ML14211A213
ML14211A213 | |
Person / Time | |
---|---|
Site: | Oconee ![]() |
Issue date: | 07/31/2014 |
From: | Batson S Duke Energy Carolinas |
To: | Division Reactor Projects II |
References | |
Download: ML14211A213 (52) | |
Text
Regulatory Conference Oconee Nuclear Station July 31, 2014 Agenda Opening Remarks Scott Batson Site Vice President Event Summary Ed Burchfield General Manager, Engineering Risk Perspective Ed Burchfield General Manager, Engineering Paul Fisk Operations Manager Karl Fleming Consultant Cliff Lange Consultant Root Causes/Corrective Actions Leo Martin General Manager, Engineering Plant Support Regulatory Significance Bob Guy Director, Site Support Closing Comments Preston Gillespie Senior Vice President 2 For Information Only Opening Remarks Scott Batson Site Vice President 3 For Information Only Opening Remarks Opening Remarks Duke Energy agrees with the apparent violation described in the NRC letter dated June 26, 2014 Duke Energy understands importance of fission product barrier integrity Performance did not meet our expectations for identifying the flaw independent industry expert analysis Organizational and programmatic improvements have been initiated Robust corrective actions implemented for the site and for the fleet 4 For Information Only Event Summary Ed Burchfield General Manager, Engineering 5 For Information Only Event Summary Event Summary Leakage identified promptly at very low rate Aggressive actions taken to identify and resolve changes in RCS leakage Operator response actions prudent and timely Procedures Training Controlled plant shutdown in accordance with Technical Specifications RCS leakage detection threshold very low (hundredths of a gpm) Containment sump readings Process monitors Daily management review Safe conservative decision-making based on early detection and prompt actions HPI nozzle leak was on injection line that does not normally have flow For Information Only 6 7 Event Summary: Photograph For Information Only 8 Event Summary: Flow Diagram Of HPI System For Information Only Event Summary: Timeline Detection Timeline November 8, 2013 @ 1837 containment process monitor alarm received November 8, 2013 @ 2324 RCS leakage calculation shows 0.020 gpm unidentified leakage November 9, 2013 @ 0442 entered procedure for radiation monitor response, leakage increased to 0.126 gpm over next 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> November 10, 2013 @ 0005 containment entry made at power; discovered water on reactor building floor, source not identified November 10, 2013 @ 1630 containment entry made to obtain video images; source not identified November 10, 2013 @ 2141 initiated downpower to 20% to allow visual inspection November 11, 2013 @ 0520 identified RCS pressure boundary leak at HPI nozzle; entered TS 3.4.13 Condition B November 11, 2013 @ 1000 achieved Mode 3 9 For Information Only 10 Risk Perspective Ed Burchfield General Manager, Engineering Paul Fisk Operations Manager For Information Only Risk Perspective: Overview Duke Energy Perspective On Risk Duke Energy concludes very low safety significance Pipe flaws that rapidly progress to a LOCA have not occurred Common cause is a negligible contributor to overall risk Transient events do not significantly contribute to overall risk External events do not significantly contribute to overall risk Operator response to RCS leak is to perform controlled shutdown Independent, diverse and rigorous evaluations KNF Consulting Services (KNFCS) and Structural Integrity Associates (SIA) independent evaluations Multiple independent analyses support a very low safety significance 11 For Information Only Duke Energy Evaluations Conclude Very Low Safety Significance Duke Energy characterizes risk from the performance deficiency as being very low 2E-7 to 7E-7/yr Dominant risk scenario is a LOCA on the 2.5 inch HPI line with postulated failure to transition to high pressure recirculation and other random equipment failures 12 Risk Perspective: Very Low Safety Significance For Information Only Risk Perspective: Pipe Flaw Progression Pipe Flaws That Rapidly Progress To LOCA Have Not Occurred Risk from HPI nozzle weld flaws progressing to a rupture (LOCA) based on Extensive industry data analysis Probabilistic Fracture Mechanics NUREG-1829 (Estimating Loss-of-Coolant Accidents (LOCA) Frequencies Through Elicitation Process) Duke Energy calculates 2.5 inch HPI nozzle Conditional Rupture Probabilities (CRP) and LOCA probabilities Industry OE and fracture mechanics provide basis for slow crack growth Primary difference between Duke Energy and NRC 13 For Information Only Risk Perspective: Common Cause Common Cause Is Negligible Contributor To Overall Risk Root cause analysis confirms event was an independent pipe defect No industry data to support a common cause in piping welds Each weld is unique Stresses Weld installation Duke Energy concludes common cause failure of multiple HPI lines is not applicable to this configuration Result: Common cause is a negligible contributor to the risk analysis 14 For Information Only Risk Perspective: Transient Events Transient Events Do Not Significantly Contribute to Overall Risk Duke Energy calculates plant transient impact Reactor trips Loss of off-site power Loss of feedwater Support system failures, etc. Added stresses, and duration of added stresses, do not significantly challenge LOCA even when postulated leak is greater than 1 gpm The SIA analysis with applied stresses to cracks with 1 gpm and 10 gpm leak concludes significant margin before critical crack size reached Result: Plant transients are not a significant contributor 15 For Information Only Risk Perspective: External Events External Events Do Not Significantly Contribute to Overall Risk Duke Energy calculates external event impact Seismic Stress of seismic events is included in analysis, significant margin exists Duration of a seismic event is very short Fire Stress of HPI flow during events is included in analysis, significant margin exists Few fires require HPI for mitigation or cause a spurious HPI actuation Result: External events are not a significant contributor 16 For Information Only Risk Perspective: Operator Response Operator Response To RCS Leak Is To Perform Controlled Shutdown Operators are trained to minimize risk and conduct a controlled plant shutdown for this type of event Procedure AP/002 is presented in classroom, simulator labs and simulator practice scenarios for both initial and continuing training Simulator instructor guide contains explicit note that crew should not manually trip reactor for the 65 gpm leak scenario Since 2009, 29 crews have been evaluated with no instances of reactor trip outside of technical procedure criteria Duke Energy risk model input for operator response is controlled plant shutdown Result: Controlled plant shutdown at least a factor of 10 less risk significance than plant trip 17 For Information Only Risk Perspective: Operator Response 18 For Information Only Risk Perspective: Mid-Loop Risk Mitigation Mid-Loop Risk Mitigation Significant risk mitigation actions were implemented during the HPI repairs to eliminate human caused events Stopped all work on Unit 1 and restricted access to all areas of power block Used protected equipment for all methods of core cooling Restricted access to AC power sources Established RCS gravity fill vent path Suspended Unit 2 refueling outage work potentially affecting common systems 19 For Information Only Risk Perspective: Independent Evaluations Independent, Diverse And Rigorous Evaluations - State of the art code CRP for pipe with through-wall crack KNF Consulting Services update of LOCA probabilities Updated and corrected industry data on pipe defects Oconee-specific estimate CRP estimates from industry data and NUREG-1829 (Estimating Loss-of-Coolant Accidents (LOCA) Frequencies Through Elicitation Process) Risk evaluation of leak event 20 For Information Only Multiple Independent Analyses Support A Very Low Safety Significance Duke Energy risk analysis uses independent and diverse methods for determining CRP of the 2.5 inch HPI weld All methods support the same conclusion Slow crack growth Pipe flaws that rapidly progress to a LOCA have not occurred Very strong leak-before-break behavior of 2.5 inch HPI welds Consistent results of slow crack growth and low potential for sudden failure 21 Risk Perspective: Independent Analyses For Information Only Risk Perspective: Crack Growth And LOCA Probabilities Karl N. Fleming KNF Consulting Services 22 For Information Only Risk Perspective: Risk Characterization Of Event Change In Oconee HPI Line LOCA Frequency Calculated November 11, 2013 Oconee Unit 1 event included in LOCA frequency calculations Based on size of the pipe a more severe failure can produce a small LOCA (.LOCA (as defined in ONS Unit 1 PRA model 23 For Information Only Risk Perspective: Data Survey Of PWR Class 1 Pipe Defects Failure Mode Leak Rate (gpm) No. Events Non through-wall 0 268 Pin Hole Leak <<1 189 Small Leak <1 230 Leak 1 17 Large Leak 5 20 Very Large Leak 50 1 SLOCA 160 0 CCF Events - 0 Industry Events Have Not Resulted In A LOCA Survey accounts for 6,342 reactor-years through June 2014 Reactor coolant pressure boundary flaw data All Causes 24 For Information Only Risk Perspective: Data Survey Of LWR Vibration Fatigue Class 1 Pipe Defects Industry Data Shows No Vibration Fatigue Induced LOCAs Reactor coolant pressure boundary flaw data - Vibratory Fatigue 25 Failure Mode Leak Rate (gpm) No. Events Non through-wall 0 4 Small Leak << or <1 127 Leak 1 12 Large Leak 5 7 Very Large Leak 50 0 SLOCA 160 0 CCF Events - 0 For Information Only Risk Perspective: Insights From Data Review Industry Data Shows No LOCAs From Pipe Defects 725 pipe defects in 6,342 reactor-years of PWR experience through June 2014 26 pipe defects in high pressure injection lines No pipe defect involved a common cause failure of multiple components Bounding common cause failure < 7E-4 Conditional rupture probabilities supported by service data LOCA given a small leak in any location from any cause bounded by 2E-3 Oconee HPI line small LOCA frequency increase due to event 4E-5/yr Risk significance of event is change in core damage frequency associated with occurrence of leak event 26 26 For Information Only Risk Perspective: ONS Unit 1 SLOCA Initiating Event Frequency 27 Mean Change in LOCA Frequency from 2.4E-04/yr to 2.8E-04/yr is ~ 4E-05/yr For Information Only Risk Perspective: Evaluation Of Model Change In LOCA CDF Calculated For Event Impact of leak event on ONS specific SLOCA and MLOCA initiating event frequency quantified using established PRA methods CRP for SLOCA and MLOCA quantified using two methods Direct estimation from service data Information derived from NUREG-1829 (Estimating Loss-of-Coolant Accident (LOCA) Frequencies Through the Elicitation Process) CCDPs for SLOCA and MLOCA from ONS Unit 1 PRA provided by Duke Energy Change in risk evaluation addressed epistemic uncertainties 28 For Information Only Risk Perspective: Change In CDF Less Than 1E-6 Using Conservative NUREG-1829 CRP Model 29 For Information Only Risk Perspective: KNFCS Estimated Change Mean Change In CDF Calculated For Event 5E-7/yr based on NUREG-1829 (Estimating Loss-of-Coolant Accident (LOCA) Frequencies Through the Elicitation Process) derived CRP (~ 2E-2) 5E-8/yr based on data derived CRP (~ 2E-3) Database used in analysis is high quality with extensive worldwide input High confidence change is less than 1E-6/yr in light of epistemic uncertainties Complements the Structural Integrity Associates results with similar conclusion Change in LOCA CDF is very low safety significance 30 For Information Only Risk Perspective: Deterministic / Probabilistic Analyses And Results Cliff Lange Structural Integrity Associates 31 For Information Only Risk Perspective: Deterministic Analysis Deterministic Fracture Mechanics Analysis Objective Predict crack growth and leakage using 1B2 as-found crack geometry Methodology As-found crack modeled with ANSYS Normal, HPI full flow test and HPI split flow test vibration loads used for FCG calculations SI-PICEP code used to determine leak rates for various crack lengths Critical crack size determined by limit load analysis per ASME Section XI methodology SSE, reactor trip (high pressure) and HPI full flow test loads applied for stability checks 32 For Information Only Risk Perspective: Deterministic Results Deterministic Results From the as-found crack, it takes 23 days to develop 1.0 gpm and 33 days to develop 10 gpm leak rates under normal operating conditions Permits controlled plant shutdown From a 1.0 gpm leak rate, it takes 2.9 days to reach the critical crack size under HPI full flow vibration loads Permits controlled plant shutdown With a 10 gpm leak rate and SSE or reactor trip events, significant margin exists for additional crack growth prior to the crack becoming a critical size 33 For Information Only Risk Perspective: Probabilistic Analysis Probabilistic Fracture Mechanics Analyses Objective Develop leak and rupture probabilities for the 1B2 safe end-to-pipe weld Methodology Analyses performed using the Beyond-PRAISE Fracture Mechanics Code Normal operating and faulted load conditions included Statistical distributions based on industry data for materials and observed data for normal operating and HPI full flow test vibrations Multiple separate evaluations performed 34 For Information Only Probabilistic Fracture Mechanics Results Case #1 pipe rupture probability = 1E-6 Case #2 -wall: pipe rupture probability = 5E-6 Case #3 inches, 25% variability): pipe rupture probability = 6E-4 ): pipe rupture probability that is < 1E-7 Significant margins for leak-before-break maintained 35 Risk Perspective: Probabilistic Results For Information Only Ed Burchfield General Manager, Engineering 36 Risk Perspective: Duke Energy PRA Results For Information Only Risk Perspective: Oconee PRA Results Oconee PRA Results Reflect Very Low Safety Significance Oconee Change in CDF Results 2E-7 to 7E-7/yr Through-wall leak controlled shutdown change in CDF ~ 7E-8/yr Through-wall leak reactor trip change in CDF ~ 1E-8/yr Induced LOCA internal events change in CDF ~ 3E-8/yr Induced LOCA external events change in CDF ~ 7E-8/yr LOCA change in CDF ~1E-8 to 5E-7/yr Conclusion very low safety significance 37 For Information Only Risk Perspective: Oconee PRA Conclusion Conclusion Oconee risk evaluation Pipe flaws that rapidly progress to a LOCA have not occurred Common cause is a negligible contributor Transient and external events are not significant contributors Operator response to RCS leak is a controlled shutdown per plant procedure AP/002 Multiple and independent methods of analyses Oconee PRA results reflect very low safety significance 38 For Information Only Root Causes And Corrective Actions 39 Leo Martin General Manager, Engineering Plant Support For Information Only Root Causes/Corrective Actions: Causes Root Causes Mechanical high-cycle fatigue which caused a through-wall crack The original procedure did not sufficiently document the basis for the examination details used to inspect for the degradation previously seen, such that the superseding procedure (NDE-995) was not adequate to ensure the flaw was detected before going through-wall Contributing Causes System design resulted in higher than desired system vibration on 1B2 HPI line Inadequate administrative guidance existed in ISI Section XI Functional Area Manual for conduct of augmented examinations and appropriate disposition of UT examinations results where conditions limited the examination of weld volume 40 For Information Only Root Causes/Corrective Actions: Pictures Showing UT Scan Angles 41 For Information Only Root Causes/Corrective Actions: Extent Of Condition Extent Of Condition All locations within scope of the NDE-995 procedure where the inspections were not able to fully inspect the specified volume, i.e. limited exams Additional scope to include all procedures in the augmented inspection program pertaining to ultrasonic inspections where there is not explicit guidance for the documentation and resolution of inspection limitations DEP sites were reviewed and there is no evidence to support they have the same vulnerabilities as identified in the DEC UT NDE procedures 42 For Information Only Root Causes/Corrective Actions: Extent Of Cause Extent Of Cause Review procedures used to implement augmented UT inspections to determine if elements important to the intent of the procedure have been omitted 43 For Information Only Root Causes/Corrective Actions: Completed Corrective Actions Completed Corrective Actions The leaking weld was replaced All HPI lines have been inspected at Oconee using diverse techniques Procedure NDE-995 was revised to incorporate the root cause lessons learned Weld crowns must be reduced to allow UT probes to traverse without loss of signal Specifies higher angle probes for one-sided examinations Requirements for pre/post job briefs and brief content UT NDE procedures were subsequently revised with similar guidance Governing procedures were revised to require documentation of volume coverage limitations in the CAP, and review by the UT Level III and the Program Owner 44 For Information Only Root Causes/Corrective Actions: Actions To Prevent Recurrence Actions To Prevent Recurrence 1B2 HPI line will be modified to reduce vibration Increase frequency of 1B2 HPI piping UT to every refueling outage as an interim action Implement a means in NDE governance procedures to ensure the critical elements of the inspection procedures used to implement augmented inspections are incorporated into superseding procedures 45 For Information Only Root Causes/Corrective Actions: Fleet NDE Performance Fleet NDE Performance Augmented Inspection Programs - Weaknesses Identified Thermal Fatigue Program Non-Destructive Examinations not successful in identifying existing flaws (Oconee, McGuire) Examples of inspections not performed (Harris, Robinson) Reactor Vessel Head Inspections Indication not identified during data review at Harris during Spring 2012 refueling outage Subsequent review identified the flaw which led to maintenance outage Improvements in Program Oversight Engineering leadership NDE Services leadership Three UT Level IIIs with external industry experience added Increased Level III involvement in pre/post briefs and data review Implement annual program review (Augmented Inspection and NDE) 46 For Information Only Root Causes/Corrective Actions: Fleet NDE Performance Improvements in Nuclear Oversight External UT Level III being added to NDE Program Audit Separate NOS oversight from ISI and Augmented ISI NDE perform functions to ensure independence NOS participation in the annual program review Planned Augmented Inspection Processes and Procedure Improvements Focused Assessment using external peers to identify gaps Commitments are documented and met Technology is appropriately applied Deficiencies are identified in CAP Self evaluation practices are properly deployed Processes and organization aligned to support successful implementation of the program Create a stand-alone Augmented Inspection Program Identify a program owner Implement industry best practices during program merger NDE Examiners to demonstrate methods proficiency annually prior to outage season Separate NDE Training Program Committee 47 For Information Only Regulatory Significance Bob Guy Director, Site Support 48 For Information Only Regulatory Significance Apparent violation of 10 CFR Part 50, Appendix B, Criterion XVI, Corrective Action, for failure to identify and correct a significant condition adverse to quality; in 2004, a procedure was developed for augmented in-service inspection program ultrasonic examinations which effectively removed reasonable assurance that HPI nozzle component weld cracking would be identified and corrected Duke Energy agrees with the apparent violation Full compliance has been achieved 1B2 weld was repaired November 19, 2013 Post repair RT was completed November 23, 2013 NDE-995 was revised March 3, 2014 Comprehensive actions have been taken to prevent recurrence Other actions are taken or planned analysis concludes that the finding is of very low safety significance 49 For Information Only Closing Comments Preston Gillespie Senior Vice President 50 For Information Only Acronyms 51 AP Abnormal Procedure MLOCA Medium Loss of Coolant Accident ASME American Society of Mechanical Engineers NDE Non-Destructive Evaluation CAP Corrective Action Program NOS Nuclear Oversight CCDP Conditional Core Damage Probability OE Operating Experience CDF Core Damage Frequency ONS Oconee Nuclear Station CRP Conditional Rupture Probability PRA Probabilistic Risk Assessment DEC Duke Energy Carolinas PWR Pressurized Water Reactor DEP Duke Energy Progress RCS Reactor Coolant System FCG Fatigue Crack Growth RT Radiographic Testing GPM Gallons Per Minute SIA Structural Integrity Associates HPI High Pressure Injection SLOCA Small Loss of Coolant Accident ISI In-Service Inspection SSE Safe Shutdown Earthquake KNFCS KNF Consulting Services TS Technical Specifications LOCA Loss of Coolant Accident UT Ultrasonic Testing LWR Light Water Reactor For Information Only 52