ML24012A042

FAQ 23-03 Turkey Point IE01 Proposed NRC Response

ML24012A042
Person / Time
Site:	Turkey Point
Issue date:	11/16/2023
From:	Mack K Florida Power & Light Co
To:	Ronald Cureton, Theodore Fanelli NRC/NRR/DRO/IRAB
References
Download: ML24012A042 (7)

Text

Plant: Turkey Point Unit 3

Date of Event: 9/22/23

Submial Date: 11/16/23

Licensee

Contact:

Ken Mack Tel/email: Kenneth.Mack2@FPL.com

NRC Contacts: Theo Fanelli Tel/email: Theodore.Fanelli@NRC.gov

Ron Cureton Ronald.Cureton@NRC.gov

Performance Indicator: IE01, Unplanned Scrams per 7000 Crical Hours

Site-Speci"c FAQ (see Appendix D)? (X)Yes or ( ) No

FAQ requested to become eecve (X) when approved or (other dat e) ____________

Queson Secon

NEI 99-02 Guidance needing interpretaon (include page and line citaon):

• NEI 99-02, Revision 7, Page 9, Lines 6-7 Licensees can reduce the likelihood of a reactor accident by maintaining a low frequency of these iniang events.

• NEI 99-02, Revision 7, Pg 9 lines 21-22 This indicator monitors the numb er of unplanned scrams. It measures the rate of scrams per year of operaon at power and provides an indicaon of iniang event frequency.

• NEI 99-02, Revision 7, Page 9, Lines 25-26 The number of unplanned scrams during the previous four quarters, both manual and automac, while crical per 7,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br />.

• NEI 99-02, Revision 7, Appendix D, Page D-1, Lines 14-16 Then, if the licensee believes that there are unique circumstances sucient to warrant an excepon to the guidance as wrie n, the licensee should submit a Frequently Asked Queson to NEI for consideraon at a public meeng with the NRC.

• NEI 99-02, Revision 7, Appendix E, Page E-1, Line 12, 18-19 There are several reasons for subming an FAQ: 3. To request an exempon from the guidance for plant-speci"c circumstances, such as design features, procedures, or unique condions.

Event or circumstances requiring guidance interpretaon:

On September 22, 2023 with Unit 3 at 100% power, a grid disturbance caused by a lightning strike in the switchyard, resulted in a generator lock out. The reactor protecon system automacally actuated, and the unit tripped as designed. Subsequent analysis iden"ed that due to the magnitude of the lightning strike, an induced voltage was created through the switchyard ground grid that arced across the relay contact that drives the string bus dierenal relay, resulng i n a generator lockout.

To beer understand the magnitude of the September 22, 2023, li ghtning strike, Florida Power and Light (FPL) used measurements provided by Vaisala weather services. Vaisala is a large weather services company with over 80 years of experience, currently operang in over 150 countries. The September 22, 2023, strike was of signi"cantly greater magnitude than others historically observed in the area. The lightning strike that caused the scram was measured as 162,000 amperes (kA). The mean lightning strike magnitude was 37 kA. The magnitude of the September 22, 2023, l ightning strike is illustrated below.

Figure-1 is based on a data set of lightning strikes occurring within a half mile radius of the Turkey Point switchyard over the last ten years. As is shown, this lightning strike was greater than four standard deviaons (99.85 percenle) above the mean making this lightnin g strike an extreme stascal outlier.

On September 22, 2023, the lightni ng arrestor system was fully operaonal. Grounding grid integrity tesng is performed every six years in accordance with Staon A rea Operaons Instrucons. The latest tesng results from 2022 showed n o signs of degradaon of the e quipment that could have contributed to the event. Addionally, the switchyard and lightning arresto r system was in full compliance with design standards; namely, IEEE 998, IEEE Guide for Direct Lightning Stroke Shielding of Substaons and IEEE 80, IEEE Guide for Safety in AC Substaon Grounding. As expected, the substaon lightning protecon system absorbed and diverted the lightning strike to the substaon grounding system, prevenng it from reaching the primary conductors. This was evi dent by no faults on the primary conductors. As the lightning strike was being dissipated into the earth ground, the ground grid voltage in the staon rose with respect to remote earth. The switchyard gr ound grid funconed as designed allowing all the equipment in the switchyard to rise at the same relave rate and voltage.

The direct cause of the scram was the unprecedented magnitude of the lightning strike which created an induced voltage through the switchyard ground grid that arced across the string bus dierenal relay contact, resulng in a generator lockout. As the lightning stri ke itself was an extreme stascal outlier, this scram was caused by unique circumstances beyond FPLs ability to foresee or control and there is sucient basis to warrant excepon to the guidance as wrien in accordance with NEI 99-02, Revision 7, Appendix D, Page D-1.

According to SECY 99-007, Recommendaons for Reactor Oversight Process Improvements, an independent regulatory oversight process is one in which the agencys decisions are based on unbiased assessments of licensee performa nce. The ROP indicators are designed to provide the NRC with a framework to idenfy declining performance. NRC Inspecon Manua l Chapter 0305, Page 1, states that the Operang Reactor Assessment Program collects informaon fr om inspecons and performance indicators (PIs) to enable NRC to develop objecve conclusions about a licensees safety performance.

FPL recognizes that Performance Indicator IE01, Unplanned Scrams per 7000 Crical Hours does not generally dierenate scrams which should or should not be coun ted based on the cause of the scram.

However, FPL seeks clari"caon of whether counng the September 22, 2023, scram would align with the purpose of the indicator. In general, IMC 308 and SECY-99-007 provide a framework for establishing NRC indicators that are risk informed and performance based. IMC 308 states, The concept for seng performance thresholds includes consideraon of risk and regula tory response to dierent levels of licensee performance. SECY-99-007, Appendix H, provides the basis for the establishment of risk-informed performance indicator thresholds. In these basis docum ents, consideraon was given to liming false posives which would unnecessarily trigger increa sed regulatory oversight. As described in SECY-99-007, Secon H.2, When est ablishing the thresholds, it was taken as guiding principles that they should not result in a large number of false posives (resource concern), and that thresholds should be set to capture meaningful changes. Counng the September 22, 2 023, lightning strike scram would bring Turkey Point one step closer to increased regulatory oversight and the associated resource burden without a corresponding nexus to safety.

According to NEI 99-02, Revision 7, Page 9, the frequency of iniang events is an indicaon of the licensees ability to minimize plant vulnerability. Page 9 also states the purpose of the Unplanned Scrams per 7000 Hours indicator is to provide an indicaon of iniang events frequency. The only similar event to this one is from a potenal lightning induced scram in 1985. The Licensee Event Report (LER) for the 1985 scram (ML20134G691) does not state a lightning strike was the de"nive cause of the scram, only that heavy electrical and rain storms were in progress and that lightning was the most probable cause.

Prior to September 22, 2023, a lightning strike in the switchyard inducing a spurious scram such as this has not occurred at Turkey Point. Therefore, this scram is not an indicaon of Turkey Point Nuclear Plants ability to minimize plant vulnerability. Addionally, a s the lightning strike itself was an extreme stascal outlier, nor does it provide an indicaon of Turkey P oints performance. As such, the September 22, 2023, scram is a unique condion caused by plant-speci"c circumstances as described by NEI 99-02, Revision 7, Appendix E, Page E-1 and FPL believes counng this scram against the Unplanned Scrams per 7000 Crical Hours indi cator would not align with the indicators purpose.

This interpretaon is consistent with NRC guidance regarding Pe rformance Indicator IE03, Unplanned Power Changes per 7,000 Crical Ho urs. Indicator IE03 provides an excepon for unique environmental condions, stang:

Unique environmental condions which have not been previously e xperienced and could not have been ancipated and migated by procedure or plant modi"ca on, may not count, even if they are reacve.

This excepon was incorporated into NEI 99-02 Revision 7 even t hough there is similar language in SECY-99-007 regarding the counng of unplanned power changes and scr ams. In the case of both unplanned scrams and power changes SECY-99-007 Appendix A states the meas ure is a count of the number of events. SECY-99-007 regards both indicators as a rate and is not concerned with the cause of either scrams or power changes.

The September 22, 2023, scram is not indicave of Turkey Point s performance and counng it would not align with the purpose of the Unplanned Scrams Per 7000 Crical Hours indicator. FPL requests NRC guidance on whether this scram should be counted.

If licensee and NRC resident/region do not agree on the facts a nd circumstances, explain:

N/A

Potenally relevant FAQs:

FAQ 20-06: The licensee proposed counng two scrams as a single scram because the scrams had a related cause. In this case the NRC found the scrams in queson had suciently dierent causes such that they should count as separate scrams. The NRC cited IMC 308 and SECY-99-007 indicang the IE01 performance indicator is a count of the number of unplanned scrams, manual or automac, that have occurred regardless of cause. FPL understands and agrees performance indicator IE01 does not generally dierenate scrams which should or should not be counted based on the cause of the scram. However, FPL believes there are cases when the cause of a scram should be considered when determining whether or not to count a scram. For example, scrams that are not performance based, beyond the licensees ability to foresee and control, and represent outlier events, should not be counted. FPL believes this excepon is in alignment with the guidance contained in NEI 99- 02, the purpose of the IE01 performance indicator, and the bases for the indicator described in IMC 308 and SECY-99-007.

Resp ons e S e con

Proposed Resoluon of FAQ:

The NRC sta completed their evaluaon of this FAQ by reviewing the details of the event provided, LER 05000250/2023-03, informaon in Inspe con Manual Chapter (IMC) 0308, Aachment 1, Technical Basis for Performance Indicators, and guidance provided in NEI 99-02, Revision 7.

The purpose of the IE01, Unplanned Scrams per 7,000 Crical Ho urs, performance indicator, as stated in NEI 99-02, Revision 7, and IMC 0308, Aachment 1, is to moni tor the number of unplanned scrams. It measures the rate of scrams per year of operaon at power and p rovides an indicaon of iniang event frequency.

The review of this FAQ will focus on if granng an exempon for the September 22. 2023 scram counng against IE01, Unplanned Scrams per 7,000 Hours is warranted.

Guidance for this review is contained in NEI 99-02, Rev. 7 in t he following areas:

• Page 9 - 11

• Appendix C, Background Informaon and Cornerstone Development, Lines 10-33

NEI 99-02, Revision 7, Page 9, lines 11-14

Unplanned scram means that the scram was not an intenonal part of a planned e voluon or test as directed by a normal operang or test procedure. This includes scrams that occurred during the execuon of procedures or evoluons in which there was a high chance of a scram occurring but the scram was neither planned nor intended.

NEI 99-02, Revision 7, Page 10, lines 36-41 and Page 11, lines 1-18

Examples of the types of scrams that are included:

• Scrams that resulted from unplanned transients, equipment failures, spurious signals, human error, or those directed by abnormal, emergency, or annunciator response procedures.

• A scram that is iniated to avoid exceeding a technical speci"c aon acon statement me limit.

• A scram that occurs during the execuon of a procedure or evolu on in which there is a high likelihood of a scram occurring but the scram was neither planned nor intended.

Examples of scrams that are not included:

• Scrams that are planned to occur as part of a test (e.g., a rea ctor protecon system actuaon test), or scrams that are part of a normal planned operaon or evoluon.

• Reactor protecon system actuaon signals or operator acons to trip the reactor that occur while the reactor is sub-crical.

• Scrams that are iniated at less than or equal to 35% reactor p ower in accordance with normal operang procedures (i.e., not an abnormal or emergency operan g procedure) to complete a planned shutdown and scram signals that occur while the reactor is shut down.

• Plant shutdown to comply with technical speci"caon Liming Con dion for Operaon (LCO), if conducted in accordance with normal shutdown procedures which include a manual scram to complete the shutdown

During the September 22, 2023, Tur key Point Unit 3 scram, from 100% power; a lightning strike caused a grid disturbance which resulted in a generator lock out. The reactor protecon system automacally actuated, and the unit tripped. The NRC sta acknowledges the licensees asseron t hat the magnitude of the lightning strike was more signi"cant than what has historically been experienced in the area, however, due to the scram being unintenonal and not part of a planned evoluon or test, this scram would be considered unplanned under NEI 99-02, Revision 7. The NEI 99-02, Revision 7, list of examples that would not count as an unplanned scram are all related to planned evoluons that would result in a scram.

NEI 99-02 Revision 7, Appendix C, lines 10-33

INITIATING EVENTS CORNERSTONE

GENERAL DESCRIPTION

The objecve of this cornerstone is to limit the frequency of t hose events that upset plant stability and challenge crical safety funcons, during shutdown as well as p ower operaons. When such an event occurs in conjuncon with equipment and human failures, a react or accident may occur. Licensees can therefore reduce the likelihood of a reactor accident by maintaining a low frequency of these iniang events. Such events include reactor trips due to turbine trip, loss of feedwater, loss of osite power, and other reactor transients. There are a few key aributes of lic ensee performance that determine the frequency of iniang events at a plant.

PERFORMANCE INDICATORS

PRAs have shown that risk is oen determined by iniang events of low frequency, rather than those that occur with a relavely higher frequency. Such low-frequency, h igh-risk events have been considered in selecng the PIs for this cornerstone. All of the PIs used in this cornerstone are counts of either iniang events, or transients that could lead to iniang events (see T able in the main body of NEI 99-02). They have face validity for their intended use because they are quan"able, have a logical relaonship to safety performance expectaons, are meaningful, and the data are readi ly available. The PIs by themselves are not necessarily related to risk. They are however, the "rst step in a sequence which could, in conjuncon with equipment failures, human errors, and o-normal plant con"guraons, result in a nuclear reactor accident. They also provide indicaon of problems that, if unc orrected, increase the risk of an accident. In most cases, where PIs are suitable for idenfying problems, the y are sucient as well, since problems that are not severe enough to cause an iniang event (and therefore result in a PI count) are of low risk signi"cance. In those cases, no baseline inspecon is required (the excepon is shutdown con"guraon control, for which supplemental baseline inspecons is necessar y).

NEI 99-02, Appendix C discusses in iang events cornerstone per formance indicators and states that risk is oen determined by iniang events of low frequency, rather than those that occur with a relavely higher frequency and that low-frequency, high-risk events have been considered in selecng the performance indicators for the cornerstone. For the scram on September 22, 2023, the lightning strike was a low frequency iniang event that resulted in a reactor s cram. The objecve of the Iniang Events cornerstone, as described in NEI 99-02 Revision 7, is to limit the frequency of those events that upset plant stability and challenge crical safety funcons dur ing power operaons.

Appendix C also states that the PIs used in this cornerstone are counts of either iniang events, or transients that could lead to iniang events. They have face v alidity for their intended use because they are quan"able, have a logical relaonship to safety performanc e expectaons, are meaningful, and the data are readily available. The performance indicators by themselves are not necessarily related to risk.

They are however, the "rst step in a sequence which could, in conjuncon with equipment failures, human errors, and o-normal plant con"guraons, result in a nuc lear reactor accident. For the Turkey Point scram, the lighng strike was an external event that caus ed an unplanned scram as de"ned in NEI 99-02, Revision 7. The counng of this event has face validity given it is quan"able, had a logical relaonship to the safety perfor mance of the plant, is meaningful, and the data is readily available.

Exempon from this event counng against IE01 Unplanned Scrams per 7,000 Crical Hours is not granted since scrams are a "rst step in a sequence that places reliance on equipment and human performance as responding migang systems, increasing the risk of nuclear accidents.

If appropriate, provide proposed rewording of guidance for inclusion in next revision:

N/A

PRA update required to implement this FAQ?

No

MSPI Basis Document update required to implement this FAQ?

No