Spectral and correlation analysis for malfunction diagnosis in a nuclear power system

An investigation was made into the theory and application of several spectral and correlation estimators to the inherent fluctuations in nuclear power system process parameters for the purpose of malfunction detection and diagnosis. Statistics considered were the single and joint probability density distributions, the auto and cross correlation functions, and the spectra associated with single input linear transfer function analysis. Specific interest was directed to the relative sensitivity of these estimators and techniques for implementing them as part of on-line instrumentation. Experimental study was performed on records obtained from the Robinson-2 Reactor, a 739 MW(e) pressurized water reactor operated by Carolina Power and Light Company at Hartsville, South Carolina. (U.S.)