Automatic Generation of RAMS Analyses from Model-based Functional Descriptions using UML State Machines

In today's industrial practice, safety, reliability or availability artifacts such as fault trees, Markov models or FMEAs are mainly created manually by experts, often distinctively decoupled from systems engineering activities. Significant efforts, costs and timely requirements are involved to conduct the required analyses. In this paper, we describe a novel integrated model-based approach of systems engineering and dependability analyses. The behavior of system components is specified by UML state machines determining intended/correct and undesired/faulty behavior. Based on this information, our approach automatically generates different dependability analyses in the form of fault trees. Hence, alternative system layouts can easily be evaluated. The same applies for simple variations of the logical input-output relations of logical units such as controllers. We illustrate the feasibility of our approach with the help of simple examples using a prototypical implementation of the presented concepts.