A Dynamic Fault Tree Approach for Time-Dependent Logical Modeling of Autonomous Flight Systems 2019-01-1358

This paper addresses the urgent need for adequate methodologies to use in analyzing autonomous flight systems, including Unmanned Aircraft. These systems are inherently dynamic and require analysis that is explicitly time dependent. Autonomous flight systems are becoming more commonly used, especially for Part 23 aircraft including Business (Corporate) and Regional Jets or Unmanned Aircraft deployed in hazardous environment/situation. Such systems are expected to make their own decisions under uncertain conditions caused by potential system structure changes when entering a new flight phase or switching to a new system configuration due to system degradation or failure(s) [1]. This paper highlights significant modeling errors that can arise in analyzing dynamic scenarios where these time dependencies are ignored. Model-based solutions are provided by incorporating a time-dependent algebraic formalism into Fault Tree Analysis (FTA) and Dependency Diagram (DD) with updated descriptions in SAE ARP4761A and ARP4754B (Note: These are currently under development). A Dynamic Goal Tree (or alternatively, a Dynamic Dependency Diagram) provides an effective implementation of the time-dependent logic for dynamic system analysis analyzing autonomous flight systems which are inherently dynamic since decisions need to be made without human input in a very short time. The safety analysis for autonomous flight systems, including Unmanned Aircraft, can be performed by extending the traditional phased mission analysis, thus the potential system structure changes for different phases in a flight mission can be expressed by a Dynamic Fault Tree (DFT), or alternatively, a Dynamic Goal Tree (DGT), or Dynamic Dependency Diagram (DDD) [2].