In fast spectrum nuclear reactors, reactivity is directly related to the capability of the reactor to sustain a fission chain reaction for... Show moreIn fast spectrum nuclear reactors, reactivity is directly related to the capability of the reactor to sustain a fission chain reaction for power production. Historically, mechanical/structural analysis and design have been driven primarily by deterministic methods. However, reactivity is extremely sensitive to the location of the fuel within the reactor; which is subject to uncertainties. This makes deterministic models unstable and can allow manufacturing errors to contribute to uncertainties in analysis, resulting in potential safety concerns and incorrect reactor lifetime prediction. One potential means to address this challenge is the use of stochastic analysis. A framework is presented which introduces uncertainty analysis through the use of Monte Carlo Simulation. Latin Hypercube Sampling is used to reduce the number of sample runs and the computational effort and storage space requirements for the results. Geometric parameters such as the gaps at the load pad contact points, the location of the above core load pad (ACLP), and even temperature gradient profiles, that are important to the design of nuclear reactors are varied, and their effects on the overall performance are studied through sensitivity analysis. The main focus was to quantify the effects of the variation of these parameters directly on the variation of the contact forces and deformations of the fuel assemblies which house and control the movement of the fuel. Based on the results of the sensitivity study, this study found that the ACLP location has the largest effect on contact forces. And as such, any uncertainty in this parameter results in a rather large variation in the intensity of the contact force. Furthermore, specific recommendations are given to help control these variations as well as for further investigations on other parameters that may be significant to the design of fuel assemblies. Show less
