The method described in the present article for monitoring the boiling of the coolant in a VVER-1000 core was developed for use as part of the in-reactor noise diagnostics system. This system records signals by means of seven rhodium in-reactor neutron-flux sensors placed along the height of the fuel assemblies which are being monitored. Of the 164 fuel assemblies comprising the fuel load of the core, 54-64 are monitored. The neutron-flux noise is recorded by separating the instantaneous component of the current of the neutron-flux sensors [1]; this component is due to the Compton effect. The computational investigations performed with the NOSTRA program [2] of the local fluctuations of the energy-release field and the analysis of the experimental data obtained in the No. 3 unit of the Kalinin and the two units of the Tianwan (China) nuclear power plants have made it possible to determine the characteristic features of the height distribution and the amplitude-frequency characteristics of the neutron-flux noise as a function of the form, frequency, and amplitude of the fluctuations of the parameters of the coolant at different times during a fuel run. The influence of steam formation on the characteristics of the neutron-flux noise was also studied.Physical Principles of the Method. Let us consider a fuel assembly where surface boiling is observed at the top of the assembly. On the section from the entrance into the fuel assembly to the onset of surface boiling the density of the coolant is determined by the pressure, which decreases with height, and the temperature, which increases with height in the heated channel. The density decreases smoothly and almost linearly. After the onset point of surface boiling has been passed, the coolant density is determined by, aside from the pressure and enthalpy, the nonequilibrium vapor fraction. The intensity of the vaporization depends on the surface temperature of the fuel element cladding. If the liquid is not heated up to the saturation temperature, the vapor bubbles formed as a result of surface boiling collapse. The life time of the bubbles depends strongly on the temperature and pressure of the coolant. These effects influence the volume fraction of the vapor and, correspondingly, the coolant density.

The results of investigations of noise appearing in the signals from direct-charge sensors of the in-reactor monitoring system of VVÉR reactors as a result of coolant parameter fluctuations are presented. The calculations and experimental data are used to analyze the dependence of the amplitude of the neutron flux oscillations (local intensity) at the locations of sensors as a function of the magnitude and frequency of the fluctuations of a specific coolant parameter. The NOSTRA program was used to perform the calculations. The results of the analysis were used in the in-reactor noise diagnostics system in the No. 3 unit of the Kalinin and the No. 1 unit of the Tianwan (China) nuclear power plants.In-reactor noise diagnostics has been developed on the basis of an analysis of the noise in various parameters of the core of a nuclear reactor. Analysis of neutron flux noise is one direction of noise diagnostics whose main goals are detection and localization of coolant boiling sites inside a reactor core. Monitoring the moment when coolant begins to boil makes it possible to refine the real conditions for removing heat in the core and to determine the design limits more accurately. The computational and experimental works described in the present article were performed to solve the problem of instrumentation monitoring of coolant boiling.The standard in-reactor monitoring sensors distributed over a core are direct-charge sensors with a rhodium emitter [1]. Calculations and experiments have confirmed that such sensors make it possible to detect neutron flux fluctuations in a wide frequency range due to the instantaneous component of the current [2] that is due to the Compton effect. The instantaneous component of the current is proportional to the main (activation) component of the sensor current. For the same coolant density fluctuations, the amplitude of the sensor current fluctuations is proportional to the neutron flux or the main component of the sensor current. The standard deviation of the variable component of the sensor current reduced to its average main current -the dimensionless parameter ∆ -was chosen as the parameter characterizing the state of the coolant.The neutron flux fluctuations depend on the coolant density fluctuations, which influence the change in the moderating properties of the coolant and arise as a result of fluctuations of the temperature, flow rate, and pressure of the coolant in the reactor. When steam starts to form, the nonlinearity of the dependence of the coolant density on the amplitude of the fluctuations of the coolant parameters at the entrance into a fuel assembly becomes stronger. As a result, the amplitude of the coolant density fluctuations increases, which increases the amplitude of the neutron flux noise in the region of boiling.Experimental Results. The experiments confirmed that rhodium sensors make it possible to detect the moment local boiling of the coolant appears and starts to develop. The possibility of detecting boiling onset was shown in experi-

No abstract

The method developed by the Kurchatov Institute's WWER reactor department to detect local coolant boiling in the reactor core on the basis of the neutron noise analysis is currently being implemented at some WWER-1000 units within their regular in-core noise diagnostic systems. Given that WWER-1000 power uprating – as well as new designs being developed, such as AES-2006 and WWER-TOI characterized with higher core power and coolant temperature – would increase the probability of coolant boiling, the efficiency of the available boiling diagnostics method was to be re-evaluated. On the basis of relevant computations, it was suggested to modify the diagnostic model and introduce new parameters to improve the reliability of local boiling detection. Computations simulating in-core boiling were thus performed and showed that the improved method was quite efficient. This paper discusses the possibilities to detect – using the improved method – various in-core coolant boiling cases anticipated for WWERs.

No abstract