Reliability functions and optimal decisions using condition data for EDF primary pumps

There are many examples in which proportional hazards modelling (PHM) is used to accurately model the effects of the operating environment on an item´s lifetime. Using such techniques it is possible to find an economically optimal moment of replacement based on condition monitoring data. However, in a nuclear power plant, the problem is somewhat different as it is not possible to stop the process and perform preventive maintenance at the most economical moment in time. The problem of interest in such cases has more to do with whether the item will last until the next scheduled plant shutdown or not. This paper presents some developed and implemented theory to calculate the conditional reliability function of an item given the current equipment age and the current values of the condition-monitoring variables. A case study of a shaft seal for a reactor cooling pump (RCP) using data from several French nuclear power plants operated by the Electricite de France (EDF) is also presented