Study On Reactor Internals Monitoring System Based On Vibration Signals

The experience gained from practical operations has proved that the abnormal states ofreactor internals can affect the normal operation of NPP, cause high cost of maintenance, andeven result in nuclear accidents. Therefore, regularly monitoring of the vibration of the NPPreactor internals is necessary in order to achieve early fault detection. In this thesis, thevibration signals of reactor components have been studied based on the design principles ofPWR reactor internals vibration monitoring system. The signal noise reduction and staterecognition method are studied, in addition, the reactor internals monitoring system softwareis developed on LabVIEW platform.Considering the situation of absence of experiment data, ANSYS Workbench is used tomodel the reactor barrel and pressure vessel, and conduct modal analysis and transientdynamics analysis. The model has beam vibration mode and shell vibration mode, which arein accordance with engineering practice. In this way, the vertical direction of the accelerationtime response data of the pressure vessel under normal and abnormal state can be obtained.The wavelet Heuristic threshold filtering method is studied, and LabVIEW virtualinstrument software is used to design the wavelet filtering sub VI. The result shows that thismethod can reduce the noise mixed in the data acquisition process, improve the acquisitionaccuracy and signal-to-noise ratio. On this basis, fractal correlation dimension method is usedto process the vibration signals for fault diagnosis, discuss the correlation dimension selectionmethods of reactor internals vibration signals, and realize the sub VI design of correlationdimension algorithm. Results of calculation show that by comparing the correlationdimension of normal states and fault states, the failure modes of the system can be identified.LabVIEW virtual instrument is used to design the reactor internals monitoring system.Simulation analysis shows that the system can realize online monitoring and fault diagnose ofreactor internals vibration signals, and has the advantages of both small error and high SNR,provides an intuitive and accurate display as well as status analysis functions of the reactorinternals for the operators. The work of the thesis provides a new insight for nuclear devicesmodeling analysis, it brings meaningful exploration of wavelet filtering and fractal correlationdimension, enriches the methods of reactor internals abnormal diagnose, and, puts forward a new way to improve reactor operation security.