Incorporating Safety in Early (Airframe) Systems Design and Assessment

Abstract

Presented is a novel framework for incorporating safety analysis in early systems architecture design. Traditionally, a systems architecture is first defined by the architects and then passed to safety experts, who manually create artefacts such as Function Hazard Analysis (FHA) or Fault Tree Analysis (FTA) for safety assessment. The problem with this manual approach is that if the architect modifies the systems architecture, then the whole safety assessment process needs to be repeated, which is tedious and time consuming. To overcome this limitation, the proposed framework automates the creation of safety models such as FHA and FTA by utilizing the Requirement, Functional, Logical, and Physical (RFLP) systems engineering paradigm. The framework supports three main activities. First, the safety targets are determined by performing a FHA of the architecture and the Requirements view is updated. Second, compliance with the safety requirements is analyzed using dynamic fault trees, automatically generated from the Logical view. Interactive visualization techniques are proposed to interpret the safety results, e.g. highlighting the greatest contributors to the probability of failure. Third, an algorithm is developed that enables the designer to interactively improve the architecture's safety by introducing more reliable components or increasing redundancy. The concept is illustrated with a representative example, where the environmental control system of a civil aircraft is studied from a safety point of view.

Enabling Interactive Safety and Performance Trade-offs in Early Airframe Systems Design

“…In this paper, the RFLP paradigm is augmented with two domains. "Safety domain" for enabling safety assessment of the architecture, as proposed by Jimeno et al [4], and "computational domain" for providing the capability of automated systems sizing, as proposed by Bile et al [3]. Additionally, traceability between different view of the architecture is incorporated, as suggested by Guenov et al [15], enabling a more effective and interactive design process.…”

“…Fig. 2 shows an overview of the existing RFLP safety framework [4] (in light blue) and the necessary additions and modifications to cater for the proposed methodology. A Mission & Flight Conditions view of the architecture is added; the Computational View is modified to include the information provided related to mission.…”

“…The functions that map to the selected components, plus all their children are included in the FHA. More details of the method are presented in an earlier publication from the authors [4].…”

“…Recent work by Jimeno et al [4] discussed the limitations concerning existing methods for automating the safety assessment of architectures and proposed a novel safety framework, focused on automation and interactivity. The framework enables the creation of safety requirements through functional hazard analysis.…”

“… The comparison of the above approaches has shown that the potential trade-offs that appear when sizing subsystems have been neglected and only a single solution to the sizing problem is considered, except in the case of de Tenorio et al [8], who enable trade-offs thanks to their proposed subsystem sizing problem formulation. With the purpose of overcoming the above mentioned limitations, the work presented in this paper aims to integrate subsystems sizing and performance with the safety framework described in Jimeno et al [4]. The framework allows functional reasoning as an enabler of the architecting process, which the literature review has been shown to be important, and also provides the necessary safety features.…”