Integrating Existing Safety Analyses into SysML

Abstract

Migrating systems and safety engineering (often with legacy processes and certified tools) towards a model based systems engineering (MBSE) environment is a socio-technical problem. Establishing a common conceptual framework requires agreement on modelling artefacts and the integration of existing tool chains to minimise disruption. We discuss our experience integrating a SysML Safety Profile to model fault trees but which has the prerequisite requirement to continue the analysis of those models by existing tools. We demonstrate a lightweight profile that minimally captures the fault logic for a Rolls-Royce gas turbine engine controller and provides specific in-house extensions for both fault tree and engine dispatch analysis by exporting model entities and relationships from the SysML fault trees. During integration we realised a more fundamental need to reconcile the systems engineers functional view with the safety engineers focus on failure modes and fault logic in order to maximimse the longer term benefits of MBSE development.

Low-Code Engineering for the Internet of Things

“…approaches such as [175,178,181] do not provide such a feature and instead rely on manually created FT models, which are then transformed into FT graphs. In contrast to the previous approach, our proposed approach completely supports this feature by providing an environment in which system main blocks and sub-systems can be decomposed and analyzed separately.…”

“…Even though approaches such as [4,171,175,176,234] extend the SysML language in the same way that we do, our environment is more user-friendly due to the advanced component-based and multi-view modeling infrastructure, where each view has its own underlined constraints that enforce its specific privileges on model entities and properties that can be manipulated [42].…”

“…[174] an approach for manually deriving FT diagrams from the Reliability Block Diagram (RBD) was also presented, however, the qualitative and quantitative analysis are manually performed, differently from our approach where the analysis is performed automatically. Furthermore, unlike our approach which models the system architecture, annotates the model with safety-related information, and later generates and analyzes FTs, several approaches, such as [175][176][177], propose SysML profiles which are used to create FT models and later translate them into FT graphs without any support for system modeling itself. On another hand, Chaari et al [178] proposes a Meta-modeling-based Failure Propagation Analysis (MetaFPA) framework to support the synthesizing of the system failure propagation models in order to help the creation of the system FTs.…”

Introduction to the Special Issue on Software-Intensive Autonomous Systems: Methods and applications

“…Furthermore, their objectives are limited to safety analysis without a focus on how to exploit safety results and link them to the safety management process. This research goal is satisfied differently by other approaches [33] to align safety and systems models without conceptual clarification of semantic links. An approach to validate safety of perception software and system in autonomous driving systems has been proposed based on fault injection but it did not consider the safety management [34].…”

Integration of systems design and risk management through model‐based systems development

“…For example, Clegg et al used a profile to extend SysML to model fault trees and failure modes within the system model. 6 Helle created a safety profile that was realized by stereotypes to help system designers to perform safety analysis without the help of safety engineers. 7 The aim was to obtain fast safety-related feedback for design decisions.…”