ML20053B554

Forwards ACRS Consultant & NRC Questions & Comments Re Plant Probabilistic Safety Study

ML20053B554
Person / Time
Site:	Zion  File:ZionSolutions icon.png
Issue date:	05/18/1982
From:	Varga S Office of Nuclear Reactor Regulation
To:	Delgeorge L COMMONWEALTH EDISON CO.
References
NUDOCS 8205280567
Download: ML20053B554 (21)

Text

{{#Wiki_filter:._ .R DISTRIBUTION 'w/o encls Docket File ' s NRC PDR Local PDR MAY 181932 v0RB 1 File D. Eisenhut 0 ELD Docket Hos. 50-295 OI&E (1) N and 50-304 D. Wigginton 'i 9 C. Parrish NSIC h ACRS (10) ' 8 S gS 75 Mr. Louis 0. DelGeorge ~ Director of Nuclear Licensing [ s.8g9 pb R Connonwealth Edison Company M t'6 Post Office Box 767 Chicago, Illinois 60690 O V f /. gA

Dear Mr. DelGeorge:

co In the ra iew of the Zion Probabilistic Safety Study submitted by your letter dated September 8,1981, questions and comnents have been developed which we request that Cor.monwealth Edison Company address. is a summary of coments and questions generated by ACRS consultants. They are grouped topically and keyed to the pages or sections of the particular consultant reports in which they are mentioned. We have attached these reports as Enclosure 2. contains staff and consultant questions. Where there is duplication of questions between ACRS consultants and staff, there is a notation so that CECO should be able to provide a single response. If there are any questions on our comments or if you wish additional clari-fication of any of our questions, please let us know. In order for us to i complete our current review, we would appreciate a response by June 30, 1982. I 1 Sincerely, Original algned by:

s. A. varga Steven A. Varga, Chief Operating Reactors Branch No. 1 Division of Licensing

Enclosures:

1 As stated CC W/ enclosure.3 See next page I i 8205280567 820518 PDR ADOCK 05000295 P PDR l\\ l ORB 1 1.5<........... ........................l ornce> DWigg ....../.r5 r gt........ ton suanaur > l 5 [. 2 2 ......................I one> ' nac ronu ais tio.aoi nncu o24o OFFICIAL RECORD COPY uscron,u-mu}a

...=--~..=~_...........:-,.2. \\ . ~, - r -... l Mr. Louis 0. De1 George Commonwealth Edison Company i cc: Robert J. Vollen, Esquire 109 North Dearborn Street Chicago, Illinois 60602 i Dr. Cecil Lue-Hing Director of Research and Development l Metropolitan Sanitary District i of Greater Chicago 100 East Erie Street j Chicago, Illinois 60611 Zion-Benton Public Library District i 2600 Emmaus Avenue j Zion, Illinois 60099 l Mr. Phillip P. Steptoe Isham, Lincoln and Beale i Counselors at Law One First National Plaza 1 42nd Floor Chicago, Illinois 60603 Susan N. Sekuler, Esquire Assistant Attorney General Environmental Control Division -r 188 West Randolph Street, Suite 2315 Chicago, Illinois 60601 l U. S. Nuclear Regulatory Commission Resident Inspectors Office 105 Shiloh Blvd. Zion, Illinois 60099 1 James P. Keppler Regional Administrator - Region III U. S. Nuclear Regulatory Commission 799 Roosevelt Road Glen Ellyn, Illinois 60137 !1, 1 2 '} i i l j .,_7

- - - ~ " - ~ ~ ~ " " * - - - ~ F., Systems.~snalysis ENCLOSURE 3 1 ~ 1. The ZPSS gives cor.plete credit for the remaining water and refilling the RWST fol-lowing dryout to allow for continued containmer.: spray operation. Describe the plant procedure which instructs the operator to do this. Include a discussion of time CSIS will be available with respect to loss of low pressure recirculation, core ceit, vessel failure and containment cooling. (See ACRS question III.3(2)) i 2. The ZPSS assumes' that the operation of the containment fan cooler system will not i be adversely affected by the environment present in the containment following a i core celtdown. Since many fan cooler design parameters ar'e exceeded by the post core melt environment, provide an analysis to support this assumption? l 3. The ZPSS assumes that safety system pumps do not require cooling as long as they { are dra. sing from a cool water source. assumpa,on. Provide your analysis to support this 4 Past generic ATWS analyses by Westinghouse were based on the following important assumptions regarding the ability to limit the calculated peak pressure to below 3200 psig: i) Availability of pressurizer safety and relief valves. ii) Capability (actuation circuitry diverse from the protection system) to automatically trip the turbine early (30s60 seconds), iii) Capability (actuation circuitry diverse from the protection system) to i j automatically actuate all AFWS pumps (30s60 seconds). 4 t If diverse actuation circuitry to perform mi~tigating functions defined under items. ii) and iii) above are not implemented, provide your bases for the assumed peak pressures in the event of our ATWS. Why is core melt assumed to occur at 3200 psig? What would be the effects on the steam generator tube integrity? 5. The ZPSS assumes the operator has a high probability of opening a PORV, during f the early phase of an ATWS, in order to avoid exceeding 3200 psi. Since Zion j ATWS procedure does not instruct the operator to open the PORY, and only a few i minutes are available to do so during an ATWS, the high probability of success used in the ZPSS needs further explanation. Estimate the time available more i closely and give supporting evidence. Moreover, although with the resetting of the motor-torque switches (see p.1.3-40 of ZPSS) on the motors of the block valves, the motor-operated block valves are operable under primary system pressure, and it is not clear that they remain operable when there is an appreciable i increase in system operating pressure above normal. DETERMINE THE MAXIMUM OPERATING PRESSURE AT WHICH THE BLOCK VALVES ARE OPERABLE, IF THIS PRESSURE IS j LESS THAN THE PRIMARY REACTOR COOLANT SYSTEM BOUNDARY FAILURE PRESSURE, AND TAKE THIS PRESSURE INTO ACCOUNT IN CONSIDERING THE TIME AVAILABLE FOR OPERATOR ACTION. IF FAILURE OF ATWS PRESSURE SPIKE PROTECTION DOES NOT NECESSARILY LEAD TO CORE + MELT, PRESENT SUPPORTING EVIDENCE (SEE QUESTION 4), Af(D REVISE THE EVENT SEQUENCES TO REFLECT THIS. QUANTIFY THE REVISED SEQUENCES. (See ACRS question III.3(1)) 6. '4hy are the "other" contributions to safety system unavailabilities always negligible? 7. The ZPSS gives credit for " feed and bleed" core cooling following a total loss of main and emergency feedwater. Is the " feed and bleed" success criteria consistent with results of the " feed and bleed" tests which were recently conducted at the LOFT facility? Provide your technical basis for asswing adequate core coolf.cg via this mode. ~ >-M N WS = =. - -. = ~ - - ~~~ ~~ "^~

,g ).' Question on ZPSS - k ) 8. The ZPSS assumes the reactor coclant pumps will leak at a rate of 300 gpm each i approximately 30 minutes following a loss of seal cooling. Describe the analysis I conducted which supports this flow rate. l 9. Why does the ZPSS not model safety system equipment room cooling but the Zion l P& ids shows several room cooling systems? g

10. Justify the data used in the calculation of the interfacing systems LOCA (event V).

Review of the data sources referenced would wggest that this data does not ex-actly apply to the failure modes postulated. (See ACRS question III.3(4)) 3

11. Offsite power recovery is a major factor to which " station blackout" (loss pf all AC power) and other loss of offsite power accident sequences is most sensitive, Depending on whether the recovery potential used in the Zion PRA or recovery l

based on other nuclear power experience is used, a significant difference can be obtained for the frequency of loss of offsite power related sequences. Justi fi-cation should be provided for using the Zion PRA recovery model rather than the i less optimistic values obtained based on actual loss of offsite power experience. l (See ACRS question III.3(3)) +

12. One area apparently not addressed concerning the " station blackout" dominant sequence was the probability and effects of DC power loss due to battery power depletion during an extended loss of AC power.

Provide your analysis of this scenario to include: What is the battery life (in hours) using procedures currently available a. or proposed at Zion? I b. What inst'rumentation and controls, particularly affecting AFWS operation, would be available if DC bus 011-1 could.be energized from unit 2? c. Can the steam train AFWS continue to be operated following loss of AC and j OC power? How would it be controlled if steam generator level indication is not available due to loss of DC power? Are there procedures to do this? j l d. Would the loss of DC power (before AC can be restored) affect the recovery 3 of powe'r to the plant; such as (1) affecting the ability to restore offsite j power? (2) starting diesels without field flashing? (3) affect on plant j security systems so as to possibly hinder access to areas of the plant? 13. For Anticipated Transients without Scram (ATWS), there is a sequence analyzed in the German Risk Study (see EPRI-NP-1804-SR, p. 5-32) where a pressurizer safety valve fails open, after a loss of main feedwater transient followed by failure of This sequence was ass'umed to lead to core me reactor trip. The German Risk Study used a probability of 2.5x10-{t, in the German Risk i Study. per demand for failure of a pressurizer safety valve to reclose. With a mean frequency of loss of main feedwater of 5. /yr, a mean frequency of. 3. 7/yr for turbine trip, a probability of 1.8x10-4 for failure of reactor trip, and a probability of.075 that one of the three pressurizer safety valves ticks open, one obtains (5.2+3.7) (1.8x10-4) (.075)/yr = 1.2x10-4M for the probability of this sequence. IF THIS SEQUENCE WAS Oti!TTED BECAUSE IT DOES NOT LEAD TO CORE fiELT, PLEASE GIVE SUPPORTING EVIDENCE.

• e -~

-n-.nw..- ~ =,.:~.:- : . - = * ' '.. = - -

• ~ * &%

- ~- - ~ M- ~mv-~

_. c _ _ z__--- Question on ZPSS .i 14. For a Small LOCA the ZPSS study (see p.1.3-113 of ZPSS) asserts that success l for High Head Injection requires delivery of water from one-out-of-two high head safety injection pumps or delivery of water from nne-nut-of-two charging i pumps to the reactor coolant system. However, the Offshore Power Systems re-i port (An Evaluation of the Residual Risk from the Indian Point and Zion Muclear l Power Plants, Report No. 36A75, February 1980) assumed that two out of the four pump set consisting of the two high hecd safety injection pumps and the i two charging pumps were required. PLEASE PRESEiiT EVIDEiiCE FOR THE ASSERTION THAT ONLY ONE OF THE FOUR PUMPS IS REQUIRED FOR A SMALL LOCA. .15. In the Steam Generator Tube Rupture Event Tree (p.1.3-130 of the ZPSS) an { operator-action event OP-5 is considered. It is remarked on p.l.3-42 of the i ZPSS that "the operator could do nothing for a long period of time. In this mode water / steam would eventually pass through the steam generator relief val ves ". However, main steam lines are, generally speaking, not designed to take water loadings and may fail. A failure of a main steam line may result in sudden flashing of the water in the affected steam generator. The sudden cooldown of the reactor vessel may threaten reactor vessel integrity. JUSTIFY THE OMISSION OF THIS ACCIDENT SEQUENCE. 16. In the Steam Generator Tube Rupture event tree, failure of the High Pressure Injection System (HPIS) is considered only as a part of the OP-5 event. Failure of HPIS after a steam generator tube rupture is not assumed to lead to core melt, i f no other failures occur. JUSTIFY THIS ASSUMPTION. Also, provide justification for assuming depressurization capability o.f AR4 if failure of OP-5 occurs. ~ 17. After a LOCA, in the recirculation mpde of cooling the core, the residual heat removal pumps take suction from the containment sump. There are motor-operated valves (SI8811 A and SI88118) in the lines from the containment sump to the RHR pumps. The ZPSS used a value of 1.55x10-3 per demand for the mean failure frequency for these valves (failure mode: fail-to-open). The failure Trequency is supposed to include failures in the local control circuitry for these valves (see p.1.5-496 of ZPSS). However, the Offshore Power System study, following the Reactor Safety Study, used a value of.03 for the failure frequency of the local control circuitry of these val'ves. Moreover, common mode failures of the local control circuitry for these two valves may be of importance. EXPLAIN THE DIFFERENCES IN FAILURE FREQUENCY FOR THESE VALVES, AS CALCULATED BY OFF-SHORE POWER SYSTEMS USING REACTOR SAFETY STUDY DATA AND AS GIVEN IN ZPSS.

MORE0VER, ADDRESS TliE PROBLEM OF COMMON MODE FAILURE OF THE LOCAL CONTROL CIRCUITRY FOR THESE VALVES.

f 18. Provide justification for not considering events which result in overcooling and pressurization of the reactor co.olant system and would threaten reactor vessel integrity. Also provide the infomation requested on pressurized themal shock as listed on page 19 of this enclosure. HUMAN RELIABILITY ANALYSIS A major difficulty in reviewing the human reliability analysis (HRA) in the Zion probabilistic risk assessment (PRA) resulted from the lack of adequate documentation to permit full evaluation of assumptions and estimates of human error probabilities (HEPs) employed. Nevertheless, considering only those system failure events we selected as critical for core melt or other risk, it was possible to understand sufficiently well the Zion estimates of unrecovered HEPs to permit an appropriate sensitivity analyses. Zion estimates of response times and/or HEPs were increased to larger values (e.g., doubled or quadrupled) in accord with mre conservative analysis of the impact of human perfo rmance. We determined that the larger values had no material influence on the 4 psW ~" '~- -.= 2: _:-~~ ~ ~;L.*

" Question on ZPSS, sequence calcul'ations for the following events: loss of all AC power with failure o of AP3; and failure of LPI in a large LOCA. In the events involving ATWS (loss of main j feedwater ATWS and turbine trip ATWS), we did not accept the Zion HRA. No credit for 1 human intervention was allowed in our analysis whereas the Zion HRA did allow ample l credit for human intervention. We did accept the Zion estimates of human error prob-abilities for failure to initiate switchover to recirculation for either a large or a small LOCA and for failure to establish and maintain cooling by feed and bleed. We i also accepted the Zion estimate for failure to open MOVs CC9412A & B (the isolation valves for the component cooling flow to the RHR heat exchangersj in a large LOCA event even though they optimistically assumed that four people would have to fail. In this i particular case, although we disagreed with their aspect of their HRA, we judged that 1 other conservatism in their analysis made up for the optimistic assumption of four f, people. Therefore, for all of the event sequences in which human error played a significant role, we have been able to provide an independent assessment of the impact of human j performance despite many questions about the assumptions and estimated HEPs used in the t' Zion PRA. However, there are two potentially significant classes of human error that j we are not able to evaluate. These are: l. The apparent non-consideration of some possibilities for conion cause 4 failure from human errors during calibration and during restoration pro 4 .i cedures of safety-related components after test, maintenance, or calibration. 2. The possibly insufficient consideration' of errors (apart from ccmmon-cause errors) in restoring safety components after test, maintenance, or calibi ation. +j To enable us to evaluate the above two areas, we need to know what kinds of rules J the Zion PRA personnel used to dismiss common-cause failure from human errors and to c i decide that all human errors related to restoration tasks would be recovered. He further need to know what kind of tagging and administrative control (including logs j .and other paper work) are used at the Zion plant to ensure that recovery from such errors is highly likely. j j There are some other types of information which could assist us in evaluating your ) 3 i i assessments of human performance-l 3. What degree of practice on the simulator and how often are in-plant walk-throughs given to operators in responding to multiple problems, e.g., ATWS j with loss of feedwater? \\j 4. On what basis was it decided that the STA would be involved in detailed i switch manipulations related to post-accident conditions? .i j 5. At the Zion plant, is the STA to be as available as the SE? Since the STA is an SRO, does he also function at times as an RO or a.s an SE? 6. Was the dependence assumption stated on page 0.15-5, Volume 2, of the Zion J PRA used in any calculations of common-cause human error in valve restoration j tasks ? If yes, it is quite possible that some very optimistic judgments werc made. l} ~

t

.11 lh 3q

a Qusstion on ZPSS - 4 7. What type of psychological scaling procedure was used to derive the histograms of ope. ator r:sponse times and respons:: time versus probabilities of error? What were the c;aalifications of the developers of these procedures in the field of psychological scaling? ESTIMATION METHODOLOG_Y l. What is the basis for claiming the assumed prior distributions of Zion component failure rates and probabilities are " frequency distribution's" - the "known results of experiments on populations"? If this assumption is false, how reliable are 4 i ycur estimates? 2. What is the rationale by which some WASH-1400 5/95 percentiles w'ere treated on 20/80, for Zion, while others were accepted as 5/95? How sensitive are key results to these choices? (See ACRS question II.2(1)) i 3. Why was the increasing trend in the turbine-driven AFWS pump unavailability ignored? Discuss the effect of recent AFW system failures on the ZPSS AFW analysis. 4. Provide justification for your statements about completeness, particularly given 1 the crbitrary and inconsistent ways in which "other" failure probabilities were estimated? (Why was an average 8-factor of 0.14 assumed for t..e failure of similar i components in the low pressure injection and recirculation systems, but not for 1 the reactor protection, engineered safeguards actuation, containment spray, con- ] tainment fan cooling, component cooling water, service water, and auxiliary 2 feedwater sys tems?) \\ l 5. What is the basis of the assumed distributions (page 1.3-15) from wnich the electric power recovery probabilities were derived? How do you resolve your estimttes of .28 and.03 for failure to recover in 30 and 60 minutes, respectively, with industry-wide experience of.41 and.26? Since it takes a more unusual event than average j to fail Zion's offsite pcwer, shouldn't recovery be less likely than average? -l 6. Please explain the Interfacing LOCA estimates. What is the basis for using ? WASH-1400 bounds for valve rupture? How sensitive are the final risk estimates I to this assur.ption? Why is the interfacing systems LOCA estimated to occur so } much less frequently than a large or medium LOCA 1.05(-7) vs. 9.4(-4)? See ACRS question II.2(2)) t l 7. Why was pressure vessel rupture excluded from explicit quantification? I I i i = .m

'r. ., = i - 3 i i Seismic Fragility I ~ The following questions were taken from the draft report (dat'ed 22 February 1982): " Review of the Zion Probabilistic S f ta e y Study, Seismic Fragility," prepared for Sandia National Laboratories ey Jack R. Benjamin and ' Associates, l Inc. which is Appendix 0 to March 5,1982 Sandia letter report. The page numbers below each question refer to the draft report which gives background material for each question. The issues raised represent the most significant ,f concerns for the seismic fragility study which should be addressed and i resolved. i l .L J 1. What seismic hazard curve values were used in the integration with S the seismic fragility curves to obtain the frequency of core melt probability distribution? j (pp. 5, 6, 21, 23 and 94) 2. Is the definition of the damage effective ground acceleration appropriate for equipment which depends on functional operation as opposed to ductile strength capacity (e.g., service water pumps)? I (pp. 17, 56, 63 and 78) 1 3. What uncertainty was assigned to the inelastic energy absorption parameter for structures and equipment to account.for the variability I caused by using single-degree-of-freedom models for multidegree-of-frcedem prototypes? i (pp. 50 and 66) 4 e. ,mqg gee

l ~ ~- ~ ' ) 7-4- ~ l i 4. Since a more detailed analysis for the effects of inelastic energy absorption was conducted in the SSMRP for the Auxiliary Bui.lcing, ] shouldn't the results from this analysis be used to determine the fragility parameters for the concrete shear wall? (pp. 50 and 59) 5. What effect does the absence of perfect dependence havg on the fragility curves 'for piping systems and cable trays (i.e., cable systems)? (pp. 28, 73 and 79) s 6. Since the ductwork and dampers,' batteries and racks, relief tank, and transformers have relatively low capacity values based on generic data, shouldn't specific analyses for these components be performed to develop the fragility curves? l (pp.11, 76,101 and 102) 7. What effect does the coarseness of the data points for the hazard and fragility curves have on the accuracy of the tails of the a probability density function for frequency of core melt? (pp. 5, 6 and 94) 8. What is the basis for the displacement versus acceleration curves shown in Figure 4-7 of Section 7.9.2? (pp. 56 and 75) 9 9. What effect would consideration of a "best estimate" site-specific 1 3 ground respcase spectrum relative to the broad-banded spectrum used in the analysis' have on the value of the factor, F (Section 7.9.3)? (pp. 43, 92 and 93) 'l

10. How close do the electrical components, which were eliminated from Table 7.2-3, compare to the tested components that were used to develop the generic fragility data?

l (pp. 12, 26 and 78) i q e ,e -ei. .wg n g m

{.,[ -~ 8- )- .= 1 j 11. How do design and construction errors and aging affect the fragility 4 curves and the subsequent systems analysis for the effects,,cf seismic j events? l (pp. 10, 39 and 62) i

12. Was the possibility of a LOCA followed by an aftershock, or the

~ 1 occurrence of a moderate earthquake when some safety-related equip-ment is unavailable, considered in the analysis leading'to the ~ probability distribution of frequency of core melt? j (pp.13, 45 and 58) k

13. Could the split of variability into randomness and uncertainty a

components be different than assumed in the analysis. If so, what would be the effect of a different split on the tails of the frequency of. core melt density function? d (pp. 43, 51, 52, 54, 55, 58, 59, 60, 65, 66, 67 and 72) I 1 14. In developing the values for the mode shape parameter for equipment, l was the location of equipment relative to the location of the masses of the building model considered? If yes, how were they considered? n I r (pp. 52, 62, 82, 84 and 89) t

15. Why in Section 7.2.2 were maximum acceleration values assigned, while in the section on the hazard analysis (Section 7.9.1) maximum i

I acceleration values were treated as being uncerta-:n? (pp. 17 and 18) I

16. What is the basis for the median capacity of 1.169 for the reactor j

pressure vessel internals? 1 (p. 67) 1?. What is the basis for the fragility parameter values for the control rod drive mechanisms? (p. 68) e l g. ee J O ~* _s r

• w=

w==* W.

_g. Seismic Hazard (See Page 5.4 and Appendix A Pages A.2 to A.5 of Sandia letter report of March 5,1982) 1. How can the proposed Wisconsin Arch or Wisconsin Arch-Michigan Basin seismogenic zones be justified en the basis of the known seismicity or the deep-seated geological structure? 2. Why shouldn't the cumulative magnitude-recurrence curve determined j - by Nuttii and Herrmann (1978) be used? 3. Isn't the epicenter for the May 26, 1909 earthquake near Aurora, Illinois as given by Docekal (1970)? 4. Shouldn't the best estimate of m be 6.0 and not 5.8? b, max i l s i } 1 i i 1 l i 1i l i i n k 4, l l Il ll

i

'; =..f*: ~~ -- 7 *..

[ '... g Containment and Consecuence Analysis 1. In Section 9 of the ZpSS, you describe features to mitigate the consequences of core-melt accidents. On pages 9.0-3 and 9.0-4, you discuss a limitatinn of the filtered vented containment system, namely that ri.sk reduction from ~ a filtered vent is limited because of seismically induced failures. Wha t would be the r :k reduction from a filtered-vent if it was as robust (rela-tive to seismic events) as the containment building itself, that is, it maintained its performance capability and met its design requirements for those seismic events that dominate risk yet allow the containment building to remain intact? What would be the additional cost for such a seismic qualification? 2 The risk of early fatalities drops by 'two orders of magnitude when going from your point estimate of risk to your level two estimate. As we under-stand it, this reduction is believed to be justiff eil based on the recogni-tion that the source term (release category) and site analysis in the point estimate is conservative. The method used to quantitatively account for the conservatisms is the use of U-factors discussed in Sections 5 and 8. ,1 Some aspects of this approach appear to be justified (e.g., the accounting for the delayed release in the 2R category), but other aspects are unclear. If you are going to quantify the conservatisms of your point estimate and t i display them (as level two curves), then more justification as to how you i l got there is justified. Please provide justificatipn._for the U-factors k displayed on pages 8.6-3 and 4 and include your understanding of the pheno-menology and accident progressions involved. (See ACRS question V.1). .[ x ew w* -+- N 6 999 8"* %8 /Df

__._..~i._._... L... 3. The post-vessel-failure accident progressions are dominated by a bicwout of core materials from the reacter cavity into the containment propr, and c settling out on the containment floor according to the ZPS3. Assuming that 25-50% of the core inventory is so dispersed, how da.you claim that there is virtually no effect on the containment sprays in the recirculation mode? If significant amounts of this core particulate gets into the sumps, the sprays could fail or be seriously degraded. In order for us to better understand your position, please provide: a) the range of particle sizes you expect with a' full justification of why; b) the settling characteristics and final disposition of these core materials within the containment with fan coolers on and with fan i j coolers off; 3 c) the magnitude and composition of particulates that you believe would fail the sprays. t i 4 You clain very low probabilities for basemat penetration (0.02%) in your containment matrix; yet, your analysis indicates that: l a) the reactor cavity will eventually be dry for such damage stages as TE and SE; a l b) a large fraction of the core material will not be blown out of the i reactor cavitiy (in particular, the 50% remaining in the vessel after initial blowdown); j c) basemat penetration times for non-coolable debris _ beds are in the j range of one (1) day. t Please support this claim which appears to be inconsistent with the. points } raised above. Also reference where in Section 4 you perform a dry-cavity j overpressurization analysis. Please provide a containment temperature history a j plot consistent with your dry-cavity overpressurization analysis. u

.3 :. * :.: --in,.m,. -.

a-z-

P j - e-1 I i ? 5. You present three containment matrices: one representing your best estimate; .t 1 ] one, your more conservative estimate; and one, your more optimistic estimate. 1 j When we assess the impact of the differences between the three, we see a j virtually no change in risk. This implies that you have -sufficient knowledge j of containment phenomenology and failure behavior that no important uncer-( tainty exists. We do not share this position. Please clarify whethee we l have interpreted your position correctly. 6. It's not clear to us that the emergency planning assumptions used in the consequence analysis are consistent with emargency planning procedures a presently in place, or planned for the future. Please provide a compar-isen of your assumptions with in-place procedures. i 4 7. We would like to see risks presented as a function of distance for in-4 dividuals, also costs of mitigation and interdiction. Please provide j graphical displays for selected consequence categories. } ~ i i i 8. Are the supportive medical treatment assumptions used to correlate dose ~ versus early' fatality consistent with the available resources to pro-vide such treatment to the exposed rersons? Please explain the bases for 1 1 your ar.swer. i 9. Are the ear)y fatality consequences determined f6r"the release category with the largest release fractions for particulates consistent with the p early. fatality consequences estimates determined for categories with much snaller releas.e fractions? Please explain the bases for your answers. , m s # v *'* * **'

• • • -ye9
    • M. nfM y__' _ Q,4 s 2
• --o
• me==

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

f. -

1 We're not sure how accidents i,nvolving~ ih2 spant fuel pool were con- ~7~.- -~ 10. sidered. Please provide an explanation. 1 .b 4 11. The evacuation assumptions should consider external accident initiators, ] ] or should be shown to not produce substantially different consequence )1 Please provide justification for your assumptions considt r-estima tes. 3 p ing both internal and external accident initiators. 1 12.1 The LASL analyses of the Zion containment shell deformation under pressures ap-A proaching shell failure levels (in the 125 psig range) predict upward vertical displacements of over 3" at the dome apex. This displacement appears to result ' i from combined saucering of the basemat and axial strains in the cylindrical containment wall. Additional horizontal displacements (up to about 1.5" at midheight) occur due to shell distortions because of the radial growth of the s hell. 0 Question: e What potential structural failures could exist (under ] these conditions) in the fuel transfer tube at its point 3 of passage from the containment shell to the spent fuel building? e What effect would structural failures have on potential loss of containment integrity? - ) At what pressure level (below 125 psig) would any failures e .J d first appear? 'l j 12.2 The LASL analysis of Zion indicates radial displacements at the midheight of

• the cylindrical shell of about 1.5" (at about 125 psig). At the elevation level of various groups of piping penetrations in these plants, the combined vertical displacements of perhaps 2" plus radial displacements would re ult in l!

net m:ve. ment of the penetrations of several inches. If the ends of 'any piping ...c ."7 .[......... ~.j,,, ( m _ ~ -+ '-. = t -+- ~,.TMfE y.. ~a _.m,,,

j %_ __ =-. - ? runs (both inside and outside the containment) are substantially restresined 1] (i.e., attached to internal components that do not move with the containment or to adjacent tunnels or buildings) at points relatively near the penetratiun, there is a potential for placing very severe loads on penetration cocponents, k Question: e To what extent are any piping runs restrained so that they cannot move to accommodate shell deformations; what are the anchcage/ support mechanisms for each pipe run and the distances between penetrations and the closest major pipe support ir. side and outside the shell? j Are any majcr penetrations (such as equipment hatch or per-j e 1 sonnel hatch) or purge lines, etc., sufficiently restrairmd ? j outside the containment shell so as to potentially be-I heavily loaded due to the gross shell deformation? What are the loads and strains induced upon penetrations e 1 by shell deformation; what are the probabilities of failures in welds and distortions of hatch seals or valve seats, that could cause loss of isolation capability; can such failures occur concurrently inside and outside the shell (due to } shell distortion) so as to cause loss of dcuble scaling capability or of*any pressurized zones between seals? s. Could the vertical axis of the cont,ainment shell tilt under e 4 the postulated high pressures (as a result of non-syimietrical j saucering of the basemat or restraints imposed by structural i elements below the general level of' the basemat)? Could such tilting increase the loads upon penetration, due to in-creased displacement relative to external connections? ---~_~.,x.,'..,..'_ ,. _.,,,, ;.; ~ ~ ~. . ~, 5 ---.

'= if.f

• r'

-~ ~ ~ ~ us i At what p.essarc 1:.ci are any of the above failures likely j e to first appear t How would any leak paths grow as a func-tion of pressure? Does the extensive cracking that would-occur in contain-e 1 ) ment concrete on the way up toward failure pressure.have 1 any effect on pipe supports that could influ'ence the overall course of the accident sequence (e.g., secondary ruptures in pipes with additional relcase or early release of energy to the containment or failures of isolation valves)? ~ 12.3 Is there any possibility that thermal expansion of piping would cause excessive stresses in piping penetration welds (with or without any allowance' for relief slue to any concurrent growth of the containment shell)? ? i [ Question: e At what temperature would significant weld area 1 for each pipe run? What pipe runs are the most sensitive? J ~ l e What would be the potentia 1 strains in welds; are any strain h levels expected to constitute structural failure or loss of containment integrity? I-i e Are there any accident conditions in which such tempera-1 1 ] ture stresses and strains could occur without the concurrent i j presumed relieving effect of shell growth under pressure? t. 4 ) e Are the strains due to temperature increases in piping and j those due to shcIl distortions geometrically rel'ated (i.e., non-aligned or out-of-phase) in any way so as to be addi-hr tive (rather than tending to relief)? L rf ,_T< ~ ~ '. - W S* ' g : -s e w.,,, ; m.e l,, {., *. l -

-. - ~. -.. ~.. -.. -.. ,j g. -~ 16-12.4 In the Zion plant, the electrical penetrations are apparently all of tlw D. G. O'Brien type, utilizing glass hermetic seals around conductors plus epoxy ;attio.: compounds and ether materials. Limited equipment qualification testing i;'; ars to have been done on these penetration areemblies or the individual materi.'Is. All equipment testing was appar6ntly cca. ducted for pressure and temperature conditions pertinent to design bases currently reflected in licensee sar. ty analysis reports; no tests appear to be available for conditions now being analyzed for more severe degraded core sequences. For the Zion penetrations, qualification tests employed conditions of 265F to 273F for 49 hours at 4G psig. Although an IE review concluded that there was a "high likelihood" of sati:. factory operability under the design conditions, reservations were expressed (cve n for 3 design basis conditions) that, the potting ' materials might be suspect under high p, relative humidity, that the tests did not adequately simulate saturated :ti.am 11 conditions, that epoxies and polymer materials have shown aging effects und:Er 1 combined accident conditions of temperature, irradiation, and water / spray chem-istry, and that some uncertainty about service life may exist for conductor in-j

,l I

l! sulation and jacket materials. d 1 What are the specific materials used for assuring contain-Question: e j ment integrity? e What mean servicb life and standard deviation data exist for Zion penetration assemblies under combined severe accident conditions of concurrent temperat'uie, pressure, radiation, steam, and water / spray chemistry? Can any existing 1 ire i service data be extrapolated to the potentially more.c eri-l conditions of degraded core events? i 9 D*6*6 e6 +4 6 6M** 4h6 9 me 6 a.a p.

• M**#

E# 66*4 L -'N-Y. *1 1_U - m% N_. m.' MS 'Y

• I_

T _:. ' -' ^ ~ ^* ' ~** I

..~..~ca...--+-........_... k .4 K

s. w e e Since many electrical penetrations exist in Zion, to what extent can mean life and standard deviation be relied on to guarantee that no penetration will fail under accident conditions? Since only a few assemblie,s were tested, what t

is the confidence level in the available service life data? e If the severe accident temperatures do reach the 400F range, what is the probability of failure of scalants in several penetrations? What is the likely number, nature and extent of such failures? e Is there any possibility of synergistic effects on sealant failure as a result of the major displacements that occur in the concrete shell? ) e If internal isolation barriers seals) fail, what is likely to be the gradient of accident conditions toward the outer barrier? Can any estimate be made of the additional time l to failure of consecutive barriers? Can any pressurized j zones between seals be maintained in the event of failure i i of either barrier? h ~ l Are there any significant ameliorating effects (on penetra-s i tion assembly and seal life) due to the existence of large heat sinks nearby or are any portidns of penetration seal assemblies too exposed to containment ar.cident conditions to benefit from such effects? 1 I l .y. . = -. ~ ~ ~ - - ~wS --.~, -ly, h,., n, s o. ;.. m ,w p,s, p -.- -. - - - - - - - ~ - -e

=... _ _ _ _ -..,.;_._._-. -. -. - w. G i 4.t.. v. -18 1 12.5 For Zion, as ter other reactor plants, it is presumed that large double isolation valves have elasteneric seats which 1. ave been qualified for selected design basis accident conditions, that large hatches have double inflatable seals v.hich are similarly qualif ted, and that, for both valves and hatches, the zones between the double seals are pressurized to prevent 1catage. Conversations with ranu-facturers indicate that commonly used. lastomers may not be good for tempera-tures significantly higher than design basis conditions and that even high tempera-ture alter natives (e.g., fluorosilicone) may not be suitable for steam conditions. What are the resistances of Zicn seat and seal materials to Question: e higher temperature (up to 450F) with concurrent adverse conditions of steam, water / spray chemistry, radiation and ~ pressure? What is, the average service life under such conditions? Are alternative seal materials available that could resist e hat are both higher temperature and steam concurrently? their probable service lives and related confidence levels? In the event of failure of an inner barrier, can the pres-e surized zone be maintained? If it cannot be maintained, I what would be the effect on containment isolation integrity? What effect wilk failure of an inner ba.rier have upon loads imposed upon outer barrier; what ar.: the time parameters for loads imposed on outer barriers? m 9 .o. -.. = -.... -. _ _ .****~wY 7 MT ~~'"P ~.

] l 1. Identify and summarize any overcooling events that have occurred, where the cooldown rates were determined to have exceeded the limits specificed in plant procedures or technical specifications. Provide the results of any post event evaluations. 2. Provide a discussion on the limits, procedures and operator training for terminating operation of the r.ain or auxiliary feedwater pumps and/or safety injection pumps that will preclude a pressurized thermal shock event assuming an accident or other condition resulting in a cooldown rate exceeding the allowable limits is in progress. I 3. Provide a discussion on the operator guidelines, operating procedures and operator / shift technical advisor training program on pressurized thermal shock. Include a description of any testing given following training. 1 4. Identify and summarize the results of any post installation reactor vessel (internal) wall inspections. 5. Identify the instrumentation available to the operators to assist them in recognizing a potential pressurized thermal shoc'. event. 6. Identify and summarize the results of any accident analysis, transient analysis or probabilistic risk assessment study performed that could be applied to the pressurized thermal shock issue such as control system failures, steam line break accidents, small break loss of coolant accidents or other conditions resulting in a cooldown which exceeds the allowable rate limits and/or results in inside vessel wall fluid temperatures less than 300*F, with a potential for systems pressurization. 7. Provide a discussion of the actions taken thus far to lessen the probability and/or severity of pressurized thermal shock events. 8. Provide a discussion and schedule for implementation of any actions planned to resolve the pressurized thermal shock concern, such as I fuel management programs aimed at fluence reduction, increased ECC injection water temperature, and additional instrumentation. f o 'I i ~ . - n..,- n, .}}