ML20058F078

Requests Response to Encl List of Questions Re Internal Event Analysis in IPE & Containment Performance Improvement Program,Within 60 Days of Ltr Receipt

ML20058F078
Person / Time
Site:	Sequoyah
Issue date:	11/24/1993
From:	Labarge D Office of Nuclear Reactor Regulation
To:	Medford M Tennessee Valley Authority
References
GL-88-020, TAC-M74468, TAC-M74469 NUDOCS 9312070310
Download: ML20058F078 (9)

Text

t t

Docket Hos. 50-327 and 50-328 Tennessee Valley Authority ATIN:

Dr. Mark 0. Medford, Vice President Technical Support 3B Lookout Place 1101 Market Street Chattanooga, Tennessee 37402-2801

SUBJECT:

REQUEST FOR ADDITIONAL INFORMATION RELATED TO THE SEQUOYAH NUCLEAR PLANT UNITS 1 AND 2 INDIVIDUAL PLANT EXAMINATION SUBMITTAL (TAC NOS. M74468 AND M74469)

By letter dated September 1,1992, the Tennessee Valley Authority (TVA) submitted the Individual Plant Examination (IPE) for the Sequoyah Nuclear Plant Units 1 and 2 as required by Generic Letter 88-20.

Based on our review of the submittal, we have generated a list of questions related to the internal event analysis in the IPE and the containment performance improvement (CPI) program. The questions are designated according to their relationship to the IPE review process as General, Front End (F.E.), Back End (B.E.), and Human Factors (H.F.).

In order to complete our review of the submittal, we request that TVA respond to the attached questions within 60 days of receipt of this letter.

This requirement affects nine or fewer respondents and, therefore, is not subject to Office of Management and Budget review under P.L.96-511.

Sincerely, Original signed by David E. LaBarge, Sr. Project Manager Project Directorate II-4 3000.-

5. a Division of Reactor Projects - I/II Office of Nuclear Reactor Regulation

Enclosure:

Questions on Sequoyah IPE cc w/ enclosure:

See next page Distribution Docket File NRC & Local PDRs SQN Reading S. Varr,a G. Lainas F. Hebdon B. Clayton D. LaBarge OGC ACRS (10)

E. Merschoff, RII P. Kellogg, RII R. Crlenjak, RII C. Ader, RES J. Lane, RES 0FC:

PDII-4/LAa )

PDII-4/PM/ [ PDIIe4/D NAME:

BClayt W DLaBarge[

FHEEiio'n DATE:

N/8/93 l',

DOCUMENT NAME: G:\\SQN\\74468.RAI

)i /9/93 9312070310 931124 0" = " F#

EC RE WTdiCOPY$

3

I

~

Tennessee Valley Authority SEQUOYAH NUCLEAR PLANT ATTN: Dr. Mark 0. Medford cc:

Mr. Craven Crowell, Chairman TVA Representative Tennessee Valley Authority Tennessee Valley Authority ET 12A 11921 Rockville Pike 400 West Sumit Hill Drive Suite 402 Knoxville, TN 37902 Rockville, MD 20852 l

Mr. W. H. Kennoy, Director General Counsel Tennessee Valley Authority Tennessee Valley Authority ET 12A ET 11H 400 West Sumit Hill Drive 400 West Sumit Hill Drive Knoxville, TN 37902 Knoxville, TN 37902 Mr. Johnny H. Hayes, Director Mr. Michael H. Mobley, Director Tennessee Valley Authority Division of Radiological Health ET 12A 3rd Floor, L and C Annex 400 West Summit Hill Drive 401 Church Street Knoxville, TN 37902 Nashville, TN 37243-1532 Mr. Robert Fenech County Judge Site Vice President Hamilton County Courthouse Sequoyah Nuclear Plant Chattanooga, TN 37402 Tennessee Valley Authority P.O. Box 2000 Regional Administrator Soddy, Daisy, TN 37379 U.S. Nuclear Regulatory Comission Region II Mr. R. M. Eytchison, Vice President 101 Marietta Street, NW., Suite 2900 Nuclear Operations Atlanta, GA 30323 Tennessee Valley Authority 3B Lookout Place Mr. William E. Holland 1101 Market Street Senior Resident Inspector Chattanooga, TN 37402-2801 Sequoyah Nuclear Plant U.S. Nuclear Regulatory Comission i

Mr. B. S. Schofield, Manager 2600 Igou Ferry Road Nuclear Licensing and Regulatory Affairs Soddy Daisy, TN 37379 Tennessee Valley Authority 4G Blue Ridge Mr. D. E. Nunn, Vice President 1101 Market Street Tennessee Valley Authority i

Chattanooga, TN 37402-2801 3B Lookout Place 1101 Market Street Mr. Ralph H. Shell Chattanooga, TN 37402-2801 Site Licensing tianager S9quoyah Nuclear Plant Tennessee Valley Authority P.O. Box 2000 Soddy Daisy, TN 37379

i

^

ENCLOSURE i

RE00EST FOR ADDITIONAL INFORMATION SE000YAH INDIVIDUAL PLANT EXAMINATION i

GENERAL 1:

For the system walkdown conducted for the internal flood analysis, Sequoyah-specific documents were used. However, in order to familiarize yourself with the system configuration for Interfacing System Loss of Coolant Accident (ISLOCA) analyses a walk-through was conducted at the TVA Watts Bar facility (single unit), instead of the Sequoyah plant (two-unit site).

Please explain why the Watts Bar walkdown was used in one instance while the Sequoyah one was used in another. Discuss the appropriateness of the Watts Bar walkdown for this IPE submittal.

GENERAL 2:

The IPE submittal described various reviews performed by different review teams as well as the areas of review.

It concluded that the peer review did not identify major findings. However, details of the review results are not included in the submittal.

Please provide the results of the review, consistent with Section 2.4 of NUKEG-1335, which stated i

that as a ainimum, details of the review should be included in the submittal.

i GENERAL 3:

i Section 6.4 discusses insights and recommendations that you view as 1

potential enhancements.

Please discuss the status of these insights and recommendations, vis-a-vis actual implementation at Sequoyah.

If i

relevant, provide a schedule for the evaluation and/or implementation of these issues.

F.E.1:

In Section 3.1.1 the IPE discussed the importance of the plant HVAC systems, and stated that they are not significantly important as an initiating event because the loss of the systems is weil annunciated and i

heatup calculations show there is ample time for the operator to restore power prior to equipment failure causing a plant trip.

Please elaborate on the basis for this including a brief summary of the heatup calculations.

LER 25641 (6-11-93) identified the existence of redundant trains of vital power components which were inadvertently cooled by a single train of HVAC.

Presumably, this configuration could have been detected in the plant IPE walkdown.

Please discuss the impact of this finding on your conclusions regarding the loss of HVAC.

I i

J

t i

2 F.E.2:

l Section 3.1.1 states that the success criteria for plant safety

{

functions were determined 'on the basis of the engineering knowledge, and it further states that NUREG/CR-4550 analysis of Sequoyah was referenced for the core heat removal requirements of medium LOCAs (2-6").

The success criteria for medium LOCAs for Sequoyah requires one out of four HP recirculation and one out of two LP recirculation for late core heat removal according to Table 4.4-8 of the Sequoyah NUREG/CR-4550.

However, the IPE success criteria for Sequoyah medium LOCA requires only one out of two LP recirculation (Table 3.1.1-3).

Please explain the i

differences in conjunction with the FSAR criteria, and the basis for the success criteria.

F.E.3.

Loss of the control air system is not included in the support and inter-system dependency analysis. However, the submittal did include in i

the containment analysis the portion of the air system classified as

" safety-related." Please identify the "non-safety-related" portion of the system, and explain the basis for the exclusion.

F.E.4:

One of the FSAR success criteria for RHR injection is availability of i

two-out-of-four paths for 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />. However, IPE Table A-2 states that only one-out-of-four is acceptable.

Please explain the apparent.

j discrepancy.

I F.E.5:

~

Please clarify the role of accumulators for large and medium LOCAs. The success criteria for accumulators during large LOCAs was not modeled in NUREG/CR-4550.

Please provide a brief rationale for the difference in success criteria, vis-a-vis accumulator operation, between large'and-l small LOCAs.

i F.E.6:

i IPE Section 3.4.1 states that during the accident sequences with total i

loss of component cooling system and failure to trip RCP, the RCP seal l

was assumed to fail in 10 minutes due to loss of lubrication and i

subsequent vibration. What is the basis for the 10 minute failure i

criterion?

J F.E.7:

Differences between the accident sequences 1 and 3, and sequences 2 and 5 are due to refilling of the RWST. The sequence de s riptions imply

.l that sequences 1 and 2 successfully provide makeup water for the RWST refill, while sequences 3 and 5 fail to provide the refill water to the tank. However, no specific descriptions of the RWST refill. operations were given for the sequences 1 and 2.

In fact, it appears from other information that sequences 1 and 3 are identical.

Please clarify the sequence modeling.

1

I 3

F.E.8:

The ATWS contribution to core damage frequency in the submittal is more l

than three times higher than the NUREG-ll50 study for Sequoyah, and five j

times smaller than that for the Watts Bar IPE.

Please provide a brief l

discussion on these relatively wide disparities.

f.E.9-Please provide the rationale for not modeling feed-and-bleed as an i

alternate decay heat removal function, or alternatively employing steam generator depressurization and feedwater.

F.E.10:

The Sequoyah reactor coolant pump seal LOCA model in the NUREG-ll50 study stated that loss of seal injection and thermal barrier cooling would lead to leaks greater than 250 gpm and eventually to seal j

failures. The IPE submittal (Section 3.4.1) states that the RCP seal will fail under excessive vibration caused by the failure of the lubricating cooling upon loss of the component cooling system.

Please i

explain these apparent differences in RCP seal LOCA modeling.

In addition, did the Sequoyah IPE consider seal failures subsequent to pump trip?

If so, explain the model used in the analysis.

4 F.E.ll:

For the internal flood analysis, does the equipment failure model l

consider failure due to spray effects as well as submergence-induced failure? Please describe it in greater detail than what'has been provided in the submittal.

B.E.1 (Figure 2-18)

Please explain the curves in the figure.

Is total failure supposed to

}

be the probabilistic sum of leak plus gross failure? Is the horizontal scale "psig"? What are the definitions of leakage failure, gross failure and total failure?

B.E.2 (Page 4.1-6):

Air return fans are assumed operational in the Level 2 Sequoyah analysis.

List the components of the air return system located inside containment tht must operate successfully in the severe accident environment.

Discuss how these individual components were assessed for operability during this harsh environment and summarize the sequences for which the air return fans were assumed failed.

B.E.3:

Table 4.4-1 lists the median failure pressures for various types of containment failures. The standard deviations of these values are also provided. However, these standard deviation values appear to be i

incredibly small in comparison to the median values listed, implying an extremely tight data population.

Please explain why these values are so unusually small.

i

l 4

B.E.4 (Page 4.7-4):

Case ENIll resulted in hydrogen concentrations of 8.5 percent yet no combustion occurred in containment.

Describe the hydrogen ignition criteria used in that scenario and the likely containment response had ignition occurred. The occurrence of ignition in a case like this might have altered your determination that back-up power to the igniters are not required.

B.E.5 (Page 4.7-12):

For plant damage state LNI it is considered that deflagration-to-detonation transition (DDT) could occur in the ice condenser to upper i

plenum region. However, at the same time you indicate that this phenomena was not modeled because it was not considered very likely.Please clarify the DDT conditional probability of occurrence and if it was zero, justify in light of the fact that all that appears necessary, given the hydrogen concentration present, is an ignition source which might be provided by metal door contact.

i B.E.6 (Page 4.8-5).

Section 4.8.1.2, Hydrogen Phenomena, discusses the phenomena of hydrogen

'i

" recombination" as advocated by IDCOR. This phenomena had not been i

given wide credibility previously by NRC staff or consultants.

Discuss the extent to which the assumption of occurrence of this phenomena impacted your general treatment of hydrogen combustion in the Sequoyah IPE.

B.E.7 (Page 4.8-21):

In discussing top event 11 you indicate that containment performance will not be affected by small leaks.

However, source term release would be affected and you indicate that " specific release categories were identified for Sequoyah to account. for the possible increased source terms due to small leaks prior to vessel breach."' What are these early release categories? Have they been added or included in the results for i

Release Category II, or are they simply available to be added later?

B.E.8:

Figures 4.8-3 through 4.8-7 appear to indicate that some of the MAAP runs were " borrowed" from Watts Bar.

If so, please reiterate the basis for your conclusion that no significant differences exist between the two plants.

B.E.9:

Please interpret the MAAP DCH predictions provided in Figure 4.8-10.

At the vertical line representing approximately 18 psig, what probability is predicted? Is there is a probability distribution associated with a range of possible pressures, as in NUREG/CR-4551 or dces MAAP predict a single value?

B.E.10:

Is the containment failure probability curve provided in Figure 4.8-13 taken from actual Sequoyah data or is it a representation of such a curve? (It does not appear to be very similar to'the actual curves, theraselves, such as Figure 4.8-19.)

r t

1 9

5 B.E.II (Page 4.9-6):

Clarify the meaning of Release Category IIIA as opposed to Category III.

B.E.12 CEI Top Event 3 indicates that as a result of detailed thermal-hydraulic j

analyses scenarios that were originally assumed to result in core damage.

were shown not to do so. However, Appendix C (pg. C-1) indicates that the discrepancies existing between detailed thermal-hydraulic i

calculations and the original best-estimate assumptions " generally have an insignificant impact on the evaluation results." Please clarify this apparent contradiction.

i H.F.1-The SLIM methodology (NUREG/CR 3518) uses experts to assign scores (the l

degree to which the performance shaping factors (PSF) effect the likelihood of success of an action) and weights (the relative influence f

of each PSF on the likelihood of the success of the action) to specific actions. These scores and weights are combined to produce a success l

likelihood index (SLI).

In this methodology, log (HEP)-a(SLI) + b.

In the IPE, however, the SLIM methodology was modified so that the i

operators scale the degree of difficulty, rather than the potential for success, when they score the action. This score, multiplied by the j

associated weight, results in a failure likelihood index (FLI) rather than a SLI.

In this approach, log (HEP)-a + b(FLI).

However, it is not apparent how ratings based on " degree of difficulty of an action" will result in an accurate measure of failure likelihood.

It would seem more appropriate to have experts score " likelihood of failure of an action" rather than " degree of difficulty of an action" to obtain an accurate measure of failure likelihood.

In addition, having an expert score " degree of difficulty of an action" may elicit a different rating than if a rating was obtained by having an expert score

" likelihood of failure of an activn." This difference could result i

because " degree of difficulty" and likelihood of failure" are not i

equivalent constructs.

Further, " degree of difficulty" is not the inverse of " likelihood of success." Therefore, stating that your modification of SLIM will result in a measure of the FLI appropriate for use in the SLIM model is questionable. As a result, your measure of the FLI may not be an accurate measure of failure likelihood, but simply a measure of degree of difficulty.

i Based on these considerations:

1) discuss whether this modification has been used elsewhere in the industry, and if so, provide references,

2) discuss whether this modification of SLIM received an in-house peer review, 3) provide the rationale on how your FLI is a true measure of failure likelihood rather than simply a measure of degree of difficulty.

t H.F.2:

The IPE states that a system analyst was used to evaluate the routine actions before initiating events.

Please discuss if this system analyst held any discussions or interviews with plant personnel (1.9.,

operations, training) on the interpretation and performance of the required tasks during these evaluations, i

i

+.

6

~

H.F.3:

Please discuss how you addressed the effects of dependencies, such as the adequacy of training and procedures, plant conditions, and human engineering on pre-initiator human events (e.g., miscalibration, failure to restore),

in addition:

1) provide specific examples of the dependencies included in the analysis, 2) fully discuss how these dependencies were identified (i.e., observations, interviews), and

3) describe how these dependencies were included and addressed in the analysis.

H.F.4:

The IPE does not explicitly discuss how miscalibration errors were included in the analysis of routine errors before initiating events.

Please provide a brief discussion describing how human errors from miscalibration were included in the analysis.

H.F.5:

In the IPE, you briefly address the recovery model used to modify generic error rates during the evaluation of routine actions before initiating events.

Please discuss:

1) the types of plant-specific performance shaping factors (i.e., post-maintenance tests, administrative controls) used to modify the generic error rates, and provide the values used, and 2) what process was used to determine the appropriateness of applying the performance shaping factors to each separate component.

H.F.6:

The recovery actions were qualitatively and quantitatively evaluated by only one group of licensed plant operators while the dynamic human actions were evaluated by three groups of operators.

Please discuss:

1) why only one group was used to assess the recovery actions, and

2) how using only one group versus three could have affected the results obtained.

H.F.7:

The IPE states that three groups of operators were used to evaluate the dynamic human actions.

You also stated that rating differences were observed between the groups on several evaluations.

Please discuss:

1) if all operators were from Sequoyah, and 2) if any differences in qualifications and experience of operators within the groups could have led to the observed rating differences between the groups.

H.T.8:

Recovery actions typically are plant specific.

The IPE stated that the evaluation of the recovery actions used licensed operators from both Sequoyah and Watts Bar. Even though Watts Bar is a similar plant, please discuss your rationale for using operators from both plants rather than from just Sequoyah.

H.F.9:

Post-accident human events involve those actions performed by the operator after the onset of core damage.

Please discuss how you identified, screened, and quantified these types of human events.

o.

~i a

7 i

H.F.10:

In IPE Section 1.4.4, " Comparison of Sequoyah IPE with Sequoyah NUREG-IISO," a comparison of the Watts Bar IPE, the Sequoyah IPE and the Sequoyah NUREG-IISO PRA is provided.

You state that differences in the results between the Sequoyah IPE and the Watts Bar IPE are somewhat attributable to the differences in the assessment of operator actions, l

based on input from operations department of each station.

Please discuss:

1) these differences and the subsequent affect on the IPE 6

results, and 2) how the current assessment of operator actions differed i

from the assessment and analysis in NUREG-ll50.

I i

5 J

l l

l

)

1

- -