Search

Since Cyber-Physical Systems (CPS) may exhibit different structures and emergent behaviours during different operational phases, while also being safety critical, it is useful to perform systematic safety analyses tightly relying on the functional and components topologies of such systems. Our proposal is to perform FMEA and FTA analyses as soon as possible in the CPS design process in order to identify and mitigate the risks related to some safety critical structures and behaviours. Thus, these preliminary analyses enable to propose relevant design modifications and improvements such as optional or additional redundancies, components repairability capabilities or relevant control strategies, taking into account the complexity and the potential variability of the structure and the behaviour of the systems. This work derives from previous MBSE (Model-Based System Engineering) and MBSA (Model Based Safety Analysis) integration studies, performed during the early phases of the design of safety critical mechatronic systems, including interconnection components and multi-physical interactions. It relies mainly on M2M (Model to Model) and M2T (Model to Text) transformations and appropriate SysML metamodeling. The proposed CPS safety analysis methodology is illustrated using an aeronautic industrial case study.

In order to develop a system with high safety, it is necessary to determine countermeasures for hazards that extracted using analysis methods like HAZOP, FTA, and so on. Furthermore, it is necessary to able to explain that countermeasures are surely embedded into a developed system in order to be accepted the analysis result by stakeholders. In this paper, we discuss the effectiveness of the knowledge to apply D-Case for a safety analysis result that includes hazards extracted by HAZOP, failure modes extracted by FTA and countermeasures for failure modes.

With the development of technology and hardware, it has been assumed IoT(Internet of Things) society in the future that any device leads. In the automotive industry, in order to provide advanced services, such as automatic driving, any things are expected to lead the vehicle. In the IoT society, so that the lead is more than one system each other quality characteristics are different, such as safety. For this reason, there is concern that trouble is generated from the difference in the attitude toward safety. In order to prevent the problem is to visualize the design quality of each other's systems, it is necessary to obtain a common understanding among the developers.In this paper, as a technique to visualize the design quality of the system, to create a description document of automatic operation system using the GSN(Goal Structuring Notation), to be able to objectively explained on the basis of the assumption and evidence the validity of the design quality of each other's system Check. Upon confirmation, provide a description items to be measures to hazards and threats that are expected in the relationship between the systems in automatic operation system, it showed the item should be explained to each other between the systems.Such description structure is standardized, if shared between systems, with the common understanding can be obtained between developers can predict the quality required for the product. As a result, it is considered to be able to prevent a problem that occurs from the difference of the corporate culture.

Increasing complexity of products and safety regulations combined with an increasing amount of variants complicates the process of safety analysis within systems engineering. Moreover, it is known that the early avoidance or prevention of failures saves costs and improves the quality. As methods of safety analysis, i.e. fault tree analyses require immense manual efforts and expert knowledge, the efficiency of these analyses has to be improved. Our paper thus presents an approach to generate and evaluate fault trees by the usage of matrix-based models. It is an approach tailored to the early phases of system design and provides a preliminary fault tree analysis. It automatically generates fault trees and evaluates them. Thus, it facilitates the efficient identification of safety critical elements and the assessment and comparison of alternative system architecture concepts. This paper provides a brief introduction to fault tree analysis and presents existing approaches to automate the generation or synthesis of fault trees. The limitations of these approaches during early stages of design are discussed and the need for a tailored approach is derived. The developed approach consists of four phases and six steps which each are explained in detail. The whole approach is validated within a small industrial case study and its benefits and limitations are discussed. The case study shows, that the approach successfully improves the efficiency of a preliminary fault tree analysis.