Systematic scenario test case generation for nuclear safety systems

  • Authors:
  • Wan-Hui Tseng;Chin-Feng Fan

  • Affiliations:
  • Dept. of Computer Science and Engineering, Yuan-Ze University, Chung-Li, Taiwan;Dept. of Computer Science and Engineering, Yuan-Ze University, Chung-Li, Taiwan

  • Venue:
  • Information and Software Technology
  • Year:
  • 2013

Quantified Score

Hi-index 0.00

Visualization

Abstract

Context: The current validation tests for nuclear software are routinely performed by random testing, which leads to uncertain test coverage. Moreover, validation tests should directly verify the system's compliance with the original user's needs. Unlike current model-based testing methods, which are generally based on requirements or design models, the proposed model is derived from the original user's needs in text through domain-specific ontology, and then used to generate validation tests systematically. Objective: Our first goal is to develop an objective, repeatable, and efficient systematic validation test scheme that is effective for large systems, with analyzable test coverage. Our second goal is to provide a new model-based validation testing method that reflects the user's original safety needs. Method: A model-based scenario test case generation for nuclear digital safety systems was designed. This was achieved by converting the scenarios described in natural language in a Safety Analysis Report (SAR) prepared by the power company for licensing review, to Unified Modeling Language (UML) sequence diagrams based on a proposed ontology of a related regulatory standard. Next, we extracted the initial environmental parameters and the described operational sequences. We then performed variations on these data to systematically generate a sufficient number of scenario test cases. Results: Test coverage criteria, which are the equivalence partition coverage of initial environment, the condition coverage, the action coverage and the scenario coverage, were met using our method. Conclusion: The proposed model-based scenario testing can provide improved testing coverage than random testing. A test suite based on user needs can be provided.