Session: K7-01: THERMOPHYSICAL PROPERTIES I

Paper Number: 130194

130194 - Analysis of Approximations in Thermal Diffusivity Measurements Using High Fidelity Simulations 

Abstract: 

Advances in high performance computing and simulation software are rapidly increasing reliance on digital or simulation-driven design of thermal processes or systems, often resulting in better, faster, and cheaper development. However, simulation results are only as good as the material properties specified in the simulation. Thus, simulation-driven design processes are improved by increasing the accuracy of thermophysical property measurements. This work uses high-fidelity simulations to evaluate the effects of approximations used in the flash method for measuring thermal diffusivity.

Many thermophysical properties, including thermal diffusivity, cannot be measured directly. Techniques must be employed to estimate thermal diffusivity using measurements of related properties. Furthermore, thermal diffusivity is a transient property, so measuring it requires taking measurements during transient heat transfer. Amenable conditions can be created by depositing energy into a sample and observing the evolution of the temperature profile throughout the transient process. A common method for doing this is the flash method, which uses a laser or flash lamp to deposit energy within a sample. To calculate thermal diffusivity, a mathematical model relating the evolution of the temperature profile to thermal diffusivity must be developed. Many models which have been developed use thermal diffusivity to mathematically predict temperature at a point in the sample as a function of time. The thermal diffusivity which best fits the data is determined. During model development, it is typical for some, or all, of the following approximations to be leveraged: all energy is instantaneously deposited in an infinitesimally thin layer at the surface of the sample, negligible heat is lost to the environment, and conduction through the sample is one-dimensional. The most common model used to calculate thermal diffusivity, the Parker model, uses all of these approximations. Although these approximations simplify the model, they often have adverse effects on thermal diffusivity estimates. High-fidelity simulations are used to investigate the extent to which these approximations affect estimated values for the thermal diffusivity.

Realistic simulations which incorporate heat loss to the environment, temporally and spatially varying energy deposition, and in-depth energy absorption are run to generate simulated data. Estimates of thermal diffusivity are calculated using the Parker model and other previously proposed models. The results of each model are compared to the actual thermal diffusivity. An analysis of each model’s performance is presented and the effects of their approximations on thermal diffusivity estimates are identified. Improved mathematical models which reduce these effects are being developed.

Presenting Author: Tage T. Burnett Brigham Young University

Presenting Author Biography: Tage is an undergraduate student studying mechanical engineering at Brigham Young University. His research interests include computational heat transfer and fluid dynamics. He plans on applying to graduate programs for computational science & engineering in the fall.

Authors:

Tage T. Burnett Brigham Young University
Jakob G. Bates Brigham Young University
Matthew R. Jones Brigham Young University
Christopher R. Dillon Brigham Young University
John Tencer Sandia National Laboratories