Text

Slides, Summary of January 25, 2007
	ML070570103
Person / Time
Site:	Kewaunee
Issue date:	02/26/2007
From:	; Dominion Energy Kewaunee
To:	; Office of Nuclear Reactor Regulation
	R. Kuntz LPL3-2
Shared Package
ML07050083	List: ML070570103;
References
	TAC MD1194
	Download: ML070570103 (43)
	v • d • e

Agenda Introduction Overview Purpose of this Meeting Kewaunee Issue History TORMIS - Overview Kewaunee Submittal Technical Review NRC Staffs Questions Dominion Energy Kewaunee Response to NRC Staff Questions Conclusions Questions and Answers

Introduction Introduction of Dominion Energy Kewaunee Attendees

Meeting Purpose Dominion Energy Kewaunee, Inc requested this meeting to :

1. Explain how the Kewaunee submittal is in compliance with the NRCs Approved Safety Evaluation for TORMIS

2. Show how the submittal is consistent with other approved submittals

3. Reveal the Conservatisms in the analysis

4. Answer the NRC staffs Questions

Kewaunee Issue History March 2005, plant walkdown for extent of condition on tornado issues identified that the EDG exhausts were not designed for tornado winds.

April 2005, during evaluation of the EDG exhaust ducts, identified that the ducts were not protected from tornado missiles.

Extent of condition also identified that the EDG fuel oil vent lines were not protected from tornado missiles.

June 2005, design changes implemented to shorten the EDG fuel oil tank vent lines to reduce exposure to a tornado missile and to reinforce the EDG exhausts.

July 2005, TORMIS evaluation completed for EDG exhaust ducts and fuel oil vent lines showing that a damaging tornado missile strike was not credible (< 1E-6).

April 6, 2006, LAR submitted to add use of TORMIS methodology for the EDG exhaust and fuel oil vent lines to the KPS USFAR.

The tornado extent of condition walkdown identified that the service water lines to the control room air conditioning were also susceptible to tornado missiles. TORMIS evaluation determined that a damaging missile strike was credible. The plant was modified to add a tornado missile shield around these lines.

October 24, 2006, received a request for additional information from NRC. Responded to the request on November 9, 2006.

December 21, 2006 discussed additional questions on the TORMIS submittal in a phone call between NRR and DEK.

EDG A Exhaust - Turbine Floor

EDG B Exhaust - Turbine Floor

EDG Fuel Oil Storage Tanks Vents

EDG Fuel Oil Tanks Vents - Closeup

These criteria generally specify that safety related systems be provided positive tornado missile protection (barriers) from the maximum credible tornado threat.

SRP Section 3.5.1.4 includes acceptance criteria permitting relaxation of the above deterministic guidance if it can be shown that the probability of damage to unprotected essential safety related features is sufficiently small.

NRC has approved the use of TORMIS via an SER dated November 29, 1983 for determining tornado missile damage probabilities.

The SER contains five questions the user must address in their submittal The use of TORMIS should be limited to the evaluation of specific features where additional costly tornado missile protection barriers or alternative systems are under consideration.

1.

The total impact and damage probability to each structure from all the postulated missiles.

2.

The probability that a combination of targets (e.g.,

redundant components) will be damaged during a tornado strike.

3.

The impact and damage probabilities of the entire plant.

4.

The barrier thickness corresponding to specified levels of impact and damage risk.

5.

Overall structural response damage probabilities through a impact velocity exceedance criteria in which any postulated missile that impacts a target with velocity greater than VDAM(I) results in damage to the target.

Emergency Diesel Generator Exhaust Ducts z

Emergency Diesel Generator Fuel Oil Storage Tanks Vents Follows EPRI Methodology

Damaging hit on either EDG exhaust or one or more EDG fuel oil tank vent lines is assumed to cause a loss of function.

2.

The EDG is assumed to be fully loaded vs. blackout loaded when determining the allowable crush.

3.

We did not model the Turbine Building structural steel or any of the large components in the Turbine Building that will provide shielding.

4.

All non-tornado proof buildings were treated as missile sources and provided no shielding to the targets. Our modeling also treated roofs as particular elevation sources for the major plant buildings near the targets.

5.

The plant was in a forced outage during the missile survey walkdown. The number of potential missiles on the Turbine Building operating floor used in the analysis was much higher than typical.

6.

Conservatively blended the EPRI Region B data into the Kewaunee tornado subregion path width

Treated all the plant potential missiles as N missiles (maximum transport injection mode) and none of the missiles as N missiles (first exceedance injection mode).

8.

We used a full spectrum of missiles, including missiles with good aerodynamics, as well as the heavy NRC missiles with relatively poor flight characteristics.

9.

Used conservative missile injection heights for all missile sources.

10.

Exhaust vents are assumed to be crushed without tearing, shearing, or perforation for each missile impact.

Is the V0 / V33 Tornado Wind Field Velocity Ratio a Fixed or Variable Input to TORMIS?

2.

In the TORMIS study, were comparisons made of broad tornado regions to small areas around the plant and were the most conservative values used in the DEK study?

3.

Were Variance reduction techniques used in the Kewaunee TORMIS analysis?

4.

How many missile types can be used in TORMIS? Does TORMIS have 15 or 22 missile types?

5.

How was the tumbling of missiles considered in modeling targets?

6.

Were missile impact velocities used for crushing of the EDG exhausts and EDG fuel oil tank vent lines? Are impact velocity thresholds an input to TORMIS?

7.

Were the turbine building walls considered to resist impact? How was this treated in TORMIS?

8.

How did DEK arrive at the number of missiles used in the TORMIS Analysis? Are the number of missiles used consistent with other nuclear plants?

9.

How did DEK model openings in structures?

DEK Response: Fixed The V0/V33 parameters, alpha () and zeta () are input parameters to TORMIS. Following the NRC review of TORMIS, the default parameters were identified as =10 and =30.

These values of alpha and zeta were the values used in the Kewaunee study and are a fixed ratio for all the tornado simulations.

DEK Response: Yes A Kewaunee-specific subregion was developed for the plant analysis (Figure 1).

The one degree square block that contains the plant is near the mean of the cluster. (Figure 2).

The Kewaunee TORMIS analysis used a Final Point Probability of 5.47E-4 (Yellow Line in Figure 3).

The analysis procedure automatically compares small areas to broad areas and combinations of areas in between.

Point Probability (per yr)

Plant 1 deg Block 1.00E-06 1.00E-05 1.00E-04 1.00E-03 1.E-05 1.E-04 1.E-03 Detrended Occ Rate (tor per sq mi per yr)

Point Probability (per yr)

Plant 1 deg Block Plant 1 deg Block

© 2003 Dominion 22 Figure 3 - Adjustments to Kewaunee Subregion Data Kewaunee Subregion 1 x 1 Degree Data and Final Adjusted TORMIS Point Probability 1.00E-06 1.00E-05 1.00E-04 1.00E-03 1.E-05 1.E-04 1.E-03 Detrended Occ Rate (tor per sq mi per year)

Point Probability (per yr)

Kew 1 deg Subregion PP Final Adjusted PP

© 2003 Dominion 23 Figure 4 - Comparison of Final Kewaunee Tornado Hazard Risk Curve to Best Estimate Kewaunee Hazard Curve with No Conservative Adjustments 1.00E-09 1.00E-08 1.00E-07 1.00E-06 1.00E-05 1.00E-04 1.00E-03 1.00E-02 50 150 250 350 Windsped (mph)

Windspeed Exceedance Probability (per yr)

Final Kewaunee No adjustments and F' Winds

DEK Response: Yes the following variance reduction methods were used.

z Importance Sampling on Tornado Offset Position Relative to Plant Center z

Stratified Sampling on Missile Initial Position z

Importance Sampling on Missile Injection Height z

Russian Roulette on missile trajectories that are headed away from plant These techniques are typical of those used for other TORMIS evaluations.

ZIMR = a We use ZIMR = 2 a

Missiles exiting MTC are killed with Prr = 0.5

DEK Response:

There is not a limitation of 15 missile types in TORMIS.

A basic missile spectrum of 26 missile aerodynamic sets was developed in the TORMIS research. See Appendix 2 of EPRI NP-769, Tornado Missile Risk Analysis-Appendixes.

Based on the Kewaunee Plant survey, a total of 22 missile subsets were modeled. The types of missiles that were modeled are based on numerous plant surveys and the Kewaunee survey which produced similar types and sources of missiles.

(continued)

DEK Response: The potential for offset hits was considered in the modeling of all targets of concern at Kewaunee. The dimension of each free surface of each target was increased per the original analysis in Appendix Section 4.2.3 of EPRI NP 769. A one foot increase in each free surface was used.

Hence, a target with two free surfaces in the X direction was increased in X direction length by 2 feet, etc.

DEK Response: Impact velocity thresholds were used for the crushing of the exhaust vents and fuel oil tank vent lines.

TORMIS was developed to consider hits, penetration, perforation, scabbing, and overall structural response failures.

The NRC SER states that the EPRI methodology employs Monte Carlo techniques in order to assess the probability of multiple missile strikes causing unacceptable damage to unprotected safety-related plant features. (Page 2)

In the University of Chicagos TER, under Impact Assessment, it states that from the standpoint of general missile risk assessment, the important feature is the capability of estimating probabilities that each of the postulated missiles strikes each structure with a velocity greater than some specific value. The combination of these probabilities for various velocity thresholds can be used to produce estimates of missile impact velocities for the different missile types which can be used to design for the desired probability level.

(cont)

Overall structural response failures are determined by separate finite element analyses.

Determined allowable reduction in cross-sectional area Finite element analysis done for each missile type performed to determine threshold velocity Critical impact velocity VDAM(I) input to TORMIS for each missile type.

The impact velocity threshold approach has been used for other Nuclear Plant TORMIS analyses for similar types of targets.

(continued)

The following conservatisms were used in the impact velocity evaluation.

The EDG was assumed to be fully loaded vs. blackout loading when determining allowable back pressure.

No perforation or tearing was assumed, only crushing.

A damaging hit on one or more EDG Fuel Oil Vent resulted in loss of function.

A damaging hit on either EDG exhaust vent resulted in loss of function. (Appendix Table D-3). (No Credit for Train Separation)

DEK Response: Yes The turbine building walls are metal panels with a design pressure of 63 psf.

Using ASCE 7-02, we estimated that these panels have a failure peak gust wind speed of about 165 mph (>F1 & F2 Tornados).

In TORMIS the panels were assumed to fail in F3 winds, greater than 158 mph.

Reasonable Because Missiles penetrating walls would lose a large amount of energy.

No wind inside building to provide accelerating forces to missile.

Steel columns, beams, and heavy equipment would provide some shielding.

Our modeling of the turbine building was extremely conservative for F3 and higher tornadoes 9

We assumed no metal panels would be in place.

9 The building was not present and the diesel vents stick up like two tall columns, totally exposed to all missile impacts 9

Tornado winds were able to blow through the building space as if no building were present.

Our modeling of the turbine building for F1 and F2 tornadoes assumed the turbine metal panels remained in place and shielded the vents from impacts from the debris that is produced in those lower windspeeds.

DEK believes the conservatism listed will offset the probability that a missile will penetrate the siding with enough energy to damage the exhaust duct.

DEK Response: A total of 27,826 missiles were modeled for the Kewaunee analysis. This number was based on a site-specific walkdown in 2005 during a plant forced outage.

The site specific walkdown surveyed each area, including the inside of the turbine building, and counted and cataloged the number of missiles present. It also reviewed drawings of plant structures to determine the number of potential missiles generated upon failure of the structure.

The approach was consistent with Section 4.2 of the TER.

(Continued)

Results of 14 plants surveyed by the TORMIS Developers between 1975 and 1998 determined that on average there were 29,532 missiles per reactor.

Sites with an operating reactor and a reactor under construction had more missiles.

See Table 4.

© 2003 Dominion 39 Table 4. - Results of all TORMIS Nuclear Plant Surveys Conducted by TORMIS Developers Survey Year Nuclear Plant Number of Missiles Number of Reactors Avg Per Reactor Reactor under Construction 1975-1976 Plant 1 22,766 2

11,383 no 1975-1976 Plant 2 65,685 3

21,895 yes 1975-1976 Plant 3 186,322 2

93,161 yes 1975-1976 Plant 4 64,700 2

32,350 no 1975-1976 Plant 5 9,294 1

9,294 no 1975-1976 Plant 6 2,918 1

2,918 no 1975-1976 Plant 7 4,538 1

4,538 no SubTotal 356,223 12 29,685 1998 Plant 8 51,864 2

25,932 no 1984 Plant 9 66,797 2

33,399 yes 1988 Plant 10 37,746 1

37,746 no 1982 Plant 11 10,730 1

10,730 no 1985 Plant 12 41,733 3

13,911 no 1983 Plant 13 21,453 1

21,453 no 1984 Plant 14 122,225 2

61,113 yes SubTotal 352,548 12 29,379 Grand Totals 708,771 24 29,532 2005 Kewaunee 27,826 1

27,826 no Avg/Reactor for Plants with no Construction = 17,849 Avg/Reactor for Plants with Construction = 52,391

DEK modeled all nearby buildings that were built to conventional construction standards as open buildings (imaginary walls). This means they provided no shielding and were modeled with a floor slab on the ground with missiles sources above the slab. For many such buildings, we also modeled the roof as a separate source of missiles, so they could be generated from the appropriate elevation.

DEK modeled the turbine building as open for F3 and higher tornadoes.

v t e Letter Sequence Other
TAC:MD1194, (Withdrawn, Closed)
Initiation Acceptance... Administration Meeting Questions 11 October 2006 Answers Results Withdrawal Other: ML070570103
MONTHYEAR2007-02-26 [Table View]

ML070570103

Text

Navigation menu

Search