safeTbox is a multi-viewpoint and multi-analysis tool framework aimed at supporting engineers during the development and certification phases of safety-critical systems. It tightly integrates activities and models from safety engineering with those from systems engineering, thus fostering the interaction between safety engineers and system engineers and bridging the gap between these two. Moreover, modularization is a core concept in safeTbox and is applied to (or a characteristic of) many of the implemented methods and techniques. In putting such strong emphasis on modularization, safeTbox facilitates maintainability and reusability, as well as supplier-OEM relationships and integration of third-party components.

 

safeTbox is being developed as an extension of the commercial modeling tool Enterprise Architect (EA). It comprises a set of profiles for the definition of specialized modeling languages (in support of the typical safety-engineering activities) as well as several add-ins to extend the functionality of EA, primarily for the support of computations, usability, and interoperability.

An appropriate documentation of the system's architecture is a critical input for the realization of any safety related activity. Therefore, safeTbox integrates standardized approaches like SysML (1.3 and 1.4) as well as new experimental approaches like the SPES modeling framework. These approaches allows the specification of the system on a modular and hierarchical basis, on which the safety related artifacts might be structured. Moreover, a type assignment approach allows the user to create heterogeneous architecture models, as well as defining with precision component interfaces.

Hazard and risk assessments have been traditionally performed with the help of  tools like Excel. In safeTbox we follow an integrated model-based approach to overcome the typical problems that arise from the use of such a tool, like it is for instance: the lack of maintainability and traceability. safeTbox provides the best trade off: it offers an intuitive table-base user interface while being able to link existing model elements. The current implementation of supports primary the ISO 26262.

Safety analysis in safeTbox is focused in the modeling of failure models, in which the faults of individual components can be specified, as well as their interaction to depict the failure behavior of the entire system. This can be achieved by using different techniques like Fault Tree Analysis, FMEAs and Markov chains. All of them can be integrated in a heterogeneous model that can afterwards be analyzed from a qualitative and quantitative point of view. 

safeTbox integrates several computation engines that can be used for this purpose. Specially, an integration with Fault Tree Plus from Isograph.

Safety Concepts: Specially in the automotive domain, but not exclusively to it, a safety concept is demanded as part of the certification process. Traditionally, this is a document in which a set of safety requirements is delivered. In safeTbox we have evolved this concept by defining a model-based implementation that overcome typical issues related to text-based approaches like traceability, modularity and poor argumentation. 

Safety Cases: Safety cases are becoming more and more relevant for the safety assessments of systems and for building assurance cases. safeTbox supports the user in this respect by integrating the Goal Structuring Notation as best known technique to support the specification of safety cases. Moreover, special traceability features have been built in to facilitate the assessment process.