HL-4781, Submits Addl Info Requested by NRC Re Recent Core Shroud Mod on Plant Unit 1
| ML20080H884 | |
| Person / Time | |
|---|---|
| Site: | Hatch  |
| Issue date: | 02/20/1995 |
| From: | Beckham J GEORGIA POWER CO. |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| Shared Package | |
| ML20080H886 | List: |
| References | |
| HL-4781, TAC-M91091, NUDOCS 9502240062 | |
| Download: ML20080H884 (11) | |
Text
'
Georgia Power Company
~ 40 invemass Center Parkway
, " '. Post Office Box 1295
'~
.,+
Birmingham, Alabama 35201 -
i'
. Telephone 205 877 7279
)
't
\\
Georgia Power J. T. Beckham, Jr.
Vice Presi t - Nuclear g g, g p Febniary 20, 1995 Docket No. 50-321 HI 4781 TAC No.
M91091 U.S. Nuclear Regulatory Commission ATTN. Document Control Desk Washington, D.C. 20555 Edwin I. Hatch Nuclear Plant - Unit 1 Response to Request for Additional Information Regardina Core Shroud Modification Gentlemen:
By letter dated January 19,1995, the Nuclear Regulatory Commission (NRC) staff requested Georgia Power Company (GPC) to provide additional information regarding the core shroud modification recently installed on Unit 1. The enclosure provides GPC's response.
Please be advised that the attached General Electric documents do not _ represent proprietary information. Consequently, this information may be placed in the public document room.
Should you have any questions in this regard, please contact this office.
Sincerely,
):
<N
$~ ~
J. T. Beckham, Jr.
j
.i JKB/eb
Enclosure:
Request for AdditionalInformation Regarding Core Shroud Modification I
Attachments: (See next page.)
" *
- Q' 1 r, *('.
~
l I
950224oo62 95o220 DR ADOCK 0500o321 p
GeorgiaPower A U.S. Nuclear Regulatory Commission Page 2 February 20, 1995 Attachments:
i 4
- 1. GENE-771-39-0794, Revision 1, dated September 26,1994
- 2. Supplement to GENE-771-39-0794, dated February 15,1995 1
cc; Georzia Power Comvany 1
Mr. N. L. Sumner, Nuclear Plant General Manager.
l NORMS (w/o attachment)
U.S. Nuclear Rentatory Commissim Washineta D.C.
Mr. K. Jabbour, Licensing Project Manager - Hatch U.S. Nuclear Regulatory Commission. Rezion H Mr. S. D. Ebneter, Regional Administrator Mr. B. L. Holbrook, Senior Resident Inspector - Hatch HL-4781
~.
s m -
Enclosure EdwinI. Hatch Nuclear Plant - Unit 1 Response to Request for Additional Information Renar&no Core Shroud Modification
Background
j.
By letter dated January 19,1995, the Nuclear Rf =*~y Commission (NRC) Staff requested Georgia Power Company (GPC) to provide additional information regarding the core shroud modification recently installed on Unit 1. The NRC questions and GPC's responses are provided below:
- 1. NRC Question If welds H2 and H3 fail 360* throughwall during normal plant operation, the preload on the tie rods at Hatch Unit I may be insufficient to prevent separation at the H6B, H7 or H8 weld locations if any of these welds also fail 360 throughwall during normal plant operation. A gap of about 0.08 to 0.10 inches can develop at these lower weld locations during normal operation. Therefore, please provide the analyses, hiiption of the model used, and support calculations to estimate the size ofgaps which are expected to develop during normal operation and a main steam line break accident for postulated failures of these weids.
GPC Resnonse GPC's evaluation of the stated technical concern concluded that a small gap is calculated to occur as a result of the crack scenario A small gap of approximately 0.008 inches is calculated. This value bounds all load cases corresponding to the current licensed power and core flow. The gap results from a reduction in the total preload on the tie rods, caused by the 360 degree throughwall cracking at H2 and H3.
As a result of this cracking, the total preload is slightly less than the resultant uplift force in the shroud during normal operating conditions. As a result of these new evaluations, GPC has determined that the previous conclusions relative to the '
prevention of upward motion of the shroud during normal operation for all crack cases is inaccurate. This error occurred due to a failure to properly account for the loss of preload in the tie rods resulting from failure of welds H2 and H3 during the design of the repair. As a result of the error, the repair currently does not meet the Boiling i
Water Reactor Vessel Internals Project's criteria for no separation during normal operation. GPC has completed an evaluation of the loss of preload'and the corresponding small gap and concludM that there is no impact on any accident or transient analysis and no impact to normal operation. Also, the safety design basis of the shroud repair is not adversely affected. The gap does not inhibit the ability of the repaired core shroud to perform its safety functions and to meet its power generation objectives. The evaluation considered the increased leakage from the gap, the change HL-4781 E-1
L W for AdditionalInformation l
Regarding Core Shroud Modification to assumptions in the seismic analysis, and conformance with the design basis. A l
discussion of the evaluations performed by GPC and General Electric follow:
The shroud preload is primarily applied by differential thermal expansion between the l
tie rod assembly and the shroud. The alloy X-750 lower spring and upper bracket, which are part of the overall tie rod load path, have a smaller coefficient of thermal expansion than the 304 stainless steel shroud In addition, a small mechanical preload was applied to each tie rod with a mechanical torque during installation of the tie rod.
If all horizontal welds are intact, the preload is greater than the net vertical upliA forces applied to the shroud during normal operation.
The magnitude of the preload in the tie rods and shroud is a function of the stiffnesses of the tie rods and the shroud. The dffnes of an uncracked Hatch Unit I shroud is 3.6E7 pounds per inch. GENF-7 /1-39-0794, " Hatch Unit 1 Shroud Repair Hardware i
Stress Analysis," Revision 1, states that the stiffness of the set of four tie rods is 1.94E6 pounds per inch. GENE-771-39-0794, Revision 1 is provided as
'. This combination of an uncracked shroud with the rods results in a preload of 95 percent of the maximum possible pieload, which would result for a rigid shroud.
The NRC question postulates the complete failure of three horizontal welds (H2, H3, and either H6B, H7 or H8). The combined failure of H2 and H3 reduces the axial stiffness of the shroud. The failure of a weld below the core plate would then transfer j
the upward force due to the pressure drop across the core plate through the shroud to the tie rods. This postulated condition was evaluated using the same computer program as used in the initial stress analysis. ' ' A finite element model of the shroud was created using three dimensional brick elements. Welds H2 and H3 were W by connecting a single row of nodes ir the shell to the ring to simulate the postulated 1
cracks. The weld joint configuration and assumed failure of the shell and the ring at H2 and H3 was modeled considering the metallurgy at those welds. A detailed j
discussion follows:
i Boiling Water Resctor (BWR) shroud horizontal weld heat affected zone (HAZ) related intergranular stress corrosion cracking (IGSCC) has been most extensive when it occurs in type 304 stainless steel plate (rather than forged) rings at the various ring i
to shell welds. In all cases where the ring to shell weld cracking has exceeded s
approximately 180 degrees of circumference, the cracking has occurred in the plate rings rather than the adjacent shell. This is expected to be the case at the top guide and core plate rings.
l The cracking pattern observed in other operating BWRs is consistent with the understanding that cracking occurring in sensitized type 304 stainless steel plate weld HAZs is exacerbated by the presence of surfaces having the short transverse HL-4781 E-2
5
. Request for AdditionalInformation Regarding Core Shroud Modification c
f orientation exposed to the coolant. The acciising effect of this orientation on crack initiation in sensitized type 304 stainless steel has been previously documented based
)
on laboratory testing and has been attributed to the presence of elongated inclusions which occur during fabrication in the plate rolling operation. Where these elongated indneinns intersect the ring surfaces they can act as IGSCC initiating sites (crevices and/or stress risers) by forming pits and/or corrosion tunnels through dissolution or h
micro fracture. In addition, the ring surfaces at the ring outer and inner radii tend to develop layers of relatively thick (approximately 5-30 mils) cold work during machining to final size, which also facilities crack initiation. Both these effects have 4
been seen during evaluation of boat samples removed from cracked shroud welds at l
several plants and together with sensitization, they provide the conditions leading to the field cases where greater than 180 degrees of cracking has been observed at the ring to shell welds.
Although surface cold work may also be present in rings fabricated from forging, the i
inclusions tend to be elongated in the directions of maximum metal flow and they -
l normally do not intersect perpendicular to the ring outer and inner surfaces. -This is also true for elongated inclusions present in the shroud shell material although the extent and depth of surface cold work present would be expected to be considerably r
less than for the ring inner and outer surfaces.
^
In addition to the shroud weld HAZ IGSCC seen in plants with type 304 stainless material, in two cases, significant but much less than 180 degrees, cracking has been' seen at ring to shell welds in plants with type 304L shroud materials. For these cases, l
l no HAZ sensitization is expected and crack initiation is likely' driven by the presence of surface cold work. lone, since the neutron fluence tends to be relatively low at the various ring locations and Irradiation Assisted Stress Corrosion Cracking (IASCC) would not be expected. Evaluation of a boat sample from an overseas BWR with
[
304L shroud cracking indicated that the cracking did initiate in a surface cold worked layer (present at the shell HAZ inner surface) and propagated into the unsensitized E
substrate. For the unsensitized low neutron fluence case, crack growth rates are
[
expected to be significantly lower than for sensitized material, especially considering E
the relatively low coolant conductivity levels at which essentially all BWR's are currently operating.
Based on the above metallurgical discussion, the most likely location for extensive cracking at H2 and H3 is in the welded plate ring and not in the 304 shell. Therefore, the crack locations were modeled at the top surface of the ring for H2 and the bottom surface for H3. Both welds H2 and H3 are full penetration groove welds with a large 1.00 x.75 inch fillet on one side. Figure 1 shows the welds and the assumed cracks.
The point of connection between the shells and the ring was chosen as the natural pivot point for a through wall crack. This natural pivot point is the toe of the fillet welds. No tensile load carrying ability was included for any of the groove or fillet L
HL-4781 E-3 w
.+
Request for AdditionalInformation j
Regarding Core Shroud Modification weld in H2 or H3. Therefore, at H2 the nodes that were connected were the innermost nodes of H2 and at H3 the outermost nodes of H3 were connected. A uniform load was applied to the top of the shroud and the bottom of the shroud was restrained in the vertical direction to calculate the vertical stiffness. Because of symmetry, only 90 degrees of the shroud was modeled.
3 In addition to welds H2 and H3, there are two other features which stiffen the ring between H2 and H3. These are the top guide aligner brackets and the upper stabilizer springs themselves. W upper stabilizer springs are initially preloaded against the top guide ring in the radial inward direction. They also will resist ring outward motion with a spring constant of 20,000 pounds per inch. However, for conservatism, both the stabilizers and the aligners were neglected in the final stiffhess calculation.
If the aligner bracket welds are assumed to be completely cracked so that they can not carry any tensile load, then the axial stiffness of the entire shroud, with the postulated
+
cracking, was calculated from this model to be 1.13E7 pounds per inch. The total axial stiffness of the four tie rod assemblies is 1.94E6 pounds per inch. These stiffnesses yield a thermal preload value of 185,000 pounds, which is 85 percent of the maximum possible. h mechanical preload is 13,600 pounds.
A calculation was performed with the above values of preload plus the normal operating forces of weight and differential pressure. This calculation was for the currently licensed power and core flow, h result of this calculation for different assumed shroud cracks show a small gap of approximately 0.008 inches. Table 1 provides the results.
There are three potential concerns for separation at a weld during normal operation.
These concerns are; increased leakage, changes to assumptions in the seismic analysis, and conformance with the design basis. Each of these is discussed below.
The safety evaluation contained in GENE-771-42-0894, Safety Evaluation for Installation of Stabilizers on the Hatch Unit 1 Core Shroud, Revision 1,is based on leakage through eight cracks each 0.001 inch wide. This document shows that the assumed leaks have no impact on plant safety or operation. h cumulative gap is within the recent calculated shroud gap of 0.008 inches. This leakage is single phase (below core plate) and would have no impact on steam separating system performance, jet pump performance or anticipated abnormal transients. There would be no effect on ECCS because the gap would close as core flow decreased.
Consequently, there is no impact on plant operation.
i The calculated gap is at or below approximately 0.008 inches. - The seismic analysis assumed that tight cracks through the shroud would best be represented by a hinge
[
which carries shear but no moment. All of the gaps in Table 1 are small enough, that a hinge isjudged to be the best boundary condition. Large gaps such as that which i
i HL-4781 E-4
l.
l Request for Additional Information I
Regarding Core Shroud Modification momentarily exists aAer a main steam line Loss of Coolant Aen(LOCA) were modeled as a roller, which could not carry shear or moment The imunding condition in the Safety Evaluation which opened a large gap was the main steam line LOCA.
This bounding condition remains unchanged. The combination of a LOCA and a Design Basis Earthquake is a faulted event. The faulted load case is not changed by a i
small normal operating gap. To assure that the' desi n bounds all possible seismic I
8 conditions, an Operating Basis Earthquake (OBE) analysis _was performed with the H7
)
weld assumed to be a roller. This case represents the limiting case. This analysis
. showed that all stresses and displaces in the shroud and stabilizers are still within the allowable values.
The structural analyses discussed above are further documented in a supplement to GENE-771-39-0794, Revision 1. The supplement is provided as Attachment 2. The behavior of the shroud during a main steam line LOCA is not changed by cracks at H2 1
l and/or H3. As stated in the Safety Evaluation, the tie rods clastically stretch during the LOCA. AAer a few seconds, the loads are reduced and the shroud gap is closed.
In summary, the existence of the calculated gap has no adverse consequences on plant operation or plant performance in design bases accidents or transients. The conclusions reached in GENE-771-39-0794 remain valid with the exception of the postulated small upliA under extremely unlikely conditions. This small uplia has no adverse effect on plant safety. Since the existence of the small gap has no impact on i
any accident or transient analysis or normal operation, GPC is currently evaluating options related to achieving conformance with the criteria established by the Boiling Water Reactor Vessel Internals Project. GPC will provide timely notification of the j
results of these evaluations to the NRC. Additionally, the issue of conformance with j
the industry repair criteria will be resolved prior to implementation of Power Uprate, Notification relative to the need for additional actions or modifications, if any, will be i
made prior to the next scheduled Unit One refueling outage currently scheduled to begin on approximately March 20,1996.
HL-4781 E-5
i l ' '-
Request for AdditionalInformation Regarding Core Shroud Modification Table 1 Summary of Calculation of Gans Load Case Calculated Gap Size Inches (Approximate)
Cracked Power / Flow H6B H8 Welds
- 100/105 H2,H3,H6B X
0.008 N/A H2,H3,H6B,H8 X
0.008 N/A H2,H3,H8 X
N/A 0.0 (1) Identified welds are assumed to be cracked through wall for 360 degrees.
(2) The calculated gap size for the load case with H2, H3, and H7 cracked is a gap of 0.001 inches at H7.
o e
HL-4781 E-6 i
Request for AdditionalInformati1n K
Regarding Core Shroud Modification WELD GEDMETRY MODEL
.i CRACK LOCATIONS Y
1 i
l l
l FIGURE 1 HL-4781 E-7
~
~
e i
Request for AdditionalInformation Regarding Core Shroud Modification
- 2. NRC Question There appears to be an arithmetical error in the cahd=* ion of the tie rod stress that could be encountered during a cold feedwater transient. (Refer to lines 23 and 2 on pages 40 and 41, respectively, Shroud Repair Hardware Stress Analysis, GENE-771-39-0794). If the tie rod preload to maintain an acceptable gap at weld location H6B in Item I, above, is determined to be higher than the existing preload, then the combined stress due to the higher initial cold preload and thermal loads during the cold feedwater transient could result in excMag the allowable stress in the tie rods. If the cahd=+ ion is in error, also ' provide the following:
- b. An assessment of the significance of this error,
- c. An assessment of how this error was introduced and any actions taken to assure no other errors exist in GENE-771-39-0794 and related submittals made to the staff regarding the core shroud repair.
GPC Respong There is indeed a numerical error on page 43 of GENE-771-39-0794, Revision 1, which corresponds to the page number location identified in the NRC question for Revision 0. The value of 100 degrees was incorrectly used instead of the stated intent to use 130 degrees.
The error has been corrected in the supplement to GENE-771-39-0794, Revision 1.
All criteria stated in the stress analysis are satisfied with the corrected calculation.
There are no additional changes required as a result of this error. Thus, there is no technical significance to the error.
The error was caused by an unintentional change in the temperature of the shroud -
during a cold feedwater transient from 430 degrees to 400 degrees. To assure that this was an isolated occurrence, an independent engineer has performed a complete review of GENE-771-39-0794. The review has not identified any additional significant technical errors. One dimensional error, a material property error, and some typographical errors were found. All of the errors have been cc5 acted. There are no changes to any conclusions and all criteria are satisfied.
i HL-4781 E-8
l
~
. Request for AdditionalInfonnation W
Regarding Core Shroud Modification 3.~ NRC Question In the upper stabilizer assembly, the two parts of the upper bracket act in parallel to provide axial stiffhess It appears that the calculated axial stiffness is based on only one-half of the bracket. Please provide an explanation for the apparent discrepancy and, if needed, revised calculations to include both halves of the bracket. (Refer to page 30, Shroud Repair Hardware Stress Analysis, GENE-771-39-0794, dated
. August 1994). Also, indicate what impact this would have on the gap calculation in Item 1.
GPC Response The calculated stiffness in GENE-771-39-0794 is correct. The upper bracket is symmetric about a plane through the center of the tie rod. As stated on page 32 of GENE-771-39-0794, Rev.1, the analysis used one half of the bracket and one half of.
the applied load (326,500/2 = 163,250 pounds). Since the two halves of the bracket act as parallel springs, the analysis yields precisely the correct stiffness as is shown as follows:
K = F/D Where.
K = Stiffhess F = Force D = Deflection For springs in parallel the total stiffness is the sum of the individual stiffnesses. The stiffness of each half of the bracket is Kl. The total stiffness is then 2*Kl.
K =2*K1 The analysis used an applied force F/2 on one half of the bracket. The analysis determined D for a load of F/2. The stiffness of the total bracket was then calculated as K = 2E = F/D 2D Substitution of the correct numerical values results in the correct stiffness as used in GENE-771-39-0794.
HL-4781 E-9