A RULE-BASED APPROACH FOR SAFETY ANALYSIS USING STAMP/STPA

摘要

Systems-Theoretic Process Analysis (STPA) is a safety analysis technique based on Systems-Theoretic Accident Model and Processes (STAMP) and it is useful in identifying causal factors and hazardous scenarios. One of the most critical tasks of STPA is to define which contexts are hazardous. The task is manual, time consuming and demands exhaustive and careful analysis of all system contexts. Contexts are combinations of control actions, variables and states and therefore result in a significant number. We propose a rule-based approach to define hazardous contexts in order to assist the identification task. The approach corresponds to a workflow in which some activities can be automated. We implemented a tool to automate these activities through rules. We employed the approach in an experiment with a simple system and simple rules, using the tool, and we were able to identify up to 75% of the contexts as hazardous.