ML20055E005
| ML20055E005 | |
| Person / Time | |
|---|---|
| Site: | Haddam Neck File:Connecticut Yankee Atomic Power Co icon.png |
| Issue date: | 06/25/1990 |
| From: | Mroczka E NORTHEAST NUCLEAR ENERGY CO., NORTHEAST UTILITIES |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| Shared Package | |
| ML20055E006 | List: |
| References | |
| B13555, NUDOCS 9007100266 | |
| Download: ML20055E005 (28) | |
Text
...
.,,,0 L
o General Offices e Seiden Street, Berlin, Connecticut RTFORD CONNECTICUT 061410270 L
L 1J UZZ,7[,%C (203) 665 $000 June 25, 1990 Docket No. 50 211 B13555 Re:
10CFR50.90 U.S. Nuclear Regulatory Commission Attention: Document Control Desk Washington, DC 20555 Gentlemen:
Haddam Neck Plant Cycle 16 Fuel Recovery Program Proposed Chanaes to Technical Soecifications Pursuant to 10CFR50.90, Connecticut Yankee Atomic Power Company (CYAPC0) hereby proposes to amend Operating License DPR 61 by incorporating the attached (Attachment No. 5) changes into the Technical Specifications for the Haddam Neck Plant.
These proposed changes establish a limit of 160 failed fuel rods (of any type) for Cycle 16 operation and are being provided as a followup to a meeting that CYAPC0 had with the NRC Staff on Thursday, May 24, 1990.
Backaround On September 2,1989, CYAPC0 shut down the Haddam Neck Plant for the fifteenth (15) refueling and maintenance outage.
During the shutdown, primary system radiochemistry indicated a significant number of potential fuel rod failures in the reactor core.
Subsequent ultrasonic, v'.sual, and eddy current inspec-tions of the fuel revealed 456 failed fuel rods in 133 fuel assemblies of the Cycle 15 core.
In a meeting with the NRC Staff on October 15,1989, CYAPC0 provided preliminary information with respect to the identified fuel failures and CYAPCO's effort in defining corrective actions, in a letter dated 13,1989,gg CYAPC0 provided the NRC Staff with additional details December regarding fuel inspections and repair and the projected impact on the Cycle 16 reload analysis and operations.
This was followed up by another meeting between CYAPC0 and the NRC Staff on December 21, 1989.
(1)
E. J. Mroczka letter to the U.S. Nuclear Regulatory Commission, "Haddam Neck Plant, Cycle 16 fuel Recovery Program," dated December 13, 1989.
1 J
f 90
-Rg72co266,00623 ADoct o5000213 L ' j p
i u
PLC j
w
t o
o U.S. Nuclear Regulatory Commission B13555/Page 2 June 25, 1990 Proposed License Condition In a letter dated r bruary 12,1990,I2) CYAPC0 submitted to the NRC Staff a e
proposed license endition requiring implementation of an augmented primary system radiochemistry monitoring program for Cycles 16 and 17 to evaluate fuel cladding performance.
CYAPCO stated that we had evaluated the impact on the accident analysis of Cycle 16 operation with a number of damaged fuel rods, it was determined that this situation constituted an unreviewed safety ques-tion (USQ) in accordance with 10CFR50.59 and, therefore, NRC Staff approval would be required prior to startup for Cycle 16.
The details regarding the VSQ determination were outlined in Attachment No. 2 to the February 12, 1990 letter.
Review of the proposed license amendment (license condition) consti-tuted the process by which CYAPC0 was pursuing NRC Staff approval of startup j
for Cycle 16 with damaged fuel rods in the Haddam Neck Plant reactor.
CYAPC0 is hereby withdrawing our February 12, 1990 license amendment request.
4 Based on the results of testing by the fuel vendor of fuel rods with simulated debris damage, as described in Attachment No. I to this letter, CYAPC0 has now revised our evaluation of Cycle 16 operation with potentially damaged fuel in the reactor.
CYAPC0 has concluded that operation of Cycle 16 with damaged fuel don agi constitute a USQ.
This conclusion has a two fold basis.
- First, CYAPC0 has concluded, based on the fuel rod testing described below, that there is no increased risk of fuel rod failure during an Anticipated Opera-tiona'i Occurrence (A00).
Second, the 160 failed fuel rod limit proposed herein is based on the current technical specification limit of 1 #Ci/gm Dese Equivaler t lodine for primary syste.n radionuclide activity and ensures that the radiological consequences of a postulated steam generator tube rupture are not increased.
Therefore, NRC Staff approval for startup is no longer required and the proposed license condition is not necessary.
Fuel Rod Testina by Fuel Vendor The results of the testing of simulated fuel rod segments with debris damage have demonstrated that damaged fuel rods subjected to limiting mechanical loading will not fail with defects up to 90% throughwall.
The results of the testing program are provided in Attachment No. 1.
Testing of defects greater than 90% throughwall was limited by the ability t'.
accurately machine the defect on the fuel rod surface.
The combination of tests performed with and without the backing of a simulated pellet stack demonstrates that the presence of the pellei stack inhibits the failure of the cladding during overpressure conditions.
There is no reason to believe that this phenomenon would not be effective in rods with defects greater than tested defect depths.
If there (2)
E. J. Mroczka letter to the U.S. Nuclear Regulatory Commission, "Haddam Neck Plant, Cycle 16 fuel Recovery Program, Proposed License Condition,"
dated February 12, 1990.
i U.S. Nuclear Regulatory Commission B13555/Page 3 June 25, 1990 were a failure threshold in the 90% throughwall range, it is highly probable that the rod would fail during normal operation.
But as previously stated, no failures were observed during the testing of rods with simulated fuel
)ellet backing.
The elimination of these incipient failures leaves a population of damaged fuel rods whose risk of failure during an A00 has been shown to be equivalent to an undamaged fuel rod.
This significant result indicates that the radiological consequences during an A00 will be unaffected by the presence of damaged fuel rods.
Therefore, CYAPC0 has concluded that there is no USQ associated with Cycle 16 operation.
The expected number of failed rods is representative of failure rates experienced by other observed failure modes during normal operation.
Fuel Recovery Proaram Summary The onsite fuel recovery program was completed by successfully reconstituting 92 fuel assemblies for reuse in Cycle 16 and six additional fuel assemblies for reuse in future cycles.
A final summary of the onsite activities and lessons learned is provided in Attachment No. 2.
The program evolved from a traditional inspection and repair program to a recovery program that included:
1.
two independent fuel assembly ultrasonic (UT) inspections.
2.
visual inspections and debris removal.
3.
eddy current testing (ECT) inspections of failed rods, neighbors of failed rods, and rods at debris sites; randomly selected rods were also inspected to conservatively estimate the number of reinserted damaged rods.
4.
first-of-a-kind application of fuel rod inspection techniques using a rotating UT prose.
5.
replacement of _ failed and damaged fuel rods with donor fuel rods of equivalent burnup.
The extent of the fuel failures and damage warranted taking actions beyond current industry standards.
The actions taken provide a high degree of confidence that all identified failed and damaged rods were replaced and all identified and retrievable debris was removed from the fuel assemblies, the reactor coolant system, and connected systems.
Statistical Evaluation of Fuel Rod Samolina The estimated number of damaged rods that will be present in the Cycle 16 core has been revised, and is provided in Attachment No. 3.
The revised estimate includes data that became available from the onsite sampling activities subsegnt to the evaluation provided in CYAPC0's letter dated February 12, l
- 1990, and a reevaluation of ECT data based on a revised evaluation l
(3)
Ibid.
I 1
p 8
[
U.S. Nuclear Regulatory Commission B13555/Page 4 June 25, 1990 standard.
The estimated number of fuel rods with damage greater than 20%
throughwall is 375 rods.
The estimated number of rods with damage greater than 90% throughwall is less than 50 rods. These revised results are based on best estimates, and assume a uniform damage distribution.
Auamented Radiochemistry Monitorina Proaram CYAPCO will be implementing an augmented radiochemistry monitoring program for Cycles 16 and 17 to monitor fuel cladding performance and support adherence to the proposed technical specification limit of 160 failed fuel rods contained heroin.
The augmented radiochemistry monitoring program is described in Attachment No. 4.
This program will be used to monitor debris induced failures, more traditional type of failures, and any combination thereof.
Proposed Chanaes to Technical. Specifications As discussed with the NRC Staff in our May 24, 1990 meeting, CYAPC0 is propos-ing technical specification changes, including limiting conditions for opera-tion (LCO) and surveillance requirements (SR), that limit the number of allowable fuel failures for Cycle 16 to a maximum of 160 fuel rods.
This limit would apply to debris induced failures, traditional type failures or any combination thereof.
The augmented radiochemistry monitoring program described in Attachment No. 4 provides the means for implementing the surveil-lance requirement.
CYAPC0 hereby commits to abide by these proposed technical specifications via 1
administrative procedure until such time as the license amendment is issued by the NRC Staff.
The proposed changes (Attachment No. 5) consist of a new LCO, 3.4.12 (new page 3/4 4 51).
The LC0 specifies that the total estimated number of failed fuel rods shall not exceed 160 for more than 7 consecutive days of steady state power operation, if this LC0 cannot be met, the reactor is required by the action statement to be in hot standby within 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br />. A 7-day sampling period is provided to preclude a spurious analysis result from causing a plant shutdown.
The proposed SR describes the testing to demonstrate compliance with the LCO. The initial test of Specification 4.4.12.1.a is performed after 20 days of steady state power operation to ensure that the radioiodines and noble gas concentrations in the primary system have reached equilibrium.
In addition, the SR is written to provide for an increased testing frequency if i
there is an increase in the calculated number of failed fuel rods.
Also a new Bases section, 3/4.4.12 Failed fuel Rods, is being added to page B 3/4 4-13.
Sianificant Hazards Consideration In accordance with 10CFR50.92, CYAPC0 has reviewed the attached proposed changes (Attachment No. 5) and has concluded that they do not involve a
a U.S. Nuclear Regulatory Commission B13555/Page 5 June 25, 1990 significant hazards consideration.
The basis for this conclusion is that the three criteria of 100FR50.92(c) are not compromised.
The proposed changes do not involve a significant hazards consideration because the changes would not:
1.
Involve a significant increase in the probability of occurrence or consequences of an accident previously analyzed.
These proposed changes impact design basis accidents for which no fuel failures are assumed and which have radiological consequences as a result of releases of normal coolant activity, or coolant activity with post-transient spikes. The SGTR is the limiting transient for this case.
Certain design basis accidents, such as the SGTR, assume an initial RCS specific activity consistent with the 1.0 pCi/gm Dose Equivalent Iodine (DEI) limit in the Technical Saecifications.
The proposed limit of 160 failed fuel rods was chosen to se consistent with the DEI limit, based on the experience that the Haddam Neck DEI is a factor of two higher than the I-131 concentration, and that all of the rod failures conservatively release iodine in the traditional manner.
The Cycle 15 experience with the debris induced f ailures has demonstrated that this failure mode releases very little iodine.
The algorithm developed for the augmented monitoring program has demonstrated the capability of conservatively being able to identify a debris type failed fuel rod.
Therefore, if all failed rods were debris type failures, the resulting initial condition DEI and expected spiking factor yield radiological consequences bounded by the current design basis.
Similarly, if all 160 failed rods were traditional type failures, the initial condition assump-tions remain valid.
The fuel rod testing program described in Attachment No.1 shows that no new failures are expected as a result of Anticipated Operational Occur-rences.
Therefore, there is no impact on the radiological consequences of these accidents.
The presence of failed fuel rods cannot initiate a design basis accident.
Therefore, there is no impact on the probability of any accident.
Al so, there are no safety systems affected by the presence of failed fuel rods.
2.
Create the possibility of a new or different kind of accident from any previously evaluated.
The possibility for an accident or malfunction of a different type than any evaluated previously in the Safety Analysis Report is not created.
Since there are no changes in the way the plant is operated, the poten-tial for an unanalyzed accident is not created.
No new failure modes are introduced.
c U.S. t' : lear Regulatory Commission B13555/Page 6 June 25, 1990 The presence of failed fuel rods would only affect potential offsite doses.
The system response to accidents is unaffected.
There are no failure modes associated with the failed fuel rod limit and since the presence of failed rods does not initiate an accident, there is no possibility of generating a new design basis event.
3.
Involve a significant reduction in a margin of safety.
The proposed changes do not have any adverse impact on the protective boundaries.
The margin of safety, as defined in the basis for any i
Technical Specification, is not reduced.
The proposed changes do not adversely impact any of the safety systems, nor do they increase the number of challenges to the safety systems.
The limit of 160 failed rods was chosen to be consistent with initial conditions assumed for the radiological design basis when the fuel failures are of the traditional type.
If the failures are debris induced, the dose equivalent iodine will be significantly lower and the resulting doses will be bounded by the initial condition and s?iking assumptions.
The protective bou..Jaries are not directly affected by tb proposed failed fuel rod limit, since the limit is consistent with the radiological design basis assumptions.
Since the number of allowable failed rods is consistent with the design
' basis radiological analyses, there is no reduction in the margin of sinfety as defined in the basis of any Technical Specification.
The Commission has provided guidance concerning the application of the standards in 10CFR50.92 by providing certain examples (51FR 7751, March 6,1986) of amendments that are considered not likely to involve a significant hazards consideration.
The changes proposed herein are not enveloped by a specific example.
As described above, the aroposed changes do not constitute a significant hazards consideration s nce the proposed failed fuel rod limit was chosen to be consistent with the initial conditions assumed for the radiological design basis for the Haddam Neck Plant.
Therefore, these changes do not involve an increase in the probability of occurrence or consequences of an accident previous-ly analyzed, do not create the possibility of a new or different kind of accident, nor involve a reduction in a margin of safety Based upon the information contained in this submittal and the environmental assessment for the Haddam Neck Plant, there are no significant radiological or nonradiological impacts associated with the proposed action, and the proposed license amendment will not have a significant effect on the quality of the human environment.
The Haddam Neck Plant Nuclear Review Board has reviewed and approved the proposed changes and has concurred in the above determination.
L
I U.S. Nuclear Regulatory Commission B13555/Page 7 June 25, 1990 In accordance with 10CFR50.91(b), CYAPC0 is providing the State of Connecticut with a copy of this amendment request.
CYAPC0 is not requesting a particular schedule for this license amendment.
However, as stated previously, CYAPC0 will adhere administratively to these proposed technical specifications until the license amendment is issued by the NRC Staff.
We trust you will find this information satisfactory and we remain available to discuss this with you at your convenience.
Very truly yours, CONNECTICUT YANKEE ATOMIC POWER COMPANY bYZf+<.
/
E.J.Mrot2K ff Senior Vice President Attachments cc:
T. T. Martin, Region 1 Administrator A. B. Wang, NRC Project Manager, Haddam Neck Plant J. T. Shediosky, Senior Resident inspector, Haddam Neck Plant Mr. Kevin McCarthy Director, Radiation Control Unit Department of Environmental Protection Hartford, Connecticut 06116 STATE OF CONNECTICUT)
) ss. Berlin COUNTY OF HARTFORD )
Then personally appeared before me, E. J. Mroczka, who being duly sworn, did state that he is Senior Vice President of Connecticut Yankee Atomic Power Company, a Licensee herein, that he is authorized to execute and file the foregoing information in the name and on behalf of the Licensee herein, <1d that the statements contained in said information are true and correct to the best of his knowledge and belief.
Uthn:
Pb 7
r Notary Pt@ic MyCommissionEhlre:Mhch3I,ihh[
1 4
o.
Docket No. 50 213 B13555 i
i Attachment No. 1 Haddam Neck Plant Fuel Rod Failure Analysis and Testing l
r June 1990 i
[ 7. = Fuel Rod Failure Analyses and Testing Fuel Rod Failure Annivses The Haddam Neck fuel rods are backfilled with helium at relatively low pressure (40 psig).
During operation, the increase in fuel temperature, change in the rod gas volume, and the release of gaseous fission products increase the rod internal pressure. However, the fuel rod internal pressure is less than the Reactor Coolant System (RCS) pressure during normal operation. In addition, the rod internal pressure will reduce due to reactor shutdown. For all transients, except the large break LOCA, the fuel rod will remain in a compressive state. The rod will be exposed to a tensile stress only during the normal RCS depressurization and cooldown required for entry into Modes 4 through 6. The decrease in RCS temperature and decay heat with increasing time from reactor shutdown reduces the rod internal pressure even further.
The maximum differential pressure expected to result in a tensile stress in the cladding is
~390 psi. Only a limited number of rods will have a severe power history and burnup to experience the highest intemal pressures. The maximum differential pressure expected to result in a compressive stress in the cladding is -2600 psi. All of the fuel rods experience high external pressure during operation. The only significant risk to fuel rod integrity, therefore, is from external pressure loading.
A conservative stress analysis was performed to determine the minimum required cladding wall thickness. The analysis limit was set at the minimum unirradiated tensile strength.
Irradiated cladding increases in strength and decreases in ductility. Exceeding the limit does not necessarily result in failure, since the cladding is in compression. A summary of the results is provided below:
Pressure. nsi Matimum Defect Depth. inch 2750 0.008 2500 0.009 2000 0.011 A fuel rod testing program was developed to establish more realistic failure limits. This testing program is discussed below.
Fuel Rod Failta Testing Babcock and Wilcox (B&W) was requested to perform testing of mock up fuel rod segments with simulated debris scars to investigate the possibility of fuel rod failure due to limiting mechanical loads associated with an overpressure Abnormal Operating Occurrence l
l t
u Fuel Rod Failure Analyses and Testing (AOO) Phase 1 testing was performed on rod samples that were each made from two lower end caps and a six inch long piece of cladding. The nominal cladding wall thickness is 0.0165 inch. The rodlets were not backfilled and thus had an initial internal pressure of 14.7 psia. No fuel rod internals (pellets or plenum spring) were included in the Phase 1 testing. Phase 2 testing was performed on rod samples similar to the Phase I samples except that the fuel pellets were conservatively simulated by stainless steel stock, and a plenum spring and upper end cap were included. Each phase of the testing program is t
described below:
Phase 1 Testine The rod samples for Phase 1 testing included simulated debris wear defects that were machined into the cladding one inch from one end. The range of individual defect depths was from 0.006 to 0.014 inch in 0.002 increments. A total of four rods for each defect depth were prepared.
A total of five tests were cenducted. For each test, the samples were loaded into an autoclave which was mised in temperature and pressure to 600 F at saturation pressure.
The pressure was then increased to the test pressure and held for one hour. The pressure was then reduced and the autoclave cooled. The samples were removed and inspected. Samples which had failed were removed and the next test performed. The test sequence is shown below:
Rn Pressure (nsic) 1 2750 2
2000 3
2250 4
2500 5
2750 Five samples (one at each defect depth) were evaluated in Test 1. Only the rods with 0.014 and 0.012 inch defects failed at the defect site. The other three rods failed near one of the end caps where the cladding loses support from the endcap. The remaining samples were subjected to Tests 2 through 5. A summary of the Phase 1 testing results is provided below:
- Attachment 1 - Fuel Rod Failum Analyses and Testing Defect Test Pressure (psig) fiDdd 2000 225.0 2f00 2.7$.0
.006 ok ok ok ok 1
.008 ok ok ok ok l
.010 ok ok ok' ok
)
.012 ok ok failed failed i
.014 ok failed failed failed
- One of the four specimen failed at 2500 psig.11owever,its failure is believed to have been caused by the shock wave generated by the failure of the specimens with 0.012 defects.
A defect as deep as 0.010 inch can withstand pressures up to 2750 psig. The defects that did fail, failed by cracking after denting at lower pressmes. The failure mechanism 1
does not appear to be buckling, but a bending defonnation similar to ovalization. The cladding deforms plasticly and then exceeds the tensile yield strength. For irradiated cladding, the higher yield and tensile strengths should allow defects deeper than 0.010 inch to survive the maximum transient system pressures. This represents 60% of the specified cladding thickness. It should be noted that 15 of the 20 samples also failed at the end cap due to the lack of support from the pellet stack or plenum spring. The Phase 2 testing was devised to address this shortcoming.
Phase 2 Testine The Phase 2 testing program was developed in order to funher assess the potential for fuel rod failure when the simulated fuel rods contained a mock up of the fuel rod internals. The results of the Phase I testing indicated that some of the fuel rods subjected to compressive loadings associated with an overpressure event failed due to the lack of pellet support at the endcap/ cladding interface, not at the simulated debris defect site. The rod samples for the Phase 2 testing included the presence of a simulated pe!!et stack and the upper plenum spring. The pellet stack was simulated by a stainless steel bar, conservatively sized to give a gap that is greater than the maximum gap between the cladding and the pellet in an actual fuel rod.
The cecond set of tests were perfonned in a similar manner to the first set of tests, except only the deeper defects were niodeled and a set of control samples was also tested. Twelve rods were tested, with three rods each having nominal defect depths of -
0.010,0.012, or 0.014 inch. The as built defect depths ranged from 0.009 to 0.015 inch. The remaining three rods were the control samples, which did not have simulated defects. The testing sequence and conditions were similar to the Phase 1 testing. The first test at 2750 psig included one control sample and one rod at each of the three L
' + '
... Attachment 1 - Fuel Rod Failum Ana.yses and Testing i
defect depths. Subsequent tests neluded all three control samples and all remaining sample rods until a failure mode was observed. Tube collapse, even though not initated at the defect site or necessarily throughwall, was considered a failure mode for the purpose of these tests.
All rod samples were examined at the completion of the testing. The defect site of all samples was examined under high magnification (10 to 70X) with a stereo microscope.
None of the samples showed evidence of failure at the defect site. Most of the samples showed local bending (denting) at the defect sites. The -support provided by the simulated pellet prevented rod failure at the defect. Similar support would be expected from the actual fuel pellets in core. The samples with the two deepest defects were sectioned and examined with a scanning electron microscope. Even though there was evidence of tearing and microcracks, no throughwall cracks were observed. All but one sample showed cladding collapse over the plenum spring section, away from the defect site. None of the control samples, however, collapsed. The collapsed sections all appear to have started in the middle of the rodlet,90 degrees and several inches from the defect site. This is the location where the samples were gripped when the defects were machined. The grip force was excessive and resulted in a slight pinch in the cladding. The site of the pinch buckled under the high test pressures. Therefore, this test failure mode was unique to the testing configuration and unrelated to the simulated defect site.
The effects of irradiation and plant operational changes must be assessed and combined with the results of the test programs. All of the rods with potential debris damage have one or two cycles ofirradiation, which has resulted in the cladding yield and ultimate tensile strength increasing and the ductility decreasing. The increase in yield strength and the higher ultimate tensile strength means that the rods will not dent or crack until a greater differential pressure than the conditions experienced by the test samples. The buildup of caustic fission produce such as iodine and cesium could result in stress corrosion cracking. This failure mode is not expected since the highest stresses are on the rod OD, not the ID. Cesium and iodine form solid compounds at temperatures representative of the bottom pellets of the fuel rod. When combined with the low release rate from the low power bottom pellets, few of these fission products would migrate to the cladding ID surface.
The changes in pressure loadings due to operational changes has also been assessed.
The highest external pressure loads have been shown to occurr during normal operation and overpressure events. A much smaller load in the opposite direction occurrs when the plant is shutdown and depressurized. This alteration ofloading slightly fatigues the
\\
i ip I
ii I
ll h
- ~
- . Attachment 1. Fuel Rod Failure Analyses and Te: ting cladding. The additional fatigue caused by pressure changes during two more cycles of exposure is not expected to result in failure by fatigae.
The results of the analyses and two testing programs are summarized below:
1.
Transients which result in a pressure dmp represent much less of a risk.
The maximum loading in tension is ~390 psi compared to a compression loading between 1800 and 1900 psi during normal operation. None of the rods which can survive normal operation should be at risk during an AOO which results in a pressure reduction. Pressure drops can, however, result in radiochemistry spikes from failures which have occurred during normal operation.
2.
An unsupported defect site with a depth up 10 0.014 inch can survive at the normal operating pressure of 2000 psig, but could fail between 2250 and 2500 psig. Any unsupported defect greater than 0.010 inch could fail at a transient pressure of 2750 psig. These results represent limiting cases, since all of the observed debis induced defects were in the portion of the rod where the fuel pellets provide cladding support.
3.
The presence of the fuel pellets at the debris induced defect sites proddes support to the cladding and limits the amount of buckling so that rod failure is precluded. A defect as deep as 0.015 inch (90%
throughwall) experienced 2750 psig without a throughwall failure.
1 1
g..
- g a.
v
-f
's)
[,
.h.
r.<
7a
- e; '
y 9.,
+ -,
3,-g <
^
4
- Docket Nor 50 213 1JH1 1
a 5
.t
>l
,eo i
.j F
4
- - I i
.' l
- I r
+
i r
r j
r, l: '
I_
i t
~ Attachment No. 2 j
' Haddam Neck Plant
.i Final-Results-Fuel-Recovery Program j
o
,)
-- i 1
1, j
1-5
.)
'r.
r o
't is g.
a, 5.'
o i
-- i c;
i r
June 1990 1
4
,Iy'.
a.
a
.-.w-r-....
. -e4
..w.-s
..-4-4.-
-~,e-.. - -. - < < - - -.v,e-,,*,.+-..-mem.v -r+4 e-,
re-=-4--+.-==~.-,
--r.--e-
-w
~~*r-s*
e Anachment 2. Final Results. Fuel Recovery Program i
j The ongoing status of the Haddam Neck fuel recovery program was provided to the NRC i
in the Reference 1 and 2 submittals as well as presentations to the NRC on October 15, 1989 and December 21,1989. The onsite recovery efforts were completed on February 9, 1990 with the successful reconstitiution of 92 fuel assemblies whleh will be re inse the Cycle 16 core and six additional fuel assemblies for use in future cycles. The final results of the inspection, cleaning and reconstitution activities identified 456 failed rods in 133 fuel assemblies of the Cycle 15 core. The failures in the once and twice burned fuel assemblies (348 rods in 94 fuel assemblies) were confirmed during reconstitution. The failures in the thrice bumed fuel assemblies (108 rods in 39 fuel assemblies) were b the initial ultrasonic inspection (UT) results. Only five of the once and twice burned fuel assemblies had no observed debris and no failed rods. The recovery program involved the inspection of over 3600 fuel rods and replacement of over 500 failed and damaged rods.
The following discussion provides a summary of the onsite activities.
Oricinal Workscope The original workscope of the fuel related activities for the Cycle 15 refueling outage 4
was established based on the results of the Cycle 15 radiochemistry predictions of six to twelve failed rods in high burnup fuel. The expected failure mode was pellet-cladding interaction. The Haddam Neck fuel has been susceptible to this type of failure mode (~five rods per cycle) in the past. Traditional UT and reconstitution capabilities were mobilized prior to the outage. An additional feature of the original program not typically found in reconstitution campaigns was fuel rod Eddy Current inspection (ECT) capability to confirm that a rod identified as a leaker by UT was indeed failed,
and to characterize the failure as part of a root cause fuel failure investigation. The ECT inspection, however,is unable to charaterize failures in the end plug and upper plenum spring regions due to signal interference from these components. Only rods identified i
j as failed would be removed and inspected, and replaced with a solid stainless steel
~
dummy rod if confirmed failed. A fuel management limit of one dummy rod per fuel assembly quadrant was imposed, i
Initial UT Inmeetion The vendor for the initial UT inspection (Innovative Technologies Incorporated, i ~
formally ABB and BBR) was selected based on their experience evaluating the Haddam Neck stainless steel clad fuel during the Cycle 14 refueling outage. The Cycle 15 data obtained were unique, almost to the point of being unrealistic due to a combination of the number of potentially failed rods and the characteristics of the UT signals. The initial data evaluation identified 25 failed rods in 22 fuel assemblies of the Cycle 15 l
l
I
'. Attachment 2 Final Results. Fuel Rec;very Program h
core, and many questionable rods.
Reconstitution Phase 1 The initial reconstitution activities were sequenced to repair fuel assemblies with rods failed on the fuel assembly periphery to allow a visual inspection of the failed rod prior to di: assembly. The first two fuel assemblies inspected showed significant amounts of metallic debris between the bottom nozzle and the bottom spacer grid. At this point it was recognized that reconstitution would be ineffective without removing the debris from the fuel assemblies. Reconstitution was performed on five fuel assemblies with an expanded scope of evaluating rods with questionable UT signals and neighbors of failed rods. These early reconstitution efforts confirmed that the questionable rods were failed and that neighbor rods were damaged. These results led to a re evaluation of the UT dtaa to yield failure predictions of over 300 rods in the Cycle 15 core. The reconstitution activities were suspended during the fuel assembly visualinspection and cleaning actisities.
Fuel Assembly Visualinspection and Cleaning All 109 of the planned re insert fuel assemblies for the Cycle 16 core were inspected and cleaned. The visual inspection confirmed that the debris was limited to the region between the bottom spacer grid and the bottom nozzle and the underside of the bottom nozzle. During the inspection and cleaning activities, debris maps were developed for input to the subsequent reconstitution activities. All visible debris was removed with the exception of one piece each in nine fuel assemblics that was woven into the periphery of the bottom spacer grid. Attempts to remove this debris would have resulted in irreparable damage to the spacer grid. Subsequent inspection of the affected rods showed that all rods were undamaged and that the debris posed little risk during future operation.
The finalinspection and cleaning results showed that 97 of the original 109 re-insert fuel assemblies had visible debris and that in general, there was a correlation between debris sites and failed rods identified by the UT inspection.
Independent UT Inmeetion The combination of the extensive amount of debris removed from the planned re insert fuel assemblies and the uncertainties in the original UT data evaluation identified the need for an independent UT inspection. Babcock & Wilcox (B&W) provided the
-J
\\
-. Attachment 2 - Final Results - Fuel Recovery Program a
i inspection service which consisted of using a different UT device, data collection system and evaluation guidelines. The inspection results of the originally planned 109 re-insert fuel assemblies confirmed the extent of the rod failures. The results also identified an additional 44 failed rods. The majority of these failed rods, however, would have been discovered during reconstitution since they were located adjacent to other failed rods or at debris sites.
The cause of this discrepancy is related to the nature of the debris type of failure. Very little water appears to be present in the failed rods due to the location of the defect.
Subsequent reconstitution activities identified several rods that were still releasing gas up to eight weeks after shutdown, which substantiates the radiologically benign nature of this type of defect and the inability to detect its presence using traditional radiochemistry monitoring. The elapsed time between the two UT inspections (six weeks) allowed more water to enter the failed rods. The accuracy of the combined results is proven due to the fact that only three failed rods were discovered that were undetected by a combination of both UT results.
Reconstitution - Phase 2 Reconstitution was restarted after completing the debris removal and independent UT inspection. The original strategy of only inspecting and replacing failed rods with dummy fuel rods and preserving the original Cycle 16 core design was abandoned due the extent of the fuel damage. Twelve fuel assemblies originally planned for re insertion were selected to provide donor fuel rods to replace failed and damaged fuel rods. An additional once burned fuel assembly from the fuel pool inventory was used to provide donor fuel rods. Donor fuel rods were selected to match the burnup of the failed or damaged rod as closely as possible. The Cycle 16 core inventory was established assuming the purchase of eight new fuel assemblies and the re use of four fuel assemblies discharged at the end of Cycle 14 to replace the twelve donor fuel assemblies. The successful reconstitution of 92 once and twice burned fuel assemblie was also assumed.
L l
Fuel rod visual and ECT inspections wm expanded to include neighbors ofidentified fuel rods, rods at the mapped debris sites and rods selected on a random basis to esdmate the extent of the damage in rods that would not be inspected. If any of these
/-
rods were found to be failed or damaged, the inspection was further expanded to include the new neighbor rods.
A new fuel inspection technique was introduced during the last month of the reconsdtution activities. The inspection device consisted of a rotating UT probe that I
- - - - - - - - - - - ~ -
)
'
- Attachment 2. Final Results - Fuel Recovery Program was adapted from control rod inspection technology. This developmental tool was used to provide a better inspection at the lower end cap region were the ECT inspection is inconclusive due to signal interference. In general, these results showed good agreement with the ECT results when damage was present above the lower end cap weld. The results also indicated that rods were being conservatively discharged as damaged based on a misinterpretation of some the ECT signals in the end cap weld region. The ECT evaluation indicated defects significantly greater thsn 20%
throughwall, while the rotating UT and visual inspections showed that the anamoly was not damage, but a feature of the end cap wcld geometry from the original rod fabrication ptceess.
At the completion of the onsite fuel recovery activities, the final fuel assembly configurations were provided to the the fuel cycle designer to evaluate the impact of all of the fuel rod transfers on the the revised Cycle 16 design. In all cases, the impact was shown to be insignificant. The details of the revised Cycle 16 design, associated safety analysis and core operating limits were provided in Reference 3.
+
i i
Lwons Learned The extensive fuel rod failure and damage that occurred during Cycle 15 operation was due to metallic debris remaining in the Reactar Coolant System at the startup of Cycle 15. The importance of debris removal and control has resulted in comprehensive programs implemented as part of the thermal shield removal and debris cleanup efforts. These programs are described in detail as provided in Attachment 3 of Reference 4.
A contingency that was investigated in parallel with the fuel recovery program was the development of a bottom nozzle debris filter. This filter was intended to be retrofit to all fuel assemblies in the Cycle 16 core. Two independent design concepts were pursued until it became clear that technical and performance concems could not be adequately addressed to support Cycle 16 implementation.
l t
Debris resistant features, however, are being developed for the first batch of zircaloy clad fuel for the startup of Cycle 17. The fuel rod will be re-designed to include a solid lower endcap that extends into the bottom spacer grid. Any debris present between the bottom spacer grid and the bottom nozzle will not cause a fuel rod failure. This design feature will l
preserve hydraulic compatability with the stainless steel clad fuel assemblies during the transition cycles.
'. Final Results Fuel Recovery Program The debris control and cleanup programs implemented during the Cycle 15 refueling outage and fuel rod design changes proposed for the next batch of fuel will minimize the risk of debris induced fuel failures in future operating cycles.
References
- 1. E. J. Mroczka letter to U. S. Nuclear Regulatory Commission, "liaddam Neck Plant, Cycle 16 Fuel Recovery Program", dated December 13,1989,
- 2. E. J. Mroczka letter to U. S. Nuclear Regulatory Commission, "liaddam Neck Plant, Cycle 16 Fuel Recovery Program, Proposed License Condition", dated February 12, 1990.
. E. J. Mroczka letter to U. S. Nuclear Regulatory Commission, "lladdam Neck Plant, Revised Technical Report Supporting Cycle Operation, Proposed Changes to Technical Specifications", dated May 9,1990.
- 4. E. J. Mroczka letter to U. S. Nuclear Regulatory Commission, "Haddam Neck Plant, Reactor Vessel Surveillance Program, Proposed Changes to Technical Specifications",
dated February 16,1990.
r
w.b
+-
L%_ w t,
- a
.1
_%, t -.
er. -
m
,1 3.. w
- r
- t 1
'_1 3
t
(
+
'st F
'l_'-
5
.j i..
[j:,
^
l r;
s 5
{
/
[t
'N(
f' )'.
Docket No. 5% 213 m'
n
-13
.- y
,a 4:
r
?
^#
E I
I-l 1
I
- x..
- h. %.
,z b
4 n
' ':p i\\?!
'a
,y
_j 4..
-1 I
i i
.)
..]
i ~- l
'l
- o E
Attachment'No. 3
'\\
Haddam Neck Plant-4 s
ll Revised Damage-Estimate -Cycle'16 Core-t
-i
'f
. ;q E!
i 5}
q s
s
};'--
i e-
- }
?
.)!
I -
+
't4~ 5
'.+^;
4 H
3; 1
's
_o a
1 1
l
+
l,.....
-c, f5..,,
i!
_A-
' E
- f i
- -
- p t_
5 y,
l..
t i
g-9
['
5 r
t)( (
e,.
'f 4
r~
June 1990 1
g1 3
i..
t t
=
g
q
- . Attachment 3 Revised Damage Estiman Cycle 16 Core A statistical evaluation of the random fuel rod sampling data performed as a part of the onsite fuel recovery program was provided in Reference 1. This ewtuation used Monte Carlo methods to provide a conservative estimate of damaged rods at a 95% confidence level. The results of the upper bound estimate (95% level) yielded 939 damaged rods, with a mean value of 640 damaged rods.
Several additional fuel assemblies worth of data can now be included in the data base. An additional significant input to the number of damaged fuel rods is a review of the Eddy Current (ECT) data using a revised calibration standard. As described in Attachment 2, a new fuel rod inspection technique was implemented during the last month of the onsite repair activities. This new inspection technique (rotatinF ultrasonic technique) was developed because the ECTdata are significantly affected by me presence of the solid lower end cap and the weld heat affected zone. Since the daw.c4.d not be reliably interpreted in this region, Me ECT technicians evaluated the data conservatively.
The inspection data obtained from the combination of ECT, visual and rotating UT exams shov/cd that many fuel rods had been discharged as damaged (greater than or equal to 20%
throughwall) based on a conservative interpretation of ECT data near the end cap weld region. The rotating UT and visual inspection results showed that the ECT data were responding to an endcap weld geometry phenomenon prior to losing a meaningful signal in the endcap/ weld region. The endcap weld geometry was replicated in the laboratory, and provided an additional standard to be used during the ECT data evaluation in the field.
The revised evaluation standard was used during the last three weeks of the repair activities. After the onsite activities were completed, B&W was requested to review the ECT data using the me.lified evaluation standard for the 33 fuel assemblies that were part of the random sampling program. The results of this review showed that the number of damaged rods in the sample could be reduced from 35 to 22. The additional fuel assembly data and the revised damage evaluations are provided below:
l l
l
r
'. Attachment 3. Revised Damage Estiman - Cycle 16 Core
(
Pin Damage Data by Fuel AEKmblX
!bi Da T.'s DA Donors Da
$02 0/16 701 3/40 S14 1/115 505 0/27 TV2 0/10 S15 0/35 S08 1/21 704 0/16 S22 4/47 Sil 0/21 TV5 0/16 S47 2/103 S17 0/16 708 3/33 T31 0/61 S29 0/15 T12 0/16 T34 0/74 S30 0/21 T13 2/20 T36 3/53 S37 0/16 T22 1/6 T39 2/31 S38 0/25 123 6/13 T35 0/20 T37 0/23 T42 0/16 T50 0/16 T51 0/29 T54 0/16 T55 0/14 D/l = number rods damaged / number rods inspected A revised estimate of the number of damaged rods (greater than 20% throughwall)in the Cycle 16 core was determined by multiplying the revised damage frequency (22 rods damaged /1000 rods inspected) by the total number of uninspected rods reinserted in the Cycle 16 core. The revised best estimate of the number of damaged rods in the Cycle 16 core is 375 rods. Assuming a uniform damage distribution, the number of damaged rods with damage greater than 90% throughwall is less than 50 rods.
References
- 1. E. J. Mrocz.ka letter to U. S. Nuclear Regulatory Commission, "Haddam Neck Plant, Cycle 16 Fuel Recovery Program, Proposed License Condition", dated February 12, 1990.
1
"in(e 'q w,. %' l 'c ka
,m
..h ?,.l ei;-
6 g
,:y
/ (. p.lj,gj $ " S k i I, ', 4 s.
i
.t 4
i i.
ana=m 4./ f..j
't..l.
- . ; c;t. o,, i j,.
't,',
-,4 I
h p *g e. l.,,
a
,c
( th '.
cii (gy 14' l.:
7 i
p; ;
o
- ,}'.
(
I c
1
- 9. r4.M..
W Docket No. 50 j
z ' s t
. _~.$.
4.K, r ',.,
t ii.ds ' i f,3 44 b.i i
h b.u.e,,,
~ d hTW e sa ; -
,., 1) ; } ' 4 h ';
4 '
6, r,
r
- s.. g7
'kl h $.
e < t ;Ic'h)..
,r.
y t-R ".' l '
- l,?
1,
, y 0 /-;.s e
[p k0 I
. g.g.w 1
QM, ceb p;,t[- g!'j {.i 8
- /
t./ 2.w(.
i m
k'd ;.
'4 w; J, it L F!
. \\.
hjh
,\\
't, f
((~
g s=
w e
.n.
f.?Sl m
'2[g 'e--
)
p.
.,i 4
7
[
l g-,
-+f 3,
1 i
o " ~ ~~
- Attachment No. 4 q
y
.)
m Haddam Neck Plant-
.u.
Ih m
3-
+
- Augmenteo Radiochemistry. Monitoring Program L.
,e,
. 'o ",
8 1
^ !
- t
- ', n
f
{,
t g.
a.
WP a
xx m
t J
4
- q..
..,5-r 1 }..
3 i;
i?
,4 I: k j-[kd l i, '
, t.
s
'x.
)
'; n-i g
y s t I
'-s Is 3
- - 4-I..
s'
.. t
,, s w
- n...
m 4
I k
-June 1990
,-,s'.#.
.f t
,--F
,'s,, f".h.-.
b, b 1m. '
,)N
-d:
- .- w 5 L 4.,.' be -w
.--.--e--,*
_~h,-.,,.
--<+-ww----w--=**-***WS--+=*~+++'"***-'"-*'-"*"****""*-*
- * * " ' * * ' ' " " ' * - ~ " " " " * * ' ' ' " ' " " '
M ""
' ^- '
- Augmented Radiochemistry Monitoring Program Debris present in the Reactor Coolant System (RCS) during Cycle 15 produced fuel pin defects that did not radiochemically behave as expected. Calculations were performed to determine an accurate model by which Cycle 15 and previous radiochemistry data could be explained. Data for Cycle 15, as well as previous cycles, were reviewed to test the model.
From the model, a derived Xenon Pin Equivalent (XPE), similar to the Iodine Dose Equivalent, could then be calculated in order to relate the gaseous activity concentration of Xe 133 to the number of defects. The following discussion provides a summary of the bases for the development of a Technical Specification based on monitoring xenon radiochemistry data.
Defect Characterization The defects that occurred during Cycle 15 were a result of debris induced fretting at the i
bottom of the fuel rod. When similar defects occur in zircaloy clad fuel rods, secondary failures usually follow at a higher elevation on the rod. As with other defects at this i
higher elevation, the gas in the pin will escape, allowing water to enter the rod, and in turn, facilitate iodine and other soluble fission product transport into the bulk coolant.
In the case of Cycle 15, these secondary defects did not occur, primarily because the stainless steel cladding is relatively impervious to hydriding. As a result, RCS iodine concentrations were not indicative of the number of failures.
Defects at this location created a bell jar effect, where gases remained trapped and pressurized in the fuel rod and slowly escaped. However, since a gas / water interface did exist at the defect location, some osmotic diffusion of fission gases could take place. Diffusion across this interface is time consuming, thus shorter lived gases would decay prior to being released to the bulk coolant, while the longer lived gases would survive the transit time.
Model Development M Xenon Contribution Using an equation similar to the iodine based equation for fission product transport, a Xe-133 contribution to the RCS for one rod failure was determined:
E Y
4 m
E
- Augmented Radiochemistry Monitoring Program Y e Vxe + Y V; A (pCi/gm/ rod) = AxeF X
g A
A.
Xe l
(Axe +B )(3.7E4 dps/pCi)(RCS vol)(# rods) 1 e
when::
A = decay constant for 1 133 = 9.26E 6 sec-1 1
Axe = decay constant for Xe 133 = 1.53E 6 sec-1 VXe = escape rate coefficient for Xe 133 for pellet to gap = IE 7 sec 1 V = escape rate coefficient for 1 133 for pellet to gap = 1.44E 8 sec 1 I
YXc = Y = fission yield for 133 line = 0.0670 atoms / fission I
B = purification rate constant for1133 = 3.67E 5 sec'l I
- rods = 157x204 = 32028 RCS volume = 2.305E+8 cc F = fission rate = 5.66E19 fission /see which yields a theon:tical release rate of 0.0373 pCi/gm/md.
Since the gas phase would be a more consistent value, this value is adjusted to the concentration of gas, assuming a concentration of 33.3 cc/kg (average from previous cycles), or 1.119 pCi/cc/ rod. This would be the xenon contribution for defects which behave normally, that is, release noble gas without hold up and also release iodine.
b) Xenon.Ratics The direct application of a single Xe-133 value would not accurately estimate the number of failures which occurred during Cycle 15. Recalling that the difference in iodine ratios can depict the difference between open failures and tight failures (cracks or pinholes), a review of long lived Xe 133 to short lived noble gases shows that Cycle 15 experienced a Xe 133/138 ratio a factor of ten higher than previous cycles. The other ratios did not experience such a large deviation.
Therefore, it is appropriate to use the comparative ratio between the Cycle 15 ratio and the normally observed ratio to determine an associated diffusion factor by which to adjust the Xe-133 contribution per failed pin. This contribution per pin l
value is then compared to the Xe 133 value to determine the defective pin count.
1
f
' ' ' $. Augmented Radiochemistry Monitoring Program 4
3 c) Model Testine This xenon model was tested using data from Cycles 8,13,14, and 15 with favorable results. Some conservatism exists with the xenon model, but it is not excessive The results are shown below:
hindel Predictions Versus Actual Failures I. Based Xe Based Failure Failure Actual Failed Defect Tyne Prediction Xe Ratio Prediction Eini Cycle 8 Open Defects 100 Rods 7
150 Rods 100 Rods
- Cycle 13 Tight Cracks 2 Rods 7
3 Rods No Inspections Cycle 14 Tight Cracks 6 Rods 7
9 Rods 6 Rods Cycle 15 Debris 12 Rods 70 450 Rods 456 Rods
- Estimated, based on fuel assembly sipping results d) Xenon Pin Equivalent Derivation The modelis sufficient to explain and quantify defects of either type, but does not adequately quantify a mixture of defects. To obtain a value which adequately represents total pins, regardless of defect type, an accounting approach was used to derive the Xenon Pin Equivalent (XPE).
First, the theoretical Xe 133 concentration as a result of failures which allow iodine release to the bulk coolant is quantified:
Xet = (.0373uCi/gm/ rod)(1000 cm/kg) [ (I 131 pCi/gm)(.01)(32028)]
(1)
(33.3 cc/kg) where:
- Of Rods In The Core = 32028 Xe-133 Per Pin = 0.0373 pCi/gm/ rod (derived)
Gas Concentration = 33.3 cc/kg The term [I-131 pCi/gm)(.01)(32028)]is the estimate of the number of traditional failures, since past experience has shown that the I 131 concentration at Haddam Neck approximates the number of defects (%). Subtracting this value from the
k*
' Attachment 4 - Augmented Radiochemistry Monitoring Program
'o measured xenon concentration in the RCS, (XeRCS), yields the Xe activity J
resulting from the Cycle 15 debris type defects:
XeC15 = XeRCS Xet (2)
Dividing both sides by the theoretical Xe 133 activity as a result of one Cycle 15 failure would result in an equation which estimates the tamber of Cycle 15 debris type defects, f
A e 133 (pci/cc/ rod) = 00373 uCi/cm/rody.1V1000 cm/kc)
(3)
X 33.3 cc/kg i
Adding the number of failures which release iodine, (I 131 pCi/gm)(.01)(32028)
(4) yields, when reduced, the Xenon Pin Equivalent.
Xenon Pin Equivalent (XPE) =9.0 Xe 133 pCi/cc 29901 131 pCi/gm (5)
To evaluate this equation, several mixes of failures were examined. The I 131 and Xe 133 concentrations were determined using the calculated contribution based on the two types of failures. The results of the comparison are provided below:
I Xenon Pin Equivalent Compsif.on Traditional Debris Total Xe 133 I131-L Egilures Failures Failures gCi/.cg uCi/cm XEE
% Error 160 0
160 179
.50 161
+1 80 80 160 98.48
.25 161
+1 40 120 160 58.20
.12 176
+10 20 140 160 38.06
.06 168
+5 10 150 160 27.99
.03 165
+3 5
155 160 22.95
.016 160 1
159 160 18.93
.003 162
+1 0
160 160 17.90 0.0 161
+1 6
450 456 70
.025 557
+22
. t;.. -
'E Attachment 4 Augmented Radiochemistry Monitoring Program e
The above data shows that even a small number of traditional failures would conservatively skew the prediction at levels below 160 failures. In all cases, the XPE prediction is within 10% of theoretical. Cycle 15 data were also added for an evaluation of actual data. The 70 pCi/cc concentration of Xe 133 was an end of cycle value. The number of traditional failures was estimated, since the discharge fuel was not examined and would be expected to contain the traditional pellet / cladding interaction type of failure. An additional conservatism in the calculation is that the original fission yield for Xe is used as a fixeo value. Later in core life, this will increase, caosing a higher actual Xe-133 concentration and result in an overestimate, rather than an underestimate.
Summarv A de of a Xenon Pin Equivalent of 160 pCi/cc is approximately equal to 160 pin faih.m. If all of these fA+ released iodine, an Iodine Dose Equivalent of 1 uCi/gm would result, based on the. cycrience that the Haddam Neck lodine Dose Equivalent is a factor of two higher ti.mi the I 131 concentration. This number of failures is also ebout one third of the number of failures that were present at the end of Cycle 15.
Therefore, the desi;;a basis radiological evaluations for the Iodine Dose Equivalent would be consistent with a Xenon Pin Equivalent of 160 pCi/cc.
l l
..