| For nuclear reactors that use clad fuel, the detection of breach of clad and the determination of location and of size of breach is a troublesome and time consuming process, especially for reactors with liquid metal coolant. Breach of cladding can lead to safety problems if coolant is channeled away from the fuel pin. At the least, contamination of the primary system can lead to radiation protection problems.; When breach of cladding occurs, fission fragments created in the fuel below the breach can be transported into the coolant. Some of these fission fragments are delayed neutron precursors, that is, after undergoing a beta decay, these fragments eject a neutron from the nucleus as a mode of decay toward stability. Several reactors monitor for these delayed neutrons to detect failed fuel and to try to locate the breach, notably, the second Experimental Breeder Reactor (EBR-II), the Fast Flux Text Facility (FFTF), and the French fast reactors, Rapsodie, Phenix, and Superphenix. The detection and location systems of these reactors are reviewed and the limitations of each are discussed.; The analyses of all of these reactor system signals are based largely on empirical data, without a complete understanding of the mechanisms involved. In this thesis, each step of the process that leads to a delayed neutron signal is analyzed and a complete theoretical model is developed for each step. These steps include: formation of the delayed neutron precursor in the fuel, transport through the breach, transport out into the bulk coolant flow, transport by the coolant to regions near the detector, delayed neutron production rate, and detection by the delayed neutron monitor. The models for each step are combined into an overall algorithm to determine breach size and determine breach location. Transport by the coolant, although ignored or grossly simplified by all other analysis methods, is shown to be a dominant factor in the delayed neutron signal.; This algorithm is tested on data collected on two different breaches in the FFTF. The location of one of the breaches is correctly predicted by the algorithm, and on the other, the signal between two of the monitors appears to be switched. Experiments are proposed using the FFTF to confirm the algorithms developed in this thesis. |
