Sensitivity Analysis of the 9975 Packaging Thermal Response Using Factorial Design Methods

A method is presented for using the statistical design of experiment technique (2^k Factorial Design) in the sensitivity analysis of the thermal (temperature) response of the 9975 radioactive material packaging where multiple thermal properties of the impact absorbing and fire insulating material Celotex and certain boundary conditions are subject to uncertainty. 2^k Factorial Design method is very efficient in the use of available data and is capable of analyzing the impact of main variables (Factors) and their interactions on the component design. The 9975 design is based on detailed finite element (FE) analyses and extensive proof testing to meet the design requirements given in 10CFR71 [1]. However, the FE analyses use Celotex thermal properties that are based on published data and limited experiments. Celotex is an orthotropic material that is used in the home building industry. Its thermal properties are prone to variation due to manufacturing and fabrication processes, and due to long term environmental exposure. This paper will evaluate the sensitivity to variations in thermal conductivity of the Celotex, convection coefficient at the drum surface, and drum emissivity (herein called Factors) of the thermal response of 9975 packaging under Normal Conditions of Transport (NCT). Application of this methodology will ascertain the robustness of the 9975 design and it can lead to more specific and useful understanding of the effects of various Factors on 9975 performance.