Quantified fault tree techniques for calculating hardware fault metrics according to ISO 26262

Since its introduction in 2011, the ISO 26262 standard has provided the state-of-the-art methodology for achieving functional safety of automotive electrical and electronic systems. Among other requirements, the standard requires estimation of quantified metrics such as the Probabilistic Metric for Hardware Failure (PMHF) using quantitative failure analysis techniques. While the standard provides some brief guidance, a complete methodology to calculate the PMHF in detail has not been well described in literature. This paper will draw out several key frameworks for successfully calculating the probabilistic metric for hardware failure using Fault Tree Analysis (FTA). At the top levels of the analysis, methods drawn from previous literature can be used to organize potential failures within a complex multifunctional system. At the lower levels of the FTA, the effects of all fault categories, including dual-point latent and detected faults, can be accounted for using appropriate diagnostic coverage and proof-test interval times. A simple example is developed throughout the paper to demonstrate the methods. Some simplifications are proposed to estimate an upper bound on the PMHF. Conclusions are drawn related to the steps and methods employed, and the nature of PMHF calculation in practical real-world systems.