Inverse Process Analysis for the Acquisition of Thermophysical Data

Frankel, Jay ; Sabau, Adrian

University of Tennessee

关键词: Sample Holders;    Aluminium;    Thermocouples;    Temperature Distribution;    Heat Flux;   

DOI  :  10.2172/861453
RP-ID  :  ORNL/TM-2005-132
RP-ID  :  FC07-01ID14249
RP-ID  :  861453

美国|英语

来源: UNT Digital Library

【 摘 要 】

One of the main barriers in the analysis and design of materials processing and industrial applications is the lack of accurate experimental data on the thermophysical properties of materials. To date, the measurement of most of these high-temperature thermophysical properties has often been plagued by temperature lags that are inherent in measurement techniques. These lags can be accounted for with the appropriate mathematical models, reflecting the experimental apparatus and sample region, in order to deduce the desired measurement as a function of true sample temperature. Differential scanning calorimeter (DSC) measurements are routinely used to determine enthalpies of phase change, phase transition temperatures, glass transition temperatures, and heat capacities. In the aluminum, steel, and metal casting industries, predicting the formation of defects such as shrinkage voids, microporosity, and macrosegregation is limited by the data available on fraction solid and density evolution during solidification. Dilatometer measurements are routinely used to determine the density of a sample at various temperatures. An accurate determination of the thermophysical properties of materials is needed to achieve accuracy in the numerical simulations used to improve or design new material processes. In most of the instruments used to measure properties, the temperature is changed according to instrument controllers and there is a nonhomogeneous temperature distribution within the instrument. Additionally, the sample temperature cannot be measured directly: temperature data are collected from a thermocouple that is placed at a different location than that of the sample, thus introducing a time lag. The goal of this project was to extend the utility, quality and accuracy of two types of commercial instruments -a DSC and a dilatometer - used for thermophysical property measurements in high-temperature environments. In particular, the quantification of solid fraction and density during solidification was deemed of critical importance. To accomplish this project goal, we redesigned sample holders and developed inverse mathematical methods to account for system lags. The desired property could then be correlated to the proper sample temperature. For the NETZSCH DSC 404C instrument with a high-accuracy heat capacity sensor, a mathematical model was developed by assuming that each component was isothermal and that the heat transfer among components occurred by conduction and radiation. Model parameters included effective conduction time constants and radiation time constants. Several model cases were investigated to assess the effect of heat transfer interactions. New features that have not been considered in previous DSC models were included in the present study. These new features included (a) considering the sensor platform, (b) accounting for the heat loss through the stem, and (c) considering the lag between furnace temperature and set point temperature. Comparisons with experimental results showed that temperature lags in heat flux DSC instruments could be determined by performing a heat transfer analysis based on a comprehensive model. The proposed mathematical model yielded accurate results over a wide temperature range during heating and cooling regimes. The induced thermal lag in the Theta Industries dual push-rod horizontal dilatometer is apparent owing to the distance of the thermocouple from the actual sample. In a near steady-state mode of operation, this apparent problem is minimal. However, in a transient situation, where the density is varying as a function of time, the temperature output from the remote temperature sensor must be adjusted in order to reflect the sample temperature. The conventional push-rod dilatometer insert was modified significantly to allow an accurate correlation of the measured density to the predicted sample temperature of alloys in the phase-change regime. This new configuration made use of a standard furnace assembly; however, the specimen was symmetrically encased in a well-instrumented cylindrical graphite shell. The new insert was designed, fabricated, instrumented, calibrated, and tested. It was demonstrated, using an array of high-precision thermocouples, that a nearly uniform temperature distribution existed in the sample holder. The combination of system geometry and high-conductivity sample holder material promoted the development of a simplified but highly effective inverse heat transfer model. The prediction of this model properly correlated the measured density in the phase change regime to that of the actual sample temperature based on using remote, sample-holder temperature measurements. It was demonstrated (using A356) that accurate modeling of the solid fraction was important for accounting for thermal lags in the phase-change regime. Additionally, the average heat transfer coefficient during phase change was also calculated and compared to existing results.