ML14204A708

Enclosure 2, Description of Closure Evaluation, Non-Proprietary Information - Class I (Public)

ML14204A708
Person / Time
Site:	Millstone, Hatch, Monticello, Dresden, Peach Bottom, Browns Ferry, Nine Mile Point, Perry, Fermi, Oyster Creek, Hope Creek, Grand Gulf, Cooper, Pilgrim, Susquehanna, Columbia, Brunswick, Limerick, River Bend, Vermont Yankee, Duane Arnold, Clinton, Quad Cities, FitzPatrick, LaSalle
Issue date:	07/23/2014
From:	GE-Hitachi Nuclear Energy Americas
To:	Office of Nuclear Reactor Regulation
Shared Package
ML14204A706	List: ML14204A707 ML14204A708 ML14204A709
References
MFN 14-010 R1
Download: ML14204A708 (9)

Text

ENCLOSURE 2 MFN 14-010 R1 Description of Closure Evaluation Non-Proprietary Information - Class I (Public)

IMPORTANT NOTICE This is a non-proprietary version of Enclosure 1 to MFN 14-010 R1, from which the proprietary information has been removed. Portions of the enclosure that have been removed are indicated by open and closed double square brackets as shown here

[[ ]].

to MFN 14-010 R1 Non-Proprietary Information - Class I (Public)

July 23, 2014 Page 1 of 8 Summary GE Hitachi Nuclear Energy (GEH) has evaluated the discovery of a Scram Solenoid Pilot Valve (SSPV) with a disengaged valve spring based on available information and determined the condition is not reportable for the Columbia Generating Station (Columbia) and the Enrico Fermi Unit 2 Nuclear Facility (Fermi 2). However, GEH does not have sufficient information to fully evaluate the safety significance of the condition under 10 CFR Part 21 for the remaining potentially affected plants. Therefore, GEH will close this 10 CFR Part 21 evaluation with a Transfer of Information to all potentially affected licensees per 10 CFR 21.21(b). In parallel with the GEH evaluation, the SSPV manufacturer has issued an Interim Report and a Final Report, providing information resulting from their evaluation.

Effects of the condition were observed at the Columbia and Fermi 2 plants and attributed to a single affected SSPV at each plant, with extents limited to only those affected SSPVs.

Because the cores for those plants were designed to ensure adequate shutdown margin with the strongest control rod remaining fully withdrawn, the effects of the condition were accommodated by design, and there was no impact on the ability to shutdown the reactor.

For the remaining plants, GEH does not know the quantity of any affected SSPVs or their plant-specific locations, which are necessary to determine impact of the condition on the ability to shutdown an affected plant. In some scenarios adequate shutdown margin is assured regardless of postulated control rod behavior, but for other scenarios impact on shutdown margin would depend on the number of affected SSPVs, plant-specific core configuration, and the location and cycle-specific worth of any affected control rods. In all cases, diverse backup functions, which are not safety-related, but are proven to be reliable, would likely activate or be available for manual action to ensure control rod insertion.

The GEH evaluation of potential effects consisting of slow scram insertion of affected control rods during a full-core scram indicates that compensatory measures and defense in depth provided by the backup scram function and the alternate rod insertion function, which are not safety-related but are proven reliable, ensure that affected plants will be able to shutdown in accordance with requirements. A previous GEH evaluation indicates that [[° ° ° °

°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°° °°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°

° ° ° ]] would not lead to violation of the Technical Specifications safety limit for Minimum Critical Power Ratio (MCPR).

Introduction This communication includes background material provided in MFN 14-010 R0 (for completeness), but updates the evaluation and conclusions.

On January 15, 2014, GEH was informed of the failure of a control rod to scram (i.e., no control rod motion) during a single rod scram test at a U.S. BWR/5 plant (Plant E in Table 1).

Plant personnel conducted an investigation and determined that the 118 SSPVs in the Hydraulic Control Unit (HCU) for the affected control rod did not function. They inspected the valve and discovered that the spring, normally connected to valve plunger (or core), had to MFN 14-010 R1 Non-Proprietary Information - Class I (Public)

July 23, 2014 Page 2 of 8 become detached. Such a condition would not place the valve into its isolated/vented configuration upon deactivation in response to the Reactor Protection System (RPS) scram signal, and scram performance would be degraded, with a delayed insertion proceeding through function of the backup scram valves (or, if necessary, manual activation of the Alternate Rod Insertion (ARI) system or by other manual methods). In the case of a single rod scram test, the backup scram valves do not open and the condition would leave the control rod unmoved in the pre-scram position; this was the behavior observed at Plant E.

The valve was returned for inspection to the SSPV manufacturer (ASCO Valve, Inc.), who initiated their own 10 CFR Part 21 investigation into the apparent condition of the valve and the extent of condition. The manufacturer completed their investigation and followed their Interim Report (March 11, 2014, available on the U.S. NRC website under ADAMS accession number ML14079A026) with their Final Report, submitted to the U.S. NRC on June 16, 2014 (ADAMS accession number ML14169A258), and results from the investigation have been incorporated into the GEH evaluation of safety significance in accordance with 10 CFR Part 21. Details provided herein present the information available to GEH and support the GEH determination that GEH does not have sufficient information to completely evaluate the safety significance of the condition, as required by10 CFR Part 21 for some potentially affected plants.

Description of Discovery A review of available industry operating experience and recent GEH experience indicates that this condition was observed and reported 5 times: 1993, 1994 (two occurrences), 2012, and 2014. The salient information from the review is summarized in Table 1.

Summary of Observed Behaviors In observations from 1993, 1994 (2), 2012, and 2014, the condition was discovered during normal plant operation within months after valve installation but in only one case during post-maintenance testing just after installation of the affected valve.

o The 2012 and 2014 occurrences of the condition were evident within 7 months of valve installation, after cycling roughly 60 to 200 times in half-scram tests (during which the condition would not be evident, because only one of the pair of SSPVs is cycled at a time).

o Two of the three occurrences from 1993 and 1994 also appear to have been discovered within a short time after valve installation (immediately in post-installation testing in one case, and shortly after HCU maintenance in another case).

When the condition is present:

o During a full-core scram, the rod fully inserts but is delayed, often outside of the insertion time required by Technical Specifications. This is because the backup scram valves depressurize the scram air header, which allows the scram valves to open and the rod to insert, but the rate is slower than to MFN 14-010 R1 Non-Proprietary Information - Class I (Public)

July 23, 2014 Page 3 of 8 insertion triggered with the SSPV function. (If the backup scram function, which is activated by the RPS were to not function for some reason, the operator could then manually activate the ARI system, which would also depressurize the scram air header to insert affected rods, but at a delayed rate and after the duration needed to trigger the system manually. Note that the backup scram function and the ARI system are not safety related.)

o During a single-rod scram test, the control rod fails to move upon scram signal.

Table 1 - Summary of SSPV Spring Disengagement History Date Plant Observations March 1993 Plant A Slow scram during full-core scram from operation BWR/3 Shortly after Control Rod Drive (CRD) HCU maintenance April 1994 Plant B Slow scram during full-core scram from full-power operation BWR/2 Had operated successfully on previous occasions after installation (time to failure not provided)

November 1994 Plant C Failure to scram in low-power Tech Spec testing during post-outage BWR/3 startup Immediately after installation/refurbishment October 2012 Plant D Passed scram-time test during start-up after installation BWR/4 Cycled 87 times in half-scram testing over 2 months Normal insert from notch 12 during full-core scram 2 months after installation Cycled 112 additional times in half-scram testing over next 3 months Slow to scram during full-core scram (2.5-sec delay) 5 months after installation (scheduled for testing)

Failure to scram during single-rod test 6 months after installation January 2014 Plant E Passed two scram-time tests during start-up after installation BWR/5 Cycled roughly 60 times in half-scram testing over 7 months Failed to scram (no movement) in single-rod test 7 months after installation Extent of Condition Based on the manufacturers bounding of the condition, GEH has defined the suspect population as 437 SSPVs assembled during a period extending from 2010, when the assembly operation was moved to a new location, to November 2012, when an assembly inspection step was added to confirm spring engagement. GEH has correlated internal records with those of the manufacturer to identify the serial numbers of SSPVs delivered to plants sites, as provided in Table 2. From the population of 437 SSPVs, a total of 399 SSPVs remain at the plant sites, based on the latest information provided to GEH.

(Excluded from Table 2 are an additional 38 valves from the suspect population that were received by a plant but later returned to the manufacturer.)

to MFN 14-010 R1 Non-Proprietary Information - Class I (Public)

July 23, 2014 Page 4 of 8 Table 2 - Serial Numbers and Receiving Plants for SSPVs from Suspect Population S/N Plant Name Customer PO Base S/N Extensions Ship Date QTY 00073419 140715 A762654 -001, -002, -003 4-Apr-11 3 Browns Ferry Plant Total: 3

-008, -012, -021, -027 -028, -031, -034, 337332 A272718 -035, -036, -045, -046, -048, -049, -050, 3-May-13 20

-052, -053, -057, -060, -062, -065 Columbia Note: This quantity for Columbia is for SSPVs remaining at the site, all of which were inspected for the condition and cleared. An additional 36 valves from the suspect population were initially received and then returned by Columbia, cleared by inspection, and are not Plant Total: 20 included in this table.

00000707 13418 A171767 -001, -002 25-Oct-10 2 Dresden Plant Total: 2

-001 thru -009, -011 thru -014, -016 thru -024, A113888 -026 thru -050, -052 thru -067, -069, -070, 8-Feb-12 118

-072, -073 thru -081, -083 thru -125

-007 thru -009, -014, -025, -031, 31-Jan-12 9 A171767 -034, -037, -045 4700381095 -030 22-Aug-11 1 Fermi A321861 -021 thru -029 22-Aug-11 9

-005, -006, -010, -012, -030 thru -037, A321861 31-Jan-12 20

-041 thru -047, -050 A351454 -001 thru -005 28-Sep-12 5 Note: This quantity for Fermi is for suspect SSPVs remaining at the site, and does not include two suspect valves previously returned to the manufacturer and replaced with two valves not Plant Total: 162 from the suspect population. All SSPVs at Fermi were inspected and cleared of the condition.

-003 thru -006, -010 thru -013, -015, -020, -

022, -023, -024, -026, -027 thru -029, A171767 23-Aug-10 33

-032, -033, -035, -036, -038 thru -044,

-046 thru -050 80 034650

-001 thru -004, -007 thru -009, A321861 -011, -013, -014, -017, -018, -038, 23-Aug-10 17

-039, -040, -048, -049 Oyster Creek 80 045972 A276160 -006 thru -011, -013 thru -033 26-Jun-12 27 A272718 -001 thru -006 5-Nov-12 6 80 048306 A276160 -002, -004 11-Sep-12 2 80 045972 A272718 -022, -023, -024 6-May-13 3 Plant Total: 88 A171767 -016 thru -019 5-Jan-11 4 90 258497 427 Peach Bottom A321861 -015, -016, -019, -020 5-Jan-11 4 Plant Total: 8 00000707 13765 A272718 -066 thru -071 11-Sep-13 6 Quad Cities Plant Total: 6 8981120150 A496905 unknown 9-Dec-10 2 Non-U.S. 8991120188 A793691 -001 thru -108 26-Sep-11 108 Plant Note: The A496905 quantity for the Non-U.S. Plant is for suspect SSPVs remaining at the site and does not include 104 SSPVs from that original shipment that were returned to the manufacturer for a different issue and received a spring inspection prior to reshipment under Plant Total: 110 a new S/N. Those 104 SSPVs are not included in the evaluation scope.

Combined Total: 399 to MFN 14-010 R1 Non-Proprietary Information - Class I (Public)

July 23, 2014 Page 5 of 8 Implications for Safety Significance The limited number of observations of the condition from the current suspect population and the consistency of time to identify the condition for the five occurrences (recent and 1990s) suggest this condition occurs with low frequency and relatively early identification. However, the possibility of other affected SSPVs from the suspect population is indeterminate and cannot be quantified without full inspection of each suspect SSPV. For that reason, the degree to which the suspect population is affected by the condition is uncertain, and the safety significance of the condition, which depends on the number of control rods affected and their plant-specific locations, cannot be assessed. However, some comments on conditions and potential effects can be made.

For cases of any affected but undetected SSPVs, the ability to safely shutdown the plant and maintain shutdown will depend on whether the functioning control rods maintain sufficient shutdown margin. In the U.S., and in other countries with similar regulations and practices, reactor cores and fuel cycles are designed to maintain adequate shutdown margin at all times with the highest-worth control rod fully withdrawn. This ensures that a plant unknowingly affected with a single control rod of this condition (typical of the cases observed thus far), or other conditions with similar effects on scram, will meet shutdown requirements throughout the operating cycle, during cold shutdown, and during refueling; in such a case a single control rod affected by a malfunctioning SSPV will not prevent the safe shutdown of the reactor. If that provision is not incorporated into the core design or if multiple control rods are concurrently impacted, shutdown requirements are likely to be met due to the function of the backup scram valves. The backup scram function is activated by the RPS and ensures the scram air header is depressurized to allow the scram valves to open and control rods with non-functioning SSPVs to insert. If the backup scram function were to fail, operator action to activate the ARI system would also depressurize the scram air header, allowing affected control rods to insert. Activation of the ARI system must be initiated in sufficient time to allow the rods to insert before the scram discharge volume fills. Control rod insertion by either of these functions is slower than that triggered through the SSPVs, due partly to the delay in start of motion and also to the slower rate of insertion. Finally, the operator has the option of using other, longer-term manual insertion methods, such as the normal control rod positioning function. Although some equipment used for the backup scram function, the ARI system, and the other insertion methods is not safety-related, these functions are redundant, independent, and/or diverse to the normal RPS scram function.

Furthermore, many years of reactor operation have demonstrated these functions to be reliable under normal and encountered off-normal conditions. If those systems were to fail with multiple control rods concurrently affected by this condition, then shutdown requirements might not be met; this would depend on plant-specific details, including the core locations of the affected rods.

For cases in which affected control rods are inserted by the function of the backup scram valves, the scram insertion time could exceed Technical Specifications scram time limits (i.e., > 7 seconds). However, a previous GEH evaluation of control rod performance [[° ° ° ° ° °

°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°

°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ]] will to MFN 14-010 R1 Non-Proprietary Information - Class I (Public)

July 23, 2014 Page 6 of 8 have negligible effect on critical power ratio for the limiting fuel assembly. [[° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° °

°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°

°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°° ° ° ° ° ° ° ° ° ° ° ]] will not pose a concern for shutdown of the plant and will not lead to violation of the Technical Specifications safety limit for MCPR.

Synopsis What is known about the condition can be summarized as follows:

The effects of the condition result from a latent tendency for an affected SSPV spring to disengage from the valve plunger due to an assembly error that escaped detection by the manufacturer.

If a misassembled SSPV spring disengages, the affected control rod is observed to insert slowly during a full-core scram or to not move at all during a single-rod scram test.

A number of BWR plants have received GEH-shipped SSPVs from a population identified by the manufacturer as being associated with valve spring disengagement.

o Two occurrences of valve spring disengagement have been discovered from the identified population of 437 potentially susceptible SSPVs. (At the time of the GEH Interim Report notification of this issue in March 2014, 399 of those suspect SSPVs remained at plant sites.)

o Industry operating experience indicates three other known or likely occurrences of valve spring disengagement in the 1990s, which are not associated with the currently identified population.

Of the five known occurrences to date (three in the 1990s and two within the past two years):

o Valve spring disengagement was detected in only one occurrence at each of five plants.

o The frequency at which this defect exists in any of the current suspect SSPVs is estimated at just under 1%, based on inspection results from 220 suspect SSPVs from the two plant subpopulations that contained the 2 observed defects. The probability that this defect exists in more than one suspect SSPV in a single plant is also unknown, but is lower than the probability of the defect existing in a single suspect SSPV at the same plant. Finally, the probability of two or more control rods with affected SSPVs concurrently failing to scram is lower yet.

Discovered occurrences of the disengaged spring were detected within a relatively short time after SSPV replacement installation and specifically within 7 months (or 60 to 200 valve cycles) for the two recent occurrences.

to MFN 14-010 R1 Non-Proprietary Information - Class I (Public)

July 23, 2014 Page 7 of 8 o However, the manufacturers investigation of the failure mechanism was not able to decisively conclude whether the condition would reliably become apparent within a short time or a limited number of valve cycles after installation (i.e., it could not consistently replicate behavior consistent with an SSPV spring disengaging within a single plant operating cycle as observed in at least 4 of the 5 occurrences from 1993, 1994, 2012, and 2014).

Safety significance of this condition, if present, depends on the number of affected control rods and, if multiple control rods in a plant are concurrently affected, on their location. But some general points can be made:

o For U.S. plants, with core loadings designed to maintain adequate shutdown margin with the strongest rod fully withdrawn, failure of one control rod to insert, for whatever reason, will not prevent the safe shutdown of the reactor.

o Shutdown margin is otherwise provided reliably by the backup scram function and the ARI system, which serve to depressurize the scram air header and allow control rods with SSPV failure to insert. Operators might also choose to insert un-scrammed rods using the normal control positioning function.

These alternate methods use equipment that is not safety-related, but has proven reliable over many years of reactor operation.

o Scram insertion initiated by backup scram function or ARI could exceed Technical Specification scram times, but in a previous GEH evaluation, a condition [[° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° °

°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°° °°°°°°°°°°

° ° ° ° ° ° ° ° ° ° ° ° ° ]] was shown to have negligible effect on critical power performance of the limiting fuel assembly.

Because the condition affected only a single control rod each at the Columbia and Fermi 2 plants, and because the cores for those plants were designed to ensure adequate shutdown margin with the strongest control rod remaining fully withdrawn, the effects of the condition were accommodated by design and there was no impact on the ability to safely shutdown the reactor.

GEH is not able to evaluate the safety significance of this condition with the limited information available for the remainder of the potentially affected plants. Specifically, GEH does not know the following:

the number of shipped suspect SSPVs at those plants that are actually affected; the time needed for the condition to manifest with an impact on scram performance; and the specific core locations and relative worth of control rods and HCUs with potentially affected SSPVs.

to MFN 14-010 R1 Non-Proprietary Information - Class I (Public)

July 23, 2014 Page 8 of 8 ABWR and ESBWR Design Certification Documentation Applicability The issue described herein has been reviewed for applicability to documentation associated with 10 CFR 52, and determined to have no effect on the technical information contained in either the ABWR certified design or the ESBWR design in certification. This is true because this condition is not a deficiency in design.

Recommendations GEH suggests the following:

Customers who received shipments of the suspect SSPVs but have not inspected those SSPVs for spring engagement should consider coordinating with GEH to return suspect SSPVs to the manufacturer, ASCO Valve, Inc., for inspection, as practical and appropriate. Inspections can be performed at the plant site, but the supplier cautions against that because of the potential for valve nylon thread lock debris to impact operation of the valve.

Plant staff can consider whether more-frequent testing of CRDs with suspect SSPVs is beneficial to identify presence of the condition for appropriate action. However, staff should be advised that effects are not gradual to allow early identification before SSPV malfunction (i.e., the observable characteristic is scram performance affected by the malfunction).

GEH assistance on this issue can be obtained by contacting Rich Jones (richard2.jones@ge.com; 910-819-6043, or 910-228-4602).

Corrective/Preventive Actions The manufacturer, ASCO Valve, Inc. issued an Interim Report, (Interim Report on HVL26600000010J Valve, March 11, 2014, U.S. NRC ADAMS Accession Number ML14079A026) with stated intention to complete their investigation by the end of June 2014.

Their investigation was completed and their final report transmitted to the U.S. NRC and to GEH on June 16, 2014. The final report can be found on the U.S. NRC website under ADAMS Accession Number ML14169A258.

In the suppliers Interim Report, the supplier stated that future occurrence of this condition will be prevented by an additional inspection step, incorporated November 9, 2012, to confirm proper spring engagement during assembly.