Application based dynamic fault tree method for AFDX network

Abstract

The traditional static fault tree cannot capture the dynamic behavior of system fault mechanisms, so the dynamic fault tree has become one research hot spot now. Many scholars focus on the research of using the dynamic fault tree model and analyze for dynamic behavior of system fault mechanisms. However, in their studies they generally separate the software and hardware and analyze independently. But for the AFDX (Avionics Full Duplex Switched Ethernet) network, the software and hardware function interactively keep the application running on the network, which means we cannot simply separate the software and hardware and analyze independently. Therefore, it is essential to find a reasonable method to illustrate the faults caused by the coupling relationship between hardware and software. In this paper, we analyze the dynamic faults caused by the coupling relationship between hardware and software, and put forward a dynamic fault tree modeling method to model the faults. Therefore, we need to study the relationship between the software and hardware before modeling and analyzing the faults. Therefore we come up with a dynamic fault tree modeling method based on the network applications. To model and analyze the dynamic fault caused by the coupling relationship between the software and hardware, we first study the coupling relationship between the software and hardware in a certain application operation. Secondly we put forward the concept of Function-Time order-Dependency block diagram. With this basis, we then put forward a transformation method from the Function-Time order-Dependency block diagram to the dynamic fault tree, and realize this method with software. Finally, a case study shows that the method is effective and the fault caused by coupling relationship can be easier modeled. Application based Dynamic fault tree modeling method does not separate the software and hardware. It concerned with their coupling relationships and functions in a certain application operation. Then we can model the dynamic fault tree for those faults which have the coupling relationship with this method. Further research on a complex coupling relationship between the software and hardware in many different applications will be taken to acquire consideration.