ML070120395

SAND2007-0055, Structural Integrity Analysis of the Degraded Drywell Containment at the Oyster Creek Nuclear Generating Station.

ML070120395
Person / Time
Site:	Oyster Creek
Issue date:	01/12/2007
From:	Petti J Sandia
To:	Office of Nuclear Reactor Regulation
Dudley N, NRR/DLR/RLRA, 415-1154
References
DE-AC04-94AL85000 SAND2007-0055
Download: ML070120395 (102)

Text

SANDIA REPORT SAND2007-0055 Unlimited Release Printed January 2007 Structural Integrity Analysis of the Degraded Drywell Containment at the Oyster Creek Nuclear Generating Station Jason P. Petti Prepared by Sandia National Laboratories Albuquerque, New Mexico 87185 Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of Energys National Nuclear Security Administration under Contract DE-AC04-94AL85000.

Approved for public release; further dissemination unlimited.

Issued by Sandia National Laboratories, operated for the United States Department of Energy by Sandia Corporation.

NOTICE: This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government, nor any agency thereof, nor any of their employees, nor any of their contractors, subcontractors, or their employees, make any warranty, express or implied, or assume any legal liability or responsibility for the accuracy, com-pleteness, or usefulness of any information, apparatus, product, or process disclosed, or represent that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise, does not neces-sarily constitute or imply its endorsement, recommendation, or favoring by the United States Gov-ernment, any agency thereof, or any of their contractors or subcontractors. The views and opinions expressed herein do not necessarily state or reflect those of the United States Government, any agency thereof, or any of their contractors.

Printed in the United States of America. This report has been reproduced directly from the best available copy.

Available to DOE and DOE contractors from U.S. Department of Energy Office of Scientific and Technical Information P.O. Box 62 Oak Ridge, TN 37831 Telephone: (865)576-8401 Facsimile: (865)576-5728 E-Mail: reports@adonis.osti.gov Online ordering: http://www.osti.gov/bridge Available to the public from U.S. Department of Commerce National Technical Information Service 5285 Port Royal Rd Springfield, VA 22161 Telephone: (800)553-6847 Facsimile: (703)605-6900 E-Mail: orders@ntis.fedworld.gov Online order: http://www.ntis.gov/help/ordermethods.asp?loc=7-4-0#online 2

SAND2007-0055 Unlimited Release Printed January 2007 Structural Integrity Analysis of the Degraded Drywell Containment at the Oyster Creek Nuclear Generating Station Jason P. Petti Systems & Structures Department Sandia National Laboratories P.O. Box 5800 Albuquerque, NM 87185-0744 Abstract This study examines the effects of the degradation experienced in the steel drywell containment at the Oyster Creek Nuclear Generating Station. Specifically, the structural integrity of the con-tainment shell is examined in terms of the stress limits using the ASME Boiler and Pressure Ves-sel (B&PV) Code,Section III, Division I, Subsection NE, and examined in terms of buckling (stability) using the ASME B&PV Code Case N-284. Degradation of the steel containment shell (drywell) at Oyster Creek was first observed during an outage in the mid-1980s. Subsequent in-spections discovered reductions in the shell thickness due to corrosion throughout the contain-ment. Specifically, significant corrosion occurred in the sandbed region of the lower sphere.

Since the presence of the wet sand provided an environment which supported corrosion, a series of analyses were conducted by GE Nuclear Energy in the early 1990s. These analyses examined the effects of the degradation on the structural integrity. The current study adopts many of the same assumptions and data used in the previous GE study. However, the additional computa-tional recourses available today enable the construction of a larger and more sophisticated struc-tural model.

3

Acknowledgment The U.S. Nuclear Regulatory Commissions (NRC) Office of Nuclear Reactor Regulation (NRR) sponsored this analysis program under NRC Project J3312. The author would like to acknowl-edge Hansraj Ashar, Noel Dudley, Sujit Samaddar, Samir Chakrabarti, Donnie Ashley and Sally Adams of NRC for their technical and administrative oversight of this project. In addition, the author would like to thank Sandia management and colleagues, Michael Hessheimer, Matt Tur-geon, and Jeff Smith, for their helpful discussions and advice.

4

Contents Abstract...........................................................................................................................................3 Acknowledgment............................................................................................................................4 Contents ..........................................................................................................................................5 Figures.............................................................................................................................................6 Tables ..............................................................................................................................................8 Executive Summary .....................................................................................................................11

1. Introduction..............................................................................................................................15

2. Oyster Creek Drywell Finite Element Model........................................................................17 2.1 Finite Element Program and Modeling Procedures ......................................................17 2.2 Geometry .......................................................................................................................18 2.2.1 Drywell Head, Cylinder, Stiffeners, and Knuckle ..............................................20 2.2.2 Drywell Sphere & Personnel Lock/Equipment Door .........................................24 2.2.3 Ventline and Ventline Jet Deflector....................................................................27 2.3 Boundary Conditions.....................................................................................................29 2.4 Loading..........................................................................................................................34 2.4.1 General Loads: Gravity, Dead, Penetration, and Compressible Material Loads......................................................................................................................35 2.4.2 Seismic Load.......................................................................................................42 2.4.3 Refueling Condition Specific Loads: Live, External Pressure, and Refueling Loads......................................................................................................................43 2.4.4 Accident Condition Specific Loads: Internal Pressure and Thermal Loads.......45 2.4.5 Post-Accident Condition Specific Load: Hydrostatic Load ...............................45 2.5 Material Properties ........................................................................................................46 2.6 Degraded Model ............................................................................................................46 2.7 Mesh Size ......................................................................................................................50

3. Stress Analysis..........................................................................................................................55 3.1 Refueling Condition ......................................................................................................56 3.2 Accident Condition........................................................................................................58 3.3 Post-Accident Condition ...............................................................................................64 3.4 Conclusion.....................................................................................................................66

4. Stability Analysis......................................................................................................................67 4.1 Refueling Condition ......................................................................................................68 4.2 Post-Accident Condition ...............................................................................................72 4.3 Conclusion.....................................................................................................................76

5. Sandbed Region Minimum Thickness Study ........................................................................77

6. Summary of Assumptions .......................................................................................................81

7. Conclusions...............................................................................................................................83

8. References.................................................................................................................................85

9. Appendix A - Natural Frequency Extraction .......................................................................86

10. Appendix B - Sandbed UT Measurement Data and Shell Thickness Development .......91 5

Figures Figure 2-1. Oyster Creek Reactor Building and Containment......................................................19 Figure 2-2. Extent of Drywell and Ventlines Including the Current Model (Approximate Elevations) ...................................................................................................................20 Figure 2-3. Head and Cylinder Shell Thickness and Dimensions ................................................21 Figure 2-4. Model with Head Removed for Refueling .................................................................21 Figure 2-5. Cylinder Stiffener Layout ..........................................................................................22 Figure 2-6. Knuckle Region Shell Thickness ...............................................................................23 Figure 2-7. Upper and Middle Sphere Shell Geometry ................................................................24 Figure 2-8. Thickened Middle Sphere Geometry .........................................................................25 Figure 2-9. Personnel Lock and Equipment Hatch Geometry......................................................26 Figure 2-10. Lower and Bottom Sphere Geometry ......................................................................27 Figure 2-11. Drywell Geometry near Ventline Penetration..........................................................27 Figure 2-12. Ventline Geometry ...................................................................................................28 Figure 2-13. Ventline Deflector Geometry...................................................................................29 Figure 2-14. Boundary Condition at the Bottom of the Drywell with Cross-section View of Embedded Drywell Shell .............................................................................................30 Figure 2-15. Boundary Condition at the Ends of the Ventlines....................................................31 Figure 2-16. Ventline Header Submodel ......................................................................................32 Figure 2-17. Ventline Spring Locations .......................................................................................32 Figure 2-18. Boundary Condition at the End of the Hatch Penetration........................................33 Figure 2-19. Boundary Condition at the Stabilizers .....................................................................33 Figure 2-20. Upper and Lower Spray Header Locations..............................................................36 Figure 2-21. Weld Pad Locations .................................................................................................36 Figure 2-22. Flange and Stabilizer Locations...............................................................................37 Figure 2-23. Upper and Lower Beam Seat Locations ..................................................................38 Figure 2-24. Personnel Lock and Equipment Door Loads Application Region...........................38 Figure 2-25. Penetration Load Application Regions in the Drywell Sphere ................................40 Figure 2-26. Penetration Load Application Regions in the Drywell Cylinder .............................41 Figure 2-27. Elevation 16 Penetration Load Application Region Between the Ventlines..........41 Figure 2-28. Refueling Load on Drywell Cylinder.......................................................................44 Figure 2-29. Bay 1 UT Measurement Locations Taken from Outside of the Containment (Image Extracted from GPU Nuclear Calculation Sheet, 1993)..................................48 Figure 2-30. Lower Sphere Bay Combination Regions (Ventlines Removed for Clarity)...........48 Figure 2-31. Detailed View of Local Bay 1 Region (Ventline Removed for Clarity) .................49 Figure 2-32. Degraded Thicknesses in the Lower Sphere (inches) ..............................................50 Figure 2-33. Finite Element Mesh for the Refueling Load Case..................................................51 Figure 2-34. Finite Element Mesh in the Drywell Cylinder for the Refueling Load Case...........52 Figure 2-35. Finite Element Mesh for the Accident and Post-Accident Load Cases ...................52 Figure 2-36. Finite Element Mesh in the Drywell Cylinder and Head for the Accident and Post-Accident Load Cases ...........................................................................................53 Figure 2-37. Finite Element Mesh in the Upper and Middle Sphere............................................53 Figure 2-38. Finite Element Mesh in the Lower Sphere, Bottom Sphere, and Ventlines ............54 Figure 2-39. Finite Element Mesh for the Local Thin Regions under the Ventlines in Bay 1 and 13...........................................................................................................................54 6

Figure 3-1. Meridional Membrane Stress Distribution in the Lower Sphere for the Refueling Load Case with No Degradation (ksi) .........................................................................57 Figure 3-2. Meridional Membrane Stress Distribution in the Lower Sphere for the Refueling Load Case with Degradation (ksi) ...............................................................................57 Figure 3-3. Meridional Membrane Stress Distribution in Local Bay 13 Region for the Refueling Load Case with Degradation (ksi) ..............................................................58 Figure 3-4. Circumferential Membrane Stress Distribution in Sandbed for the Accident Load Case with No Degradation (Internal Pressure without Thermal Load) (ksi)...............61 Figure 3-5. Circumferential Membrane Stress Distribution in Sandbed for the Accident Load Case with No Degradation (Internal Pressure with Thermal Load) (ksi) ....................62 Figure 3-6. Circumferential Membrane Stress Distribution in Sandbed and Local Thin Region Under the Ventline in Bay 13 for the Accident Load Case with Degradation (Internal Pressure without Thermal Load) (ksi)......................................62 Figure 3-7. Circumferential Membrane Stress Distribution in Sandbed for the Accident Load Case with Degradation (Internal Pressure with Thermal Load) (ksi)..........................63 Figure 3-8. Meridional Membrane Plus Bending Stress Distribution (Tension on the Inside Surface of the Drywell Shell) in Sandbed for the Accident Load Case with No Degradation (Internal Pressure with Thermal Load) (ksi)...........................................63 Figure 3-9. Circumferential Membrane Stress Distribution in Sandbed for the Post-Accident Load Case with No Degradation (ksi) .........................................................................65 Figure 3-10. Circumferential Membrane Stress Distribution in Sandbed and Local Thin Region Under the Ventline in Bay 13 for the Post-Accident Load Case with Degradation (ksi) .........................................................................................................65 Figure 4-1. Buckling at the Upper Beam Seat for the Refueling Case with No Degradation ......68 Figure 4-2. Buckling in the Sandbed Region for the Refueling Case with No Degradation........69 Figure 4-3. Buckling at the Upper Beam Seat for the Refueling Case with Best Estimate Degradation..................................................................................................................70 Figure 4-4. Buckling in the Sandbed Region for the Refueling Case with Best Estimate Degradation..................................................................................................................71 Figure 4-5. Buckling in the Cylinder for the Post-Accident Load Case with No Degradation ....73 Figure 4-6. Buckling in the Sandbed Region for the Post-Accident Load Case with No Degradation..................................................................................................................74 Figure 4-7. Buckling in the Sandbed Region for the Post-Accident Load Case with Best Estimate Degradation...................................................................................................75 Figure 5-1. Drywell Lower Sphere for Establishing a Minimum Thickness in the Sandbed Region (Ventlines and Hatch Removed for Clarity) ...................................................78 Figure 5-2. Effective Factor of Safety Values Computed for Various Thicknesses in the Sandbed Region for the Refueling Load Combination................................................79 Figure 5-3. Buckling in the Sandbed Region with a Thickness of 0.844 for the Refueling Load Combination .......................................................................................................80 Figure 9-1. Modified Model for Natural Frequency Extraction ...................................................87 Figure 9-2. Lower Sphere Region (Highlighted in Red) Set to a Thickness of 0.835 for the Degraded Natural Frequency Extraction .....................................................................88 Figure 9-3. Base State and the First 5 Frequencies and Mode Shapes for the Drywell Containment with No Degradation ..............................................................................89 7

Figure 9-4. Base State and the First 5 Frequencies and Mode Shapes for the Drywell Containment with Degradation ....................................................................................90 Figure 10-1. Bay 1 and Bay 3 UT Measurement Locations Taken from Outside of the Containment (Images Extracted from GPU Nuclear Calculation Sheet, 1993) ..........91 Figure 10-2. Bay 1 UT Measurement Locations Taken from Outside of the Containment (Image Extracted from GPU Nuclear Calculation Sheet, 1993)..................................96 Figure 10-3. Bay 3 UT Measurement Locations Taken from Outside of the Containment (Image Extracted from GPU Nuclear Calculation Sheet, 1993)..................................96 Figure 10-4. Bay 5 UT Measurement Locations Taken from Outside of the Containment (Image Extracted from GPU Nuclear Calculation Sheet, 1993)..................................97 Figure 10-5. Bay 7 UT Measurement Locations Taken from Outside of the Containment (Image Extracted from GPU Nuclear Calculation Sheet, 1993)..................................97 Figure 10-6. Bay 9 UT Measurement Locations Taken from Outside of the Containment (Image Extracted from GPU Nuclear Calculation Sheet, 1993)..................................98 Figure 10-7. Bay 11 UT Measurement Locations Taken from Outside of the Containment (Image Extracted from GPU Nuclear Calculation Sheet, 1993)..................................98 Figure 10-8. Bay 13 UT Measurement Locations Taken from Outside of the Containment (Image Extracted from GPU Nuclear Calculation Sheet, 1993)..................................99 Figure 10-9. Bay 15 UT Measurement Locations Taken from Outside of the Containment (Image Extracted from GPU Nuclear Calculation Sheet, 1993)..................................99 Figure 10-10. Bay 17 UT Measurement Locations Taken from Outside of the Containment (Image Extracted from GPU Nuclear Calculation Sheet, 1993)................................100 Figure 10-11. Bay 19 UT Measurement Locations Taken from Outside of the Containment (Image Extracted from GPU Nuclear Calculation Sheet, 1993)................................100 Tables Table 2-1. Cylinder Stiffeners .......................................................................................................22 Table 2-2. Load Combination Components...................................................................................34 Table 2-3. Dead Load Tractions ....................................................................................................39 Table 2-4. Penetration Load Tractions ..........................................................................................42 Table 2-5. Live Load Tractions .....................................................................................................44 Table 2-6. Main Drywell Shell Model Thicknesses, Original and Degraded ...............................47 Table 2-7. Degraded Lower Sphere Shell Model Thicknesses......................................................49 Table 3-1. Refueling Load Case Peak Stresses with No Degradation, Primary Stresses (Percentage of ASME Limit in Parenthesis)................................................................56 Table 3-2. Refueling Load Case Peak Stresses with Degradation, Primary Stresses (Percentage of ASME Limit in Parenthesis)................................................................56 Table 3-3. Accident Load Case Peak Stresses with No Degradation, Primary Stresses (Percentage of ASME Limit in Parenthesis)................................................................59 Table 3-4. Accident Load Case Peak Stresses with No Degradation, Primary + Secondary Stresses (Percentage of ASME Limit in Parenthesis)..................................................60 Table 3-5. Accident Load Case Peak Stresses with Degradation, Primary Stresses (Percentage of ASME Limit in Parenthesis)................................................................60 Table 3-6. Accident Load Case Peak Stresses with Degradation, Primary + Secondary Stresses (Percentage of ASME Limit in Parenthesis)..................................................60 8

Table 3-7. Post-Accident Load Case Peak Stresses with No Degradation, Primary Stresses (Percentage of ASME Limit in Parenthesis)................................................................64 Table 3-8. Post-Accident Load Case Peak Stresses with Best Estimate Degradation, Primary Stresses (Percentage of ASME Limit in Parenthesis)..................................................64 Table 4-1. Buckling Evaluation at the Upper Beam Seat for the Refueling Load Case with No Degradation..................................................................................................................69 Table 4-2. Buckling Evaluation in the Sandbed Region for the Refueling Load Case with No Degradation..................................................................................................................70 Table 4-3. Buckling Evaluation at the Upper Beam Seat for the Refueling Load Case with Best Estimate Degradation...........................................................................................71 Table 4-4. Buckling Evaluation in the Sandbed Region for the Refueling Load Case with Best Estimate Degradation...........................................................................................72 Table 4-5. Buckling Evaluation in the Cylinder for the Post-Accident Load Case with No Degradation..................................................................................................................73 Table 4-6. Buckling Evaluation in the Sandbed Region for the Post-Accident Load Case with No Degradation............................................................................................................75 Table 4-7. Buckling Evaluation in the Sandbed Region for the Post-Accident Load Case with Best Estimate Degradation...........................................................................................76 Table 5-1. Main Drywell Shell Model Thicknesses Outside of Sandbed Region .........................78 Table 5-2. Buckling Evaluation for the Refueling Load Case with a Thickness of 0.844 in the Sandbed..................................................................................................................80 Table 7-1. Comparison of Conclusion Between GE Study (GE, 1991a and b) and the Current Study ............................................................................................................................83 Table 9-1. Drywell Shell Thicknesses for Natural Frequency Extraction Analyses .....................87 Table 9-2. Summary of the First 5 Natural Frequencies for Drywell with and without Degradation..................................................................................................................90 Table 10-1. UT Measurement Data for Bay Combinations 1-3, 3-5, 5-7, and 7-9........................92 Table 10-2. UT Measurement Data for Bay Combinations 9-11, 11-13, and 13-15. ....................93 Table 10-3. UT Measurement Data for Bay Combinations 15-17, 17-19, and 19-1. ....................94 Table 10-4. UT Measurement Data for Local Bay 1 and 13 Regions. ..........................................95 9

(Page Intentionally Left Blank) 10

Executive Summary The Oyster Creek Nuclear Generating Station is a GE Mark I BWR which began operation in 1969. It is located in New Jersey and is operated by AmerGen/Exelon. The drywell portion of the containment vessel consists of a free-standing welded steel shell with an upper cylindrical section atop a lower spherical section. The steel containment rests on a reinforced concrete base mat and is surrounded by a reinforced concrete reactor building.

Corrosion of the steel drywell containment shell at Oyster Creek was first observed during an outage in November 1986 (GE, 1991a). Subsequent inspections discovered reductions in the shell thickness due to general corrosion in many regions of the drywell containment. Significant corrosion occurred in the sandbed region of the lower sphere. The sandbed is located below the ventlines that lead down to the torus section of the containment and just above the concrete base mat. A small pocket of sand was originally placed adjacent to the steel shell at the base to pro-vide a transition, or cushion, as the shell emerges from being embedded in concrete. Inspec-tions concluded that water leakage occurred through the gap between the reactor building and the drywell shell and collected in the sandbed region. Since the wet sand provided an environment which supported corrosion, the Licensee embarked on a series of corrective actions including removing the sand from the sandbed region, cleaning and coating the affected surfaces, and seal-ing the gap between the containment vessel and the concrete to prevent further penetration by water. The Licensee also implemented periodic re-inspections of selected areas of the vessel to monitor the progression, if any, of the corrosion damage.

Prior to the removal of the sand from the sandbed region, the Licensee tasked GE Nuclear with assessing the vessel in its degraded state to determine whether or not the degradation prevented the vessel from performing its intended design function. They concluded that the degraded dry-well shell, with the sand removed, still satisfied the ASME Boiler & Pressure Vessel (B&PV)

Code stress and stability limits, albeit with a reduced design pressure. The sand was removed and based on subsequent inspections, the Licensee has claimed that there is no on-going corro-sion in the sandbed region of the drywell shell. Inspections have, however, discovered ongoing corrosion in the portions of the drywell above the sandbed region (sphere and cylinder).

In July of 2005, the Licensee submitted an application to the U.S. Nuclear Regulatory Commis-sion (NRC) to extend the operating life of the plant from 40 to 60 years (extend from 2009 to 2029). The NRC Office of Reactor Regulation (NRR) commissioned Sandia National Laborato-ries (SNL) to perform an evaluation of the degraded containment vessel to determine if the Li-censees contention, that the current known condition of the vessel and the progressive damage expected over the extended service life did not compromise the design function or licensing ba-sis, was reasonable. The scope of the analyses performed by Sandia was defined by NRC staff and the procedures employed were discussed with NRC staff throughout the project.

In this evaluation, Sandia developed a detailed three-dimensional (3D) finite element model of the drywell containment vessel using information provided by the NRC and the Licensee.

Analyses for the governing load combinations were performed for the vessel in its original, as-designed state and for a representation of the vessel in an approximation of the current degraded state. Based on previous work performed at Sandia (Cherry and Smith, 2001, Spencer et. al, 11

2006), modeling of the corrosion damage was represented by uniform shell thinning. The de-graded condition of the sandbed region in the model is based on the measurements performed in 1993 (GPU Nuclear, 1993). These measurements were taken prior to the application of the pro-tective coating. The shell thicknesses of the model in the sandbed region are based on averages of the available measurements. Assuming these measurements made in the accessible portions of the sandbed are representative of the entire region, the average of the measurements should be conservatively biased since the thickness measurements were only made at the thinnest points (by visual inspection). No statistical analysis of the Licensees in-situ thickness measurements was performed. Rather, the averaging procedure used to develop thicknesses was based on engi-neering judgment. No additional reduction in thickness due to ongoing corrosion during the 20-year plant life extension was considered in the sandbed region, accepting the Licensees conten-tion that corrosion processes have been arrested. The thicknesses in the upper portions of the degraded drywell model were based on the additional thickness measurements performed by the Licensee over the past 20 years and included an estimate of future corrosion by linear extrapola-tion of past corrosion rates.

The models were then used to evaluate the structural integrity of the vessel in terms of the stress limits specified in the ASME Boiler and Pressure Vessel (B&PV) Code,Section III, Division I, Subsection NE, and in terms of buckling (stability) limits specified in ASME B&PV Code Case N-284. The analyses performed in this study aim only to independently confirm the general con-clusions reached in a previous study performed by GE Nuclear Energy in the early 1990s. Two important points regarding the current analysis are important to recognize:

• The original design of the containment based on the analyses by the Licensee and GE and subsequent analyses of the degraded vessel have been accepted by the NRC and are part of the current licensing basis.

• The current analysis by Sandia cannot, and is not intended to, reproduce the results of the original licensing basis analyses. As such, the baseline (i.e. un-degraded) analysis was performed so that the effects of the degradation could be clearly isolated. The results of the current analysis should, therefore, focus more on the relative reduction in design mar-gin due to the corrosion modeled, than the absolute stresses or stability limits which are calculated. This relative reduction in margin, examined together with the current licens-ing basis and additional relevant information, should be considered by the NRC staff in the development of the basis to accept or reject the Licensees application for an ex-tended license. By itself, the analysis performed by Sandia cannot be used for this deci-sion.

A significant amount of data, primarily regarding the external loads on the drywell shell, was extracted directly from the GE analyses due to insufficient plant information to allow independ-ent calculation of these loads. Every effort was made to use the best available information for the current models and analyses. However, since the GE analyses and the current analyses use a different modeling approach, the data taken directly from the GE analysis was of necessity modi-fied to fit the current approach.

12

The purpose of the Sandia analyses was to assess the effects of degradation on the stress and buckling behavior for the drywell containment. In this context, the results of the analyses show that the degradation does not result in a definitive violation of the stresses or buckling criterion in the ASME code given the modeling procedures and assumptions outlined in this report.

13

(Page Intentionally Left Blank) 14

1. Introduction This study examines the effects of the degradation experienced in the steel drywell containment at the Oyster Creek Nuclear Power Plant. Specifically, the structural integrity of the containment shell is examined in terms of the stress limits using the ASME Boiler and Pressure Vessel (B&PV) Code,Section III, Division I, Subsection NE, and examined in terms of buckling (sta-bility) using the ASME B&PV Code Case N-284.

The analyses performed in this study aim to independently confirm the general conclusions reached in a previous study performed by GE Nuclear Energy in the early 1990s. Since the GE analyses and the analyses performed here use different models, and in some cases, different as-sumptions, a direct comparison to the previous GE analysis is not the intent of this effort. In ad-dition, a significant amount of data was taken directly from the GE analysis and applied or modified as required for the current study. This was necessary when information was not avail-able, or was not made available, to be independently verified. Within the project schedule, all efforts were made to use the best available information for the models and analysis used in the current study. All stress and buckling analyses were performed for both a representation of the containment in its degraded condition and in its original, as-built, condition. The study of the as-built conditions provides base-line analyses to assess the effects of degradation on the stress and buckling behavior for the containment.

Degradation of the steel drywell containment shell at Oyster Creek was first observed during an outage in November 1986 (GE, 1991a). Subsequent inspections discovered reductions in the shell thickness due to corrosion throughout the containment. Specifically, significant corrosion occurred in the sandbed region of the lower sphere. The sandbed is located below the ventlines that lead down to the torus section of the containment. The small pocket of sand was originally placed adjacent to the steel shell at the base to provide a transition as the shell emerges from be-ing embedded in concrete. Water leakage through the gap between the reactor building and the drywell shell collected in the sandbed region. Since the presence of the wet sand provided an environment which supported corrosion, a series of analyses were conducted by GE Nuclear En-ergy to examine the effects of removing the sand. GE determined that the degraded drywell shell with removal of the sand was acceptable based on ASME B&PV stress and stability limits.

Therefore, the sand was removed and the surface of the drywell shell epoxy coated to protect the surface from additional degradation. Subsequent inspections have supported the claim that there is no on-going corrosion in the sandbed region of the drywell shell. However, inspections have shown the existence of ongoing corrosion in the upper portions of the drywell (sphere and cylin-der).

Thickness measurements have been performed during refueling outages at the plant over the last 20 years. The UT measurement data used to estimate the thickness of the containment shell was limited to a few selected regions in the sandbed and throughout the remaining containment.

Since only a very small percentage of the total shell surface has been measured, a number of as-sumptions were made in this study to assign appropriate shell thicknesses throughout the drywell model. These are described in more detail in subsequent sections.

15

The degraded Oyster Creek drywell shell was analyzed in this study using a full three-dimensional (3D) finite element model. The previous analyses by GE employed both an axi-symmetric and a 36o slice model of the drywell. These analyses were conducted in the 1990-91 timeframe and were constrained by the computational limits of the day. Due to a significant in-crease in computational power relative to the time of the GE analysis, a full 3D model was cre-ated here and is described in detail in this report.

2. Oyster Creek Drywell Finite Element Model A full three-dimensional (3D) finite element model of the Oyster Creek drywell was developed for this study. A full 3D, 360o, model enables a more sophisticated analysis which includes structural detail that account for the asymmetries of the containment vessel. It also provides for a more realistic representation of the boundary conditions, thicknesses transitions, and the spatial variation of the degradation.

Two reports summarizing the work performed by GE (GE, 1991a and 1991b) along with a par-tial set of drywell structural drawings (CB&I, 1980) were the two resources used to develop the model geometry. Unfortunately, many of the resources available to the GE analysts were not available, or were not made available in time for use in this study. In a number of instances, this has led to the need to assume information required to complete this program. For example, many items related to the structural loads documented by GE could not be confirmed or recre-ated. In these cases, the information that was available from the GE study and/or other sources was used, combined, or adapted for use in the current analysis. These assumptions and proce-dures are documents throughout this report, and are summarized in a section at the end of this report.

2.1 Finite Element Program and Modeling Procedures The finite element modeling conducted in this study uses the ABAQUS (ABAQUS, 2004) suite of analysis software. Specifically, Version 6.5-6 of the ABAQUS/Standard general-purpose fi-nite element program and the ABAQUS/CAE interactive environment are used to perform the analyses and to create the solid models and finite element meshes, respectively.

ABAQUS/Standard is employed since all of the analyses performed here are static. The CAE component of ABAQUS provides an interface for defining the model geometry, material proper-ties, shell thicknesses, boundary conditions, loadings, and meshing. After the analysis is com-pleted using ABAQUS/Standard, the Visualization module within CAE (also identified as ABAQUS/Viewer) is used to examine the analysis results.

The analyses performed here include geometric nonlinearities, also known as large-displacement or finite strain analyses. When applying geometric nonlinearities to the analysis, the element formulation at each load step is performed using the current configuration (e.g. deformed shape).

A combination of standard, S4R, 4-noded, and S3R, 3-noded, reduced integration shell ele-ments are used here to model the drywell. The meshing technique used is identified as quad-dominated in ABAQUS/CAE. The method meshes the geometry using quad (4-noded) ele-ments, but does introduce tri (3-noded) elements in regions where introducing a quad element would result in a severely distorted element.

Shell elements are used in modeling when the thickness dimension is significantly less than the in-plane dimensions. Typically, the reference surface of the shell element is set at the mid-section, or centerline, of the structure being modeled. The thickness of the shell is set in the 17

Section definitions within ABAQUS. Each nodes in a given shell elements have six degrees-of-freedom, three translational and three rotational.

The use of shell elements introduces discontinuities at the interface between plates of differing thickness. The actual structure also included discontinuities at these locations due to the inter-face of plates of differing thickness. These interfaces often include a small tapered region. Here, the thicker plate is gradually reduced in thickness over a length on the order of the plate thick-ness, and welded to the thinner plate. In the models developed in this study, a small region is included at the interface of plates of differing thickness to represent the transition region in the actual structure. This transition region is set to a thickness equal to the average of the plates on either side. The length of the model transition is based on the actual, or estimated, transition length given in the structural drawings (CB&I, 1980).

2.2 Geometry The Oyster Creek reactor building contains a GE BWR Nuclear Steam Supply System with a steel Mark I containment vessel. Figure 2-1 illustrates the pressure suppression system which includes the pressure suppression chamber (torus) and the drywell (containment vessel) con-nected with a series of ventlines. Figure 2-1 also shows the positioning of the containment ves-sel within the reactor building (one half of the reactor building has been removed to view the containment vessel) and a detailed view of the sandbed region below the ventlines. Since the drywell is not exactly a symmetric structure, it is modeled in full for this study. The series of ventlines which connect the drywell with the torus includes a flexible bellow (not shown) at the interface between the ventline and the torus. Since these bellows prevent significant structural interaction, the torus shell was not included in the model and is shown in Figure 2-1 for illustra-tive purposes only. As stated previously, the ventlines are modeled down past the interface with the torus, ending at the intersection with the ventline header.

Figure 2-2 shows the extent of the structure modeled for the current analysis. As stated above, the torus is not included in the model. The drywell is modeled from an elevation of 2-3 (2 feet, 3 inches) to an elevation of 107-9. At the top of the drywell, the head region is a 2:1 ellipse.

Below the head, the drywell cylinder has an inside diameter of 33 feet (33) and the drywell sphere has an inside diameter of 70 feet (70). The cylindrical and spherical regions are joined by a thickened knuckle. The equator of the drywell sphere is located at elevation 37-3. The largest drywell penetration is the personnel lock/equipment hatch located at an elevation of 27-

6. The centerline of the ventlines extends down to an elevation of 0-6. The sandbed region is located in the lower sphere of the drywell shell just below the ventlines. Below the sandbed, part of the lower sphere and the entire bottom sphere are completely contained within concrete on both sides below elevation 8-11.25 (lower sphere extends down to elevation 6-10.25). Addi-tional details related to the geometry, shell thickness, boundary conditions, and loadings are pro-vided throughout the next several subsections. The plate thicknesses given in these sections are for the drywell in its as-built state. The thinning due to the corrosion that exists in the shell is described in Section 2.6 18

Reactor Building (one half removed to view containment)

Ventline Removed to View Sandbed Region Drywell Elevation 8-11.25 Sandbed Region Torus Drywell (head)

Drywell (cylinder)

Drywell (knuckle)

Drywell (sphere)

Ventlines (10 at 36o)

Torus (not modeled)

Figure 2-1. Oyster Creek Reactor Building and Containment 19

Elevation 107-9 Head Iso Elevation 94-9 View Cylinder 0o Azimuth 90o Azimuth Elevation 71-10 270o Azimuth Elevation 65-4 Knuckle Elevation 50-11 Upper Sphere Middle Sphere Elevation 37-3 Personnel Lock Elevation 27-6 & Equipment Hatch Lower Sphere Elevation 23-6 Elevations Ventlines 8-11.25 (10 at 36o) 6-10.25 Ventlines End at 2-3 Ventline Header Sandbed Region Bottom Sphere Figure 2-2. Extent of Drywell and Ventlines Including the Current Model (Approximate Elevations) 2.2.1 Drywell Head, Cylinder, Stiffeners, and Knuckle Figure 2-3 shows the drywell head residing at the top of the structure up to an elevation of 107-

9. The 2:1 ellipse that defines the geometry of the head region extends down to an elevation of 99-6. The head region has a shell thickness of 1.1875. In the region below the head, the flange assembly includes a double tongue-and-groove seal at an elevation of 94-9. At this ele-vation, the head separates from the drywell during refueling as shown in Figure 2-4. For the analyses of the refueling load case, a separate model was created that has an identical geometry to the full model with the exception of the head being removed. In the full model, the flange as-sembly region is assigned the same thickness at the head, 1.1875. The geometry of the flange assembly is complex with the actual thickness varying from 1.25 to 1.5.

Since the thickness dimension is not represented when using shell elements, the location of the shell in the model is defined in space at the mid-section of the actual shell. This leads to the ra-dius of the flange assembly to be 16-6.59375. This number is computed by adding the actual inside radius in this region, 16-6, to one half of the shell thickness, or 1.1875/2.

Underneath the flange assembly, the shell thickness is reduced to 0.64 below elevation 92-2.75. The model also includes a thin transition region between the flange assembly region 20

and the lower cylinder. In the actual structure, the steel plate is tapered from one thickness to the next over a short distance. The transition region represents this tapered region and is assigned a thickness equal to the average, 0.91375, of the two surrounding plates (e.g. 1.1875 and 0.64).

Since the inside radius of the cylinder remains constant and the thickness of the lower cylinder is less than the flange assembly region, the centerline of the shell is shifted inward producing a ra-dius of 16-6.32.

Figure 2-3. Head and Cylinder Shell Thickness and Dimensions Figure 2-4. Model with Head Removed for Refueling 21

The cylinder region of the drywell also contains several stiffeners. Figure 2-5 and Table 2-1 summarizes the stiffeners dimensions and positions. Figure 2-5 gives an inside cut view of the cylinder. Half of Stiffener-0 resides within the cylinder and half resides outside the cylinder.

Stiffeners 1, 3, 4, and 5 are positioned completely within the cylinder. Only Stiffener-2 and 2a are attached completely to the outside of the cylinder. Stiffener-2 is connected directly to the outside surface of the cylinder shell. Stiffener-2a is thinner than Stiffener-2 and is attached to the outer extent of Stiffener-2.

Figure 2-5. Cylinder Stiffener Layout Table 2-1. Cylinder Stiffeners Stiffener Elevation Length (inches) Thickness (inches) Orientation Stiffener-0 96-7.875 12.5 2.25 half & half Stiffener-1 94-3 12 1.0 inside Stiffener-2 92-8.5 7 2.75 outside Stiffener-2a 92-8.5 7.38 1.0 outside Stiffener-3 88-8.5 6 0.5 inside Stiffener-4 84-11.8 6 0.75 inside Stiffener-5 80-6.3 6 0.75 inside 22

The knuckle illustrated in Figure 2-6 connects the drywells cylindrical region to the upper sphere. A thin transition region is introduced between the cylinder and knuckle and between the knuckle and upper sphere. The upper fillet portion of the knuckle has a 72 radius. Below an elevation of 66-5.77, the knuckle fillet is joined to the upper sphere with a linear section of the knuckle. The thickness of the entire knuckle (elevation 65-4.27 to 71-6.28) is set at 2.5625.

This is the minimum specified thickness in this region as stated in the previous GE study (GE, 1991a). However, the structural drawings (CB&I, 1980) and other sections of the GE study in-dicate a knuckle thickness of 2.625. The lower value of 2.5625 is adopted for the undegraded thickness of the knuckle since that value was confirmed1 and is more conservative.

Figure 2-6. Knuckle Region Shell Thickness 23 1

June 21, 2006, conference call between Sandia National Laboratories, NRC, and Exelon.

2.2.2 Drywell Sphere & Personnel Lock/Equipment Door The largest section of the drywell is the spherical region which lies below the cylinder. The sphere has an inside radius of 35 and is composed of four main regions of different thickness.

Figure 2-7 shows the upper and middle sphere regions. The upper sphere has a thickness of 0.722 and the middle sphere was constructed with a thickness of 0.77. As mentioned previ-ously, the position of the shell in the model created here is set at the mid-section of the shell in the actual structure. Therefore, the radii of the upper and middle sphere are 35-0.361 and 35-0.385, respectively. The 0.746 thick transition region between the upper and middle sphere lies between elevations 50-11.25 and 50-10.8. At the lower extent of the middle sphere, a 0.962 transition region connects the middle sphere with the 1.154 thick lower sphere between elevations 23-6.74 and 23-4.82.

Figure 2-7. Upper and Middle Sphere Shell Geometry 24

A section of the middle sphere is thickened between azimuths 2.5o and 317.5 o due to the pres-ence of the personnel lock and equipment hatch penetration as shown in Figure 2-8. (The values for the azimuths were assumed from an examination of the structural drawings (CB&I, 1980).)

This thickened region is 1.0625 and extends from the lower sphere to the upper sphere (23-6.74 to 50-10.8). Transition regions surround the thickened middle sphere on all sides. The transition along the top is 0.89225, along the vertical sides is 0.91625, and along the bottom (outside of the hatch) is 1.10825. There are also two small transition regions at the top corners (0.819125) and two small transition regions at the bottom corners (1.035125) of the thickened middle sphere. The thickness of these corner regions are weighted averages of the surrounding plates.

2.5o Azimuth 317.5o Azimuth Elevation 50-11.25 50-10.8 Middle Sphere Thickened Middle Sphere Thickness = 0.77 Thickness = 1.0625 Azimuth Middle Personnel Lock 342o Sphere

& Equipment Hatch Thickness Elevation = 0.77 27-6 23-6.74 23-4.82 Lower Sphere Thickness = 1.154 Figure 2-8. Thickened Middle Sphere Geometry Figure 2-9 illustrates the personnel lock and equipment hatch penetration. The penetration is 10 in diameter and extends from the thickened middle sphere down into the lower sphere. The cen-ter of the penetration is located at an elevation of 276 and an azimuth of 342o. Embedded within the drywell shell and surrounding the penetration is a 2.625 thick plate. The outer di-ameter of this thickened region is approximately 14-1.5. A thin transition region lies between this thickened plate surrounding the penetration and the surrounding thickened middle sphere (t

= 1.84375) and lower sphere (t = 1.8895).

25

Figure 2-9. Personnel Lock and Equipment Hatch Geometry The penetration extends away from the drywell shell to a distance of 41-6 from the centerline of the drywell. This is the location of a vertical support within the reactor building. This is dis-cussed in additional detail in the following section on boundary conditions. The penetration has a thickness of 2.625 at the connection with the drywell shell. The outer 5-9 length of the penetration has been set to a thickness of 0.5. In the actual structure, the thickness of this outer region varies and has been set to 0.5 to simplify the model. Only this outer shell of the penetra-tion is modeled here. The internals of the personnel lock and equipment hatch are included through applied loads and are described in the loading section.

Below the middle sphere and the hatch penetration is the 1.154 thick lower sphere region of the drywell as shown in Figure 2-10. The lower sphere extends from an elevation of 23-4.82 down to 6-10.25 and has a radius of 35-0.577. The section of the lower sphere below an elevation of 8-11.25 is embedded within concrete on both sides. The lowest extent of the drywell is the bottom sphere with a thickness of 0.676 and a radius of 35-0.338. The entire bottom sphere is also embedded within concrete. The sandbed region is located at the bottom of the lower sphere, from elevation 8-11.25 up to 12-3.

26

Figure 2-10. Lower and Bottom Sphere Geometry 2.2.3 Ventline and Ventline Jet Deflector Within the lower sphere of the drywell, 10 ventlines spaced at 36o connect the torus to the dry-well. As shown in Figure 2-11, the elevation of the center of the ventline penetration into the drywell shell is 15-6.8. (The actual elevation is 15-7.25. The difference is due to round-off error in constructing the geometry). The ventline is 7-10 in diameter at the intersection with the drywell shell and transitions down to a diameter of 6-6.25. As with the personnel lock and equipment hatch penetration, the thickness of the drywell shell surrounding the ventline penetra-tion is thickened. Here the thickened region is 2.875 with a thin transition zone of 2.0145.

Figure 2-11. Drywell Geometry near Ventline Penetration 27

Figure 2-12 illustrates the extent of the ventline modeled in this study. At the intersection with the drywell, a 2.5 thick section of the ventline at a diameter of 7-10 extends approximately 1-3.2 away from the drywell shell. The diameter of the ventline then transitions down to 6-6.25 with a 0.4375 thick region which extends approximately 1-11.8. A 0.25 thick region then extends another 14-0.7 to a 0.3125 thick section. This section extends approximately 4-2.3 to a point where the angle of the ventline changes from 38o21 to 17o from horizontal. The next section of 0.3125 thick ventline is approximately 2-5.2 in length with a 0.25 thick section extending the final 4-1. The center of the end of the ventline is at an elevation of 0-6. The ventline ends at the connection with the ventline header. Springs are attached to the end of the ventline to account for the additional stiffness provided by the ventline header. This is discussed in detail in the next section. Part of the lower section of ventline modeled here is actually con-tained within the torus and connected with a bellow. It is assumed that the bellows prevent any meaningful structural interaction, and therefore the torus and bellow are not modeled here.

Figure 2-12. Ventline Geometry Figure 2-13 shows the ventline jet deflector included in the current model. The deflector in-cludes 20 - 0.875 thick gusset plates that connect the inside of the drywell shell to the 2.31 thick deflector plate. The thickness of the gusset plate could not be identified on the structural drawings (CB&I, 1980), and was taken from the value given in the GE report (GE, 1991b). The actual deflector plate is 2.5 thick and includes 189 holes through the thickness of the plate.

Since including the holes explicitly is beyond the fidelity of this model, the plate was modeled as solid with a reduced thickness to maintain a constant volume with the actual plate. This reduced thickness solid plate approximates the membrane stiffness exhibited by the perforated plate due to the consistent cross-sectional area (on average).

28

Figure 2-13. Ventline Deflector Geometry The only penetrations explicitly modeled here are the ventlines and the personnel lock and equipment hatch. Other penetrations are included through loads applied to the structure and are discussed in the subsequent loading section.

2.3 Boundary Conditions The boundary conditions applied to the current model attempt to approximate the conditions within the actual structure. At the same time, it must be acknowledged that all finite element models are idealizations. The boundary conditions that can be applied to a given model, while increasingly realistic and complex, will never exactly represent the complexities in an actual structure.

Here, four boundary condition regions have been created and applied to the model. Figure 2-14 shows the fixed region of the drywell shell below elevation 8-11.25. This region is fixed since the drywell shell is surrounded by concrete on both sides. Outside of the drywell, concrete rises up to an elevation of 8-11.25. Above the concrete, the sandbed region extends up to an eleva-tion of 12-3. The sand has been removed from this region and is currently open space. Within the interior of the drywell shell, a concrete floor extends up to an elevation of 10-3 with curbs extending up to 11-0 below the ventlines and up to 123 between the ventlines. Since the cur-rent state of the bond between the drywell shell and the concrete inside of the drywell is not know to the analyst and because of the absence of concrete outside of the drywell shell, the con-crete inside of the drywell above an elevation of 8-11.25 is not accounted for in the model.

This is believed to be a realistic assumption since the shell deforms outward, away from the inte-rior concrete, for the load cases examined in this study.

29

Concrete Curb Drywell Shell Inside of Shell Elevation 10-3 Elevation 8-11.25 Sandbed Region Concrete Outside of Shell Elevation 8-11.25 Fixed Region Figure 2-14. Boundary Condition at the Bottom of the Drywell with Cross-section View of Embed-ded Drywell Shell Figure 2-15 shows the highlighted ends of the ventlines where the degrees of freedom are fixed against rotation and lateral displacement. Springs are attached to the ends of the ventlines in the vertical and radial directions. Since the spring constant used by GE at the ends of the ventlines to represent the compliance of the ventline header connection were not documented, a sub-analysis of the ventline header was performed for this study to estimate the stiffness provided to the ends of the ventlines. Figure 2-16 shows the ventline header and the submodel used to de-termine the spring constants to be applied to the ends of the ventlines in the main model. A sec-tion of the ventline header was extracted and analyzed with symmetry boundary conditions at one end and fixed displacement at the location of the ventline header columns. The end of the ventline header submodel that intersects with the ventline is fixed laterally and unit displace-ments are imposed in the radial and vertical directions. The reactions along this edge are summed and multiplied by two to account for the section of the ventline header on the other side of the ventline. The summed reactions in the radial and vertical directions are the resistance that would be applied to the ventline from the ventline header. The springs acting vertically are ap-plied at two points with magnitudes of 2332 kips/in. The vertical springs are located on each 30

side of the end of the ventline as shown in Figure 2-17. Figure 2-17 also shows the springs that act radially at the top and bottom of the end of the ventline. These springs have a magnitude of 519.9 kips/in. These points of application were selected since the largest reactions resisting the imposed displacements are the located near these locations.

Fixed Against Rotation and Lateral Displacement, Springs Act Against Radial and Vertical Displacement Figure 2-15. Boundary Condition at the Ends of the Ventlines 31

Ventline Header Locations of Ventline Intersection with Ventline Header Fixed Against Lateral Displacement Symmetry Boundary Conditions Fixed Against Vertical Displacement Imposed Displacement in Radial and Vertical Directions Figure 2-16. Ventline Header Submodel Locations of Locations of Vertical Springs Radial Springs Figure 2-17. Ventline Spring Locations 32

The outer extend of the personnel lock and equipment hatch is shown in Figure 2-18. The end of the penetration included in the model extends 41-6 from the centerline of the drywell. At this point, the penetration reaches a roller support within the reactor building. The end of the hatch is constrained against vertical displacement at this point.

Fixed Against Vertical Displacement Elevation 27-6 Figure 2-18. Boundary Condition at the End of the Hatch Penetration Finally, Figure 2-19 illustrates the boundary condition at the seismic lateral stabilizers. These stabilizers are centered at an elevation of 82-9 and have a diameter of 5-3. There are 8 stabi-lizers spaced at 45o around the circumference of the drywell cylinder. The structural details in these regions allow the steel shell to move radially and vertically, but constrain the shell against lateral displacement. Lateral motion for a cylindrical shell can be described as a twisting or rota-tion in the azimuth direction (see CB&I, 1980, for structural detail).

Figure 2-19. Boundary Condition at the Stabilizers 33

2.4 Loading The load combinations for the Oyster Creek drywell stress and stability analyses are provided in the Technical Specification for Primary Containment Analysis - Oyster Creek Nuclear Generat-ing Station (Reference 1-4 of GE, 1991a) and summarized in the previous GE analyses (GE 1991a and b). Based on the detailed discussion of the different load combinations in the GE re-ports and the previous acceptance of their calculations, the following three load combinations are explored in this study:

• Case IV - Refueling Condition,

• Case V - Accident Condition,

• Case VI - Post-Accident Condition.

GE determined that these three load combinations essentially envelope all other scenarios, and therefore, define the governing set of load combinations. Stress analyses are performed for all three of the above load combinations. In addition, only Case IV - Refueling Condition, and Case VI - Post-Accident Condition, are examined for the stability (buckling) analysis. The cur-rent analysis assumes that these two conditions govern the potential buckling in the sandbed re-gion since the accident condition does not produce significant compressive stresses in the containment.

Each of the above load combinations includes a specific set of load types. Among these, the dead, live, and equipment loads were applied in the GE analysis using calculated loads from an earlier study by Chicago Bridge & Iron (Reference 2.4.3 of Reference 1-4 of GE, 1991a). This reference was not made available for the current study, and therefore, the loads documented by GE (Tables 2-5a through 2-5c of GE, 1991a) were adapted and applied to the current model.

In addition, to the loads mentioned above, several other load types are required to complete the load combinations of interest. These include seismic, water loads, and internal pressure, among others. The set of loads applied for each load combination was extracted from the previous GE analysis (GE, 1991a) and the FSAR (FSAR, 2003) and is summarized in Table 2-2. A descrip-tion of each load type is given in following subsections.

Table 2-2. Load Combination Components Load Type Load Combinations Load Source Refueling Condition Accident Post-Accident Dead Load - Gravity of Shell x x x General Dead Loads - Shell Attachments x x x GE Report Penetration Loads x x x GE Report Compressible Material x x x GE Report Live Loads x GE Report Internal Pressure x FSAR External Pressure x GE Report Hydrostatic Internal Pressure x GE Report Seismic Loads x x x (flooded) FSAR Refueling Loads x GE Report o

Thermal Load at 292 F x FSAR 34

2.4.1 General Loads: Gravity, Dead, Penetration, and Compressible Material Loads This section describes the general loads that are applied in each of the load combinations consid-ered in this study. The first of these loads employs a distributed body force to apply gravitation forces to the model. In ABAQUS, the user must define the material density, the model geome-try, and the value for the acceleration of gravity to enable the simulation of gravity. Since the current model is defined in units of inches, the gravity constant is defined as 386.4 in/s2. In addi-tion to the gravity load, a 0.0694 psi (10 psf) vertical load is applied to the exterior of the entire drywell shell. This represents the weight of the compressible material that lies in the approxi-mately 3 gap between the drywell shell and the surrounding concrete shield wall.

The dead load for components attached to the drywell shell, but not explicitly modeled, are in-cluded through the application of a series of surface traction loads. The current study uses the loads defined in Table 2.5a of the previous GE analysis (GE, 1991a). As mentioned earlier, these loads were compiled by an even earlier study by Chicago Bridge & Iron. In the GE analy-sis, these dead loads were applied by smearing the load from a specific item attached to the drywell shell along the circumference of the shell at the elevation the item is located. In other words, the total load from an item or series of items was summed together and distributed along the entire 360o of the drywell. Since the GE model was only a 36o slice of the drywell, 10% of the total load was then distributed along the slice as nodal loads applied at the appropriate eleva-tion. Here, the current model contains the entire 360o extent of the drywell shell. Therefore, the location of these applied loads can be as specific as the information available. Here, the region of application was defined on the drywell shell by imprinting the shape of the attachment.

This imprinting creates surfaces within ABAQUS that can be used to define where a specific load is applied. The load is applied by smearing it along the defined surface as a surface trac-tion. This smearing is similar to the method used in the GE analysis, but the load is smeared over the actual location on the shell where a piece of equipment or other items are attached in the real structure. This method provides a more realistic loading condition in the model.

In applying the surface tractions for the dead loads given in the GE analysis report (Table 2.5a of GE, 1991a), the drywell surface was imprinted with the locations of each item listed. These lo-cations were determined from a set of structural drawings of the drywell (CB&I, 1980). Figure 2-20 through Figure 2-24 illustrate the regions of application for each of the loads defined in the GE analysis report (GE, 1991a). Figure 2-20 shows the region of application for the upper and lower spray headers. The center of the application region is located at elevation 64-6 and 37-3 for the upper and lower headers, respectively2. Since the drawing or schematic showing the exact regions of attachment to the drywell shell was not provided, it was assumed that the region of attachment spans 3 in elevation both above and below the center points given above. There-fore, the total width of the regions of load application is 6 in elevation. The actual width of the region depends on the curvature of the drywell shell at each location. The load is also assumed to extend around the entire circumference.

35 2

August 3, 2006, conference call between Sandia National Laboratories, NRC, and Exelon.

Figure 2-20. Upper and Lower Spray Header Locations Figure 2-21 illustrates the regions of load application for the upper, middle, and lower weld pads (CB&I, 1980). Each of the weld pads covers a 8 diameter region imprinted onto the drywell shell. In the actual structure, the weld pads are attached to interior surface of the drywell. Based on the structural drawings of the weld pad layout, the center of the upper, middle, and lower weld pads are located at elevations 66-3.2, 61-2, and 54-9, respectively. The number of weld pads and spacing along the drywell circumference also varies: 15 pads at 24o, 20 pads at 18o, and 24 pads at 15o, respectively.

Figure 2-21. Weld Pad Locations Figure 2-22 shows the regions of load application for the top and bottom flanges, as well as the stabilizers (CB&I, 1980). Each of these items is located in the cylinder region of the drywell.

The top flange spans from and elevation of 96-7.878 down to 94-9. The bottom flange ex-tends from 94-9 down to 92-8.5. Both of these loads are applied along the entire circumfer-36

ence of the cylinder. The stabilizer load is applied at 8 circular regions spaced at 45o and cen-tered at elevation 82-9. Each of the stabilizer regions is 5-3 in diameter.

Figure 2-22. Flange and Stabilizer Locations Figure 2-23 illustrates the load application regions for the upper and lower beam seats (CB&I, 1980). These are the attachment points of beam within the drywell sphere. The imprinted region for the upper beam seats is approximately 12 wide and 51 high and centered at an elevation of 46-4.5. The spacing of the 20 seats around the circumference varies from seat to seat, and range from 12o to 25o30. These dimensions and spacings were derived using the structural drawings. The imprinted region for the lower beam seats is approximately 12 wide and 13.5 high and centered at an elevation of 20-11.125. The spacing of the 20 seats around the circum-ference varies from seat to seat, and range from 11o45 to 29o40. Since the surface imprints for 6 of the lower beam seats overlapped other surface partitions for the thickened regions around the personnel lock and several ventlines, the height of the region of application was reduced in slightly to 10. This modification was introduced to avoid oddly shaped surfaces which can be problematic during the meshing of the geometry. In addition, the load for the beam seats was distributed evenly among the 20 seats for both the upper and lower seats. Due to the varying spacing of the beam seats, the load could have been distributed using tributary areas. Since the exact makeup and details of the total load are unknown here, a simple even distribution was ap-plied.

37

Figure 2-23. Upper and Lower Beam Seat Locations Figure 2-24 shows the load application region for the personnel lock and equipment doors. This area is essentially the thickened region of the drywell shell surrounding the penetration. The penetration is centered at an elevation of 27-3.

Figure 2-24. Personnel Lock and Equipment Door Loads Application Region 38

The final load listed in the GE dead load Table 2-5a (GE, 1991a) is for the vents. It is assumed that this additional load accounts for the portion of the ventline that was not modeled explicitly in the GE model. Since the entire ventline is modeled in the current model, no additional load was applied to the structure.

Table 2-3 summarizes the dead loads described above. The total load given in Table 2-5a of the GE report (GE, 1991a), the total surface area from the current ABAQUS model, and the result-ing applied traction are all provided. The ABAQUS total area is the summed surface area for all of the regions of application for a given dead load case. The traction is simply the total load di-vided by the area. These tractions are applied to the appropriate regions on the drywell shell in the vertical direction.

Table 2-3. Dead Load Tractions GE Total Load* ABAQUS Total Area** Traction***

Dead Load Case kips in2 ksi Upper Header 36 15847.2 0.00227 Lower Header 41 15848.1 0.00259 Upper Weld Pads 52**** 754 0.06897 Middle Weld Pads 59.2**** 1005.3 0.05889 Lower Weld Pads 56.2**** 1206.4 0.04675 Top Flange 20.1 28543.4 0.00070 Bottom Flange 20.7 30571.1 0.00068 Stabilizers 21.65 12508.7 0.00173 Upper Beam Seats 1102 12688.3 0.08685 Lower Beam Seats 556 --- ---

- Standard Size 389.2 2563.3 0.15184

- Reduced Size 166.8 811.7 0.20549 Equipment Doors, Lock 169.1 11938.3 0.01416

• GE Total Load - This is the total load reported in the Table 2-5a of the GE analysis report (GE, 1991a).

• • ABAQUS Total Area - This is the total summed surface area from the current ABAQUS model for each of the dead load items listed in the GE report.

• • • Traction - This is the GE Total Load divided by the ABAQUS Total Area. These tractions are applied to the appropriate regions for each dead load case. The tractions are applied in the down-ward or vertical direction.

• • • • The GE Total Loads for the three weld pad loads given in Table 2-3 are the sum of two sepa-rate loads for each set of weld pads in GE Table 2-5a.

In addition to the above dead loads, the Oyster Creek has numerous penetrations that were not modeled explicitly in the current model. These penetration loads are listed in Table 2-5b of the GE report (GE, 1991a). Unfortunately, the penetration identification numbers provided in this table do not correspond to the penetration identification numbers given in the structural draw-ings penetration schedule (CB&I, 1980). Since the correlation between these two numbering systems could not be readily provided to the analyst, the loads from each penetration in the GE Table 2-5b (GE, 1991a) were summed to give a total load at each elevation and distributed along 39

the entire drywell circumference. GE Table 2-5b documents penetration loads at 17 different elevations: 16, 20, 26, 30, 31, 32, 33, 34, 35, 36, 40, 54, 60, 70, 73, 87, and 90.

These elevations were assumed to be the centerline of the application region and extend 6 in elevation in each direction. For example, the region of application for the penetration load at 33 is from 32-6 to 33-6. The regions of application for each of the penetration elevations are shown in Figure 2-25 and Figure 2-26. Typically, the total penetration load for a given elevation is distributed along the entire circumference of the drywell. Gaps in the application regions do exist near the personnel lock and equipment hatch. These regions are excluded from the applica-tion region since the hatch is an explicitly modeled penetration and other penetrations do not pass through that region. Figure 2-27 shows the application region for the penetration load at the 16 elevation. The load is distributed in the drywell shell between the ventlines including within a portion of the thickened region around the ventlines. The 16 elevation penetration load is dis-tributed along this identical region between each of the ventlines.

Figure 2-25. Penetration Load Application Regions in the Drywell Sphere 40

Figure 2-26. Penetration Load Application Regions in the Drywell Cylinder Figure 2-27. Elevation 16 Penetration Load Application Region Between the Ventlines 41

Table 2-4 provides a summary of the penetration load tractions applied to the current ABAQUS model. The total load given in Table 2-5b of the GE report (GE, 1991a) is provides along with the total surface area from the current ABAQUS model and the resulting applied traction. The ABAQUS total area is the total surface area for the region of application for a given penetration elevation. The traction is simply the total load divided by the area. These tractions are applied to the appropriate regions on the drywell shell in the vertical direction.

Table 2-4. Penetration Load Tractions Penetration Load GE Total Load* ABAQUS Total Area** Traction***

Elevation kips in2 ksi 16 168.1 24169.7 0.006955 20 11.2 23809.4 0.000470 26 11.1 29530.9 0.000376 30 50.5 29688.2 0.001701 31 16.5 29836.7 0.000553 32 0.75 30044.2 0.000025 33 15.45 30342.4 0.000509 34 28.05 30805.2 0.000911 35 1.5 31616.3 0.000047 36 1.55 31696.3 0.000049 40 43.35 31702 0.001367 54 **** 7.85 31694.5 0.000248 60 0.7 31694.5 0.000022 70 5.75 15651.8 0.000367 73 8.85 14953 0.000592 87 1.0 14953 0.000067 90 15.0 14953 0.001003

• GE Total Load - This is the total load reported in the Table 2-5b of the GE analysis report (GE, 1991a).

• • ABAQUS Total Area - This is the total summed surface area from the current ABAQUS model for each of the penetration load elevation listed in the GE report.

• • • Traction - This is the GE Total Load divided by the ABAQUS Total Area. These tractions are applied to the appropriate regions for each penetration load elevations, and act in the downward or vertical direction.

• • • • 54 Elevation Loads - The loads for this elevation are centered at 53-10 to avoid creating oddly shaped surfaces at the intersection with the lower weld pads.

2.4.2 Seismic Load A full dynamic simulation of the governing seismic loading would be ideal in determining the resulting stresses. GE applied this method by performing a dynamic using an appropriate time history. In addition, the Oyster Creek FSAR (FSAR, 2003) states that a dynamic seismic analy-sis was also performed by John A. Blume & Associates. Neither this report nor the seismic ground motions were available for the current study. Although, the FSAR states that this dy-namic analysis by John A. Blume & Associates confirmed that the original static coefficients used by Chicago Bridge & Iron in the design of the structure were acceptable. These static coef-ficients are 22% laterally and 10% vertically (acting simultaneously) of the permanent gravity load. The use of the static coefficients to simulate the seismic loading is justified due to the con-firmatory nature of this study.

42

Since the degraded drywell containment (degradation described in Section 2.6) may potentially exhibit a different dynamic behavior than the original, as-designed containment, the suitability of using the static coefficients to approximate the seismic loading is uncertain. In order to address this issue, a short study was conducted which compares the natural frequencies and associated mode shapes for the drywell in its original and degraded conditions. The details of this study are included in the Appendix A (Section 9) of this document. Since the frequencies and mode shapes proved relatively insensitive to the levels of degradation experienced in the Oyster Creek drywell, the use of the static seismic coefficients to simulate the seismic loading for the degraded structure is assumed to be acceptable.

The static coefficients are applied to the current ABAQUS model using body forces. The gravity loading in ABAQUS was utilized for this purpose. In addition to the standard 1g gravity load, an additional 0.1g was applied downward and 0.22g was applied in one lateral direction, as per-formed in the original design by CB&I. Several orientations of the seismic lateral load were ex-amined to determine the case that produced the highest stresses in the sandbed region. The direction for the 0.22g lateral load that extends from the 180o azimuth to the 90o azimuth was determined to produce the highest stresses, in general, throughout the sandbed region.

The 0.22g lateral seismic load was applied in the Accident and Refueling load cases. For the Post-Accident load case, the drywell is flooded with water up to an elevation of 74-6. The ad-ditional seismic load from the mass of the water is introduced into the analysis by increasing the value of the acceleration of gravity for the lateral seismic load and applying it to the drywell shell model that does not include the water explicitly. To determine the appropriate increase in the acceleration of gravity, the total mass of the drywell shell (degraded and undegraded) was computed within ABAQUS. The total weight of the water flooding the drywell (20% removed for the reactor vessel, GE, 1991a) was computed and added to the weight of the drywell shell.

The weight of the combined drywell shell and water for the degraded containment was deter-mined to be 10.6 times the weight of the drywell shell allow, and 10.0 times for the undegraded shell. Therefore, the lateral seismic load for the degraded analysis uses 2.3g, and the undegraded shell uses 2.2g. These loads are applied to the entire drywell shell. This method is extremely approximate, but judged appropriate based on the limited seismic information available. It is assumed that the vertical seismic loads are unaffected by the presence of the water during the Post-Accident load condition.

2.4.3 Refueling Condition Specific Loads: Live, External Pressure, and Refueling Loads For the refueling load condition, the head of the drywell is removed as illustrated in Figure 2-4.

The additional weight on the cylindrical portion of the drywell is given as 561 lbs/in along the circumference (Ref. 2.4.3 of Ref. 1-4 of GE, 1991a). This load is applied in the current model as a shell edge traction as shown in Figure 2-28.

43

Figure 2-28. Refueling Load on Drywell Cylinder In addition to the refueling cylinder load, the refueling load combination includes a 2 psi external load. This load was applied to the entire exposed exterior surface of the drywell shell.

The final refueling load combination specific item is the live loading. The live loads for the dry-well are provides in Table 2-5c of the GE report (GE, 1991a). Table 2-5 summarizes the live loads and the applied tractions. The region of live load application for each item is identical to the region of application for the dead load.

Table 2-5. Live Load Tractions GE Total Load* ABAQUS Total Area** Traction***

Live Load kips in2 ksi Upper Header 4.2 15847.2 0.000265 Lower Header 7.15 15848.1 0.000451 Upper Weld Pads 20 754 0.026525 Middle Weld Pads 20 1005.3 0.019895 Lower Weld Pads 24 1206.4 0.019894 Equipment Doors, Lock 115 11938.3 0.009633

• GE Total Load - This is the total load reported in the Table 2-5c of the GE analysis report (GE, 1991a).

• • ABAQUS Total Area - This is the total summed surface area from the current ABAQUS model for each of the live load items listed in the GE report.

• • • Traction - This is the GE Total Load divided by the ABAQUS Total Area. These tractions are applied to the appropriate regions for each live load case. The tractions are applied in the down-ward or vertical direction.

44

2.4.4 Accident Condition Specific Loads: Internal Pressure and Thermal Loads The accident condition includes an internal pressure within the drywell of 44 psi at a temperature of 292oF (FSAR, 2003) due to the design basis accident (LOCA, loss of coolant accident). This pressure is applied to the interior surface of the drywell shell above elevation 8-11.25, or the bottom of the sandbed. However, the concrete floor within the drywell does extend up to an ele-vation of 10-3 with curbs extending between the ventlines up to an elevation of 12-3. In the previous discussion of the boundary conditions, the interior concrete above the bottom of the sandbed is ignored and only the shell below elevation 8-11.25 is fixed since it is surrounded by concrete on both sides. The actual condition of the bond between the drywell shell and the con-crete floor inside the drywell is not known. If a small gap exists, it would be likely that gas could enter and pressurize the shell below the level of the concrete floor.

In the previous GE accident condition analysis, the thermal stresses in the sandbed region were determined using a heat transfer analysis. Specifically, this region is below the concrete floor at an elevation of 10-3 and extends down to the bottom of the sandbed at 8-11.25. Since there is not sufficient information and/or explanation provided in the GE report to reproduce the heat transfer analysis or apply the temperatures given in this region, the entire shell in the current analysis is set to 292oF down to an elevation of 8-11.25. As stated above, the condition of the bond between the drywell shell interior and the concrete below 10-3 is not known to the ana-lyst. A small gap would allow the temperature of the shell below 10-3 to be heated uniformly.

The internal pressurization and heating of the shell down to the fixed boundary condition at ele-vation 8-11.25 produces a severe discontinuity in the drywell shell. At the point in the shell just above 8-11.25, the increase in the temperature causes the steel shell to expand and the in-ternal pressure forces the steel shell outward. The high bending stresses in this region were originally designed to be tempered by the sand outside of the shell above elevation 8-11.25.

Based in part on the previous reviewed and approved study be GE (GE, 1991a and b), the shell was determined to resist the potential accident condition without the sand present. The sand in the sandbed region was subsequently. Since the focus of this study was not to assess the deci-sion to remove the sand, potentially conservative boundary conditions and applied loads (pres-sure and thermal) were used here.

2.4.5 Post-Accident Condition Specific Load: Hydrostatic Load The only post-accident condition specific loading is the hydrostatic load from the flooding of the drywell interior. In this condition, the water fills the drywell from the top of the concrete floor at 10-3 up to the 74-6 elevation. Assuming a density of water at 62.3 lbs/ft3, the hydrostatic pressure in the drywell interior at 10-3 is 4003 psf (0.02780 ksi). This load reduces linearly to zero at the 74-6 elevation. Since the elevation that the water reaches in the ventlines extends below the 10-3 elevation, the hydrostatic load in the ventlines increases appropriately with the distance from the top of the water at 74-6.

45

2.5 Material Properties The drywell shell was constructed out of A-212-61T Grade B pressure vessel steel. The modulus of elasticity, E, has been reported as 29,500 ksi at temperatures from 70oF to 100oF, 28,800 ksi at 200oF, and 28,300 ksi at 300oF (IPE, 1992). The yield stress for the material is 50.7 ksi from 70oF to 100oF, 46.1 ksi at 200oF, and 45.1 ksi at 300oF (IPE, 1992). The coeffi-cient of thermal expansion is assumed to be 6.5E-6oF-1. The density of the steel is 0.283 lb/in3 (GE, 1991b), which is equivalent to its value in the required ABAQUS density units, 7.324E-7 kips-sec2/in4.

2.6 Degraded Model Section 2.2 provides the steel plate thicknesses throughout the drywell in Oyster Creeks as-build state. For over 20 years, the drywell has experienced extensive thinning due to corrosion.

Since UT measurements have only been taken at a limit number of locations throughout the shell, the current analysis adopts average measured thickness values for different regions of the drywell reported by AmerGen. Average values have been adopted to establish a realistic model that reflects the current conditions.

Since uniform thinning was used in this analysis, any additional stress concentration that might occur at the location of a crack-like pit or a highly non-uniform region was not captured in this analysis. While some pit data has been documented, it is not detailed enough to make any as-sessment of these types of local defects.

The cylinder, upper sphere, and middle sphere degraded thicknesses are based on the minimum average thickness values from recent documentation on the condition of the drywell shell up to 2004 (AmerGen, April 7, 2006). The minimum average values reported at any location within each of the cylinder, upper sphere, and middle sphere are 0.604, 0.676, and 0.678, respec-tively. Due to ongoing corrosion, the thicknesses of the cylinder and middle sphere were further reduced. A location in the cylinder shows a corrosion rate of 0.0003/yr. Based on 25 years of additional corrosion (2004 to 2029), the cylinder was modeled at a thickness of 0.585 (0.604 -

0.00075/yr x 25yr = 0.585). One location in the middle sphere shows an ongoing corrosion rate of 0.00075/yr. This leads to a thickness of the middle sphere of 0.670 (0.678 -

0.0003/yr x 25yr = 0.670). The knuckle is reduced slightly in thickness from 2.5625 to 2.54 (AmerGen, April 4, 2006). These thicknesses are taken as uniform throughout the entire region and are summarized in Table 2-6.

The middle sphere and thickened regions around penetrations are decreased in thickness by the same magnitude as the surrounding regions. For example, the thickened middle sphere is re-duced in thickness by 0.1 since the middle sphere is reduced by 0.1. The thin transition re-gions that fall between the main regions are set typically to a thickness equal to the average of the surrounding plates, as described previously for the geometry without degradation. Thick-nesses in the cylinder stiffeners, hatch, and ventlines do not include any degradation and are equal to their as-built values.

46

Table 2-6. Main Drywell Shell Model Thicknesses, Original and Degraded Original Degraded Original Degraded Section Thickness, Thickness, Section Thickness, Thickness, in in in in Head 1.1875 N/C Reinforcing Around Ventlines 2.875 2.618 Upper Cylinder 1.1875 N/C Lower Sphere Below Sandbed 1.154 N/C Main Cylinder 0.640 0.585 Bottom Sphere 0.676 N/C Knuckle 2.5625 2.54 Middle Sphere Thickened 1.0625 0.9625 Upper Sphere 0.722 0.676 Reinforcing Around Hatch 2.625 2.525 Middle Sphere 0.770 0.670 Lower Sphere 1.154 See below N/C - No Change For modeling the degradation in the sandbed region, the lower sphere was divided into 10 re-gions to be assigned uniform thicknesses. These regions extend from the centerline of one ven-tline to the centerline of the adjacent ventline. Each of these newly defined regions contains one-half of the two different, but adjacent, bays. This was done in order to avoid placing the thickness discontinuity at the centerline between the ventlines, since this is typically the location of the highest stresses. If the thickness jump was placed at this location, the stresses of interest would be difficult to interpret.

The thickness values used in these 10 regions were defined based on a set of UT measurements from a study performed in 1993 (GPU Nuclear, 1993). In these calculations, a selected set of thickness measurements were taken from the outside of the containment before the application of the epoxy coating. Measurements are provided for each bay of the sandbed as shown in Figure 2-29 for Bay 1. The image in Figure 2-29 was extracted from the 1993 GPU Nuclear Calcula-tion Sheet. Since the set of thickness values are reported to be the thinnest areas (by visual in-spection) in each bay, the averages used here are still biased conservative. As stated above, the 10 regions used in the analysis combine one-half of two adjacent bays. For example, the thick-nesses for points in the right half of Bay 3 are combined with the thicknesses for points in the left half of Bay 1 (Points 8, 9, 15, 18, and 19 in Figure 2-29). This is continued around the cir-cumference of the sandbed as shown in Figure 2-30. In addition, the effects of locally thinner regions were explored by introducing two 30 by 18 regions under the ventlines of Bay 1 and Bay 13 as shown in Figure 2-31 for Bay 1 (labeled as the Bathtub in Figure 2-29). These two Bays showed a concentration of thin points within a local region. The GPU Nuclear Calculation sheet provided the approximate dimensions of the local thin region in Bay 1, but not for Bay 13.

Due to a lack of information for Bay 13, the dimensions and placement of the local region in Bay 13 were assumed to be identical to the region shown for Bay 1.

47

Local Bay 1 Region Bay Combination 1-3 Bay Combination 19-1 Elevation 8-11.25 Figure 2-29. Bay 1 UT Measurement Locations Taken from Outside of the Containment (Image Extracted from GPU Nuclear Calculation Sheet, 1993)

Bay Combinations Elevation 3-5 8-11.25 15-17 1-3 17-19 19-1 Local Bay 1 Region Figure 2-30. Lower Sphere Bay Combination Regions (Ventlines Removed for Clarity) 48

Local Bay 1 Region Bay Combination 1-3 Bay Combination 19-1 Elevation 10-11 18 Elevation 9-11.4 18 12 Elevation 8-11.25 Region Embedded in Concrete Figure 2-31. Detailed View of Local Bay 1 Region (Ventline Removed for Clarity)

The average of the datapoints that fall within each bay combination (e.g. Bay 1-3) was computed and assigned to the thickness in that defined region of the model.

Table 2-7 summarizes the thicknesses throughout the lower sphere based on the average UT measurements. Figure 2-32 illustrates the layout of the thicknesses prescribed to the bay combi-nations in the lower sphere. To explore the effects of significant local thinning, the lowest meas-ured value at any point within the two local regions (Bay 1 and 13) was assigned as the uniform thickness throughout the entire 30 by 18 section. The measurement values that fall within each of these local regions were not used in the averaging to define the uniform thickness as-signed to the surrounding bay combinations. A detailed description of the computation of these thicknesses is provided in Appendix B (Section 10).

Table 2-7. Degraded Lower Sphere Shell Model Thicknesses Bay Combination Thickness, Degraded, inches Bay 1-3 0.894 Bay 3-5 0.922 Bay 5-7 0.998 Bay 7-9 0.998 Bay 9-11 0.835 Bay 11-13 0.859 Bay 13-15 0.842 Bay 15-17 0.857 Bay 17-19 0.904 Bay 19-1 0.858 Local Bay 1 Region 0.705 Local Bay 13 Region 0.618 49

Sphere Equator - Elevation 37-3 Bay Combinations Ventline Locations Elevation 23-6 Lower Sphere 11-13 9-11 7-9 5-7 3-5 1-3 19-1 17-19 15-17 13-15 11-13 0.859 0.835 0.998 0.998 0.922 0.894 0.858 0.904 0.857 0.842 0.859 Elevation 8-11.25 Local Bay 13 Region Sandbed Region Local Bay 1 Region 0.618 0.705 Figure 2-32. Degraded Thicknesses in the Lower Sphere (inches) 2.7 Mesh Size The solid geometry describes in Section 2.2 was meshed within the ABAQUS/CAE utility. A nominal mesh seed size of 4 was applied to the geometry. Typically, this leads to elements sizes that have a 4 by 4 square dimension. Due to the unique shape of the model and the sur-face partitions introduced for application of the boundary conditions, loadings, and to divide the shell sections of different thickness, some elements contain edges slightly larger that 4 with some edges much smaller that 4. In the two local regions where the effects of more extensive degradation is explored, smaller elements (1 by 1) are employed to better capture the poten-tially high stresses. As stated previously, the mesh used throughout the model adopts a quad-dominated scheme. This enables the meshing utility to insert 3-noded, or tri, elements when needed to avoid creating a poorly shaped quad element.

A 4 element size was employed based on a limited mesh convergence study. Models with nominal element sizes of 3, 4, and 5 were constructed using the accident load conditions. For each of these meshes, the hoop stresses at the same location in the sandbed were compared at one point. In addition, the meridional stresses at the same location in the sandbed were com-pared at one point. The meridional stresses at the point examined were not sensitive to the mesh size. For the hoop stresses at the point examined, the percentage of area reduction for a typical element was compared to the percentage of hoop stress increase as the element size was reduced.

The area reduction percentage when going from a 5 nominal mesh to 4 nominal mesh was in excess of one order of magnitude larger than the percentage increase in the hoop stress. In other 50

words, a significant reduction in the element size only lead to a slight increase in the stress. The reduction of the element size from 4 to 3 produced a percentage ratio that was nearly two or-ders of magnitude. The percentage ratio of one order of magnitude was judged to be acceptable, and therefore, the 4 nominal element size mesh was adopted for all analyses in this study.

Figure 2-33 illustrates the finite element mesh for the refueling load case. This mesh contains 245,192 shell elements. Figure 2-34 shows a detailed view of the drywell cylinder mesh with the head removed. The same identical mesh is used for perform the stress analyses of the contain-ment in its original and degraded states, as well as for the eigenvalue buckling analyses.

Figure 2-33. Finite Element Mesh for the Refueling Load Case 51

Figure 2-34. Finite Element Mesh in the Drywell Cylinder for the Refueling Load Case Figure 2-35 illustrates the finite element mesh used for the accident and post-accident load cases.

The mesh contains more elements than the refueling mesh with 263,446. The additional ele-ments are required due to the including of the head as shown in Figure 2-36. As with the refuel-ing mesh, this mesh is used for all accident and post-accident analyses.

Figure 2-35. Finite Element Mesh for the Accident and Post-Accident Load Cases 52

Figure 2-36. Finite Element Mesh in the Drywell Cylinder and Head for the Accident and Post-Accident Load Cases Figure 2-37 illustrates the mesh in the upper and middle sphere regions of the drywell. The por-tion of the personnel lock/equipment hatch modeled is also visible as well as the upper portion of the lower sphere and ventlines. The mesh for the refueling case and the mesh for the accident and post-accident analyses are similar in these regions.

Figure 2-37. Finite Element Mesh in the Upper and Middle Sphere 53

Figure 2-38 shows the finite element mesh in the lower sphere, bottom sphere, and ventlines. As stated for the upper and middle sphere, the meshes in these lower drywell regions are similar for the refueling and accident/post-accident models.

Figure 2-38. Finite Element Mesh in the Lower Sphere, Bottom Sphere, and Ventlines The mesh in the local thinned regions is shown in Figure 2-39. While the meshes for the local thinned regions under the ventline for Bay 1 and 13 are not identical, they are similar with a typi-cal element size equal to 1 x 1. The elements in these local thinned regions have been reduced in size compared to the surrounding mesh to better capture any potential stress concentrations.

No detailed mesh convergence study was performed to determine the optimum element size in these regions.

Figure 2-39. Finite Element Mesh for the Local Thin Regions under the Ventlines in Bay 1 and 13 54

3. Stress Analysis In this analysis, the structural integrity of the drywell shell is examined in terms of the stress lim-its using a combination of the values used in the previous analysis by GE (GE, 1991a) and the current ASME code. The GE analysis provided a description of the allowable stresses per the original design code (1962 ASME Code,Section VIII). Using this code, the appropriate code case (1272-N-5) was used to define the allowable primary and secondary stresses for the differ-ent loading conditions. Since the original pressure vessel steel used to construct the drywell shell has been designated, it was determined to be appropriate to adopt the original stress crite-rion while also considering the current code. The allowable stress based on the re-designated steel in the current 2004 ASME B&PV Code,Section III, Division 1, Subsection NE (ASME, 2004) is slightly higher that the original value used by GE. The use of the original value is therefore a conservative assumption.

As reported in the GE report (GE, 1991a), the allowable stress, S, for the SA-212 Grade B steel used for the drywell is defined at 17.5 ksi. The primary stresses include the general membrane stress and the general membrane plus bending stress. For the refueling load case (Service Level B) and the accident load case (Service Level C), the allowable general membrane stress, Smc, were set equal to 1.1 times the allowable stress, or 1.1 x 17.5 = 19.3 ksi. The general membrane plus bending stress is set equal to 1.5 times the general membrane allowable stress, or 1.5 x 19.3

= 29 ksi. The primary plus secondary stress is set equal to 3 times the allowable stress, or 3 x 17.5 = 52.5 ksi. Secondary stresses include thermal stresses and bending stresses at gross struc-tural discontinuities (e.g. the intersection of two plates of different thickness). For the post-accident (Service Level D) load case, the general membrane, general membrane plus bending, and the primary plus secondary stress allowables are 38 ksi, 57 ksi, and 70 ksi.

In the FSAR (FSAR, 2003), it is stated the steel designated SA-212 Gr. B has been superseded in the ASME code by SA-516 Gr. 70. In the 2004 ASME code,Section II, Part D, Subpart 1, Table 1A, the allowable stress at room temperature is given as 20 ksi. Since this value is slightly larger than the 17.5 ksi used previously, the lower value of 17.5 ksi is used here. When using this value for the allowable stress, S, the 2004 ASME Section III, Division I, Subsection NE, Class MC, Article NE-3000 (ASME, 2004), as well as the Standard Review Plan (SRP, 1996) pro-duces similar values as compared to the allowables defined above by GE. The allowables in the current code for not integral and continuous structures are slightly more conservative than those for integral and continuous structures, and produce the same allowables as describes above. Therefore, these values are conservatively adopted here. In addition, the allowables for primary plus secondary stresses for Level C and Level D do not need to be evaluated per the ASME Code. For consistency with the previous GE analysis, these stresses are evaluated using the limits of 52.5 ksi and 70 ksi for Level C and D, respectively.

The stress analyses comparison with the code allowables treats the peak surface stresses for the shell elements used in this analysis as membrane plus bending stresses. If a case was encoun-tered where the surface stress exceeded the membrane plus bending stress allowable, the stress value was explored further to determine if the surface stress resided at a gross structural discon-tinuity. In these cases, the stress values were considered to be primary plus secondary values and assessed using the higher stress limits defined in the ASME code. The results of the elastic 55

ABAQUS stress analyses for the Refueling, Accident, and Post-accident load cases are summa-rized below.

3.1 Refueling Condition The analyses of the refueling load condition employed the model and loadings described in Sec-tion 2. Two stress analyses were performed for the refueling load case. These included the con-tainment with and without degradation. The thicknesses used for the upper portions of the degraded drywell are outlined in Table 2-6. In the lower sphere of the drywell, the average UT measurement data was used to assign shell thicknesses as outlined in Table 2-7. Table 3-1 and Table 3-2 summarize the peak stresses for each of the analyses. In each case and for each region of the containment, the peak membrane stresses are reported as well as the peak membrane plus bending stresses. The membrane plus bending stresses are the surface stresses provided in the analysis output for each shell element. The membrane stresses are taken at the midsection output value for each shell element. The peak stresses in both the meridional and circumferential direc-tions are provided. Values given as positive represent tensile stresses, and values given as nega-tive are compressive stresses. The percentage of the ASME limit for each stress value is provided in parenthesis. For each analysis, the stresses remain within ASME code allowables (Service Level B).

Table 3-1. Refueling Load Case Peak Stresses with No Degradation, Primary Stresses (Percentage of ASME Limit in Parenthesis)

Drywell Region Membrane Stresses, ksi Membrane + Bending Stresses, ksi Meridional Circumferential ASME Limit Meridional Circumferential ASME Limit Cylinder -1.31 (6.8) -1.33 (6.9) 19.3 -1.59 (5.5) -1.53 (5.3) 29 Knuckle -0.59 (3.1) -2.06 (10.7) 19.3 -2.33 (8.0) -2.45 (8.4) 29 Upper Sphere -2.49 (12.9) -0.88 (4.6) 19.3 -6.27 (21.6) -4.62 (15.9) 29 Middle Sphere -4.45 (23.1) -2.08 (10.8) 19.3 -7.94 (27.4) -8.65 (29.8) 29 Thickened Middle Sphere -2.71 (14.0) 3.89 (20.2) 19.3 -5.05 (17.4) -5.66 (19.5) 29 Lower Sphere -5.02 (26.0) 6.05 (31.3) 19.3 -12.14 (41.9) 9.64 (33.2) 29 Positive values are tension, negative values are compression. ASME Limits based on stress magnitude.

Table 3-2. Refueling Load Case Peak Stresses with Degradation, Primary Stresses (Percentage of ASME Limit in Parenthesis)

Drywell Region Membrane Stresses, ksi Membrane + Bending Stresses, ksi Meridional Circumferential ASME Limit Meridional Circumferential ASME Limit Cylinder -1.43 (7.4) -1.44 (7.5) 19.3 -1.72 (5.9) -1.64 (5.7) 29 Knuckle -0.60 (3.1) -2.08 (7.2) 19.3 -2.38 (12.3) -2.48 (8.6) 29 Upper Sphere -2.71 (14.0) -1.01 (3.5) 19.3 -6.94 (36.0) -5.18 (17.9) 29 Middle Sphere -5.51 (28.5) -2.58 (13.4) 19.3 -9.72 (33.5) -10.65 (36.7) 29 Thickened Middle Sphere -3.15 (16.3) 4.99 (25.9) 19.3 -5.78 (19.9) 7.06 (24.3) 29 Lower Sphere -6.37 (33.0) 8.00 (41.5) 19.3 -14.70 (50.7) 14.32 (49.4) 29 Local Region 1 -5.01 (26.0) 3.94 (20.4) 19.3 -7.25 (25.0) 4.42 (15.2) 29 Local Region 13 -5.02 (26.0) 3.91 (20.3) 19.3 -7.31 (25.2) 4.39 (15.1) 29 56

Figure 3-1, Figure 3-2, and Figure 3-3 show the meridional membrane stress distributions in the lower sphere regions for the refueling case without degradation, and with degradation. Figure 3-3 shows a detailed view of the local thin region in under the ventline in Bay 13. Note that the scales for the color stress contours are not the same for the no degradation case and for the deg-radation case. The regions in light gray have tensile meridional stresses which are typically much lower in magnitude than the compressive stresses for this loading condition. The merid-ional membrane stress distribution is similar for each case, with the highest stresses near the bot-tom of the sandbed and between the ventlines. The local thin area in Figure 3-3 does not experience significantly higher stresses since the compressive load is typically lower beneath the ventlines and the load in that region is easily redistributed around the thin region.

Figure 3-1. Meridional Membrane Stress Distribution in the Lower Sphere for the Refueling Load Case with No Degradation (ksi)

Local Bay 13 Region Figure 3-2. Meridional Membrane Stress Distribution in the Lower Sphere for the Refueling Load Case with Degradation (ksi) 57

Local Bay 13 Region Figure 3-3. Meridional Membrane Stress Distribution in Local Bay 13 Region for the Refueling Load Case with Degradation (ksi) 3.2 Accident Condition The analyses of the accident load condition employed the model and loadings described in Sec-tion 2. Two analyses were performed for the stress analysis of the accident load case. These in-cluded the containment with and without degradation. The thicknesses used for the upper portions of the degraded drywell are outlined in Table 2-6. The degraded shell thicknesses for the lower sphere are outlined in Table 2-7. Table 3-3 through Table 3-6 summarize the peak stresses for each analysis. Table 3-3 and Table 3-5 include the peak membrane stresses and the peak membrane plus bending stresses. In addition, the peak primary plus secondary stresses are provided in Table 3-4 and Table 3-6. These values are typically surface stresses that include the thermal stress component from the increase of the drywell shell from 70oF to the accident temperature of 292oF. As for the refueling case, the peak stresses in both the meridional and circumferential directions are pro-vided. Values given as positive represent tensile stresses, and values given as negative are com-pressive stresses.

For each analysis, the stresses remain within ASME code allowables (Service Level C) with a few potential exceptions which required additional discussion. The meridional membrane plus 58

bending allowable stress for the degraded analysis was exceeded in the upper sphere at the inter-section with the knuckle. This was then determined to be a gross structural discontinuity, and therefore, the stress in this region was well below the primary plus secondary stress allowable.

The only remaining stress potentially exceeding the allowable is for the meridional and circum-ferential primary plus secondary stresses at the bottom of the lower sphere. These values are ex-tremely large, exceeding the assumed allowable even for the case with no degradation. The high stresses in this region are caused by a combination of the bending due to the internal pressure and the thermal expansion due to the increase in the temperature from 70oF to the accident tem-perature of 292oF. While the introduction of degradation does increase these stresses, it appears to be a secondary effect. The model constructed in this study uses several approximations of the geometry and loading in this region. These include the assumption that beginning the increase in temperature from 70oF while the service temperature is closer to 150oF. Any potential stress re-laxation due to the higher service temperature has been neglected. In addition, the temperature in the entire sandbed region is raised to 292 oF and the internal pressure is applied to the inside of the drywell shell down to an elevation of 8-11.25. The previous GE analysis included a heat transfer analysis to determine the thermal gradient in the drywell shell in the sandbed region due to the concrete slab within the drywell extending up to an elevation of 10-3. Since the present condition of the bond between the drywell shell and the concrete between the elevations of 10-3 and 8-11.25 is not currently know, the temperature and internal pressure were conserva-tively extended down to 8-11.25. These assumptions, especially the extension of the tempera-ture down to the point of fixity (elevation 8-11.25), imposes a severe discontinuity in the shell as discussed in Section 2.4.4. The potential conservativeness of the assumptions adopted here should be considered when interpreting the analysis results. It should be noted that the sand that originally filled the sandbed was included in the original design to mitigate the bending stresses in this location. The sand was removed based in part by the previous analysis by GE (GE, 1991a). In addition, the intent of this study was not to reinvestigate the acceptability of remov-ing the sand since this was performed in the approved analyses by GE. Finally, the ASME code does not require an evaluation for primary plus secondary stresses (stresses including thermal effects) for Level C loading. The evaluation is performed here to remain consistent with the stress evaluation in the previous GE analysis.

Table 3-3. Accident Load Case Peak Stresses with No Degradation, Primary Stresses (Percentage of ASME Limit in Parenthesis)

Drywell Region Membrane Stresses, ksi Membrane + Bending Stresses, ksi Meridional Circumferential ASME Limit Meridional Circumferential ASME Limit Cylinder 6.80 (35.2) 14.02 (72.6) 19.3 12.15 (41.9) 14.90 (51.4) 29 Knuckle 3.59 (18.6) 13.79 (71.5) 19.3 10.33 (35.6) 15.83 (54.6) 29 Upper Sphere 12.73 (66.0) 13.68 (70.9) 19.3 28.86 (99.5) 16.43 (56.7) 29 Middle Sphere 12.57 (65.1) 13.98 (72.4) 19.3 16.35 (56.4) 16.01 (55.2) 29 Thickened Middle Sphere 10.61 (55.0) 11.13 (57.7) 19.3 13.68 (47.2) 12.27 (42.3) 29 Lower Sphere 9.44 (48.9) 10.95 (56.7) 19.3 14.42 (49.7) 17.39 (60.0) 29 Positive values are tension, negative values are compression. ASME Limits based on stress magnitude.

59

Table 3-4. Accident Load Case Peak Stresses with No Degradation, Primary + Secondary Stresses (Percentage of ASME Limit in Parenthesis)

Drywell Region Primary + Secondary Stresses, ksi Meridional Circumferential ASME Limit Cylinder 12.19 (23.2) 15.01 (28.6) 52.5 Knuckle 10.37 (19.8) 15.89 (30.3) 52.5 Upper Sphere 28.97 (55.2) 16.46 (31.4) 52.5 Middle Sphere 17.34 (33.0) 15.99 (30.5) 52.5 Thickened Middle Sphere 13.00 (24.8) 12.28 (23.4) 52.5 Lower Sphere 82.51 (157.2) -62.71 (119.4) 52.5 Table 3-5. Accident Load Case Peak Stresses with Degradation, Primary Stresses (Percentage of ASME Limit in Parenthesis)

Drywell Region Membrane Stresses, ksi Membrane + Bending Stresses, ksi Meridional Circumferential ASME Limit Meridional Circumferential ASME Limit Cylinder 7.46 (38.7) 15.36 (79.6) 19.3 13.59 (46.9) 16.28 (56.1) 29 Knuckle 3.63 (18.8) 13.96 (72.3) 19.3 10.47 (36.1) 16.02 (55.2) 29 Upper Sphere 13.61 (70.5) 14.70 (76.2) 19.3 30.77 (58.6) 17.61 (60.7) 52.5 / 29 Middle Sphere 14.45 (74.9) 16.59 (86.0) 19.3 20.12 (69.4) 18.29 (63.1) 29 Thickened Middle Sphere 11.83 (61.3) 12.31 (63.8) 19.3 16.07 (55.4) 13.63 (47.0) 29 Lower Sphere 13.24 (68.6) 14.73 (76.3) 19.3 27.11 (93.5) 24.62 (84.9) 29 Local Region 1 8.91 (46.2) 13.46 (69.7) 19.3 15.46 (53.3) 15.36 (53.0) 29 Local Region 13 10.13 (52.5) 14.41 (74.7) 19.3 17.29 (59.6) 16.45 (56.7) 29 Positive values are tension, negative values are compression. ASME Limits based on stress magnitude.

Table 3-6. Accident Load Case Peak Stresses with Degradation, Primary + Secondary Stresses (Percentage of ASME Limit in Parenthesis)

Drywell Region Primary + Secondary Stresses, ksi Meridional Circumferential ASME Limit Cylinder 13.60 (25.9) 16.31 (31.1) 52.5 Knuckle 10.48 (20.0) 16.04 (30.6) 52.5 Upper Sphere 30.80 (58.7) 17.61 (33.5) 52.5 Middle Sphere 21.50 (41.0) 19.52 (37.2) 52.5 Thickened Middle Sphere 14.79 (28.2) 14.21 (27.1) 52.5 Lower Sphere 88.55 (168.7) -63.13 (120.2) 52.5 Local Region 1 32.59 (62.1) 12.52 (23.8) 52.5 Local Region 13 34.59 (65.9) 13.54 (25.8) 52.5 60

Figure 3-4 through Figure 3-7 illustrate the circumferential membrane stresses in the sandbed region of the cases without and with degradation. Both cases are show with and without the ap-plication of the thermal loading. With dead loads and the internal pressure load, the sandbed re-gion is in tension circumferentially. The addition of the thermal loading causes sections of the sandbed region to go into compression due to the constraint at the point the drywell shell is em-bedded within concrete below elevation 8-11.25 and below the ventlines. The sections of the sandbed and lower sphere that remain in tension are at significantly lower values due to the con-straint provided by the ventlines. For the degraded case prior to the application of the thermal load, the local thinned region in Bay 13 does experience higher stresses than the surrounding area as shown in Figure 3-6. The thermal loads cause a significant reduction in the tensile stresses in this region. As discussed previously, the meridional membrane plus bending stresses also experience significantly higher stresses. Figure 3-8 illustrates the meridional membrane plus bending stresses, or the tensile stresses on the inside surface of the drywell shell, for the case without degradation after application of the internal pressure and thermal loads. The ther-mal expansion below the ventlines causes the sandbed region of the drywell shell to extend out-ward. This produces a significant stress concentration at the point the drywell shell becomes fixed within the concrete below elevation 8-11.25. This bending stress concentration is high-lighted in Figure 3-8 by the ring of red, orange, and yellow elements. The bending at this loca-tion is so severe the outside surface of the drywell shell is in significant compression, exceeding 60 ksi in some regions. It should be noted that the analyses performed here are elastic, and therefore, the stress reported do not include the effects of material yielding and plastic deforma-tion. As mentioned previously, the addition of degradation does increase the bending stresses in this region, but the degradation appears to be secondary to the basic geometry and the modeling assumptions in this location.

Figure 3-4. Circumferential Membrane Stress Distribution in Sandbed for the Accident Load Case with No Degradation (Internal Pressure without Thermal Load) (ksi) 61

Figure 3-5. Circumferential Membrane Stress Distribution in Sandbed for the Accident Load Case with No Degradation (Internal Pressure with Thermal Load) (ksi)

Figure 3-6. Circumferential Membrane Stress Distribution in Sandbed and Local Thin Region Un-der the Ventline in Bay 13 for the Accident Load Case with Degradation (Internal Pressure without Thermal Load) (ksi) 62

Figure 3-7. Circumferential Membrane Stress Distribution in Sandbed for the Accident Load Case with Degradation (Internal Pressure with Thermal Load) (ksi)

Figure 3-8. Meridional Membrane Plus Bending Stress Distribution (Tension on the Inside Surface of the Drywell Shell) in Sandbed for the Accident Load Case with No Degradation (Internal Pres-sure with Thermal Load) (ksi) 63

3.3 Post-Accident Condition The analyses of the post-accident load condition employed the model and loadings that are de-scribed in the previous section. Two analyses were performed for the stress analysis of the post-accident load case. These included the containment with and without degradation. The thick-nesses for the upper portions of the degraded drywell are outlined in Table 2-6. The thicknesses in the lower sphere of the drywell are outlined in Table 2-7.

For the Post-Accident condition, Table 3-7 and Table 3-8 summarize the peak stresses for each analysis. In each case and for each region of the containment, the peak membrane stresses are reported as well as the peak membrane plus bending stresses. The membrane plus bending stresses are the surface stresses provided in the analysis output for each shell element. The membrane stresses are taken in the midsection output value for each shell element. The peak stresses in both the meridional and circumferential directions are provided. Values given as positive represent tensile stresses, and values given as negative are compressive stresses. For each analysis, the stresses remain within ASME code allowables (Service Level D).

Table 3-7. Post-Accident Load Case Peak Stresses with No Degradation, Primary Stresses (Per-centage of ASME Limit in Parenthesis)

Drywell Region Membrane Stresses, ksi Membrane + Bending Stresses, ksi Meridional Circumferential ASME Limit Meridional Circumferential ASME Limit Cylinder -1.68 (4.4) -4.76 (12.5) 38 -4.25 (7.5) -6.96 (12.2) 57 Knuckle -0.43 (1.1) -1.29 (3.4) 38 -1.99 (3.5) -1.58 (2.8) 57 Upper Sphere 1.41 (3.7) 5.37 (14.1) 38 -4.65 (8.2) 6.39 (11.2) 57 Middle Sphere 2.75 (7.2) 12.27 (32.3) 38 -5.44 (9.5) 12.61 (22.1) 57 Thickened Middle Sphere -5.03 (13.2) 13.43 (35.3) 38 -10.22 (17.9) 15.90 (27.9) 57 Lower Sphere -10.10 (26.6) 18.34 (48.3) 38 -25.00 (43.9) 21.36 (37.5) 57 Table 3-8. Post-Accident Load Case Peak Stresses with Best Estimate Degradation, Primary Stresses (Percentage of ASME Limit in Parenthesis)

Drywell Region Membrane Stresses, ksi Membrane + Bending Stresses, ksi Meridional Circumferential ASME Limit Meridional Circumferential ASME Limit Cylinder -1.80 (4.7) -4.87 (12.8) 38 -4.49 (7.9) -6.94 (12.2) 57 Knuckle -0.40 (1.1) -1.19 (3.1) 38 -1.91 (3.4) 1.58 (2.8) 57 Upper Sphere 1.44 (3.8) 5.92 (15.6) 38 -5.12 (9.0) 6.93 (12.2) 57 Middle Sphere 3.19 (8.4) 14.13 (37.2) 38 -6.49 (11.4) 14.48 (25.4) 57 Thickened Middle Sphere -5.58 (14.7) 17.25 (45.4) 38 -13.05 (22.9) 19.35 (33.9) 57 Lower Sphere -13.21 (34.8) 24.04 (63.3) 38 -28.60 (50.2) 29.51 (51.8) 57 Local Region 1 -7.24 (19.1) 17.31 (45.6) 38 -15.93 (27.9) 20.20 (35.4) 57 Local Region 13 -8.87 (23.3) 20.31 (53.4) 38 -18.75 (32.9) 23.67 (41.5) 57 64

Figure 3-9 and Figure 3-10 illustrate the circumferential membrane stresses in the sandbed re-gion for the post-accident load case, without degradation and with degradation, respectively.

Note that the color stress contours used in these two figures are not set at the same scale. The stresses in the degraded analysis are much larger than the case with no degradation. The local thin region under the ventline in Bay 13 experiences higher stresses, but do not approach the al-lowables.

Figure 3-9. Circumferential Membrane Stress Distribution in Sandbed for the Post-Accident Load Case with No Degradation (ksi)

Figure 3-10. Circumferential Membrane Stress Distribution in Sandbed and Local Thin Region Under the Ventline in Bay 13 for the Post-Accident Load Case with Degradation (ksi) 65

3.4 Conclusion The ASME allowable stresses are met for all three load cases examined here given the modeling and loading procedures outlined in Section 2. The only potential exception is for the primary plus secondary stresses located at the base of the sandbed region of the accident condition due to the thermal expansion of the shell. The primary cause of these high stresses is the number of modeling and loading assumptions in this region, with the introduction of degradation producing only a secondary effect. In addition, the primary plus secondary stresses (includes thermal stresses) were compared to the allowables use in the previous GE analysis (GE, 1991a). The current code does not require an evaluation of the primary plus secondary stresses for Service Level C, but were performed here for consistency with the previous study and since some evaluation of the shell was judged to be appropriate. Beyond the stresses at the base of the sand-bed region for the accident condition, the introduction of the degradation does cause a noticeable increase in the stress levels throughout the drywell shell for each load condition. In general, the accident condition causes the largest stress increases throughout the containment when degrada-tion is introduced.

66

4. Stability Analysis In this analysis, the structural integrity of the drywell shell is examined in terms of stability using the ASME Code Case N-284, Metal Containment Shell Buckling Design Methods,Section III, Division 1, Class MC. This stability analysis used the stresses computed through the stress analysis outlined in the previous section. The refueling and post-accident load cases were as-sumed to be the governing load combinations for potential buckling in the drywell. The effec-tive factors of safety against buckling were computed and compared to the required ASME code allowables.

Here, the theoretical elastic buckling stress, ie, is computed using a combination of the stress analyses described in the previous section and a separate eigenvalue extraction analysis in ABAQUS. The eigenvalue buckling analysis provides the load factors, , that cause buckling given the applied loads. For each eigenvalue, or load factor, the analysis provides the resulting buckling mode or displaced shape. Each load factor defines the multiplier on the applied loads that would cause the given buckling mode. For example, a load factor of 4 indicates that the ap-plied loads would need to be increased by a factor of 4 to cause that buckling mode to occur.

The load factor can also be applied to the compressive stress value, c, located in the buckling region to compute the buckling stress. Therefore, the stress determined from the stress analysis of a specific load case and level of degradation is multiplied by the load factor computed in the eigenvalue buckling analysis to produce the theoretical elastic buckling stress, ie = c. The same models used for the stress analyses in the previous section are used in the eigenvalue buck-ling analyses.

Since the theoretical elastic buckling stress does not take into account the imperfections that ex-ist within any fabricated shell structure, ie is modified in N-284 by capacity and plasticity re-duction factors. This is necessary due to the buckling phenomenon being highly sensitive to imperfections.

The capacity reduction factor, , for an unstiffened sphere in uniaxial compression equals 0.207.

In the previous analysis by GE (GE, 1991b), they employed an increased capacity reduction fac-tor due to the tensile stresses in the circumferential direction. Article 1500 of N-284 and a refer-ence by Johnson (Johnson, 1976), among others, were used to justify the use of an increased capacity reduction factor. Article 1500 and the Johnson reference explain that an increase in buckling capacity have been observed in cases where circumferential tensile stresses are pro-duced due to internal pressure. This internal pressure has the effect of smoothing out the initial imperfections that are often the site of buckling initiation. GE applied the method provided in the Johnson reference to increase the capacity reduction factor for examining buckling for both the post-accident and refueling load cases. While the post-accident case includes an internal pressure from the flooded drywell, the refueling case has no internal pressure. The circumferen-tial tensile stresses in the sandbed region for the refueling case stem from the geometry of the structure. Article 1500 of N-284 states clearly that an increased capacity reduction factor may be justified due to internal pressure. Since no further justification was provided in the previous GE analysis to use this increased factor for cases with circumferential tensile stresses not due to in-ternal pressure, this method was not adopted for the refueling load case. However, since the 67

post-accident load case includes internal pressure, a modified version of the method used by GE is applied and described in a section describing the post-accident buckling results.

The plasticity reduction factor, , for spheres under uniaxial compression is provided in N-284.

For values of < 0.55, = 1.0, and for values of 0.55 < < 1.6, = 0.45/ = 0.18, where =

ie/y and y is the material yield strength.

The compressive buckling stress, c, can be evaluated using the reduced theoretical elastic buck-ling stress that equal ie/FS, where FS equals the factor of safety. The factor of safety equal 2.0 for Service Level B (refueling) and 1.67 for Service Level D (post-accident).

4.1 Refueling Condition For the refueling load case with no degradation, the fist buckling mode occurs at the upper beam seats in the middle sphere. These locations are shown in Figure 2-23. The load that the beam applies to the drywell shell is applied to these locations with surface tractions. The original thickness of the middle sphere was 0.77 inches. Figure 4-1 illustrates the buckled displaced shape for this mode with a load factor of 13.36. The drywell shell buckles inward and down due to the load of the attached beam. In the previous GE analysis, the load for the beam seats was smeared along the entire circumference of the drywell, and therefore did not predict this type of buckling mode. Buckling modes are extremely dependent on the constraint conditions. This model does not account for the possible constraint by the beam attached to the interior surface of the shell. Without further study, it is not know if the attached beam would prevent the buckling in this region. Even so, the N-284 buckling evaluation in Table 4-1 indicates that the compres-sive stress in this region does not exceed the allowable stress for the case with no degradation.

The effective factor of safety (inelastic instability stress divided by the applied compressive stress) equals 2.77 which is larger that the factor of 2 required for Service Level B loadings.

Here, the compressive stress used in the buckling evaluation was taken at the element that shows the maximum buckled displacement (red region in Figure 4-1). Subsequent buckling modes oc-cur in other locations throughout the middle sphere, the cylinder, and then in the sandbed region of the lower sphere.

Figure 4-1. Buckling at the Upper Beam Seat for the Refueling Case with No Degradation 68

Table 4-1. Buckling Evaluation at the Upper Beam Seat for the Refueling Load Case with No Deg-radation Sphere Radius, in 420 Sphere Thickness, in 0.77 Material Yield Stress, ksi 38 Elastic Modulus, ksi 29500 Factor of Safety, FS 2 Applied Meridional Compressive Stress from Analysis, c, ksi 4.45 Load Factor from Bucking Analysis, 13.36 Theoretical Elastic Buckling Stress, ie = c, ksi 59.452 Capacity Reduction Factor, 0.207 Reduced Elastic Instability Stress, e = ie, ksi 12.307 Yield Stress Ration, = e/y 0.324 Plasticity Reduction Factor, 1.0 Inelastic Instability Stress, i = e, ksi 12.307 Allowable Compressive Stress, all = i/FS, ksi 6.153 Applied Compressive Stress Percentage of Allowable, c/all

• 100 72.3%

Effective Factor of Safety, FSE = i/c 2.77 For the refueling load case with no degradation, buckling is eventually predicted in the sandbed region as shown in Figure 4-2 with the evaluation outlined in Table 4-2. The buckling occurs in two different regions of the sandbed, between the ventlines in Bays 1 and 3, and between the ventlines in Bays 17 and 19. The largest displacements occur in the 1.154 inch thick shell be-tween Bays 1 and 3. Therefore, this location is used to evaluate the compressive buckling s-tresses. Table 4-2 shows that the effective factor of safety is 3.85 which exceeds 2.

Figure 4-2. Buckling in the Sandbed Region for the Refueling Case with No Degradation 69

Table 4-2. Buckling Evaluation in the Sandbed Region for the Refueling Load Case with No Deg-radation Sphere Radius, in 420 Sphere Thickness, in 1.154 Material Yield Stress, ksi 38 Elastic Modulus, ksi 29500 Factor of Safety, FS 2 Applied Meridional Compressive Stress from Analysis, c, ksi 4.32 Load Factor from Bucking Analysis, 18.61 Theoretical Elastic Buckling Stress, ie = c, ksi 80.374 Capacity Reduction Factor, 0.207 Reduced Elastic Instability Stress, e = ie, ksi 16.637 Yield Stress Ration, = e/y 0.438 Plasticity Reduction Factor, 1.0 Inelastic Instability Stress, i = e, ksi 16.637 Allowable Compressive Stress, all = i/FS, ksi 8.319 Applied Compressive Stress Percentage of Allowable, c/all

• 100 51.9%

Effective Factor of Safety, FSE = i/c 3.85 Figure 4-3 and Table 4-3 illustrate the buckling in the upper beam seat for the refueling load case with degradation. In this case, the thickness of the middle sphere has been reduced to 0.67.

Therefore, the stresses in this region increase leading to a decrease in the load factor (9.49). This indicates that the applied loads are closer to causing the shell to buckle. The N-284 evaluation produces an effective factor of safety equal to 1.96 which is just under the require value of 2. As discussed previously, the constraint provided by the beam may affect the buckling predicted here.

Figure 4-3. Buckling at the Upper Beam Seat for the Refueling Case with Best Estimate Degrada-tion 70

Table 4-3. Buckling Evaluation at the Upper Beam Seat for the Refueling Load Case with Best Es-timate Degradation Sphere Radius, in 420 Sphere Thickness, in 0.67 Material Yield Stress, ksi 38 Elastic Modulus, ksi 29500 Factor of Safety, FS 2 Applied Meridional Compressive Stress from Analysis, c, ksi 5.39 Load Factor from Bucking Analysis, 9.49 Theoretical Elastic Buckling Stress, ie = c, ksi 51.15 Capacity Reduction Factor, 0.207 Reduced Elastic Instability Stress, e = ie, ksi 10.59 Yield Stress Ration, = e/y 0.279 Plasticity Reduction Factor, 1.0 Inelastic Instability Stress, i = e, ksi 10.59 Allowable Compressive Stress, all = i/FS, ksi 5.29 Applied Compressive Stress Percentage of Allowable, c/all

• 100 101.8%

Effective Factor of Safety, FSE = i/c 1.96 Figure 4-4 and Table 4-4 predict buckling in the sandbed region for the refueling load case with degradation. In this analysis, sandbed Bay Combination 13-15 was the first to buckle at a thick-ness of 0.842 inches. This region is just adjacent to the local thin region (t = 0.618 inches) under the ventline in Bay 13. Since Bay Combination 9-11 (t = 0.835 in) is thinner than 13-15, it is possible the local thin region adjacent to Bay Combination 13-15 aids in the initiation of the buckling of the entire region. The effective factor of safety for this buckling mode is 2.15 which just exceeds the required value of 2.

Figure 4-4. Buckling in the Sandbed Region for the Refueling Case with Best Estimate Degrada-tion 71

Table 4-4. Buckling Evaluation in the Sandbed Region for the Refueling Load Case with Best Es-timate Degradation Sphere Radius, in 420 Sphere Thickness, in 0.842 Material Yield Stress, ksi 38 Elastic Modulus, ksi 29500 Factor of Safety, FS 2 Applied Meridional Compressive Stress from Analysis, c, ksi 4.47 Load Factor from Bucking Analysis, 10.40 Theoretical Elastic Buckling Stress, ie = c, ksi 46.49 Capacity Reduction Factor, 0.207 Reduced Elastic Instability Stress, e = ie, ksi 9.62 Yield Stress Ration, = e/y 0.253 Plasticity Reduction Factor, 1.0 Inelastic Instability Stress, i = e, ksi 9.62 Allowable Compressive Stress, all = i/FS, ksi 4.81 Applied Compressive Stress Percentage of Allowable, c/all

• 100 92.9%

Effective Factor of Safety, FSE = i/c 2.15 4.2 Post-Accident Condition The analysis of the post-accident load case with no degradation produces numerous spurious buckling modes prior to those determined to be realistic in nature. These spurious modes occur at the ends of the ventlines and equipment hatch and are judged to be caused by the approximate boundary conditions used in those regions. The first realistic buckling mode for the no degrada-tion case occurs in the cylinder. From the displaced shape for this buckling mode in Figure 4-5, it appears that it is caused by a combination of the additional lateral seismic load used for the flooded condition and the lateral constraints applied to the stabilizers.

Table 4-5 summarizes the buckling evaluation. Here the applied meridional compressive stress is actually taken as the minimum principal stress since the maximum compressive stresses in this region are slightly rotated from the meridional axis. The effective factor of safety for this mode is 2.85 which exceeds the required 2. When degradation is introduced, buckling first occurs in the critical sandbed region and not in the cylinder. Therefore, an evaluation of buckling in the degraded cylinder has not been included here.

72

Figure 4-5. Buckling in the Cylinder for the Post-Accident Load Case with No Degradation Table 4-5. Buckling Evaluation in the Cylinder for the Post-Accident Load Case with No Degrada-tion Sphere Radius, in 198 Sphere Thickness, in 0.640 Material Yield Stress, ksi 38 Elastic Modulus, ksi 29500 Factor of Safety, FS 1.67 Applied Meridional Compressive Stress from Analysis, c, ksi 2.3 Load Factor from Bucking Analysis, 13.75 Theoretical Elastic Buckling Stress, ie = c, ksi 31.625 Capacity Reduction Factor, 0.207 Reduced Elastic Instability Stress, e = ie, ksi 6.546 Yield Stress Ration, = e/y 0.172 Plasticity Reduction Factor, 1.0 Inelastic Instability Stress, i = e, ksi 6.546 Allowable Compressive Stress, all = i/FS, ksi 3.920 Applied Compressive Stress Percentage of Allowable, c/all

• 100 58.7%

Effective Factor of Safety, FSE = i/c 2.85 73

For buckling in the sandbed for the post-accident case, the allowable compressive stress is in-creased to account for the additional buckling capacity due to the internal pressure. A modified version of the procedure used by GE (GE, 1991b) is applied here. The only difference in the standard N-284 procedure is in the computation of the reduced elastic instability stress, e.

Based on the method outlined by Johnson (Johnson, 1976), e = ie + C(Et/r), where and ie are computed the same as in N-284 with C determined from a chart provided in Johnson (John-son) and reprinted by GE (GE, 1991b). The chart of C requires the computation of the X pa-rameter, where X = (P/4E)(2r/t)2. Here, P is the internal pressure within the vessel and is taken as the maximum hydrostatic pressure near the bottom of the sandbed, 0.0278 ksi. GE applied a slightly modified version of this procedure by using the computed tensile stress in the buckled region to back-out an equivalent internal pressure. They then used the C chart to compute a modified capacity reduction factor. The method used in the current study produces slightly lower allowable compressive stresses, and is therefore more conservative.

Table 4-6 shows the buckling calculations in the sandbed region for the post-accident case with no degradation and is illustrated in Figure 4-6. The largest displacement magnitudes for this buckling mode occur between the ventlines in Bays 17 and 19. After adjusting for the circum-ferential tensile stresses caused by the internal water pressure, the effective factor of safety is 3.47 which exceeds the required 1.67 for Service Level D loading.

Figure 4-6. Buckling in the Sandbed Region for the Post-Accident Load Case with No Degradation 74

Table 4-6. Buckling Evaluation in the Sandbed Region for the Post-Accident Load Case with No Degradation Sphere Radius, r, in 420 Sphere Thickness, t, in 1.154 Material Yield Stress, ksi 38 Elastic Modulus, E, ksi 29500 Factor of Safety, FS 1.67 Applied Meridional Compressive Stress from Analysis, c, ksi 6.25 Load Factor from Bucking Analysis, 13.94 Theoretical Elastic Buckling Stress, ie = c, ksi 87.12 Capacity Reduction Factor, 0.207 Internal Pressure, P, ksi 0.0278 X Parameter, X = (P/4E)(2r/t)2 0.125 C (from Johnson, 1976) 0.095 Reduced Elastic Instability Stress, e = ie+C(Et/r) , ksi 25.73 Yield Stress Ration, = e/y 0.677 Plasticity Reduction Factor, 0.844 Inelastic Instability Stress, i = e, ksi 21.73 Allowable Compressive Stress, all = i/FS, ksi 13.01 Applied Compressive Stress Percentage of Allowable, c/all

• 100 48.0%

Effective Factor of Safety, FSE = i/c 3.47 Figure 4-7 and Table 4-7 illustrate buckling in the sandbed for the post-accident load case with degradation. Buckling first occurs in Bay Combination 13-15 at a thickness of 0.842 inches.

This is just adjacent to the local thin region (t = 0.618 inches) under the ventline in Bay 13. Af-ter adjusting for the internal pressure effects, the effective factor of safety is 2.6 which exceeds the required 1.67.

Figure 4-7. Buckling in the Sandbed Region for the Post-Accident Load Case with Best Estimate Degradation 75

Table 4-7. Buckling Evaluation in the Sandbed Region for the Post-Accident Load Case with Best Estimate Degradation Sphere Radius, r, in 420 Sphere Thickness, t, in 0.842 Material Yield Stress, ksi 38 Elastic Modulus, E, ksi 29500 Factor of Safety, FS 1.67 Applied Meridional Compressive Stress from Analysis, c, ksi 7.99 Load Factor from Bucking Analysis, 7.58 Theoretical Elastic Buckling Stress, ie = c, ksi 60.53 Capacity Reduction Factor, 0.207 Internal Pressure, P, ksi 0.0278 X Parameter, X = (P/4E)(2r/t)2 0.234 C (from Johnson, 1976) 0.14 Reduced Elastic Instability Stress, e = ie+C(Et/r) , ksi 20.81 Yield Stress Ration, = e/y 0.547 Plasticity Reduction Factor, 1.0 Inelastic Instability Stress, i = e, ksi 20.81 Allowable Compressive Stress, all = i/FS, ksi 12.46 Applied Compressive Stress Percentage of Allowable, c/all

• 100 64.1%

Effective Factor of Safety, FSE = i/c 2.60 4.3 Conclusion The buckling evaluation performed here using ASME N-284 show that based on the loadings and the model described in Section 2, both the refueling and post-accident load combinations met buckling requirements with a one exception. The buckling at the upper beam seat for the refueling load case with degradation does not met the required factor of safety of 2. As de-scribed earlier, the potential constraint provided by the attached beam has not been included in this analysis. In all cases, the introduction of degradation causes a significant decrease in the ef-fective factor of safety against buckling. In the sandbed region, the degraded state analyzed in this study predicts an effective factor of safety of 2.15. This model includes spatial variation in the degradation and two local areas with increased thinning. In order to establish a minimum acceptable uniform thickness, an additional study was performed and is described in the next section.

76

5. Sandbed Region Minimum Thickness Study In addition to the stress and stability analysis of the drywell shell using the average UT meas-urements in the sandbed region (thicknesses described in Section 2.6, and analyses outlined in Sections 3 and 4), a minimum sandbed thickness study was also performed. These analyses aim to establish the minimum uniform thickness in the sandbed region that maintains compliance with the ASME B&PV code. The minimum acceptable shell thickness established here is based on a buckling (stability) analysis for the refueling load case. The refueling load case appears to govern the potential for instability since a relatively low effective factor of safety was produced in the average UT measurement analysis at 2.15. For Service Level B (refueling condition), a factor of safety of 2.0 is required by ASME N-284.

The previous GE analysis (GE, 1991b) assumed a uniform sandbed shell thickness of 0.736.

Their analyses produced an applied compressive stress of 7.58 ksi in the sandbed region and an inelastic buckling stress of 21.30 ksi (per ASME N-284). This produces an effective factor of safety of 2.81. A subsequent calculation documented in a 1993 GPU Nuclear Calculation Sheet (GPU Nuclear, 1993) shows an applied compressive stress of 7.58 ksi in the sandbed region for a shell thickness of 0.736, but with a lower value for the inelastic buckling stress at 15.18 ksi.

This produces an effective factor of safety of 2.0, or at the required ASME N-284 value.

The inconsistency between the two calculations appears to stem from a difference in the applica-tion of the increased capacity reduction factor due to the tensile stresses in the circumferential (hoop) direction. This issue was discussed in detail in the previous stability analysis section. Ar-ticle 1500 of ASME N-284 states clearly that an increased capacity reduction factor may be jus-tified if an internal pressure loading is present and causes tensile stresses in the circumferential direction. This internal pressure aids in smoothing the initial imperfections and increased the buckling capacity under compressive meridional stresses. The lack of an internal pressure load for the refueling load case prevents the justified use of an increased capacity reduction factor.

As with the buckling calculations for the refueling load case in the previous section, the mini-mum thickness study does not employ any increase in the capacity reduction factor.

The shell thicknesses used in the minimum thickness study are summarized in Table 5-1 for re-gions outside of the sandbed region. The degraded thickness values for the majority of the dry-well are equivalent to the values used in the average UT measurement analysis. The only exception being the thickness assigned to the lower sphere above an elevation of 15-6.8, or the center of the ventlines. In this region of the lower sphere (see Figure 5-1), the thickness is set to 1.154, or the nominal as-built value. This remains consistent with inspections of the upper por-tions of the lower sphere. In the average UT measurement analysis, additional conservatism was introduced by degrading the entire lower sphere uniformly in each bay combination. However, several confirmatory analyses performed during this study showed that the thickness assigned to the lower sphere above elevation 15-6.8 has only a negligible effect since the buckling occurs in the sandbed below 15-6.8.

In the lower sphere below elevation 15-6.8 (sandbed region), the drywell shell is set to a uni-form thickness. This region is shown in Figure 5-1. While the same finite element mesh was used as for the average UT measurement analyses, the local thinned regions under the ventlines 77

for Bays 1 and 13 are uniformly thinned consistent with the surrounding shell. In addition, this study only examined the minimum thickness required in the sandbed region and not in the upper portions of the sphere or in the cylinder.

Table 5-1. Main Drywell Shell Model Thicknesses Outside of Sandbed Region Original Degraded Original Degraded Section Thickness, Thickness, Section Thickness, Thickness, in in in in Head 1.1875 N/C Reinforcing Around Ventlines 2.875 2.618 Upper Cylinder 1.1875 N/C Lower Sphere (below Sandbed) 1.154 N/C Main Cylinder 0.640 0.585 Bottom Sphere 0.676 N/C Knuckle 2.5625 2.54 Middle Sphere Thickened 1.0625 0.9625 Upper Sphere 0.722 0.676 Reinforcing Around Hatch 2.625 2.525 Middle Sphere 0.770 0.670 Lower Sphere (above El. 15-6.8) 1.154 N/C N/C - No Change Lower Sphere Above Elevation 15-6.8 Thickness = 1.154 Elevation 15-6.8 Elevation 8-11.25 Lower Sphere Below Elevation 15-6.8 (Sandbed)

Thickness = Varied to Establish Minimum Figure 5-1. Drywell Lower Sphere for Establishing a Minimum Thickness in the Sandbed Region (Ventlines and Hatch Removed for Clarity) 78

The thickness values assigned to the sandbed region were varied from 0.800 up to 1.050 with a concentration of analyses performed between 0.800 and 0.860. In the previous buckling analysis section, a buckling analysis was also performed for the undegraded drywell containment which included a uniform thickness of 1.154 throughout the sandbed region. The results of each of these analyses are summarized in Figure 5-2. Here, the effective factor of safety is plot-ted against the associated shell thickness in the sandbed region. This study shows that a thick-ness of 0.844 is required in the sandbed region to produce an effective factor of safety equal to the ASME N-284 value of 2.0.

Figure 5-2 also plots the datapoint established in the previous buckling analysis section using average UT measurement data. In that analysis, the bay combination that buckled first was set to a thickness of 0.842 and resulted in an effective factor of safety equal to 2.15. Although the thicknesses used in the minimum thickness analysis and the average UT measurement analysis are essential equivalent, there are several important factors that produce the difference in safety factors. First, the average UT measurement analysis included two locally thinned regions that, in general, cause lower effective factors of safety for buckling in the adjacent bays than without the locally thinned regions. However, the effect of the locally thinned is outweighed by the exis-tence of bay combinations with thickness far exceeding 0.842 (see Figure 2-32). For the aver-age UT measurement analyses, 5 out of the 10 bay combinations were assigned thicknesses near or above 0.9. The existence of thicker bays enables a redistribution of the compressive loads leading to buckling. Therefore, the average UT measurement analysis produced an effective fac-tor of safety of 2.15 with a thickness of 0.842, while the minimum thickness study produced an effective factor of safety of 2.0 with a thickness of 0.844. In the minimum thickness study, the entire sandbed region was uniformly thinned which prevents any redistribution of the load through thicker shell regions. The effect of the locally thinned regions was not rigorously ex-plored in the average UT measurement analyses, but it is likely that the effective factor of safety of 2.15 would increase without the presence of the locally thinned region.

4 Effective Factor of Safety 3.5 Minimum Thickness Study 3

Average UT Measurment Analysis 2.5 2

1.5 0.6 0.7 0.8 0.9 1 1.1 1.2 Sandbed Shell Thickness, in Figure 5-2. Effective Factor of Safety Values Computed for Various Thicknesses in the Sandbed Region for the Refueling Load Combination 79

Figure 5-3 and Table 5-2 illustrate the buckling location and ASME N-284 calculations for the sandbed with a thickness of 0.844. The major displacements for the first buckling mode in the sandbed are located between the ventline in Bays 1 and 3.

Figure 5-3. Buckling in the Sandbed Region with a Thickness of 0.844 for the Refueling Load Combination Table 5-2. Buckling Evaluation for the Refueling Load Case with a Thickness of 0.844 in the Sandbed Sphere Radius, in 420 Sphere Thickness, in 0.844 Material Yield Stress, ksi 38 Elastic Modulus, ksi 29500 Factor of Safety, FS 2 Applied Meridional Compressive Stress from Analysis, c, ksi 4.78 Load Factor from Bucking Analysis, 9.67 Theoretical Elastic Buckling Stress, ie = c, ksi 46.19 Capacity Reduction Factor, 0.207 Reduced Elastic Instability Stress, e = ie, ksi 9.56 Yield Stress Ration, = e/y 0.252 Plasticity Reduction Factor, 1.0 Inelastic Instability Stress, i = e, ksi 9.56 Allowable Compressive Stress, all = i/FS, ksi 4.78 Applied Compressive Stress Percentage of Allowable, c/all

• 100 100.0%

Effective Factor of Safety, FSE = i/c 2.00 80

6. Summary of Assumptions The study performed for this program required a number of assumptions. A summary of the most significant assumptions is provided below.

• The Accident and Post-Accident load combinations are assumed to govern the stress analysis.

• The Refueling and Post-Accident load combinations are assumed to govern the buckling (stability) analysis.

• Information of the loads applied to the finite element model was taken from the previous study by GE. These loads were not independently verified.

• The seismic loading was applied using static coefficients provided in the Final Safety Analysis Report (FSAR, 2003). The static coefficients were applied using body forces in both the vertical and lateral directions. The displacement time histories (ground motions) were not made available for this study. The body forces used for the seismic loads were increased in the post-accident load combination to account for the mass of the water flooding the drywell.

• The ventlines were modeled down to the intersection with the ventline header. Here, springs acting in the radial and vertical directions were added to approximate the compli-ance of the ventline header. The spring constants were based on a simple submodel analysis of the ventline header. Since the ventline is connected to the torus with a flexi-ble bellow, all interaction between the ventline and torus was neglected.

• The ventline jet deflector was modeled as a solid plate. In reality, the deflector has mul-tiple holes throughout the plate. The thickness of the solid plate in the current model was reduced to account for the holes.

• In a number of cases, the exact location that a specific load acts upon the drywell shell was not known. The magnitude and elevation of these loads were provided in the GE re-port, but the azimuth locations remained unknown. In these cases (mainly in the case of the penetration loads), the loads were distributed along the entire circumference of the drywell as a surface traction.

• The loads applied to the drywell shell were smeared along a region defined on the shell surface. Typically, the region of application was taken as the area where an item is actu-ally attached to the shell in the real structure. As mentioned above, the penetration loads were smeared along the entire circumference since loads for individual penetrations were not provided.

• The spacing of the upper and lower beam seats around the circumference is not constant, but the appropriate load distribution at each seat was not known. The loads for the upper 81

and lower beam seats were distributed equally at each point of attachment to the drywell shell.

• The concrete that fills the drywell shell interior from an elevation of 8-11.25 to 10-3, and the additional curbs, have not been accounted for in this model. The drywell shell is assumed encased in concrete below elevation 8-11.25 (bottom of sandbed).

• For the accident load combination, the internal 44 psi pressure and the thermal load of 292oF (starting at 70oF) were applied to the entire drywell shell down to an elevation of 8-11.25. The concrete within the interior of the drywell shell extends up to 10-3 with curbs extending up to 12-3. Since the bond between the steel shell and the concrete is not known, it was assumed that a gap could exist which would enable gas to pressurize and heat the shell down to 8-11.25, or the bottom of the sandbed region on the exterior of the shell. Even if no gap exists initially, it is likely that the initial pressurization (pres-sure << 44 psi) acting on the shell above elevation 10-3 would cause a gap to open.

This would allow heated gas to flow between the shell and concrete.

• The Personnel Lock & Equipment Hatch penetration geometry (extent modeled and thicknesses assigned) was approximated and the outer surface fixed against vertical dis-placement.

• The coefficient of thermal expansion for the A-212-61T Grade B pressure vessel steel used for the drywell was assumed to be 6.5E-6oF-1.

• A number of assumptions were made to develop the thicknesses assigned to the model in its degraded state. Section 2.6 provides a detailed discussion of these items.

• A very limited mesh convergence study was performed which led to the use of a 4 nomi-nal element size. It was assumed that this mesh size was acceptable even though all load combinations were not examined in the convergence study and no checks on buckling were performed using different mesh sizes. In addition, a 1 nominal mesh size was used in the two local regions under the ventlines in Bay 1 and 13. No checks were performed to assess the mesh size in these regions.

• Several assumptions were made in developing the ASME stress limits. These are dis-cussed in Section 3. .

• ASME Code Case N-284 was used to assess the stability of the degraded drywell shell.

It was assumed that since the refueling case does not include any internal pressure, that the increase in buckling capacity used by GE for cases with circumferential (hoop) ten-sion was not appropriate. Since the post-accident load case includes internal pressure, an increase in the capacity was applied.

82

7. Conclusions The structural integrity of the degraded Oyster Creek drywell shell has been analyzed in this study. The allowable stresses and the buckling stability were both examined in accordance with the ASME B&PV code. The ASME allowable stresses are met for all three load cases examined here given the modeling and loading procedures outlined in Section 2. The only potential excep-tion is for the primary plus secondary stresses located at the base of the sandbed region of the accident condition due to the thermal expansion of the shell. There are a number of modeling and loading assumptions in this region that may contribute to the stress magnitudes recorded in the current analysis. In addition, the primary plus secondary stresses were compared to the al-lowables use in the previous GE analysis (GE, 1991a). The current code does not require an evaluation of the primary plus secondary stresses for Service Level C. However, these stresses were assessed in this report to be consistent with the previous evaluation by GE. The buckling evaluation performed here using ASME N-284 show that based on the loadings and the model described in Section 2 both the refueling and post-accident load combinations met buckling re-quirements with a one exception. The buckling at the upper beam seat for the refueling load case with degradation does not meet the required factor of safety of 2. As described in Section 4, the potential constraint provided by the attached beam has not been included in this analysis. Table 7-1 summarizes the major conclusions for this study and for the previous GE analyses.

Table 7-1. Comparison of Conclusion Between GE Study (GE, 1991a and b) and the Current Study Current Study Conclusion GE Study Conclusion The ASME B&PV stress analysis of the de- The ASME B&PV stress analysis of the de-graded Oyster Creek drywell shows all values graded Oyster Creek drywell shows all values within code limits. The current study uses within code limits. The GE study assumed a average UT measurement data to assign conservative uniform thickness of 0.736 in the thicknesses in the sandbed region. (Note that sandbed region.

some primary plus secondary stresses for the accident condition are of concern as dis-cussed in Section 3.)

ASME B&PV Code Case N-284 stability analy- ASME B&PV Code Case N-284 stability analysis sis of the degraded Oyster Creek drywell of the degraded Oyster Creek drywell shows shows that acceptable factors of safety are that acceptable factors of safety are met. The met. The current study uses average UT GE study assumed a conservative uniform measurement data to assign thicknesses in thickness of 0.736 in the sandbed region.

the sandbed region. (Note that the buckling at the upper beam seats produces an effec-tive factor of safety slightly less than 2 for the refueling load case, but this may be affected by the modeling of that specific detail.)

The minimum uniform shell thickness re- The minimum uniform shell thickness required quired to meet the ASME N-284 buckling to meet the ASME N-284 buckling safety factor safety factor was determined to be 0.844 in was determined to be 0.736. This thickness the sandbed region. This thickness was es- was established using the refueling load case.

tablished using the buckling analysis for the (The thickness of 0.736 was established in a refueling load case. calculation by GPU Nuclear, 1993. This calcula-tion included an increase in the capacity reduc-tion factor not used in the current study.)

83

The assessments performed here employ a uniform thinning of the drywell shell over large sec-tions of the surface. The thicknesses assigned in each region were based on limited measure-ment data since a very small percentage of the shell has been examined. In many cases, the raw data was not available. This led to the use of averages provided by AmerGen throughout the relevant documentation.

84

8. References ABAQUS Analysis Users Manual, Version 6.5. 2004. Pawtucket, RI: Hibbit, Karlsson & Soren-son, Inc.

AmerGen April 4, 2006, Letter to U.S. NRC, 2130-06-20284, NRC Docket No. 50-219.

AmerGen April 7, 2006, Letter to U.S. NRC, 2130-06-20289, NRC Docket No. 50-219.

ASME (American Society of Mechanical Engineers). 2004. Boiler and Pressure Vessel Code.

Sections III and VIII. New York, NY: American Society of Mechanical Engineers.

Cherry, J.L., and Smith, J.A., Capacity of Steel and Concrete Containment Vessels with Corro-sion Damage, NUREG/CR-6707, SAND2000-1735, U.S. Nuclear Regulatory Commission, 2000.

Chicago Bridge & Iron (CB&I, 1980) Structural Drawings of the Oyster Creek Drywell, Torus, and Ventlines, attached to AmerGen letter to NRC, July 7, 2006, NRC Docket No. 50-219.

Final Safety Analysis Report (Updated), Oyster Creek Nuclear Generating Station (FSAR),

AmerGen, Exelon, Revision 13, April 2003.

GE Stress Analysis Report, An ASME Section VIII Evaluation of the Oyster Creek Drywell for Without Sand Case - Part I - Stress Analysis, GE Index #9-3, DRF #00664, February 1991a.

GE Stability Analysis Report, An ASME Section VIII Evaluation of the Oyster Creek Drywell for Without Sand Case - Part II- Stability Analysis, GE Index #9-4, DRF #00664, February 1991b.

GPU Nuclear, Calculation Sheet, Oyster Creek Drywell External UT Evaluation, C-1302-187-5320-024, 1993.

Individual Plant Examination Submittal Report, Oyster Creek, (IPE), June 1992.

Johnson, B,G. editor, Guide to Stability of Design Criteria for Metal Structures, Third Edition, John Wiley & Sons, 1976.

Spencer, B.W., Petti, J.P., and Kunsman, D.M., Risk-Informed Assessment of Degraded Con-tainment Vessels, NUREG/CR-6920, SAND2006-3772P, U.S. Nuclear Regulatory Commission, 2006.

Standard Review Plan for the Review of Safety Analysis Reports for Nuclear Power Plants (SRP), NUREG-0800, U.S. Nuclear Regulatory Commission, 1996.

85

9. Appendix A - Natural Frequency Extraction An eigenvalue extraction was performed to calculate the natural frequencies for the degraded drywell shell in addition to the drywell shell in its original condition. This analysis was per-formed to assess the potential effects of degradation on the dynamic behavior of the drywell con-tainment during a seismic event. The change in the natural frequencies when introducing degradation can provide justification for using a simplified static seismic analysis. The FSAR (FSAR, 2003) for Oyster Creek include the static seismic coefficients used in the original design of the drywell. If only minor differences exist between the natural frequencies for the contain-ment in its original condition (the condition used to establish the static coefficients) and the con-tainment in its degraded state, the dynamic behavior of the containment would not be significantly altered by the degradation. Therefore, the same static seismic coefficients could be applied when simulating the seismic loading for the degraded containment as for the as-built containment.

The model used previously for the stress and buckling analyses was modified and used to per-form a natural frequency extraction in ABAQUS. Initial attempts to extract the natural frequen-cies for the drywell structure used the same model described in Section 2. That model included a section of the personnel lock/equipment hatch and the 10 ventlines down to the intersection with the ventline header. The natural frequency extraction analyses that included these penetrations resulted in spurious modes where the displacements of each mode concentrated at the ends of the penetrations. At these locations, the boundary conditions are approximated and applied to the structure as described in Section 2. These applied boundary conditions and approximated ge-ometry cause these spurious, or unrealistic, mode shapes. In order to avoid these spurious modes, the geometry of the drywell was simplified by removing the hatch and ventline penetra-tions. The resulting holes in the drywell shell were subsequently filled-in with solid mate-rial to avoid spurious mode shapes with deformation concentrated around the holes. This results in the geometry illustrated in Figure 9-1. The nodes along the bottom of the sphere below eleva-tion 8-11.25 are fixed in all directions and the seismic stabilizers are fixed against lateral dis-placement as described for the full model in Section 2.

The thicknesses for the drywell with and without degradation are summarized in Table 9-1. The degraded thicknesses in the cylinder and upper sections of the sphere are the same as in the aver-age UT measurement analysis. The thickness of the entire lower sphere is set to a uniform value of 0.835 for the natural frequency extraction of the degraded drywell. This region is high-lighted in Figure 9-2. The thickened reinforcing plates surrounding the hatch and ventlines in the actual structure are not thickened in the frequency extraction performed here. These simpli-fications in the geometry enable a general assessment of the effects of degradation on the natural frequencies. This analysis was not intended to provide the exact frequencies for the drywell structure, but only to justify the use of the static seismic coefficients for the seismic loading component of the stress and buckling analyses. Therefore, the assumptions (e.g. the use of 0.835 for the thickness in the lower sphere and the simplified geometry) are judged to be ac-ceptable in order to study the general effect of degradation on the natural frequencies.

86

Figure 9-1. Modified Model for Natural Frequency Extraction Table 9-1. Drywell Shell Thicknesses for Natural Frequency Extraction Analyses Original Degraded Original Degraded Section Thickness, Thickness, Section Thickness, Thickness, in in in in Head 1.1875 N/C Middle Sphere 0.770 0.670 Upper Cylinder 1.1875 N/C Bottom Sphere 0.676 N/C Main Cylinder 0.640 0.585 Middle Sphere Thickened 1.0625 0.9625 Knuckle 2.5625 2.54 Lower Sphere Below Sandbed 1.154 N/C Upper Sphere 0.722 0.676 Lower Sphere in Sandbed and Above 1.154 0.835 N/C - No Change 87

Figure 9-2. Lower Sphere Region (Highlighted in Red) Set to a Thickness of 0.835 for the De-graded Natural Frequency Extraction The lowest 5 frequencies and mode shapes from the ABAQUS eigenvalue extraction are illus-trated in Figure 9-3 and Figure 9-4 for the containment without and with degradation, respec-tively. The frequencies show only minimal decreases with the introduction of degradation with the lowest frequency dropping from 20.46Hz to 19.12Hz. The differences are smaller at the higher frequencies (modes 2 through 5). The displacements for each of the mode shapes are nearly identical. The first mode is a vertical extension of the drywell, or stretching mode. The second and third modes are overturning modes where the drywell is bent between the cylindri-cal and spherical sections of the structure. The forth and fifth modes are compressive modes where the cylinder is compressed down vertically toward the drywell sphere. These shapes are generally consistent between the analyses with and without degradation. Table 9-2 summarizes the comparison of the frequencies for the two analyses. Since the effects of the degradation on the frequencies and associated mode shapes are minimal, the use of the original design static seismic coefficients is judged to be acceptable.

88

Mode 0 Mode 1 Base State freq = 20.46Hz Mode 2 Mode 3 freq = 23.78Hz freq = 23.80Hz Mode 4 Mode 5 freq = 27.89Hz freq = 27.93Hz Figure 9-3. Base State and the First 5 Frequencies and Mode Shapes for the Drywell Containment with No Degradation 89

Mode 0 Mode 1 Base State freq = 19.12Hz Mode 2 Mode 3 freq = 23.26Hz freq = 23.26Hz Mode 4 Mode 5 freq = 27.52Hz freq = 27.56Hz Figure 9-4. Base State and the First 5 Frequencies and Mode Shapes for the Drywell Containment with Degradation Table 9-2. Summary of the First 5 Natural Frequencies for Drywell with and without Degradation Mode Frequency - No Degradation, Hz Frequency - Degraded Model, Hz 1 20.46 19.12 2 23.78 23.26 3 23.80 23.26 4 27.89 27.52 5 27.93 27.56 90

10. Appendix B - Sandbed UT Measurement Data and Shell Thickness Development For modeling the degradation in the sandbed region, the lower sphere was divided into 10 re-gions to be assigned uniform thicknesses. These regions extend from the centerline of one ven-tline to the centerline of the adjacent ventline. Each of these regions contains one-half of the two different, but adjacent, bays. This was done in order to avoid placing the thickness discontinuity at the centerline between the ventlines, since this is typically the location of the highest stresses.

If the thickness jump was placed at this location, the stresses of interest would be difficult to in-terpret. An example of the bay combinations is illustrated in Figure 10-1. Here, half of Bay 1 and half of Bay 2 are combined to create Bay Combination 1-3. The measurement points indi-cated on the images (GPU Nuclear, 1993) were taken from the outside of the containment shell prior to the application of the epoxy coating. For Bay Combination 1-3, Points 8, 9, 15, 18, and 19 were taken from the left half of Bay 1 and Points 1, 2, 3, and 7 were taken from the right half of Bay 3, and averaged. The thicknesses for these points were reported in the GPU Nuclear cal-culations (GPU Nuclear, 1993) and are provided in Table 10-1. This average was assigned as a uniform thickness to the region highlighted in light red in Figure 10-1 and shown on the model in Figure 2-30. The points that fall within the bathtub region (Points 1, 2, 3, 4, 5, 10, 11, 12, 13, 20, and 21) under the ventline in Bay 1 were not included in the average for the adjacent bay combinations. The minimum measured thickness (Point 3) in this region was assigned to the en-tire Local Bay 1 region as outlined in Figure 10-1 and shown on the model in Figure 2-31.

Ventlines Bay Combination 1-3 Local Bay 1 Region Elevation 8-11.25 Figure 10-1. Bay 1 and Bay 3 UT Measurement Locations Taken from Outside of the Containment (Images Extracted from GPU Nuclear Calculation Sheet, 1993) 91

Table 10-1 through Table 10-4 and Figure 10-2 through Figure 10-11 provide the individual datapoints (GPU Nuclear, 1993) and the grouping used to compute the averages for all of the bay combinations summarized in Table 2-7. The bay combinations are assembled and averaged in the same manner as for Bay Combination 1-3 in Figure 10-1. The Local Bay 13 is shown in Figure 10-8 with thickness provided in Table 10-4. As with the Local Bay 1 region, the mini-mum measured value (Point 7) in the defined region was assigned as a uniform thickness.

Table 10-1. UT Measurement Data for Bay Combinations 1-3, 3-5, 5-7, and 7-9.

Bay Combination 1-3 Bay UT Point Shell Thickness, in 1 8 0.805 1 9 0.805 1 15 1.156 1 18 0.917 1 19 0.89 3 1 0.795 (min) 3 2 1.00 3 3 0.857 3 7 0.826 1-3 average 0.894 Bay Combination 3-5 Bay UT Point Shell Thickness, in 3 4 0.898 3 5 0.823 3 6 0.968 3 8 0.78 (min) 5 1 0.97 5 2 1.04 5 3 1.02 5 4 0.91 5 5 0.89 3-5 average 0.922 Bay Combination 5-7 Bay UT Point Shell Thickness, in 5 6 1.06 5 7 0.99 5 8 1.01 7 1 0.92 (min) 7 2 1.016 7 3 0.954 7 4 1.04 5-7 average 0.998 Bay Combination 7-9 Bay UT Point Shell Thickness, in 7 5 1.03 7 6 1.045 7 7 1.00 9 1 0.96 9 2 0.94 (min) 9 3 0.994 9 4 1.02 7-9 average 0.998 92

Table 10-2. UT Measurement Data for Bay Combinations 9-11, 11-13, and 13-15.

Bay Combination 9-11 Bay UT Point Shell Thickness, in 9 5 0.985 9 6 0.82 9 7 0.825 9 8 0.791 9 9 0.832 9 10 0.98 11 1 0.705 (min) 11 2 0.77 11 7 0.831 11 8 0.815 9-11 average 0.835 Bay Combination 11-13 Bay UT Point Shell Thickness, in 11 3 0.832 11 4 0.755 11 5 0.831 11 6 0.800 13 1 0.672 (min) 13 2 0.722 13 3 0.941 13 4 0.915 13 9 0.924 13 13 0.932 13 17 0.807 13 18 0.825 13 19 0.912 13 20 1.17 11-13 average 0.859 Bay Combination 13-15 Bay UT Point Shell Thickness, in 13 12 0.885 13 16 0.829 15 1 0.786 (min) 15 2 0.829 15 3 0.932 15 4 0.795 13-15 average 0.842 93

Table 10-3. UT Measurement Data for Bay Combinations 15-17, 17-19, and 19-1.

Bay Combination 15-17 Bay UT Point Shell Thickness, in 15 5 0.85 15 6 0.794 15 7 0.808 15 8 0.77 15 9 0.722 15 10 0.86 15 11 0.825 17 1 0.916 17 2 1.15 17 3 0.898 17 4 0.951 17 5 0.913 17 9 0.72 (min) 17 10 0.83 15-17 average 0.857 Bay Combination 17-19 Bay UT Point Shell Thickness, in 17 6 0.992 17 7 0.97 17 8 0.99 17 11 0.77 19 1 0.932 19 2 0.924 19 3 0.955 19 4 0.94 19 5 0.95 19 8 0.753 (min) 19 9 0.776 17-19 average 0.904 Bay Combination 19-1 Bay UT Point Shell Thickness, in 19 6 0.86 19 7 0.969 19 10 0.79 1 6 0.76 1 7 0.70 (min) 1 14 1.147 1 16 0.796 1 17 0.86 1 22 0.852 1 23 0.85 19-1 average 0.858 94

Table 10-4. UT Measurement Data for Local Bay 1 and 13 Regions.

Local Bay 1 Region Bay UT Point Shell Thickness, in 1 3 0.705 (min) 1 4 0.76 1 5 0.71 1 12 0.724 1 13 0.792 1 1 0.72 1 2 0.716 1 10 0.839 1 11 0.714 1 20 0.965 1 21 0.726 1 min 0.705 Local Bay 13 Region Bay UT Point Shell Thickness, in 13 5 0.718 13 10 0.728 13 14 0.868 13 6 0.655 13 7 0.618 (min) 13 8 0.718 13 11 0.685 13 15 0.683 13 min 0.618 95

Local Bay 1 Region Bay Combination 1-3 Bay Combination 19-1 Elevation 8-11.25 Figure 10-2. Bay 1 UT Measurement Locations Taken from Outside of the Containment (Image Extracted from GPU Nuclear Calculation Sheet, 1993)

Bay Combination 3-5 Bay Combination 1-3 Elevation 8-11.25 Figure 10-3. Bay 3 UT Measurement Locations Taken from Outside of the Containment (Image Extracted from GPU Nuclear Calculation Sheet, 1993) 96

Bay Combination 5-7 Bay Combination 3-5 Elevation 8-11.25 Figure 10-4. Bay 5 UT Measurement Locations Taken from Outside of the Containment (Image Extracted from GPU Nuclear Calculation Sheet, 1993)

Bay Combination 7-9 Bay Combination 5-7 Elevation 8-11.25 Figure 10-5. Bay 7 UT Measurement Locations Taken from Outside of the Containment (Image Extracted from GPU Nuclear Calculation Sheet, 1993) 97

Bay Combination 9-11 Bay Combination 7-9 Elevation 8-11.25 Figure 10-6. Bay 9 UT Measurement Locations Taken from Outside of the Containment (Image Extracted from GPU Nuclear Calculation Sheet, 1993)

Bay Combination 11-13 Bay Combination 9-11 Elevation 8-11.25 Figure 10-7. Bay 11 UT Measurement Locations Taken from Outside of the Containment (Image Extracted from GPU Nuclear Calculation Sheet, 1993) 98

Local Bay 13 Region Bay Combination 11-13 Bay Combination 13-15 Elevation 8-11.25 Figure 10-8. Bay 13 UT Measurement Locations Taken from Outside of the Containment (Image Extracted from GPU Nuclear Calculation Sheet, 1993)

Bay Combination 15-17 Bay Combination 13-15 Elevation 8-11.25 Figure 10-9. Bay 15 UT Measurement Locations Taken from Outside of the Containment (Image Extracted from GPU Nuclear Calculation Sheet, 1993) 99

Bay Combination 17-19 Bay Combination 15-17 Elevation 8-11.25 Figure 10-10. Bay 17 UT Measurement Locations Taken from Outside of the Containment (Image Extracted from GPU Nuclear Calculation Sheet, 1993)

Bay Combination 17-19 Bay Combination 19-1 Elevation 8-11.25 Figure 10-11. Bay 19 UT Measurement Locations Taken from Outside of the Containment (Image Extracted from GPU Nuclear Calculation Sheet, 1993) 100

DISTRIBUTION 1 P.T. Kuo U.S. Nuclear Regulatory Commission Mailstop: O-11F1 Washington, DC 20555-0001 1 Louise Lund U.S. Nuclear Regulatory Commission Mailstop: O-11F1 Washington, DC 20555-0001 1 Donnie Ashley U.S. Nuclear Regulatory Commission Mailstop: O-11F1 Washington, DC 20555-0001 1 Noel Dudley U.S. Nuclear Regulatory Commission Mailstop: O-11F1 Washington, DC 20555-0001 1 Patrick Hiland U.S. Nuclear Regulatory Commission Mailstop: O-9E3 Washington, DC 20555-0001 1 Sujit Samadar U.S. Nuclear Regulatory Commission Mailstop: O-9D3 Washington, DC 20555-0001 1 Hans Ashar U.S. Nuclear Regulatory Commission Mailstop: O-9D3 Washington, DC 20555-0001 1 Samir Chakrabarti U.S. Nuclear Regulatory Commission Mailstop: O-12D5 Washington, DC 20555-0001 101

1 Sally Adams U.S. Nuclear Regulatory Commission Mailstop: O-12E5 Washington, DC 20555-0001 1 Randolf Blough U.S. Nuclear Regulatory Commission 474 Allendale Road King of Prussia, PA 19406-1415 1 Richard Conte U.S. Nuclear Regulatory Commission 474 Allendale Road King of Prussia, PA 19406-1415 1 Michael Modes U.S. Nuclear Regulatory Commission 474 Allendale Road King of Prussia, PA 19406-1415 1 Timothy OHara U.S. Nuclear Regulatory Commission 474 Allendale Road King of Prussia, PA 19406-1415 5 MS 0744 Jason Petti, 6764 1 0744 Mike Hessheimer, 6764 1 0748 Matt Turgeon, 6764 1 0718 Jeff Smith, 6765 1 0736 Marianne Walck, 6760 2 9018 Central Technical Files, 8944 2 0899 Technical Library, 4536 102