Development of A Statistical Testing Approach for Quantifying Safety-Related Digital System on Demand Failure Probability (NUREG/CR-7234)

On this page:

• Publication Information
• Abstract

Download complete document

• NUREG/CR-7234 (PDF – 6.61 MB)

Publication Information

Manuscript Completed: December 2015
Date Published: May 2017

Prepared by:
Tsong-Lun Chu1, Athi Varuttamaseni1, Joo-Seok Baek¹, Meng Yue¹, Tim Kaser², George Marts², Paul Murray², Bentley Harwood², Nancy Johnson², and Ming Li³

¹Brookhaven National Laboratory
²Idaho National Laboratory
³U.S. Nuclear Regulatory Commission

Ming Li, NRC Project Manager

NRC Job Code V6196

Office of Nuclear Regulatory Research
U.S. Nuclear Regulatory Commission
Washington DC 20555-0001

Availability Notice

Abstract

A statistical testing approach for quantifying on-demand failure probabilities for safety-related digital systems has been developed and applied to the loop-operating control system (LOCS) of an Advanced Test Reactor (ATR) experimental loop at Idaho National Laboratory (INL). This work is the result of a collaboration between Brookhaven National Laboratory (BNL), INL, and the Korea Atomic Energy Research Institute (KAERI).

The objectives of the study include:

1. development of a statistical testing approach for estimating digital system failure probability, the results of which are suitable for including in a probabilistic risk assessment (PRA); and
2. application of this approach to the LOCS, and insights into the feasibility, practicality, and usefulness of the estimation in models of digital systems for inclusion in nuclear power plants' PRAs.

The study used the ATR's PRA to define the testing environment, that is, the conditions under which the safety system would be called upon to initiate a safety function. Based on the PRA accident sequence information, a thermal-hydraulic model (RELAP5) was used to simulate the experimental loop conditions (e.g., pressure, temperature, and flow) during the selected accident sequences in order to provide realistic input signals to the LOCS test platform. To ensure that the test cases provided adequate coverage of operational conditions, thirteen probabilistic failure process models (PFPMs) were developed to represent the varieties associated with timing, component failure modes, and process variable control. An automated test platform was developed to supply input signals for each test case to the LOCS digital system and monitor when a trip signal was generated. The testing results were then used to quantify the on-demand failure probability of the digital LOCS system.