Reliability Assessment of Anti-SEU Computer System with Timed Refresh and Backup Fault Tolerance

Spacecraft often fail due to single event upset (SEU) events. To enhance system reliability, a variety of anti-radiation reinforcement measures are used in combination. This paper proposes a reliability assessment method for a computer system that adopts timed refresh and backup fault tolerant measures. This method establishes a failure probability model for the SEU-prone RAM devices in computer systems by considering the state transition characteristics of the timed refresh task and the error correction rules of Hamming codes. Additionally, the Bootstrap small sample data processing method is utilized to obtain the confidence interval estimation of the failure probability. Furthermore, a dynamic fault tree is established according to various backup measures and logical architectures of the anti-SEU computer system. The reliability of the system during the mission is estimated using the Monte Carlo method. At the end of the paper, a case study is carried out. The reliability data of components or systems obtained by this method will provide a reference for designers to evaluate whether the spacecraft is operating safely in orbit.