Title: Keeping the Safe in New Digital Safety System Designs
1Keeping the Safe in New Digital Safety System
Designs
Dr. Peter B. Lyons, Commissioner U.S. Nuclear
Regulatory Commission IAEA International
Conference on Common-Cause Failures of Digital
Instrumentation and Control Systems in Nuclear
Power Plants June 19, 2007
2Common-Cause Failure (CCF) of Digital IC Safety
Systems
- Practical Solutions Already Implemented
Throughout the World - Increasing Worldwide Interest in All Digital
Nuclear Plants - Continue Identifying Practical and Safe Solutions
3U.S. Historical Perspective
- 1980s some digital components within larger
analog safety systems - CCF not a significant concern
- 1990s questions during reviews of advanced LWR
designs - Could failure probabilities be estimated?
- Was analog backup needed for CCFs?
4National Academies of Science and Engineering
Study Panel (1997)
- To address CCFs, emphasize diversity of inputs
and algorithms, hardware, and operating systems - Agreed with NRC position that CCF could be
addressed using a variety of diversity approaches
dependent on each plant design, including diverse
digital systems
5Historical Summary
- Very real safety benefits can be achieved
- Regulatory questions have been addressed in past
designs, but continue to persist today - Future Senior NRC/industry attention to
establish a project plan to address these
questions for new designs
6Defense-in-Depth
- Greater inter-connection and coupling of digital
channels raise regulatory questions regarding the
adequacy of defense-in-depth - Are familiar defense-in-depth approaches
applicable to digital IC?
7Defense-in-Depth
Redundancy
Diversity
Independence
8The Basic Rules for Inter-connections
- 1 Determine there will actually be a safety
benefit - 2 The greater the inter-connection, the more
attention must be paid to preventing adverse
results while preserving the intended benefits
9Key Safety Concepts
- Independence and Diversity
- Presently no other safety concepts to take their
place
10What is Sufficient?
- How timely can we address all the new questions
raised by using digital safety system design
concepts that differ significantly from past
digital system designs? - NRC is actively updating regulatory guidance, but
will we become too prescriptive for this rapidly
changing technology?
11The Big Picture
- PRA is a Big Picture tool that helps us better
understand overall system failure as a function
of individual failures where it is impossible to
test overall system reliability - PRA requires that we know how the various parts
of a system can fail
12A Catalogue of Failures
- Hardware Failures
- Software Failures
- Combined/Interactive Hardware/Software Failures
- A SYSTEMATIC APPROACH
13Operating Experience
- Gathering, Sharing, and Using It
- Safety-related and Non-Safety-Related
- Nuclear Industry and Other Industries
14Operating Experience
- Important for systems where testing cannot be
expected to shake out all the potential failure
causes and modes
15Test and Evaluation Facilities
- Todays NRC approach
- National Laboratories
- Universities
- International Research Centers
- Case-by-case
- Not Efficiently Integrated
- Not Keeping Pace with the Technology
16A New Concept?
- A U.S. Research, Test, and Evaluation Facility
for digital safety systems - Attract new graduates and experienced
professionals - Participation of other government agencies and
industries
17A New Concept?
- Public Workshop to discuss the concept of a U.S.
Research, Test, and Evaluation Facility - Location Atlanta, Georgia (tentative)
- Dates September 6-7, 2007 (tentative)
- Further Information Steven Arndt,
NRC/RES - (301) 415-6502
18Closing Key Points
- Digital Technology can improve safety
- Achieving Defense-in-Depth requires redundancy,
diversity, independence - Need to identify and catalogue the ways that
digital systems can fail and increase our
knowledge-base with operating experience - Need to find better ways of evaluating new
digital safety system designs
19THANK YOU!