Application of a bayesian belief network model to reliability assessment of nuclear safety-related software - 207

Conference · Thu Jun 15 04:00:00 EDT 2017

OSTI ID:23035332

Lee, Seung Jun ^[1]; Lee, Sang Hun; Kang, Hyun Gook ^[2]; Chu, Tsong-Lun; Varuttamaseni, Athi; Yue, Meng ^[3]; Eom, Heung Seop; Cho, Jaehyun ^[4]; Li, Ming ^[5]

---

1. School of Mechanical and Nuclear Engineering, Ulsan National Institute of Science and Technology, 50, UNIST-gil, Ulsan (Korea, Republic of)
2. Department of Mechanical, Aerospace, and Nuclear Engineering, Rensselaer Polytechnic Institute, 110 8th St, Troy, NY (United States)
3. Brookhaven National Laboratory, Brookhaven Avenue, Upton, NY (United States)
4. Korea Atomic Energy Research Institute, 111, Daedeok-daero, Daejeon (Korea, Republic of)
5. U.S. Nuclear Regulatory Commission Washington, DC (United States)

As the instrumentation and control (I and C) systems in nuclear power plants (NPPs) have been replaced with digital-based systems, the need to incorporate software failures into NPP probabilistic risk assessments has arisen. In order to assess the probability of software failure on demand, a Bayesian belief network (BBN) model was developed which estimates the number of defects and the resulting probability of software failure on demand in nuclear safety-related software. To assess the feasibility of the BBN framework, the BBN model was applied to the prototype Integrated Digital Protection System-Reactor Protection System (IDiPS-RPS) to estimate the number of remaining faults and the software failure probability of a target software. The developmental- and V and V-activities carried out during the IDiPS-RPS development process were evaluated based on the well-defined checklist derived by the V and V team and were estimated based on expert elicitation. In addition, the attribute evaluations and the number of FPs of the target software is provided as the inputs for the BBN model. The application results showed the feasibility of using BBNs for quantifying software failure probabilities and several insights were gained from the applications of the BBN model. The proposed BBN framework can be applied to estimate the software failure probability for other safety-related NPP software and provide an insight on modeling the software development process that involves iterations between different development phases. (authors)

• Research Organization: American Nuclear Society - ANS, 555 North Kensington Avenue, La Grange Park, IL 60526 (United States)
• OSTI ID: 23035332
• Country of Publication: United States
• Language: English

Related Subjects

22 GENERAL STUDIES OF NUCLEAR REACTORS
46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND TECHNOLOGY
COMPUTER CODES
COMPUTERIZED SIMULATION
EVALUATION
NUCLEAR POWER PLANTS
PROBABILISTIC ESTIMATION
RADIATION PROTECTION
REACTOR CONTROL SYSTEMS
REACTOR INSTRUMENTATION
REACTOR PROTECTION SYSTEMS
REACTOR SAFETY
RELIABILITY
RISK ASSESSMENT