Application of Reliability Theory to a Reactor Safety Circuit

Abstract

The reactor safety circuit considered in this paper is that of a r from m'> design. Such a circuit comprises m guard lines in parallel, continuously monitoring the reactor performance. When r or more lines signal the presence of a dangerous reactor condition the reactor is automatically shut down. Two types of circuit failures are examined, a safe failure which results in a spurious reactor shutdown, and an unsafe failure when no tripping signal can be produced even if the reactor is in a dangerous condition. The repair of a guard line which failed safe starts at once, after the occurrence of a fault, provided that an engineer is available. No circuit failure will occur if the fault is eliminated before r-1 other faults have occurred. The effects of using a single short circuiting facility in the servicing of circuit faults have also been investigated. In this paper formulae for the calculation of the frequency of safe and unsafe circuit failures are derived, treating time as a continuous variable. The results obtained differ from those based on time represented by a series of discrete intervals. Some illustrative examples are given.