; others employ additional (or fewer) stages .and provide a more (or less) detailed breakdown of particle j

sizes. This section contains values of TF p for a specific set of particle ,

sizes. Calculated transmission factors for other particle sizes can be

). provided to match the radioactive particle size distribution determined by ,

different impactor designs.

, For some situations there will be no measurement data that define the radioactivity distribution versus particle size; post-accident aerosols are-f in this category and are perhaps the most difficult to characterize. In-the absence of measurements of the fractionation of radioactivity among the  ;

various particle sizes in the ambient aerosol, a reasoned analysis must be

) used to estimate TF,.

Results of the calculations of particle transmission through the four SGTS sampline-lines are presented in Table 10 for six particle diameters in the range 0.1 gm to 5 pm. Results are given for five particle densities, ranging from 1.0 to 10 g/cm3 .

%e results for the SPING monitor sampling lines indicate the TFp is ,

0.9 or greater for particles with diameters $1 pm, regardless of dencity.

h For the normal range of particle densities,1--3 g/cm3, the values of TFp are uniformly high for submicron particle diameters, but decrease. rapidly for larger sizes. The -values of TF pfor the AXM monitor sampling lines also decline rapidly for particle diameters greater than 1 gm. Values of TFp

. for submicron particles e):ceed 0.7, regardless of particle density.

At Fermi 2, all of the normal and post-accident SGTS effluent discharges pass through HEPA filters prior to tronitoring and release to the environment.

)

The particles most likely to penetrate such filtration systems are small ,

(-0.3 pm). The results in Table 10 show that transmission losses for 0.3-ym particles will be negidgible in all four SGTS lines for any of the densities considered.

25

.b

} js , .\

f ~

s i 1

Table 10. COMPUTED TRANSHISSION FACTORS FOR PARTICLES FOR FERMI 2-SGTS SAMPLING LINES

): ,

TRANSMISSION FACTOR (Tcp ) FOR i

. spi'CIFIED PARTICLE DIAhETER AND DENSITY  ;

';

(a/cm 3-

) Particle Density 3

}". Particle  ;

Diameter (um) 1.d., R L_0_ ,L L lt Results for SP!NG-I line _

)..

0.1 1.0 1.0 1.0 1.0 1.0 i 0.3 1.0 1.0 1.0 1.0 0.99- l;

. 0.5 1.0 1.0 0.99 0.99 0.98 1.0 'O.99 0.98 0.98 0.96 0.89 i

3.0 .0.90 0.4A 0.16 0.013 a j l

5.0 a a a a a

. .:l Results for SPING-II Line l 0.1 1.0 1.0 1.0 1.0- 1.0 0.3 1.0 1.0 1.0 1.0 1.0

[ 0.5- 1.0 1.0 1.0 0.99 0.99 1.0 1.0 0.99 0.99 0.97 0.94 3.0' O.94 0.58 0.29 0.054 a' 1

• 5.0 a a a a a

i 1

a. Calculated transmission factor was <10-2 ,

t 26

)  ;

t.

f. C? 'w g

• i i

s ,

Table 10 (Continued)

m COMPUTED TRANSMISSION FACTORS FOR PARTICLES FOR FERMI 2 SGTS SAMPLING LINES ,

t TRANSMISSION FACTOR (TFp ) FOR .,

SPECIFIED PARTICLE DIAMETER AND DENSITY.

)

Particle Density (a/cm 33 Particle L Diameter (um) Q Q Q Q 1 Results for AXM-I Line

. 0.1 1.0 1.0 1.0 1.0 l '. 0 0.3 1.0 0.99 0.99 0.99 0.97 k 0.5 0.99 0.98 0.98 0.96 0.93

'1.0 0.97 0.9A 0.91 0.86 0.73 3.0 0.76 0.55 0.37 0.11 a .

5.0 a a a a a

, r Results for AXM-II Line 0.1 1.0 1.0 1.0 1.0 1.0 0.3 1.0 -1.0 0.99 0.99 0.98 0.5 0.99 0.99 0.98 0.97 0.93 ,

1.0 0.97 0.95 0.92 0.87 0.76

}

3.0 0.78 0.59 0.40 0.11 a 5.0 a a a a a

)

a. Calculated transmission factor was <10-2 ,

)

27 y, ,

r __

e .. d l

f.

a 5., REFERENCES h ~1. J. H. Keller, L. G. Hof fman, and P. G. Volllequd, Wet Deoosition Processes for Radioioclinsi, NRC Report NUREG/CR-2438 (August 1982). ,

2. . C. A. Pelletier, E. O. Barefoot, J. E. Cline, R. T. Hemphill, -1 W. A. Emel, and P. G. Voillequd, Sources.__of Radioiodine at Boilina l1 Water Reactors, EPRI Report NP-495 (February 1978).  ;

1 .

l

3. P. G. Voillequd, Iodine Soecies in Reactor Effluents and in the Environment, EPRI Report NP-1209 (December 1979).

4. M. J. Kabat, Deposition of Airborne Radiciodine Species on Surfaces

[. of Metal and Plastics,'in Proceedinas of the 17th DOE Air Cleanina ;i

) Conference, DOE Report CONF-820833 (February 1983).

131

5. P. G. Voilleque and J. H. Keller, Air-to-Vegetation Transport of 1 as Hypoiodous Acid (HOI), Health Rvt.,,, M, 91-(1981).

6. J. W. Mandler, B. G. Motes, C. A. Pelletier, A. C. Stalker, T. E. Cox, P. G. Voillequs, S. T. Cror.ey, D. W. Akers, C. V. McIsaac, N. K. Bihl, G. A. Soli, S. W. Duce, ,'. K. Hartwell, sl. W. - Tkachyk, and L. S. Loret, ,

In-Plant Source Term Me'.surements at Four PWRs, NRC Report NUREG/CR-1992 (August 1981).

7. R. T. Hemphill and C. A. Pelletier, lurface Effects in the Transoort of

)J Airborne Radiciodine at ti,qhi Mater Nuclear Power Plants, t."RI Report

[ NP-876 (September 1978).

U 8. P. G. Voillequ6, Reactor Containment Soecies Samoles, Memo Report of I

Analytic.nl Results, Science Applications, Inc. (July 1979).

9. J. E. Cline, P. A. Roy, J. W. Ho11 croft, J. Hobaugh, T.' t. M::Vey, C. Th of g$91omas, Jr., C. A. Pelletie and Radioactive , and Particulate P. G. Voillequd, Concentrations Mgiturements in the TM;-l I Containment Atmotphg.rt Durina and After the Ventino, 00E Report ,

GEND 009 (April 1981)

10. A. K. Postma and L. C. Schwendiman, Tut bulent Deposition in Sr.mpling

[- Lines, in Proceedinos of the AEC Air Cleanina Conference, AEr Report TID-7627 (1961), i

'11. H. Matsui, Y.'Yashida, M. Murata, and T. 0hata, Measurement of Deposi- l L tion Fraction of Aerosol Particles in a Horizontal Straight Metal Pipe,  !

). J. Nucl . Sci . Tech. , 11, 300 (1974).

12. N. Yoshicka, C. Kanoaka, and H. Emi, Kaaaku Buturi, s, 89 (1971) (in I Japanese).

13. S. K. Beal, Deposition of Particles in Turbulent Flow on Channel or Pipe

) Walls, Nucl . Sci . Enar. , M (1970) . i i

28 l

).

e