July-August 2019 NPJ

42 NuclearPlantJournal.com Nuclear Plant Journal, July-August 2019 Implementation of Technology A number of neutron detectors exist on the outside of the reactor vessel in a PWR plant. Four of these detectors (NI-41, NI-42, NI-43, and NI-44) are referred to as power range monitors and used to measure neutron flux to indicate reactor power (Figure 3a). The output of these detectors can be sampled at a high frequency (1000 Hz or greater) and analyzed to provide the vibrational signature of the reactor vessel and its internals including the core barrel, thermal shield, and fuel assemblies. Figure 3b shows the result of analysis of noise data from a neutron detector in a PWR plant. This plot is referred to as the Power Spectral Density (PSD) of the neutron noise data. A PSD is a graph of the variance of the noise signal in a narrow frequency band plotted as a function of frequency for a wide range of frequencies. The plot is generated automatically using Fast Fourier Transform (FFT) analysis and existing software packages. The peaks in the plot of Figure 3b show the frequency and amplitude of vibration of the reactor internal components. These components are distinguished by analysis of the phase between pairs of cross-core neutron detectors. If data from other sensors such as the existing temperature, pressure, level, and flow sensors are added to the mix and cross-correlated, the diagnostics capabilities of the noise analysis and cross-correlation techniques will increase dramatically to the point that a “motion picture” of the movement of all components within the reactor vessel can be produced. AMS has a library of noise data from over 40 years of testing in nuclear power plants involving in-core and ex-core neutron detectors, core exit thermocouples, process sensors such as resistance temperature detectors (RTDs), and pressure, level, and flow transmitters. This database can be used to demonstrate the full capabilities of the noise analysis and cross-correlation techniques to provide reactor diagnostics and enable the nuclear industry to monitor the condition of the reactor vessel and its internals while the plant is operating. Through a collaboration agreement soon to be established between AMS and the Consortium for Advanced Simulation of Light Water Reactors (CASL) at ORNL, our inventory of noise data for nuclear power plants will be used with CASL models to elevate the diagnostic capabilities of noise analysis and cross- correlation techniques and thereby help optimize plant maintenance activities. Case Study In the early 1990s, AMS was contracted to identify the root cause of neutron signal alarms which increased daily near the end of an operating cycle at the Diablo Canyon Nuclear Power Station. The authorities had asked the plant to identify the root cause of the problem or shut the plant down for fear of excessive vibration of reactor internals. In response, AMS measured the amplitude and frequency of vibration of the reactor internals using the existing neutron detectors and determined that the vibration levels were normal. This did not satisfy the authorities who insisted that the root cause of the problem had to be identified to allow the plant to continue to operate. Subsequently, AMS performed cross-correlation analysis of noise signals from in-core and ex-core neutron detectors, core exit thermocouples, and pressure transmitters together with analytical modeling of the reactor system which produced a “motion picture” of the reactor internals. This helped identify the root cause of the problem as core flow anomalies resulting from a change in geometry of new fuel assemblies that were installed in the plant prior to the emergence of the neutron signal alarms. The authorities then became concerned that flow anomalies can cause flow fluctuations with high amplitude and low frequency in such a combination to cause the fuel elements to overheat. Fortunately, further analysis by AMS demonstrated that the frequency of the flow fluctuations was high enough and their amplitude was low enough so as to eliminate any concern about fuel overheating. Consequently, the plant was allowed to continue to operate with a simple adjustment of neutron signal alarm setpoints. This is just one example of how signals from existing sensors may be used to extract more information about the plant equipment and processes and thereby provide a variety of diagnostics as to the condition and health of the reactor and its components. This type of measurement has been used anecdotally by the nuclear industry for troubleshooting such as the one just described but not for routine predictive maintenance, diagnostics, prognostics, or aging management. Today, with the U.S. nuclear fleet aging and license renewals Figure 3. (a) Ex-Core Neutron Detectors, (b) Reactor Internal Vibration Signature. Plant Condition... ( Continued from page 41)