ML021160410

Safety Significance Assessment of the Davis-Besse Nuclear Power Station, Unit 1 Reactor Pressure Vessel Head Degradation - Non-Proprietary Version

ML021160410
Person / Time
Site:	Davis Besse
Issue date:	04/08/2002
From:	Bergendahl H FirstEnergy Nuclear Operating Co
To:	Dyer J Region 2 Administrator
References
1-1268 W-DB-01Q-301, Rev 1
Download: ML021160410 (51)

Text

FENOC5501 Davis-Besse Nuclear Power Station North State Route 2 Oak Harbor,Ohio 43449-9760 FirstEnergyNuclear OperatingCompany Howard W. Bergendahl 419-321-8588 Vice President- Nuclear Fax: 419-321-8337 Contains Proprietary Information Pursuant to 10 CFR 2.790 Docket Number 50-346 License Number NPF-3 Serial Number 1-1268 April 8, 2002 Mr. J.E. Dyer, Administrator United States Nuclear Regulatory Commission Region III 801 Warrenville Road Lisle, IL 60532-4351

Subject:

Safety Significance Assessment of the Davis-Besse Nuclear Power Station, Unit 1 Reactor Pressure Vessel Head Degradation Ladies and Gentlemen:

As part of the investigation of the degradation of the Davis-Besse Nuclear Power Station, Unit 1 (DBNPS) Reactor Pressure Vessel (RPV) head, the safety significance assessment as outlined in FirstEnergy Nuclear Operating Company (FENOC) letter Serial Number 1-1267, dated March 27, 2002, has been completed.

Structural integrity analyses of the as-found RPV head were performed independently by Framatome ANP and Structural Integrity Associates, Inc (SIA). These analyses compared favorably with one another and showed that the structural integrity of the as found RPV head, though degraded, would have functioned to maintain the DBNPS within its design basis during anticipated operational occurrences and postulated accidents. The more conservative of these analyses, the SIA analysis, was used as the basis for the probabilistic assessments and is provided in Attachment 1. This document is considered to be proprietary to Framatome ANP and is requested to be withheld from public disclosure. In accordance with 10 CFR 2.790, an affidavit providing the basis for withholding this information from public disclosure is provided in Attachment 4. A non proprietary version of this document is provided as Attachment 2.

A deterministic safety assessment has concluded that in the unlikely event of RPV failure considering the as-found degraded RPV head condition: a) adequate core cooling could have been established and maintained for the long term, b) the reactor could have been placed and maintained in a safe shutdown condition, and c) the integrity of the c1 r\

Docket Number 50-346 License Number NPF-3 Serial Number 1-1268 Page 2 of 2 containment would not have been compromised. In addition, a probabilistic safety assessment concluded, per Regulatory Guide 1.174 guidelines, that there was a small increase in core damage frequency and a very small increase in large early release frequency. The safety assessments are provided in Attachment 3.

If you have any questions or require additional information, please contact Mr. David H.

Lockwood, Manager - Regulatory Affairs, at (419) 321-8450.

Very trullrours, Attachmen s cc: USNRC Document Control Desk S.P. Sands, DB-1 NRC/NRR Project Manager D.V. Pickett, DB-1 NRC/NRR Backup Project Manager C.S. Thomas, DB-1 Senior Resident Inspector Utility Radiological Safety Board

Docket Number 50-346 License Number NPF-3 Serial Number 1-1268 Structural Integrity Associates, Inc.

File Number W-DB-01Q-301 "ELASTIC-PLASTIC FINITE ELEMENT STRESS ANALYSIS OF DAVIS-BESSE RPV HEAD WASTAGE CAVITY" (23 Pages Follow)

(THE FOLLOWING DOCUMENT CONTAINS PROPRIETARY INFORMATION

Docket Number 50-346 License Number NPF-3 Serial Number 1-1268 Structural Integrity Associates, Inc.

File Number W-DB-O1Q-301 "ELASTIC-PLASTIC FINITE ELEMENT STRESS ANALYSIS OF DAVIS-BESSE RPV HEAD WASTAGE CAVITY" (23 Pages Follow)

NON-PROPRIETARY

S~FILE No.: W-DB-O1Q-301 CALCULATION STRUCTURAL INTEGRITY PACKAGE PROJECT No.: W-DB-o1Q Associates, Inc.

nd Rot PROJCTAME:Opeabilty auseEvauaton o th Damge f th RectorPresur PROJECT NAME: Operability and Root Cause Evaluation of the Damage of the Reactor Pressure Vessel Head at Davis-Besse CLIENT: First Energy Corporation CALCULATION TITLE: Elastic-Plastic Finite Element Stress Analysis of Davis-Besse RPV Head Wastage Cavity Project Mgr. Preparer(s) &

Document Affected Revision Description Approval Checker(s)

Revision Pages Signature & Signatures &

Date Date 0 1-23 Original Issue N.G. Cofie R.L. Bax 4/2/02 4/2/02 Project S.S, Tang CD-Romn 4/2/02 1 1,3,4,5,8 Incorporated Client Comments ý;z,,wý Page 1 of 23

1.0 INTRODUCTION

During recent in-service inspections of the reactor pressure vessel (RPV) head and penetrations at Davis-Besse, significant wastage was observed in the vicinity of control rod drive mechanism (CRDM) No. 3. A root cause investigation is on going to determine the cause of the wastage damage.

The purpose of this calculation is to perform a finite element stress analysis to determine the limiting pressure load of the damaged RPV head.

2.0 TECHNICAL APPROACH The analysis will be performed using a three-dimensional finite element analysis and by accurately modeling the reactor pressure vessel head, the wastage area and the penetrations in the vicinity of the wastage. In this evaluation, elastic-plastic material properties of the materials of the various components will be used to determine the limiting pressure.

2.1 Finite Element Model The finite element model was constructed using the ANSYS finite element software package [1]. A full 3600 model was created and includes the following:

"* Closure head

"* Closure head cladding

"* Upper closure flange

"* CRDM housing tubes 1, 3, 6, 7 and 11

"* J-groove attachment weld and butter for CRDM tube No. 3 only

"* Wastage cavity and thinning in the low alloy steel.

Details of the model are described in the following sections and the resulting model is presented in Figures 1 and 2.

2.1.1 Hemispherical Head/Upper Closure Flange References 2 and 3 provided the closure flange and hemispherical head dimensions used in the finite element model. Figure 3 provides the dimensions used for the hemispherical head and closure flange. The flange and head were constructed using the ANSYS 8-node SOLID45 elements. The following assumptions were made during construction of the finite element model:

• The clad material thickness was modeled as a constant value over the inner surface of the head and flange.

Revision 0 1 VPreparer/Date Checker/Date RLB 4-02-02 SST 4-02-02 RLB 4-08-02 SST 4-08-02 File No. W-DB-O1Q-301 Page 2 of 23

"* The bottom face of the closure flange does not specifically model the contact surface. The compression surface identified in Figure 3 was held such that no vertical translation was possible (additional details are provided in Section 2.1.4).

"* The hemispherical head to closure flange fillet radius on the outside surface was assumed to be~nches.

"* The closure bolt holes were not specifically modeled, resulting in a solid closure flange.

"* Only the local CRDM tubes around the wasted area (tubes 1, 3, 6, 7, 11) were modeled in order to simulate the ligament effect between adjacent CRDM tubes.

0 Two stainless steel clad thickness were considered in this evaluation. The first thickness of 0.297 inches represents the average thickness of the clad in the wasted region. The second thickness of 0.24 inches represents the minimum measured thickness [4].

2.1.2 CRDM Housing Tube Penetrations A total of five CRDM housing tube penetrations were modeled. They also were modeled using the ANSYS 8-node SOLID45 element. They include Tubes 1, 3, 6, 7 and 11. Figure 4 shows the locations of the penetrations modeled. The dimensions were provided in Reference 3.

2.1.2.1 Hemispherical Head CRDM Penetration Dimensions and J-Groove Weld Based on Reference 3, the CRDM penetration holes in the hemispherical head are inches in diameter. Figure 5 provides additional details.

2.1.2.2 CRDM Tube Dimensions Per Reference 5, the (CDMtube outside diameter at the penetration is inches with an inner diameter o -inches.

For this analys,,the tube outside diameter is reduced to rnches and inner diameter retained as to nches (which results in a slight decrease in the CRDM tube thickness of Ulm nches).- effect of the interference loading between the tubes and the vessel closure head was not considered since this is not expected to have significant impact on the analysis results. In addition, the tubes were not modeled to interact with the closure head except through their attachment welds.

Revision 0 1 "V Preparer/Date RLB 4-02-02 RLB 4-08-02 Checker/Date SST 4-02-02 SST 4-08-02 File No. W-DB-01Q-301 Page 3 of 23

Finally, neither the tube stub inside the vessel nor the bolted flange connections outside the vessel were explicitly modeled. The tube lengths used for both the inside section and the outside were chosen such that they provide adequate representation in the model. See Figure 5 for additional details.

2.1.2.3 CRDM-to-Head Weld The J-groove attachment weld and butter were explicitly modeled only for tube 3. For tubes 1, 6, 7, and l Ithe CRDM tubes are attached only at the stainless clad. Figure 5 indicates the J-groove dimension used for the evaluation as provided by Reference 3.

2.1.3 Low Alloy Steel Wastage and Thinning Reference 13 provided a basic layout of wastage in a series of 3 drawings. The first outlined the region of wastage where only the cladding remained. The second was an outline of those locations where 75% of the original low alloy steel wall remained and the final indicated a basic shape at which the wall thickness had returned to nominal.

Each image provided was scanned into a CorelDraw 7.0 file in which a series of radial lines were generated from the centerline of the tube to the edge of the defined wastage. These lines were scaled in an EXCEL spreadsheet against known distances to provide a radial location at each azimuth. Locations of thinning were measured at 5' degree increments around the CRDM tube.

An example of the CorelDraw image is shown in Figure 7.

As a result, there are four primary levels of wastage. The first represent the wastage that is through the thickness of the low alloy RPV material. The second represents the area of transition from the full thickness wastage to a point where 75% of the wall remains. The third wastage represents thinning in the upper right hand quadrant that is approximately 45% of wall thickness deep. The fourth and final region represents thinning in the upper left quadrant that is to approximately 10% of the wall thickness. Figure 8 provides additional detail. The basis for the 45% and 10% regions of wastage are conservatively defined in Reference 14.

The transition from the clad only portion to the 75% wall thickness is expressed linearly at which point the transition to 100% wall is instantaneous. No curvature of smoothed transition was attempted.

The other regions of thinning were simulated as instantaneous reduction in thickness to the maximum depth of the thinning. No transitioning was performed.

Revision 0 1 VPreparer/Date Checker/Date RLB 4-02-02 SST 4-02-02 RLB 4-08-02 SST 4-08-02 File No. W-DB-01Q-301 Page 4 of 23

2.1.4 Mechanical Boundary Conditions The flange was restrained from vertical translation along the compression surface (see Figure 3).

Four additional nodes at the base of the flange restrained the structure in the lateral translations in order to avoid rigid body motion. See Figure 6 for the resulting applied boundary conditions.

3.0 MATERIALS The materials of the various components are as follows:

Component Material Upper Head Closure Flange CRDM Housing Tube J-Groove Weld Weld Butter Clad It should b noted that tmaterial properties were conservatively assumed for the J-groove l weld meta since t e stress-strain properties for the weld metal are more favorable than base metal.

3.1 Basic Material Properties The elastic material properties used for this evaluation are based on the 1989 ASMIE Code [8] for a temperature of 600'F. The properties used are indicated in Table 2.

A Poisson's Ratio oewas used for all of the materials, as was the metal density o 3.2 Elastic-Plastic Material Properties The analysis will be performed using elastic-plastic material properties for all materials involved with the exception of the SA-508 Class II closure flange. Specifically, the true stress-true strain curves for all of the materials were derived and input in the ANSYS program.

For the stainless steel weld material, the stress-strain curve, the Type 308/308L weld metal derived by Framatome Tech. Inc. [7] was used. This stress-strain curve is shown in Table 1.

For the Alloy 600 tube/attachment weld and the low alloyan so n material, the stress-strain curves were determined based on Ramberg-Osgood parameters deiid for ASME Code material properties shown in Table 2.

Revision 0 1 "V Preparer/Date RLB 4-02-02 RLB 4-08-02 Checker/Date SST 4-02-02 SST 4-08-02 File No. W-DB-01Q-301 Page 5 of 23

The true stress-true strain curve can be represented by the relationship:

- +a (1)

E0 go"0 where, E, a are true strain and true stress, E0, o0 are yield strain and yield stress, and a, n are the Ramberg-Osgood parameters.

The values of a and n were determined based on the methodology described in EPRI Report No.

NP-5531 [11].

From Reference 11, the Ramberg-Osgood true stress-true strain parameters, a and n, can be determined from the following relations: ()

n - 1 (2) 1n(1 + e nln(1+e,,)) S,(I+e)jS,(1+e,) (3)

In I +E S,, I+- 1+ E I*

where, e,, is the engineering strain at the ultimate tensile strength, conservatively assumed as e,, -0.05 e,,, is the total elongation Sy is the engineering yield strength S,, is the engineering ultimate tensile strength E is the modulus of elasticity.

The value of E is provided in ASME Section II, Part D [12] and was obtained at a temperature of 600'F. The total elongation, ett,, is provided in ASME Section II, Part A [12], but only at room temperature.

Using Equation 2 and Equation 3 and the parameters listed above, the Ramberg-Osgood parameters, n and cc, are determined and presented in Table 2. Having determined the Ramberg-Osgood Revision 0 1 "V Preparer/Date RLB 4-02-02 RLB 4-08-02 Checker/Date SST 4-02-02 SST 4-08-02 File No. W-DB-01Q-301 Page 6 of 23

parameters, the true stress-true strain curve can be determined using Equation 1 and is shown in Tables 3 and 4.

4.0 LOADING Two limit load elastic-plastic evaluations were performed with the only variation being the change in the stainless clad thickness from inches tW inches. A uniform temperature of 605'F [9] was applied over the entire model with the stress free temperature being 70'F.

The pressure was applied incrementally to the inside surface of the hemispherical head, the external surfaces of the CRDM tube sections inside the vessel, the inside surfaces of the CRDM tubes and to the closure flange face out to a radius oflo inches [2] until instability is reached.

In addition, a cap pressure was applied to the outside free end of the CRDM tubes to simulate line load in each tube. Note that the applied cap load was actually applied in the negative direction in ANSYS, thus providing a traction load.

5.0 FAILURE CRITERIA In this elastic-plastic analysis, the failure criterion is set such that the maximum strain cannot exceed the ultimate tensile strain. Hence for the stainless steel cladding where the maximum strain is expected to occur, the maximum equivalent total strain is limited to the maximum strain of 11.15%

(corresponding to the ultimate strain for the stainless steel cladding in Reference 7) through the thickness of the component.

6.0 RESULTS AND CONCLUSION For them inch clad thickness evaluation, the pressure at which the total Von-Mises strain through the clad thickness first exceeds is approximately m This load step of the elastic-plastic analysis produced th -wall"tota-strain whose lowestwvalue was & slightly below the failure criteria of" The strain throughout the thickness is shown in able 5.

For the *inchclad thickness evaluation, the pressure at which the total Von-Mises strain through the cladWthickness first exceeds lis approximately This load step of the elastic-plastic analysis produced through-wall ain whose lowest value was 4 again slightly below the failure criteria of The strain throughout the thickness is shown in Table 5.

7.0 REFERENCES

1) ANSYS Mechanical, Revision 5.7, ANSYS Inc., December 2000.

2) Framatome Technologies Technical Document 33-1201205-02, "Stress Report Summary for Reactor Vessel, Toledo Edison Company, Davis-Besse Unit No. 1," SI File W-ENTP- 11Q-219P.

3) B&W Drawing 154628E, Rev. 5, "Closure Head Sub-Assembly," Davis-Besse Document Number M-503-212, SI File No. W-ENTP-1 1Q-219P.

4) FAX from "Davis-Besse Reactor Vessel Head ID Clad Thickness Measurements in Region of Wastage Between Nozzles 3 and 11," March 18, 2002, SI File W-DB-O1Q-205P.

5) Email from Prasoon Goyal (First Energy) to Richard Bax (SI), dated Thursday, September 20, 2001, 4:45 AM, "Additional Design Input," Referencing B&W Drawing 154632E, Rev. 2, "Control Rod Mechanism Housing," Davis-Besse Document Number M-503-213-2, SI File No.

W-ENTP-11Q-219P.

6) Framatome Document 51-5013435-02, "CRDM Nozzle/Bore Dimensional Analysis," SI File No.

W-ENTP- 11Q-219P.

7) Email of from B.R. Grambau (Framatome ANP) to N. Cofie (SI), "308 Stress -Strain Curve,"

March 15, 2002, SI File W-DB-O1Q-202.

8) ASME Boiler and Pressure Vessel Code, 1989 Edition,Section III, Appendices.

9) Letter DBE-01-000133, Dated September 13, 2001 from Prasoon Goyal (First Energy) to Dick Mattson (SI), SI File W-ENTP-1 1Q-219P.

10) Email from Prasoon Goyal (First Energy) to Richard Bax (SI), dated Wednesday, September 19, 2001, 12:39 PM, "Additional Design Input," Referencing B&W Drawing 54628E "Closure Head Sub-Assembly," SI File No. W-ENTP-1 1Q-219P.

11) EPRI Report NP-5531, "Evaluation of High-Energy Pipe Rupture Experiments," January 1988.

12) ASME Boiler and Pressure Vessel Code, 1992 Edition.

13) FAX from Prasoon Goyal (First Entergy) to Nat Cofie (SI), dated Friday, March 15, 2002 at 2:49 PM, SI File W-DB-O1Q-206.

14) Email from Mark A. McLaughlin (First Energy) to Nat Cofie (SI), dated March 8, 2002, "Update on Nozzle 3 Cavity," with attachments, SI file W-DB-O1Q-201.

Revision 0 1 V Preparer/Date Checker/Date RLB 4-02-02 SST 4-02-02 RLB 4-08-02 SST 4-08-02 File No. W-DB-O1Q-301 Page 8 of 23

Table 1 True Stress-Strain Strain Curve Data for 308/308L Clad Material (Reference 7)

True Strain (in/in) II True Stress (ksi) 0 0 0.0020 31.0 0.0050 37.2 0.0100 42.8 0.0150 46.5 0.0200 49.3 0.0300 53.4 0.0400 56.6 0.0500 59.2 0.0600 61.5 0.0700 63.4 0.0800 65.1 0.0900 66.7 0.1000 68.1 0.1115 69.7 Revision 0 1 VPreparer/Date Checker/Date RLB 4-02-02 SST 4-02-02 RLB 4-08-02 SST 4-08-02 File No. W-DB-01Q-301 Page 9 of 23

Table 2 Material Stress-Strain Parameters @ 600'F Based on ASME Code (1989 Edition) Properties Material Property Low Alloy Steel Alloy 600 308/308L SA-508 Class II SA-533 Gr. B Clas*sl Qni -1KI7 I'

l/r-E (ksi)

Mean cc (in/in/°F)

Sy (ksi)

Su (ksi)

F,, (in/in) __

RAMBERG-OSGOOD PARAMETERS 0.193 0.434 (2)

_()

0C n 8.182 4.481 (1) _(2)

Oo 43.8 23.9 _(1) _(2) 6o 1.659e10.3 0.833e10 3 _ _(2)

(1) Stress-Strain curve parameters ineerence 7 were used.

Table 3 True Stress-Strain Strain Curve Data fo True Strain (in/in) True Stress (ksi) 0 0 0.000166 4.38 0.000332 8.76 0.000498 13.14 0.000664 17.52 0.000831 21.9 0.001000 26.28 0.001179 30.66 0.001379 35.04 0.001628 39.42 0.001979 43.8 0.002523 48.18 0.003414 52.56 0.004897 56.94 0.007347 61.32 0.011324 65.7 0.017635, 70.08 0.027422 74.46 0.042257 78.84 0.064273 83.22 0.096312 87.6 0.142104 91.98 0.206479 96.36

Table 4 True Stress-Strain Strain Curve Data for Alloy 600 True Strain (inlin) True Stress (ksi) 0 0 8.33E-05 2.39 0.000167 4.78 0.000251 7.17 0.000339 9.56 0.000433 11.95 0.000536 14.34 0.000656 16.73 0.000799 19.12 0.000975 21.51 0.001194 23.9 0.001470 26.29 0.001817 28.68 0.002254 31.07 0.002798 33.46 0.003473 35.85 0.004302 38.24 0.005312 40.63 0.006533 43.02 0.007996 45.41 0.009736 47.8 0.011792 50.19 0.014203 52.58 0.017013 54.97 0.020268 57.36 0.024019 59.75 0.028317 62.14 0.033219 64.53 0.038784 66.92 0.045074 69.31 0.052156 71.7 0.060099 74.09 0.068975 76.48 0.078862 78.87 0.089840 81.26 0.101992 83.65 0.115406 86.04 0.130173 88.43 0.146389 90.82 0.164152 93.21 0.183565 95.6 0.204736 97.99 0.227775 100.38 0.252798 102.77 0.279924 105.16 Revision 0 1 VPreparer/Date Checker/Date RLB 4-02-02 SST 4-02-02 RLB 4-08-02 SST 4-08-02 File No. W-DB-O1Q-301 Page 12 of 23

Table 5 Through Wall Total Von Mises Strain at Failure Criteria Clad Thickness Through-Wall Nodes (1) OD Surface (2) H41 (3)

(4) Mid Span

-'I.

(5)

(6)

(7) ID Surface Revision 0 1 SPreparer/Date Checker/Date RLB 4-02-02 SST 4-02-02 RLB 4-08-02 SST 4-08-02 File No. W-DB-O1Q-301 Page 13 of 23

Figure 1 - Finite Element Model Revision 0 1 VPreparer/Date Checker/Date RLB 4-02-02 SST 4-02-02 RLB 4-08-02 SST 4-08-02 File No. W-DB-O1Q-301 Page 14 of 23

Figure 2 - Shaded Finite Element Model (Colors by Material)

Revision 0 1 VPreparer/Date Checker/Date RLB 4-02-02 SST 4-02-02 RLB 4-08-02 SST 4-08-02 File No. W-DB-O1Q-301 Page 15 of 23

..............N S. at.n.

Mating ..............................................................

Surface [3] 1*

Figure 3 - Top Hemispherical Head / Closure Flange Dimensions Revision 0 1 S Preparer/Date Checker/Date RLB 4-02-02 SST 4-02-02 RLB 4-08-02 SST 4-08-02 File No. W-DB-01Q-301 Page 16 of 23

CRDM Penetrations (69 total [3])

/ I S/ I .

/ I

/  !

/ 1

/ /

/ /

/ /

/ /

/ /

/ /

Figure 4 - CRDM Penetration Locations Revision 0 1 VPreparer/Date Checker/Date RLB 4-02-02 SST 4-02-02 RLB 4-08-02 SST 4-08-02 File No. W-DB-O1Q-301 Page 17 of 23

Figure 5 - CRDM Tube/Weld Arrangement Revision 0 1 VPreparer/Date Checker/Date RLB 4-02-02 SST 4-02-02 RLB 4-08-02 SST 4-08-02 File No. W-DB-O1Q-301 Page 18 of 23

1 AN Figure 6 - Mechanical Boundary Conditions Revision 0 1 VPreparer/Date Checker/Date RLB 4-02-02 SST 4-02-02 RLB 4-08-02 SST 4-08-02 File No. W-DB-O1Q-301 Page 19 of 23

00 Figure 7- Example of Wastage Sizing

J-Groove Weld /

/

Open Clad J Transition from "No Back Wall to 75%

Figure 8 - Layout of Wastage as Modeled Revision 0 1 VPreparer/Date Checker/Date RLB 4-02-02 SST 4-02-02 RLB 4-08-02 SST 4-08-02 File No. W-DB-O1Q-301 Page 21 of 23

Figure 9 - Von Mises Strain in Clad (Top/Bot) form inch Clad at psi Revision 0 1 V Preparer/Date Checker/Date RLB 4-02-02 SST 4-02-02 RLB 4-08-02 SST 4-08-02 File No. W-DB-O1Q-301 Page 22 of 23

Figure 10 - Von Mises Strain in Clad (Top/Bot) folinch Clad alwpsi Revision 0 1 VPreparer/Date Checker/Date RLB 4-02-02 SST 4-02-02 RLB 4-08-02 SST 4-08-02 File No. W-DB-O1Q-301 Page 23 of 23

Docket Number 50-346 License Number NPF-3 Serial Number 1-1268 Page 1 of 19 Safety Assessment of the Davis-Besse Reactor Vessel Head Corrosion Introduction During the thirteenth refueling outage (13RFO), an inspection of the Davis-Besse reactor vessel head and its control rod drive mechanism (CRDM) nozzles was completed. The inspection identified several CRDM nozzles with cracking. While preparing to repair nozzle number 3, a region of substantial boric acid wastage adjacent to the nozzle was discovered. As a result of this as-found condition, an assessment of the plant's safety has been prepared. A deterministic assessment evaluates the potential consequences associated with a postulated pressure boundary breach from the boric acid wastage. In addition, a risk-based assessment examines the change in plant core damage frequency and large early release frequency. The purpose of this paper is to present the results of the assessments.

The deterministic assessment evaluates three critical aspects associated with a breach of the Reactor Coolant System (RCS) pressure boundary. These include the adequacy of the overall core cooling, the ability to establish and maintain safe reactor shutdown conditions, and the adequacy of the containment, the outer fission product barrier. The core cooling evaluation examines peak clad temperatures, inventory control, long term cooling, and boron concentration control. The reactor shutdown evaluation examines total core reactivity with consideration for potential loss of insertable reactivity. The potential failure mechanisms that could reduce the insertable reactivity are also discussed.

Finally, the assessment of the integrity of the containment building evaluates the potential for damage due to generation of a missile and the potential for over-pressurization due to the RCS pressure boundary breach.

The risk assessment evaluates the change in Core Damage Frequency (CDF), the change in Large Early Release Frequency (LERF), and the effect on public health risk. These are based on calculations that consider the frequency of having a transient that could trigger the reactor vessel head to fail, the probability of the head failing, based on as-found conditions, and the conditional core damage probability, once a failure of the Reactor Coolant pressure boundary has occurred.

A. Deterministic Safety Assessment A. 1. Boundary Conditions and Assumptions In evaluating the consequences of a breach of the pressure boundary, it is necessary to define the initial boundary conditions and assumptions included in the assessment. It is

Docket Number 50-346 License Number NPF-3 Serial Number 1-1268 Attachment 3 Page 2 of 19 assumed that a small break LOCA is initiated from nominal plant conditions, i.e. an average coolant temperature of 582°F and a hot leg pressure of 2170 psia. The local fluid condition at the top of the reactor vessel (RV) head is approximately 2230 psia and 606.5°F. The nominal break area is based on the area of the exposed cladding material as depicted in Figure 1. The nominal break area, excluding CRDM# 3 is estimated to be approximately 25 in 2. However, a break area, ranging from a pinhole sized opening in the exposed stainless steel cladding up to an area 1.5 times the estimated area of exposed RV cladding material plus the area of a CRDM nozzle that may also have been lost, is considered in this evaluation. For the core reactivity aspects, it is assumed that the control rod located in CRDM #3 is not available for core reactivity shutdown. In addition it is assumed that the Technical Specification minimum Borated Water Storage Tank (BWST) and Core Flood Tank (CFT) volumes and boron concentrations are available.

A.2. Single Failure Considerations Single failures do not play a particular role in this assessment. This assessment only evaluates the consequences of the initiating event. The failure effects analyses have been included in the various response analyses used to support this assessment. For example, the plant response due to a failure of one Emergency Core Cooling System (ECCS) train has already been evaluated and shown acceptable. This assessment does not alter that failure effects analysis. Where standard failure modes may not bound the as-found conditions, this assessment will include appropriate discussions. As an example, the potential for the failure of multiple CRDMs due to the postulated scenario is discussed within this assessment.

A.3. Capability to Maintain Adequate Core Cooling A.3.1 Peak Clad Temperature LOCA analyses performed to demonstrate compliance to 10 CFR 50.46 (Reference A. 1) for Davis-Besse do not specifically include the potential failure of the reactor vessel or any of the attached CRDM nozzles. The analyses do include a spectrum of reactor coolant system (RCS) pipe break sizes from 0.01 ft2 to 14.2 ft2 in area, Reference A.2.

The as-found boric acid wastage area on the top of the reactor vessel head exposed a portion of the stainless steel cladding on the inside of the vessel head that was estimated to be approximately 25 in2 , based on Figure 1. If the cladding had failed, the RCS break area could be anywhere from a pinhole up to the area of the exposed cladding. The smaller a break in the upper head, the less challenging to core cooling, since mass loss rates are lower. This keeps the core covered with froth, thereby limiting cladding temperature excursions. Because of the uncertainty in the size of the wastage area, a hole size up to 50 percent larger than the preliminarily estimated size will be considered (i.e.

Docket Number 50-346 License Number NPF-3 Serial Number 1-1268 Attachment 3 Page 3 of 19 37.5 in2). The effects of a break of this size conservatively bounds the actual wastage area identified in References B.3 and B.4. It is also possible that a rupture near the CRDM #3 nozzle opening could result in that CRDM nozzle being dislodged from the head, thereby increasing the hole size by 12.5 in2 . This represents a maximum hole size of approximately 50 in2 (-0.35 ft2) to be considered. The consequences of any break size up to this maximum break area in the RV upper head has been reviewed and is concluded to be bounded by the consequences determined in the existing Davis-Besse LOCA analyses that demonstrate compliance to 10 CFR 50.46.

A break in the RV upper head is favorable from a core cooling standpoint because it is on the hot side of the core, such that all emergency core cooling system (ECCS) fluid is available for core cooling before it can be discharged out of the break. Consequently, the RV liquid inventory would be sufficient to keep the core covered throughout the transient with a two-phase mixture level that would prevent heat up of the fuel pins or cladding.

Therefore, the maximum peak cladding temperature would be equal to the initial steady state cladding temperature of roughly 715°F.

Davis-Besse also has a continuous RV head vent line from CRDM #14 to Steam Generator 1-2. The relative location of the continuous RV head vent line is shown in Figure 2. CRDM #14 is one nozzle removed from CRDM #3 as shown in Figure 1 and Figure 2. If ejection of a CRDM damages this line, the effective break area would not be significantly increased because the pipe area is less than 3.6 in2 (-0.025 ft2), Reference A.3, and the break is still within the phenomenological range being considered in this evaluation. There would be no increase in the expected peak clad temperature if the continuous head vent line were damaged.

A.3.2 Long Term Decay Heat Removal/Boron Concentration Control The maximum break size is large enough to depressurize the RCS well within the low pressure injection (LPI) pressure range within the first six minutes, Reference A.2.

Reduced break sizes would take longer to depressurize, but the RCS inventory loss would be lower such that HPI and the remaining RCS inventory are capable of supplying adequate core heat removal. Eventually, the RCS will depressurize to the point that the ECCS injection rate exceeds the leak rate allowing the vessel to begin to refill. The RCS and vessel will refill to the break elevation and the ECCS fluid that passes through the core and out the break can reestablish core exit subcooling.

Eventually, the ECCS injection will empty the borated water storage tank (BWST). If the break size is large, the RCS will depressurize below the core flood tank (CFT) pressure, allowing it to discharge. The containment spray flow, if actuated, and the RCS inventory that exits the break will flow through designed drain paths into the reactor building lower

Docket Number 50-346 License Number NPF-3 Serial Number 1-1268 Attachment 3 Page 4 of 19 elevations and raise the containment water level until it spills into the emergency sump.

The liquid in the emergency sump provides the inventory that the Containment Spray and ECCS pumps use to provide long-term recirculation injection. Consequently, an assured long-term core cooling liquid source from the emergency sump will be available.

Switching the ECCS pump suction source from the BWST to the sump establishes the path for continuous long-term cooling.

The break size and location are also favorable from a boron precipitation perspective.

The specific break location will inherently provide adequate passive boron concentration control since flow will be maintained through the break. While not needed for this break location, one of the active boron dilution methods may also be initiated, if conditions specified by Reference A.4 are met. Both the passive dilution provided by the break and the active boron dilution control ensure that the post-LOCA core boron concentration will be maintained within acceptable limits.

A.4 Reactor Shutdown A.4.1 Immediate Shutdown Margin In addition to effective RCS inventory control to establish and maintain core cooling, adequate reactor shutdown capability must also be demonstrated for this event. During the early portion of a small break LOCA transient, shutdown is ensured by the insertion of the control rods. In the LOCA analyses, a minimum inserted control rod worth is credited. The analyses model a control rod worth that is sufficient to offset the moderator and Doppler deficits from hot full power conditions to hot zero power conditions plus 1 percent for shutdown margin. For Cycle 14 this is 2.3 %Ak/k (Reference A.5). [The calculation was based strictly on the Cycle 14 core design, with beginning of life core parameters, and the expected post-LOCA temperatures. The Cycle 14 core design will be more limiting than Cycle 13 because the Cycle 14 design includes four additional fresh fuel assemblies surrounding the interior of the core (i.e., core locations H07, G08, H09, and K08 in Figure 2). This will result in a relatively high worth for CRDMs #1, 3, 6, 7, and 11 when these control rods are not placed in the core, which results in a lower total inserted worth.] The minimum inserted rod worth modeled in the LOCA analyses is less than the total available, even after accounting for calculation uncertainties and a penalty for the maximum stuck control rod, Reference A.6. No credit is taken for other reactivity sources. While the total available rod worth would be somewhat less with the loss of CRDM #3, the shutdown margin modeled in the analyses would still be preserved.

Evaluation shows that with the five rods around the corrosion area, and one other randomly stuck rod, there would be an insertable reactivity of 3.6 %Ak/k (Reference A.7). This exceeds the amount of reactivity needed to shut down the reactor to hot zero power conditions, as discussed above.

Docket Number 50-346 License Number NPF-3 Serial Number 1-1268 Attachment 3 Page 5 of 19 A.4.2 Potential for Collateral Damage to Adjacent CRDMs In a previous evaluation, Reference A.8, performed for the B&W Owners Group (BWOG) it was assumed that an ejected CRDM would rise vertically. If there were an impact on an adjacent CRDM, the contact would not cause failure of those adjacent control rods because there was no indication of significant wastage near the adjacent CRDMs and the J-groove welds were not structurally impaired. Based on the as-found condition of the J-groove weld and the CRDM nozzle #3, this position is not necessarily invalidated but can not be rigorously demonstrated. Therefore, assessment of the potential for causing sufficient damage to prevent the adjacent rods from inserting is evaluated below. The BWOG study also evaluated the failure of an adjacent control rod to insert due to loose parts or metal fragments in the upper head region associated with a control rod ejection. This condition is also examined.

In order to address the potential damage to adjacent control rods, an assessment was performed (Reference A.9) to determine the effects of steam flowing out of a hypothesized rupture of the cladding at CRDM nozzle #3 into the service structure on the adjacent CRDMs. The break area, 37.5 in2 assumed based upon the size of the exposed cladding with a 50% increase. The break area considered the cladding rupture along with the catastrophic loss of CRDM nozzle #3. There are no jet loads on the adjacent CRDMs since the CRDMs are parallel to the jet. As such, the loadings on the CRDMs have been treated as cavity pressures. The loadings are conservatively approximated based on ANS 58.2/1988, Reference A.10, by calculating the pressure as a function of the length from the RV head. The most conservative CRDM location is chosen to maximize the pressure loadings on the CRDM. The pressure loadings are applied to a structural model of the CRDM. The structural model is fixed at the RV Head and pinned at the service support structure (SSS) tie plate location. The bending moments are a maximum at the RV head location. Reviewing the section properties of the CRDM and housing, this is the critical location. Using the housing cross-section, just above the RV head, the bending stress is less than the yield strength at the normal full-power temperature. Therefore, it can be concluded that the adjacent CRDMs loaded from a hypothesized rupture of the as-found wastage area and the catastrophic loss of the CRDM nozzle stay elastic and the control rods can be inserted. Since the rods rapidly drop into the core, there is insufficient time for the ejected CRDM to leave the service structure, hit the missile shield, and fall back to the structure before rod insertion is complete. Therefore, subsequent CRDM damage would not affect the insertion of the rods.

The "effective" break area considered in previous evaluations was small, such that it would take roughly 15 to 20 seconds for the RCS to depressurize to the low-pressure reactor trip setpoint. This timing could allow the loose parts that may be generated as the rod is ejected to enter the adjacent control rods and potentially impair their operation. It

Docket Number 50-346 License Number NPF-3 Serial Number 1-1268 Attachment 3 Page 6 of 19 was postulated that a nozzle fragment Could be generated if a circumferential crack joined up with axial cracks in the nozzle base metal below the weld. In the BWOG report (Reference A.9), it was noted that about 25 percent of the plenum cover area is comprised of control rod column weldments. If debris falls within a guide column weldment, to interfere with control rod movement, the debris would have to be trapped between a control rod guide brazement and the control rod spider. The area of the control rod guide brazement that debris must be trapped against to interfere with the control rod spider movement is a small portion of the column weldment area. The control rod guide brazements are at several elevations and blockage at each decreasing brazement elevation would permit an increasing length of control rod to enter the core. Consequently, the probability of impeding multiple control rods from being inserted is unlikely.

The indication surrounding Davis-Besse CRDM nozzle #3, however, is more severe than previously found for other CRDM nozzles or J-groove welds, in that the potential break area could be much larger than previously considered. Based on the as-found conditions, if the CRDM is lost and the stainless steel cladding gives way, a break size of approximately 0.35 ft2 is postulated. The low RCS pressure reactor trip signal will be reached between 1 to 2 seconds. The control rods will begin falling, and will be fully inserted within two seconds after the trip signal. This timing will not allow enough time for the fragments from a potentially failed control rod assembly to inhibit the insertion of any adjacent control rods. As a result of the early reactor trip, fragments from a potentially failed control rod assembly would not inhibit the insertion of any adjacent control rods.

A.4.3 Maintaining Safe Shutdown Conditions As discussed above in Section 4.1, sufficient negative reactivity exists to immediately shutdown the reactor. Early in the transient, voiding in the core region and xenon buildup augments the inserted control rod worth. Because of this specific break location, core subcooling may be regained and over time, xenon will decay, eliminating these additional sources of negative reactivity. Depending on the inserted control rod worth and the moderator temperature coefficient, the real transient core power may become somewhat cyclic. If the power does increase, subcooling margin will again be lost, localized boiling may be initiated, and the power will decrease. This cycle will continue until the mixed-mean boron concentration has been increased sufficiently to suppress the core reactivity.

Post-LOCA long-term shutdown of the core is provided by the soluble boron in the ECCS fluid and total inserted worth. The required boron concentration in the ECCS fluid is based on only having 50 percent of the control rod worth available. As a result, even

Docket Number 50-346 License Number NPF-3 Serial Number 1-1268 Page 7 of 19 with a stuck rod penalty and the loss of a single control rod (CRDM nozzle #3), adequate long-term shutdown would be demonstrated.

Notwithstanding the conclusions developed by the CRDM loading evaluation, a determination of the overall available control rod worth was made assuming that CRDMs

1. 1, 3, 6, 7, and 11 fail to insert. (See Figure 2 for details on the CRDM locations.) The calculation was based strictly on the Cycle 14 core design, with beginning of life core parameters, and the expected post-LOCA temperatures. The Cycle 14 core design will be more limiting than Cycle 13 because the Cycle 14 design includes four additional fresh fuel assemblies surrounding the interior of the core (i.e., core locations H07, G08, H09, and K08 in Figure 2). This will result in a relatively high worth for CRDMs #1, 3, 6, 7, and 11 when these control rods are not placed in the core, which results in a lower total inserted worth. Using the smaller inserted rod worth from the Cycle 14 core design and applying it to the boron requirements for both Cycle 13 and Cycle 14 ensures that a conservative evaluation is performed. The results of the calculation (Reference A.11) concluded that the minimum total inserted worth of the remaining control rods, less an additional penalty for the highest worth remaining rod to be stuck out of the core, was found to be approximately 3.1% Ak/k.

In parallel with the evaluation contained in Reference A. 11, a calculation was performed to determine the control rod worth required to maintain adequate shutdown margin after considering the negative reactivity addition from the boron in the RCS, BWST, CFTs, and makeup tank for the Cycle 13 and 14 core designs, Reference A.7. Using the minimum Technical Specification BWST and CFT volumes and boron concentration of 2600 ppm, Reference A.12, the required control rod worth for long term shutdown, including a 1-percent shutdown margin, was between approximately 1.0% to 1.9% Ak/k, depending on the fuel cycle and fluid temperature being evaluated.

Based on a simple comparison of the available control rod worth versus the required worth, there would have been sufficient negative reactivity in the inserted control rods and the effective mixed-mean boron concentration to adequately shutdown the core following this postulated LOCA scenario. The negative reactivity would maintain the core in a subcritical state even if CRDMs #1, 3, 6, 7, and 11 did not insert along with the highest remaining worth control rod stuck out of the core. In other words, there would be sufficient boron in the available fluid systems and the control rods that were inserted for long-term core reactivity control following this location-specific small break LOCA.

Docket Number 50-346 License Number NPF-3 Serial Number 1-1268 Page 8 of 19 A.5 Containment Integrity The containment integrity must also be considered. Davis-Besse has a reactor vessel missile shield installed above the reactor vessel that would preclude an ejected CRDM from hitting the containment structure and impairing its fission product barrier function.

The mass and energy release to the containment associated with the spectrum of LOCA break sizes has been evaluated from a peak pressure and temperature perspective, Reference A.13. The rapid blowdown from a double-ended hot leg pipe break (14.14 ft2) is limiting for Davis-Besse and the consequences bound this smaller break size because the estimated area is more than an order of magnitude smaller. The resulting rate of energy release to the reactor building will be much slower, such that the peak allowable containment pressure and temperature will not be challenged for this postulated reactor vessel break.

This size break may or may not actuate the Containment Spray system. That system is set to start automatically if the containment pressure limit is approached. The system can be started manually if deemed appropriate by plant operators. This may be done to speed depressurization of containment and to scrub radioactive releases from the atmosphere, however, manual actuation will not affect the consequences of the accident.

B. Risk Based Safety Assessment B. 1 Boundary Conditions and Assumptions An assessment of the change in core damage frequency was made due to the degradation of the reactor vessel head. This evaluation is based, in part, on the deterministic evaluation described above and Reference B. 1. This includes assuming sufficient negative reactivity is available to shut down the reactor and maintain it shut down throughout the event. The basis for this assumption is described in the Section A.4, above. It is also assumed that no impairment to containment integrity occurs as a result of the head failure, including ejection of a CRDM. The basis of this assumption is discussed in the Section A.5, above.

It is not necessary to postulate specific equipment failures in this probabilistic portion of the Safety Assessment. Equipment failures are addressed in the core damage probability calculation though the normal failure rates included in the PSA evaluation model.

Docket Number 50-346 License Number NPF-3 Serial Number 1-1268 Attachment 3 Page 9 of 19 B.2 Event Initiators The only events that could have challenged the degraded RCS pressure boundary include overpressure transients and seismic events. The Davis-Besse seismic PSA analysis (Reference B.2) was reviewed to estimate the frequency of pressure boundary failure due to a seismic event. Based on this review it was determined that a seismic event of sufficient magnitude to initiate failure of the degraded RCS pressure boundary would have a frequency sufficiently low to justify neglecting seismic events in this analysis.

It is therefore postulated that only an overpressure condition initiated by a plant transient could have challenged the RCS integrity. This assessment evaluates the risk contribution from a LOCA due an overpressurization transient causing failure of the wastage area in the head.

B.3 Event Frequency The frequency of a transient that actually causes failure of the reactor vessel head was calculated. This frequency is based on the probability of head failure and the frequency of the overpressure transient. The probability of head failure and the frequency of the over pressurization transient are discussed below.

B.3.1 Probability of Pressure Boundary Failure due to an Overpressure Transient A finite element stress analysis of the Davis-Besse head wastage cavity was performed by Framatome ANP (Reference B.3) and independently by Structural Integrity Associates (Reference B.4). These analyses determined a maximum pressure for the cladding without loss of pressure boundary function. The definition of maximum pressure was chosen to be the pressure at which the strain value reaches 11% or more through the full thickness of the cladding. A maximum pressure of 5600 psig was assumed for the risk analysis based on the more limiting average of the analytical results.

Exceeding the maximum pressure of 5600 psig is not considered credible. Overpressure protection is provided by the Pilot Operated Relief Valve (PORV) and the two Code Safety Valves, with setpoints of 2450 psig and 2500 psig, respectively. Furthermore, the pressure sufficient to lift the reactor vessel head from the vessel and break the seal between the head and the vessel flanges has been estimated to be in the range of 3000 psig to 4000 psig. Consequently, the head would lift and relieve pressure through the flanges before exceeding the capability of the wastage area. However, to calculate the risk of the as-found condition, it is recognized that the clad in the wastage area could fail before the calculated maximum pressure. The probability of failure before reaching the maximum calculated pressure was estimated by applying the method taken from

Docket Number 50-346 License Number NPF-3 Serial Number 1-1268 Attachment 3 Page 10 of 19 NUREG/CR-5603, Reference B.5. This approach is also described in NUREG/CR-5604, Reference B.6 and NUREG/CR-2300, Reference B.7. Based on these references, the failure probability of the cladding on the wasted area of the reactor head can be expressed as follows:

Where f Probability that failure occurs at pressure p P = Median pressure capability (psig) p = random pressure capacity (psig) 13c = Composite logarithmic standard deviation for randomness 00 = Standard gaussian cumulative distribution function The median pressure capacity represents the internal pressure level for which there is a 50% probability of failure. This value is assumed to be 5600 psig based on the results in Reference B.3 and B.4.

The logarithmic standard deviation 03, is a function of various uncertainties. Uncertainties can exist due to difference in modeling and material properties. Modeling uncertainties are associated with the assumptions used to develop the analytical models and differences between analytical idealization and the real conditions. Material uncertainties involve variability in material properties. Reference B.5 provides values of logarithmic standard deviations for various materials and applications. For this analysis, a very conservative logarithmic standard deviation of 0.33 was used based on the values reported in References B.5 and B.6 for a temperature of 600 degrees.

Applying the assumed median pressure of 5600 psig and logarithmic standard deviation of 0.33, the equation can be solved for probability of failure as a function of pressure.

The results of this analysis are as shown in Table 1.

Docket Number 50-346 License Number NPF-3 Serial Number 1-1268 Page 11 of 19 Table 1 - Failure Probabilities as a Function of Pressure Pressure (psi g) Failure Probability 2250 2.86E-3 2300 3.50E-3 2350 4.25E-3 2400 5.12E-3 2450 6.12E-3 2500 7.27E-3 B.3.2 Transient Frequencies The transient history as documented in Davis-Besse Transient Assessment Program Reports was reviewed to collect data on pressure excursions at Davis-Besse from July 1979 through February 2002. The yearly frequency of pressure excursions was calculated based on total operating hours from July 1979 to February 2002. The number of pressure transients and the yearly frequencies are shown in Table 2. No pressure excursions of 2450 psig or greater have occurred in the Davis-Besse operating history.

Therefore, the frequency of these events was estimated using a Bayesian update with a non-informative prior as described in NIJREG/CR-2300, Reference B.7. The frequency for events greater than 2450 psig was further divided into two categories representing events with pressures greater than 2450 psig and greater than 2500 psig. Reference B.8 provides the details of this calculation.

Table 2 - Frequency of Pressure Transients Pressure (psig) Number of Events Frequency (yr")

> 2250 4 2.54E-1

> 2300 8 5.08E-1

> 2350 2 1.27E-1

> 2400 1 6.35E-2

> 2450 0 1.59E-2

> 2500 0 1.59E-2

Docket Number 50-346 License Number NPF-3 Serial Number 1-1268 Attachment 3 Page 12 of 19 B.3.3 Frequency of Pressure Boundary Failure The frequency of failure for the affected area of the reactor vessel head was calculated using the information in Tables 1 and 2. The frequency of each category of overpressure transient was multiplied by the probability of failure of the head area for that pressure.

The results were then summed to calculate a total failure frequency.

Table 3 - Frequency of Pressure Boundary Failure Pressure Overpressure Head F Failure P sFrequency Head Failur Frequency (psig) (yr'l) Probability (yra')

> 2250 2.54E-1 2.86E-3 7.27E-4

> 2300 5.08E-1 3.50E-3 1.78E-3

> 2350 1.27E-1 4.25E-3 5.40E-4

> 2400 6.35E-2 5.12E-3 3.25E-4

> 2450 1.59E-2 6.12E-3 9.71E-5

> 2500 1.59E-2 7.27E-3 1.15E-4 3.58E-3 B.4 Conditional Core Damage Probability Possible LOCAs that could result from a failure of the wasted area range from very small to 0.35 ft2 . This includes breaks that would be categorized as small or medium in the Davis-Besse PSA. The conditional core damage probability for small LOCAs is smaller than for medium LOCAs. Therefore, for this analysis, all failures will be assumed to be in the medium LOCA range.

Medium LOCAs, as defined in the PSA, cover a range of break areas of 0.02 ft 2 to 0.5 ft2 .

The largest LOCA size, postulated for failure of the head wastage area, is less than the maximum of the medium LOCA range. Therefore, this analysis will evaluate the conditional core damage probability for a 0.35 ft2 LOCA. The largest LOCA within the postulated range allows the shortest time to transfer to recirculation, but exceeds the LOCA size that would require high pressure injection. Therefore, a smaller LOCA that requires high pressure injection could be more limiting. A LOCA of 0.1 ft2 represents the upper range of the LOCAs that require high pressure injection so this size LOCA was also evaluated.

The results of the analysis documented in Reference B.9 demonstrates that the highest conditional core damage probability (CCDP) of any LOCA in the range from very small

Docket Number 50-346 License Number NPF-3 Serial Number 1-1268 Page 13 of 19 through 0.35 ft 2 occurs for a break of about 0.1 ft2 . The conditional core damage probability of this size break is 2.9E-3. This result differs from the conditional core damage probability in the Davis-Besse Individual Plant Examination (IPE) based on the following:

"* The CCDP in the IPE represents the worst case conditions for any medium LOCA and combines the human error probability of the largest possible medium LOCA with the more restrictive success criteria of a smaller LOCA.

"* The human error probability (HEP) for the action required to transfer to recirculation was based on a specific calculation for a 0.1 ft2 LOCA. Additionally, the methods used for timing calculations have been improved by the use of the Modular Accident Analysis Program (MAAP) which replaces more simple calculations based on rated pump flows.

"* The Davis-Besse Emergency Procedure, Reference A.4 was improved based on feedback from the PSA human reliability analysis. This improvement involved adding the procedural requirement to initiate outside of the control room actions necessary to initiate low pressure recirculation prior to reaching low level in the BWST. This change greatly expands the time window for outside the control room actions. Additionally, the actions have been included on a check off list requiring a specific written entry.

B.5 Core Damage Frequency 2

The more limiting core damage probability was evaluated to be associated with the 0.1 ft LOCA. Therefore, the result for this smaller LOCA will be used in this analysis.

Based on the conditional core damage probability of 2.9E-3 and the failure frequency of 3.6E-3/yr, the core damage frequency can be calculated as follows:

CDF = 2.9E-3 x 3.6E-3/yr CDF = 1.OE-5 /yr.

B.6 Core Damage Probability Evaluation The calculation of a core damage probability requires information concerning the duration of the head wastage condition, which is difficult to assess. However, to estimate a result, the vessel head condition can be assumed to have existed at the as-found condition for one year. The total core incremental damage probability for this time would be 1.OE-5. Based on the guidance in the EPRI PSA Procedures Guide ,Reference B.10, this core damage probability is conservatively estimated at the value of 1E-5, which is considered the lower threshold of being risk significant for a temporary change.

Docket Number 50-346 License Number NPF-3 Serial Number 1-1268 Attachment 3 Page 14 of 19 B.7 Large Early Release Frequency Contribution For the assessment of large early release frequency (LERF), it is conservatively assumed that all pressure boundary failures result in a medium LOCA. The conditional large early release probability (CLERP) for a medium LOCA is about 4.OE-6 based on the results of the Davis-Besse PSA (Reference 11).

Using the frequency of failure calculated for the reactor vessel head from Table 3 and the conditional large early release probability, LERF for this condition is calculated as follows:

LERF = 3.6E-3 /yr. x 4.OE-6 LERF = 1.43E-8 /yr.

The CLERP for a medium LOCA is very small resulting in a LERF contribution from nozzle failures that is much less significant than the core damage frequency contribution.

However, this small LERF contribution is expected for a medium LOCA due to specific plant features. Davis-Besse has a large dry containment with a large containment to core power ratio. Consequently, the most significant contributors to release from containment for Davis-Besse are containment bypass events, such as steam generator tube ruptures or interfacing system LOCAs. Other significant contributors to LERF include sequences where core damage results from loss of support systems that also support containment cooling. These sequences include loss of cooling water systems or station blackout.

Medium LOCAs are not representative of either of these dominant LERF sequences.

Additionally, LOCAs depressurize the reactor coolant system prior to core damage, which reduces the possibility of high pressure melt ejection.

B.8 Public Health Risk The Davis-Besse Level 3 PSA, Reference B.12, was used to estimate the impact on the public, in terms of person-REM, from the condition of the head. Based on the results of the Level 3 PSA analysis, a conditional population dose given a medium LOCA is 1.57E2 person-REM.

Using the frequency of failure calculated for the reactor vessel head from Table 3 and the conditional population dose, the health consequences for this condition are calculated as follows:

Person-REM = 3.6E-3 /yr. x 1.57E2 Person-REM Person-REM = 0.56 Person-REM/yr.

Docket Number 50-346 License Number NPF-3 Serial Number 1-1268 Page 15 of 19 Conclusions Based on the deterministic safety assessment, it is concluded that

"* Adequate core cooling can be established and maintained for the long term.

"* The reactor can be placed and maintained in a safe shutdown condition.

"* The integrity of the containment will not be compromised as a result of the damage found on the reactor vessel head.

The results of the risk based safety assessment show that:

"* The core damage frequency (CDF) contribution of the as-found condition is conservatively estimated to be 1.OE-5 / yr. or less. This is a value that would be considered small per Regulatory Guide 1.174 (Reference B. 13).

"* The large early release frequency (LERF) contribution of the as-found condition is estimated to be 1.4E-8 / yr. This is a value that would be considered very small per Regulatory Guide 1.174 (Reference B. 13).

"* The public health risk was calculated to be 0.56 person-REM/yr which is a very small contribution to the total risk for Davis-Besse.

" Core damage probability is difficult to assess since the duration of the condition is not well known. However, it is expected to be approximately 1E-5. This is the lower threshold that is considered risk significant per the EPRI PSA Applications Guide (Reference B. 10).

References A.1 Title 10 of the Code of Federal Regulations, Part 50.46.

A.2 Framatome ANP Document 86-5006232-00, "DB-1 LOCA Summary Report."

A.3 Davis-Besse System Description SD-039A, "Reactor Coolant System."

A.4 Davis-Besse Operating Procedure DB-OP-02000, "RPS, SFAS, SFRCS Trip or Steam Generator Tube Rupture."

A.5 Framatome ANP Document No. 86-5006232, "DB-1 LOCA Summary Report."

Docket Number 50-346 License Number NPF-3 Serial Number 1-1268 Page 16 of 19 A.6 Framatome ANP Document No. 103-2368-00, "Davis-Besse Nuclear Power Station Unit 1, Cycle 13 Reload Report," BAW-2368.

A.7 Framatome ANP Document No. 86-5017531-00, "D-B Post LOCA Sump Support From Task 11."

A.8 Framatome ANP Document 51-5011603-01, "RV Head Nozzle and Weld Safety Assessment."

A.9 Framatome ANP Document No. 32-5017526-00, "CRDM Integrity Due to RV Head Rupture."

A. 10 American National Standard, "Design Basis for Protection of Light Water Nuclear Power Plants Against the Effects of Postulated Pipe Rupture, ANSI/ANS-58.2-1988."

A.11 Framatome ANP Document No. 32-5017530-00, "Davis-Besse RV Head SBLOCA Required Rod Worth."

A.12 Davis-Besse Nuclear Power Station, Unit No. 1, Docket No. 50-346, Technical Specifications, Through Amendment 255, Limiting Condition for Operation 3.5.1 and 3.5.4.

A. 13 Davis-Besse Calculation C-NSA-60.05-009, "Containment Analyses for Caldon Power Uprate."

B. 1 Framatome ANP, 51-5017529-00, Davis-Besse RV Head SBLOCA Assessment."

B.2 Davis-Besse Nuclear Power Station Seismic Risk Calculation B.3 Framatome ANP, 32-5017543-00, DB1- Damage Assessment of CRDM 3 B.4 Structural Integrity Associates Calculation W-DB-O1Q-301, Elastic Plastic Finite Element Stress Analysis of Davis-Besse Head Wastage Cavity.

B.5 NUREG/CR-5603, Pressure-Dependent Fragilities for Piping Components, October 1990.

B.6 NUREG/CR-5604, Assessment of ISLOCA Risk-Methodology and Application to a Babcock and Wilcox Nuclear Power Plant, April 1992.

Docket Number 50-346 License Number NPF-3 Serial Number 1-1268 Page 17 of 19 B.7 NUREG/CR-2300, PSA Procedures Guide, January 1983.

B.8 Calculation C-NSA-99.16-49, Revision 0, Success Criteria and Human Action Timing for Rx Head LOCA, March 2002.

B.9 Calculation C-NSA-99.16-050, Rev. 0, Risk Assessment for Reactor Head Wastage, April 2002.

B.10 Electric Power Research Institute (EPRI) Report TR-105396, PSA Applications Guide.

B.11 Davis-Besse PSA, Rev. 3.

B.12 BAW-2382, Level 3 PRA for Davis-Besse, January 2001.

B. 13 Regulatory Guide 1.174, An Approach for Using Probabilistic Risk Assessment in Risk Informed Decisions on Plant Specific Changes to the Licensing Basis, July 1998.

Docket Number 50-346 License Number NPF-3 Serial Number 1-1268 Page 18 of 19 Figure 1 - Exposed RV Cladding Area at CRDM Nozzle #3 7

SNozzle (Nozzle 1)

Noze1 Note: This drawing was based on a preliminary investigation of the exposed reactor vessel cladding area. For the small break LOCA evaluation, break areas anywhere from a pinhole sized opening in the exposed stainless steel cladding up to an area that is 1.5 times larger than the estimated area, plus the area of a CRDM nozzle, was considered.

Docket Number 50-346 License Number NPF-3 Serial Number 1-1268 Attachment 3 Page 19 of 19 Figure 2 - Davis-Besse Fuel and CRDM Locations (Cycle 14)

"X" Axis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 S 1 1 1 1 I I I '1 "WUI .Y1, Axis Axis "Z" Axis o Control Rod

• CRDM Nozzle with Indications O Open Control Rod Location *Continous RV Head Vent Location

• APSR Location rF Fresh Fuel Assembly

Docket Number 50-346 License Number NPF-3 Serial Number 1-1268 Framatome ANP Affidavit for Structural Integrity Associates, Inc. File W-DB-01Q-301 (Attachment 1)

(3 Pages Follow)

AFFIDAVIT COMMONWEALTH OF VIRGINIA )

) ss.

CITY OF LYNCHBURG )

1. My name is Raymond W. Ganthner. I am Vice President, Engineering &

Licensing for Framatome ANP ("FRA-ANP"), and as such I am authorized to execute this Affidavit.

2. I am familiar with the criteria applied by FRA-ANP to determine whether certain FRA-ANP information is proprietary. I am familiar with the policies established by FRA-ANP to ensure the proper application of these criteria.

3. I am familiar with the information contained in the calculation package prepared by SIA (File No. W-DB-01 Q-301) on the subject of "Elastic-Plastic Finite Element Stress Analysis of Davis Besse RPV Head Wastage Cavity." This package, known herein as "Document," contains information that is considered by Framatome ANP to be proprietary. Information contained in this Document has been classified by FRA-ANP as proprietary in accordance with the policies established by FRA-ANP for the control and protection of proprietary and confidential information.

4. This Document contains information of a proprietary and confidential nature and is of the type customarily held in confidence by FRA-ANP and not made available to the public. Based on my experience, I am aware that other companies regard information of the kind contained in this Document as proprietary and confidential.

5. This Document has been made available to the U.S. Nuclear Regulatory Commission in confidence with the request that the information contained in this Document be withheld from public disclosure.

6. The following criteria are customarily applied by FRA-ANP to determine whether information should be classified as proprietary:

(a) The information reveals details of FRA-ANP's research and development plans and programs or their results.

(b) Use of the information by a competitor would permit the competitor to significantly reduce its expenditures, in time or resources, to design, produce, or market a similar product or service.

(c) The information includes test data or analytical techniques concerning a process, methodology, or component, the application of which results in a competitive advantage for FRA-ANP.

(d) The information reveals certain distinguishing aspects of a process, methodology, or component, the exclusive use of which provides a competitive advantage for FRA-ANP in product optimization or marketability.

(e) The information is vital to a competitive advantage held by FRA-ANP, would be helpful to competitors to FRA-ANP, and would likely cause substantial harm to the competitive position of FRA-ANP.

7. In accordance with FRA-ANP's policies governing the protection and control of information, proprietary information contained in this Document has been made available, on a limited basis, to others outside FRA-ANP only as required and under suitable agreement providing for nondisclosure and limited use of the information.

8. FRA-ANP policy requires that proprietary information be kept in a secured file or area and distributed on a need-to-know basis.

9. The foregoing statements are true and correct to the best of my knowledge, information, and belief.

SUBSCRIBED before me this day of /, 2002.

lreiida C. 'lWjaddox NOTARY PUBLIC, STATE OF VIRGINIA MY COMMISSION EXPIRES: 07/31/03

Docket Number 50-346 License Number NPF-3 Serial Number 1-1268 Page 1 of 1 COMMITMENT LIST The following list identifies those actions committed to by the Davis-Besse Nuclear Power Station (DBNPS) in this document. Any other actions discussed in the submittal represent intended or planned actions the DBNPS. They are described only for information and are not regulatory commitments. Please notify the Manager - Regulatory Affairs (419-321-8450) at the DBNPS of any questions regarding this document or associated regulatory commitments.

COMMITMENTS DUE DATE None