ML22143A854
| ML22143A854 | |
| Person / Time | |
|---|---|
| Site: | Nuclear Energy Institute |
| Issue date: | 05/20/2022 |
| From: | Andy Campbell Nuclear Energy Institute |
| To: | Bhagwat Jain NRC/NRR/DORL/LPL4 |
| References | |
| Download: ML22143A854 (29) | |
Text
NEI Common Cause Failure Policy Input Alan Campbell Technical Advisor
©2022 Nuclear Energy Institute
State of Digital I&C The Digital I&C Integrated Action Plan (IAP) has improved regulatory guidance clarity and consistency
- RIS 2002-22 Supplement 1 provided criteria for qualitative assessments of Common Cause Failure (CCF) in low safety significant safety-related systems.
- BTP 7-19 Revision 8 incorporated graded approach assessments into staff review guidance
- NEI 96-07, Appendix D and Reg. Guide 1.187 Rev. 3 provided enhanced guidance for digital systems under 50.59
- DI&C-ISG-06 Rev. 2 provided an Alternate Review Process to improve regulatory confidence for digital safety systems upgrades.
©2022 Nuclear Energy Institute 2
Why Digital Safety Systems?
Existing systems are reaching (or have already reached) obsolescence Enhances safety via system diagnostic capabilities to identify and respond to issues Improves plant performance via improved accuracy, processing time, and automated capabilities Provides more data available to Operations, Maintenance and Engineering resulting in better real-time knowledge Reduces hardware inventory compared to existing systems Supports long-term, safe operation of our plants
©2022 Nuclear Energy Institute 3
Todays Digital Landscape Digital I&C technology has design features that provide for deterministic behaviors through the use modern standards International standards, such as IEC/IEEE, are widely accepted and have stable processes to reflect current understanding Hazard analysis techniques have matured and are used extensively in non-nuclear safety industries (such as aviation/aerospace, defense, automotive, and chemical industries)
NRC needs a modernized digital CCF policy that reflects todays technology, experience, and understanding
©2022 Nuclear Energy Institute 4
Applicable Regulation 10 CFR 50.55a(h) - Codes and Standards, Protection and safety systems
- Requires compliance with either IEEE 603-1991 or IEEE 279-1971 IEEE requirements
- Both IEEE standards require means to implement manual initiation of protection actions
- Neither IEEE standard requires diversity RG 1.62 - Manual Initiation of Protective Actions
- Provides guidance for manual initiation/control to meet IEEE requirements
- Provides a staff position that diversity is required to meet BTP 7-19.
Required codes and standards specify a means for manual initiation of protection actions, BUT do not specify diversity as a requirement.
©2022 Nuclear Energy Institute 5
Applicable Regulation 10 CFR 50.62 - Anticipated Transient Without SCRAM (ATWS)
- 1) Must have diverse means of automatic Auxiliary (or Emergency) Feedwater Initiation and Turbine Trip
- 3) Must have diverse Alternate Rod Injection system
- 4) Must have standby liquid control system (no diversity requirement)
- 5) Must have reactor coolant recirculation pump trip (no diversity requirement)
ATWS requirements for diversity are limited to specific functions and do NOT require manual, system-level actuation.
©2022 Nuclear Energy Institute 6
Applicable Regulation 10 CFR 50 Appendix A, General Design Criteria 22 - Protection System Independence
- The protection system shall be designed to assure that the effects of natural phenomena, and of normal operating, maintenance, testing, and postulated accident conditions on redundant channels do not result in loss of the protection function, or shall be demonstrated to be acceptable on some other defined basis. Design techniques, such as functional diversity or diversity in component design and principles of operation, shall be used to the extent practical to prevent loss of the protection function. [emphasis added]
Design techniques are required to prevent loss of the protection function.
©2022 Nuclear Energy Institute 7
How Are We Addressing CCF Today?
Branch Technical Position 7-19, Rev. 8
- Eliminate Diversity within system or component Testing Alternative Methods
- Mitigate Existing System Manual Operator Action New Diverse System
- Acceptance Bounding acceptance criteria
©2022 Nuclear Energy Institute 8
How Are We Addressing CCF Today?
Branch Technical Position 7-19, Rev. 8
- Eliminate Diversity within system or component Testing Alternative Methods
- Mitigate Existing System - Requires sufficient diversity Manual Operator Action - SSCs used to support the manual operator action are diverse New Diverse System - Requires sufficient diversity
- Acceptance Bounding acceptance criteria
©2022 Nuclear Energy Institute 9
How Are We Addressing CCF Today?
Branch Technical Position 7-19, Rev. 8
- Eliminate Diversity within system or component
- Mitigate Diversity using Existing System Diversity using Manual Operator Action Diversity using New Diverse System
- Acceptance Bounding acceptance criteria
©2022 Nuclear Energy Institute 10
How Are We Addressing CCF Today?
Primary System #1 NRC Digital Instrumentation & Control Training, Module 3.0 Regulatory Concerns, Figure 3-22
©2022 Nuclear Energy Institute 11
How Are We Addressing CCF Today?
Primary System #1 System Interactions (Controlled and Uncontrolled)
NRC Digital Instrumentation & Control Training, Module 3.0 Regulatory Concerns, Figure 3-22
©2022 Nuclear Energy Institute 12
How Are We Addressing CCF Today?
Primary System #1 Diverse System #2 Based on the same understanding of system and interactions.
System Interactions (Controlled and Uncontrolled)
©2022 Nuclear Energy Institute 13
How Are We Addressing CCF Today?
I&C OE (nuclear and non-nuclear) indicates that most systematic failures are a result of:
- Latent design defects due to inadequate requirements
- Uncontrolled system interactions An EPRI study on nuclear events1 indicate that the primary contributing factor is requirements errors Diversity MAY be useful in addressing hazards (e.g., CCF), BUT:
- 1. Diversity CAN increase plant complexity and errors.
- 2. Diversity MAY NOT address all sources of systematic failures.
- 1. EPRI 3002005385 ©2022 Nuclear Energy Institute 14
How Are We Addressing CCF Today?
I&C OE (nuclear and non-nuclear) indicates that most systematic failures are a result of:
- Latent design defects due to inadequate requirements
- Uncontrolled system interactions An EPRI study on nuclear events1 indicate that the primary contributing factor is requirements errors Industry solution to CCF is a diagnostic approach to addressing systematic failures proven effective in other industries and research.
- 1. EPRI 3002005385 ©2022 Nuclear Energy Institute 15
Proposed Implementation Guidance NEI 20-07 Rev. D
- Leverages EPRI Hazards and Consequence Analysis for Digital Systems2 and Digital Reliability Analysis Methodology3
- Provides a diagnostic approach to addressing systematic failure beginning during early stages of design process Identifies missing, inadequate, or incorrect requirements
- Diagnoses system architecture for unsafe control actions
- Uses risk-insights to address hazards commensurate with plant risk
- 3. EPRI 3002018387 ©2022 Nuclear Energy Institute 16
Research Basis EPRI investigated strengths and limitations of various hazard and failure analysis techniques4 EPRI HAZCADS and DRAM combines Fault Tree Analysis (FTA) and Systems Theoretic Process Analysis (STPA)
- Complementary strengths
- Reduces limitations of each method used on its own
©2022 Nuclear Energy Institute 17
Proposed Implementation Guidance The applicant will:
- apply Systems Theoretic Process Analysis (STPA) to diagnose the system architecture and determine specific loss scenarios leading to hazards
- perform a Fault Tree Analysis (FTA) to determine the risk impact of loss scenarios
- map results of FTA to RG 1.174 Figures 4 and 5 regions (graded approach)
- apply control methods to address each loss scenario of STPA commensurate with results from FTA mapping
©2022 Nuclear Energy Institute 18
Systems Theoretic Process Analysis Diagnostic tool that iteratively analyzes requirements, design and system interactions STPA5
- 1) Define Losses and 2) Model the Control 3) Identify Unsafe Identify Loss Hazards Structure Control Actions Scenarios
- 5. STPA Handbook, https://psas.scripts.mit.edu/home/get_file.php?name=STPA_handbook.pdf
©2022 Nuclear Energy Institute 19
Systems Theoretic Process Analysis Efficacy proven through blind studies Example blind study6
- Real incident caused by digital I&C system analyzed
- Participants were familiar with STPA and blind to the selected OE
- Participants provided general description of the system as it existed prior to the incident
STPA also identified ~9 other scenarios unaccounted for in the design.
- 6. EPRI 3002000509 ©2022 Nuclear Energy Institute 20
Systems Theoretic Process Analysis Utilized in non-nuclear industries (automotive, aviation, chemical, defense, etc.)
Automotive Standards: Standards in Progress:
- ISO/PAS 21448, SOTIF: Safety of the Intended
- ASTM WK60748, Standard Guide for Application Functionality of STPA to Aircraft
- SAE J3187, Recommended Practice for STPA in
- SAE AIR6913, Using STPA during Development Automotive Safety Critical Systems and Safety Assessment of Civil Aircraft Aviation Standards:
- IEC 63187, Functional Safety - Framework for
- RTCA DO-356, Airworthiness Security Methods safety critical E/E/PE systems for defence and Considerations industry applications Cyber Security Standards:
- IET 978-1-83953-318-1, Code of Practice: Cyber
- NIST SP800-160 Vol 2, Developing Cyber Security and Safety Resilient Systems: A Systems Security Engineering Approach
©2022 Nuclear Energy Institute 21
Systems Theoretic Process Analysis NuScale used STPA to perform a hazards analysis of I&C systems
- DCA7 describes how STPA was used to analyze I&C systems
- SER8 provides NRC acceptance of hazards analysis SER, Chapter 7 Section 7.1.8.6 The NRC staff concludes that the application provides information sufficient to demonstrate that the proposed HA has identified the hazards of concern, as well as the system requirements and constraints to eliminate, prevent, or control the hazards. The NRC staff also concludes that the HA information includes the necessary controls for the various contributory hazards, including design and implementation constraints, and the associated commitments. The QA measures applicable to HA for developing the I&C system design conform to the QA guidance in RG 1.28 and RG 1.152. [] On this basis, the NRC staff concludes that the application provides information sufficient to demonstrate that the QA measures applied to the HA for I&C system and software life cycle meet the applicable QA requirements of GDC 1 of Appendix A to 10 CFR Part 50; Appendix B to 10 CFR Part 50; and Section 5.3 of IEEE Std. 603-1991. [Emphasis added]
- 8. https://www.nrc.gov/docs/ML2020/ML20204B028.pdf ©2022 Nuclear Energy Institute 22
Benefits of Risk Risk-Informed v. Risk-Insights Better system function allocation between components Better understanding of the impacts of system architectural decisions Inform the use of measures to address a potential common cause failure based upon risk significance Understand risk impact to specific loss scenarios
©2022 Nuclear Energy Institute 23
Proposed Risk Guiding Principles Common-Cause Failure (CCF) SECY Paper Outline, Guiding Principles
- All five principles of risk-informed decision making, as listed in RG 1.174, need to be addressed satisfactorily.
- The PRA used for risk-informed approaches needs to be technically adequate (e.g., meets the guidance in RG 1.200) and include an effective PRA configuration control and feedback mechanism.
- The expanded policy needs to ensure that the introduction of digital I&C does not significantly increase the risk of operating the facility.
©2022 Nuclear Energy Institute 24
Proposed Risk Guiding Principles Due to challenges modeling Digital I&C software reliability in PRA:
- The absolute risk impact of software reliability cannot be quantitatively measured without substantial uncertainties
- The effectiveness of applied design techniques cannot be quantitively measured without substantial uncertainties
- There are no means of comparing design techniques to using diversity without substantial uncertainties NEI 20-07 Rev. D leverages concepts from RG 1.174; however, it is not completely applicable
- This RG is used in the context of licensing basis changes, not design decisions
©2022 Nuclear Energy Institute 25
How Can We Use Risk Insights?
NEI 20-07 utilizes Fault Tree Analysis to assess the risk sensitivity of each loss scenario The result of the sensitivity analysis is mapped to the CDF/LERF regions and used in a graded approach to apply control measures
©2022 Nuclear Energy Institute 26
Policy Considerations Allow for graded approaches based upon plant risk-insights to ensure applicants focus on the most risk-significant functions and to provide flexibility in meeting established system performance criteria.
Consider the full plant defense-in-depth strategy to prevent (to the degree practicable), mitigate, or respond to a digital common cause failure.
Allow for the use of modern hazards and/or reliability analysis techniques to examine the system for adverse conditions and identify appropriate system requirements to prevent systematic failures.
Expand the ability to use design techniques, including diversity when applicable, to prevent (to the degree practicable), or mitigate a digital common cause failure in accordance with GDC 22.
©2022 Nuclear Energy Institute 27
Example Policy
- 1. The applicant shall assess the impact of the proposed digital instrumentation and control Reactor Protection System (RPS) and Engineered Safety Features Actuation System (ESFAS) on the plants defense-in-depth systems and procedures to demonstrate that vulnerabilities to digital common cause failures have been adequately addressed.
- 2. The applicant shall identify each digital common cause failure that could adversely impact a safety function using risk-insights, and hazards and/or reliability analysis techniques.
©2022 Nuclear Energy Institute 28
Example Policy
- 3. The applicant shall demonstrate commensurate with the risk significance of each identified digital common cause failure adequate measures to address the identified digital common cause failure that could adversely impact a safety function. The measures may include non-safety systems or components if they are of sufficient quality to reliably perform the necessary functions and with a documented basis that the measures are unlikely to be subject to the same common cause failure. The measures may also include monitoring and manual operator action to complete a function.
©2022 Nuclear Energy Institute 29