MEADEP and its application in dependability analysis for a nuclear power plant safety system

Abstract

Although there are several measurement and model based approaches to assessing the compliance of critical computing systems with reliability requirements, applying these approaches requires sophisticated data analysis and mathematical skills so that reliability engineers often hesitate to perform such a task. The need to develop cost effective, credible, and easy-to-use tools to reduce difficulties in performing such tasks has thus been apparent. This paper presents a tool of this kind-MEADEP. MEADEP integrates techniques in graphical user interface programming, database engineering, dependability modeling, and statistical/numerical analysis, and provides a user-friendly interface on Windows 95 for non-expert users. Use of MEADEP on failure data from measurements produces quantitative evaluations of dependability for critical systems while greatly reducing requirements for specialized skills in data processing, statistical analysis, dependability modeling and model solution from the user. The application of MEADEP on safety systems is demonstrated by modeling dependability for a nuclear power plant safety system based on the Eagle 21 architecture and its early field failure reports.