Dependability modeling and evaluation of IPS architectures in the railway sector

Abstract Rail transport is increasingly preferred for its safety and reliability, which can be achieved through a variety of advanced design solutions and technological supports. Among them, ensuring a continuous power supply is crucial for the reliable execution of energy-dependent critical operations. Inspired by the Italian railway infrastructure, here the focus is on Integrated Power Supply (IPS) systems that support critical functionalities, such as the signaling system that regulates railway traffic movement. Fault-tolerant IPS designs are implemented to address unforeseen fault events at the IPS level, which could lead to safety or availability issues. This paper presents a refined stochastic model-based evaluation framework to aid in comparing various IPS redundant architectures with respect to dependability characteristics, mainly focusing on reliability and availability metrics. Major contributions pertain to i) the modeling of four IPS architectures; ii) the integration of models to analyze the IPS behavior under normal and critical conditions; iii) the consideration and modeling of battery failures; iv) a comprehensive analysis campaign to compare the considered IPS architectures in terms of both dependability and cost. A designer can effectively utilize the findings from this analysis to determine the most appropriate IPS organization that meets dependability criteria while also considering potential energy consumption savings. The analyses conducted under realistic parameter settings demonstrate that architectures featuring redundancy at the level of individual IPS components yield the best performance in terms of both dependability and cost. For instance, under specific parameter settings, the CR and IR architectures outperform the SR and MR architectures by roughly three orders of magnitude in terms of Mean Time Between Failures (MTBF), achieving MTBF values of up to approximately $$6\times 10^7$$ h.