Qualitative analysis of state/event fault trees for supporting the certification process of software-intensive systems

Abstract

For the certification of modern safety critical systems tree based failure models, like standardized fault trees (FTs), are frequently used methodologies. But when it comes to software-intensive systems these techniques have some crucial disadvantages, especially in modeling timing behavior. To deal with these weak points state/event fault trees (SEFTs) [6] were developed. However, these kind of fault trees can only be analyzed in a quantitative way. In this paper we propose an approach to analyze them qualitatively as well. This results in ordered event sequences which represent different ways for triggering a critical event of the underlying SEFTs, which can be seen as a time-dependent equivalent of the minimal cut set (MCS) analysis of standardized FTs. To evaluate our approach, we implemented the SEFTAnalyzer to apply it on a software-controlled fire alert system.