Semi-empirical method for measuring thermal conductivity of fibrous insulation materials

An experimental apparatus has been constructed to explore and isolate the different modes of heat transfer through fibrous insulation materials. This apparatus utilizes a cut-bar methodology to measure conductivity, and current processing uses a 1-D heat transfer assumption. However, there are radial losses due to natural convection and radiation which challenge this assumption. In order to validate and quantify this assumption, a Computational Fluid Dynamics model has been developed to study these losses. The models includes a laminar model to quantify losses from natural convection, and a discrete ordinates model to determine the radiative losses. It was found that there are non-negligible losses that occur, negating the 1-D assumption. Non-linear models were investigated to identify methods of improving the thermal conductivity measurement accuracy, finding that a quadratic polynomial resulted in more consistent experimental uncertainty than the other models in the study. The error from the 1-D assumption analysis ranged from 31.90% to 66.92%, while the quadratic model ranged from-8.77% to-9.70%. The improved consistency of the quadratic error could be used as a correction factor for additional experimental conditions, although further study will need to be done. The CFD models also informed improvements that can be made to reduce losses and better predict the thermal conductivity of the sample.