Алгоритм парирования сбоев в бортовых цифровых вычислительных комплексах со структурно-временной избыточностью

Abstract

Currently, the problem of ensuring reliable operation of onboard digital computing systems is becoming increasingly important. When designing on-Board digital computing systems of spacecraft, it is necessary to provide a combination of a number of indicators: high performance, minimal mass-size and energy characteristics, availability of diagnostic tools and parry the consequences of failures. In addition, one of the main requirements for the onboard digital computing complex of spacecraft is the high total reliability of the system, the minimum time for detection and parry of failures. To improve the reliability of control systems, there are many ways, the main of which is structural redundancy. The use of reservations makes it possible, through the introduction of redundancy, by voting to identify the failed system and exclude it from the configuration. However, in onboard digital computing complexes the failure rate is much lower than the failure rate. This can cause the failure to be perceived by the system as a failure and cause the system to exit the configuration of a healthy channel. By returning the program to the previous recovery point and passing the program fragment again, you can classify the type of failure/failure. Under this approach, the reconfiguration of the system will occur only after unsuccessful attempts to pass a fragment of the program n times (as a rule, the technique is used repeat 7–8 times). The aim of the work is to develop an algorithm for parrying failures in onboard computing systems with three-channel architecture, which involves the complex application of test control and recovery points with different dominance depending on the number of serviceable channels.