Automation Trustworthiness in Nuclear Power Plants: A Literature Review

Hammad Khalid, Ha Bui, Pegah Farshadmanesh, Ahmad Al Rashdan, Zahra Mohaghegh

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In the pursuit of enhanced efficiency and operational safety, the U.S. nuclear industry is progressively integrating automation technologies into Nuclear Power Plants (NPPs). However, the adoption of these automation technologies presents challenges associated with establishing reliable business cases for their implementation, ensuring their trustworthiness, enhancing their transparency, and addressing licensing process burdens. These hurdles elevate costs and increase schedule uncertainties, thereby complicating the broader use of automation in NPPs. Before committing to a significant investment in the deployment of these technologies, decision-makers require sufficient evidence to verify that the automation would be transparent, trustworthy, and operationally acceptable. Existing methodologies for evaluating automation trustworthiness and enhancing automation transparency are limited in three primary aspects: (i) lack of consensus in definitions of automation trustworthiness and automation transparency concepts across different domains, leading to the absence of widely accepted methodologies for evaluating automation trustworthiness and improving automation transparency; (ii) existing methodologies for enhancing automation transparency are context-specific and lack technical generality, making them unjustifiable for contexts other than those being tested; and (iii) existing methodologies for evaluating automation trustworthiness rely heavily on qualitative/semi-quantitative approaches and empirical validation. This paper is part of an ongoing project, funded by the Department of Energy (DOE) under the Nuclear Energy University Program (NEUP), which aims to develop a probabilistic validation and risk importance ranking methodology for automation trustworthiness and transparency in NPPs. This paper presents (1) a comprehensive literature review on automation trustworthiness and transparency within the nuclear power domain; and (2) proposed working definitions of “automation trustworthiness” and “automation transparency” specifically for the nuclear power domain. Limitations of existing definitions are explored, with proposed adaptations to better suit the nuclear power domain. To illustrate the utility of these proposed definitions, the authors present a case study centered on an Artificial Intelligence (AI)-based automated firewatch system. Lastly, the paper examines the scientific relationship between the evaluation of automation trustworthiness and uncertainty analysis, as encapsulated in the proposed definition of automation trustworthiness. This examination informs the development of a “generic” (rather than automation technology-specific) methodology for automation trustworthiness evaluation, a subject extensively discussed in a companion paper submitted to this same conference as part of the ongoing DOE NEUP project.
Original languageEnglish (US)
Title of host publicationProbabilistic Safety Assessment and Management (PSAM) International Topical Meeting on Artificial Intelligence (AI) and Risk Analysis
StatePublished - Oct 25 2023

Keywords

  • Automation trustworthiness
  • Uncertainty analysis
  • AI-based firewatch
  • Working definitions
  • Automation transparency

Fingerprint

Dive into the research topics of 'Automation Trustworthiness in Nuclear Power Plants: A Literature Review'. Together they form a unique fingerprint.

Cite this