Vibration monitoring of a four-loop PWR: Model-investigations of the sensitivity of the monitored signals on mechanical failures

Abstract Since years Vibration Monitoring Systems (VMS) are used in German PWR with more and more success. Experience has been gained with respect to detect mechanical failures at a very early stage of development. However, the influence of incipient failures on the monitored vibrations of the primary components are known only for few cases, when mechanical failures occurred at PWRs in reality and when at the same time a VMS was in operation there. In order to get more information on the sensitivity of the monitored vibrations on failure caused changes in the functions or patterns deduced from the VMS signals, model investigations including mechanical failures have been performed. As a first step an analytical vibration model has been formed, which can calculate “analytical vibration patterns”, matching the “measured vibration patterns” quite well. The monitored functions are the APSDs of the VMS-signals. Using this model it was possible, to change certain stiffness and damping parameters in such a way, as we assumed that any mechanical failure could happen in a real structure. By calculating the analytical vibration patterns using the failure model and by comparing these results with the reference patterns, the changes of the monitored signals caused by the failure in question could easily be evaluated. In this way also the effects of a growing failure like more and more relaxing hold down springs at the pressure vessel flange were investigated. In addition, constrained supports of components or internals and broken tie-rods of steam generator or main coolant pump supports have been simulated. A great number of incipient mechanical failures and breaks of supports were such investigated with respect of their effect on the different VMS-signals and patterns respectively. The calculations were performed for the primary system of a 1300 MW el. power PWR with four loops. The work was sponsored by the Minister of Research and Development of the FRG. In the paper an overview of the results as well as some examples with details of interest will be given. In addition, the most importing lessons learned from these sensitivity investigations will be presented.