NRC-96-4617, Forwards Info Provided to Address NRC Open Issues Related to AP600 Steel Containment.Draft Rev of Subsection 3.8.2 of Ssar,Draft Rev of Chapter 42 of AP600 PRA & Responses to NRC Requests for Addl Info Also Encl
| ML20096B289 | |
| Person / Time | |
|---|---|
| Site: | 05200003 |
| Issue date: | 01/04/1996 |
| From: | Mcintyre B WESTINGHOUSE ELECTRIC COMPANY, DIV OF CBS CORP. |
| To: | Quay T NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| NTD-NRC-96-4617, NUDOCS 9601160184 | |
| Download: ML20096B289 (59) | |
Text
m.
c l
Westinghouse Energy Systems ecx355 ;
Pinsburgh PennsyNada 15230 0355 l Electric Corporation NTD-NRC-96-4617 DCP/NRCN45 l Docket No : STN-524X)3 January 4,1996 Document Control Desk U.S. Nuclear Regulatory Commission Washington, D.C. 20555 1 A'ITENTION: T. R. QUAY
SUBJECT:
AP600 CONTAINMENT I
Dear Mr. Quay:
Enclosed are two copies of information provided to address NRC open issues related to the AP600 steel containment. Included are a draft revision of Subsection 3.8.2 of the SSAR, a draft revision of i Chapter 42 of the AP600 PRA, and responses to NRC requests for additional information.
The SSAR Subsection 3.8.2 includes changes in the text to address DSER open items and questions i raised in meetings and on phone calls. The figures have also been updated. Once the revision of the sections is finalized, a number of figures will be deleted and unnecessary details removed from the remaining figures as shewn in the draft figures. This action should resolve the question about proprietary information in this section. You may note that the information in this transmittal is nonproprietary.
The sections in Chapter 42 related to methodology have been updated to resolve open items. The results in Sections 42.5 and 42.6 have not been requantified pending NRC concurrence on the resolution of methodology open items.
The responses to the NRC requests for additional information are in response to a letter dated. f September 14,1995. i If you have any questions, please contact D. A. Lindgren at (412) 3744856.
04ff Brian A. McIntyre, h ana . r Advanced Plant Safety and Licensing 7nj.a 9601160184 960104 PDR ADOCK 05200003 A PDR _
Enclosures cc: Diane Jackson, NRC Nicholas J. Liparuto, Westinghouse (w/o Enclosures) r 160016 ,
' 'I I
o 4
- 3. Design of Structures, Components, Equipment, and Systems l
l 3.8 Design of Category I Structures 3.8.1 Concrete Containment 1
nis subsection is not applicable to the AP600. .
l 3.8.2 Steel Containment 3.8.2.1 Description of the Containment 3.8.2.1.1 General l This subsection describes the structural design of the steel containment vessel and its parts and l appurtenances. The steel containment vessel is an integral part of the containment system l whose function is described in Section 6.2. It serves both to limit releases in the event of an i accident and to provide the safety related ultimate heat sink.
l The containment vessel is an ASME metal containment. The information contained in this I subsection is based on the des ~ign specification and preliminary design and analyses of the l l
vessel. Final detailed analyses will be documented in the ASME Design Report.
The containment arrangement is indicated in the general arrangement figures in Section 1.2.
The portion of the vessel above elevation 132'-3"is surrounded by the shield building but is exposed to ambient conditions as part of the passive cooling flow path. A flexible watertight and airtight seal is provided at elevation 132'-3" between the containment vessel and the shield building.De portion of the vessel below elevation 132'-3"is fully enclosed within the shield building.
Figure 3.8.2-1 shows the containment vessel outline, including the plate configuration and crane girder. It is a freestanding, cylindrical steel vessel with ellipsoidal upper and lower heads. The containment vessel has the following design characteristics:
Diameter: 130 feet Height: 189 feet 10 inches Design code: ASME III, Div. I Material: SA537, Class 2 Design Pressure: 45 psig Design Temperature: 280*F l Design External Pressure: 3.0 psid De wall thickness in most of the cylinder and the heads is 1.625 inches. He wall thickness of the lowest course of the cylindrical shell is increased to 1.75 inches to provide margin in the event of corrosion in the embedment transition region. The heads are ellipsoidal with a major diameter of 130 feet and a hei;;ht of 37 feet 7.5 inches.
Revision: Draft W Westiflgh00S8 3.8 1 January 4,1996
O O-
- 3. Design of Struntures, Components, Equipment, and Systems DRATT The containment vesselincludes the shell, hoop stiffeners and crane girder, equipment hatches, personnel airlocks, penetration assemblies, and miscellaneous appurtenances and attachments.
l The polar crane is designed for handling the reactor vessel head during normal refueling. The crane girder and wheel assemblies are designed to support a special trolley to be installed in the event of steam generator replacement.
The containment vessel supports most of the containment air baffle as described in subsection 3.8.4. The air baf0e is arranged to permit inspection of the exterior surface of the containment vessel. Steel plates are welded to the dome as part of the water distribution system, described in subsection 6.2.2. The polar crane system is described in subsection 9.1.5.
3.8.2.1.2 Containment Vessel Support ne bottom head is embedded in concrete, with concrete up to elevation 100' on the outside
-l and approximately elevation 108' on the inside. The containment vessel is assumed as an independent, free-standing structure above elevation 100' The thickness of the lower head is the same as that of the upper head. There is no reduction in shell thickness even though credit could be taken for the concrete encasement of the lower head.
Vertical and lateral loads on the containment vessel and intemal structures are transferred to the basemat below the vessel by friction and bearing. Seals are provided at the top of the concrete on the inside and outside of the vessel to prevent moisture between the vessel and concrete. A typical cross section design of the seal is presented in Figure 3.8.2-8, sheets 1 and 2. Furthermore, the concrete Door area and curb inside containment near Elevation 108' cnd the concrete slab outside containment at Elevation 100' slope away from the steel containment vessel to preclude water ponding adjacent to the vessel.
3.8.2.1.3 Equipment Hatches Two equipment hatches are provided. One is at the operating Door (el.135'-3"). The hatch has an inside diameter of 22 feet sized to permit replacement of a steam generator. The other is at Door elevation 107'-2" to permit grade-level access into the containment with an inside diameter of 16 feet. De hatches, shown in Figure 3.8.2 2, consist of a cylindrical sleeve with a pressure seated dished head bolted on the inside of the vessel. De containment intemal pressure acts on the convex face cf the dished head and the head is in compression. The Ganged joint has double O ring or gum-drop seals with an annular space that may be pressurized for leak testing the seals.
3.8.2.1.4 Personnel Airlocks Two personnel airlocks are provided, one located adjacent to each of the equipment hatches.
Figure 3.8.2 3 shows the typical arrangement. Each personnel airlock has about a 10 foot extemal diameter to accommodate a door opening of width 3 feet,6 inches and height 6 feet, 8 inches. The airlocks are long enough to provide a clear distance of eight feet which is not Revision: Draft January 4,1996 Y W96t10gh00S8 3.8 2
a e
- 3. Design of Structures, Components, Equipment, and Systems DRAi'T impaired by the swing of the doors within the lock. The airlocks extend radially out from the containment vessel through the shield building. They are supported by the containment vessel.
Each airlock has two double gasketed, pressure-seated doors in series. The doors are l mechanically interlocked to prevent simultaneous opening of both doors and to allow one door l to be completely closed before the second door can be opened. The interlock can be bypassed l by using special tools and procedures. !
1 3.8.2.1.5 Mechanical Penetrations l l
The mechanical penetrations consist of the fuel transfer penetration and mechanical piping penetrations and are listed in Table 6.2.3-4.
Figure 3.8.2-4, sheet I, shows typical details for the mainsteam penetration. This includes bellows to minimize piping loads applied to the containment vessel and a guardpipe to protect l the bellows and to prevent overpressurization of the containment annulus in a postulated pipe !
rupture event. Similar details are used for the feedwater penetration. I Figure 3.8.2-4, sheet 2, shows typical details for the startup feedwater penetration. This includes a guardpipe to prevent overpressurization of the containment annulus in a postulated pipe rupture event. Similar details are used for the steam generator blowdown penetration. l l
i Figure 3.8.2-4, sheet 3, shows typical details for the normal residual hear removal penetration. l Similar details are used for other penetrations below elevation 107' 2" where there is concrete inside the containment vessel. The flued head is integral with the process piping and is welded to the containment sleeve. The welds are accessible for inservice inspection. The containment sleeve is separated from the concrete by compressible material.
Figure 3.8.2-4, sheet 4 shows typical details for the other mechanical penetrations. These consist of a sleeve welded to containment with either a flued head welded to the sleeve (detail A), or with the process piping welded directly to the sleeve (detail B). Flued heads are used for stainless piping greater than 2 inches in nominal diameter and for piping with high operating temperatures.
Design requirements for the mechanical penetrations are as follows:
- Design and construction of the process piping follow ASME Section III, Subsection NC. Design and construction of the remaining portions follow ASME Code.
Section III, Subsection NE. The boundary of jurisdiction is according to ASME Code,Section III, Subsection NE.
- Penetrations are designed to maintain containment integrity under design basis accident conditions, including pressure, temperature and radiation.
- Guard pipes are designed for pipe ruptures as described in subsection 3.6.2.1.1.4.
Revision: Draft
[ W85tingh00S8 3.8 3 January 4,1996
. i
++.+.gma..= -
! 3. Design of Structures, Components Equipment, and Systems 7dy .#%I . i gJ [ l
- Bellows are stainless steel or nickel alloy and are designed to accommodate axial and l lateral displacements between the piping and the containment vessel. These l displacements include thermal growth of the mainsteam and feedwater piping during plant operation, relative seismic movements, and containment accident and testing conditions. Cover plates are provided to protect the bellows from foreign objects during construction and operation. Rese cover plates are removable to permit in-service inspection.
l The fuel transfer penetration, shown in Figure 3.8.2-4, sheet 5, is provided to transfer fuel between the containment and the fuel handling area of the auxiliary building. The fuel transfer tube is welded to the penetration sleeve. The containment boundary is a double-gasketed blind flange at the refueling canal end. The expansion bellows are not a part of the containment boundary. Rather, they are water seals during refueling operations and accommodate differential movement between the containment vessel, containment interior structures, and the auxiliary building.
3.8.2.1.6 Electrical Penetrations Figure 3.8.2-4, sheet 6, shows a typical 12 inch diameter electrical penetrations. The penetration assemblics consist' of three (or six modules in a similar 18 inch diameter penetration) passing through a bulkhead attached to the containment nozzle. Electrical design of these penetrations is described in subsection 8.3.1.1.5.
Electrical penetrations are designed to maintain containment integrity under design basis I accident conditions, including pressure, temperature, and radiation. Double barriers permit i testing of each assembly to verify that containment integrity is maintained. Design and test is according to IEEE Standard 317 83 and IEEE Standard 323-83.
3.8.2.2 Applicable Codes, Standards, and Specifications The containment vessel is designed and constructed according to the 1992 edition of the ASME Code,Section III, Subsection NE, Metal Containment. The Combined License applicant may update the Code edition and addenda as defined in subsection 5.2.1.1.
Stmetural steel nonpressure parts, such as ladders, walkways, and handrails are designed to the requirements for steel structures defined in subsection 3.8.4.
Section 1.9 discusses compliance with the Regulatory Guides and the Standard Review Plans.
3.8.2.3 Loads and Load Combinations Table 3.8.2-1 summarizes the design loads, load combinations and ASME Service Levels.
They meet the requirements of the ASME Code,Section III, Subsection NE. The containment vessel is designed for the following loads specified during construction, test, normal plant operation and shutdown, and during accident conditions:
Revision: Draft January 4,1996 [ "Liii$00S8 3.8-4
-w E
- 3. Design of Structures, Components, Equipment, and Systems y4
% ~
D Dead loads or their related internal moments and forces, including any permanent piping and equipment loads.
L Live loads or their related intemal moments and forces, including ctane loads.
Po Operating pressure loads during normal operating conditions resulting from pressure variations either inside or outside containment.
To Thermal effects and loads during normal operating conditions, based on the most critical transient or steady-state condition.
Ro Pipe reactions during normal operating conditions, based on the most critical transient or steady-state condition.
W Loads generated by the design wind on the portion of the containment vessel above elevaticn 132', as described in subsection 3.3.1.1.
E, Loads generated by the safe shutdown earthquake (SSE) as described in Section 3.7.
Wt L ads generated by the design tomado on the portion of the containment vessel above elevation 132' as described in subsection 3.3.2.
I P
t Test pressure.
P d Containment vessel design pressure that exceeds the pressure load generated by the postulated pipebreak accidents and passive cooling function.
P, Containment vessel external pressure Ta Hermal loads under thermal conditions generated by the postulated break or passive cooling function and including To. His includes variations around the shell due to the surrounding buildings and maldistribution of the passive containment cooling system water.
Ra Pipe reactions under thermal conditions generated by the postulated break, as described in Section 3.6, and including Ro.
Yr L ads generated by the reaction on the broken high<nergy pipe during the postulated break, as described in Section 3.6.
Y; Jet impingement load on a structure generated by the postulated break, as described in Section 3.6.
Ym Missile impact load on a structure generated by or during the postulated break, as from pipe whipping, as described in Section 3.6.
Revision: Draft Westingh00$8 3.8-5 January 4,1996
i l
- 3. Design of Structures, Components, Equipment, and Systems FOiU' ^W i Note that loads associated with Gooding of the containment below elevation 107' are resisted by the concrete structures and not by the containment vessel.
3.8.2.4 Design and Analysis Procedures The design and analysis procedures for the containment vessel are according to the requirements of the ASME Code,Section III Subsection NE. I l The analyses are summarized in Table 3.8.2-4. The detailed analyses will use a series of general-purpose finite element, axisymmetric shell and special purpose computer codes to I conduct such analyses. Code development, verincation, validation, configuration control and error reporting and resolution are according to the Quality Assurance requirements of Chapter ;
- 17. I 3.8.2.4.1 Analyses for Design Conditions 3.8.2.4.1.1 Axisymmetric Shell Analyses
]
The containment vessel is modelled as an axisymmetric shell and analyzed using general shell of revolution computer programs. A model used for static analyses is shown in :
l Figure 3.8.2 36. The programs calculate displacements and stresses in thin walled, elastic, l
shells of revolution when subjected to edge, surface, and/or temperature loads with arbitrary distribution over the surface of the shell. Three different types of analyses are available in these computer programs:
- 1) Static analysis can be performed for any arbitrary loading distribution. Longitudinally, concentrated loads may be applied at panel ends and distributed loads may be applied varying between specined points within each panel. The circumferential distribution is obtained through the use of Fourier Series.
- 2) Natural frequencies, mode shapes (displacements and forces), and participation factors for any loading that can be handled statically can be calculated and output to a Die for use in either a spectral analysis, when a response spectrum is available, or in a modal superposition analysis, when transient forcing functions of the form g(s,0).f(t) are available. The program can handle extra concentrated and distributed mass acting in any or all directions plus Guid structure interaction.
- 3) Direct time integration analysis for general transient problems can be calculated where the forcing function cannot readily be separated into separate spatial and temporal functions. As in natural frequency analysis, additional concentrated and distributed masses acting in any and all directions can be applied. Pressures can vary in an arbitrary fashion versus time and damping may be included.
Revision: Draft January 4,1996 T %%iiiiiiouse 3.8-6
- 3. Design of Structures, Components, Equipment, and Systems
.w . i- u l Dynamic analyses of the axisymmetric model, which is si.nitar to that shown in l Figure 3.8.2-6, are performed to obtain frequencies and mode shapes. These are used to confirm the adequacy of the containment vessel stick model as described io subsection 3.7.2.3.2. Static stress analyses are performed for each of the following loads.
- Dead load
- Internal pressure
- Equivalent static seismic accelerations
- Polar crane wheel loads
- Wind loads
- Thermal loads The equivalent static accelerations applied in the seismic analysis are the maximum acceleration responses based on the envelope of the results for each soil case. These accelerations are applied as separate load cases in the east-west, north-south and vertical directions. The torsional moments, which include the effects of the eccentric masses, are increased to account for accidental torsion and are evaluated in a separate calculation.
The results of these load cases are factored and combined in accordance with the load combinations identified in Table 3.8.2-1. These results are used to evaluate the general shell away from local penetrations and attachments, i.e. for areas of the shell represented by the axisymmetric geometry. The results for the polar crane wheel loads are also used to establish local shell stiffnesses for inclusion in the containment vessel stick model described in subsection 3.7.2.3. The results of the analyses and evaluations are included in the containment vessel design report.
Design of the containment shellis primarily controlled by the intemal pressure of 45 psig. The meridional and circumferential stresses for the internal pressure case are shown in Figure 3.8.2-5. The most highly stressed regions for this load case are the portions of the shell away from the hoop stiffeners and the knuckle region of the top head. In these regions the stress intensity is close to the allowable for the design condition.
Major loads that induce compressive stresses in the containment vessel are intemal and external pressure and crane and seismic loads. Each of these loads and the evaluation of the compressive stresses are discussed below.
- Intemal pressure causes compressive stresses in the knuckle region of the top head and in the equipment hatch covers. The evaluation methods are similar to those discussed in subsection 3.8.2.4.2 for the ultimate capacity.
- Evaluation of extemal pressure loads is performed in accordance with ASME Code,Section III, Subsection NE, Paragraph NE 3133.
- Crane wheel loads due to crane dead load, live load, and seismic loads result in local compressive stresses in the vicinity of the crane girder. These are evaluated in
! accordance with ASME Code Case N-284, Revision 1.
Revision: Draft 3 W8Sthg!100S8 3.8 7 January 4,1996
~
g
- 3. oes,,n er st,.<t.,es. cem,enents, e,u>, ..t. .nd s,,t.ms
- DRAF
. Overall seismic loads result in axial compression and tangential shear stresses at the base of the cylindrical portion. These are evaluated in accordance with ASME Code Case N 284, Revision 1.
The bottom head is embedded in the concrete base at elevation 100 feet. This leads to circumferential compressive stresses at the discontinuity under thermal loading associated with the design basis accident. The containment vessel design includes a Service Level A combination in which the vessel above elevation 100' is conservatively specified at the design I
temperature of 280*F and the portion of the embedded vessel (and concrete) is specified at a temperature of 70*F. Containment shell buckling close to the base is evaluated against the l l criteria of ASME Code Case N-284, Revision 1, using a BOSOR-5 model of the portion of l the shell above elevation 100' extending up to the horizontal stiffener at elevation 132' 3" l Material yield and stiffness properties are based on properties at the design temperature of 280*F. Temperature differences are raised by small increments until buckling is predicted. l Buckling occurred 20 inches above elevation 100' for a circumferential wave number, N = l 190, at a factor of 6.0 times the design differential temperature condition. The half buckling ;
wave length is less than 0.5 V(rt). Thi.s is not a significant buckling issue; buckling did not occur for wave numbers below N = 60, which is the critical range for the cylinder and top head under external and internal pressure.
3.8.2.4.1.2 Local Analyses The penetrations and penetration reinforcements are designed in accordance with the rules of ASME III, Subsection NE. He dynamic response of the local concentrated tratss is considered in local analyses of the shell and is included in the design.
Finite element analyses are performed to confirm that the design of the penetration in accordance with the AShG code results in acceptable buckling safety factors. A finite
! element ANSYS model, as shown in Figure 3.8.2-7, represents the portion of the vessel close to the embedment with the lower equipment hatch and personnel airlock. This is analyzed I for external pressure and axial loads and demonstrates that the penetration reinforcement is sufficient and that the lowest buckling mode occurs in the shell away from the penetrations and embedment.
3.8.2.4.2 Evaluation of Ultimate Capacity The capacity of the containment vessel has been calculated for internal pressure loads for use I in the probabilistic risk assessment analyses and severe accident evaluations. Each element of the containment vessel boundary was evaluated to estimate the maximum pressure at an ambient temperature of 100 F corresponding to the following stress and buckling criteria:
- Deterministic severe accident pressure capacity corresponding to ASME Service Level C limits on stress intensity, Code Case N 284 for buckling of the equipment hatch covers, and sixty percent of critical buckling for the top head.Re deterministic severe accident pressure capacity corresponds to the approach in SECY 93-087, to maintain a reliable Revision: Draft January 4,1996 y L iiigtiouse 3.8-8
- 3. Design of Structures, Components, Equipment, and Systems I' A C I L ,; h s% J .
leak-tight bamer approximately 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> following the onset of core damage under the more likely severe accident challenges. This approach was approved by the Nuclear Regulatory Commission as outline in the Staff Requirements Memorandum on SECY-93-087 - Policy, Technical, and Licensing issues Pertaining to Evolutionary and l Advanced Light Water Reactor (ALWR) Designs, Dated July 21,1993. l 1
- Best estimate capacity corresponding to gross membrane yield at the ASME-specified minimum yield stress (SA537 Class 2, yield stress = 60 ksi, ultimate stress = 80 ksi),
and critical buckling for the equipment hatch covers and top head.
l The results are shown in Table 3.8.2-2. He analyses at a temperature of 100 F are described in the following paragraphs for each element. The critical regions identified in this table are !
then examined further for their response at higher temperature. This results in the best-l estimate capacity based on the ASME-specified minimum yield properties. He evaluation l considered the containment boundary elements including:
- Cylindrical shell ,
- Top and bottom heads l
- Equipment hatches and covers I
- Personnel airlocks
- Mechanical and electrical penetrations l
The evaluation identified the most likely failure mode to be that associated with gross yield of the cylindrical shell. Loss of containment function would be expected to occur because the large post-yield deflections would lead to local failures at penetrations, bellows, or other local discontinuities.
3.8.2.4.2.1 Tensile Stress Evaluation of Shell Results of the axisymmetric analyses of the cylinder and top head described in subsection 3.8.2.4.1 for dead load and intemal pressure were evaluated to determine the l pressure at which stresses reach yield at ambient temperature of 100 F. The analyses assume the shell is fixed at elevation 100', where the bottom head is embedded in concrete. He steel bottom head is identical to the top head and has a pressure capability greater than the top head due to the additional strength of the embedment concrete.
De allowable stress intensity under Service Level C loads is equal to yield. This corresponds to an intemal pressure of 125 psig. The critical section is the cylinder, where the general primary membrane stress intensity is greatest.
The best-estimate yield analysis uses the von Mises criterion to establish yield rather than the more conservative ASME stress intensity approach. This increases the yield stress by about 15 percent for the cylinder, where the longitudinal stress is equal to one-half of the hoop stress resulting in first yield at an intemal pressure of 144 psig. At this pressure, hoop stresses in the cylinder reach yield. The radial deflection is about 1.6 inches. As pressure increases further, large deflections occur. For a material such as S A537, where the yield plateau extends Revision: Draft
[ W8Stingh0054 3.8 9 January 4,1996
- 3. Design of Structures, Components, Equipment, and Systems k ., s # D 9 from a strain of 0.2% to 0.6% deflections would increase to 4.8 inches at yield without a substantial increase in pressure. Strain hardening would then permit a further increase in
, pressure with large radial deflections, as described in subsection 3.8.2.4.2.6.
3.8.2.4.2.2 Buckling Evaluation of Top Head The top head has a radius-to-height ratio of 1.728. This is not as shallow as most ellipsoidal or torispherical heads, which typically have a radius-to-height ratio of 2. The ratio was specifically selected to minimize the local stresses and buckling in the knuckle region due to ;
internal pressure. As the ratio decreases, the magnitude of compressive stresses in the knuckle region decreases; for a radius-to-height ratio of 1.4 or smaller, there are no compressive stresses and therefore there is no potential for buckling.
Theoretical Buckling Capacity The top head was analyzed using the-BOSOR-5 computer code (Reference 1). This code permits consideration of both large displacements and nonlinear material properties. It calculates shell stresses and checks stability at each load step. Yield of the cylinder started at a pressure of 144 psig using elastic - perfectly plastic material properties, a yield stress of 60 ksi, and the von Mises yield criterion. Yield of the top of the crown started at an intemal pressure of 146 psig. Yield of the knuckle region started at 152 psig. A theoretical plastic buckling pressure of 174 psig was determined. At this pressure, the maximum effective prebuckling strain was 0.23 percent in the knuckle region where buckling occurred and 2.5 percent at the crown. He maximum deflection at the crown was 15.9 inches. A similar analysis was performed using non linear material properties considering the effects of residual stresses; buckling did not occur in this analysis, and failure would occur once strains at the crown reach ultimate. ,
Predicted Pressure Capacity The actual buckling capacity may be lower than the theoretical buckling capacity because of effects not included in the analysis such as imperfections and residual stresses. This is considered by the use of capacity reduction factors that are based upon a correlation of theory and experiment. De capacity reduction factor for the top head was evaluated based on comparisons of BOSOR-5 analyses against test results of ellipsoidal and torispherical heads.
His evaluation is described below and concludes that no reduction in capacity need be considered; i.e. a capacity reduction factor of 1.0 is appropriate.
The knuckle region of ellipsoidal and torispherical heads is subjected to meridional tension and circumferential compression. Re meridional tension tends to stabilize the knuckle region and reduces its sensitivity to imperfection. De radius-to height ratio of 1.728 of the AP600 head results in a larger ratio of meridional tension to circumferential compression than on shallower heads, further reducing the sensitivity to imperfection.
Revision: Draft January 4,1996 y Westfrighouse 3.8 10
l
- 3. Design of Structures Components, Equipment, and Systems l
1
,, . . .i -a E Welding Research Council Bulletin 267 (Reference 22) shows a comparison of BOSOR-5 predictions of buckling against the results of 20 tests of small head models. Dese results are summarized in Table 4 of the reference and show ratios (capacity reduction factors) of actual buckling to the BOSOR 5 prediction with an average of 1.2. Only one of the 20 cases shows a capacity reduction factor less than 1.0.
Table 3.8.2 3 shows the key parameters, test results, and BOSOR 5 predictions for two large, fabricated 2:1 torispherical heads tested and reported in NUREG/CR-4926 (Reference 23). The theoretical plastic buckling pressure predicted by BOSOR-5 represents initial buckling based on actual material properties. The initial buckling did not cause failure for either of the tests, and test pressure continued to increase until rupture occurred in the spherical cap. He collapse pressures were three to four times the initial buckling pressures.
- Test Head 1 - The test result of 58 psig is 79 percent of the predicted theoretical plastic buckling pressure of 74 psig. Many of the buckles occurred directly on the meridional weld seams of the knucide. The knuckle welds were noticeably flatter than the corresponding welds of the Test 2 head. The as-built configuration extended inside the theoretical shape at some of the meridional weld seams and was most pronounced at the location of the first observed buckle. Model I exceeded the tolerances for formed heads specified for containment vessels in NE-4222.2 of ASME Section III, Subsection NE.
- Test Head 2 - ne test result of 1% psi is 100 percent of the BOSOR-5 predic.ted theoretical plastic buckling pressure. For test head 2, the welds had no noticeable flat spots and there was a smooth transition between the sphere and knuckle sections. Test head 2 was well within the Code allowable deviations.
The low-capacity reduction factor of 0.79 for test head I is attributed to excessive imperfections associated with the fabrication of relatively thin plate (0.196 inch). These imperfections were visible and were outside the tolerances permitted by the ASME Code.The results of test head I are therefore not considered applicable to the AP600. The results of test head 2 and of the small-scale models described in the Welding Research Council Bulletin support the application of a capacity reduction factor of 1.0.
The capacity of the AP600 head was also investigated using an approach similar to that permitted in ASME Code Case N284. His code case provides attemate rules for certain containment vessel geometries such as cylindrical shells. The theoretical clastic buckling pressure was calculated to be 536 psi using the linear clastic computer code, BOSOR-4 (Reference 24). A reduction factor (defined as the product of the capacity reduction factor and the plastic reduction factor) was established as 0.385 based on the lower bound curve of test results of 20 ellipsoidal and 28 torispherical tests specimens, which also include the two large i fabncated heads previously discussed. His resulted in a predicted buckling capacity of 205 psig.
He preceding paragraphs addressed incipient buckling. It is concluded that buckling would not occur prior to reaching the pressure of 174 psig predicted in the BOSOR 5 analyses. Tests indicate that pressure can be significantly increased prior to rupture after the formation of the l
l Revision: Draft !
Westinghouse 3.8. I I january 4,1996 i
}
- 3. Design of Structures, Components, Equipment, and Systems 99fb n 4 WI initial buckles. - The best estimate capacity of the head is taken as the theoretical plastic buckling pressure of 174 psig predicted in the BOSOR-5 analyses l The Deterministic severe accident pressure capacity is taken as sixty percent of critical buckling. This is consistent with the safety factor for Service Level C in ASME Code Case j N 284 and results in a containment head capacity of 104 psig.
l 3.8.2.4.23 Equipment Hatches l l
The equipment hatch covers were evaluated for buckling according to ASME Code Case !
N-284. The containment internal pressure acts on the convex face of the dished head and the i hatch covers are in compression under containment intemal pressure loads. The critical l buckling capacity is based on classical buckling capacities reduced by capacity reduction l factors to account for the effects of imperfections and plasticity. Rese capacity reduction i factors are based on test data and are generally lower-bound values for the tolerances specified in the ASME code.
The critical buckling pressures are 195 psig for the 22 foot diameter hatch and 160 psig for the 16 foot diameter hatch at ambient temperature of 100 F. For the Service Level C limits l a safety factor of 1.67 is specified, resulting in capabilities of 117 psig (22' dia) and 96 psig !
I (16' dia).
Typical gaskets have been tested for severe accident conditions as described in NUREG/CR-5096 (Reference 25). The gaskets for the AP600 will be similar to those tested l with material such as Presray EPDM E 603. For such gaskets the onset of leakage occurred (
at a temperature of about 600 F.
3.8.2.4.2.4 Personnel Airlocks The capacity of the personnel airlocks was determined by comparing the airlock design to that l tested and reported in NUREG/CR 5118 (Reference 3). Critical parameters are the same, so the results of the test apply directly. In the tests the inner door and end bulkhead of the airlock withstood a maximum pressure of 300 psig at 400*F. De capacity of the airlock is therefore at least 300 psig at ambient temperature. The maximum pressure corresponding to Service Level C is conservatively estimated by reducing this capacity in the ratio of the minimum specified material yield to ultimate.
1 3.8.2.4.2.5 Mechanical and Electrical Penetrations l Subsections 3.8.2.1.3 through 3.8.2.1.6 describe the containment penetrations. Penetration reinforcement is designed following the area replacement method of the ASME Code. He insert plates and sleeves permit development of the hoop tensile yield stresses predicted as the limiting capacity in subsection 3.8.2.4.1. Capacities of the equipment hatch covers are discussed in subsection 3.8.2.4.2.3 and of the personnel airlocks in subsection 3.8.2.4.2.4.
Revision: Draft January 4,1996 y Westinghouse 3.8-12
~
- 3. Design of Structures, Components, Equipment, and Systems LR::i DRAFT Mechanical penetrations welded directly to the containment vessel are generally piping systems with design pressures greater than that of the containment vessel. Thicknesses of the Oued head or end plate are established based on piping support loads or stiffness requirements.
The capacities of these penetrations are greater than the capacity of the containment vessel cylinder.
Mechanical penetrations for the large-diameter high-energy lines. such as the main steam and ;
feedwater piping, include expansion bellows. The piping and flued head have large pressure l capability. The response of expansion bellows to severe pressure and deformations is I described in NUREG/CR-5561 (Reference 4). The bellows can withstand large pressure loading but may tear once the containment vessel deflection becomes large. Testing reported in NUREG/CR-6154 (Reference 26) has shown that the bellows remain leaktight even when i subjected to large denections sufficient to fully compress the bellows. Such large deGections j do not occur so long as the containment vessel remains elastic. As described in subsection 3.8.2.4.2.6, the radial deflection of the shell increases substantially once the containment cylinder yields. The resulting deflections are assumed to cause loss of containment function.
The containment penetration bellows are designed for a pressure of 90 psig at design temperature within Service Level C limits, concurrent with the relative displacements imposed on the bellows when the containment vessel is pressurized to these magnitudes.
Electrical penetrations have a pressure boundary consisting of the sleeve and an end plate containing a series of modules. He pressure capacity of these elements is large and is greater than the capacity of the containment vessel cylinder at temperatures up to the containment l design temperature. Electrical penetration assemblies are also designed to satisfy ASME Service Level C stress limits under a pressure of 90 psig at design temperature. Tests at pressures and temperatures representative of severe accident conditions are described in NUREG/CR 5334 (Reference 5), where the Westinghouse penetrations were irradiated, aged, then tested to 75 psia at 400*F. Other electrical penetration assemblies were tested to higher pressures and temperatures. Rese tests showed that the electrical penetration assemblies withstand severe accident conditions. The electrical penetration assemblies are qualified for the containment design basis event conditions as described in Appendix 3D. The assemblies are similar to one of those tested by Sandia as reported in NUREG/CR 5334 (Reference 5).
l The ultimate pressure capacity of the electrical penetration assemblies is primarily determined by the temperature. The maximum temperature of the containment vessel below the operating deck during a severe accident is approximately equal to the containment design temperature of 280*F. His temperature is significantly below the temperature at which the assemblies from the three suppliers in the Sandia tests were tested.
3.8.2.4.2.6 Material Properties The containment vessel is designed using SA537 Class 2 material. His has a specified minimum yield of 60 ksi and ultimate of 80 ksi. Test data for materials treeting SA537 or having similar chemical properties were reviewed. In a sample of 122 tests for thicknesses equaling or exceeding 1.50 inches and less than 1.75 inches, the actual yield had a mean value of 69.1 ksi with a standard deviation of 3.3 ksi. Thus, the actual yield is expected to be about Revision: Draft nW W65tingh00$8 3.8-13 January 4,1996
i
- 3. Design of Structures, Components, Equipment. and Systems l
DRAFT 15 percent higher than the minimum yield. Membrane yield of the cylinder is predicted to occur at an intemal pressure of 166 psig. j A stress strain curve for material with chemistry similar to SA537, Class 2, indicated constant yield stress of 81.3 ksi from a strain of 0.002 to 0.006 followed by strain hardening up to a maximum stress of 94.5 ksi at a strain of 0.079. The first portion of the strain hardening is nearly linear, with a stress of 90 ksi at a strain of 4 percent. This strain occurs at a stress 10 percent above yield. Thus, a pressure load 10 percent higher than that corresponding to yield of the shell would' result in 4 percent strain and a 31 inch radial deflection of the containment cylinder. Such a deflection is expected to cause major distress for penetrations, the air flow path, and local areas where other structures are close to the containment vessel.
Loss of function is therefore assumed for the containment once gross yield of the containment cylinder occurs.
3.8.2.4.2.7 Effect of Temperature The evaluations described in the preceding subsections are based on an ambient temperature of 100'F. Nonmetallic items, such as gaskets, are qualified to function at the design temperature. He capacity of steel elements is reduced in proportion to the reduction due to temperature in yield stress, ultimate stress, or elastic modulus. He cylinder is govemed by ;
d yield stress, and elastic buckling of the hatch covers is governed by the elastic modulus. He reduction in capacity is estimated using the tables given for material properties in the ASME Code. At 400 F, the yield stress is reduced by 17 percent and the pressure capacity corresponding to gross yield is reduced from 144 to 120 psig. l l
3.8.2.4.2.8 Summary of Containment Pressure Capacity The ultimate pressure capacity for containment function is expected to be associated with leakage caused by excessive radial deflection of the containment cylindrical shell. This radial deflection causes distress to the mechanical penetrations, and leakage would be expected at the expansion bellows for the main steam and feedwater piping. There is high confidence that this failure would not occur before stresses in the shell reach the minimum specified material yield. This is calculated to occur at a pressure of 144 psig at ambient temperature and j 120 psig at 400 F. Failure would be more likely to occur at a pressure about 15 percent higher l based on expected actual material properties. l l
De deterministic severe accident pressure that can be accommodated according to the ASME l Service Level C stress intensity limits and using a factor of safety of 1.67 for buckling of the i top head is determined by the capacity of the 16 foot diameter equipment hatch cover and the !
ellipsoidal head. He maximum capacity of the hatch cover, calculated according to ASME l Code Case N 284 Service Level C, is % psig at ambient temperature of 100*F and 93 psig at 280 F. He maximum capacity of the ellipsoidal head is 104 psig at 100*F and 92 psig at 280*F.
Revision: Draft January 4,1996 3 WOStingh0088 3.8 14
m=
- 3. Design of Structures, Components, Equipment, and Systems J
- 7. .. - ;
$ , L. J 3.8.2.5 Structural Criteria The containment vessel is designed, fabricated, installed, and tested according to the ash 1E Code,Section III. Subsection NE, and will receive a code stamp.
Stress intensity limits are according to ash 1E Code,Section III Paragraph NE-3221 and Table NE-3221-1. Critical buckling stresses are checked according to the provisions of ASNIE Code,Section III. Paragraph NE-3222, or ASNIE Code Case N 284.
3.8.2.6 Staterials, Quality Control, and Special Construction Techniques 5faterials for the containment vessel, including the equipment hatches, personnel locks, penetrations, attachments, and appurtenances meet the requirements of NE-2000 of the AShiE Code. The basic containment material is SA537, Class 2, plate. This material has been selected to satisfy the lowest service metal temperature requirement. This temperature is established for the portion of the vessel exposed to the environment as the minimum ambient air temperature which is site specific. Temperatures as low as -40 F (see Table 2.0-1) are acceptable for S A537, Class 2 material. Impact test requirements are as specified in NE 2000.
The containment vessel is coated with an inorganic zinc coating, except for those portions fully embedded in concrete.The inside of the vessel below the operating floor and up to eight feet above the operating floor also has a phenolic top coat. Below elevation 100' the vessel is fully embedded in concrete with the exception of the few penetrations at low elevations l (see Figure 3.8.2-4, sheet 3 of 6 for typical details). Embedding the steel vessel in concrete protects the steel from corrosion.
The AP600 configuration is shown in the general arrangement figures in Section 1.2 and in Figure 3.8.2-1. The exterior of the vessel is embedded at elevation 100' and concrete is placed against the inside of the vessel up to elevation 108'-2" Above this elevation the inside and outside of the containment vessel are accessible for inspection of the coating. The vessel is coated with an inorganic zine primer to a level just below the concrete. Seals are provided at I the surface of the concrete inside and outside the vessel so that moisture is not trapped next to the steel vesseljust below the top of concrete. The seal on the inside accommodates radial growth of the vessel due to pressurization and heatup.
The plate thickness for the first course (elevation 104't.5" to 116'10") of the cylinder is 1.75 inches, which is cne eighth of an inch thicker than the rest of the vessel. His provides margin in the event that there would be any corrosion in the transition region despite the coatings and seals described above. Equivalent margin is available for the 1.625 inch thick bottom head in the transition region ( elevation 100'-104't.5"). The plate thickness for the head is a constant thickness and is established by the stresses in the knuckle. As a result, the pressure stresses in the transition zone are well below the allowable stress providing margin in the event of corrosion in this region.
Revision: Draft 3.8-15 January 4,1996
% W95tingh00S8
- 3. Design of Structures, Components. Equipment, and Systems 1
The quality control program . involving welding procedures, erection tolerances, and nondestructive examination of shop- and field- fabricated welds conforms with Subsections ,
NE-4000 and NE.5000 of the ASME Code. l l
The containment vesselis designed to permit its constmetion using large subassemblies. These subassemblies consist of the two heads and three ring sections. Each ring section comprises three courses of plates and is approximately 38 feet high. These are assembled in an area near l the final location, using plates fabricated in a shop facility. l 3.8.2.7 Testing and Inservice Inspection Requirements l Testing of the containment vessel and the pipe assemblies forming the pressure boundary within the containment vessel will be according to the provisions of NE-6000 and NC-6000 respectively.
Subsection 6.2.5 describes leak rate testing of the containment system including the containment vessel.
Inservice inspection of the containment vessel will be performed according to the ASME Code Section XI. Subsection IWE, and will be described in the Combined License application.
Revision: Draft January 4 IM $@
~
- 3. Design of Structures, Components, Equipment, and Systems 3.8.7 References
- 1. D.Bushnell, "BOSORS - Program for buckling of elastic-placic complex shells of revolution including large deflections and creep," Computers and Structures, Vol.6 pp 221-239, 1976
- 2. C.D Miller, R.B. Grove, J.G.Bennett, " Pressure Testing of Large Scale Torispherical Heads Subject to Knuckle Buckling," NUREG/CP-0065. Transactions of the International Conference on Structure Mechanics in Reactor Technology August 1985.
- 3. " Leak and Structural Test of Personnel Airlock for LWR Containments Subjected to l Pressures and Temperatures beyond Design Limits," NUREG/CR-5118, S AND88-7155, l May 1989.
- 4. " Analysis of Bellows Expansion Joints in the Sequoyah Containment," NUREG/CR-5561, SAND 90-7020, December 1991. l l
- 5. " Severe Accident Testing of Electrical Penetration Assemblies," NUREG/CR-5334, SAND 89-0327, November 1989.
l
- 6. ASTM C 94-1990, " Specifications for Ready Mixed Concrete." I
- 8. ASTM C 150-1989, " Specification for Portland Cement."
- 9. ASTM C 33-1990," Specification for Concrete Aggregates."
- 10. ASTM C 131-1989, " Resistance to Abrasion of Small Size Coarse Aggregate by Use of the Los Angeles Machine."
- 11. ASTM C 535-1989, " Test Method for Resistance to Degradation of Large-Size Coarse Aggregate by Abrasion and Impact in the Los Angeles Machine.
- 13. ASTM D 1888-1978, " Particulate and Dissolved Matter in Industrial Water."
- 14. ASTM C 618-1989, " Fly Ash and Raw or Calcined Natural Pozzolans for Use in Portland Cement Concrete."
- 15. ASTM C 311-1990, " Sampling and Testing Fly Ash or Natural Pozzolans for Use as Mineral Admixture in Portland Cement Concrete."
- 16. ASTM C 260-1986, " Air Entraining Admixtures for Concrete."
Revision: Draft January 4,1996 3 W95tingh00S8 3.8-46
4 n.= :q
- 3. Design of Structures, Components Equipment, and Systems .
l L ,s(N. h
- 17. ASTM C 494-1986, " Chemical Admixtures for Concrete "
- 18. ACI 211.1-1989, " Standard Practice for Selecting Proportions for Normal, Heavy Weight, and Mass Concrete."
- 19. ASTM A 615-1990, " Deformed and Plain Billet Steel Bars for Concrete Reinforcement." l l
- 20. ASTM A 706-1990, " Low Alloy Steel Deformed Bars for Concrete Reinforcement."
- 21. ANSYS, Version 4.4. A135 for SILICONGRAPHICS 4D and INDIGO.
- 22. D. Bushnell, " Elastic-Plastic Buckling of Intemally Pressurized Ellipsoidal Pressure i Vessel Heads," Welding Research Council Bulletin 267, May 1981. I
- 23. J. G. Bennett, "An Assessment of Loss-of-Containment Potential Because of Knuckle Buckling for 4:1 Steel Containment Heads," NUREG/CR-4926, LA-10972 MS, April 1987 ( 'Ihis Report also includes the Contractor's Report: " Buckling and l Rupture Tests of Two Fabricated Torispherical Heads under Inte nal Pressure," by R. l B. Grove, S. W. Peters, C. D. Miller, and M. F. Eder, C.B.I. Industries, Inc. Research j Laboratory.)
- 24. D. Bushnell, " Stress, Stability, and Vibration of Complex Branched Shells of Revolution: Analysis and User's Manual for BOSOR 4," LMSC-D243605, Lockheed Missiles and Space, Palo Alto, Cal.,1972.
- 25. " Evaluation of Seals for Mechanical Penetrations of Containment Buildings",
NUREG/CR 5096 SAND 88-7016.
- 26. " Experimental Results from Containment Piping Bellows Subjected to Severe Accident Conditions", NUREG/CR-6154, SAND 94-1711, Vol.1, September,1994 Revision: Draft Wm;tifWh00Sa 3.8-47 January 4,1996
- -E
- 3. Design of Structures. Comgwnents, Eqvipment, and Systems
- ., , u . . ' -
Table 3.8.2-1 LOAD COMBINATIONS AND SERVICE LIMITS FOR CONTAINMENT VESSEL Load Description Load Combination and Service Limit
! Con Test Des. Des. A A A C C C D l NM D x x x x x x x x x x x l Live L x. x x x x x x x x x x
! Wind W x x l l
SSE E3 x x x '
Tornado W' t x Test pressure Pt X Test temperature Tt X Operating pressure Po x x x Design pressure Pd X X X X l External pressure (3.0 psid) Pe x x x l l
Normal reaction Ro x x x x x Normal thermal To x x x x x Accident thermal reactions Ra X X X X Accident thermal Ta X X X X Accident pipe reactions x Yr Jet impingement Yj x Pipe impact Ym x Notes:
- 1. Service limit levels are per ASME-NE.
- 2. Where any load reduces the effects of other loads, that load shall be taken as zero, unless it can be demonstrated that the load is always present or occurs simultaneously with the other loads.
Revision: Draft January 4,1996 ] Westinghouse 3.8-48
E
- 3. Design of Structures, Components, Equipment, and Systems lT. I.T h I.I PI
' ~ ~'
Table 3.8.2 2 CONTAINMENT VESSEL PRESSURE CAPABILITIES Containment Element Pressure Capability Deterministic Severe Accident Maximum Pressure Capacity (I) Capability (2) l Temperature 100*f- 280*F 400*F 100*F 400*F
! Cylinder 125 psig i10 psig IM psig 144 psig 120 psig l Ellipsoidal Head IN psig 92 psig 87 psig 174 psig 145 psig l 22 foot equipment hatch 117 psig i14 psig 110 psig 195 psig 184 psig l 16 foot equipment hatch 96 psig 93 psig 90 psig 160 psig 151 psig l Personnel airlocks(3) >l63 psig >l63 psig >l63 psig >300 psig >300 psig Note:
- 1. The buckling capacity of the ellipsoidal head is taken as sixty percent of the critical buckling pressure calculated by the BOSOR-5 non linear analyses; the buckling capacity at higher temperatures is calculated by reducing the capacity at 100*F by the ratio of yield at 100*F to yield at the higher temperature. Evaluations of the other elements are according to ASME Service Level C and include use of Code Case N284.
- 2. The estimated maximum pressure capability is based on minimum specified material properties. ,
- 3. He capacities of the personnel airlock are estimated from test results. l l
l l
Revision: Draft 3.8 49 January 4,1996 y Westinghougg
- 3. Design of Stnictures, Components Equipment, and Systems C22
!i. . ..
Table 3.8.2 3 l ANALYSIS AND TEST RESULTS OF FABRICATED HEADS (Reference 23)
Test Model #1 Test Model #2 Cylinder radius 96.0 inches 96.0 inches Knuckle radius 32.64 inches 32.64 inches Spherical radius 172.8 inches 172.8 inches Thickness 0.196 inches 0.27 inches Head height / radius 0.5 0.5 Radius / thickness 490 356 Test initial buckling pressure 58 psig 1% psig Test collapse pressure 229 psig 332 psig I
Collapse pressure / initial buckling pressure 3.95 3.13 BOSOR 5 predicted buckling pressure 73.6 psig 106.6 psig ;
I l
l l
l l
l l
Revision: Draft January 4,1996 sum [ h ihiliOUS8
l l
l 4 ,..
g "
- 3. Design of Structures, Components Equipment, and Systems -
l l
1 mm
- y
. .. 4 Table 3.8.2 4 SU5151ARY OF CONTAIN5 TENT VESSEL 510DELS AND ANALYSIS SIETilODS Analysis 5todel Stethod Program Purpose Axisymmetric shell hfodal analysis CBI / To calculate frequencies and mode shapes for l Kalnins comparison against stick model Lumped mass stick Modal analysis CBI / To create equivalent stick model for use in model SAP nuclear island seismic analyses Axisymmetric shell Static analyses using CBI / To calculate containment vessel shell stresses l Fourier harmonic Kalnins loads Axisymmetric shell Non-linear BOSORS To calculate buckling capacity close to base bifurcation under thermal loads.
To calculate pressure capacity of top head.
Finite element shell Linear bifurcation ANSYS To study local effect of large penetrations and 5.1 embedment on buckling capacity for axial and external pressure loads Finite element shell Modal analysis ANSYS To calculate frequencies and mode shapes for 5.1 local effects of equipment hatches and personnel airlocks Finite element shell Static analyses ANSYS To calculate local shell stress in vicinity of the 5.1 equipment hatches and personnel airlocks Revision: Draft
% W95tingh0llS8 3.8 51 January 4,1996
- 1 OF STRUCTURES, COMPONENT 3. EQUIPMENT, AND $YSTEMS u1
>: 01/13/94 .
DRAFT 6 PIECE E xpatt(0 CWI 24 PLATES AAQA0 ELLIP CAL 4 A0 T
\
M[ihD\\ !l }
~
32PLATESA90tec j l
l.
l .
tu t:-(
! it. 218 ' 8'/r EL 201' 0
'b f0P 7 RAIL l
I CRA4. 1lC LR l
l
'. 130' 0 1.0.
(12 PLATES ARal m i EL 110' 0 I, ,,
SilFFFO
$\
El14*4 .
O f, , ,,w isi EL 04 7 ni il 135' 3 OP. F.00R t i 1 I g( g, . 9 4 g . 0 1.0.
g i3.lP(Mi NA'CH k 1 16 3.3
()*10!.D.M I PERSML LOCK g ,
TE0llF4Nf4fCH EL 110' 6l h \
ngi II Et 107' 2 7. FLOT '.
ill s,
'n '. MC wg s T l I
(L 04 ' /r
- i. \. ;t ,3o. 3 I
I 32 h4'E5 All0M
\ i i _
EaiP50lCAL *( AD i4g I$
% \lU f/ l
/u ti 6 _
24 P.ATES ARQAQ 6 PIECE EXPAPOE0 W WATER'4Li 5453/ %2 Figure 3.8.21 (Sheet 1 off)f"it I Westinghouse
@p2Om0 t7ced @czmI3 Outtine
2
- 3. DESIGN OF STRUCTURES, COMPONENTS, EQUIPMENT, AND SYSTEMS Revision: 1 Effective: 01/13/94 WESPNGHOUSE 2,0 DRAFT ,
t I
32 PLATES AR3UNO l
24 FLATES AROUM)
Y igg. _ , _ . _ . . _ . _._._._._._, _._.
0'
_, I __
l I
! \
i i
1!
1 l -
90' TOP ANC BCITCH llEAC L AYCLI l
Figure 3.8.21 (Sheet 2 ofk l
Containment Vessel General Outline Westinghousa -
OPGETARY CLASS 2 51
- 2,0 DRAFT I
I I !
l 1
l I
I I i l 1
I.
.l I l 1
'I.
I l
l.
l l80* - -*- - -*- -
m ----
- 0'
- - --} --- --- - - -~~ --
l.
1 4, !
I l.
l I
I7* -
% -._ J \
i \
- 6 -C l .0.
NM+d i ' A'w* -
1
/ < ?, ' ,, -g,
. ';*S T MATCH , l
. rec.
8 / 30
& i 1 22 1 !.0. %
EQUIPMENT HAICH Q L
s
' ( (2 ) 9 ' 10 1.0.
PERSONNEL LOCK ORIENTAT10N OF MAJCR PENETRAT'.ONS 3 #'t Rgure 3.8.2 1 (Sheet 2'of A)
Containment Vessel General Outline M
3, DESIGN OF STRUCTURES, COMFONENTS, EQUIPMENT, AND SY$TEMS Re d Effecttve: 06/26/92 r)
(
6 12 ITPICAL SPECAlt
. OF CRANE RAIL l __ _. _____ _ _'
_ _ CL ATS Ai ( RAll WELD l'/2 RAD TYP l ! I I I I
._p_. ._... ._ _
_.p. _._ -._. .
p_. ._.. . . _. . _
' j . .
I l l l 1 I
/_ WELD b W b k ) k SECTION m W
l u
i I
( ,
f' '
I I
oS l 9 El 218 ' 8% I AN L'NE E bg I j ( CRANE RAIL bf l El 209 ' -0 TOP OF RA!L [--=-
7- ... . ..
A-l 62'-0 E_, l i 24*8 ACCESS OPENINC g
5 I 65' 0 INS RAD
- a. =
-' $ l
= -, .
I o a .
a 1 D I I l =
I l -+
l 8
...... . _ . . _ .El 135.'. 3. . _OP FL.O.O.R. . . . .. . . . . _ .
62' 010 VESSEL (
f ELEVA110N l
- s e W-Westinghouse e# 'M m__ ____ _ _ _ _ _ _ _
DRAFT M i
( RADIAL WEB t 1801 REG)
- N 7 san ,
- f, j
\
I, wo f HP
- 1. ).1 '
.L 'M, c, g 5
\$k i i !
,- 73 g- [--+---- + -- gg gg, b ,*
i i i l j I E F o 24*l ACCESS OPENX g IIYP I o PARTIAL PLAN e
E lT5sn3 CRuE RAll
\l i
MTAt:. m -
w !
= l IOP OF RAll I
'__~ . - . . - - . .. I i i i ir i
d i
, i m i i V-SHN io ASant REQUIRED nAit
/ og , , ,, toeless 8
}inw i
nP -
I i xrAit m ANSTEC
- v APERTURE CARD 0
]- CRANE CIRDER MA*ER:AL' 5A537 Cl2 0 d 3' C _,
H h4 Figure 3.8.21 (Stret/off)
Containment Vessel General Ottfline
=
%g&o/8 -0/ -
tk; vision: 1
\
_ , Ufectiv : 01/13/94 1 i
, 8 8 i #
m l
- N ./ /
I
/
/
65'-0 RAD l}s I
il l 6.691 t
30'f,4 X 4h THK lh5ERT PLAIE t
l*-0}s AT YERilCAL 4 Or HATCH v- ' 8 ll 10 l l El 3 il 18 If ,3 ll g 11 1 ll '
ll l l
? ' 22 '#
I g
ll ll Il
%o b l
~
. t
q 10
- . . ._. '.- -._jl.... ,
El (44'-6
, , 7, R i iI
( EQUIPPENT HATOF l ll AZ 6?'
, I, g 4l l 1I g iI
- il l Ii g ii g ii g Il
- Ii g II i ll II k ll g
5' 3'h LONG X 40 THK DARREL l A
4 61 ' 0 TO 4 22' 0 I .D. E0JIPMEAT HAICH r 66'-3'h'O(
c
's.
W
- Westinghouse v.y .
~
DRAFT f.- ,
t 6 6
, ._.,e._.._
~!
65 3 RAD _ _ _ _ l ie 4.833 a
23' 0 4 X 4 THK INSERT PLATE 7o io 4r vtaricat < or HarcH 4; ;;
- i. : :
ll
! N
~
4 l
_.._.U._
. LL H2'4
<EQUIPMENTHATch
- :: n a6 l ll l ll l
ll t !! 2 -6 +' t >: x 4 rm BAaaEt L rm t me tt io4 i e',
hs, N Ett:P50lDA. 80ll]M HEA)
ANSTEC
_ e 6)4 ro< AMA mm
_ (6 ' 0'ts 10 ( ;
CARD Also Availaole on Aporturo Card
,is 0_1.0. E0JI3piNI FATOl DIERIAL : 5A53' CL2 Figure 3.8.2 2 Equipment Hatches
&//h(jh $ 3 *-5 4 -
.- ... . . . .. . . ..... --.... via .ia, cwvwMeru, Arm d vsnMS Revision: 0 E#ectives 06/26/92 64* I1%10 ( C:
I I'4% os s u
_ d
__ -ee.
/ ~'~~'
I:I
- h. . . ._
,%/ --- -..
A
/ % -
' .i:
BAJ000 g s s
n -- - u TESTARI CAsrET w
- WTERSPACE TEST PAE55uit TAP 00m CASXET C3ACUtAfl0N nEssmE ca
,c. .
I
!. g s
+0 @
~ ' ' , N)lCATOR UQlT5 E-------- --1 TERMXAL B01 FOR DOCDCT COMNCATlG6 -
Dd
=w I I
, !' , . . ,m -
I l
l RD40TE H)lCATas H:
--t-- ,! - @ - -
i ma En FM OfDim DOCe l l l % FOR UCHTS .
.p [/ R i i l DOCDCT AR ACCESS 1 . I rm man oom !h d'- ~g x_. 7 TEST rocTunas -
l I I I
,, TE DOVNS STCRD tif trTons Do mv --
15' 8 MS T(
b T Westinghouse ~ ~
_,-6 s
-. 1 1
-s - ,
DRAFT i, _,
5
.cm, r -,-
a
~
- l
..is.-.-
E
&- lI Wl 4w w.-
s < atut s an
,- I - s *-io%cta m m wx Tl'4 Rio M
N ],J ,. . __
4'55 65'4s N
't M. e N CAGE 1
- -7%d W4
! l ~ l'l _I
'Y
_a i H --
s
~ ~ I
'N3M '
@l 7 ~(I ICATOR LICHT 5
[
- " "Aa"/ m g msg 3 -YALYE LMT
~
~\\ h fot LMT $WTTOI TV m m3
_ -. y . - ._ x... . . _ . 4
.+ta
$ $- ~
- q. 1 4 ~
7
( B~( (,, , a, 6=gMT ( Am mera m ANSTEC
-r " " m emu m tal e ~._ l- APERTURE s sunas
~ I CARD 40L a,.u au -"" "" ""i Tn_ usa a usn 12 not urE AISO A val!Oble or7 Aperture Card 9
4 Figure 3.8.2 3 k PersonnelNrfock iGC #4D/8f $ 8 ~'
1
. A ,
l DRAFT l l
l
'> 1 1
j 12* 11' _
=
/
I .
. ,* . .dos.*
4 o 4 l ieu, m .
404 .s 4 j i
l _ _ _. \ 3 ido.. ,
I. j l
,' 1.TS* _ _ _ _ . . , _ , _ ___,
1'
\, gay, g,o m yy,f ym-tres -m wirr7DM9Ar?.sys,gysyruscKUMuux'mvictuga-m a
8 d'
g
- W7E 6 m
- 1. . *k ; En\MR AR\n\RRR\RRRRRRRRRRR\MRRRR\MRRRRRRRRR\RRR\hu d\hWRH1.
I r m m m x m,mwwv.xmumxmmxmmwmxmmxmmmmummmmww I 9'.le's _ _ 13 *
\
- g ; ,
g .._...;.__.....i._._........_.._1..........._,_._.__._.._..._._._..,_--....___ q i
a o . . . i n.v \
I
, i i . i *j. 4 i *o*
i i .
J
- i.i5- *
=
s surus i
u < , *j. 4
/ E ,...... . . . . 4. .i . .__.'y _
40 4 l
, p. 4 */
- o
_ , v.e. twstat h att ,. ,4 e xstaate y4 y.
y i~ scar rtate 4-
%ia'c k ,.
. 4 4o4 do*
i CTMT.
/t i,
FLUED W ao SUPP0at Oul0E 4
f16 IN $7g A.m ~~P y n d NT1e si fiduA C f. 8,2 -af SNfW I *f(,
e.oNTAmm FM 'PF N FTAAT7eg
2 DRA'FT 1 l
l i
I 1
1 1
I
- I te' .P - !
2
- ai ie outor rue ,<.o ,coua
- PCA MTAIL 1 Te8LE 1
! (10197 a$ WTED l
l .c.ev a r , A.!
- t <o4
. 4 y.
4o e
/p
/ ,
l I.,[/gf mnuunuuaumumuuuuuummuoumnumuuuuuuumuuuunctS ...... .
l
,! \ .
_\ ,
.Nt
<o4 a ok i .
5rkkru? FftubxrSA wrinez n R c o A L 3 e.2-Y G a w ~
corThnmVM ~? f usrR w p n
e O
DRAi~T L
'EL 100' 0*
7 o* o m Se r
\*o f
- o* ,
, * * [
's o C> e '
/ o O EL 98' O' o' ,
- C> *
'1
'q% S 'n ssiett y ttniat routo wno .coutr' a y
' '- s o g ..
,\ .
. s I ,4o 9o O e
\ _
o .,
hSMuygu
._ ptt g.w .e.4
.3.g
...._ .._. .i ._... .._. .. _ .... . ...._._ .. . _ . _
' WffmfWM8WermMBE6W8E6m o
- s. M INsVLafION z
9-EL.94* B'
=
n.s- -
p i
.y '
sitto catale at sa - o V -
- 4 O'
EL. 92' 6' ,
. . . .;o o
- on o poAAAL 6HR. 7 Fa'4 PFA's7R5i*o f>lv R E. 3.8.2 ~4 (SHW1 e om. ,yr w , tam s
. - . . . . - .- . =..--.
. l DRAFT l
La lt' L8 -
l l
T !
Fs . .
LXXXXXX\\N \1p" / /
-. E l j,
MTAIL T
N I Efell 'A' g' AI TAIL .
) V//////////////2 k
g a I
t #
l 4 m Y 3 Fdll A INSIDE CONTAINMENT /
t em l
1 1
l l
_ x
- t __ ,
c '.
. .. z.
]iVhl L b 1
1 i l
l l
I l
1 fjfoAC 3.6.2-4 b N W h*I ge,ygg ysur 757 EW47~/* ^'#
l a~ m l 3 DESIGN OF STRUCTURES, COMPONENTS, EQUIPMENT, AND SYS11lMS Revision: 0 Effecifve: 06/26/92
~i i i i ! ...s......
i !
i !
i !
i !
i !
i !
i !
i !
i !
i
.L 111 \
l l 'N . . .
! !9 i !
i !
- i. cu *u l
- = _ -- r3 ___ ;
e - e
( a .
a to
= -
i
,, ; cao ao o o . . ses .
7._.
i !
i !
i !
i !
c.... ! -
+ . L- .
i
=d i ! i i
.,',. . .. , i Q* . '
.L. EA. .s7.G
. 5.'.CO. ne.o l
(.c ,
=
l l W-Westinghouse W *wr* e
WW bm oc u~.1-T B
/
s ,-
/ !
es -a -
/. . . .. . . e
-.......-...... /
Ud h
a flLf fs, ANSTEC
\
APERTURE
....i.---*'.*:'2*.**'L---------,n'- --- -
CARD n- .' ,
Also Available on Apertura Card p li
?
fQ ,'1:
1 4... f.. , ,
's'.
x.,
y -
i / '
, *', , s ,' .
. ,/
5
, , ... / __.-
A L_ _. __-
{ _ . . _ .
e ,-
.e2 . s ,
- \ '
. .-.....'~~,.- ,
l ~
s f . .. ... . -
I ' '
\ {
\, g#<
5, ;
,/
l \
\
V//A ,/
- t cG Figure 3.8.2-4 (Sheet /off)
'~s - 1 '
... . . . . . . . . ~' . --*...
,g g,; ,7;=;;;~,;
Containment Penetrations
.?.O '" ".'*T" .0" 11. Fuel Transfer PenetTation
~
(so//bo/8 -
,j_,_..,,,~e "'#
6 l
.s
'O '
O l 1 4 e n i t 4 k "i i i um
, / a
-un
( e r, ha we oo
,o, yr.
- r?. o-- 4 N b,,
0 m x
+
0 - ' --, - - .
o-F as 1
. =L_.'
o, ,
TJ+7 - -
- _}
- s F
- o , a 4 D*
M o,
- o0
$. [
. s s\N j,
)
b k th h (T.
O - s
,- O0 0
0o 3, s',
.O G
' " O e -
Q 0a 0
-4 m (%
, o l l o QQ C
e d
- -e -=
- D o
u LN ~-
- oo a <
_, 3o R ,J:-f. -%e RN l
2 1 p k b
s
\
i ,
b ; .
k la f y* -&. ..... _.
- e* ,
11
\ ,
i y[? 1xx h
_ r""'AMN3 -
- 1 ;:
L
- sxxe
]lgl k,
N; p:~ '- '
s
.C' ..:: R,- _
A QCP' . ////Jtr1
.{,
. 6
&?
I O., It 9
m
,a
s
+ i: ::
- 3. Design of Structures, Compone1ts, Equipment, and Systems DT3 f '[ +
L/t ~< A" '. . c 280 i i i i
-- DEFLECTIONS NORMAL TO SHELL
-- DEFLECTIONS ALONG SHELL 260--
~
N
',f./
\
240- - ,4,/ ' / , ,. .
\ -'
/* I
/ /
/ /
220-- N- . .(, --
r I./'
w w
LL. 2ee --
I s -
I i z I B l 4188-- I ;
l a I I
c,M l W
l 160--
I N,I -
I I
I l 14e-- I I
i %,1 12e-- I I l J
'), 1e i;s 2.0 t ee.. s . ~e . 's DEFLECTION - INCHES Figure 3.8.2-5 (Sheet 1 of 4)
Containment Vessel Response to Internal Pressure of 45 psig ikflections Revision: Draft January 4,1996 3 W85tingt100S8 3.8-64
( --v
- 3. Design of Structures, Compone ts, Equipmelt, r.nd Syst ms p pw .-1
.'7 4'8 4
- w. a 280 i j i j j j j MERIDIONAL STRESS
-- CIRCUMFERENTIAL STRESS 260-- -.
/
/ ' / ;T/ /
240-- ,/, / -.
,s ,/
/
t
/,
220-- --- -- --
___ l H
~~~~s "
1 w et" w
200-- L t' ' ' ~ ~ -.
z /
o ,
M l l H \
< 1e0 -- -.
> )
w -
a ,'
W
~~s i N
160-- f l
I i
140-- ) -.
/
\
120- -
,I -.
\
J see t I C'T~~ ~
l I
-15 -10 -5 0 5 i0 t'5 2'O 25 MEMBRANE STRESS - KSI Figure 3.8.2-5 (Sheet 2 of 4)
Containment Vessel Response to Internal Pressure of 45 psig Membrane Stresses Revision: Draft W Westinghouse 3.8-65 January 4,1996
. i i -- M
- 3. Design of Structures, Components, Equipmert, tnd Syst:ms
._.- m -
)l% .t 2ee i j ; j i j ; i
- INSIDE SURFACE
-- DUTSIDE SURFACE 260-- --
/
240--
/ --
e#' '
220- -
'],- - -
--f,,, * --
i K ,. '
~*
---s=
ua u.
2e0- -
- ~ ~ " -
._. 1 ".* ~ Tz_y- = =: :-- " ' - '
- - - - z'l
/
z !
O 1
>.4 l H l
< 180-- --
l J
/N '
gj 7~~--*M*---..-.
~- .-- -
l
[ 's l 160-- --
140-- --
/\
, = - N -- p-T-~l-~
120- - --
2
~C ^ 2O-1,,le
~
- r z- E ' Z m == i ~ ~ ~
i
-5 0 5 1's 15 20 25 3'O 35 SURFACE MERIDIONAL STRESS - KSI Figure 3.8.2-5 (Sheet 3 of 4)
Containment Vessel Response to Internal Pressure of 45 psig Surface Meridional Stress Revision: Draft January 4,1996 W W86tingh00S8 3.8-66
- 3. Design of Structures, Compone1ts, Equipme t, tnd Systems
'., , 'i
- s. a 2 280 i j j j j j j
-- INSIDE SURFACE
-- DUTSIDE SURFACE 260-- --
/
240-- s/ --
220-- '--'N= -
" l H _.
y 7.-
~"~~ -J --
200-- %,
1 1 z
c3 w
H g 1B0-- }--
u J _
- -f p *' l i
w
- 4,%
160-- f--
l 14e--
~<>
) --
~~.'
120-- --
- )
' ' '_' ~~ ' ' '
100
-15 -10 -5 0 5 1'O 15 2'O 25 SURFACE CIRCUHFERENTIAL STRESS - KSI Figure 3.8.2 5 (Sheet 4 of 4)
Containment Vessel Response to Internal Pressure of 45 psig Surface Circum. Stresses Revision: Draft T Westinghouse 3.8-67 January 4,1996
-u
- 3. Design of Structures, Compone ts, Equipment, r.nd Systems
- 3. , - - ,
l
_c.
b .E(25{f,f,jns(de[y,a(Pg = 180,*
qi -
El. 249 4 a
U n155.g2,0*
. . i! I ;
a . _ E l . 2,4,{4, _ , , , , , , , , ,, ,
_ y_=, ,149,5g 6,*_
5 ._.il,. 035 ' t = t .6 25- -
44' A ei. 227-61/4 . , = 117.399*
-p -
@ --+- @
,/
l Q
El. 210'-81/2,9 = 90* T . L. .a
@ j 1
Q / @ l
. / .
mm
@_ . ~._
. . .I. . . _ . . _El 2_0.s. 4 . . . . _ . ._.
. _ _ . ... .. . ,. y 3 e ._...:.," g ', e
. / goe,, w,e, . .. .-. a_
! h
_.L..7_ . .E!;, } 9g-O_ , , _ , , _ , , _ , , _ , , _ Girder (80 Ard )
, , , Plaes ,8* , 36* j
,# h s t = 1.5* (typ.) 'I
. 9: Mendianal (along 's') h .
30*
! 8- Circumferential
@ h: Q
_.,_.._..E_l.17_7..0_.._.._.._..._.._'.p,
'. m h ... ....
- %... . .
- N Q10 5, i
, El. 16 2' .0_ . . _ . .\ . . _ . . _ .(.[L_/. . esy_
i
.+,
- i
- t = 1.75*
., INSIDE RADIUS, R = 65'-0 = 700* '
t = 1 '625' p REFERENCE SURFACE '
_ . . _ . . _ . .E_l . 14_4.' .6_ . . _ . . _ . . _t . 44*
l j l' i @ @
. . . ;_ . . _ . . EI; ,1p;9 , , , , , , , , , , , _ , , _ , , .. . .
b ]@" g m,
@, .+- t = 0 75*
3
. . . . . _ . . E_l 1_16.
. '. _10. 9_/32. _ . . _ . . _ . . _ / aper T 1:3 (tW)
. _Im. _ . . _E. .l_ . 1 1 2_'-6. . _ . . _ . . _ . . _ . . _ . 3- , , , 73 1
47 O e PARTEND
' '.104'.1 El 1/2 - k_2./.. T. L.
7 ' \ El.107-0 ' ' .(1jf )
I
\ g' # 4 gf C PARTNO 9 = 79 212* N Q l /
- . w=u,=p,=uga0 /
I
/
S l /
- /
/
l j /
.x. _ _;_
Figure 3.8.2-6 Containment Vessel Axisymmetric Model Revision: Draft January 4,1996 W W8Stingh00S8 3.8-68
i._ .
i I n-w g
l
- 3. Design of Structures, Components, Equipment, end Systems j r. n 4 ~ c w- ,
a n
l I i i
i ,
t l
I l
I l
i l
l l
{ Figure 3.8.2 7 4
l Finite Element Model for Local Buckling Analyses 3
I i
Revision: Draft
! 3 W85tingh00S8 3.8-69 January 4,1996
~ _ _ _ _ _ _ _ _ _ _ _ _
- 3. Design of Structures, Compone1ts, Equipment, and Syst:ms
("s y *r r%
.. . . , . u DETAIL '
/
= EL 108 -2' y ,,
v
) EL.100*-0*
Et. 94'-0*
Nm DETAIL EL 92*-s*
tt. e3 -o* tt. e2 -s'
- l l
l i
i tt. so.-s' l c
Figure 3.8.2-8 (Sheet 1 of 2) l Location of Containment Seal Revision: Draft i January 4, IM 3 W95tiligh00S8 3.8 70 ,
- 3. Design of Structures, Components, Equipmelt, and Systems P. O A T"
. . r '. e STEEL CONTAINMENT VESSEL
/
2"X SILICONE
/ ADHESIVE SEALA
'2", / PRIMER
. / / (STEEL FACE ON 108'-2" h /
$ * ',R.'. *.** . .\
\ BOND BREAKER Tl G;,M;,/, OR CERAMIC FEL'.
l@,,j/
. %* l
' '.* i l l
DETAIL 1 (DETAIL 2 SIMILAR)
Figure 3.8.2 8 (Sheet 2 of 2)
Seal Sections and Details Revision: Draft
[ W8Sfingfl00S8 3.8-71 January 4,1996
4
.P
- 42. Conditional Containment Failure Probability Distribution 11 mm -
CHAPTER 42 CONDITIONAL CONTAINMENT FAILURE PROBABILITY DISTRIBUTION ;
l 42.1 Introduction -
The probability distribution for containment failure due to intemal pressurization of the containment has been developed for the AP600 containment vessel.
The AP600 containment and its structural propenies are described in subsection 3.8.2 of the AP600 Standard Safety Analysis Report (SS AR). He limiting containment failure modes that have been identified include:
. General yielding of the cylindrical shell
- Buckling of the ellipsoidal head
. Buckling of the 22-ft. equipment hatch
. Buckling of the 16-ft. equipment hatch
- Yielding of the personnel airlock .
Other containment failure modes are examined, as discussed in subsection 3.8.2 of the SSAR.
Other failure modes, such as general yielding of the ellipsoidal head and failure of the piping penetrations, are not considered to be independent containment failure modes. Rather, these other failure modes are bounded by the failure criteria for the limiting failure identified above.
Failures of the mechanical penetration bellows, and leakage of the equipment hatches due to ovalization, do not occur prior to general yielding of the cylinder. Failures of the electrical penetration assemblies do not occur prior to general yielding of the cylinder for temperatures equal to or less than 400*F.
Each of the limiting failure modes is examined to determine the best estimate of the failure pressure. In addition. the random and subjective uncertainties associated with each of the failure modes are identified. These failure characteristics are then used to develop a probabilistic model to predict the containment failure due to intemal static pressurization. The details of the model development and the results of the analysis are presented in the following sections.
42.2 Probabilistic Model To define the probability of a containment failure due to internal pressurization,it is necessary to select a statistical distribution with the correct properties. De engineering justification for a panicular probability density function is made based on the gathering and evaluation of relevant infonnation that can serve to characterize the nature of the random data and physical pn> cesses that lead to the random data. The nature of the random data, and in particular any limits or bounds on the data. are as important as the predicted means and variances from statistical analysis of the data. Thus, specific limits in the data and characteristics, such as g Revision: x N sym b , Draft December 1,1995
4
.p
- 42. Conditional Containment Failure Probability Distribution ru-skewness, are utilized to specify the probability density function. Five potential probability 4 distributions are considered: the Gaussian, the Gamma. the Gumbel, the lognormal. and the l Weibull. )
Based on a review of the characteristics of the five potential probability distributions, it was i detennined that both the Weibull and the lognormal distributions would be suitable to desenbe the containment failure probabilities. An additional review of the containment failure probability distributions reported in a number of the Individual Plant Examinations submitted to the Nuclear Regulatory Commission (in response to the Commission's Generic Letter 88-20) indicates that the lognormal distribution is the most commonly used distribution form for predicting containment failure from intemal pressurization. Therefore, the lognormal distribution is selected to construct the conditional containment failure probability distnbution.
42.3 Containment Failure Characteristics l
The characteristic parameters for containment failure due to intemal pressurization are derived from detailed analyses of the containment vessel. supplemented by applicable test data for certain design features of the containment, as described in subsection 3.8.2 of the SSAR. For the construction of the conditional contamment failure probability distribution, the required" characteristic parameters are the mean failure pressure and the statistical variance that represent the uncertainty associated with the mean value. ;
1 42.3.1 Mean Values for Containment Failure The development of the cond,itional containment failure probability distribution requires the specification of the mean value for contamment failure for each of the possible containment I failure modes. Subsection 3.8.2 of the SSAR provides values for the ultimate containment pressure capability at 100*F and 400*F. These failure pressures are based on code specified minimum material properties. To obtain the mean and median values for the probability distnbution. the failure pressures of subsection 3.8.2 of the SSAR are adjusted to account for the expected material properties and this adjusted best estimate failure pressure is considered as the median value of the lognormal distribution.
The contamment vessel is designed using SA537. Class 2 material. This has a specitied minimum yield of 60 ksi and minimum ultimate of 80 ksi. Test data for materials meeting l SA537 or having similar chemical pmperties was reviewed from two United States and two Japanese steel suppliers. Some of the data was from tests of steel procured to S A537, Class 2, while the remaining data was identified by the steel supplier as having similar chemistry to SA537, Class 2 and being representative of SA537, Class 2. In a sample of 122 tests for thicknesses equaling or exceeding 1.50 inches and less than 1.75 inches, the actual yield had l a mean value of 69.1 ksi with a standard deviation of 3.3 ksi, giving a mean yield value equal to 1.15 times the specified minimum yield with a coefficient of variation of 0.048. Test data l for 389 tests for thicknesses from 0.31" to 3.16" showed a mean yield value of 71.7 ksi with l a standard deviation of 5.7 ksi, giving a mean yield value equal to 1.19 times the specifieJ l mirumum yield with a coefficient of variation of 0.079.
JLXXL mm _ Westingt100se
9
- 42. Conditional Containment Failure Probability Distribution s l l
i l
Reference 42-1 confirms that the actual yield strength of containment construction matenal can typically be expected to be 9 to 22 percent higher than the specified minimum material strength with coefficients of variation of 6 to 137c (the lower yield strengths reported are not appplicable to containment material, the ABS steel is procured to an ABS specification and not to ASTM; the static yield stress is a little lower than ASTM test yield). Since the mean )
containment strength is used to construct a probability distribution that includes random I uncertainties in material properties as well as subjective uncertainties in modelling of the l containment strength, it is appropriate to use the expected, as-built containment strength (i.e..
the 15 percent increase over the nominal value) for the mean value of the distnbution.
~
J However, since as-built information is not available for the AP600, the fragility is conservatively calculated using only a 107c increase above the specified yield and this value is used as the median value of the lognormal distribution.
The containment intemal pressure value used for the median value of the containment failure pmbability distribution is the expected failure pressure, evaluated at 400*F. A review of the severe accident sequences in which the contamment intemal pressure approaches the failure l' pressure of the contamment (inckling the decomposition event trees in Section 43 of the AP600 Level 2 PRA) leads to the conclusion that the contamment shell is likely to be at the' ,
containment saturation temperature (for the intemal contamment pressure) for most severe l accident sequences. With the passive containment cooling provided by the containment design, the highest likelihood of containment failure due to overpressurization is due to extreme cases of severe accident phenomena (e.g., hydrogen bums and noncoolable ex vessel core debris). As described in Section 34, the temperature of the containment vessel steel does not significantly exceed the design temperature of 280*F. Therefore, the use of a uniform containment shell temperature of 400*F for evaluation of the containment material properties l is bounding for the prediction of the conditional contamment failure probability distribution. l The values used to construct the conditional containment failure probability distribution are identified in the next section.
42.3.2 Uncertainties in Containment Failure The uncertainties identified and examined include both random uncertainties and subjective uncertainties. The broad categories of uncertainties considered are:
Geometric Propertiest This category of uncertainty is principally concemed with the vanations between the as-built containment vessel and the design utilized in the analysis.
Some of these variations include contamment dimensions, placement of stiffeners, and thickness of the steel plates used to make up the contamment vessel. Also included in this category are construction practices such as the strength of weldments, etc. It has been reported (Reference 42 11 that the overall uncertainty in the containment strength is generally insensitive to variations in geometric properties, except for buckling mode of failure.
g Revision: x
_______ _ __ ____ _ _HF h_ Draft December 12 1995
~
1-:
- 42. Conditional Containment Failure Probability Distribution Structural Analysis: The uncertainties in the overall containment strength can be sensitise to uncenainties in assumptions and models used in the structural analysis of the ulttmate strength l of the containment structure. Some of the sources of uncertainty include: the detimtion of l containment failure used in the analysis, the simplified geometric model used in the analysis, the analysis method, the analysis focus of failure locations and modes. the yield entenon for .
biaxial stresses, the rate of loading, the effect of non uniform geometnes, the effect of local l temperature and the interpretation of test results to construct the analytical model. These l uncertainties are subjectively evaluated since no complete investigation of these uncenainues l
'is available. Reference 421 provides several estimates of the actual to-predicted results. l which vary according to the failure mode assumed in the analysis, the person doing the evaluation, and the method of analysis. The range suggested by these values is a mean value for the actual to-predicted results tending to unity and a standard deviation in the range between 0.08 and 0.24. Reference 42 2 suggests that a value of 0.12 be used in constructing the probability distribution for the ultimate strength of the contairunent.
Afaterial Properties: Uncenainties in material properties can be important in estimating the overall containment failure uncertainty. De total uncertainties in materials consider the estimation of statistical propenies from a small sample (e.g., is the calculated mean the real mean) and assumptions on uniformity of propenies. There is a wide range of application of' material properties to estimate uncertainties and, except for the buciding mode of failure, most analyses neglect all uncertainties except the random, measurable variations in material propenies. Reference 42 1 provides several estimates of the uncertainty in material properties that show a coefficient of variation in the range of 0.044 to 0.1i for conditions that may be applicable to the passive containment shell. Reference 42 2 recommends that a coefficient of variation between 0.06 and 0.08 be used to define the random variance in matenal properties for the containment shell. Finally, based on sampling of test results of material similar to that specified for the AP600 containment shell as described in subsection 42.3.1.
the coefficient of variation was found to be 0.048.
Gross Errors: Gross errors in construction and / or design are not quantifiable since they lead to catastrophic results that are not predictable by reliability methods.
The values used to constmct the conditional containment failure probability distnbution are identified in the next section.
42.4 Containment Failure Predictions 42.4.1 Containment Cylindrical Shell ne response of the cylindrical portion of the containment vessel to internal pressurization has been analyzed, and the results reported in Table 3.8.2 2 of the SSAR. The best estimate of the pressure at which failure would occur is 120 psig, based on specified material properties at a uruform containment wall temperature of 400*F. This was adjusted to 132 psig to account for the expected actual material properties.
n mh.e t toeg 4 Westirigt10ljse
_ - -. - ~ _. .
- 42. Conditional Containment Failure Probability Distribution 1
l A coefficient of variation of 0.06 is used to represent the random uncertainty in matenal ,
properties. consistent with the variability reported in subsection 42.3.1 for the sample of 122 l i ests of similar steel and with References 42-1 and 42 2. For the subjective uncertainty associated with modelling of the ultimate containment failure pressure. a coetficient at vanation of 0.10 is used. This value is derived from Reference 42 1 (Table 5) for the analysis error (in-place / analytical result). It is the average of the coefficients of vanation suggested for analyses based on limit analyses and axisymmetric finite element models. The predicted failure is based on the occurrence of large radial deflections of the shell and was estimated from both axisymmetric finite element analyses and hand calculations. The mean and coefficient of variation are supported by testing of scale model containment vessels, such as Reference 42-3 where gross yielding occurred within a few percent of the predicted pressure.
42.4.2 Ellipsoidal Upper Head .
1 i
The response of the ellipsoidal upper' head of the containment vessel to intemal pressurization l has been analyzed and the results are reponed in subsection 3.8.2.and Table 3.8.2-2 of the l SSAR. Failure is predicted to occur either in the knuckle region or at the crown and may be l initiated by buckling in the knuckle region. Failures due to tensile stresses (plastic collapser l l are bounded by the vanations considered for yield of the cylindrical shell. Only the buckling failure mode at the knuckle region can be considered to be an independent failure mode that l must be separately considered in determining the conditional containment failure probability l distribution.
4 The best-estimate intemal pressure at which the ellipsoidal head of the containment vessel would fait due to post-yield buckling in the knuckle region is 145 psig. using minimum l specified yield strength of the containment materiah at 400*F. Since this buckling is l associated with yield in the knuckle region the capacity was adjusted to account for actual material propenies by the ratio of actual to minimum yield to give a predicted pressure of 160 psig.
A coefficient of variation of 0.06 is used to represent the random uncertainty in matenal 2 propenies, consistent with the variability reported in subsection 42.3.1 for the sample of 122 tests of similar steel and with References 421 and 42 2. For the subjective uncertainty associated with modelling of the ultimate contamment failure pressure, a coefficient of variation of 0.14 is used. This value is consistent with References 421 and 42-2 for the buckling mode of contamment failure. It allows for uncertainties including those due to geometric propenies (as-built condition) and residual stress.
42.4.3 Equipment Hatches The response of the 22-ft. and 16 ft. diameter equipment hatches to intemal pressurization has been analyzed, and the results are reponed in subsection 3.8.2 of the SSAR. The containment intemal pressure acts on the concave face of the dished head and the hatch covers are in compression under containment intemal pressure loads. The predicted failure mode is elastic g Revision: s N g,.m%., Draft December 1,1995
-1 i
- 42. Conditional Containment Failure Probability Distribution l
buckling of the hatch covers. The best estimate of the pressure at which fatiure would occur for the 16-ft. equipment hatch is 226 psig at a unifonn containment wall temperature of 400 F. based on 150 percent of the critical buckling pressure as indicated by a review of test data for buckling of spherical caps. The matching value for the 22 ft. hatch is 276 psig.
A coefficient of variation of 0.06 is used to represent the random uncertainty in matenal properties, consistent with the variability reported in subsection 42.3.1 for the sample of 122 tests of similar steel and with References 42-1 and 42 2. For the subr~ive uncenainty associated with modelling of the ultimate containment failure pressure. a coefficient of variation of 0.14 is used. His value is consistent with References 42-1 and 42-2 for buckling failures. It allows for uncertainties including those due to geometric propenies (as-built condition) and residual stress.
42.4.4 Personnel Airlock The response of the personnel alriock to intemal pressurization has also been analyzed, and the results are reponed in subsection 3.8.2 of the SSAR. The estimated pressure at which failure would occur is in excess of 300 psig, based on test results. Since the mean failure pressure is far above the mean failure estimates for the other contamment failure modes, ne further analysis of the personnel airlock is performed. Since its expected contribution to the overall contamment failure probability distribution is negligible, it is not included further in the development of the conditional containment failure probability distribution.
42.5 Overall Failure Distribution LATER 42.6 Summary and Conclusions The cumulative contamment failure probability distribution has been developed, using lognonnal distribution. which is based on best-estimate predictions of containment strength and accounts for random uncertainties in material properties and subjective modelling uncertamues. Based on this model, the median intemal pressure at which the AP600 contamment vessel is predicted to fail is (later) psig. This is the best-estimate or expected containment failure pressure. This value is comparable to, or slightly higher than. the expected contamment failure probability for other conventional pressurized water reactor (PWR) plants using pre stressed or post tensioned concrete contamment structures.
De 5th and 95th percentile failure probabilities are (later) psig and (later) psig, respectively.
The cutoff for consideration of conuunment failure due to intemal pressurtzstion during a severe accident is defined as the pressure at which the containment failute probability is less than 10". Below this point the failure probability is so low that. when combined with the small core damage frequency numbers. tne overall probability of a core damage accident resulting in containment failure is in the 104' range. This is generally considered to be a ember 1,1995 sm NO
l w
- 42. Conditional Containment Failure Probability Distribution m- l negligible calculated number. From the lognonnal distribution, the containment press 're l corresponding to a 10" probability of failure is (later) psig.
42.7 References 42.l " Reliability of Containments Under Overpressure." L. Greimann and F. Fanous. Pressure L'essel and Piping Technology.1985. pp. 835 - 856.
42 2 " Reliability of Steel Containment Strength." L. Greimann. et. al., NUREG/CR 2442. June 1988.
42 3 " Comparisons of Analytical and Experimental Results from Pressurization of a 1:8 Scale Steel Containment Model." Clauss. D.B. and Horschell. D.S.. Proceedings 8th Intl. Conf. on SMiRT. Aug. 19 - 23. 1985 m6 Draft Decem r1 5
- 42. Conditional Containment Failure Probability Distribution Table 42-1 PARASIETERS USED IN THE CONSTRUCTION OF THE AP600 CONTAINMENT FAILURE PROBABILITY DISTRIBUTION Failure Location Failure Mode Niedian Failure Coefficient of variation Pressure' (psig) Material Modelling Cylindrical Shell Membrane Yield 132 0.06 0.10 Ellipsoidal Head Buckling 160 0.06 0.14 22 Ft. Equipment Hatch Buckling 276 0.06 0.14 16-Ft. Equipment Hatch Buckling 226 0.06 0.14 Personnel Hatch 300 0.06 0.14 All median failure pressures are those specified at 400 *F Table 42 2 CUMULATIVE CONTAINMENT FAILURE PROBABILITY Containment Probability of Costalament Failure Cylinder Head 22-ft. Hatet 16-ft. Hatch Total LATER Revision: x g Draft, December 1,1995 WJamb. DM
l
. e NTD NRC-96-4617 Enclosure 3 I AP600 Containment CCFP Calculations Request for Additional Information, dated 9/14/95 l
Ouestion # 1 ,
In Chapter 42 of PRA SSAR, Rev. 4, the mean failure pressure is mentioned for each failure mode. As I stated in DSER, the staff recommended the best estimate pressure be median for containment CCFP l 2
calculation. (If lognormal distribution is used, the mean is median times exp(B /2) where B is ogarithmic I standard deviation.) For the failure pressure estimates, the staff is not in a position to accept the 32% l increase using both von Mises criterion and mean yield strength of SA537 Class 2 material. See Open Items 3.8.2.4-19 and 19 2.6.2-3.
- A comparison between experimental and theoretical yield stresses in Engineering Design, Faupel, l J.H., pp. 249-258, John Wiley & Sons,1964 shows that the von Mises yield criterion does not I always give a 15 percent higher yield stress than that obtained from the maximum shear stress criterion,
- The material test data uses only 122 specimen and they are neither exactly the same as the SA 537, Class 2 material nor as-built material.
l
- In " Comparisons of Analytical and Experimental Results from Pressurization of a 1:8 - Scale Steel Containment Model," Clauss, D.B. and Horschell, D.S., Proceedings 8th Intl. Conf. on SMiRT, Aug.19 - 23,1985, and NUREG/CR-4209, the measured yield pressure was reported 15% less -
than that predicted yield pressure (r = 84", t = 0.197", yo = 57.1 ksi. P = cyt/r = 134 psig) using MARC FEM code with large displacement, nonlinear material property obtained from standard uniaxial tensile tests (test coupons were machined from remnants and cutouts for the penetrations),
and von Mises yield criterion due to (1) strain rate effects (5% reduction), (2) Bauschinger effect (5 to 10% reduction) referring to the phenomenon whereby the yield stress in tension or compression is reduced if the material has been previously yielded in the opposite sense (when the plates comprising the cylinder were rolled into the cylindrical shape, the internal surface underwent compressive yielding and intemal pressurization results in tensile yielding in the cylinder), and (3) difficulties in applying uniaxial data to multiaxial strain states,
- From an American Iron and Steel Institute (AISI) survey of test results for thousands of individual product samples, it has been found that strength levels vary as much as 20 percent from the l certified material test reports (CMTR) test values. It has been the staff's position that minimum specified strength values (e.g., ASME Code minimum strength values) should be used as the basis for allowable stresses as described in the letter from G. Bagchi and C. Cheng to J. Stolz,
Subject:
l L Review of Oyster Creek Drywell Containment Structural Integrity, dated June 14, 1990.
l
Response
PRA Chapter 42 will be revised to use the lognormal distribution with the best estimate pressure considered as median for the containment CCFP calculation. The best estimate pressure for the cylinder will be based on the von Mises yield criterion and median actual material properties 10% above the ASME specified yield. The examples quoted in the RAI are discussed below. These examples are considered in the selection of the material and modelling uncertainty coefficients of variation.
l
_ _. _ 7 NTD-NRC-96-4617 Enclosure 3
. The von Mises yield criterion is a theoretical criterion that does give a 15 percent higher yield stress than that obtained from the maximum shear stress criterion for the stresses due to pressure in a cylindrical vessel (longitudinal stress equal to one half hoop stress). Test data shown in the reference indicate "slightly better agreement with the distortion energy theory (von Mises) than with the shear theory". In the tension-tension interaction portion the test data is between the shear theory (1.00 factor) and the distortion energy theory (elliptical interaction curve with maximum equal to 2/43 = 1.155. Most text books recommend the distortion energy theory and this is the basis included in the BOSOR-5 analyses performed for the AP600.
The uncertainty in yield criterion is a part of the modelling uncertainty.
. The material test data quoted in the SSAR refers to 122 tests where the plate thickness was between 1.50" and 1.74" and gave a mean value of 69,100 psi with a standard deviation of 3,300 psi. This was quoted since it was most representative of the 1.625" thickness for the AP600 Test data for 389 tess for thicknesses from 0.31" to 3.16" showed a mean value of 71,700 psi with a standard deviation of 5,700 psi. This test data came from two United States and two Japanese steel suppliers. Some of the data is from tests of steel procured to SA537, Class 2, while the remaining data was identified by the steel supplier as having similar chemistry to SA537,
. Class 2 and being representative of SA537- Class 2.
The results for the 122 tests give a mean value equal to 1.15 times the specified yield with a coefficient of variation of 0.048. His is similar to that fo- A 516 grade 50 material (mean =
1.19 times the specified yield, COV = 0.08) quoted in Table 1 of " Reliability of Containments under Overpressure" (PRA Chapter 42, reference 42-1)
. NUREG/CR-4209 reports a predicted membrane yield pressure of 180 psig for the 1:8 scale model. His prediction is described in NUREG/CR-4137, where it is described as general yielding of the cylinder (page 15). He prediction is based on an axisymmetric finite element analysis described in Section 6.1 of NUREG/CR-4137, where the ring stiffeners and shell are modelled. Typical analysis results of the ring stiffened cylinder are shown in Figures 8 and 9.
The effect of the rings is to restrain a portion of the shell while the cylinder midway between ring !
stiffeners expands further; the rings therefore increase the pressure capacity corresponding to gross j yielding of the cylinder as a whole. As the pressure is increased, local yield initiates and loads are shared by the stiffener rings. General membrane yield does not occur until both the shell and the rings reach yield. This is best observed from radial displacement plots from which yield pressure corresponds to the pressure at which the radial displacement rapidly increases.
Unfortunately deflection data for the test was not complete and only strain data is available.
The yield capacity of 134 psig calculated in the question uses the shear yield criterion and average thickness for an unstiffened cylinder. De capacity may also be estimated from the uniaxial yield stress (57,100 psi), the radius (84 inches), the nominal thickness (0.1875") and the stiffener rings (2.75" x 0.1875",15" spacing). The shell is in biaxial tension and yield is calculated using the von Mises criterion. The ring is in uniaxial tension.
Unstiffened cylinder: pus = c / R x 2/43 x t
= 5Y100 / 84.09 x 1.155 x 0.1875
= 147 psig 2
o o NTD-NRC-%-4617 Enclosure 3
}
i Stiffened cylinder: ps = ay/ R x (2N3 x t + A3)
= 57100 / 84.09 x (1.155 x 0.1875 + 2.75 x 0.1875 /15)
= 170 psig i
- NUREG 4209 reports an average plate thickness of 0.197". This would increase the above capacities to 155 psig and 179 psig. The stiffeners are far enough apart that the shell mid way between ring stiffeners acts elastically as an unstiffened cylinder. Thus, the expected behavior would be that the cylinder near the midpoint between ring stiffeners will yield when the pressure reaches 155 psig and additional load is carried by the shell bending between the ring stiffeners.
- This leads to yield at the ring stiffener due to combined hoop and longitudinal strains, and thereby reaches the general yield pressure at 179 psig. This is consistent with the predicted capacity of 180 psig given in NUREG/CR-4209. Note that yield would be predicted to start at a pressure of 155 psig which is equal to the pressure at which the NUREG reports that it started.
NUREG/CR-4209 reports that membrane yielding for most of the cylinder occurred between 155 ,
1 and 171 psig. The lower end of this range is believed to be yield of the cylinder between l stiffeners and does not represent general membrane yielding of the stiffened cylinder. Yielding l of one ring stiffener is shown in Figure 10 at a pressure of about 163 psig. General membrane i yielding (shell / ring combined) leading to large radial deflections does not occur until all of the
, cylinder reaches membrane yield and can be assumed to occur between the 163 psig for the ring in Figure 10 and the 171 psig magnitude reported as the upper end of the yield range (the third paragraph of Section 9 of NUREG'/ CR 4216 also states that, up to the pressure of 170 psig, a
- single yield plateau was not observed). This is 5 - 10% below the predicted capacity for the smeared stiffeners.
The difference of 5 - 10% in the test and analytical results for gross yield of the 1:8 model'is reported in the referenced NUREGs as being due to a combination of the strain rate and the Bauschinger effect. It could also be due to inaccuracies in the von Mises criterion as discussed in the first bullet. The difference of 5 - 10% in the test results is small and is accounted for in the AP600 fragility estimates in the coefficient of variations assigned to materials and modelling uncertainties.
l Steel yield test data is summarized in the paper " Reliability of Containments under Overpressure" (PRA Chapter 42, reference 42-1). The steel for the AP600 will be procured to ASME requirements which should result in better quality than the range of data reported in the AISI survey. Specific examples are described in previous paragraphs.
The staff's position that minimum specified strength values (e.g., ASME Code minimum strength values) should be used as the basis for allowable stresses is the basis for allowable stresses for design loads and not to the "best estimate failure analysis". Other fragility evaluations have used best estimate yield properties based on 10% above the specified values. Westinghouse will also use this assumption in updating the fragility calculation.
Question # 2 In Section 42.2, is lognormal distribution applicable for the 16 ft and 25-ft equipment hatches? Due to their convexity, these are under compression when subjected to containment internal pressure as mentioned, further justification is necessary for these equipment hatches.
3
$D NTD-NRC-96-4617 Enclosure 3
Response
The 16-ft and 25-ft equipment hatch covers are arranged such that, due to their convexity, they are under i compression when subjected to containment internal pressure. The internal pressure is defined for the
!- whole containment vessel; however, the hatch covers are designed using the ASME rules for extemal pressure loading (NE-3133 and Code Case N 284). In sections of the SSAR and PRA report discussing equipment hatch pressure capacity, a statement has been added that the internal pressure acts on the convex face and the cover is in compression.
Distribution is primarily influenced by imperfection. Measured imperfection must be less than ASME specified limit. Use of lognormal distribution is considered appropriate, similar to its use for material yield which must also exceed a minimum ASME specification.
Ouestion # 3
- In Section 42.3.1, Ref. 42-1 did not provide data showing that the actual yield strength of containment construction materials could be 12 to 22 percent higher than the specified minimum material strength.
The range is 2.5 to 22 percent in Table 1. Also, there is no data for SA537. Class 2 material in Ref. 42-1.
Response
Section 42.3.1 has been revised. The reference includes two sets of relatively low data. One set is ABS 4
steel which is not covered by an ASTM specification, and may not have a specified yield. The other set is static yield values which are at a lower strain rate than the ASTM test specification.
t Question # 4 i
In Section 42.4, provide uncertainties for ge'ometric properties (as-built condition) and residual stress for buckling of knuckle area and equipment hatch covers. Imperfection for internal pressure is insensitive,
- however, for extemal pressure, it should be significant. (from N-284, capacity reduction factor is considered for imperfection and plasticity of nonlinear material properties.)
Response
Section 42 has been revised to state that the COV includes consideration of these uncertainties.
Ouestion # 5 i
In Section 42.4.1, how is the COV of 0.1 derived from Ref. 42-l? He Table 4 of Ref. 421 shows only the thickness of 1-1/4" thickness (mean = 1.277", o = 0.012", COV = 0.01) and it assumes normal distribution, not lognormal distribution. Also, this COV of 0.01 represents the uncertainty for geomet,ric properties, not modeling error. The Ref. 42-1 shows the modeling error COV of 0.144 from (0.12 +
j 0.08')h in Table 7. De COV for all practical purposes of modeling error which should include nonsymmetric features such as penetrations and other reinforced openings, longitudinal stringers, etc. as well as circumferential variations in thickness, ring and stringer sizes, amount of reinforcing steel, and 4
o.
NTD-NRC-96-4617 Enclosure 3 shell imperfection is 0.12 (Ref. 42-2). The staff believes that the use of the COV of 0.1 results in unconservative CCFP calculation. See Open Items 3.8.2.4-21 and 19.2.6.3-1.
Response
ne value of 0.10 is derived from Reference 42-1 (Table 5) for the analysis error (in-place / analytical result). It is the average of the coefficients of variation suggested for analyses based on limit analyses and axisymmetric finite element models. The predicted failure is based on the occurrence of large radial deflections of the shell and was estimated from both axisymmetric finite element analyses and hand calculations. The mean and coefficient of variation are supported by testing of scale model containment vessels, such as Reference 42-3 where gross yielding occurred within a few percent of the predicted pressure.
Question # 6 In Section 42.4.2, provide mean (median) failure pressures with modeling and material uncertainties for crown yield, knuckle area yield, and knuckle area buckling. Imperfection uncertainty is insignificant due to internal pressure buckling (See N-284).
How is 192 psig derived in knuckle area? Is it 146*l.15*l.15?
How is 144 psig derived for ellipsoidal head buckling failure mode in Table 42-l? It is not given in SSAR. Is it derived from 174 x 138/166?
For the ellipsoidal head, there are two pxsible failure modes, i.e., asymmetric buckling (Per) and plastic collapse (Pc). Therefore, the plastic collapse pressure information should be considered in SSAR.
Response
SSAR and Section 42 have been revised to clarify that failure mode of the top head is represented by knuckle buckling. The other failure modes, including plastic collapse, are bounded by the case for knuckle buckling.
The value of 192 psi in paragraph 42.4.2 should have beer.174 psi. The value of 144 in Table 42-1 should have been 163. Both have been revised due the change in mean value of the actual material yield (see RAI # 1).
Question # 7 In Section 42.4.3, Westinghouse increases 50% critical pressure for the best estimate failure pressure based on N-284 curve which was derived from lower bound of tests. However, there was only one test performed for AP600 containment configuration (M = 14.5). Therefore, it is believed that 50% curve from tests might be appropriate use for AP600 containment. (N 284 does not provide 50% and upper bound curves.) Are test data in N 284 applicable to AP600? They seem to be stiffened spheres.
Also, in NUREG/CR-4209 and -4137, equipment hatch has critical pressure of 3,000 psig and the predicted response of the cover and tensioning ring was elastic up to 360 psig. In this case, only up to 12% of critical pressure is elastic. After that, equipment hatch cover will experience plastic deformation.
5
i NTD-NRC-96-4617 Enclosure 3 Herefore, Westinghouse's claim that the failure pressure is 150% of critical pressure is questionable. See Open Items 3.8.2.4-26 and 19.2.6.3-6.
Equipments hatches are subjected to extemal pressure, not intemal pressure.
Response
ne test data on spherical caps was included in the response to RAI 220.32. This data is representative for a large range of "M" values. The mean value of the data is greater than 1.50 times the Code Case N-284 criteria. The value of 1.50 was used conservatively in the AP600 fragility evaluation. For the 2
AP600, buckling of the hatch covers is elastic.
Question # 8 In Section 42.5, provide the sample CCFP calculations for head at 100 psig. You have constructed the containment failure probability distribution for a particular failure mode by first developing the failure distribution assuming only random error and then developing another distribution assuming only subjective error. He staff believes this me,thod may not be conservative in comparison with the combination of random and subjective errors (Be=Smaterial + 0modelling ) in the left tail region.
Response
Westinghouse will update the fragility analysis using an SRSS combination of random and subjective errors.
Question # 9 In Section 42.6, what is the definition of mean internal pressure? Should it be median pressure? See Open Items 3.8.2.4-27 and 19.2.6.3-7.
Response
Agreed. Revise to median intemal pressure.
Question # 10 In Table 42-1, does " Structural" under COV heading imply " Material"?
4
Response
Agreed. Revise " structural" to " material".
Question # 11 6
NTD-NRC-96-4617 Enclosure 3 In Table 42 2,50% failure pressure for head seems to be around 156 psig. Where does this pressure come from?
Response i Table 42-2 will be revised. Revised methodology outlined in responses to other RAls should resolve this issue.
Ouestion # 12 In Section~42.4, coefficient of variation, not coefficient of variance, should be used.
Response
Agreed. Section 42 has been revised.
Ouestion # 13 In SSAR Subsection 3.8.2.4.2.5, EPAs to be used will be one of those tested by Sandia in NUREG/CR-5334:
D.G. O'Brien: 182.8 C (361*F) and 1,068.7 kPa (155 psia) for 10 days, Westinghouse: 204. 4*C (400*F) and 517.1 kPa (75 psia) for 10 days, Conax: 371.l*C (700 F) and 930.8 kPa (135 psia) for 10 days If Westinghouse EPAs will be used for AP600, they do not satisfy ASME Service Level C limits (90 psig at 400*F). Also, in fragility curve, the dominant failure mode is cylindrical shell with 138 psig at 400*F.
Therefore, if they are used, they control the whole design both in deterministic and probabilistic. The fragility curve for EPAs should be provided.
Response
i SSAR and PRA report have been revised. Based on latest severe accident analyses documented in PRA 0
Section 34, the shell temperature does not exceed design temperature of 280 F. Hence, the qualification for Design Basis plus Sandia tests provides basis that temperature does not affect capacity. Structural evaluation of plate capacity is then sufficient.
Ouestion # 14 In Section 42.4, if the bellow capacity is 90 psig at 400 F, what is probability of failure beyond this pressure? Westinghouse should provide the mean (or median) failure pressure, and uncertainties of geometric properties, modeling, and material for complete CCFP calculations.
Response
PRA report now states that bellows fragility is included in cylinder yield fragility.
7
4 i
4 NTD-NRC-96-4617 Enclosure 3 i
Ouestion # 15 i In Section 3.8.2.4.2.2, the maximum deflection at crown is 15.9" at 174 psig and corresponding strain is 2.5%. Therefore, radius is 15.9/0.025 = 636". Where does this radius come from? The radius, R s, is 1,347.5".
i
' Response In Section 3.8.2.4.2.2, the reported strain is the effective (von Mises) strain. The circumferential and meridional strain at the crown are about one half of the effective strain or 1.25%. The radius of curvature at the crown, calculated based on equilibrium, is 1105" at the internal pressure of 174 psig. This is consistent with the radius of curvature at this pressure in the large displacement non-linear BOSOR-5 analysis.
l 1 j
l I
8
_ ,_