ML13308A905
| ML13308A905 | |
| Person / Time | |
|---|---|
| Site: | Duane Arnold  |
| Issue date: | 11/05/2013 |
| From: | NextEra Energy Duane Arnold |
| To: | NRC/RGN-III |
| References | |
| EA-13-182 | |
| Download: ML13308A905 (35) | |
Text
Duane Arnold Energy Center Regulatory Conference Emergency Diesel Generator Failure November 5, 2013
2
3
4 Richard Anderson, Site Vice President Ken Kleinheinz, Site Engineering Director Mark Gilbert, Operations Shift Manager Anil Julka, Fleet Risk and Reliability Manager Mike Davis, Site Licensing/EP Manager Jim Petro, Managing Attorney - Nuclear Larry Lee, ERIN Engineering Attendees
5 Opening Remarks - Rich Anderson Performance Deficiency - Rich Anderson Root Cause Evaluation - Rich Anderson Corrective Actions - Rich Anderson Engineering Analysis - Ken Kleinheinz Operations Response to Lube Oil Leak - Mark Gilbert Risk Significance - Anil Julka Closing Remarks - Rich Anderson Agenda
6 Rich Anderson Site Vice President Opening Remarks
7
- Importance of Emergency Diesel Generators
- Any failure is unacceptable Lessons Learned from Previous Failures
- Site took prompt action in response to Leak Risk Significance of March 2013 Event New Information Will be Discussed Today Opening Remarks
8
- A Emergency Diesel Generator (EDG) lube oil cooler tube bundle replaced in October 2012 EDG removed from service on March 6, 2013 for cable inspections Operator identified small leak on subsequent maintenance run
- Leak increased from 5 drops per minute (DPMs) to 60 DPM over course of the nearly 2 hour2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> run EDG restarted for maintenance investigation and resulted in significant increase in leakage Flange leak repaired and EDG returned to service on March 10, 2013 Event Summary
9 Performance Deficiency
- Licensee failed to prescribe a work order appropriate to the circumstances Cause Analysis
- RC1 - Misalignment between flange mating surfaces of the A EDG Heat Exchangers led to uneven compression of a gasket, leakage, and failure of the lube oil channel head connection, specifically at the 10-11 oclock position
- CC1 - The evaluation completed in 2008 lowered the torque value to ensure the channel head did not crack during reassembly. The torque values were lowered to an overly conservative value to protect the cast iron channel head from cracking without taking into account additional factors that would contribute to adequate compression (gasket change, increased monitoring, torque checks, and sealing surface imperfections).
This action allowed for pre-existing conditions/weaknesses to manifest Performance Deficiency and Cause Analysis
10 Cause Analysis Continued
- CC2 - A lack of detail in the work packages for the unique EDG Heat Exchanger arrangement has led to inconsistent assembly practices.
- CC3 - Multiple opportunities to incorporate internal and external operating experience and identified EPRI Best Practices for re-torquing have been missed in the planning of EDG heat exchanger work.
- CC4 - Joint temperature differences between shutdown and operating conditions for lube oil system allowed thermal forces to reduce gasket compression.
Performance Deficiency and Cause Analysis Cont.
11 EDG lube oil cooler was repaired and returned to service on March 10, 2013 Extent of Condition immediately evaluated
- B EDG Lube Oil Cooler Flange was installed with higher torque value using correct torqueing sequence Immediate Corrective Actions
12 Sustainability Corrective Actions
- Detailed Heat Exchanger Re-assembly Procedure
- Torque values with additional margin for gasket sealing
- Generic guidance for re-torqueing after relaxation period
- Generic guidance for managing misalignment Corrective Actions
13 Ken Kleinheinz Site Engineering Director Engineering Failure Analysis
14 Engineering Failure Analysis
15 Engineering Failure Analysis
16 Timeline of Events Engineering Failure Analysis
17 Fundamental Strategy for Evaluating Gasket Failure
- Understand the forces that are holding the gasket in place
- Understand the forces and conditions that would cause leakage and/or displace the gasket Forces/Conditions that Could Lead to Failure
- Gasket size/material AP Services/Garlock Sealing Technologies performed a material analysis on the failed gasket. Concluded that the gasket material was adequate for this service condition Engineering Failure Analysis
18 Forces/Conditions that Could Lead to Failure
- Vibration - Measured during normal run and found no forces at work at this joint.
- Surface Imperfections causing gasket damage or leak-path Recent inspection of the flange surface revealed no irregularities in gasket area, no warping Engineering Failure Analysis
19 Forces Holding Gasket In-place
- AP Services/Garlock Sealing Technologies Measured Compression of gasket. Identified that the gasket crush ranged from 1-2% to 9%, which is below the expected crush range of 10-20%
Engineering Failure Analysis
20 MPR Associates performed a structural analysis of the lube oil cooler gasket joint prior to the gasket failure
- Analysis cases were run with reduced preload in the failure location and assumed coefficients of friction to assess the effect of the oil leakage on the gasket failure
- Analysis provided a quantitative confirmation of the Garlock conclusions (classical case of gasket blowout)
- Finite element model of the joint was developed
- Program controlled under Appendix B QA program Engineering Failure Analysis Finite Element Model
21 Methodology
- Gasket Preload Following gasket failure, the permanent compression of the failed gasket was measured Preload in FEM was adjusted to reproduce conditions
- Friction Calculations showed that the friction required to restrain the gasket is near zero
- Analysis indicates a friction force as small as 0.034 would prevent gasket extrusions
- Acceptance Criteria Gasket failure occurs when the gasket extruded from the joint by the lube oil pressure
- FEM backed up by a hand-calculation showing similar result Engineering Failure Analysis Finite Element Model
22 Why did the gasket extrude on March 8th, 2013
- Performed thermal analysis of the heat exchanger joint Materials that make up the Heat Exchanger, Flanges, and Bolting have different thermal expansion coefficients Compression forces on the gasket in the low-crush area drops to zero when heat exchanger cooled to 70F (believe it was lower, but did not monitor at the time)
- With zero compression, joint could be expected to leak with minimal standby pressure in the heat exchanger Engineering Failure Analysis Thermal Model
23 Conclusions
- Disassembled inspection of the flange surface performed after the SERP showed no irregularities, removing the uncertainty of one of the potential root causes
- The forces holding the gasket in place are greater than the forces acting to displace the gasket under normal and startup conditions
- Gasket extrusion is predicted at zero friction Zero friction indicates a lubricated surface
- Initial oil leak on March 8, 2013 contributed to the gasket failure Reduced the friction force in the low preload area (10 to 11 oclock)
- The failure of the gasket was a result of a thermal transient on the joint and an area of low gasket compression due to misalignment of the joint Engineering Failure Analysis Finite Element Model
24 The A EDG lube oil system was inoperable from 3/6/13 to the time of repair and subsequent operability run on 3/10/13
- Revised Past Operability evaluation concluded 3.69 days of EDG unavailability vs. initial determination of 22 days
- A EDG would have run for > 5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> Up to the time the system was removed from service on March 6, 2013, the EDG would have been capable of starting and performing its required safety functions Engineering Failure Analysis Summary
25 Mark Gilbert Operations Shift Manager Operations Response to Lube Oil Leak
26 Event Setup
- B SBDG OOS with restoration time >24hrs Loss of offsite power with A SBDG powering essential bus Event Onset
- Loss of Offsite Power Several annunciators received in Control Room CRS Enters AOP 301, Loss of Essential Electrical Power
- Approximately 2 minutes into event Operations Response to Lube Oil Leak
27 Balance of Plant Operator Dispatches Plant Operator to EDG Room
- Approximately 3 minutes into event Action directed by AOP 301 Action is to complete EDG Operating Checklist, OI 324A9
- Checklist includes step to check for lube oil leaks Plant Operator Inspects Running EDG
- Approximately 5 minutes into event Plant Operator Identifies Leak
- Expected that leak identified approximately 8 minutes into event Plant Operator Directed to Monitor Leak
- Direction would occur approximately 10 minutes into event Operations Response to Lube Oil Leak
28 ERO Mobilized Per Emergency Plan
- Per Emergency Plan, activation of ERO occurs within 15 minutes of declaration of an Emergency Classification of an Alert TSC functional
- TSC functional within 45 minutes
- Repair teams and engineering sent to investigate/repair within 60 minutes Conclusion
- Operators/SROs have the training and procedures to maintain the EDG available Operations Response to Lube Oil Leak
29 Anil Julka Nuclear Risk and Reliability Manager PRA and Risk Significance
30
- Compared the results with the use of SPAR model
- ERIN engineering assisted with the changes to SPAR model that would emulate NRC assumptions
- Limitations exist regarding the ability to incorporate SPAR model assumptions consistent with the NRC quantification
- As such, NEE could not fully reproduce the NRCs SPAR model results EVALUATION OF RISK SIGNIFICANCE
31
- SPAR: Assumes long term mitigating actions are unreliable and therefore not credited
- NEE: Core Damage can be prevented without AC Recovery by implementing AOPs, EOPs, EMG and SAMPs Comparison of Results
32 Comparison of Key Model Inputs Issue NRC SPAR Model NextEra Position Comments Exposure Time 22 Days 3.69 Days Engineering RCE Credit for A EDG Recovery No Yes A was repairable Could be recovered in < 9 hours1.041667e-4 days <br />0.0025 hours <br />1.488095e-5 weeks <br />3.4245e-6 months <br /> HEP for Aligning Diesel Fire Pump (Long Term SBO)
HEP = 0.3 HEP < 0.01 15-20 hours before RCIC trips on backpressure Aligning Diesel Fire Pump Time Frame After 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> After 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> Implementation begins in 5-30 minutes; 30-60 minutes to implement River Water System Train Maintenance Unavailability 19.4 Days/Yr/Train (5E-2)
<3 Days/Yr/Train (6E-3)
DAEC Maintenance Rule Red threshold is 7.3 days AC Power Recovery Time Frame 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br />
> 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> DAEC credits PDFP, Containment Venting, TSC Diesel EDG Run time > 5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> Number of MCR Panels with potential to cause Fire-Induced LOOP 9 (out of 74) 2 (out of 74)
DAEC Cable Database THESE NEW INPUTS HAVE A SIGNIFICANT IMPACT ON THE RESULTS
33 Key Actions Mitigating Consequences of Long Term Station Black Out Core Damage prevented during Station Black Out (>24 hours) without AC recovery based on these actions:
- Aggressive Cooldown
- Technical Support Center (TSC) Diesel - Charge Batteries > 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />
- Alternate RPV Injection with Portable Diesel Fire Pump (PDFP)
- Containment Venting
34 CDF Results Internal Events PRA + Fire PRA Model CDF 3.69 Days 22 Days SPAR -NRC 6.2E-07 3.7E-06 SPAR - NEE 1.5E-07 9.0E-07 DAEC RG 1.200 6.2E-08 3.7E-07
35 Rich Anderson Site Vice President Closing Remarks