Inference of core barrel motion from neutron noise spectral density. [PWR]

A method was developed for inference of core barrel motion from the following statistical descriptors: cross-power spectral density, autopower spectral density, and amplitude probability density. To quantify the core barrel motion in a typical pressurized water reactor (PWR), a scale factor was calculated in both one- and two-dimensional geometries using forward, variational, and perturbation methods of discrete ordinates neutron transport. A procedure for selection of the proper frequency band limits for the statistical descriptors was developed. It was found that although perturbation theory is adequate for the calculation of the scale factor, two-dimensional geometric effects are important enough to rule out the use of a one-dimensional approximation for all but the crudest calculations. It was also found that contributions of gamma rays can be ignored and that the results are relatively insensitive to the cross-section set employed. The proper frequency band for the statistical descriptors is conveniently determined from the coherence and phase information from two ex-core power range neutron monitors positioned diametrically across the reactor vessel. Core barrel motion can then be quantified from the integral of the band-limited cross-power spectral density of two diametrically opposed ex-core monitors or, if the coherence between the pair is greater than or equal to 0.7, from a properly band-limited amplitude probability density function. Wide-band amplitude probability density functions were demonstrated to yield erroneous estimates for the magnitude of core barrel motion.