Development of a stochastic temperature treatment technique for Monte Carlo neutron tracking: Dissertation

Thermal motion of nuclides has a significant effect on the reaction probabilities and scattering kinematics of neutrons. Since also the nuclides in nuclear reactor materials are in constant thermal motion, the temperature-induced effects need to be taken into account in all neutron transport calculations. This task is notably complicated by the fact that the temperature distributions within operating power reactors are always non-uniform. With conventional transport methods, accurate modeling of temperature distributions within a reactor is cumbersome. The temperature distributions that are in reality continuous in space need to be approximated with regions of uniform temperature. More importantly, pre-generated temperature-dependent data on reaction probabilities must be stored in the computer memory at each temperature appearing in the system, which restricts the feasible level of detail in the modeling of temperature distributions. This thesis covers the previous development of a temperature treatment technique for modeling the effects of thermal motion on-the-fly during Monte Carlo neutron transport calculation. Thus, the Target Motion Sampling (TMS) temperature treatment technique is capable of modeling arbitrary temperature distributions such that the memory footprint of the interaction data is unaffected by the resolution of the temperature discretization. As a very convenient additional feature the TMS technique also provides for modeling of continuous temperature distributions as-is, making the discretization of temperature distributions unnecessary altogether. The basic idea of the TMS technique is introduced, and the results are shown to be in accordance with reference solutions calculated with conventional neutron transport methods. The TMS method is developed further by optimizing its implementation, and the performance is compared against conventional neutron transport methods in different reactor systems. The results show that the TMS method significantly facilitates the modeling of complex temperature distributions in nuclear reactors without compromising the accuracy of the calculations. The method also proves to be well-feasible in terms of performance, especially as long as the number of temperature-dependent nuclides remains relatively small.