Risk Informing the Commercial Nuclear Enterprise Promise of a Discipline: Reliability and Risk in Theory and in Practice University of Maryland Maria Korsnick Constellation Energy Nuclear Group, LLC April 2, 2014
How our Business is Risk-Informed I. Managing Risk to the Business II. Managing the Risk of Normal Plant Operation III. Defining Extreme External Events IV. Risk-Informed Lessons for External Events V. The Path Forward 2
I. Managing Risk to the Business Each CENG nuclear plant and the corporate office maintains a risk Heat Map An easy-to-read summary of the risks associated with a business unit A method for communicating the risks being managed Delphi Method for forecasting risk is used - experts come together to perform periodic assessments of Company risks Subjective (non-analytical) probability and impact assessment of each risk Identifies mitigating actions 3
Operating Fleet Heat Map (example) Top risks. Probability Very Likely 95% Likely 80% Moderate 50% Remote 20% Rare 5% Tritium Flood analysis White finding (G) Fire Protection/ NFP 805 New NRC Regulations Cyber Security Short term Output/ Forced outages Key Staffing Post-Fukushima Response Extended Refueling Outage GSI 191 Prolonged Forced Outage EPA Cooling Water Intake regulation Regulatory Compliance Nuclear Risk Corporate / Generation Industrial / Radiological Environmental Level of Control High Medium Low Insignificant Minor Significant Major Impact Critical Significant risks from site maps grouped / assigned based on significance to fleet 4
Heat Map Risk Table (example) Issue Risk Category Impact Probability Level of Control Fukushima Response EPA 316b Rule, Clean Water Act Key Staffing High cost of studies, modifications, uncertainty of outcomes. Impact on emergency planning Potential for significant modifications to intake structures at NY and MD sites High rate of retirements over next ten years, loss of expertise/talent Mitigation Regulatory Major Likely Medium Active engagement with industry and NRC Regulatory Critical Remote Low Industry proposing alternatives to federal and state EPA Corporate Significant Moderate Medium Implement Knowledge Transfer and Retention program 5
II. Managing Risk during Normal Operations Plant-specific PRAs model core damage and large early release frequency Risk impact of scheduled maintenance, plant evolutions, and system outages are analyzed Four risk levels used to communicate to plant staff and set controls GREEN YELLOW ORANGE RED Pre-established risk mitigation measures applied as higher risk conditions are entered 6
Example Plant PRA Risk Initiating Event Distribution Potential Risk Increase Factor for Key Equipment Condenser, 2% Sei smic, 4% MSIV, 3% Feedwater, 4% Other, 8% Fire, 42% Div I Emergency Switchgear Div II Emergency Switchgear Div I Emergency DC Div I 120V Emergency AC Div II 120V Emergency AC Lake Intake, 4% Loss of 2 SWP Pumps, 5% Di v I AC, 7% LOOP, 7% 7.0% 6.0% 5.0% 4.0% 3.0% 2.0% 1.0% 0.0% Div II AC, 8% Flood, 6% System Percentage Contribution to CDF Colors correspond to the associated System Health Report status as of 4th quarter in 2013 2RHS A/LPCS Supp Pool Return Div 1 600V Emergency Switchgear Risk Thresholds >x30 125V DC Switchgear >x15 Div 2 600V Emergency Switchgear >x3 x3 2RHS B/RHS C Supp Pool Return 1 10 100 1000 Key Operator Actions %CDF 30% 15% 9% 4% 3% 3% 2% 2% 2% 1% Description Respond to Control Room Fire Control Service Water and Open Room Doors (HVAC) Align Containment Heat Removal Vent PC (Air or Div I AC lost) Isolate SW Header Flood in RB Control Level to Prevent Boron Washout Align RHR During ATWS Align Fire Water for EDG Cooling Manually Depressurize (Transient) Vent PC (Local Actions including use of Port. Powerpack)
Hypothetical PRA Risk Planetary Charts Plant 1 Plant 3 Plant 2 Plant 4 Every Plant is Unique design, internal / external events Risk insights are gained by comparing plant risk profiles Physical Modifications Protective Barriers Procedures Operator Response Times Maintenance Practices Housekeeping
III. Defining Extreme External Events Original plant design for external events (security, seismic, flood, fire) based on regulations and best state of knowledge of risk at time of licensing Industry understanding of risk has been highly dynamic 1975 Browns Ferry fire 2001 terrorist attacks 2011 Japan earthquake and tsunami (Fukushima) Evolving risk insights from new data creates constant churn in design and operation of our plants Fire: industrial fire code - to - Appendix R - to - NFPA 805 Revised design basis security threat, robust defenses, cyber Post-Fukushima reassessment of earthquake frequency and intensity for central and eastern US plants (NRC GSI-199) Post-Fukushima reassessment of design basis flood/frequency 9
IV. Risk-informed Lessons for External Events The uncertainties are real and unavoidable Extrapolation from internal event modeling experience is not applicable to other models Reliance on numerical mean values is not sufficient Data supporting rare events may have large uncertainty (e.g., floods) Undue focus on numerical outcomes leads to a reduced emphasis on important insights Adding conservatism in PRA is not an antidote, it can significantly distort sound risk-informed decision-making 10
Case in Point NFPA-805 FPRA Challenge: Deterministic PRA mentality distorts risk perspective Conservatisms added at every major step of the process to bound uncertainties Results do not match operating experience benchmarks Risk-significant fires over-predicted Fires with significant spurious operations over-predicted Outcome: Disproportionately large resources spent on model refinements and plant modifications 11
Large Conservatism in Fire PRA Building Blocks Significant Departure from Realism = Ineffective Decision-Making + Conditional Core Damage Probability Conservatism + Fire Suppression Conservatism + Fire Severity Conservatism + Fire Frequencies Conservatism Compounding conservatism reduces effectiveness of decision making tool
V. The Path Forward Objective Gain a more complete and balanced understanding of important risk contributors Clarify risk-informed decisionmaking process that can deal with uncertainties Educate decision-making stakeholders on risk-informed decision-making Develop technical resources to support better risk-informed understanding Industry Continue development of more realistic and complete plant-specific PRAs Propose a practical integrated decision-making process Provide focused PRA training to industry staff and decision-makers Expand EPRI/OG commitment to training and technology Proposed Actions NRC Move away from imbedding conservatism in PRA models - Starts with fire PRA Adapt/adopt a practical integrated decision-making process consistent with RG 1.174 Provide focused PRA training to NRC staff and decision-makers Expand training on truly risk-informed decisionmaking 13
Key Takeaway PRA has added tremendous value to the Nuclear Industry allowing us to operate plants safer. Addressing very low probability / high consequence events can be as important as addressing high probability / high consequence events. Challenges remain with the tools: Risk insights are masked by over conservatism or deterministic approach back to basics. Uncertainty matters what can we do to address and reduce uncertainty? 14