ML20053C708

Accident Risk Assessment of Toxic Chemical Shipments Transported on Il Central Gulf Railroad Near Clinton Power Station. Response to NRC Question 311.6 Encl

ML20053C708
Person / Time
Site:	Clinton
Issue date:	05/28/1982
From:	Riley T ILLINOIS POWER CO.
To:
Shared Package
ML20053C699	List: ML20053C698 ML20053C708
References
U-0494, U-494, NUDOCS 8206020487
Download: ML20053C708 (26)

Text

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

Attachment 1 to U-0494 i i

t k

1

'I  ;

f i

j '

i l

.: ACCIDENT RISK ASSESSMENT 1

0F T0XIC CHEMICAL SHIPMENTS TRANSPORTED ON THE ILLINOIS CENTRAL GULF RAILROAD NEAR THE CLINTON POWER STATION I-l 1

1 l

i i

l 4

1

, Terry L. Riley i Staff Specialist  ;

Illinois Power Co.

Clinton Power Station i

i i

i J

U 8206020487 820528 PDR ADOCK 05000461 E pyg

% e , ..~..----.c--.y- -,r._ ,._.,$_-. _._ _ _ _ .7, yp,_,, , ,f. , , _,..,.,_9 ,.,.,y, ,,, ,,,,,~py ,,,, . ~ _ _ , . . , , , , , , _ _ , ,, , _ ,.gy,.,,,, ,

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

l' ABSTRACT l

A study of the hazards associated with rail trans-portation of toxic chemicals via the Illinois Central i Gulf Railroad "Gilman Line" (ICGRGL) was performed by

. the Illinois Power Company (IP). The probability of an l accident involving the release of toxic chemicals resulting in deleterious effects on control room habitability was determined to be 9.1x10-9 per year for the most limiting hazardous chemical shipped on the ICGRGL. Regulatory Guide 1.70 defines design basis toxic chemical accidents external to nuclear power plants as those accidents having a probability of occurrence of 10-7 per year or greater.

Therefore, those hazardous chemical shipments transported near the Clinton Power Station (CPS) were determined to

! have such a low probability of occurrence that they should not be considered design basis events, f

I f

l l

I

• I I

l

_ - . . . - . _ _ - _ _ _ , - . . _ . . . . . . . _ . . _ _ _ _ ~ . _ . . - - . _ - , _ , , _ _ _ . . . _ _

l i List of Illustrations / Tables Figures Page

1. Transportation Routes within a 5-mile Radius of the Clinton Power Station..... 10

2. Same as Figure 1 (shows quadrants used

] in calculation of meteorological term,

, "M").................................... 11 i

Tables Page

1. Toxic Chemical Shipment j Frequency on the ICG Gilman Line............................. 12 4

2. Toxic Chemicals that Remain '

as Potential Problems Following the Hazchem Analysis.................... 13

3. Distribution-Joint Frequency"M" Calculation of Term................. 14 a

i i

h

TABLE OF CONTENTS Section Page

1. Introduction.......................... 1

2. Hazardous Chemicals Analyzed.......... 2

3. Risk Assessment....................... 6

4. Conclusions........................... 9 l

l t

I

=

(

o

LIST OF REFERENCES

1. " Safety Evaluation Report related to the operation of Clinton Power Station, Unit No. 1"; NUREG-0853; February 1982 (CPS-SER).

2. Regulatory Guide 1.78, " Assumptions for Evaluting the Habitability of a Nuclear Power Plant Control Room During a Postulated Hazardous Chemical Release";

June 1974.

. 3. CPS Final Safety Analysis Report (CPS-FSAR), Section 6.4.

4. Regulatory Guide 1.70; Revision 3; November 1978.

5. C. D. Bossard, Superintendent, Hazardous Materials, Illinois Central Gulf Railroad, letter dated March 5, 1982 to S. A. Hallaron of Sargent & Lundy.

6. C. D. Bossard, Superintendent, Hazardous Materials, Illinois Central Gulf Railroad, letter dated November 2, 1981 to S. A. Hallaron of Sargent & Lundy.

7. CPS-FSAR, Section 2.3; Table 2.3-15.

8. N. Irving Sax, " Dangerous Properties of Industrial Materials,"

Third Edition, Reinhold Book Corp., New York, N.Y. (1968).

9. " Patty's Industrial Hygiene and Toxicology," '?olume 2A, Third Revised Edition, G. D. Clayton and F. E. Clayton, Editors,. John Wiley and Sons, New York, N.Y. (1981).

10. Material Safety Data Sheets for each chemical supplied by manufacturers.

11. T. Riley & Erich Kant of IPC and C. D. Boussard of ICGR, Record of Coordination, dated May 17, 1982.

1

, . - - .m., _ . - - - , _y, -- , _ _ ,

1. INTRODUCTION

The Nuclear Regulatory Commission (NRC) Staff has required IP to submit an analysis of the hazards associated with rail transportation of toxic chemicals via the Illinois Central Gulf Railroad Gilman Line (ICGRGL) near the CPS.

The Staff's position was recently stated in the CPS-Safety Evaluation Report (SER) as follows:

1 "The nearest railroad is a line of the Illinois Central i Gulf Railroad which runs parallel to State Route 54 and traverses the site approximately 0.75 mi (1.21 km) north of the station. The Illinois Central Gulf Railroad also has a line approximately 3.5 mi (5.6 km) south of the station.

The hazards associated with rail transportation of toxic and explosive materials are still being evaluated. Based on 1976 and 1980 transportation data obtained from Illinois Central Gulf Railroad, the applicant has identified several materials requiring further analysis. These will be addressed in a future SER supplement."

IP has studied data available on hazardous chemical

shipments for various transportation nodes near CPS. The i

data was obtained from the Illinois Commerce Commission (IOC), ICGR, and other agencies or companies involved in

~

such shipments.

Figure 1 shows all transportation routes within a 5-mile radius of CPS. A radius of 5 miles was chosen to determine compliance with Regulatory Guide 1.78 (Ref. 2).

The nearest highways to the station are Illinois State Routes 54, 48 and 10 and U. S. Highway 51. The ICGR

! "Gilman Line" runs parallel to State Route 54 and transverses the IP property approximately 3/4 mile north of CPS. The ICGR also has a line approximately 3.5 miles south of the

! station.

The potential for the release of toxic chemicals in the vicinity of CPS has been evaluated for all modes

, of transportation. Since the frequency of transportation l of toxic materials via highway routes was determined to be less than the Reg. Guide 1.78 value of 10 per year, l any further analysis of accidents via this transportation i mode was not justified. Through conversations with the ICGR, it was determined that the railroad line 3.5 miles i south of CPS is not used for shipment of toxic chemicals and is currently being considered for abandonment in the near future. Therefore, this report provides a discussion of the toxic chemicals shipped via ICGRGL,.and the Potential risk associated with cuch shipments.

2. HAZARDOUS CHEMICALS ANALYZED Reference 3 provides a detailed discussion of the design bases forthe CPS main control room habitability systems. The NRC Staff has requested that IP consider accidents involving all hazardous material shipments on the ICGRGL to determine if deleterious effects on control room habitability exist. This concern was re-cently stated in NRC Question 311.6 as follows:

"Regarding toxic chemical shipments on the Gilman Line near Clinton Station the applicant indicated that he re-quested and received from Illinois Central Gulf Railroad information on all of the toxic chemicals listed in Table C-1 of R. G. 1.78. Table C-1 is only illustrative and as such is incomplete. The applicant should consider all hazardous material shipments in the plant vicinity.

Based on the frequency of shipment and the specific hazard the applicant should develop a list of potentially hazardous shipments requiring detailed evaluation and conduct such an evaluation."

R. G. 1.78 (Ref. 2) provides detailed criteria for determining if an evaluation of the habitability of a nuclear power plant control room during a postulated hazardous chemical release needs to be performed. These assumptions are listed briefly below:

1. hazardous chemicals frequently shipped within a 5-mile radius of the plant.

2. shipments are defined as frequent if there are 30 or more per year for rail traf#ic.

3. determine toxicity limits from Table C-1 (or from appropriate authoritative sources).

4. two types of accidents should be considered:

(a) maximum concentration chemical accident; (i) chemicals that are not gases at 100 F and normal atmospheric pressure but are liquids with vapor pressure less than 10 torr, need not be considered.

(ii) the atmospheric diffusion model should be the same or cimilar to that of Appendix B.

(b) maximum concentration-duratibn accident; (i) guidance on atmospheric diffusion model given in R.G.1.3 and R.G.1.4.

5. detection instrumentation, control room isolation systems, filtration equipment, air supply equipment and protective clothing should be evaluated as needed.

The ICC provided a list of chemicals transported on the "Gilman Line" for the months of October, November and December, 1981. Frequency of shipment information was not included. The ICC advised that this list should be fairly representative of all chemicals shipped through-out the year.

In addition, the ICC provided a record of the annual average quantity of materials shipped throughout the state of Illinois. Any toxic chemical shipped with a stated frequency greater than 100X per year was evaluated by IP along with the chemicals listed on the 3-month "Gilman Line" record. It was assumed that a chemical transported less than 100X per year throughout the entire state would be transported less than 30X per year on the "Gilman Line". This approach is believed to be conservative.

In addition to the above criteria, slightly toxic or non-toxic chemicals were removed from consideration. Those liquids with less than 10 torr vapor pressure at 100 F (per R.G.1.78) were eliminated.

A list of all chemicals not eliminated from consideration was submitted to ICGR. For these chemicals ICGR provided shipmenc frequency information, for the Gilman Line, for the entire year of 1981. Those chemicals that were shipped on the "Gilman Line" more than 30X in 1981 are shown in Table 1. In addition, some chemicals were shipped more than 30X for the years 1976 and 1980. These chemicals are also listed on Table 1 and are evaluated in this re-port. This approach was considered to be conservative for two major reasons:

1. The frequency data obtained was stated in terms of cars / year. It was assumed that one car represented one shipment for comparison with the R.G.l.78 shipment frequency criteria. This is conservative, since in many cases, more than one car per shipment was probably made. On this basis alone, many of the chemicals listed would not present hazards from a shipment frequency standpoint, and,therefore,could have been eliminated from concern.

2. Yearly average shipment frequencies were not determined. Any chemical that exceeded the 30X per year criteria, for any year that data was available, was considered further.

1  ;

1 Each of the chemicals listed in Table 1 was considered further to determine if hazards associated with an accidental release existed.

IP's Architect / Engineer, Sargent & Lundy, has de-veloped a computer analysis program, entitled HAZCHEM, l that evaluates the habitability of a power plant control

room in the event of a toxic chemical spill at or near a power plant. The HAZCHEM model has been developed from R.G.1.78. There are two HAZCHEM programs: one pro-gram is applicable to materials that are gaseous at ambient conditions and one program evaluates the spill of a toxic liquid.

1 The following assumptions are included in the program-for gaseous releases:

1. Instantaneous spill of total contents of a tank containing the chemical.

, 2. Ground release of tank car contents.

3. Control room intake is modeled as being directly downwind of the point of chemical release with no intervening structures.

4. The chemical is a gas at the input temperature and 14.7 psia but is stored or transported as a liquid under pressure.

1 5. Instantaneous release results in a puff of finite volume described by the puff model for atmospheric dilution in Appendix B of Regulatory Guide 1.78.

6. The diffusion equation for an instantaneous (puff) ground level release used in the program was taken directl The "y"yand from"z"Appendix terms in B of theRegulatory diffusion Guide equation 1.78.

were assumed to be zero. This assumption centers the puff at the control room intake in the hori-j zontal crosswind and vertical directions.

The relationship x = D - ut as defined in Appendix B was directly substituted for the "x" term in the equation.

7. The value calculated by the equation represents the chemical concentration at the intake to the control room. The program uses the control room ventilation characteristics to determine the chemical concentration inside the control room from the intake value. Concentration levels are calculated for various equally spaced wind speeds up to the maximum wind speed supplied as input.

The HAZCHEM program for liquid spills uses the diffusion equation of Regulatory Guide 1.78 in the same manner as the program for gaseous releases. The same assumptions apply, except that the chemical is assumed to be a liquid at ambient conditions. The mass of chemical vaporized is calculated as a function of the spill radius, the molecular weight, density, vapor pressure, and molecular diffusivity in air of the chemical, and the ambient temperature and pressure.

The spill radius may be input directly or calculated by the program from input values for spill thickness, chemical surface tension, viscosity, density, and diffusivity.

As in the HAZCHEM gas program, all necessary cheaical properties, weather conditions, and ventilation values are given as inputs to the program.

As can be seen from the above description of the HAZCHEM model, the assumptions made in the analysis are very conservative. The basic idea is, that if the accident occurs, it will occur in the worst possible way, and under the worst possible meteorological conditions, such that the effects on the control room habitability will be worse than what would be anticipated.

Based on the HAZCHEM analysis, propylene oxide, butyl acetate, ethyl acetate, methanol, methyl ethyl ketone, and Styrene (monomer) were determined to present no severe toxic hazards to the control room operators per the require-ments of R.G.l.78. For each of these chemicals, the maximum control room concentration would be less than the two-minute exposure limit. Where two-minute exposure limits were not available from published sources or chemical manufacturers, a value of two times the published threshold limit value for a chronic 8-hour per day exposure was used. IP believes that this approach is quite conservative.

Table 2 lists those chemicals that remained for further consideration. The next section of this report presents a risk assessment of the accidental release of any of the remaining chemicals being considered.

3. RISK ASSESSMENT Reference 4 (R.G.1.70) provides a method for determining if a toxic chemical accidental release need be considered a design basis event. The position of R.G.1.70 is stated below:

"2.2.3.1 Determination of Design Basis Events. Design basis events external to the nuclear plant are defined as those accidents that have a probability of occurrence on the order of about 10-7 per year or greater and have poten-tial consequences serious enough to affect the safety of the plant to the extent that Part 100 guidelines could be exceeded. The determination of the probability of occurrence of potential accidents should be based on an analysis of the available statistical data on the frequency of occurrence for the type of accident under consideration and on the transportation accident rates for the mode of transporation used to carry the hazardous material. If the probability of such an accident is on the order of 10-7 per year or greater, the accident should be considered a design basis event, and a detailed analysis of the effects of the accident on the plant's safety-related structures and components should be provided.

Toxic Chemicals. Accidents involving the release of toxic chemicals (e.g., chlorine) from onsite storage facilities and nearby mobile and stationary sources should be considered.

If toxic chemicals are known or projected to be present onsite or in the vicinity of a nuclear pl ant or to be frequently transported in the vicinity of cne plant, releases of these chemicals should be evaluated. For each postulated event, a range of concentrations at the site should be determined for a spectrum of meteorological conditions.

These toxic chemical concentrations should be used in evaluating control room habitability in Section 6.4."

The risk assessment analysis used in this report

employs a simple, but conservative, approach to calculating the probability of a toxic chemiccl release that results

! in potential consequences serious enough to affect the habitability of the CPS Control Room. The railroad accident i probability is defined by the following equation:

Pa = Pr xLxMxN I

where, Pa = accident probability (year -1)

Pr " Probability of rupture ( car -mile -1)

L = length of ICGRGL track inside a 5-mile radius of the CPS = 9.7 miles

M = meteorological term representing probability that atmospheric conditions are such as to carry toxic chemical release directly to Control Room air intake N = number of tank cars shipped per year 1 (cars / year) ,

ICGR has provided IP with the data necessary to deter-mine the probability of a rupture resulting in a toxic chemical release. The data needed is listed below:

1. Number of ruptures of hazardous material railcars in the history of ICGR = 2 ruptures.

2. Average number of hazardous material railcars shipped per year over all ICGR lines = 130,000 cara / year.

3. Number of miles of ICGR railroad track = 7900 miles.

In applying the data in item 2 above it was assumed that this value was applicable to the last 10 years of ICGR operation.

Thus, the probability of rupture becomes:

Pr = 2 ruptures (10 years)(130,000 cars / year)(7900 miles)

Pr = 1.95 x 10-10 ruptures / cars -mile.

There are several conservatisms present in the value of Pr. The two major items are:

1. The value used for the number of ruptures does not diffentiate between a rupture which results in a total release of the toxic chemical or one

! shich results in a partial release of the toxic

! chemical involved. Ruptures resulting in small partial releases of the chemical concerned, would 3 not have a significant effect on the habitability of the control room.

4

2. The 2 ruptures represent data over the entire history of the ICGR. This represents well over 50 years of operation. However, in applying the data to the calculation, only the last 10 years was assumed to apply.

The data used to determine the meteorological term (M) is discussed in detail in the CPS-FSAR Section 2.3 (Reference 7). This data is listed in Table 3, with a brief discussion of how this data was utilized to calcu-late "M". The method used to calculate "M" is a realistic approach that utilizes the following bases:

1. The total length of ICGRGL track inside the 5-mile radius was partitioned into sections of varying lengths, as a function of wind direction sector.

This was based on the fact that, depending upon where the accident occurred, the wind direction had to be such that the toxic chemical was carried directly to the control room air intake in order for a problem to exist.

2. Stability classes A, B, and C were deleted from consideration since these Pasquill Categories result in highly unstable atmospheric conditions that would not be conducive to a slow diffusion of the toxic chemicals. This maximizes the theoretical concentration of the toxic chemical at the control room air intake.

3. One control room air intake faces Lake Clinton.

A significant impact of the lake will be the warm surface it presents to the atmosphere which, during nighttime and the winter, will be signifi-cantly warmer than the surrounding ground. This increase in temperature will cause the layer of air in contact with the lake to achieve a neutral lapse rate, especially when stable conditions prevail over the land. Thus, material released from a ground-level source would receive additional diffusion in the vertical over the lake than would be computed using a stable delta T stability

, cctegory determined from the meteorological tower.

l IP did not take credit for this factor in determining "M".

4. The length of each track section was determined from a marked-up copy of Figure 1 (presented.as Figure 2).

k .

The accident probabilities were calculated for each of the chemicals shown in Table 2. The limiting value is for ammonia, and is shown calculated as follows:

Pa - Pr xLxMxN P

ammonia

= (1.95 x 10-10 ruptures / car-mile) x (9.7 miles) x (0.051) x (94 cars / year)

Thus, P ammonia

= 9.1 x 10-9 per year The results are listed for all 5 chemicals below:

Chemical Pjt Ammonia 9.1 x 10-9per year Vinyl Acetate 4.5 x 10 per year Ethylene Oxide 4.3 x 10 per year Bromine 4.2 x 10 per year Hydrogen Fluoride 3.9 x 10 per year

5. CONCLUSIONS R.G.l.70 defines design basis toxic chemical accidents external to nuclear power plants as those accidents having a probability of occurrence greater than or equal to 10-7 per year. Those chemicals that remained for further consi-deration, following the initial data reduction and the subsequent HAZCHEM analysis, were found to present accidental occurrence probabilities 1-2 orders of magintude smaller than the R.G.l.70 value. Based upon the fact that none of these toxic chemicals for ICGRGL transportation, result in accident probabilities of 10-7 or greater per year, none of these accidents need to be considered as design basis events for CP.S.

O.. ' . . , i l. . - f_ S'j M. @,7.n'~

~

p ,7

. A _

. ,e 4 _ ".

~.

'. O. - -

. . _ _ ^_

s

. ~y .- .. _ .

.:~. .-. , ~ n, ,.. m .

~ .,, . y;s

~- ._ x: ... L *. ,..:..: - . ~- 2 f : ~ *' '

.3,

_ffQ.K _

y. .

.as 1_.- -

- d "- _ .._ u . 2 2 ,,

?

. +

. - }, n - -

i I .

\-

3 l 1 I \. / .

I m '

\,

N' g

g } l I

jlV,r

.\ l [ I '  ! l

, g __ _g__ _ _ _ _ _ , _ _ _ _ _ . - s. -

.t .l 1

'g 3 / ~e -

? \l*

J_. g / /

l g l /

, t l.\, . .--

! \ I f ,

.! l

• g I f *.' p

/

d l l .f*/

lp,-

%. I g I I s  !

s.

- o g, _______2_____,_ y! ,

w

, g p

.j .

E -

'p' f g - - -

f g [ LE GE ND: ,

I 7 l q ,/g ._. .- ASHL A ND OIL PIPELINE -8**

g -/ \ CLIN T O N ST A TION g

• i CAPPED AND ABANDONED o

[ g ,

/ ^-

N .. . ... 4LLINOIS POWE R G A S LINE -2**

.~ J O

! / r*\.

__-- SHE LL OIL PtPE LINE - 14**

Ppp :i ---- PHILLIPS G A S PIPELINE -8**

f- g g , .f () ..--- E X PLORE R PIPE LINE -24**

l \, ILLINOIS CE NTR AL GULF R AILRO ADS

'!. o

/ / \* .'-g HE AVY DUTY ROADS

'f s,r%. / 4 \, i IJ p* .%. .___- MEDIUM DUTY ROADS Y t)9 O ~ '

I

\  !] -

x f I e \  :

l - .

\,

\*

.f l g .' e ur o

~

$- a 2 6

j s %1 *# f. ,

3,....._., O10 y %ts i

. /~j ' .83 ,. , # \. I i

/ / \ I_

a a A #4l q g .

r f CLtNTON POWER STATION g 3

g , , _ - .

fin AL s AFETY AN ALYSIS REPORT l L _ _ _ _ _ _ _, <./. _ _ _ _ _ .Lli.

4 q

i  : t i 1 .

/

I

\ TRANSPORTATION ROUTES AND PIPELlh[5 [

l* WITHIN A 5-MILE RADIUS OF THE I I / #

l . CLINTON POWER STAi!ON t

,_-__._.__'.1_'________/ _

=----_--- J-\*--------

. _ _ _ _ _ _ _ _ _.\_..____ _ _.____ _

~

..'.~_,e l f6 . ;.mo s' s .A --z ,

. 7; .

4 ., f _ . _ ,

n .yA ,.,w u .~ ,, .- ,

r "r '

,s.m.s,

s"#

g -- _y

)- b e ~ p

^ '3.t. .

~

n' ,

.., . g;

' ~

• ,A

~q..

'^g

'7<M'r

" , .,y

..m,.

5 . - -

g. .

, / . ." L .; ' ' * .

f

^

. , - 'f ,

~ 1 s '. ; , ,

' w i,l '

_.;w ' _

, ... -r_v y r.: -

i e

~

_. ~ ,' ~ --.a.<..! ..y* * ~

l g

. - l- >*J _ ', ; . . .

N\. s

\. _ _ ,

,____.:._g;.

_ _ _ _ w_ _ _ _ _ _ ;_

,e 7

sj __\____ll ,

s N i

. ,. _ i . .i

,/ ..,

j

.- i.' '

ix.

!@ g f O=v;6,, -p ,.

g #

s. l'

<

N s\ /1 \'

N r # x 4k i s /!

v x! x i

i , ,

,l ,,, ,  :

. n -

>--- g -,

e

/

ii 's .

/,' r.

/ \

i l /

/ ,

p _ d *-rNl-]-

/ g I CLINTON POWER STATION 4

i i

)

_s /

_ # _ _ _ _ _ .s t.

..I riNat sartiv aNatvsis arront

, 3 2 / \ TRANSPORTAT!DN ROUTES AND PIPELINES

i [' I 1 *g WITHIN A 5-MILE RAD!uS Of THE l / f l . CLINTON P0h1R STATION J I ' <

_ _ _ _ _ _ _ _ _ _ J_ \

, .i.

,. / \ -

g s e t ,C

~ _ . . u- - . .j -1 ,~.-.1 - , - , - - *i- *-

1 Table 1 Chemical Shipment Frequency on the ICG Gilman Line Chemical Shipped (1981) Frequency (cars / year)

I

1. Styrene (monomer) 106

2. Propylene oxide 48 3 Vinyl Acetate 47

i. Butyl Acetate 46

5. Bromine 44

6. Hydrogen Fluoride 41 7 Ethyl Acetate 39 8 Methylethyl Ketone 35 1

l Chemical Shipped (1980) Frequency (cars / year)

1. Methanol 115

2. Ammonia 94 Chemical Shipped (1976) . Frequency (cars / year)

1. Ethylene Uxide 45 l

! s i

a i

6-Table 2 i

Toxic Chemicals that Remain as 3 Potential Problems Following the

HAZCHEM Analysis ,

Chemical Shipped Frequency (cars / year) Year 4

l 1. Ammonia 94 1980 4

l

2. Vinyl Acetate 47 1991 i

3. Ethylene Oxide 45 1976

4. Bromine 44 1981 4

5. Hydrogen Fluoride 41 1981 T

4 4

T I

i J

l i

I e

{

l l

t l

l yv--n-- - w e my y-m~r -w e ng r.wg-re-*-*,.-3 w ,;- - - , -, , = - - . - --,rv.%ev -- -. -w-w e -v er . .m --w-e , - - - . ,-e. - < - - - , - + ~ + -a

~~

ABl_E 3 deinT~ EREQUEMC4 )1STRIBUTloAl - ALculATloAT OF W IERM.

IcGRGL @Ch WIND

, STABluTV class * .

! secr$ t @Y 3 E F G hjs C MMEMTS

1. Total Trade Len C "ches 1 2.06 W 3.50 1.'79 0.70 o,48 6.47 ,,,i,) = u , g, q ,,,,j Oowi Fipre. 2. .

i

, E 0.56 WNW 4.05 1.4B 0l10 0.4B fo.71 a.ReM = meteorlobc..) i d te i

fu!kesedr r >>i fermsa Fseck

, 3 02.9 NW 290 1.01 0.50 0.44 6% $l > "ji,'d $"d!Q,*g'

^

mi+; vale w .ae. ~

4- 0.21 NNW 2..%o 0.77 0.32. o.25 3.M v

"'d.  ;/-k.!'7chdTlh.?

eslumn for eab tenen s ra' 5

5 03.1 N 2.03 o.91 0.39 0.16 3.61. +g-k,4*'****I m

t ;-loo 6 0.22 NNE 1M o.7B 0.35 0.37 3.36 h.000.*0 +@s00.70 +'

Co.29)C4.w.) + Co.206.LO +

(o.22X3.61) + Co.za)Dat)+

7 0.50 hE 2.08 1.13 0.5+ 0.M 4.39 n= co sec4.vo +0 aoD.sq 6.06 8  :.94- ENE 1.69 1.io o.s4- 0,50 3.B3 .

,. = 0.051

• RePre5eds % of tisine. he wind n Nou;ny fr= h yvea dinc%.CrsaR, Table 2 3-15 ').

s *

.',, o.

n,, , , Attachment 2 to U-0494 NRC QUESTION 311.6 '

N

' RESPONSE The potential hazard to the station as a result of delayed ignition of a liquid petroleum gas '(LPC) cloud formed as a result of an accidental rupture of a tank car on the nearest railroad has been evaluated..

Since propane is representative of the more common LPG's, a large capacity tank car loaded with 160,000 lbs liquid propane gas was considered. The release rate at which the gas escapes from the ruptured tank depends on the size of the rupture and the pressure and temperature conditions inside and outside the tank. The cloud volume and its' location relative to the station depend on the rate and duration of the gas release and the meterological con-ditions which would possibly exist at the time of release. Thus, to estimate the pr~obability of damage to the plant from this hazard, all combinations of the following factors and their effects on the station were evaluated:

1. Ruoture Size

a. A 23.5-inch puncture hole releasing 2,666.7 lbs/sec gas for a period of 60 sec (Ref. 1).

b. A 14.5-inch puncture hole releasing 888.9 lbs/sec gas for a period of 180_sec.

c. A 7.5-inch puncture hole releasing 266.7 lbs/sec gas for a period of 600 sec. (Ref. 1).

e

d. A 4.75-inch puncture hole releasing 102.6 lbs/sec gas for a period of 1560 sec.

~

2. Meteorological Conditions

a. Entire range of atmospheric stability classes given in FSAR 2.3.2.1.6; i.e., stability classes A,B,C,D,E,F, and G.

b. Six wind speeds, selected as the averages of the six inter-vals given in FSAR 2.3.2.1.6; i.e., 2.8 ft/sec; 7.4 ft/sec; 13.3 ft/sec; 21.5 ft/sec; 30.4 ft/sec; and 36.7 ft/sec.

c. Sixteen wind directions; i.e, N, NNE, NE, ENE, E,...,NNW.

The' motion and the configuration of the gas cloud in the horizontal plane were determined for each rupture size from a gas dispersion analysis using'a Gaussian plume model. The model determines the extent of the flammable region of the cloud for a given meteorological condition by calculating the gas concentration as a function of time and space coordinates relative to the point of release. i .

Knowing the concentration limits of the flammable region to be s

between 2.8% and 7% gas by volume for propane, the volume and <

centroid location of the flammable cloud relative to the station were calculated. '

/

The equivalent TNT mass yield of 240% recommended in Regulatory Guide 1.91 was used to calculate the weight of the flammable volume of the cloud. '

The' detonation hazards were determined by calculating the yearly probability of exceeding one psi overpressure at the plant. Combi-

! ~

f. .

/ .

' nations of various rupture locations along the railroad, meteoro-logical conditions, and detonation times were evaluated in the estimation of this probability. The probability of exceeding the one psi overpressure at the plant for each rupture size was calculated by dividing the railroad in question into a number of segments and performing an ' analysis similar to the one described as acceptable for pipeline analysis in NUREG-0014 (Ref. 3). On this basis, the proability of overpressure hazard, P, at the plant per year from a railroad is calculated by:

P=PF. E Py(S,V,D)xP d (Ty _y ,T7) xL (N) xd (S,V,D,I,N)

N=1 S=1 V=1 I=1 D=1 gy) where P = probability of rupture per tank mile r

F = frequency of shipment of tanks carrying LPG, in shipments per year Py(S,V,D) = probability that wind of Stability Class S, speed V, and direction D is blowing when _ , . _

detonation occurs P d(Tg _t,T 7) = probability that detonation occurs between~

times T7 _y and T7, given that a rupture has occurred

. L (N) = length of railroad segment N, in miles NP = number of railroad segments considered in the analysis NT = number of detonation time intervals d (S ,V,D ,I ,N) :

=

1 if overpressure exceeds the one psi ' criterion for S,V,D,I,N

.t

=

0 if overpressure does not exceed the one psi criterion

Analysis of the accident data applicable to all 7900 miles of Illinois Central Gulf lines (Ref. 4 and 6) indicates the following:

1. The average number of hazardous material railcars per year is 130,000.

2. The number of LPG tanks shipped during calendar year 1981 is 31,105.

3. In the history of this railroad there were 2 ruptures of hazardous material cars.

4. None of the ruptured cases involved an explosion. One of the ruptured tanks involved ignition of the escaping gas.

The probability of detonation, given a rupture, is calculated as:

1 rupture & esplosion P(D/R) = 2 ruptures

= 0.5 (2) where it is conservatively assumed that the number of ruptures with explosion is equal to the number of ruptures with ignition.

Conservatively, assuming that the 2 ruptures are applicable over the last 10 years of ICGR operation, the probability of rupture for LPG is calculated from the above data as follows:

(3) 2 ruptures X

31105 LPG tanks /vr (10 years)(130,000 tanks /yr)(7900 miles) 130,000 tanks /yr

= 4.66 x 10-11 LPG ruptures per tank , mile

Possible ruptures were considered to occur in a 13800-ft length of railroad near the plant. The probability of a rupture in a given railroad segment, Pr, was assumed to be proportional to the length of the segment. Segment lengths of 300 to 800 f t were used to discretize the railroad and to calculate the exposure distance. For each segment, the point closest to the plant is chosen as the point at which rupture is assumed to occur.

The probability of detonation time was assumed to follow the exponential law. Using this assumption, the probability of a detonation occuring between two selected time instants, T y and T2, was calcluated from the equation

-T 1 /8 -T / 6 Pd (Ty,T2 ) = P (D/R) [e -e 2] (4) where 8 is the mean time to detonation.

When applying Equation 4, eight discrete time intervals were selected such that for a given rupture size, the time range covered the time needed to discharge the contents of the tank and to dissipate the flammable concentrations of gas. In all cases, the range was increased until the computed probability showed

, no increase for increases in total time. Based on the observed l

l time to ignition reported in Reference 1 for two rail tank car accidents, the mean time to detonation used in the analysis is nearly 300 seconds. In addition, for the 23.5- and 14.5-inch l

i ruptures, average times to detonation of 90 and 120 seconds, respectively, were used in the analysis, because of the rapid dumping of tank contents at these rupture sizes.

r

The value of F used in this evaluation is 96 tank cars (Ref' 5).

This value represents the number of LPG tank cars whose intended routing was over the Illinois Central Gulf Gilman line near the plant in 1980. The results show that the worst condition would occur af ter the punctur'ing of the 23.5-inch hole. For this case, the probability of a rail tank car accident in which detonation occurs resulting in an overpressure at the plant in excess of 1 psi is 8.9 X 10 -11 per year.

Since this probability is less that the 10

-7 per year given in Section 2.2.3 of Regulatory Guide 1.70, it is concluded that the delayed ignition of LPG gas clouds caused by the accidental rupture of a rail tank car need not be considered a design event.

The following conservative assumptions were used in the preceding evaluation:

1. The entire contents of the tank are assumed to flash vaporize, whereas, in actuality, only 1/3 of the contents may vaporize, the remaining 2/3 staying in the form of liquid droplets.

2 No rise of the plume due to bouyancy has been assumed.

3. The equivalent weight of TNT was taken as 240% of the gas contained in the flammable cloud region. RG 1.91 indicates that the 240% represents an upper limit for hydrocarbons.

4 The safe overpressure for the safety-related structures was taken as 1.0 psi, whereas the maximum safe overpressure is 1.65 psi, which is one-half the total tornado wind design load.

~ '

REFERENCES

1. T. V. Eichler and H. S. Napadensky, " Accidental Vapor Phase Explosions on Transporation Routes Near Nuclear Power Plants,"

report prepared for the NRC, NUREG/CR-0075, April 1977.

2. U.S. Army, " Structures to Resist the Effects of Accidental Explosions," TM5-1300, June 1969.

3. NUREG-0014, " Safety Evaluation Report for Hartsville Nuclear Plants."

4. C. D. Bossard, Superintendant, Hazardous Materials, Illinois Central Gulf Railroad, letter dated March 5, 1982, to S. A.

Hallaron of Sargent & Lundy.

! 5. C. D. Bossard, Superintendent, Hazardous Materials, Illinois Central Gulf Railroad, letter dated November 2, 1981 to S. A. Hallaron of Sargent & Lundy.

6. T. Riley and E. Kant of IPC and C. D. Bossard of ICGR, Record of Coodination, dated May 17, 1982.

i t

s

• -_, e