skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Development of an integrated system for estimating human error probabilities

Abstract

This is the final report of a three-year, Laboratory Directed Research and Development (LDRD) project at the Los Alamos National Laboratory (LANL). This project had as its main objective the development of a Human Reliability Analysis (HRA), knowledge-based expert system that would provide probabilistic estimates for potential human errors within various risk assessments, safety analysis reports, and hazard assessments. HRA identifies where human errors are most likely, estimates the error rate for individual tasks, and highlights the most beneficial areas for system improvements. This project accomplished three major tasks. First, several prominent HRA techniques and associated databases were collected and translated into an electronic format. Next, the project started a knowledge engineering phase where the expertise, i.e., the procedural rules and data, were extracted from those techniques and compiled into various modules. Finally, these modules, rules, and data were combined into a nearly complete HRA expert system.

Authors:
; ;
Publication Date:
Research Org.:
Los Alamos National Lab., NM (United States)
Sponsoring Org.:
USDOE Assistant Secretary for Human Resources and Administration, Washington, DC (United States)
OSTI Identifier:
296684
Report Number(s):
LA-UR-98-2649
ON: DE99001272; TRN: AHC29903%%82
DOE Contract Number:
W-7405-ENG-36
Resource Type:
Technical Report
Resource Relation:
Other Information: PBD: [1998]
Country of Publication:
United States
Language:
English
Subject:
29 ENERGY PLANNING AND POLICY; 99 MATHEMATICS, COMPUTERS, INFORMATION SCIENCE, MANAGEMENT, LAW, MISCELLANEOUS; HUMAN FACTORS; PROBABILISTIC ESTIMATION; SAFETY ANALYSIS; EXPERT SYSTEMS; RISK ASSESSMENT; ERRORS

Citation Formats

Auflick, J.L., Hahn, H.A., and Morzinski, J.A. Development of an integrated system for estimating human error probabilities. United States: N. p., 1998. Web. doi:10.2172/296684.
Auflick, J.L., Hahn, H.A., & Morzinski, J.A. Development of an integrated system for estimating human error probabilities. United States. doi:10.2172/296684.
Auflick, J.L., Hahn, H.A., and Morzinski, J.A. 1998. "Development of an integrated system for estimating human error probabilities". United States. doi:10.2172/296684. https://www.osti.gov/servlets/purl/296684.
@article{osti_296684,
title = {Development of an integrated system for estimating human error probabilities},
author = {Auflick, J.L. and Hahn, H.A. and Morzinski, J.A.},
abstractNote = {This is the final report of a three-year, Laboratory Directed Research and Development (LDRD) project at the Los Alamos National Laboratory (LANL). This project had as its main objective the development of a Human Reliability Analysis (HRA), knowledge-based expert system that would provide probabilistic estimates for potential human errors within various risk assessments, safety analysis reports, and hazard assessments. HRA identifies where human errors are most likely, estimates the error rate for individual tasks, and highlights the most beneficial areas for system improvements. This project accomplished three major tasks. First, several prominent HRA techniques and associated databases were collected and translated into an electronic format. Next, the project started a knowledge engineering phase where the expertise, i.e., the procedural rules and data, were extracted from those techniques and compiled into various modules. Finally, these modules, rules, and data were combined into a nearly complete HRA expert system.},
doi = {10.2172/296684},
journal = {},
number = ,
volume = ,
place = {United States},
year = 1998,
month =
}

Technical Report:

Save / Share:
  • Human Reliability Analysis (HRA) is currently comprised of at least 40 different methods that are used to analyze, predict, and evaluate human performance in probabilistic terms. Systematic HRAs allow analysts to examine human-machine relationships, identify error-likely situations, and provide estimates of relative frequencies for human errors on critical tasks, highlighting the most beneficial areas for system improvements. Unfortunately, each of HRA's methods has a different philosophical approach, thereby producing estimates of human error probabilities (HEPs) that area better or worse match to the error likely situation of interest. Poor selection of methodology, or the improper application of techniques can producemore » invalid HEP estimates, where that erroneous estimation of potential human failure could have potentially severe consequences in terms of the estimated occurrence of injury, death, and/or property damage.« less
  • This report reviews probability assessment and psychological scaling techniques that could be used to estimate human error probabilities (HEPs) in nuclear power plant operations. The techniques rely on expert opinion and can be used to estimate HEPs where data do not exist or are inadequate. These techniques have been used in various other contexts and have been shown to produce reasonably accurate probabilities. Some problems do exist, and limitations are discussed. Additional topics covered include methods for combining estimates from multiple experts, the effects of training on probability estimates, and some ideas on structuring the relationship between performance shaping factorsmore » and HEPs. Preliminary recommendations are provided along with cautions regarding the costs of implementing the recommendations. Additional research is required before definitive recommendations can be made.« less
  • This report describes and evaluates several procedures for using expert judgment to estimate human-error probabilities (HEPs) in nuclear power plant operations. These HEPs are currently needed for several purposes, particularly for probabilistic risk assessments. Data do not exist for estimating these HEPs, so expert judgment can provide these estimates in a timely manner. Five judgmental procedures are described here: paired comparisons, ranking and rating, direct numerical estimation, indirect numerical estimation and multiattribute utility measurement. These procedures are evaluated in terms of several criteria: quality of judgments, difficulty of data collection, empirical support, acceptability, theoretical justification, and data processing. Situational constraintsmore » such as the number of experts available, the number of HEPs to be estimated, the time available, the location of the experts, and the resources available are discussed in regard to their implications for selecting a procedure for use.« less
  • This two-volume report presents the procedures and analyses performed in developing an approach for structuring expert judgments to estimate human error probabilities. Volume I presents an overview of work performed in developing the approach: SLIM-MAUD (Success Likelihood Index Methodology, implemented through the use of an interactive computer program called MAUD - Multi-Attribute Utility Decomposition).