ML20024C364

Limited Appearance Statement of TB Cochran Re Issues Raised in CP Proceeding.Discusses Radiological Consequences of Crbr Core Disruptive Accident & Site Suitability.Certificate of Svc Encl

ML20024C364
Person / Time
Site:	Clinch River
Issue date:	07/08/1983
From:	Cochran T National Resources Defense Council, Sierra Club
To:
References
NUDOCS 8307120557
Download: ML20024C364 (80)

Text

. . - -- ..

< r

~

, 50 .N-l July 8, 1983 BEFORE THE U.S. NUCLEAR REGULATORY COMMISSION ATOMIC SAFETY AND LICENSING BOARD - ~

10i

) / Dx .w v.

In the Matter of ) /2'sY ' C

) c'N UNITED STATES DEPARTMENT OF ENERGY ) id JUL A 1983 T PROJECT MANAGEMENT CORPORATION TENNESSEE VALLEY AUTHORITY

) \Q 't f:

)

} g/y 1f^ f

/-

(Clinch River Breeder Reactor Plant) 6f  %[

LIMITED APPEARANCE STATEMENT OF DR. THOMAS B. COCHRAN REGARDING ISSUES RAISED IN THE CONSTRUCTION PERMIT PROCEEDING Pursuant to 10 CFR $2.715, and in accordance with the Board's order of June 29, 1983, Dr. Thomas B. Cochran hereby submits a limited appearance statement on behalf of the Natural Resources Defense Council. Inc., and the Sierra Club, regarding several issues raised by the Board for resolution in the upcoming CRBR construction permit hearings.

I. The Radiological Consequences of a CRBR Core Disruptive Accident Staff has evaluated th~e radiological consequences of' Applicants' postulated CRBR core disruptive accident (CDA) l i scenarios and reported the results in Appendix A.5 of the March 1983 Safety Evaluation Report (SER) and the May 20, 1983, SER Supplement No. 2. According to Staff, the evaluation used

" realistic (albeit conservative) assumptions" (SER Suppl. No. 2 at A.5-1), including'50% X/Q meteorology (SER Suppl. No. 2 at

, 8307120557 830700 PDR ADOCK 05000537 T PDR

A.5-3). The low population zone (LPZ) thyroid dose was reported to be 192 rem (SER Suppl. No. 2 at A.5-4) using thyroid dose conversion factors taken from TID 14844 (NRC Staff's Response to Intervenors' Third Set of Construction Permit Interrogatories and Request to Produce to Staff, Resonse to Interrogatory 1(d), p. 2, May 20, 1983)

Staff claims the 192 rem thyroid dose at the LPZ

. . . gives the staff confidence in the applicants' claim that the critical organ dose for a CDA would be within the 10 CFR Part 100 dose guidelines and that

. . . the comparison to 10 CFR 100 dose guidelines is made here to provide perspective regarding the relative severity of Ehe CDA consequences and to provide assurance that if such an event were to occur that adequate accommodation has been provided to limit the consequences of such an event, so that doses would not exceed dose guidelines in

• 10 CFR 100.

(SER Suppl. No. 2 at A.5-1.)

I dispute these claims on several counts:

First, Staff has calculated the thyroid dose for an adult, but the infant thyroid should be considered the critical organ of interest. Infants can be expected to receive a thyroid dose twice that of an adult or, in this case, approximately 400 rem --

i some 100 rem (or one-third) higher than the 300 rem guideline value for thyroid used by Staff. Evidence for this is as follows:

The thyroid dose conversion factor for inhalation of I-131 given in TID 14844 at p. 25 is 1.48 x 106 rad /Ci inhaled (= 1.48 i

, _ . - + . - . , , - - , , . - , . . , . , - - . , , - - - , ..

x 10-3 mrad /pci inhaled), the same value as that given for an adult in NUREG-0172 at Table 8, p. 2 of 4. Likewise, the breathing rate used in TID 14844 at p, 23 is the value for adults, 20 m3 / day (= 7300 m3 /yr), as indicated in NUREG-0172, Table 8-4, at p. B-4.

The ratio (infant dose / adult dose) for inhalation of I-131 can be calculated from data in NUREG-0172 (at Table 5, p. 2 of 4; Table 8, p. 2 of 4; and Table B-4, p. B-4) as follows:

D131(infant) , 1.06 x 10-2 mrad /pci X 2045 m3 /yr 2 D131(adult) 1.49 x 10 3 mrad /pci 7300 m 3 /yr Similar calculations can be made for other halogen isotopes.

Second, for purposes of judging the adequacy of CRBR containment to mitigate CDAs, Staf f uses as a benchmark the 10 CFR 100 dose guideline values developed for siting analysis (SER Suppl. No. 2 at A.re-1). In '.he 10 CFR 100 site suitability analysis at the CP licensing stage, Staff requirements are to reduce the guideline values by approximately a factor of two to account for uncertainties in final design detail and meteoro.!ogy and new data and calculational techniques that might influence the final design of engineered safety features or the dose reduction factors allowed for those features.

NRC Staff's Supplemental Answers to Intervenors' Twenty-Sixth Set of Interrogatories to Staff, at pp. 19-20. Staff, for example, uses a thyroid dose guideline of 150 rem at the CP stage, rather than 300 rem used at the OL stage. (1982 Site Suitability Report (SSR) at p. III-9.) Staf f fails to apply the same logic --

although it applies equally -- to the CDA analysis, realizing of course that, if they did so, the calculated " realistic" CDA adult thyroid dose of 192 rem would exceed the 150 rem guideline value. Staff's failure to apply the same logic in the two cases is arbitrary and serves only to ensure licensability of the current CRBR design rather than to protect the public health.

Third, I believe the estimated severity of a CDA at CRBR, assuming " realistic (albeit conservative)" conditions, namely, thyroid doses of 192 rem to adults and 400 rem to infants at the LPZ boundary, is excessive and should not be tolerated for CRBR, or for any reactor. In effect, this is also a challenge to Staf f's use of 300 rem (to the adult thyroid) as a benchmark to judge the adequacy of CDA mitigation based on a realistic CDA scenario (c*., SER Suppl. No. 2 at A.5-1).

The basis for this view is a direct comparison of these thyroid doses against the observed medical findings in a Marshall Islands population accidentally exposed to fallout from the 15-megaton thermonuclear device (code named Bravo) tested at Bikini Atoll on March 1, 1954. A summary of the thyroid abnormalities that have appeared as of 1981 are reported in the attached table (Attachment 1) taken from Robert A. Conard, M.D., et al., " Review of Medical Findings in a Marshallese Population Twenty-Six Years After Accidental Exposure to Radioactive Fallout," Brookhaven National Laboratory, BNL 51261, Jan. 1980, Table 1 of Chapter IX,

p. 59. These data speak for themselves. I note only that exposure occurred only 26 years ago, that many of the victims are

still in their early years, and that additional thyroid abnormalities can be expected as the survivors grow older.

II. Combined Probability and Consequences of CRBR Core Disruptive Accidents In the LWA proceeding, Intervenors Natural Resources Defense Council, Inc., and the Sierra Club presented an affirmative case regarding their Contentions 1 and 3, namely that Staff's and Applicants' analysis of the consequences of CDAs, coupled with Staff's analysis of the probability of a CDA ( Appendix J of the CRBR FSFES; Staff Exhibit 8), demonstrate that the Commission Standard Review Plan criterion for identification of design basis events is not met and consequently the CDA should be a containment DBA. (See Intervenors' Proposed Findings of Fact for the LWA-1 Proceeding, January 24, 1983, at TT l-23). I hereby reaffirm and incorporate that testimony in my. statement today and request that the Board take that evidence into account in the current proceeding.

Using CRBR design-specific information generated by Applicants and Staff, I am able to provide additional evidence in support of our earlier claim. Any of the documents cited below will be made available to the Board upon reque,st.

Pursuant to Intervenors' CP discovery request, Applicants made available the bulk of Applicants' CRBR probabilistic risk assessment (PRA) analysim that were by Board Order ruled beyond the scope of the LWA-1 proceeding (see Natural Resources Defense Council, Inc., and the Sierra Club First Set of Construction 1

Pe,rmit Interrogatories and Request to Produce to Appicants, April 7, 1983; Letter from Thomas A. Schmutz to Dr. Thomas B. Cochran, June 20, 1983, with enclosure). Among the documents produced was EG&G Idaho, Inc., Wood-Leaver and Associates, Inc., and Fauske &

Associates, Inc. " Clinch River Breeder Reactor Plant Probabilistic Risk Assessment -- Phase I Main Report," EGG-EA-6162, January 1983. (Selected pages of this voluminous work are attached as Attachment 2.) I wish to call attention to two aspects of this work. First, as evidenced by the Abstract (reproduced in Attachment 2), this PRA has as its overall objective a " realistic evaluation of the risk" associated with CRBR, with the caveat that, since the entire PRA must await Phase II, the results of Phase I must be interpreted cautiously.

The second aspect of this work that I call to your attention is its estimate of the cumulative probability of dominant Core Damage Sequences (i.e., CDAs), of 1.1 x 10 4/yr (see Attachment 2,

p. 8-11), which is dominated by loss of offsite power scenarios (Attachment 2, p. 11-2).

In sum, whereas Staff in Appendix J of the Final Supplement to the CRBR Final Environmental Statement (FSFES) estimated that a " conservative," or upper bound, estimate of a CDA at CRBR was 10-4/yr, Applicants' consultants calculate the same frequency based on " realistic," as opposed to Staff's conservative, assumptions. The sensitivity analysis performed by Applicants' l

consultants suggests the upper limit in the probability of a CDA at CRBR may be even higher than 10-4/yr ( Attachment 2, pp.11-8 and 11-9).

l

One can combine the Applicants' consultants' best estimate of CDA frequency of 10-4/yr with Staff's " realistic (albeit conservative)" estimate of the thyroid dose at the LPZ boundary of 192 rem to adults (400 rem to infants) in order to compare the results against the Commission's Standard Review Plan guidance for identifying DBAs.1,/ Certainly by this test the CDA ought to be a containment DBA; the probability of exceeding 10 CFR 100 guidelines is approximately three orders of magnitude too high to exclude the CDA from the containment DBA envelope.

I wish to anticipate several responses to this observation.

First, if the Board were to reject my view that the appropriate CP thyroid dose guideine is 150 rem rather than 300 rem, and were also to reject the argument that the infant thyroid dose should be examined as the critical organ dose, the Board might conclude that the calculated 192 rem to the adult thyroid 1/ The Commission's Standard Review Plan for light water reactors (Staff Exh. 6, p. 2.2.3-2) states:

[T]he identification of design basis events resulting from the presence of hazardous materials or activities in the vicinity of the plant is acceptable if the design basis events includo each postulated type of accident for which the expected rate of occurrence of potential exposures in excess of the 10 CFR Part 100 guidelines is estimated to exceed the NRC staff objective of approximately 10 ' per year. .

. . LTJhe expected rate of occurrence of potential exposures in excess of the 10 CFR Part 100 guidelines of approximately 10-6 per year is acceptable if, when combined with reasonable qualitative argument, the realistic probability can be shown to be lower.

(Emphasis added.) (

,l I

l

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

is well within the 300 rem guideline. The response to this is straightforward. The 192 rem estimate is based on the median of some 8600 X/O values in all cardinal directions (i.e., 50%

X/Q). At a somewhat higher X/Q, the adult thyroid dose at the LPZ boundary would exceed 300 rem. I do not know at what X/O percentage this would occur since the X/O spectrum is not reported. However, Staff could readily produce this figure. For purposes of argument, I will assume that, at the 80% X/Q level, the adult thyroid dose would in fact exceed the 300 rem guideline. If so, then the dose guideline value would be exceeded for 20% of all CDAs, or 0.2 x 10-4/yr, still well above the 10-7/yr requirement for excluding CDAs from the DBA envelope.

A second response is likely to be that the quantitative probabilities, or the PRA results, are highly uncertain and therefore should not be used as a basis for determining the DBA envelope in lieu of the " judgmental" approach taken by Staff.

This is a correct response to the wrong question. PRAs are indeed highly uncertain, and their primary function should be to identify previously unrecognized risks to health and safety. As a prudent health and safety practice, one should use great

, caution in applying highly uncertain PRA results to argue against 1

the application of additional safety equipment or procedures, such as excluding CDAs from the DBA. Prudence dictates, however, that, if the PRA results support the application of additional equipment or procedures, such as including CDAs within the DBA envelope, then one should be extremely apprehensive about

rejecting the results in favor of higher public health risks. In other words, prudence dictates an asymmetry in the application of PRA.

Staff is taking just the reverse approach to public safety in the case of the CRBR. Staff has applied their PRA results in the LWA-1 proceeding to " demonstrate" that CRBR risks are comparable to LWR risks and to eliminate alternative sites, but apparently Staff does not want to apply the PRA results to test whether the CDA should be included in the DBA envelope, realizing that to do so would force a safer design or a rejection of the CRBR site.

PRA should be used ae a check on the " judgmental" approach taken by Staff. In this case, Staff's conclusions have been checked and found not to wash.

III. CRBR Site Suitability The Board resolved Intervenors' Contention 2 for purposes of the LWA-1 by finding that The containment / confinement design of the CRBRP has been shown capable of performing its intended function to accommodate all credible deaign base threats and hold doses to the public below guideline values, without requiring any technological innovations. . . .

The Staff's final position on the adequacy of the containment / confinement design will be presented when its SER is published. . . . The Board is not persuaded by the evidence of record to date . . . that the CRBR will be built and operated in a manner that precludes the necessity for considering CDAs within the design basis. . . . [W]e foresee a heavy burden upon these parties at the construction permit phase of evidentiary hearings to

w ni provide sufficient evidence to permit a resolution of these questions.

ASLB Partial Initial Decision (Limited Work Authorization),

February 28, 1983, at p. 22. At the CP stage, the Board must resolve these open issues and make a finding, based on reasonable assurance, that the proposed facility can be constructed and operated without undue risk to the health and safety of the public, taking into consideration 10 CFR Part 100. 10 CFR

{50.35(a). I therefore offer the following new information for further consideration by the Board of one of the issues raised finder Contention 2.

In the LWA-1 proceeding, Intervenors argued that, in assessing the suitability of the CRBR site, the effects of the containment vent / purge system on offsite doses must be considered. Had the effects of the vent / purge system been incorporated into Staff's and Applicants' calculation of offsite doses in the site suitability analysis, Intervenors demonstrated that the bone surface doses would have exceeded the 10 CFR 100 dose guideline values.

Through discovery in separate litigation, it has come to my attention that there is a precedent for incorporating the effect of the containment vent system in 10 CFR 100 analyses. This precedent, which lends further weight to the arguments made by Intervenors in the LWA-1 proceeding, is found in analyses DOE has performed for the airborne activity confinement systems used by DOE production reactors at the Savannah River Plant (SRP).

l There are both differences and similarities between the confinement system of SRP production reactors and the containment / confinement system of CRBR. I will not elaborate on the containment / confinement, annulus filtration, and vent / purge systems of CRBR, since these systems are known to the Board. The SRP reactors are somewhat different in that they do not have a dual containment / confinement building and consequently do not have an annulus filtration system. The SRP reactors do not have a containment building capable of withstanding 10 psi (design pressure); rather the reactor is housed in a reactor building that can be sealed to withstand a 2 psi differential. The two systems are similar in that they both utilize a filtered vent system to mitigate offsite doses in the event of CDAs (core melting). Both filter systems filter halogens and particulates, but are ineffective with regard to noble gases. The SRP reactor Airborne Activity Confinement Systems are described further in J.A. Smith, et al., " Safety Analysis of Savannah River Production Reactor Operation," DuPont Savannah River Laboratory, DPSTSA-100-1, Rev. 12/76, pp. IV-43 to IV-49 (Attachment 3); and Memorandum from G.F. Merz to S.P. Tinnes, " Airborne Activity Confinement System Base Case Design Basis Accident," July 19, 1979 l ( Attachment 4) .

As can be seen from the Merz Memorandum ( Attachment 4), DOE has selected as a DBA for the Airborne Confinement System a fission- product release consistent with the 10 CFR 100 site suitability source term for LWRs, namely, a full core meltdown

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

i

! with release of 1004 noble gases and 50% of the halogens to the reactor building.,2/

[ Since DOE reactors are not licensed by the Commission, there l

is no requirement that they meet the requirements of 10 CFR 100, and in fact, as evidenced by the Safety Analysis Report (DPSTSA-t 100-1, Rev. 12/76), they do not; but,that is an issue for another proceeding. Nevertheless, it is clear from several SRP documents

made available to me, including DPSTSA-100-1 and the Merz Memorandum, that DOE conducts 10 CFR 100 analyses to assess the l

adequacy of the SRP reactor confinement system and alternative containment concepts. In each offsite dose analysis of the I production reactor airborne confinement system, including the I design basis accident based on use of the 10 CFR 100 site suitability source term, the effect of the filtered vent system is treated in the offsite dose calculation.

It would be interesting to know, and the Board might wish to determine, whether a second precedent for inclusion of the filtered vent system in the 10 CFR 100 dose calculations is found

, 2/ It is perhaps worth noting that for SRP production reactors

, the DBA for the emergency core cooling system is dif ferent from i the DBA for the Airborne Activity Confinement System. For purpose of analysis of the operation of the emergency core l cooling system for SRP reactors, the DBA is a double-ended pipe break in'one of the six primary lines supplying heavy water to the reactor plenum with the simultaneous failure of a single active component, a second emergency cooling water addition system. Under these conditions core damage is limited to one percent. (See J.W. Joseph and R.C. Thornberry, " Analysis of the Savannah River Reactor Emergency Core Cooling System," DuPont

Savannah River Laboratory, DPST-70-463, October 1970).

l t

l in the site suitability analysis of the Ft. St. Vrain reactor, which, I am told, uses a filtered vent / confinement approach.

In the LWA-1 proceeding, Intervenors indicated that the record was inadequate to determine the effect of including the vent / purge system in the 10 CFR 100 CRBR site suitability source term analysis on organ doses such as bone surface, lung, thyroid, and liver. With Staff publication of its " realistic" CDA dose results in the SER Suppl. No. 2, additional evidence can now be provided relative to the effect of inclusion of the vent / purge system on the SSST thyroid and bone surface dose. I will examine the effect on the thyroid dose first, followed by the bone surface.

In the 1982 SSR, the LPZ thyroid dose was estimated to be 7 rem, with no consideration given to the effect of containment venting, but with other parameters conservatively chosen, including the following 95% X/O values:

95% X/O (sec/m3 )

0-8 hours 1.2 x 10-4 8-24 hours 8.4 x 10-5 24-96 hours 3.9 x 10-5 l 96-720 hours 1.4 x 10-5 l

l (Staff Exhibit 1, p. III-ll.)

l In the SER Suppl. No. 2, the adult LPZ thyroid dose was estimated to be 192 rem, with the vent / purge system modeled, and with other parameters " realistically (albeit conservatively)" chosen, including the following 50% X/O valuese

. - - - - , , e - -- - - - - ,- , , - - r,-,

50% X/O (sec/m3 )

0-8 hours 1.1 x 10-5 8-24 hours 1.0 x 10-5 .

24-96 hours 8.0 x 10-6 96-720 hours 5.7 x 10-6 (SER Suppl. No. 2, p. A.1-3.)

The ratios of the 954 X/Q values (used in the SSST analysis) to the 50% X/Q values (used in the " realistic" analysis) are:

I !O Ratio '50% X/Q 0-8 hours 11 8-24 hours 8.4 24-96 hours 4.9 96-720 hours 2.5 Staff's computer modeling output indicates that 153 rem of the 192 rem total LPZ thyroid dose occurs between 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> and 96 hours0.00111 days <br />0.0267 hours <br />1.587302e-4 weeks <br />3.6528e-5 months <br />, where 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> is the time venting commences; the additional 39 ren occurs between 96 hours0.00111 days <br />0.0267 hours <br />1.587302e-4 weeks <br />3.6528e-5 months <br /> and 130 hours0.0015 days <br />0.0361 hours <br />2.149471e-4 weeks <br />4.9465e-5 months <br />, where 130 hours0.0015 days <br />0.0361 hours <br />2.149471e-4 weeks <br />4.9465e-5 months <br /> is the time of sodium pool dryout. (NRC Staff's Response to Intervenors' Third Set of Construction Permit Interrogatories and Request to Produce to Staff (May 20, 1983);

Staf f's Computer Run for Halogens, Noble Gases, and Sodium, dated 3/19/83.) Thus, if the " realistic" assumptions were selected but 954 X/Q values were used instead of 50% X/Q values, the adult thyroid dose would be determined by multiplying the thyroid dose

i

i for each of the two time periods of interest (24-96 hours and 96-t 720 hours0.00833 days <br />0.2 hours <br />0.00119 weeks <br />2.7396e-4 months <br />) by the 95t/50% X/O ratio for that time period, and

{

adding the two doses together. The result would be 153 x 4.9 + 39 x 2.5 = 850 rem,

over five tin.es the 10 CFR 100 CP guideline value for thyroid.

1 The infant thyroid dose would be 1700 rem, or over 11 times the

guideline values.

Turning now to the bone surface dose, in the 1982 SSR the

] Staff estimated the LPZ bone dose at 9 rem, with no consideration given to the effect of containment venting, but with other parameters conservatively chosen. In the SER Suppl. No. 2 (p.

A.5-4), the bone dose was estimated to be O rem, 'with the vent / purge system modeled and other parameters " realistically 1

(albeit conservatively)" chosen.

I j Staff's computer modeling output indicates that 5.7 rem of i -

! the 9 rem total LPZ bone dose occurs between 24 and 96 hours0.00111 days <br />0.0267 hours <br />1.587302e-4 weeks <br />3.6528e-5 months <br />, j with the additional 1.9 rem between 96 and 130 hours0.0015 days <br />0.0361 hours <br />2.149471e-4 weeks <br />4.9465e-5 months <br /> (Staff l Computer Run for Solids Only, dated 3/11/P3). Thus, if the

" realistic" assumption were used, but 954 X/Q values were used instead of 50% X/Q values, the LPZ bone dose would be calculated using the same technique as the thyroid dose, thus yielding (5.7)(4.9) + (1.9)(2.5) = 33 rem.

l

! The bone surface dose is three times this value, or 100 rem.

In their " realistic" CDA analysis, Staff assumed 0.16% of the plutonium is initially available to the sodium pool in the reactor cavity, whereas in the 10 CFR 100 site suitability

I analysis Staff makes the more conservative assumption that 1% of the plutonium is available to the containment building. If the Plutonium available in the pool is increased from 0.16% to 1%,

the bone surface dose is increased from 100 rem to 625 rem, well above the 10 CFR 100 bone surface guideline value of 150 rem used by Staff at the CP (1982 SSR, Staff Exh'. 1, p. III-9).

In the LWA-1 proceeding, Intervenors noted that other corrective factors should be applied as well, including:

Factor Basis 4.3 to correct for potential use of plutonium from high burnup spent fuel 1.53,/ to convert from a 50-year dose commitment to an 80-year dose commitment Staf f and Applicants argue that because in their judgment the CDA is not a DBA they are free to ignore the vent / purge system in the 10 CFR 100 site suitability analysis, since no

" credible" accident would ever challenge the containment and require activation of the vent system. If the CDA is a DBA, then of course this argument has no merit, and the CRBR site is not suitable for the CRBR containment design.

Even if the Board concludes the CDA is outside the DBA envelope, we believe Staff's and Applicants' argument is still incorrect. In the 20+ year history of 10 CFR 100, it has always been assumed that, for purposes of assessing Whether 10 CFR 100 3j If the Board chooses to follow the EPA, NRC, and DOE precedent of using 70-year rather than 80-year dose commitment, this factor would be 1.35.

requirements are met, one should assume a gross fission product release following full meltdown (cf., TID 14844, p. 10) and the use of substantial conservativisms in the analytical methodology for estimating offsite doses. Staff's and Applicants' approach

-- to ignore the concomitant effects of the core melting -- is simply ludicrous. Ignoring the implications of fuel melting (i.e., failure to model the vent / purge system), rather than conservatively treating them, results in site suitability source term thyroid and bone surface doses that are some two orders of magnitude less than the dose associated with the most benign full core melt event " realistically" calculated.

When the site suitability source term analysis is performed l

properly and 10 CFR 100 requirements are not met, there is simply no basis for granting a CP for this reactor design at the CRBR site.

I t

BEFORE THE U.S. NUCLEAR REGULATORY COMMISSION ATOMIC SAFETY AND LICENSING BOARD

)

In the Matter of )

)

UNITED STATES DEPARTMENT OI' ENERGY )

PROJECT MANAGEMENT CORPORATION )

TENNESSEE VALLEY AUTHORITY )

)

(Clinch River Breeder Reactor Plant) )

)

AFFIDAVIT OF DR. THOMAS B. COCHRAN i City of Washington )

( ) ss:

District of Columbia )

DR. THOMAS B. COCHRAN hereby deposes and says:

The foregoing limited appearance statement prepared by me and dated July 8, 1983, is true and correct to the best of my knowledge and belief.

/

2 :-2 T_ -

Dr. Thomas B. Cochran Signed and sworn to before me this 8th day of July 1983.

l M N Notiary Public My C: nmiskn Expires July 31,1937

D-a CERTIFICATE OF SERVICE I hereby certify that copies of LIMITED APPEARANCE STATEMENT OF DR. THOMAS B. COCHRAN REGARDING ISSUES RAISED IN THE CONSTRUCTION PERMIT PROCEEDING and AFFIDAVIT OF DR. THOMAS B.

COCHRAN were served this 8th day of July 1983 by hand

• or by first class mail upon:

• Marshall E. Miller, Esq.

Chairman Atomic Safety & Licensing Board U.S. Nuclear Regulatory Commission 4350 East West Highway, 4th floor Bethesda, MD 20014

• Gustave A. Linenberger Atomic Safety & Licensing Board U.S. Nuclear Regulatory Commission 4350 East West Highway, 4th floor Bethesda, MD 20014

• Sherwin E. Turk, E sq.

Stuart Treby, Esq.

Geary S. Mizuno, E sq .

Elaine I. Chan, Esq.

Office of Executive Legal Director U.S. Nuclear Regulatory Commission Maryland National Bank Building 7735 Old Georgetown Road -

Bethesda, MD 20014 Atomic Safety and-Licensing Appeal Board U.S. Nuclear Regulatory Commission 1717 H Street, NW, Room 1121 Washington, D.C. 20555

• Atomic Safety & Licensing Board Panel U.S. Nuclear Regulatory Commission 1717 H Street, NW, Room 1121 Washington, D.C. 20555 l

• Indicates hand delivery.

l

d

, certificate of Service - 2

• Docketing & Service Section Office of the Secretary U.S. Nuclear Regulatory Commission l 1717 H_ Street, NW, Room 1121

Washington, D.C. 20555 (3 copies)

• R. Tenney Johnson, Esq.

Leon Silverstrom, Esq.

Warren E. Bergholz, Jr., Esq.  ;

William D. Luck, Esq. l Office of General Counsel U.S. Department of Energy 1000 Independence Ave., SW, Rs. 6A245 Washington, D.C. 20585

• George L. Edgar, Esq.

Irvin N. Shapell, Esq.

Thomas A. Schmutz, Esq.

Gregg A. Day, Esq.

Frank K. Peterson, Esq.

Morgan, Lewis & Bockius 1800 M Street, NW, 7th Floor Washington, D.C. 20036 Dr. Cadet H. Hand, Jr., Director Bodega Marine Laboratory University of California 1 P.O. Box 247 West Side Road Bodega Bay, CA 94923 .

(Federal Express Mail)

Herbert S. Sanger, Jr., Esq.

Lewis E. Wallace, Esq.

James F. Burger, Esq.

W. Walker LaRoche, Esq.

Edward J. Vigluicci, Esq.

Office of the General Counsel Tennessee Valley Authority 400 West Summit Hill Drive Knoxville, TN 37902 William M. Leech, Jr., Esq.,

Attorney General William B. Hubbard, Esq.,

Chief Deputy Attorney Gen 6ral Michael D. Pearigen, Esq.

State of Tennessee Office of the Attorney General 450 James Robertson Parkway Nashville, TN 37219

--. _ , _ _ - . , _ _ _ _ __ . .. _-. _ ___..__._.~-...___,___.. _ _ __. ..__,

- Certificate of Service - 3 Lawson McGhee Public Library 500 West Church Street Knoxville, TN 37219 William E. Lantrip, Esq.

City Attorney Municipal Building P.O. Box 1 Oak Ridge, TN 37830 Oak Ridge Public Library Civic Center Oak Ridge, TN 37830 Joe H. Walker 401 Roane Street Harriman, TN 37748 Commissioner James Cotham

, Tennessee Department of Economic and Community Development Andrew Jackson Building, Suite 1007 Nashville, TN 32219 J% tg,i %ln -

B'arbara A. Finamore l

I m _ _ . . . -, . - - . _m , . . - , .... .,_.__._, _ ...__. , _ , ,_..,,,.c_,,_, , . . . _ . ,

ATT%CHHstJhi-Table 1. Summary of thyroid abnormalities in the Marshallese, 1981. l i

(Corrected = matched control value subtracted.)* f Total Total nodules Carcinoma Hypo func t ion ** 1esions Group Es t. thy. Correc- Correc- Correc- Corree-age 1954 No. dose, rad No.  % ted % No. % ted % No.  % ted % ted %

Rongelap

<10 22 810-1150t 17 77.3 74.7 1 4.5 3.6 3 13.6 13.6 88.0 10-18 12 335- 810 3 25.0 17.4 1 8.3 8.3 17.4

> 18 33 335 3 9.1 1.2 2 6.1 5.4 4 11.8 11.9 13.2

{

Total 67 556 av. 23 34.3 28.5 4 6.0 5.2 7 10.4 10.1 38.7 Ailingnae

<10 7 275-450 2 28.6 26.0 26.0 10-18 1 190

>18 11 135 4 36.4 28.5 9.0 8.8 1 37.3 Total 19 217 av. 6 31.6 35.8 5.2 1 5.0 30.8 l Rong. + Ail.

<10 29 275-1150 19 65.5 62.8 1 3.4 2.5 3 10.3 10.3 73.1 10-18 13 190- 810 3 23.1 15.5 1 7.7 7.7 15.5

>18 44 135- 335 7 15.9 8.0 2 4.5 3.8 5 11.4 11.1 19.1 Total 86 482 av. 29 33.7 27.9 4 4.7 4.0 8 9.3 9.0 36.9 Utirik

<10 64 60-95 5 4.7 5.2 1 1.6 0.7 5.2 10-18 21 30-60 3 14.2 6.7 1 4.8 4.8 6.7

>18 79 30 9 11.4 3.5 1.3 0.6 1 3.5 Total 164 51 av. 17 9.1 4.6 3 1.8 1.1 4.6 Matched controls, unexposed

<10 229 6 2.6 2 0.9 2.6tt 10-18 79 6 7.6 1.3 1.3

>18 292 1 1 8.9 23 7.9 2 0.7 1 0.3 8.2 Total 600 35 5.8 5 0.8 0.3 2 6.2

• The cancer estimates are'possibly underestimated since all unoperated nodules were considered benign for these calculations. Occult carcinomas were not

' included under carcinoma.

• • No nodule cases with hypofunction are included. (See Tables 1 and 2, Appendix IV.)

tThe lower dose estimates for the <10 group were used for in utero esses.

ttNo correction necessary. ""~

1 .

l J.i EGG-EA-6:52 j . . _ . - .

l JA.' LUAU 1983

'TLINCH RIVER BREEDER REACTOR PLANT PROBABILISTIC' RISK ASSESSMENT - PHASE I ATTACHMENT ~ R MAIN REPORT l

e -

' #' EG&G Idaho, Inc. l Wood-Leaver and Associates, Inc. Fauske and Associates, Inc.  ;

~

~

.; S. Z. Bru F W. A. Brinsfield H. K. Fauske

' (, ~

J. ,J. Einerson M. F. Hinton B. A. Brogan S. E. Mays R. E. Henry J. T. Madell i D. P. Mackowiak J. E. Trainer .

J. 4. Stoffel J. L. vonHerrmann  :

M. S. Tawfik P. J. Wood D. A. Weber l R. E. Wright l Idaho National Engineering Laboratory Operated by the U.S. Department of Energy

. = = .

' h43 &?fl? ' ' ~ ,

L,

. A.,I Y. -

v. -: J' .

• W .,

"f, d' M' _

4 ammet mwout ammung ammmme m :4,}g- [

.- it _z. , ._ _ __, n.~ _ mMsWWEEmr '

IMDbE '"

& ,~.#

4'3 ~~ - - :-Q&~k .

. &r?.W4%:L_ a. ;31h~*

• _, 's; _ - ~ - _

~_L . .. 37 T "sainems---

_ [-jM I. -

- [ .- ,

, 1 l

l l

h7 b4$s 11 - - -- Em 1.

This is an informal report intended for use as a preliminary or working document Prepared for the l U.S. DEPARTMENT OF ENERGY Idaho Operations Office Under DOE Contract No. DE-AC07-76ID01570 l'l l g 6 6 E 6 1daho

.. . . . .. =-

- - -----a , - - - - ,._, ,,_,--

,.e ,, ,-w,,m, .p_--._w,, .,.m ,_y  %-y,97,_w,,.f._

EGG-EA-6162 CLINCH RIVER BREEDER REACTOR PLANT PROBABILISTIC RISK ASSESSMENT--PHASE I MAIN REPORT EG&G IDAHO, INC.

S. Z. Bruske J. J. Einerson M. F. Hinton D. P. Mackowiak

, J. W. Stoffel M. S. Tawfik D. A. Weber R. E. Wright WOOD-LEAVER AND ASSOCIATES, INC.

W. A. Brinsfield B. A. Brogan .

5. E. Mays

• J. E. Trainer J. L. vonHerrmann P. J. Wood FAUSKE & ASSOCIATES, INC.

H: K. Fauske-R. E. Henry

, s,,

J. 7. Madell

, ~.

g Published January 1983 EG&G Idaho, Inc. l Idaho Falls, Idaho 83415 Prepared for the U.S. Department of Energy Idaho Operations Office Under DOE Contract No. DE-AC07-76ID01570

l: ..

ABSTRACT

~

The overall objective of a probabilistic risk assessment (PRA) is to provide a realistic evaluation of the risk to the public from a nuclear power plant. Of the many steps in a PRA, some of the major steps are:

identification of accident initiators, the development of logic models that describe the way a plant responds to the accident initiators, the l

determination of the probability of systems failing to terminate the-l accident sequences, identification of dominant accident sequences which l have the potential to produce core damage, analysis of core and containment response to the accident sequences, and evaluation of radionuclide release quantities and of the risk to the gen.eral public.

The Clinch River Breeder Reactor Plant (CRBRP) PRA is being performed in two phases. Phase I of the CRBRP PRA has completed on a preliminary basis the identification of dominant accident sequences and the analysis of core and containment responses. Phase II will refine the analysis of the dominant accident sequences in order to evaluate the radionuclide release quantities and the risk to the general public from the CRBRP.

This report is a summary of the work performed.during Phase I. It is not a complete PRA and the results presented in this report are not final.

Significant conclusions and use of portions of this report should be avoided until completion of the CRBRP PRA. This report should only be used as a starting point for further refinement and investigation during Phase II.

• e 11 l

l

,0 FOREWORD This report is a summary of the work performed during Phase I of the CRBRP PRA. This study is being sponsored by the Department of Energy per c

CRBRP Project office request.

EG&G Idaho Inc. was contracted to perform the first phase of this risk assessment. Wood, leaver and Associates, Inc. (WLA) along with Fauske and Associates Inc. (FAI) were subcontracted for portions of the Phase I work.

EG&G Developed and quantified the CRBRP system fault trees. W1.A performed the accident sequence delineation and quantification. FAI provided best estimate analysis of plant operating characteristics and core and containment phenomenology.

This report is contained in eight volumes: a main report and seven appendices. The main report discusses the major areas of work performed during Phase I. It also discusses the study methods and their

~

limitations. The appendices contain detailed information and calculations relating to initiating events, accident sequence delineation, systems

- analyses, core damage sequence quantification, containment systems analyses and phenomenological analyses. .

Acronyms used in this report are defined in the list ir ediately following the table of contents. -

e 0

111

1

.- 1

. l ACKNOWLEDGMENTS

- The authors wish to express their thanks to several individuals and organizations who have made important contributions to this study report:

theCRBRPProjectOfficepersonnelandWestinghouse,BurnsandRoe, Atomics International and General Electric Engineering personnel for their review -

and technical comments as the study progressed; Neldon Marshall, Rob Fitch and Larry Fackrell of EG&G for their computer engineering assistance; Dale Evans, Wes Rhudy and Guy Robinson of EG&G for their assistance on the Computer Aided Design system; and T. Marcintak, M. Epstein, M. Grolmes, A. Sharon, G. Hauser, M. K(nton, J. Kabor,.R. MacDonald, K. Cady, D. Kuhaneck of FAI for engineering assistance.

m

• r*

a IV

CONTENTS i 1

11 ABSTRACT ..............................................................

iii FO R EWO RD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

iv ACKNOWLEDGMENTS .......................................................

ix NOMENCLATURE ..........................................................

1. INTRODUCTION ..................................................... 1-1 1.1 Overall Structure of Phase I of the CRBRP Probabilistic Risk Assessment ............................................ 1-2 1.2 Integ ra ti o n o f P RA El ement s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-2 1.3 The Nature of the Results from Phase I of the CRBRP PRA .... 1-7 1.4 Fernali sm of Presentation of the Results . . . . . . . . . . . . . . . . . . . 1-7

2.

SUMMARY

O F R ESU LTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1 2.1 Core Damage Frequency ...................................... 2-1 2.2 Eng i n ee ri n g I n si g hts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-3

3. PLANT DESIGN ..................................................... 3-1 3.1 Accident Mitigation Functions .............................. 3-1 3.2 Front-Li ne and Support Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-4

4. I N ITI ATI NG EV ENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-1 4.1 Initiating Event Identification ............................ 4-1 4.2 Initiating Event Frequencies .............................. 4-16 i,

4.3 Se f e ren -c e s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-20 .

5. ACCIDENTESEQUENCEDELINEATION..................................... 5-1

. 5.1 Event Tree Methodology ..................................... 5-1 1

5.2 Functional Event Tree ...................................... 5-1 5.3 SysltemicEventTreeModels................................. 5-4 i1 l 6. SY ST EM AkA LYS I S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-1 I

6.1 Major Heat Removal Systems Descripti on . . . . . . . . . . . . . . . . . . . . . 6-3 i

V

.- .e 6.2 Support Systems Description ............................... 6-127 6.3 Other Systems and Action s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-246 6.4 Data Description .......................................... 6-286 6.5 References ................................................ 6-301

7. D EP END ENCY ANA LY S I S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-1 7.1 Functional Dependencies .................................... 7-1 7.2 Fault Tree Dependencies .................................... 7-2 -

7.3 Phenomenological Dependencies .............................. 7-5 j 7.4 Phase II Requirements ...................................... 7-5 7.5 References ................................................. 7-6

8. CORE DAMAGE SEQUENCE QUANTIFICATION .............................. 8-1 8.1 Sequence Quantification Method ............................. 8-1

~

8.2 Dominant Core Damage Sequences ............................. 8-9

9. CONTAINMENT ANALYSIS ............................................. 9-1 9.1 Containment Event Tree Model .................'.............. 9-1 -

9.2 Containment Systems Faul t Analysi s . . . . . . . . . . . . . . . . . . . . . . . . . 9-6 9.3 Containment Sequence Quanti fication . . . . . . . . . . . . . . . . . . . . . . . 9-34

~9.4 Pl ant S tate De fi ni ti on . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-37 9.5 Analysi s of Radionuclide Rel ea ses . . . . . . . . . . . . . . . . . . . . . . . . . 9-38 9.6 References ................................................ 9-41 i

l

10. PH ENOMEN0 LOGICAL ANA LYS ES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10- 1 10.1 Introductory Remarks on Core Response Event Trees with Emphasis on Energettes Phenomenology and Quantification ... 10-1 10.2 Core Response Event Trees ................................. 10-2 l 10.3 Potential Mechanical Energy Release Mechanisms and Requirements to Cause Reactor Vessel Damage ............... 10-7 10.4 Energetics Assessment ..................................... 10-10 vi 1

l

. . , , - - - _ , ~_. , . - , + - - - - - . - - - - - ---wm v~'---- - -

- - -**',-'v-~-

c. ..

10.5 Implications of the Heterogeneous Core Design in Relation to Energetics Potential . . . . . . . . . . . . . . . . . . . . . . . . . . 10-26 10.6 Introductory Remarks on Containment Phenomenology .........10-28 10.7 Primary System Failure .................................... 10-30 10.8 Sodium Accumulation ....................................... 10-39 10.9 Reactor Cav i ty Li ner Fail ure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-41 10.10 Sedi um Re fl ux . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-56 10.11 Ultimate Heat Removal ..................................... 10-61 10.12 Model for Containment Transi ent Analysi s . . . . . . . . . . . . . . . . . . 10-62 10.13 Summary of Containment Phenomenology Study ................ 10-66 10.14 References ................................................ 10-69

11. R ES U LTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 - 1 11.1 Dominant Core Damage Sequence s . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-1 11.2 Engineering Insights ...................................... 11-3 11.3 Sensitivity Analysis of CRBRP Core Damage Frequency Results................................................... 11-8

12. P HAS E II TASK REQU IREMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-1 12.1 Accident Sequence Definition Revi ew . . . . ~ . . . . . . . . . . . . . . . . . . 12-1 12.2 Model Refinement .......................................... 12-2 12.3 Radi onucl ide Rel ease Analysi s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-7 12.4 Heal th Consequence Analysi s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-14 12.5 Ri s k An aly si s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-15 12.6' Ex-core Rel ea se As se s sment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-15

! 12.7 Uncertainty Analysis ...................................... 12-16 w

12.8 Common Cause Fail ure Analysi s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-16 12.9 References ................................................ 12-20 l

i

' v11 t

.._ _ . . . . , . _ , , , , _ . . _ _ , _ _ ,_ . -..-.r - _ . . . . ,y... . . . , , . _ _ . .,., _ -..,__. ,. .. _. , ,,.. , . _ . -.- _ ..-.

l I

l APPENDICES (Each appendix is published as a separate volu:ne)

APPENDIX A--INITIATING EVENTS ......................................... A-1 APPENDIX B--EVENT TREES ............................................... B-1 APPENDIX C-FAULT TREES AND FAULT SUMMARIES . . . . . . . . . . . . . . . . . . . . . . . . . . . C-1 APPENDIX 0--EVENT TREE QUANTIFICATION ................................. 0-1 APPENDIX E--CONTAINMENT FAULT TREES AND FAULT SUMMARIES . . . . . . . . . . . . . . . ' E-1 APPENDIX F--SUPPORTING INFORMATION FOR CORE AND CONTAINMENT PHENOMENOLOGY STUDIES AND SUCCESS CRITERION FOR PROTECTED FLOW COAST 00WN EVENTS ............................................ F-1 APPENDIX G--SENSITIVITY ANALYSIS ...................................... G-1 Vili

~.. .

b .

NOMENCLATURE ABHX r Blast Heat Exchanger ACS nulus Cooling System AFW uxiliary Feedwater i ALMS Auxilf ary Liquid Metal System BOEC Beginning of Equilibrium Core i BOP Balance of Plant BSC Backup Sodium Cooler

C/B Circuit Breaker i CAPS Cell Atmosphere Processing System CCDF Complementary Cumulative Distribution Function

• CCF Common Cause Failure CCS Containment Cleanup System CIS Containment Isolation System CLCV Cold Leg Check Valve CPS Containment. Purge System CRBRP Clinch River Breeder Reactor Plant '

CR0 Control Rod Drive DG8 - Diesel Generator Building DHRS Otract Heat Removal Service dP Otfferential Pressurt OP 01stribution Panel OST Demerating Storage Tank ECHW Emergency Chilled Water ECP Engineering Change Procedure ECT Emergency Cooling Towers ECW Emergency Chilled Water EDG Emergency Diesel Generator EPSW Emergency Plant Service Water EVS Ex-Vessel Storage '

EVST Ex-Vessel Storage Tank FAI Fauske and Associates FM Flow Meter 1

< HCDA Hypothetical Core Disruptive Accident HEDL Hanford Engineering Development Laboratory HVAC Heating Ventilation and Air Conditioning I&C Instrumentation and Control IA' Instrument Air -

IE Initiating Event IHTS Intermediate Heat Transport System IHTS FC' Intermediate Heat Transport System Forced Circulation IHTS NC Intermediate Heat Transport System Natural Circulation IHX Intermediate Heat Exchanger ISI Inservice Inspection LCO Limiting Conditions for Operations LMFBR Liquid Metal Fast Breeder Reactor ,

LOSP Loss of Offsite Power LP Low Pressure LTHS Long Term Heat Sink MCC Motor Control Center MDAFW Motor Driven Auxiliary Feedwater MFW Main Feedwater ix i -- - - - , - - . _ . _ - . _ _ _ , . , _ __ _ , _ , _ , , _ , _ , _ _ _ ,,

i NCHW Normal Chilled Water

. NCW Normal Chilled Water i NDHX ' Natural Draft Heat Exchanged '

NI Nuclear Island ,

NPSH Net Positive Suction Head i NPSW Normal Plant Service Water l

^

NREP National Reliability Evaluation Program NSSS Nuclear Steam Supply System 0FHX Overflow Heat Exchanger OHX Overflow Heat Exchanger

. OSIS Outlet Steam Isolation Subsystem OSP Offsite Power PACC Protected Air Cooled Condenser -

PCRDM Primary Control Rod Drive Mechanism j PHT Primary Heat Transport 2

PHTS Primary Heat Transport System

. PHTS FC Primary Heat Transport System Forced Circulation PHTS NC Primary Heat Transport System Natural Circulation PIC Pressure Indicator Controller PM Permanent Magnet PPS Plant Protection System PRA Probabilistic Risk Assessment PRSS Primary Reactor Shutdcwn System PSAR Preliminary Safety Analysis Report PSP Primary Sodium Pump PSW Plant Service Water i

PWST Prctected Water Storage Tank RAPS Radioactive Argon Processing System

RAT Reserve Auxiliary Transformer ,

RC Reactor Cavity RCS Reactor Containment Building RGC Recirculation Gas Cooling

RPS Reactor Protection System RPT Reactor Pump Trip i.

RS3 Reactor Service Building RSS Reactor Shutdown System RTD Resistance Temperature Detector 500 System Design Description SERC Southeastern Electric Reliability Council SGAHRS Steam Generator Auxiliary Heat Removal System SG8 Steam Generator Building

SG8-A8 Steam Generator Auxiliary Butiding SG8-IB Steam Generator Intermediate Bay SG8-SB Steam Generator Bay SGR Switchgear SGS Steam Generator System SGTR Steam Generator Tube Rupture SHR$ Shutdown Heat Removal Systems

! SRSS Secondary Reactor Shutdown System SSCCW Secondary Services Closed Cooling Water SSPLS Solid State Programmable Logic System SWR Sodium Water Reaction SWRPRS Sodium Water Reaction Pressure Relief Subsystem x .

ar, ,1 4

TBV Turbine Bypass Valve (s)

TDAFW Turbine Driven Auxiliary Feedwater TIC Temperature Indicator Controller TLS Top Logic Structure TMBDB Thermal Margin Beyond Design Basis UAT Unit Auxiliary Transformer USS ' Unit Substation VAC Vital Instrument AC WLA Wood, leaver and Associates -

O 6

m O

s i

l xi t

8. CORE DAMAGE SEQUENCE QUANTIFICATION l

Core damage sequence quantification is the process of combining initiator frequencies with mitigating system conditional failure 4 - probab.ilities in a manner prescribed by the event trees to determine the frequency of any given core damage sequence. From this evaluation, the dominant sequences leading to core damage are identified and grouped for more detailed analysis and further evaluation in the containment analysis.  !

i

- 8.1 Secuence Quantification Method An estimate of the core damage f.equency for Clinch River Breeder Reactor Plant requires quantification of the event trees constructed for this analysis. To minimize the complexity and effort of this process the event trees are divided into three distinct portions: initiators, PHT trees, and heat sinks. These portions can then be quantified and combined to arrive at an estimate of the core damage frequency at the plant. The details of this process are presented in Appendix D. The intent of this l section is to summarize and present the results of Appendix 0, Section D.2.

To quantify the three sets of trees requires that the fault trees be constructed to correspond to each event heading in these trees. It is important to note at this point that the top event for the system fault trees appearing in Section 6 of this report are quite often different from the top events quantified to produce failure probabilities for the event tree headings. This anomaly was due to a Project Office desire to have system level quantification based on the CRBRP system boundary designations- as-wel1 as-providingethe:inputrfor calculating'PRA4 core-damage sequence frequencies. The system level failure probabilities were

calculated to aid system level engineering reviews at the Project Office.

In Section 6, the fault trees were presented in accordance with CRBRP system boundary definitions. However, the PRA event tree headings in many cases do not coincide with CRBRP system boundary definitions. For example the event trees make a decision if PHTS pony motors are available or not available. Since the PHTS/IHTS fault trees as used in Section 6 cover loss of pony motors, piping ruptures and rupture disk failures, only gate BF12, 8-1 S

, . - - - -- - .-- ----.r.. . , _ , , _r,. _n_. ,n .,

1 (loss of pony motors) was used for the event tree heading "PHTS Pony Motor '

Available" and gates BFN12 and CN11 (loop ruptures and rupture disk failures) were merged with the SGS fault tree for quantification of event tree heading " Heat Sink." A table and a detailed explanation which describes the relationship between the event tree headings and the CRBRP system boundary fault trees can be found in Appendix 0, Section 1. The -

1 fault trees for each event heading were then quantified taking into account any dependencies between the event heading and previous event headings.

For example, if two systems require instrument air in order to function and

• i the event tree path of concern shows one of the systems is successful and the other failed, then the failure probability of the failed system should not include contributions from instrument air faults since these ,

contributions would have precluded successful operation of the first system. A summary of the event heading probabilities is presented in j Table 8-1.

Cach sequence in an initiator event tree results in either a transfer to a pHT tree, or in core damage. Each sequence in a PHT tree in turn i

transfers to a heat sink tree. The transfers to PHT or heat sink trees are

! determined by the number of loops available for forced or natural circulation, as well as the availability of the DHRS to remove heat. These decisions may be a function of the initiating event or a result of events (i.e., system failures) which occur subsequent to the initiation of the 1 sequences under consideration. Table 8-2 is a summary of the probabilities of transferring to a heat sink tree from a given PHT tree, and is based upon the event heading probabilities summarized in Table 8-1.

! Table 8-3 is a summary of the sequence probabilities in the heat sink trees.

After quantification of the individual event trees has been accomplished, the overall probability of core damage due to the initiating events considered in this study can be determined. As mentioned .

previously, the details of this process, as well as the results for each of the sequences included in the event trees, is given in Appendix D.

l Table S-4 summarizes the core damage frequencies for those sequences whose frequency of occurrence is calculated to be greater than 10 -7 per year.

This cut off frequency was chosen because an investigation of all of the l

8-2

l Table 8-1

(-

EVENT HEADING PROBABILITIES Probat,ility of Failure Event Heading Used in Event Trees Control Rod Insertion Signal:

Low Dependence 1.5 x 10-7

- High Dependence 7.5 x 10~7 .

-8 Control Rod Insertion 1.5 x 10 4 Guard Vessel Integrity ,

1 x 10-5 Cover Gas Pressure Limitation 3.5 x 10~7 Primary Pump Trip (per pump) 4 x 10-5 SWRPRS 3 x 10-5

i Superheater Isolation:

Break inside Outlet Valve -

Affected Loop 2 x 10"I Unaffected Loops 2 x 10-3 Break outside Outlet Yalve -

Each Loop ,

2 x 10-3 Turbi.no Bypass Valves 2 x 10-3 Main Feedwater 3 x 10~4 Auxiliary Feedwater:

Offsite Power Available - 7.7 x 10-7

. LOSP -

~

Motor Driven AN Train A' 1.1 x 10-2 Motor Driven AFW Train B- 1.1 x 10-2

~

Turbine Driven AFW Ptap 2.9 x 10-2 Diesel Generators (per diesel) 7.5 x 10-2 I

- PHTS Pony Motor (each loop) 2.5 x 10-2 1 Loop DHRS Cooling:

Offsite Power Available 3.9 x 10-2 LOSP 2.1 x 10'I (Continued on next page) 8-3 l

l Table 8-1 (Continued) ,

Probability of Failure Event Heading Used in Event Trees l 2 Loop DHRS Cooling < ,

Offsite Power Available 7 x 10'2 LOSP 2.3 x 10-I Operator Fails to Trip Pony Motor 1 x 10-2 Heat Sink, Short Term (per loop):

With MFW and MCON 5.2 x 10~4 MFW w/o MCON 5.2 x 10-4 With AFW only 1.6 x 10-3 LOSP 1.6 x 10-2 Heat Sink, Long Term (per loop):

With MFW 1.7 x 10-2 With AFW 1.7 x 10-2 LOSP 1.7 x 10-2 G

I 8-4

li l

Table 8-2 l

HEAT SINK TRANSFERS Heat Sink Transferred Into P_robability of Transfer PHT Tree A ~1 1

B 7.5 x 10-2

- C 1.9 x 10-3 D ,

l'.6' x 10-5 E -1 2

F 2.5 x 10-2 G' 5.1 x 10-2 H 6.3 x 10'4 3 8 7.5 x 10-2 C

1.9 x 10-3 0 -1 4 I ~1 J 5.1 x 10-2 X . 2.6 x 10-2 5 Core Damage 0.145 (0.31 for LOS )

F -1 6

G' 5.1 x 10-2 H 6.3 x 10-4 i;

L. 1

. 1-1 - -

l:

2.5-x-10-2 M \l N 2.5 x 10-2 l 0 6.3 x 10-4 !5 1-2 N -1 l'.

0 2.5 x 10-2 0 1 1-3 I

l 8-5

!I

Table 8-3 SIMMARY OF HEAT SINK EVENT TREE QUANTIFICATION Heat Sink Tree w/MF2 and MCON MFW w/o MCON w/AFW only LOSP ,

A 2.1 x 10~7 2.1 x 10~7 2.5 x 10-7 7.7 x 10-6 B 8.3 x 10~0 8.3 x 10-6 9.8 x 10-6 4.7 x 10-5 ,

C 1.8 x 10-4 1.8 x 10-4 1.9 x 10~4 3.7 x 10-4 D 5.3 x 10-6 5.3 x 10-6 6.3 x 10-6 3.6 x 10-5 9

E 1.2 x 10-5 1.2 x 10-5 1.3 x 10-5 ,

F 3.1 x 10~# 3.1 x 10-4 3.5 x 10-4 -

G 4.9 x 10~4 4.9 x 10-4 5.4 x 10-4 -

H 1.0 x 10~2 1.0 x 10-2 1.0 x 10-2 ,

I 6.9 x 10-4 ' 9 x 10 '4 7.7 x 10-4 -

J 1.7 x 10-2 1.7 x 10 -2 1.9 x 10-2 _

K 2.7 x 10-2 2.7 x 10-2 2.9 x 10 -2 ,

L - - -

8xW M - - -

8.6 x 10~

N - - -

1.7 x 10-2 0 - - -

9.9 x 10-2 8-6

'. 5.

Table 8-4 CORE DAMAGE SEQUEMCES WITH FREQUENCY >1 x 10~7 Initiator Event Initiating Event Tree Sequence i Frequency (per year)

General Transients 1 1.2 x 10-5

- 7 7.5 x 10-6 8 1.5 x 10-7 Loss of Main Condenser. 1 1.2 x 10-6 4 7.5 x 10-7 Loss of Main Feedwater 1 8.1 x 10-7 3 4.5 x 10~7 Loss of Offsite Pcuer 1 9.5 x 10-7 5 8.3 x 10-6 9 8.3 x 10-6 11 2.0 x 10-6 12 1.0 x 10-5 13 1.1 x 10'7 14 1.7 x 10'I 15 8.3 x 10-6 17 2.0 x 10-6 18' 1.0 x-10-5.

o

~~

19 1.1 x 10~7 20 1.7 x 10-7 21 1.0 x 10-5 22 ~

1.5 x 10-5 23 4.2 x 10-6 24 1.3 x 10-6 Reactivity Insertion NONE Loss of Normal PHTS Flow 1 9.4 x 10~7 7 1.2 x 10~7 (Continued on next page) 8-7

.- -__ m , ,

l

Table 8-4 (Continued)

Initiator Event Initiating Event Tree Sequence # Orequency (per year)

Steam Generator Tube 1 5.1 x 10-6 Inadvertant SWRPRS 1 5.1 x 10~0 l .

IHTS Leaks- 1 1.5 x 10-6 ,

Reactor Vessel or PHTS Leaks within Guard Vessel 8 6.1 x 10" Elevated PHTS Leak Disables Loop 1 6.8 x 10" Elevated PHTS Leak. No Loop Disabled. NONE

~

Steam Breaks inside SH Outlet Isolation Valve NONE .

Steam Breaks outside SH Outlet Isolation Valve 1 5.1 x 10-6 TOTAL: 1.2 x 10-4 t

9 0

8H5 .

_ - --,----,i & er ,,v-w

-,- - - -~ ,,-e---., .r-- - -- -- - - - - - - - .w.-- -v < --.-- - y

. s; -

. b.

sequences quantified in Appendix 0 reveals that >99 percent of the total l core damage frequency for the plant is contained in those sequences with individual sequence probabilities of 10 -7 per year or greater.

8.2 Dominant Core Damage Secuences E

It is customary in a PRA to list a group of dominant sequences which lead to core damage and/or public risk. Since the public risk analysis is to be done during Phase II, it is appropriate here only to address dominant sequences leading to core damage.

In PRAs to date on LWRs it was typical to find a list of about 10-20 sequences making up the dominant sequences. Each of these sequences represented an initittor and the failure or success of a few systems. The event trees for the CRBRP are much more complicated than those for a typical UWR. This necessitates a slightly different approach to presenting core damage sequences.

As noted in Section 8.1, for each initiator there are hundreds of individual sequences leading to core damage. Section 8.1 also shows how these sequences were grouped into Heat Sink event tree anc PHT event tree failure probabilities and combined with the initiator frequency and system failure probabilities for each path in the initiator event tree.

Appendix 0 shows a detailed example of how this was done. Thus, each

" sequence" quantified in AppendixsD.actually. represents.a. consolidation of several sequences. For purposes of clarity and reviewability, all af these l

l . c6nsolidated sequences are referenced to the-initiator event tree end point

which either results in core damage or transfers to a PHT event tree.

Therefore, when a core damage frequency is referenced to a sequence number on an initiator event tree and that sequence path transfers to a PHT event tree, that frequency represents the frequency of reac'aing the end point on the initiator event tree combined with the total failure probability of all the various paths through the appropriate PHT event tree and its associated Heat Sink event trees.

8-9

l l

i After quantifying all of the paths of the initiator event trees which lead directly to core damage or transfer to a PHT event tree (152 l consolidated sequences), 31 of these consolidated sequences had frequencies

~

-7 per year. Because of the differences in the greater than 1.0 x 10

timing and heat generation between sequences where the reactor shutdown ,

system operated to shutdown the reactor and those where the shutdown system failed, these 31 sequences were divided into 27 protacted sequences

(successful shutdown) and 4 non protected sequences (shutdown not successful). Each o'f tPe 27 protected sequences were then grouped by the ,

initiating event for ease of discussion.

Table 8-5 lists the protected sequences by initiator and the sum of the frequency of the saquences for that initiator which were greater than

' -7 1.0 x 10 par year. Table 8-6 lists the non protected sequances in a similar manner. Table 8-7 shows in abbreviated form, how that frequency arises from the various .nHT and Heat Sink event tree paths. The remainder of this section is a description of the "domicant sequences" as listed in the two tables. Each description section describes briefly the initiating event, its frequency, and its effect on the plant followed by a discussion of the significant sequences (significant meaning having the highest frequency) for each initiating event. The discussion of,the significant i sequences parallels the equations of Table 8-7. To the extent possible, the dominant contributors to various sequences are discussed. These are derived from the importance rankings of the cut sets for the individuai event tree headings. Appendix 0 discusses the details of the fault tree analysis for the various event tree headings.

8.2.1 General Transient Sequences I

General transients are those events whcse occurrence requires a reactor shutdown but does not disable any of the mitigating systems shown .

l on the event trees. The frequency of these events is estimated to be l

10 events in the first year. Figure 8-1 is the initiator event tree for general transients.

l l

8-10 l

l

Teole 8-5 i

PROTECTED SEQUENCES Initiator Core Damage Frequency

. General Transient 1.2 x 10-5

~

Loss of Main Condenser 1.2 x 10-6 Loss of Mai'n Feedwater 8.1 x 10-7 Loss of Offsite Power 8.1 x 10-5 Loss of PHTS Flow 9.4 x 10'I S/G Tube Rupture -

5.1 x 10-6 Inadvertent SWRPRS ,

5.1 x 10 '

IHTS Leaks 1.5 x 10-6' Rx Vessel or PHTS Leaks which are 6.1 x 10-7 inside a Guard Vessel PHTS' Leaks- (Loop.: Di sabled) - 6.8sx.10~7 Steam Break Insida 5.1 x 10-6 1.1 x 10~4 Other initiators produced sequences which were insignificant compared to those

. presented in the table.

8-11

Table 8-6 NON-PROTECTED SEQUENCES

~

Initiator Core Damage Fr equency General Transient 7.7 x 10-6 .

l l

Loss of Main Condenser 7.5 x 10-7 Loss of Feedwater 4.5 x 10-7 Loss of PHTS Flow 1.2 x 10-7 9.0 x 10-6 Other initiators produced sequences which were insignificant compared to those presented in the table.

e S

4 8-12

.e +

i

11. RESULTS The key points of the information from the preceeding sections of this report are summarized in this section. The iterative nature of a PRA ,

dictates that many changes will occur during the course of the analyses.

some of whicn may have a significant impact on the "results" of the previous iteration. This process has already occurred several times during Phase I of this study. The current state of knowledge concerning plant behavior under accident conditions, data sources, level of detail of models, and other analyses yet to be done strongly suggest that many changes will occur due to future iterations during Phase II of this study. i Since such changes have significant potential to elevate, decrease, or completely change the "results" deemed important at this point in the study, extreme care should be taken to avoid using the information contained herein for any purposes other than its sole intended function--to serve as a starting point for further refinement and investigation during Phase II. In this sense, the information in this section can be considered an interim status report rather than a listing of results.

11.1 Dominant Core Damage Secuences Section 8 of this report discussed how core damage sequences were quantified and discussed in some detail those groups of core damage sequences defined as the " dominant" sequences. Listing core damage -

frequencies by initiators produced a list of eleven protected sequences and four non protected sequences. Many of these fifteen sequences are very similar to ene another in terms of the physical response of the plant and .

the major contributors to the core damage frequency. Table 11-1 lists the fifteen " dominant" sequences in five major groups.

Group I includes all those sequences resulting from initiators that leave all three heat sink loops potentially available for heat removal.

This group includes the following initiators: general transients, loss of main condenser, loss of main feedwater, and loss of normal PHTS flow. The sum of the frequencies of the dominant sequences in this group is l

f 11-1 I

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

.~ .

r . - .

-5 per year. The dominant contributors to these sequences were 1.5 x 10 long term heat sink failures due to a loss of integrity of PHTS, IHTS, or SGS piping and components.

Group II includes those sequences where the initiator itself or the direct plant response to the initiator renders one of the heat sink loops inoperable. Included in this category are sequences from steam generator l tube rupture, inadvertent SWRPRS actuation, IHTS leaks, elevated PHTS leaks which disable the loop, and steam breaks inside superheater isolation valves initiators. The comoined frequency of these sequences is

-5 1.7 x 10 per year. As with Group I, the major contributors to these sequences are long term heat sink failures.

I Group III includes only the loss of offsite power initiator sequences. Although a loss of offsite power leaves all three heat sink loops potentially available it is different from Group I sequences because of the differences in how core damage occurs. The major contributors to LOSP sequences involve combinations of two diesel failures followed by heat sink failures (long and short term) or one diesel failurp followed by heat sink failures (long and short term). The frequency of Group III sequences

-5 is 8.1 x 10 per yeer.

Group IV also includes only one inittator, a PHTS leak within a. guard ,

vessel. In this sequence one or more of the main sodium pumps fail to trip following a leak. The resulting loss of sodium inventory causes the hot leg outlet nozzles to uncover, leaving no means of heat removal from the

-7 core to the heat sinks. This group has a frequency of 6.1 x 10 per .

year.

Group V covers those sequences where the reactor shutdown' 1 system fails in response to the initiator. There are four initiators for which these are g ra ra ns o ai co d nser o s of an feedwater, and loss of normal PHTS flow. The frequency of Group V

-6 per year.

l se % ences is 9.0 x 10 i

l I 11-2

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

11.2 Engineering Insichts In a study such as this which focuses on overall plant behavior, as well as individual system details, the analyst tends to develop a different perspective than would be expected of designers, builders, or operators.

This perspective allows the analyst to develop certain engineering insights which were not readily apparent in the body of information typically available about the plant. In general, these insights belong in one of two groups, physical relationships among components and systems, and probabilistic relationships. The former generally deal with functional capabilities and physical interfaces while the latter deal with the relative likelihood of postulated events. This section discusses some of the insights gained during the performance of Phase I of this study.

11.2.1 Physical Insichts i

There are several physical relationships of importance which were discovered in the course of this study. One of the most significant 4

insights has to do with the success criteria for shutdown heat removal. In previous studies the number of natural circulation heat sink loops and the timing when such heat sinks could provide core epoling was significantly different than the success criteria modeled in the event trees of this study. Based on the CRBRP Phase I best estimate analysis of the plant response for protected flow coastdown events, it was discovered that even with no primary pony motors operating, the decay heat power levels in the core were, insufficient to cause. fuel pin dryout.provided that. sodium remains above the core. This information combined with calculations showing that the ~ heat remov,al rate of one heat transport loop in natural circulation with SGAHRS venting and forced draft PACCs was significantly greater than the decay heat generation rate led to the conclusion that a single heat sink loop in natural circulation is adequate for core cooling.

Sensitivity analysis indicate that the previous assumption requiring two loops in natural circulation would result in an increase in the core damage frequency by a factor of six.

I l

11-3

A significant assumption in previous analyses of liquid metal fast ,

breeder reactors has been that any core melt event has a non-trivial ]

probability of generating a highly energetic disruption of the core. These energetic CDAs (cere disruptive accidents) were important because of their ability to fail the reactor vessel and the containment at an early phase of .

the accident when decay heat rates and fission pr,oduct inventories were still very high. Best estimate analysis conducted during Phase I of the core melt prccess from the event tree sequences indicates that the only credible way for the core material to exit the reactor vessel is by thermal failure of the vessel. This type of core material release into the containment is a much slower process resulting in a significantly lower decay heat rate and fission product inventory at the time of vessel failure than energetic CDAs. Preliminary analyses indicate that the timing of the thermal failure is much faster for non protected sequences (about 20 min) i than for protecte'd sequences (several hours).

Circulation of sodium in the PHTS following a shutdown depends on the -

number of primary pony motors that operate. If all three or none of the pony motors work then the flow characteristics between the three loops are the same (assuming equal heat sink performance). However., if one or two pony motors fail, flow in these loops will be greatly reduced due to back flow caused by the loop (s) with an operating pony motor (s). Therefore, the heat sinks in the loops without operating pony meters car.not remove reactor decay heat. If the heat sink (s) in the loop (s) wit

• an ope-sting pony motor (s) fails, then it will be necessary for the operator to intervene to trip all operating pony motors so that adequate natural circulation flow can exist in the leop(s) with an operational heat sink (s). The current design does not allow the operator to trip the pony motors from the control room. Emphasis should be placed in the operator training program on how to recognize the situation when the only loops vi-h crica y pony mctor flow have inoperable heat sinks so snat a: tion can ce taken to trip tne pony motors and establish natural circulation in the loops with operable heat l

sinks. Provisions for tripping the pony motors from the control room woulc facilitate these actions.

11-4 l

Another insight was discovered when analyzing the main feedwater system. This system has a deaerating storage tank (OST) at the suction of the feed pumps. One of the purposes of this tank is to provide a reservoir of water so that the feed pumps can continue to supply water to the steam drums following a shutdown when the main condenser is unavailable.

- However, the feed pumps require gland sealing water from the condensate

~

pumps which are upstream of the OST. Thus, condensate system failures could prevent the feed pumps from transferring the water from the OST to the

- steam drums. Instead, the feedwater pumps could be provided with sealing water from their own pump discharge (cooled by the same source as HVAC for the pumps if necessary) so that operation of the condensate pumps is not necessary for the feed pumps to work in this scenario.

11.7. 2 Probabilistic Insichts These insights relate to findings from the quantification of initiator frequencies, system failure probabilities, and/or sequence frequencies.

They also relate to the findings of various sensitivity analyses where the event tree models are perturbated and the resulting sequence frequencies are compared to the base-line frequencies.

One significant finding is that the probability of failure of the heat sinks in the long term is dominated by ruptures / leaks in the,PHTS, IHTS, and SGS piping and components. It should be noted, however, that there is a relatively high uncertainty associated with the failure rates for ruptures / leaks of sodium-and non-sodium components due.to-a.1ack of experience data on these events. Sensitivity analyses indicate that a

~

decrease in the long, term heat sink failure probability of one order of magnitude decreases the total core damage f equency for protected sequences by a factor of 3.2.

As noted in Section 4, the initiating event frequencies for some events were based on data summaries for the first year of operation. These values tend to be higher than the frequencies for subsequent plant years.

The core damage frequency would be reduced by a factor of 1.7 if data from subsequent plant years was used.

11-5

Loss of offsite power initiator sequences account for approximately 70 percent of the total core damage frequency. As noted in Section 8 of this report, most of the significant LOSP initiator sequences involve the failure of one er more emergency diesel generators. The failure probability for a single diesel generator was 0.075. Sensitivity analyses indicate that a factor of 75 reduction in the failure probability of diesel generators would reduce the overall core damage frequency by a factor-of 2.8. Additionally, no credit was taken in tne event tree model for the probability of recovery of offsite power. Therefore, the contribution from LOSP sequences is conservative in that respect. Sensitivity analyses indicate that the frequency of LOSP sequences can be reduced by a factor of four if offsite power is recovered in four hours and by a factor of about ten if offsite power is recovered in ten hours. These ananlyses assume that the steam generators in-the heat sink loops will not boil dry within the time to recover offsite power or that water can be reintroduced to a dry steam generator without causing failure of the steam generator. This -

assumption will require verification in Phase II.

The SWRPRS presents a classic case of design trade-offs between reliability and safety. For safety purposes the system is designed so that it has a low probability of failure to function when required. However, to the extent that the data for premature opening of the rupture discs is correct, it also has a significant probability of inadvertently operating (0.1 events / year). This frequency would correspond to an expectation of an erroneous SWRPRS actuation every ten years. Considering the effect on the plant caused by such an event (reactor trip, S/G module blowdown, IHTS loop disabled, etc.) this frequency may not be desirable from a' plant

. availability standpoint.

The SWRPRS also presents a classic case of tradeoffs between one aspect of safety and another aspect of safety. Premature opening of the I rupture discs results in icss of one heat sink icop for heat removal purposes. This failure mode accounts for approximately 17 percent any single heat sink's failure probability. Thus, approximately ten percent of the total core damage frequency is due to heat sink failures caused by premature opering of the rupture discs.

11-6

TABLE 11-1. OCMINANT CORE DAMAGE SEQUENCES Group Frecuency_ Acolicable Initiators Major Contributers .

-5 General Trans'ients Te I 1.5 x 10 Long's (rm Hiat Loss of Main Condenser Sink LTHS)

(All three

- loops . Loss of Main Feedwater available) Loss of PHTS Flow

-5 S/G Tube Rupture LTHS II 1.7 x 10 (One loop Inadvertent SWRPRS unavailasle) IHTS Leaks PHTS Leaks (loop disabled)

Steam Break Inside ,

III 8.1 x 10 -5 Loss of offsite power Combinations of 1 or 2 diesel (LOSP) failures followed by heat sink failures (short and long term)

-7 Rx vessel or PHTS leaks Failure to trip IV 6.1 x 10 within guard vessels sodium pumps (PHTS leaks without pump .

trip)

-6 General Transients Failure to signal V 9.0 x 10 (Failure to Loss of Main Condenser and insert control Loss of Main Feedwater rods scram) '

Loss of PHTS Flow

I i 'i i

i I

i I

i 11-7

l l

11.3 Sensitivity Analysis of CRBRp Core Damace Frequency Results Many of the modeling assumptions and event probabilities used in this study are subject to wide uncertainty. The initiator frequencies, component unavailpbilities, and human error probabilities used for fault ,

tree and event tree quantification are point estimates only. In order to explore the impact on core damage frequency of changes in some important assumptions and parameter values, sensitivity studies were performed. The ,

assumption or parameter value was changed and this change was propagated through the fault tree and event tree models. The change in core damage frequency due to the change in assumption or parameter value was noted.

Details of the sensitivity study results are displayed in the tables of Appendix G. In this section, these results are summarized.

11.3.1 Sensitivity to DHRS Unavailability A series of sensitivity analyses were performed to study the effect of changes in DHRS unavailability on core damage frequency. First, two studies were carried out in which DHRS unavailabilty in,all sequences was assumed to be 0.0 and 1.0, respectively. In these two studies, heat sink unavailability values were held at the values used for baseline

~

quantification. Next, three studies were performed in which all heat sink unavailability va;ues were reduced by a factor of ten, and DHRS unavailability was held at its baseline values, changed to 0.0, and 1.0, respectively. Finally, a study was performed in which DHRS unavailability was set at 1.0 for loss of offsite power (LOSP) sequences only, with DHRS unavailability for other sequences and heat sink unavailability set at baseline values. This analysis examines the impact on the core damage frequency from potential dependencies between offsite power and DHRS support systems.

The results of tnese sensitivity analyses are summarized in Table 11-2. The tasie shows that for both the baseline and reduced (baseline /10) heat sink unavailability cases, a change to perfect DHRS availability (unavailaoility equal to 0.0) did not change core damage frequency from the values derived using baseline DHRS unavailabilities.

l'-3

TABLE 11-2. SENSITIVITY OF CORE DAMAGE FREOUENCY TO DHRS UNAVAILABILITY Heat Sink Unavailability Core Damage Frecuency OHRS Unavai'. ability Baseline Baseline 1.2E-4 0.0 Baseline 1.2E-4 1.0 Baseline 2.9E-4 Baseline /10 3.8E-5 Baseline 0.0 Baseline /10 3.8E-5 1.0 Baseline /10 4.0E-5 1.0 (LOSP only) Baseline 1.3E-4 O

e e

a*

=

O l

l l

11-9 l

Using baseline heat sink unavailability values, and the assumption of no DHRS (unavailability equal to 1.0) led to an increase in core damage .

frequency of 132 percent. Using reduced heat sink unavailability values, the "no DHRS" assue.ption led to little change in the core damage frequency

-5 (4.0 x 10 vs. 3.8 x 10-5). The case in which LOSP sequence DHRS ~

unavailability was changed to 1.0 also yielded negligible change in the

~4 core damage frequency (1.3 x 10 vs. baseline value 1.2' x 10~#).

11.3.2 Heat Sink Dependencies To study the effect of possible common cause failures among heat sinks, a sensitivity study was carried out in which a " beta factor" of 0.1 was assigned to multiple heat sink failure sequences. The beta factor is the conditional probability of additional heat sink failures given a single heat sink failure. This change led to a fourfold increase in core damage

~4 A second study was performed using the heat frequency to 5.0 x 10 .

sink beta factor and a " perfect DHRS" (unavailability equal to 0.0)

~4 assumption, which resulted in a core damage frequency of 3.8 x 10 .

11.3.3 Sensitivity to Diesel Generator Unavailability Diesel Generator unavailability was changed from its baseline value of

-2 This change affected LOSP sequences only.

7.5 x 10 -2 to 3.0 x 10 .

As a result the contribution of LOSP sequences to core damage frequency was cut by a factor of four, and core damage frequency was reduced by

-5 A second stt.dy was performed in which diesel 48 percent to 6.4 x 10 .

-3 generator unavailability was reduced to 1 x 10 and DHRS unavailability was reduced to 0.0 for LOSP sequences only. The changes resulted in a

-5 65 percent reduction in core damage frequency to 4.2 x 10 , reducing the contribution of LOSP sequences by a factor of 90.

11.3.4 Sensitivity to Ooerator Failure probability The probability of operator failure to trip pony motors was varied by

~2 a factor of fifty from the baseline value of I x 10 . Reducing this 11-10

- a

~4 resulted in virtually no change in core damage parameter to 2 x 10 frequency (1.1 x 10'#). Increasing this parameter by a factor of fifty caused a factor of six increase from the baseline core damage frequency to

~4 7.4 x 10 .

11.3.5 Sensitivity to pony Motor Requirements for OHRS A series of sensitivity studies was conducted to explore the impact of changes in assumptions about pony motor requirements for successful DHRS operation. Table 11-3 shows the results of three cases that were con s,idered. The assumptions that were varied were the number of pony motors needed for DHRS and whether operator action is required to start the cony motors for OHRS. In those cases where operator action is required, it was assumed that the pony motors would not start automatically. The three cases considered resulted in roughly a twenty to thirty percent decrease in core damage frequency.

11.3.6 Sensitivity to Pony Motor Unavailability Pony motor unavailability was varied by a factor of twenty. A reduction of this parameter to 1.25 x 10 ~3 resulted in a 30 percent drop in core damage frequency (8.6 x 10-5), while an increase of 0.5 resulted in an increase in core damage frequency of a factor of 17 (to 2.1 x 10-3).

To assess the effect of possible dependent failures among pony motors, a beta factor of 0.1 was assigned to multiply pony motor failure 1

l .

sequences. ThisTresulted-in a small' increase in core damage frequency to '

~#

1.5 x 10 .

e 11.3.7 Sensitivity to Initiator Frequencies A series of sensitivity studies was carried out to assess the effect of changes in frequency of initiating events. The results of these studies are shown in Table 11-4. In the derivation of the baseline core damage frequency, the frequencies of transient initiators (initiators 1 through 6) were assigned values that were estimates for the early yeers of plant life. In the first sensitivity study, the frequencies of these initiators 11-11

i- .

1 l

l TABLE 11-3. SENSITIVITY TO DHRS PONY MOTOR REQUIREMENTS Number of Pony Motors Required Oceator Action Recuired? Core Damage Frecuency 0 No 8.4 x 10 -5

. 1 Yes 8.4 x 10 -5

-5 3 Yes 9.7 x 10

~4 Baseline core damage frequency = 1.2 x 10 TABLE 11-4. SENSITIVITY TO INITIATOR FREQUENCIES Initiator (s) Core Damage Frecuency Transient initiators (1-6) (all plant 7.3 x 10 -5 years)

-4

7. Steam generator tube rupture x 50 ,

3.7 x 10

-4

8. Inadvertent SWRPRS actuation x 50 3.7 x 10

~4

9. IHTS leak x 50 2.0 x 10

-4

10. PHTS leak in guard vessel x 50 1.6 x 10

-4

11. PHTS leak--loop disabled x 50 1.6 x 10

12. PHTS leak--loop not disabled x 50 1.3 x 10"#

-4

13. Steam break outside superheater 1.3 x 10 outlet isolation valve x 50

~4

! 14. Steam break inside superheater 3.7 x 10 outlet isolation valve x 50 Baseline CDF = 1.2 x 10'#

I i

l 11-12 I

l were lowered to values that were estimates for all years of plant life (see Section 4.2). As the first entry in Table 11-4 shows, this change results

~

-5 The eight other in a drop of core damage frequency to 7.3 x 10 .

sensitivity analyses involving initiator frequencies consisted of increasing one initiator frequency by a facter of 50, while holding the others at their baseline values. This was carried out for each of the initiators 7 through 14, as the frecuencies of these events are rather more uncertain than the transient initiator frequencies. The results are shown

in Table 11-4. The magnitude of change in core damage frequency due to the change in initiator frequency depends on the relative contribution of the initiator to the baseline core damage frequency. The change in frecuency for initiators with relatively large contributions (SGTR, Inadvertent SWRPRS, Steam Break Inside SH), resulted in a factor of three increase in core damage frequency, while changes in frequency for the other initiators resulted in relatively insignificant changes in core damage frequency.

d e

9 l

i 11-13

p.- SAI-140-79-PA FAILURE MODES AND EFFECTS ANALYSIS 0F CRBRP SHUTDOWN HEAT REMOVAL, SYSTEM

, INTERFACES 4

April 1979

(

\

Submitted to General Electric Company ,

310 DeGuigne Drive Sunnyvale, California 94086 Submitted by Richard Wilson-SCIENCE APPLICATIONS. INC

• 5 PALO ALTO SQUARE, SUITE 200, PALO ALTO, CA 94304 AL8UQUEROUE

• ANN ARBOR

• ARUNGTON

• BOSTON

• CHICAGO + HUNTSVILLE

• LA JOLLA

1. LOS ANGELES

• ROCKVILLE

• SUNNYVALE

• TUCSON ,

e 4

4

- -- - _. ._..e_r. _T _- ~7 m' _ h - -__-P. _%. ___m .r___ _- ----m.- ._._._atw.m__.

. i .

. s. ,

.i.,

. g.

{

t' ATACHHENT 3 G

~~ DPSTSA-100-1 l ,,

~ Rev.12/76 l .

SAFETY' ANALYSIS OF SAVANNAH RIVER

. PRODUCTION REACTOR OPERATION Coordinated by: -

i J. A. Smith

, j

. 1 l

Written by:

.. A. E. Evans L. R. Jones. ' '

. E. Naum .

N.- P. Olson ,

J. A. Smith '

.y. ,

. Issued: September 1972

. . 7Qg. g'::ew-RJefff:fy.' g ,4: n e... u u . : c.

..m.....,n . - .. s-. .:

l-l.f.:u.*_~f4%ys  :;.

e' ..} ;. y g : m ' yyg

-, . .. - .y. . ai. .. isi ._ e

. . ~ . .:. : .~

% :.:. .- 3' .

~- . - . -

% ~~. -..

ay

.G h ....:,

w.p .%"ti i. : .. olE&.

w

, . -WMyi&*Y. 3Y. . .

s . s .

4. .u  :. . -

m.~.. m. .y:u

. . . -- - e..

.. . .  ; v, r m: . . 3 .

am g. ,.,.2

.;.- ..=

by:

A..

-. . . ..n,~.. 'w. ..n.5. e. ,.- . w .e*%'m.s::.

' -n; -

.;  ;. : e- , .<

.: .- k

.c;. - ,

~ J. R. H11Tey, Director ~ ' - '

c - T' '

Nuclear Engineering Division -

• f -

e

.,.w.r..e" r-

• g ,-

s

%. . :.. . . e .;.

., ,.i ; ,., . .. . -J .

.'4 6 e

..,y.'m

..: . ~ , . .. ' ,* w-t. }"'". g + .,:= ' ..

~... .

1. .

...~ .. . 8+- ..'s . ,

-4 m W:,< p;N.1*f l ..g . . :. , 6.

. . . s

(,4. * >

.-.*.r

~- - .

.- . . . . ..47a . .,4*. ',

.1 ., , . .

u , g. . . .s _,;

E. 3. DU PONT DE NEMOURS AND COMPANY:

SAVANNAH RfVER LABORATORY AIKEN. SOUTH CAROUNA 29801 -

Meepeast peut THE U. 5. 88IGenev ====amm Ases GewSLaptAWef Aegeusessfmettoes upseSA COSBTRACT ATIN 311 w

d

--.% y .. --w .-y.., . .,y ryer--.ca-e---*- ".- g.r

1 DPSTSA-100-1 Rev. 12/76 i

4. ENGINEERED SAFETY FEATURES ,

4.1 Activity Confinement System .

He Savannah River ireactors are equipped'with an activity confinement system, (as an alternative to a containment. structure .

or other methods of reducing potential offsite doses)** to re-duce airborne activity releases in the unlikely event of an acci-dent. A 10, and a,schematic detailedisdescription shown in Figure of the system is given in Reference ,

IV-21.

Description of System 4.1.1 ,

In the event of an accident, any airborne fission product release would be expected to occur in the reactor room, with the .

possibility of some release in the heat exchanger bay or pump room. As shown in Figure IV-21, the air from these areas is ex-hausted through a set of confinement filters before release to the stack. During normal operation, the process areas are sealed and maintained at a negative pressure with respect to atmosphere. ,

During shutdown, air from the process areas at -20 ft and -40 ft elevations is routed directly to the stack, bypassing the filters.

This air can be rapidly switched to the online filter system in ef an accident.

}-:

Steam released from the reactor vessel during an accident

ould increase pressure in the reactor room. At pressures greater -

than 12 inches of water, the water seal in the discharge canal between the reactor room, and the disassembly area would be broken. I Typically, the disassembly areas are about twice the size of the Q

l reactor room and are connected to the remainder of the building by unsealed corridors, except below-grade process areas. H e venti-Vl lation syster for the disassembly area is separate from the main process area ventilation system. Air is discharged to the atmos-phere, and no filter or adsorber units are used.

Meals for all other vent paths from the reactor rooms are de S

signed to witns m.J . J.-~..... 1.5 pressure at z. psi. All o m er areas connectea to tne reactor room oy me 4-psi sealsTfF1rtrvuted And interconnecteu. J Although seals between the reactor room and adjacent areas aight be broken during an accident, pses that may escape in the -

l initial pressure surge were usn=ad m a- <--.h - m. '.--_-

size or une reactor room, the rel=H vahr Ione distance of the -

Iccor uum cnese openings, the likelihood that fuel melting will _

not ocus u.a. . hat ae ~iInitial s team <nna _ and the relatively s%Drious wnen positive pressures may avi Aw ^ .-

- m u.r _-

l Ov =. a negauve pressure in the reactor room would be re-established IV-43 l

l .

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

. . . . . __,~.m._.___.__.__..__m___f_'i-_Z_._1__.-____T__'___ -

__C____

l l

s 1

• l W

DPSTSA-100-1

~~ -

Rev. 12/76 (r- ,

1

\

1 e

ewir.co n 2s.soo er.

4,.e a

is,sso cem ni,,,

, , , ,- camperem ne. emeuse rans Supply Fen.

g, i

- 0-ge ,. n a a a ccH E E E H>Ha Mi x C*H 3 E E l--CC-M

- T- - - -

C*H EEEl-CO-

,, , ,, m ,, , , , ,

C*H E R E H>0- g k ze,ooo i efm c@cdL ,

p Profitters ,

4 & & ,

is .eso et L sees XM particuses miters (uomiure s , orators

-zo' a -4c' e4poo efm 5 Areas ,

Normotiy Closed FIGURE IV-21. Flow Diagram of Reactor Ventilation System ,

L j ,,  :- l: w' bc '

• . .. '- .-~ .

i l

IV-44 i

?- 3 e- .+w 'a- T g-+y3r . y- -W -g g-c-+-yu-,g gg wwpg=w.,vmm 3 w g w q g.,9,yyc--- e p- gree- ew y ' -me em,..gw----

l OPSTSA-100-1 Rev. 12/76

• ==idiv by the exhaust fans, because the duration of the pressure -

surge would be only a few secones.

, Airisexh2ustedfromthepdcessareas'usinglargecentri- )

fugal fans driven by electric motors. There are three fans '

in parallel, two of which are normally online, although only one i fan is necessary to maintain negative pressures in the process areas. The rated air flow capability of two fans is 128,000 ces at about six inches of H O 2 ' pressure drop. The fan motors can be pneered from three sources:- '

e Normal building power o Emergency building power '

e Special con +=i - t substation .

In addition, each onlina fan has a backup motor powered by diesel-generators.

Fwhannut filters are provided to remove moisture, particulate .

activity, and halogen activity. The filter banks are enclosed ,

in five separate compartments, three to five of which' are normally l online. A vertical cross section of a filter compartant is  ;

3 shown in Figure IV-22. Each compartment can be isolated for maintenance and testing. The three separate filter banks in order

of air flow treatment are

a e Naseture sepa mtors. The moisture separators (Danisters)"

are designed to remove about 99% of entrained water spherical particles of 1- to 5-me to protect against s'ignificant blinding of the water-repellent particulate filters. The separators consist of mats woven from fefIon 8 yarn (individual fibers about 0.001 inch in diameter) and stainless steel wire forming a pad 24 x 24 x 2 inches thick. The case, mat, and support mechanism are designed and tested to withstand flows equivalent.to about 10 times the rated air flow through the separator.

e Anw*# 4 *= MItm s. The narci ad =+= f r i, = m,. A-4 n,4 I

-4dLretain 99+% of all particles of 0.3-tas or larger diameter. i Filter moameriT1sgae or 15-mil-thick glass felt sheet, ~

dich is fire resistant and water repellent and has high wet  !

strength characteristics. Under accid ut conditions, the

_ filters could ha ===a==d *a hae =4 , -- --- _=r iuu '-

or fog mixture of steam and air.

a. Trademark of Otto H. York Co., Inc., Parsippany, New Jersey.

b. Trada==vk of E. I. du Pont de Nemours and Co., Wilmington, Delaware IV-45 l

l .

,.T_ . . _ _ -_..^ 7 _ ___ . 7- . . -~~.---~=--JC-'-~------!C'L--- -

CCCL i -- - -

DPSTSA-100-1 Rev. 12/76 E:

Filter Compartment l

i Air Flow ---*-

w.

/ -

= p = = J

- / C

/

/ J

(

= -

s l

Carbon Particulate Moisture I

- Beds Filters Separators FIGuitE IV-22. Filter Compartment (Elevation Cmss Section) e 9

~IV-46

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

_ _ _ _ .._ Z . i DPSTSA-100-1 Rev. 12/76 )

)

e Cebon beds. Activated carbon beds were designed to retain haineen activity that mieht be released if an accidens , , .

should occur. Each carbon unit is 7acked with about 57 lb_ L' of activated cocolRItrINETT caroon tsat is relatively dust - . .c rres._ aweAve separate pleats maintain a rigid 2 t

one-inch-thick bed, retard settling, and expose about 15 ft of perforated face area for air flow. Efficiency of the carbon beds is discussed in subsequent paragraphs.

Instrumentation is provided in the central control room to monitor the status of the confinement ventilation system, includ-ing fan operation and damper positions. A gang switch is also provided to shut off supply fans following an incident. Numerous alarms are provided to alert operating personnel to system changes.

In addition to the equipment described above, the ventilation sys- l tem contains supply fans, booster fans, control dampers, air pre- '

filters, and other miscellaneous equipment.

1 4.1.2 Iodine Radiolysis Effects 1

Theactivatedcarbonbedswereoriginallydesip*edforre- l tantion of elemental iodine. Kinetics calculations indicated

, that only negligible quantities of organic iodides could be fomed l before iodine was adsorbed on the beds. Based on manerous tests designed to simulate predicted temperature, hissidity, face ve-locity, and iodine loading conditions in a postulated reactor accident, it was concluded that the carbon in the. confinement mtems would retain 99.8S% of elemental iodine after at least-J 3.5 years service. Recent laboratory tests, however, indicate that penetrating iodine ' species are formed when carbon, iodine, and moisture are present simultaneously in an intense radiation field.as These penetrating species (primarily organic iodides) are readily desorbed from the unimpregnated carbon originally installed in the carbon beds.<

c __ -

Radiolytic desorption tests performed in a gamma radiation 7

field of 1.5 to 3.0 x 10 rad /hr (absorbed dose rate in the carbon) show that iodine penetration of unimpregnated carbon increases rapidly with, increasing carbon service age as shown in Table IV-5. ,

i .

t i

l IV-47 I

i'

l l

. l l

DPSTSA-100-1 Rev. 12/76 TA8LE IV-5. Effect of Service Aging on Unimpregnated Carbon C. -

Service Age, months Iodine Penetration.a g ,

0 0.37 21 0.61 33 1.04 35 3.61 -

46 4.38.

a. Penetration during a 5-hour test at 80*C greater than 955 rela'tive humioity, and an. absorbed dose rate of 3.0 x 10' rad /hr.

i Radiolytic desorption of iodine from carbon continues as l long as the radiation field is present (as it would be free from l radioiodine decay) for at least 100 hours0.00116 days <br />0.0278 hours <br />1.653439e-4 weeks <br />3.805e-5 months <br />. The rate of desorption 1 of iodine from carbon can be substantially reduced by using an -

impregnated carbon. Laboratory tests of a triethylenediamine (TEDA)-KI co-impregnated carbon show that the impregnated carbon retains iodine better after 6-months service in the confinement system than does new misprognated carbon (Figure IV-23). p lacement of Type 416 (unimpregnated) carbon.with, the more efficient Type GX-176 (impregnated) carbon was completed in 1975.

Y The effect of radiolytic desorption of iodine,on potential offsite thyroid dose is discussed in Section VI.

4.1.3 Postulated System Failures The exhaust fans and filter system would be online at the time of a reactor accident; therefore, no starting failures are postulated. The most critical active component is one of the three exhaust fans. - Because af the reA=imat fans and power supplies, the risk of a multiple fan failure is small. If a fan does fail while operating, the spare can be rapidly startd. while adequate air exhaust is continued by the second of the two normally operating fans. .

IV-48 l l

.___T , - . .

. . : :=.- , . - . . -  : =2 : .a======== === _ _=_. - - :::==_

4 l

DPSTSA-100-1 Rev. 12/76 l

E A. i. I i i O M '

99.3 - / - O.2

/

I 99.4 - / ,- 0.4 1

r i

x.  :

s so.4 -

o.s s >

a I_ '

M w i r e nos.sta Hm.wey E Gi 99 2 -

a w c== n o.a g g

a is ss c== .s g 2 .- -

. -us c s

=> _

i.o =

2 g

E E , a 3 99.8 -

/  : 1.2

. so.s - - i.4

"*o -

E E so Esposure Time, hours E ioo .s i

r, FIGURE IV-23. Iodine Retention Tests in Radiation Field of 15 Million rads Co/hr q

IV-49 l

T '~ . . - . - ..- ..= -.: :. :: z: ^. - - ~ = ::~ = ==:.-.:::: : :: =:: +- :~ - ~. . . . .

. vj .

>- .p rkT y ::Jqj

. .%g, ..::: CA1, DW1STC': ]i f1fAcHHgryy sf ;{

e y 0.% RIVER 1AEOPATORY CC: R. E. Naylor, Wiln. .; g C. P. Ross L, 2 --

f l H. J. Groh, SRp  ;

D. A. Eard h H. E. Kingo i H. R. Reeve , l8 I

J. P.1.*orin L. R. Jones 'r .

T. E.' Britt .

u It. R. Adams  :

J. P. Glas, SPL ,

". ~*

s c .x $ ' ro . . . . . . .,,.g ,,g; go3ye11 ,,. , , , ,,_. ]

t G. P. Merz F

- J. A. Smith l M. M. Anderson i N. H. Kuehn H. P. Olson S. P. Tinnes

~

\TISF 4 s

Ii July 19, 1979 .

N _E M _O _R _A _N _D _U _M

. j i ID: G. F. MiR2 s t.

FIO!: S. P. TINNES ,

AIP, BORNE ACTIVITY CONFIhBIENT SYSTEM I

BASE CASE DESIGN BASIS ACCIDEN'I  !

Ih'IRODllCTION  !

The. Airborne Activity Confinement System (AACS) for the Savannah River Plant (SRP) production reactors is designed to collect airbom e particulate and f -

halogen acitvity that might be released from the reactor in the highly unlikely event of a major reactor accident. The system consists of moisture separators, i Particulate filters, and halogen adsorbers in the e:iaust air portion of the t ventilation system. Air is exhausted from the resctor process areas through l the filters and adsorbers, and discharged to a stack. The system will not confinc or contain noble gases.

A study of the entire AACS' is now in progress. In order to evaluate the system, the conditions under which activity could be released from the reactor and

~

- stifferent system base cases must be defined. 'Ihis memorandtsn documents the

" Design Basis Accident (DBA) and AACS base cases that will be used in the study.

i

~

hts, conclusions, and rec %tions from the study will be doctanented in j.. .. stibsequent reports.

g..-

m .

sk

• E.

,0. . . - ~ .  ! DPST-7 5 u .

gu'.MW_ ,

Tr.c nesign Basis Accident (D3A) selected for the airborne activity confinement cvste NACS) is a full core meltdown with 50', of the iodine inventory reach-h- the Minement filters. The reactor power level prior to the accident is ag'stried to be M00 FL . A 0.05* instantaneous bypass of the confinement filters will be assumed. Two time frames after the incident will be investigated; 0 to 5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br />, which is the most critical time period, and >5 hours, which is a much  ;

Icss critical time period. Different base cases will be evaluated for each time frame.

s

~ . . - t forth, O to. Lhoor time frame,%eAhaust. Fa' w, .ko Exhaus't' Fan", s"nd "Ihre'e *

Exhaust Fan" cases will be used. The "One Exhaust Fan" case will depict the mini- }

nrt conditions of AACS operability allowed by current operating procedures. The f "Two Exhaust Fan" case will depict typical AACS operating conditions. The "Ihree Exhaust Fan" case will depict better than typical AACS operating conditions. Three [;

situations will be considered for each case. These are (1) delayed Emergency Cool- [

ing System (ECS) water addition, (2) the ECS fails, but the pmposed Confinement ,t Heat Renoval System (GIRS) is actuated, and (3) both the ECS and CHRS fail. The  :

iodine distribution among the filter compartments and amount of steam and/or smoke blinding of the high efficiency particulate air (HEPA) filters will be varied for each situation. .

For the >5 hour time frame, two base cases will be considered. -These will be - 4 similar to the "One Exhaust Fan" and "Iko Exhaust Fan" cases described above.

Hypothetical AACS malfunctions will be investigated for each case to determine if 3 the total AACS flow drops below minie= levels. ,9

?

bbre credible accident cases than the DBA will also be considered in the AACS study.

These accident cases will be discussed in subsequent reports. {

j DISCUSSION i

Desien Basis Accident . :t i

U Tiw est serious s tusuon potential er failure of thereactor ECS. accident is a major D 02 leak followed by delayed j lost in a short time and fuel melting would follow.Under these conditions, Activity would be released fuel coolant could fbe P,

first to the reactor tank and then to the above-grade process room through numerous . 3, y vents in the top shield and plenum. If ECS water is not added rapidly after the <

Ioss of D.3 fuel and 51uminum would slump to the bottom of the tank.0 coolcnt, the J

's entire c this point (delayed EGwater addition), copious quantities of steam would beIf ECS' water is add -

generated creating a pressure sur of the HEPA filters would occur. ge throughout the AACS. Partial steam blinding If no ECS water is added, most of the molten  ;

mass in P and K reactor would flow into the outlet nozzles and into the pumps et the and -40 ft. level. Eventually the molten mass would melt through the piping slump to the pumproom (-40 ft. level) floor (reference 1). In C reactor, the outlet no::les are 15 in. above the bottom of the reactor tank; thus j nc molten mass would be retained in the tank until the temperature became suffic-ticntly high to melt through the tank bottom dnd bottom shield (reference 1).

. he proposed GIRS is operational (reference 2), the -40 ft. pumpmomIffloor would W flooded with Disassembly Basin water to a depth of 8 to 32 inches prior ec1t-through to of the molten core (reference 3). When the molten core reached l

L,2. . . . .

the CHRS water, copious quantities of steam would be generated. Partial steam )i blinding of the HEPA filters would occur. If the excessive heat causes any 4?

y fires in the process areas, partial smoke blinding of the HEPA filters could J also occur. If the proposed CHRS is not operational (no water cooling of the i molten core), ventilation zir flow will tran port a fraction of the decay heat in the molten core debris on the pump room floor to the confinement filters and cxhaust fans. Calculated air temperatures are high enough to cause failure of the confinement system and release of airbome activity (reference 1). These cccident events, as described, might release as much as 100% of the equilibrium inventcry of noble gases and 50% of the iodine inventory to the confinement i ,

system (reference 4). ~ * -

r The amount of activity released E ., 3 rom an accident is a functi'on of the reactor '

l The highest power ever achieved at SRP was l power 2,915 W at (reference the time of5).

the accident.Normally, power levels range from 1,700 W to 2,500 ;W -

depending upon the charge design and coo'.ing water inlet temperatures. It is -

not anticipated that power levels for the SRP reactors would ever exceed 3,000 W.

)

Any radiciodine that penetrates the confinement sys:em carbon beds is assumed to ,

be released from the beds by desorption mechanisms (reference 6) or by filter  !

t bypass. The mechanical integrity of the assembled carbon beds is measured follow-  !

ing installation and periodically thereafter using inplace testing methods (refer- I cnce 4). The bypass (instantaneous) leakage is normally less than 0.03%. The current Technical Standard requirement is that the instantaneous bypass shall not  ;

(

exceed 0.05% (reference 7). .

Based on the above information, the DM was selected for the airborne activity confinement system study now in progress. The DBA selected is a full core melt- i i

I  ;

down with 50% of the iodine inventory Teaching the confinement filter's. The reactor power level prior.to the accident is assumed to be 3,000 W. A 0.05% t F

[

- instantaneous bypass of the confinement filters is assumed.

NRC Guidelines .

I Me DM selected meets the requirements specified in the Nuclear Regulatory Comiss- d' 1 ion 10 CFR 100 " Reactor Site Criteria" Guidelines (reference 8). The Guidelines  !

i ~!Itate that the fission product release assumed should be based upon a major accident that would result in potential hazards not exceed by those finn any accident con- i sidered credible. They further state that such accidents have generally been  !

cssumed to result in substantial meltdown of the core with subsequent release of k appreciable quantities of fission products. ' r i

Time Frames To Consider  !

l 1he most critical period for the AACS is fmm 0 to 5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> after the DBA has occurred.  ;

If significant radiciodine breakthrough is going to occur, it will probably occur i during this period. Calculations show that after 5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> (from initiation of the l DBA) the filter compartment carbon bed temperatures have peaked, and the carbon beds [

cre beginning to cool as significant radioactive decay of the iodine isotopes has e begun (reference 9). 1W time periods vare selected for consideration in the AACS fJ I study. These are from 0 to 5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> after* the DM and greater than 5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> after the DBA.

b .

• *W**-

.g s'

, .e p ,I * - - ' - j'i

3,

• r; . ;w -

' s

}

f y

_First 5 Hours After DM 3 I

1:ey variables for detemining the effectiveness of the AACS during this period are the number of exhaust fans online, the number of clean or dirty filter f

comparments online, the age of the filter c:cpartment carbon, the amount of l steam and/or smoke blinding of the HEPA filters, the distribution of iodine )

J among filter comparments, meteorological data, and the situations (after the i DM) to consider. A clean filter compartment is one that has clean demisters and O HEr.i filters at the time of the DBA and has the least possible flow resistance. 1 A dirty filter campart=ent is one that has fouled (dirty) demisters and HEPA filters at the time 3ff the DBA e.nd has the maxir.pn flo.4 resistance allowed by .. y

~

~

current operating procedures. nree base cases were" selected. These have been - '

dasignated the "One Exhaust Fan," "Two Exhaust Fan," and "Three Exhaust Fan" cases.

"One Exhaust Fan" Case The "One Exhaust Fan" case assumes the AACS was being operated at the minimum conditions of system operability allowed by present operating procedures at i the time of the DBA and for the 5 hour5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> period following the DBA. It assumes }

only one 903 exhaust fan is online, three dirty filter compartments are online, j and the carbon age in all three cmpartments is 30 months. j 4

"Two Exhaust Fan" Case ,

j s

The "Two Exhaust Fan" case assumes the AACS was being oxrated at typical j j

conditions based upon reactor operating experience at the time of the DBA and for the 5 hour5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> period following the.DBA. It assumes two 903 exhaust fans .j are online, four filter compartments are online (two clean and two dirty), (

and the filter compartment carbon age'is 1, 7,15, and 23 months, respectively.

"Three Exhaust Fan" Case .

The "Three Exhaust Fan" case assumes,the AACS was being operated at better than  :'

typical conditions for the 5 hour5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> period following the DBA. Fr'esent operating procedures specify placing all available 903 exhaust fans online after a reactor incident. The case assunes three 903 exhaust fans are online, five filter  :-

compartments are online (three clean and two dirty), and the filter compatt-ment carbon age is 1, 7,15, 23,, and 30 months, respectively.

-I De iodine fractions reaching each filter compartment, amount of steam and/or smoke blinding of the HEPA filters, and situations to be considered (after the DBA) will I

be varied for each base case as described below. ~

Iodine Distribution on Filter Compartments .

ne iodine distribution among filter compartments will be based on the  :

observed and calculated Ventilation System flow imbalance between the near l

l and far side Reactor Room exhaust headers. Under most Ventilation System conditions, more Reactor Rom air is . touted to the filter compartments through the near side exhaust header than the far side exhaust header because the Purification exhaust duct ties into the far side header. In N

,/ .

. . L _ _ _ _ _ _ _ _ . _ _ _ - .. .

y w

. - . . r .;1 sd;iticn, the Eelo'.: Grade e

'[1 the near side filter co~ par)tmein3ct header (Upp ,

4 -

side filter compartments. on the near side. ter conpartment BEPA Fil_ter Blindinc nts to receive more iodine ar tha IEPA filter perfomance tests thDe amount of IE i

different filter reactor smoke blinding ass accident tat onswere situa i performed in 1973 on sim reference 10).

, wide range of conditions. umed will be He v(ariedamountfrom of HEPA0 to 60

1. caused 60'. sr by smoke generated from any Partial HEPA filter gohen could ty to bliridi g'ppresent 8 .. -

tions.pke It do blges not represent a g bus arbitrarily picked as a " bad" The case fop that any smoke generated iodin nyeparticular wouldcase accident blinding possible bimply nor is mr eant to the thema cal e to the filter ccmpartmentse swept out of the pro. It will be assumed 3 cess areas with the j

{

Accident Situations (After the DHQ Bree situations were selected .

These are 1 k piping to the -40 ft. pumpreventing tn the molten core from t

}

GRS is actuated, and (3)proom (2 floor, g thro (ug)h the re ture transient calculations both thehECS an)d the'GRS failthe ECS fails l

been documentedave forbeen eachmade situati .

in previous studies and hA c on (references 1,10 and 11) ave i

.A base totalcase of 27 different analyses to be analyses completfor d e j e ach base case are s

>5 Hours After DBA are shown in Tables I,pecified.

h II, and III.The different li .

}

his tir.c period is much less critical than the first filter compartment forced co li CD1 for the delayed oECS and CHES S i out the period (reference 9) ours after the DBA. Tne that the total Ventilation ng He primary requirements S requirement during thi have decreased

{! AACS malfunctions ystem that flow couldcasure must be'at least that all ecrease process Cthrough- area s40time 000frame is g 40,000 CD1 will bepossibly r

investigated. reduce  ! total sVentilation (reference 12 Se ma jI Two base cases will be co \ ystem flow below).

i three dirty filter compart i q cican he second case nsidered2 assumes i ntwo e first 903 case assumes o '

ments nnd one dirty) are online at th exhaust are+ online fans and at four the ne 905 time exhaust filter malfunction.

e system comp of th fan and

. malfunctions listed ine timeTable be investi of the IV systemwill' malfunctionartments (three 1 gated for each base cas.e.H e AACS '

j ..

g .

f  :

i 2

1

.a

?.: ,i s

l i

. 4

- 9

'**-'W

~ ~~

~ ,

y= -

k h ~ ..- -

c, : . . :w e e s. .: ,,1  ; j e

jes.cer Accident Cases

~'bre credibic accident cases than the DBA will alse be considered in the AACS 3:t:.ly. Tnese cases will be docts:ented in subsecuent reports.  ;

i f

4

?

! @*Wh s-  ; . %, ". 4! g- ,i

. ;., - . -a. .g y- .%"

r

)

s I t

a f f

i I

e *

/

I

~

s # r e '

< r I -

I I

1 I i I $

1 i

i O

. 1 i, *

i a

{ h 2

7F t-

• 1 2

O J

! t

?  :

l E

1

• t ,

• l 4

I l

I .

O

• • 4 up

,, ; , ;.I 7 DPST-7.0-4"1 nEFERINCES h J. DPST-70-433, W. S. Durant and R. J. Brown, Analysis of Postulated Core f.bitdown of an SRP Reactor, October 1970. (Secret; 2

2. RTW-222, J. H. Hinton and S. P. Tinnes, FY 1977 Budcet Proposals - Improved Confinement of Radioactive Releases, January 1970. f c

3. TA-1-2049, DPS0X 9142, J. H. Hinton, Improved Confinement of_ Radioactive

, , , Releasesg. January.5,1;9796 ,x3 , g -;. . -

.s c t" 44 4 - t . L'

~

4.

DP-1071, August W. S. Durant, et al., Activity Confinement System of the SRP Reactors, 1966. J-( DPSTSA-100-1, J. A

~ Reactor Operation,y, et al., Safety Analysis of Savannah River Production December .

6.

DP-1280,Report:

Progress A. G. Evans and L. R. Jones, Confinement of Airbome Radioactivity - j January 1971 - June 1971, Octo' er 1971.

o 7.

DPSTS-105-3.01,105-Building ment, April 1, IFT5. Technical Standards - Airbome Activity Confine- 'I f

S.

Reactor Site Criteria, Title 10, Regulations - Nuclear Regulatory Cor:nission, '._

Part 100, AprH 1962 and subsequent amendments. * .

9. DPSI - L

, S. P. Tinnes, Airbome Activity Confinement System - First Five Hours After DBA, (to be issued). . :i j

10. DPST-73-482, L. R. Jones Effects of HEPA Filter Blinding on Reactor Confine- F ment System Perfomance,,Novemo' er 2,1973. - J F

11.

DP-1355, A. G. Evans and L. R. Jones, Confinement of Airbome Radioactivity 3[

Procress Report: July - December 1973, June 1974.  !

t

12. h ETR-1382, Process ArcaJ.Pressures, B. Guffy, AprH Minimum 1973.Reactor Building Flows Required for Subatmospheric i f

r-9 i

I

_ ~

.'., ws!-7.; .:.:1 .

T!JiLE I .

O to 5 Hours After DBA i i "One Exhaust Fan" Ease Case l

o One 903 Fan Online o Three Dirty Filter Cmpartments Online .

.e ,g -. ... o, Caybo0,ge y . Ql;pa_ments conpaingkg,th 3t is 30%nths g old Yar_iables o Situations and IEPA Filter Blinding Delayed ECS Water Addition o Stean Blinding To Give 40,000 CFM Flow o Smoke Blinding - Ot, 201, 40%, 60%

No ECS, O BS Functions -

o Steam Blinding To Give 40,000 O'M Flow o Smoke Blinding - 0%, 20%, 40%, 60%

No ECS or OBS .

o . No Blinding Assuned ,

o Iodine Fractions On Filter Ccrapartments (Starting From Near Side)

Flow weighted 0.50, 0.50, 0.0 0.65, 0.25, 0.10 e o Total Number of Analyses - 27 l

t o

I

~

• p ~[-Y_* .'*--.'.?.

-_ . . ,, . T' . _ . , . . . _ _ _. ~T~ ~~ P J.L. ,

li:

TAFLE II  ;

O to 5 !! curs After IBA i

e "Ttco Exhaust Far." Ease Case o "Iko 903 Fans Online \

o Four Filter Cmpartments Online (Iko Clean, Two Dirty)

- +- %;'-oc Carbon %6" 77, ,23',1,15)lpoths.01d (Stgrting Frcn Near Side) l i

Variables o Situations And HEPA Filter Blinding

}

- Delayed ECS Nater Addition

l. o Steam Blinding - 50% ,

- o Smoke Blinding - 01, 20%, 40%, 601 '

s .

- No EG, GRS , Functions .

i i o Steam Blinding - 40% .

t' -

o Smoke Blinding - 01, 20%, 40%, 60%

I.

t'

! - No ECS or GRS , ,

1 l

! o No Blinding Assumed . ,

I, o Iodine Fractions on Filter Compartments (Starting From Near Side) ,

i Flow Weighted ,

{

I -0.50, 0.50, 0.0, 0.0 i

! - 0.45, 0.35, 0.15, 0.05 o - Total Ninber of Analyses - 27 g

• P

)

W_

7 s-

. ._ =

Li'57-79 ~41 j

j .

TMILE II .'

0 to 5 Hours Eter D3A l

l "Iko Fxhaust Fan" Ease Case 1

o Two 903 Fans Online o Four Filter Caspartments Online (Tko Clean, Two Dirty) e .

1 o. Carbon ,Nie'- 7, 23, T,' IS-Pknths"Old (Star" ting From Near Sidh)

{ Variables l'

f a Situations And HEPA Filter Blinding 1

[ -

Delayed ECS Water Addition

?

i o Steam Blinding - 50%

l . .

15 o Smoke Blinding - 0%, 20%, 401, 60%

(,

No ECS, OSS Functions .

o Steam Blinding - 401 o Sruake Blinding - 01, 20%, 40%, 60% ,

?

No ECS or ORS - ,

o No" Blinding Assuned L r

o Iodine Fractions on Filter Canpartments (Starting Fran Near' Side) -

- Flow Weighted -

h 0.50, 0.50, 0.0, 0.0 t t t

{

0.45, 0.35, 0.15, 0.05

c Total Nunber of Analyses - 27 -

E

r.  !

j -

l N

i y  :

P I E 2 1 w

s _-_ y g- - --

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

DPST-79-4'l

• I . .

t TABLE 111' .

O to 5 lburs After MA h

' li "Three Exhaust Fan" Base Case }'l<

l 1 I

o "Ihree 903 Fans Online o Five Filter Cc:partments Online (Three Clean, Tko Dirty) '

. ;~ . . . '

t

. o Tdy; :.'15; T,h0,' I? 23' M6nYOld1 Starting FENeiir Side) s' f

i Variables t

)

o Situations And HEPA Filter Blinding j

' I

} -

Delayed ECS Water Mdition j ,

1 j

. I o Steara Plinding - 40% i

.g s I i

o S:noke Blinding - 01, 201, 40%, 60%

~$

. - No ECS, GiRS Tunctions p

.f? o Steam Blinding - 301 .

1 o Smoke Blinding - Ot, 20%, 401, 60%

E

3 -

No ECS Or CHRS

~*

o No Blinding Asstraed I

[!

j o Iodine Fractions Q1 Filter Campartments (Starting From Near Side) 4 '- Flow Weighted

}

? - 0.50, 0.50, 0.0, 0.0, 0.0 p i F

- 0.35, 0.30, 0.20, 0.10, 0.05

-p -

! o Total Number of Analyses - 27

-: i i l

I 0

7 L 1

l 1

1 I

. - - ~ . - .. ,

L

i .

,;l! ~

TMt!.E IV

-y

., >5 Hours After DBA q ],

lii Ii 1

AACS Malftmetions, Opemting Errors, Considerations

!I l ti I o M and N D=:pers Opca jq o M and N Dr.pers Cosed '

[

o M Damper Closed, N Da:r.per Open "W..^

4 " "

So G Damper CIdsed 3 '

% 3- - .

i o K Da::per Open - Offline 903 Fan /s o One 910 Fan Online, OfflineQ Damper Gosed o 'One Q Damper Gosed -

o Dampers G osed o Pressure Surge, GF Ikeper Interlock Fails o Pressure Surge, G Dauper Interlock Fails .

o Proposed 500,000 Gal. Contaminated Water Storage Tank Vented To AACS At +3'in.'H 2O Pressure-o Two 902 Fans Online l o Three 902 Fans Online o Two 1230 Fans Online  ;

o All 902 and 1230 Fans Online j [. ,

1, o Natural Phenomena (Tomadoes, etc) -

t i

f C

I

• i i

1 l I I

• k

\

{

t  ; -

1

. -p ,

I A s.. - -__- - _.__ __ _ _ _ _ _ _ - _ . _ , . _ _

._. _.- -- . _ _ _ .