1 1.2. Main concepts 5 The Physics of Failure (PoF) approach examines the root causes of component failures by assessing how materials, defects, and stresses a!ect reliability. It identifies and models individual failure mechanisms in components based on environmental and usage stresses (H. Wang, Liserre, Blaabjerg, et al., 2014). For example, Zhu, Huang, W. Peng, et al., 2016 proposes a PoF-based framework for fatigue reliability analysis of an aircraft turbine engine disc. System reliability models, on the other hand, evaluate the reliability of the entire system by considering the interaction and configuration of system components. Graphbased techniques, such as Reliability Block Diagrams (Signoret and Leroy, 2021), and systematic approaches, such as Failure Modes and E"ects Analysis (Stamatis, 2003), are employed to identify potential failures and their consequences within a system. One of the most prominent system reliability methods we encounter in the literature is Fault Tree Analysis, which we discuss further in the next section. Fault Tree Analysis Fault Tree Analysis (FTA) (Ruijters and Stoelinga, 2015) is a key technique in reliability engineering and risk analysis, used since the 1960s across various sectors such as automotive, aerospace, and nuclear industries (Kabir, 2017). FTA helps in modelling complex systems by illustrating logical relationships, which are crucial for understanding potential system failures, tracing root causes, pinpointing critical components, and computing probabilities of failure at both system and sub-system levels. Fault Trees (FTs) are graphical models composed of logic gates and basic events. See Section I.4.1 for formal definitions. As an example, Figure 1.4(a) illustrates a bike system. Figure 1.4(b) presents the bike system components, such as wheels, handle, chain, disc brake, and cassette. Figure 1.4(c) models the bike’s inability to ride safely using FTs. The system is divided into sub-systems and components until the desired resolution is reached. The top event: bike cannot ride safely represents the event of interest. The logic gates in Figure 1.4(c) determines how a failure propagates based on Boolean logic. For example, if the chain breaks, the bike cannot ride, triggering the top event through the “OR” gate. The “AND” gate models failure when the wheels fail, indicating that all basic events under the gate must activate for the top event to occur. If only the front wheel fails, the bike can still ride with e!ort, but if the disc brake roto also fails, it becomes unsafe. This simple example highlights the value of FT models in illustrating relationships between components and failure propagation, facilitating strategic actions to prevent system-level failures. When basic events in the FT are represented using
RkJQdWJsaXNoZXIy MjY0ODMw