Parameterization of few-group cross sections for reactor diffusion calculations

This thesis deals with the parameterization of neutron cross sections as functions of the most important core operating conditions for the purpose of predicting theoretically the reactor core response. Data reduction methods have been used to condense the volume of data required for a sufficient data base of neutron cross sections and to obtain a theoretical representation of a cross section model characterized by the varying core conditions. Regression analysis techniques have been used to obtain a second order polynomial cross section regression model. The neutron cross sections act as response variables and the most important core operating conditions (from a neutronic point of view) as predictor variables in the neutron cross section regression model. Techniques such as transforming, centering and piecewise fitting of the data were applied to simplify the neutron cross section regression model. The adequacy of the neutron cross section regression model was checked by verifying the validation of the model in its intended operating environment. The neutron cross section regression model was developed in this work specifically for use in pressurized water reactor diffusion calculations. However, the techniques developed in this work is applicable to any reactor type. (author). 37 refs., 147 figs., 46 tabs