ML031040150
| ML031040150 | |
| Person / Time | |
|---|---|
| Site: | Davis Besse  |
| Issue date: | 07/13/2002 |
| From: | Office of Information Services |
| To: | |
| References | |
| FOIA/PA-2003-0018 SIA-W-DB-01Q-301 | |
| Download: ML031040150 (38) | |
Text
f-Dr 7 ;. t, PRELIMINARY RESPONSES TO NRC STAFF COMMENTS AND QUESTIONS ON DAVIS-BESSE SAFETY SIGNIFICANCE ASSESSMENT (SIA-W-DB-01Q-301) SUBMITTED APRIL 8, 2002
- 4. Question I What is the technical basis of the failure criterion (e.g., strain exceeding 11.15%/o) used to detemine the failure conditions of the cladding layer? Provide specific technical references in the literature that support the failure criterion used in this evaluation.
Response to Question I The strain value of 11.1S%'corresponds to the uniform elongation of the stress-strain curve used in the evaluation. The use of this value as the basis for the failure criteria is based largely on tnginering judgment. The premise is that when any section in the cladding has through-wall strains greater than the uniform elongation, then that section has no more capacity of resisting any additional increase in load. This criterion is judged to be conservative because in reality, there is redistribution of stresses and strains to adjacent elements that would prevent incipient failure when the strains in a particular column of elements exceed this criterion.
Furthermore, it should be noted that the value of uniform elongation used in the evaluation (1 1.15%) is very conservative for stainless steel weld metal. Data obtained from the literature, and summarized in Table I indicates that the average uniform elongation for submerged arc welds (SAW) is 25.7% and that for shielded metal arc welds (SMAW) is 30.7%. The average for both populations is 27.3%. Most of the data shosn in Table I indicate uniform elongation greater than 20% with only two data points below this value.
Subsequent to the publication of SI Calculation W-DB-O1Q-301, the above criterion has been used in a finite element analysis of disk specimens that were burst tested and documented in Reference I. The geometry of the disk test specimens is very similar to the Davis-Besse RPV wastage cavity. The results of the evaluation arc detailed in the Reference 2 calculation (attached). The conclusion from the Reference 2 analysis is that the present criterion used for the evaluation is conservative compared to the disk burst test results. The burst test pressures were also compared to the pressures at which numerical instability occurred during the elastic-plastic analysis and it was found that the instability pressures, although slightly under-predicting the test failure pressures, are a much better predictor of failure pressure than any of the proposed strain-based failure criteria.
Based on the above discussions, it is believed that the use of the 11. 15% uniform elongation as a basis for the failure criterion is very conservative.
Question 2 How does the failure criterion (e.g., based on ultimate strain in a uniaxial tensile test) account for the effects of biaxial loading in the cladding, or triaxial loading in the cladding at the edges of the degradation cavity?
Response to QuestIon 2 The failure criterion was based solely on the uniform elongation and did not consider biaxial or triaxial effects. Nevertheless, as discussed in the response to Question 1, the criterion is conservative compared to burst tests results on test specimens that are a reasonable simile of the wastage cavity geometry.
i.l aC ;;~e ~;..t^fI re,'0om oi Information Act, exerm-ions -!
FOiA- amO-C -OIi .
PRELIMINARY Question 3 The failure criterion applied in SIA report W-DB-OIQ-301 (e.g., the ninimun cross-sectional strain exceeding the failure strain of 11.15%) allows the strain levels in the cladding to exceed the critical strain value entirely through the thickness, leading to very large strains at the surface of the cladding, up to 49%
in Table S ofthe SIA report What is the technical basis for this approach, as opposed to the average cross-sectional strain. or the maximum cross-sectional stirain?
Response to Question 3 Even though the failure criterion used resulted in some elements in the cross-section exceeding the failure strain, the criterion, as compared to actual test data, was found to be conservative (see response to Question No.1).
Question 4 Did you explore a continuum damage mechanics analysis to give guidance of the failure criterion once the strains exceed the critical strain where necking/void growth starts? If not, provide the technical basis for not using a continuum damage mechanics analysis. [Poisson's ratio of 0.5 no longer applies once this critical strain level is exceeded, so the analysis is strictly not valid. (Poisson's ratio is continuously changing as the voids grow at the strains beyond the start of necking.) This results in a stress redistribution that is not accounted for in a standard elastic-plastic analysis.
Response to Question 4 The analysis performed was judged to be conservative as validated by the disk burst test results discussed in the response to Question 1, and as such, it was judged unnecessary to consider the application of continuum damage mechanics analysis to this evaluation.
Question 5 How would the strain values change if the stress free temperature was assumed to be the stress relief temperature instead of 70F, and the analysis accounted for the differential thermal expansion of the cladding and head steel at the operating temperature of 605F?
Response to Question S As can be seen in the SI report W-DBMOIQ-301, and as further clarified by the above responses to Questions I through 4, the strains at the failure pressures from both the analyses and experiments are very large (on the order of 11% or greater). The strains corresponding to thermal expansion effects, at either temperature. arm expected to be much smaller (on the order of 0.1%). Therefore, the effects of changing from a stress free temperature of 7OF to 60SF will not have any significant impact on the results of the analysis Question A Does the size of the degradation cavity and the transition from the cladding thickness to the head thickness that %w-asused in the SIA report reflect current knowledge regarding the cavity geometry, in particular the undercut area described in Figure 13 on page 103 of the Davis-Besse Root Cause Analsis Report (CR2002-089 1), dated April 15, 2002? What is the transition geometry assumed in the analyses?
NGC.02.03b
PRELIMINARY Response to Question A The size of the degradation cavity and the transition from the cladding thickness to the head thickness used in the calculation reflected what was the best available at the time of the calculation. More work is currently in progress on the removed damaged cavity to determine the exact size and geometry of the cavity and the transition regions.
Question B is there sufficient mesh refinement through the cladding thickness to adequately capture the bending and shear strains at the edge of the cavity? Describe any sensitivity studies used to demonstrate the adequacy of the mesh refinement.
Response to Question B In the analysis of the wastage cavity, six elements were used through the thickness of the cladding. A convergence study, using both an axisymmetric model and a three dimensional model was performed in Reference 2 to evaluate the impact of the number of through-wall elements in the thickness of the test specimens. The results indicate that there is no significant difference in the burst pressure predictions when the number of through-wall elements is increased from six to 12. Therefore, it is concluded that the analyses of the wastage categories with six elements through the thickness represents a converged solution.
Furthermore, when fewer elements than six were used in the convergence study, it resulted in conservative estimates of the burst pressures.
Question C Was the cladding deposited by weld wire? Do the thinner cladding thickness measurments from UT coincide with the locations of weld bead toes? In what direction do the cladding weld beads run relative to the long axis of the degradation cavity?
Response to Question C The cladding was deposited by weld wire. It is difficult to determine if the thinner cladding thickness measurements from the UT coincided with the location of the weld bead toes since the UT measurements were taken on one-inch grids and as such, there was not adequate resolution to make such a determination.
It is also difficult to determine the direction of the cladding weld beads from the available information.
Additional investigation of the removed damaged cavity is currently in progress that might provide more information.
NOC42-03r
PRELIMINARY References
- 1) P. C. Riccardella, *'Elastic-Plastic Analysis of Constrained Disk Burst Tests," ASME Paper No.
72-PVP-12, Proceedings of Pressure Vessel and Piping Conference, New Orleans, LA, September 17-21. 1972.
4.
- 2) Stnczral Integrity Calculation W-DB-OIQ-304, Rev. 0, "Evaluation of Failure Criterion Used in Elastic-Plastic Analysis of Davis-Besse RPV Head Wastage."
NGC.02.033
-. * ;44.,--.-.- -- - - - - - - --
PRELIMINARY Table 1: Tensile Test Data for 304 Stainless Steel at 550°F Reference YS ksl UTS ksl Elone % RA % Mati Type NUREG!ICR-6235 20.8 62 38A 70.8 Base
- a. Base NUR.EG/CR-4538 22.2 67.3 39 70.8 NUREG'CR-4538 22.8 68.8 40.5 70.8 Base NUREG/CR-4687 20.1 65.2 53.8 71.3 Base EPRI NP-4768 23.1 613 47 74 Base EPRI NP-4768 24.8 62.6 45 70 Base EPRI NP-4768 33.2 72.7 42 67 Base ASME 72PVP12 34 84 54 75 Base 45.0 71.2
/41 EPRI NP-4668 EPRI NP-4768 44.8 36 Ave.Base 62.9 61.8 22 25 46 67 SAW SAW EPRI NP4768 40.8 70.3 25 69 SAW
. NlREGICR-6098 s 37.4 68 26.4 SAW NUREGiCR-639 49.1 68.1 30 46 SAW NUREG/CR-6389 45 67.1 33 42.4 SAW NUREG/CR-6389 54.3 74 15.5 63 SAW NLUREGICR-6389 1 51.8 71.8 13.7 54 SAW NhUREGICR-4878.' 471 67.6 31.5 44.2 SAW
- NUREGICR48789 28.3 67.5 34.5 47 SAW-Am I W-W- EPRI NP-4668 45.7 AweSAW 65.1 25.7 26 53.2 58 SMAW aT 1 -
EPRI NP-4768 EPRI NP-4768 46.8 49.4 61.4 64.7 37 35 48 46 SMAW SMAW U
1 NlUREG!CR-4878 40.8 70.3 24.8 68.6 SMAW t_ AveSMAW 30.7 55.2
\ NUREG!CR-4538 44.3 65A 33 74.3 Weld 5N
'NUREGCCR-4538 42.2 64.3 30 72.9 Weld Ave.SAW&SMAW 27.3 S3.8 . -
NAWC2.033
PRELIMINARY References
- 1) P. C. Riccardella, "Elastic-Plastic Analysis of Constrained Disk Burst Tests," ASME Paper No.
72-PVP-12, Proceedings of Pressure Vessel and Piping Conference, New Orleans, LA, September 17-21. 1972.
- 2) Stuctural Integrity Calculation W-DB-OIQ-304, Rev. O "Evaluation of Failur Criterion Used in Elastic-Plastic Analysis of Davis-Besse RPV Head Wastage."
NGC.0203S
PRELIMINARY Table 1: Tensile Test Data for 304 Stainless Steel at SSOF Reference YS ksI UTS ksi Elong % RA% Mat%
NUREG!CR-6235 20.8 62 38.4 70.8 Base 4.
NUREG!CRA4538 22.2 67.3 39 70.8 Base NUREG/CRA4538 22.8 68.8 40.5 70.8 Base NUREG/CR-4687 20.1 65.2 53.8 71.3 Base EPR]NP-4768 23.1 61.3 47 74 Base EPRI NP-4768 24.8 62.6 45 70 Base EPRI NP4768 33.2 72.7 42 67 Base ASME 72PVP12 34 84 54 75 Base
_ Ave.Base 45.0 71.2 EPRI NP-4668 44.8 62.9 22 46 SAW II EPRJ NPA4768 36 61.8 25 67 SAW EPRI NP-4768 40.8 70.3 25 69 SAW N- UREGCR-6098 s 37.4 68 26A SAW NUREGiCR-6389 49.1 68.1 30 46 SAW NUREGiCR-6389 45 67.1 33 42.4 SAW NUREGICR-6389 54.3 74 15.5 63 SAW N}UREGICR-6389 51.8 71.8 13.7 54 SAW N1UREGICR4878.' 47.1 67.6 31.5 44.2 SAW NtIREGCR-4878 4 28.3 67.5 34.5 47 SAW-Am AvceSAW 25.7 53.2 EPRI NP-4668 45.7 65.1 26 58 SMAW EPRI NP4768 46.8 61.4 37 48 SMAW
.. -... EPRI NP-4768 49.4 64.7 35 46 SMAW NVUREG'CR4878 40.8 70.3 24.8 68.6 SMAW t AveSMAW 30.7 55.2 NUREG'CR.4S38 44.3 65.4 33 743 Wed
. ZNUREG!CR.4538 422 643 30 72.9 Wld Ave.SAW&SMAW 273 53.8
.a.
N;C.02.03S
PRELIMINARY H STRUCTURAL CALCULATION FILE No: W-DB-O1Q-304 Associates, IPcPACKAGE PROJECT No: W-DB-01 Q PROJECT NAME: Operability and Root Cause Evaluation of the Damage of the Reactor Pressure Vessel Head at Davis-Besse CLIENT: First Energy Corporation ____
CALCULATION TITLE: Evaluation of Failure Criterion Used in Elastic-Plastic Analysis of Davis-Besse RPV Head Wastage PROBLEM STATEMENT OR OBJECTIVE OF THE CALCULATION:
Develop a finite element model to simulate actual test data to evaluate the effectiveness of the failure criteria used in the elastic-plastic stress analysis of Davis-Besse RPV head wastage cavity.
Project Mgr. Preparer(s) &
Document Affected Revision Description Approval Checker(s)
Revision Pages Signature & Signatures &
Date Date 0 1-28 Original Issue Al -A2 Bl -B9 Prqject CD-Rom PAGE .L1 of-28.
PRELIMINARY 1.0 Introduction During recent in-service inspections of the reactor pressure vessel (RPV) head and penetrations at Davis-Besse, significant wastage was observed in the vicinity of control rod drive mechanism (CRDM) No. 3. A calculation package was prepared for First Energy [I] to determine the limiting pressure load of the damaged RPV head.
Based on the review of this calculation package, the NRC raised a number of questions (See Appendix A), the majority of which were concerned with the failure criteria used in the evaluations.
The purpose of this calculation is to develop a better understanding of the failure criteria as used and its relative "conservativeness" in regards to the failure pressure.
2.0 Technical Approach The failure criterion used in Reference 1 was set such that the maximum strain could not exceed the ultimate tensile strain. Hence for the stainless steel cladding where the maximum strain is expected to occur, the maximum equivalent total strain is limited to the maximum strain of 11.15%
(corresponding to the ultimate strain for the stainless steel cladding in Reference 2) through the thickness of the component.
In order to evaluate the reasonableness of this failure criterion, the results of the failure pressures predicted with this criterion were compared against test results of very similar geometries. Disk burst test, similar to the Davis-Besse head wastage geometry were performed under the auspices of the PVRC Subcommittee and documented in and ASME publication [3] (see Appendix B for the actual publication).
Described in Reference 3 were a series of burst tests using machined disks of various materials. The test disk dimensions and the illustration of the test setup are shown in Figure 1.The materials tested included 304 Stainless Steel, A-533 Grade B Low Alloy Steel and A8S Grade C Carbon Steel. For the purposes of this calculation, only the 304 Stainless Steel testing will be reviewed.
As can be seen in Figure 1, three basic disk geometries were tested. In order to evaluate the effectiveness of the failure criteria developed for Reference 1, the same failure criteria will be used to determine the disk burst pressures. As a result, a series of finite element models were developed using the test disk dimension provided in Reference 3. The models were created and evaluated using the ANSYS finite element software [4]. The actual evaluations and subsequent failure criteria comparison arc included in the following sections.
Revision 0
/3P1Parer/Datc RLB 5/3/02 .
Checker/Date
_ File No. NV-DB-01Q304 I Page 2 of 28
PRELIMINARY 3.0 Finite Element Models A series of finite element models were constructed to determine burst pressure for the various disk configurations. Initial studies were performed using an axisymmetric model but subsequent evaluations included three-dimensional modeling similar to that used in Reference 1.
The elastic material properties for all evaluations were for 304 stainless at room temperature as defined by Reference 5. These values used were as follows:
I Modulus of Elasticity, E, el psi: 28.3 Poisson's Ration, v: 0.3 The plastic material properties for stainless per Reference 3 were:
[I] Stress Strain Curve Assumed to be of form a A (C)'
Therefore the stress-strain curve used in all of the evaluation is shown in Table 1. Any additional model specific conditions will be described in the following sections.
ePrepareFr/Date Revision Checker/Date j 0.
4B 5/31/02 :..
_ Filc No. W-DB- iQ-304 i Page 3 of 28
PDIDT IMNAARY Table I Stress Strain Curve for 304 Stainless Steel 131 Strain (in.in) Stress (psi) o.oo0 0 0.025 31208.63
- 4. 0.050 43952A9 0.075 53699.79 0.100 61900.24 0.125 69113.97 0.150 75627.79 0.175 E1611.83 0.200 87176.84 0225 92399.68 0.50 97336.26 0.275 102028.8 0.300 106510 0.325 110805.8 0.350 114937.5 0.375 111922.4 0.400 122775 0.425 126507.5 OA50 130130.6 0.475 133653.1 0.500 137083 0.525 140427.1 0.550 143691.6 0.575 146881.9 0.600 150002.7 0.625 153058A 0.6S0 1560521 0.675 158989.5 0.700 161871.6 0.725 164702.2 0.750 167483.7 0.775 17021S.7 0.S0 172909.5 0.2. 175558 0.850 173166.2 0.875 18073S.8 0.900 183263.6 0.925 185766 0.9S0 138229.5 0.975 190660A 1.000 193060 1.025 195429.4 1.050 197769.7 1.075 200082 1.100 202367.3 1.125 204626A 1.150 206360.2 1.175 209069.7 1.0 2112S5A
.23 213418.2 Revision O Pfcparcr/Datc RLB 5/31/02 I p Checkcr/Date i I File No. W-DPgOIQ-304 Pag 4 of 28
PRELIMINARY 3.1 Axisymmetric Finite Element Model The axisymmetric models were developed in ANSYS using the 2-D 8-Node Structural Solid element, PLANE82. All three geometries descnbed in Reference 3 were evaluated as was the effects of the finite element mesh density on the onset of numeric instability. A total of 5 evaluations for each disk geometry were made, the only difference between each evaluation was the mesh density, which can be simplified to the number of elements through the thickness of the thinned portion of the disk. As such, the mesh densities that were evaluated where 4, 6, 8, 10 and 12 elements through the thickness. Figure 2 shows the progression of mesh density for geometry-A.
The mechanical boundary conditions for these evaluations consisted of simple vertical restraint throughout the approximate clamp region. This region was assumed to into the entire region of the disk, which remained at the full I inch thickness. See Figure 3 for an example of the applied boundary conditions on the 4 element through thickness, geometry-A model.
3.2 Three-Dimensional Finite Element Model The three-dimensional models were developed in ANSYS using the 3-D 8-Node Structural Solid element, SOLID45. All three geometries described in Reference 3 were evaluated as was the effects of the finite element mesh density on the onset of numeric instability.
Only a 300 section of the total disk was modeled since the loading and geometries were also symmetrical. Two evaluations for each disk geometry were made; the only difference between each evaluation was the mesh density, which again can be simplified to the number of elements through the thickness of the thinned portion of the disk. As such, the mesh densities for the 3-dimensional models that were evaluated were 4 and 6 elements through the thickness. It should be noted that the stainless clad for the actual Davis-Besse cavity evaluation used 6 elements through the thickness. Figure 4 shows the two mesh densities for geometry-A.
The mechanical boundary conditions for these evaluations used the same vertical restraints as the axisymnietric evaluations. In addition, axisymmetric boundary conditions were applied to the free ends of the disk, the preventing translations in the circumferential direction. This results in the centerline of nodes being limited to translation in only the vertical direction See Figure 5 for an example of the applied boundary conditions on the 4 element through thickness, geometry-A model.
4.0 Loading All of the evaluations were loaded in the same manner. An incremental pressure was applied to the cavity surfaces until numeric instability was reached. See Figure 6 for an example of the applied pressure.
Revision 0 t 3 _
Preparer/Date Checker/Datc RLB S31/02 Filc No. W-DB-0lQ-304 a
PRELMINARY 5.0 Mesh Density Results For each evaluation, the pressure was allowed to rise incrementally until numeric instability occurred. The points of instability, as compared to the actual disk burst tests, are shown in Table 2.
Table 2 z.
Mesh Density Effects of Numeric Instability Model Pressure (psi) Predicted/
Model Type Through-Wall Numeric Actual Test Test Result ModelType__ Elements Instability Burst _
r-eai-1.X tm :r ig Axisvmrnetric 4 12725 84.8 Axisymmetric 6 13942 92.9 Axisymmetric 8 14004 93A Axisvmmetric 10 14022 15000 93.5 Axisymnmetric 12 14005 93.4 3-Dimensional 4 13979 93.2 3-Dimensional 6 13997 93.3 Axisymmetric 4 5929 87.2 Axisynunetric 6 6630 97.5 Axisymmetric 8 6695 98.5 Axisymmetric 10 6695 6800 98.5 Axisvmmetric 12 6694 98.4 3-Dimensional 4 6688 98.4 3-Dimensional 6 6671 98.1 Axisymmetric 4 6317 82.0 Axisymmetric 6 6962 90.4 Axisyrnmetric 8 6997 90.9 Axisymnetric 10 6998 7700 90.9 AxisM etric 12 6997 90.9 3-Dimensional 4 6976 90.6 3-Dimensional 6 6974 _ _.6 The results are also shown graphically in Figure 7.
Revision Preparer/Date RLB 5/31/02 . -..
Checker/Date File No. W-DB-01 Q-304 Page 6 of 28
PRELIMINARY 6.0 Total Strain Results Based on Section 5, only the highest through-wall element count cases will be further evaluated. As a result, Figures 8 though 10 show the total Von Mises Strain just prior to onset of instability for the 12 through-wall element axisymmetric model and Figures 11 through 13 show total Von Mises Strain for the 6 through-wall element 3-D model.
4.
7.0 Strain Criteria Comparison The original failure strain criterion described in Section 2.0 indicated that when the through-wall total strain exceeded the uniform elongation percentage, the structure would be considered to have failed. As a check of this criterion, the total Von Mises nodal strains as they varied with pressure were extracted from the middle of the modeled disk at the top, middle and bottom of the wall thickness. The resulting strains were then plotted versus the pressure and compared to the actual burst pressure measured in Reference 3 and the failure pressure as defined by the Failure Criterion in Section 2.0.
From the definition of material properties used in the disk burst test, the uniform elongation for 304 stainless steel was 54% (see Section 3.0). Therefore, the failure of the disk will occur when the through-wall total strain exceeds 54% throughout the thickness.
An examination of the 3 geometries for both the axisymmetric and 3-D modeling can be seen in Figures 14 though 19. The results are further summarized in Table 3.
Table 3 Failure Criteria Comparison Model Model Failure Pressure l, SO Type Geometry Burst Test 131 Instability Failure Criteria Axisymmetric A 15000 14005 -11000 Axisymmetric B 6800 6694 -5500 Axisymmetric C 7700 6997 -5750 3-Dimensional A 15000 13997 -11000 3-Dimensional B 6800 6671 -5500 3-Dimensional C 7700 6974 -5750 Revision 0 Preparer/Datc RIB 5/31/02 _. _ -
_ FilCheckNraDate . 2 iFile No. W-DB OIQ-304 IPage 7 of 28
PRELIMINARY 8.0 Conclusions Based on the summary in Table 3 of Section 7.0, the use of the uniform elongation limit as the basis of failure criteria in an elastic-plastic finite element analysis results in conservative failure pressures as compared to actual test results. For the three geometries, the uniform elongation criteria predicted A.I a failure pressure that was in the range of 73% to 84% of the actual failure pressure.
A better prediction of actual failure pressure is the pressure at which numeric instability was reached in the ANSYS program. Assuming a numeric instability criterion, failure pressure would range from 90% to 98% of actual failure pressure.
Revision 0
_______a _ _.eRLB 5s31/02 .
__ _Checker/Date
_ File No. W-DB-01Q-304 Page 8 of 28
PRELIMINARY 9.0 References I) Structural Integrity Calculation W-DB-01Q-301, Rev. 1, "Elastic-Plastic Finite Element Stress Analysis of Davis-Besse RPV Head Wastage Cavity."'
- 2) Email of from B.R. Grambau (Framatome ANP) to N. Cofie (Sl), "308 Stress -Strain 4.
Curve," March 15, 2002, Si File W-DB-01Q-202.
- 3) P. C. Riccardella, "Elastic-Plastic Analysis of Constrained Disk Burst Tests," ASME Paper No. 72-PVP-12, Proceedings of Pressure Vessel and Piping Conference, New Orleans, LA, September 17-21, 1972.
- 4) ANSYS/Mechanical, Revision 5.7. ANSYS Inc., December 2000 Revision Preparer/Date Re nii Checker/Date f0 iFile No. W-DB-01 Q-304 0
RLB 5/31/02 2:. __ _ __ _ __
P Page 9 of 28
PRELIMINARY
- STRAIN.GAGED CANTILEVER BEAM FOR CENTRAL DEFLEC-
- T7ON MEASUREMENT SCHEMATIC ILLUSTRATION OF TEST SETUP I in GEMER GEOMETRYI THICKNESS
() FILLET RADUS A 025in. 0375in.
B 0125 in. 0.125in.
C 0.125 F. 0.375.
O5RO Figure I - PVRC Disk Test Details (Reference 3)
Revision i.
_ Fieparer/DateRLB 5/31102 .g 1 gV~Checker/Date__
Fie No. W-DB OIQ-304 Page 10 of 28
I I. i I
i
-- - - . . .. --- -- -- I . . .- .--- - . . - . - . - . . .. . . . . . -... ... .
PRELIMINARY r
I I I I I, ,I ,I I I I,.I I I-11 i I I ... . . .,,,,, I I I I I,,I I I I I I I:Z7>oT 1r soxsr
+_
_%I~X~tA- , , , I I I, II NII_O_,,,,,,,
CCX 1L_ I III II I
Figure 3 - Mechanical Boundary Conditions Example for Axisymmetric Finite Element Model for Geometry-A Revision _________
I Preparer/Date IB 5/31/02. _ _ - . . __ _ -_
Checker/Date _____ __I l File No. W-DBLOIQ-304 Page 12 of 28
PRELIMINARY z 1:
'1.I
.81 I Eu;:
4 Element Through-Wall
- If I..-ll 6 Element Through-Wall ' * .
Figure 4 - Mesh Density Examplc for 3-D Finite Element Model for Geometry-A Revision 0 Preparer/Date RLI 5131/02 -. __ _ _.
Checker/Date File No. W-DB 01 Q-304 _1 ~ -_ Page 13 of 28
Figure 5- Mechanical Boundary Conditions Example for 3-D Finite Element Model for Geometry-A
PRELIMINARY Figure 6 - Applied Prcssure Example (Axisymmetric Geometry-A Model)
R 5/31/02.
5rp/3atc -- - ._. .
Checker/Date
_ File No. W-DB-01Q-304 Page 15 of 28
PRPIUThNARY
_s. of N . _
. _. Refinement .
Mesh Refinement vs. Onset of Numeric Instability Pressure 15000 14000 4.
13000 12000 11X00 E
91000 8000 7000 6000 4 5 6 7 8 9 10 11 12 Through-Wall Element Count Figure 7 - Mesh Density Effects Revision 0 Preparer/Date RLB 5/31/02 .I . . P
_ChFker/DateBIII
-F-ileNo.W-DB OIQ-304 Iae1 f2
PREI-rMINARY UNSYSY 5.72 PLOT NO. I NODAL SOIZJTIC rDlXE<.933654 E2TOEQV QP.VG)
EffNu=O SMN =.472E-03 5t:: =1216
.1347
£1354045 97 Em tIi LXX Von Mises Total Strain (12x96) - (Category-A)
Figure 8 - Total Von Mises Strain Just Prior to Numneric nstability Geometry-A - Axisymmetric Revision 0 C 3 Preparer/Dat Checker/Date R -531/02 File No. W-DB-OIQ-304 I Page 17 of 28
PRELIMINARY mSW§21/25O2 08047 PLOT NO. I NODAL SOLUTION STEP TIE A4627 EPOEQV- PRVG)
EfiSu
& =4863E-03 ac lA IdWl OX Von Mises Total Strain - 296) - (Category B) .
Figure 9 - Total Von Mises Strain Just Prior to Numeric Instability Geometry-B - Axisymmetric Revision 1 Preparer/Date RLB 5/31102 ._ . . e ::
iWChccker/Date
_ File No. W-DB-OIQ-304 I _ Page_ I_ of 28_
ANSYS 5.7 MAY 22 2002 08:26:25 PI NOD. I RODALL SOWIMC TD(SE'A6W47 Ef~%
Von Mises Tota Straln - OW6) - (CAteqosY-C Figure 10 - Total Von Mises Strain Just Prior to Numeric Instabilityr Geometry-C - Axisymmetric
PRELIMINARY ANSYS 5.7 MAY 22 2002 090M111 PLOT NO. I NODL SO=xMz iT 1 32 EPTOEQ Effl4t? QGVG)
. TOP MM -2296 891. 85
.778079 FMsX Von Kises Tota 1 Strain - (6x48) - (Category-A)
Figure II - Total Von Mises Strain Just Prior to Numeric Instability Geometry-A Dimensional Revision _
- s. Preparer/Date RLB 5/31/02 . .
Checker/Date
_ File No. W-DB IQ-304Page 20 of 28
PRELIMINARY ANSYS 5.7 M22002 PLOT NO. 1 NODAL SOIMICON
- 4. DM~1-2245 AQ gm e X i~' -
Von Mises Sotal Strain - (6x48) - (Cftegory-B)
Figure 12 - Total Von Mises Strain Just Prior to Numeric Instability Geometry-B Dimensional Revision 0 Preparer/Date RIB 5/31/02 .- ' . ..
T Checker/Date _
_ FileNo. W-DB-0IQ-304 Page 21 of 28
PRELIMINARY O9:14:07 PLOT NO. 1 NODAL SOLTION STEP-1 SUB '28 T3M=A64911 EPTOEV VG Ef flqu=OVG
- 4. TOP DaX -2157 as -=.14E-03 w~ 1.11403 fil Von Kies tot al Straisn - (6x48) - (Category-C)
Figure 13 - Total Von Mises Strain Just Prior to Numeric bstability Geomntzy-C Dimensional Revision Preparer/Date RLB 5/31102 I i3Checker/Date
_ File No. W-DB-o IQ-304 Page 22 of 28
PRELIMINARY Pressure vs Total Von Mises Strain at Center Geometry-A (Axisymmotric) 16000 Test Burst Pressure a 15.000 psI .
14000 12000
- a. 10000 i 8000 6000 4000 i
2000 0
0 02 OA 0.6 0.8 1 12 1.4 I Total Von Min Strain pnriln)
Figure 14 - Through-Wall Strain Results at Center of Disk Geometry-A (Axisymmetric)
Revision . A Preparer/Date R t -
Checker/Date
_ File No. W-DB-OIQ-304 - Page 23 of 28
PRELIMINARY I Pressure vs Total Von Mlses Strain at Center I
I Geometry-B (AxIsymmetric) 7000 I "fst Te Lust' Presur t6w0O psiI _ ^
z 6000 I
5000 ql 4000 i I I 13000 II 2000 1000 0
0 02 OA 0.6 0.8 1 1.2 Total Van Mines Sbtain (nlin)
Figure I5- Through-Wall Strain Results at Center of Disk Geometry-B (Axisynmetric)
Revision 0
.Prparer/Date .B 5/31t02 . . .. ._.
Checker/Date
_ File No. W-DB-OIQ-304 Page 24 of 28
PRELIMINARY I
Pressure vs Total Von Mises Strain at Center Geometry-C (Axisymmetric)
I ouU I.........................................i............................;.........
Test Burst Pressure m7,700 ps ja .
- 4. 7000 6000
- a. 5000 I 4000 i 3000 2000 1000 0
0 0.2 0.4 0.6 0.8 I 1.2 L Total Von Muss Straln Onln)
Figure 16 - Through-Wall Strain Results at Center of Disk Geometry-C (Axisymmetric)
Revision 0 _
Preparer/Date RLB S/31/02 -_. . . .
Checker/Date I
_ File No. W-DB-01Q-304 i Page 25 of 28
PRELIMINARY Pressure vs Total Von Mises Strain at Center i Geometry-A (3-D1menslonal)
! 16000 Test Burst Pressure x 15.000 psi
........................................... t...........
I.................................................
14000 _ _ _ _ _ ___ . .
0--
12000 ..
Am J 8000 .. _ _-
,-- , ,= F ;
-- Middle 4000 .
-- Bottom _
-'-Boto 2000 _ _
i O I
0 0.2 0.4 0.6 0.8 1 12 Total Von lises Strain fnlin)
Figure 17 - Through-Wall Strain Results at Center of Disk Geometry-A (3-Dimensional)
Revision 0 _ __
Preparer/Date RLB 5/31/02' Checker/Date I File No. W-DB-OIQ-304 Page 26 of 28
p PRELIMINARY i Pressure vs Total Von Mises Strain at Center Geometry-B (3-Dlmensional)
I 7000 . u........s.r................
~... ; . .. _ i~ ps .I , -_
A. 6000 5000 .
4000
- I{ . _
I 3000 Jt .. _ . ...
Houl V
i -Top II 2000 Middle I
1000
-- Jn!!
D.h
- Bottom i
0'! I i 0 02 0.4 0.6 40.8 I 11.2 I Total Von Mass Strain lfnln)
Figure 18 - Through-Wall Strain Results at Center of Disk Geometry-B (3-Dimensional)
Revision 0 Preparer/Date RLB 5/31/02 . . ._._ .
Checker/Date I File No. W-DB-01Q-304 Page 27 of 28
In PRELIMINARY Pressure vs Total Von Mises Straln at Center Geometry-C (3-Dimensional) 8000 t t r Y Presur s Y5ZW TestBurs si .........................................................
I 7000 4.
6000 - .I 5W00 I 4000 3000 2000 1000 0
0 0.2 .4A 0.6 0.6 I 1.2 I Total Von 11"s Strain (lnfln)
Figure 19 - Through-Wall Strain Results at Center of Disk Geometry-C (3-Dimensional)
Revision _
Preparer/Date RLB 5/31/02 ._._.__ I Checker/Date .
_ File No. W-DB-OIQ-304 Page 28 of 28
PRELIMINARY APPENDIX A NRC Staff Comments and Ouestions on Davis-Besses Safety Significance Assessment (SIA-W-DB-01O-301) Submitted Avril 8. 2002 e
Revision Preparer/Date 10_
RLW 51/02 . . j
__ FilCheckerN DateP File No, WV-DB 01Q304 IPage Al of A2!
PRELIMINARY NRC STAFF COMMENTS AND QllESTliONS ON DAVIS-BESSE SAFETY SIGNIFICAN-CE ASSESSMENT (SIA-W-DBQ-01-301) SUBMITTED APRIL 8, 2002 FAILURE CRITERION (1) What is the technical basis of the failure criterion (e.g., strain exceeding 11.15%) used to determine the failure conditions of the cladding layer? Provide specific technical references In the literature that support the failure criterion used in this evaluation.
(2) How does the failure criterion (e.g., based on ultimate strain In a unlaxial tensile test) account for the effects of biaxial loading In the cladding, or triaxial loading In the cladding at the edges of the degradation cavity?
(3) The failure criterion applied In SIA report W-DB-01Q-301 (e.g., the minimum cross-sectional strain exceeding the failure strain of 11.15%) allows the strain levels In the cladding to exceed the critical strain value entirely through the thickness, leading to very large strains at the surface of the cladding, up to 49% In Table 5 of the SIA report. What Is the technical basis for this approach, as opposed to the average cross-sectional strain, or the maximum cross-sectional strain?
(4) Did you explore a continuum damage mechanics analysis to give guidance of the failure criterion once the strains exceed the critical strain where necking/void growth starts? If not, provide the technical basis for not using a continuum damage mechanics analysis.
[Polsson's ratio of 0.5 no longer applies once this critical strain level Is exceeded, so the analysis is strictly not valid. (Poisson's ratio Is continuously changing as the voids grow at the strains beyond the start of necking.) This results in a stress redistribution that is not accounted for In a standard elastio-plastic analysis.)
(5) How would the strain values change If the stress free temperature was assumed to be the stress relief temperature instead of 700 F, and the analysis accounted for the differential thermal expansion of the cladding and head steel at the operating temperature of 605°F?
GEOMETRY/MESHING (A) Does the size of the degradation cavity and the transition from the cladding thickness to the head thickness that was used in the SIA report reflect current knowledge regarding the cavity geometry, In particular the undercut area described In Figure 13 on page 103 of the Davis-Besse Root Cause Analysis Report (CR2002-0891). dated April 15. 2002? What is the transition geometry assumed In the analyses?
(B) Is there sufficient mesh refinement through the cladding thickness to adequately capture the bending and shear strains at the edge of the cavity? Describe any sensitivity studies used to demonstrate the adequacy of the mesh refinement.
(C) Was the cladding deposited by weld wire? Do the thinner cladding thickness measurements from UT coincide with the locations of weld bead toes? In what direction do the cladding weld beads run relative to the long axis of the degradation cavity?
Revision 0 Pre~parer/Date RLB 5/3-1/0 . .. . -. 0:
Checker/Date I I I File No. W-DB-O1Q-304 I Page A2 of A2
PRELIMINARY APPENDIX' B Pressure Vessels and PiRing Division Paper No. 72-PVP-1 2. "Elasto-Plastic Analysis of Constrained Disk Burst Tests" Revision 0 7 Preparer/Date RLB 5/31/02 I____________
_F CheckerN Date o IFile No. W-DB-01Q-304 i Page BI of B9