Space -time analysis for *reactivity determination in source -driven subcritical systems

Increasing worldwide interests in accelerator-driven systems is related to their potential role in transmutation of the spent reactor fuel. Margin of safety expressed in terms of reactivity, measuring proximity to criticality, has to be properly addressed for such systems. Monitoring of reactivity enables us to predict performance of a nuclear system and prevent unforeseen accidents. However, due to the presence of a localized spallation source in an accelerator-driven subcritical system leads to a significantly different neutron flux shape than a source-free fundamental mode in critical systems. As a result, the simple point kinetics approach commonly used for determination of reactivity in critical systems does not account properly for space-time effects in accelerator-driven subcritical systems, yielding inaccurate estimates in reactivity. To overcome this problem and account properly for spatial and spectral effects in reactivity determination, a method directly combining measurements with numerical simulations of the experimental data is developed within a quasi-static formulation. This method provides space-time corrections to a variety of traditional point kinetics techniques and determines the reactivity essentially independent of the detector position, as long as sufficiently accurate information on the reactor configuration is provided. In the dissertation, the space-time corrections are derived for two well-known point kinetics methods: the area-ratio technique and the alpha-method. Numerical simulations performed with the FX2-TH diffusion theory code along with a space-time analysis of MUSE-4 pulsed source experimental data illustrate the applicability of the proposed methods for the determination of significant subcriticality levels in fast and thermal reactor systems. To perform space-time reactivity corrections at reduced computational cost, a modal-local method is developed for source-driven systems and tested with the ERANOS code.;This dissertation also investigates physical properties and mathematical characteristics of two neutron multiplication factors proposed for accelerator-driven systems. The conventional multiplication factor determined by an eigenvalue problem is compared with the source multiplication factor introduced as a ratio of reaction rates involving actual flux distributions, subject to an external source. Variational analysis and numerical calculations performed with the MCNP code indicate a significant difference between the two multiplication factors. Steady-state experimental methods are proposed for determination of the source multiplication factor.