Inverse Modeling of Laboratory Tests for Rock Thermal Property Estimation

Abstract

A laboratory experimental system was designed and validated for measuring thermal properties (heat conductivity and specific heat) on cylindrical rock samples. The experimental setup includes four semi-insulated cells holding cylindrical rock samples; four heaters at the bottom of the samples; and four temperature probe loggers, mounted in holes drilled at the top of the samples. For each of the test runs, four rock samples were used. During the first test run, the heaters are switching on for a specified time and then turned off for temperature recovery. 2-nd, 3-rd and 4-th test runs were repeated for all samples, changing their places in four cells correspondingly. Temperature records were obtained for over the entire duration of the tests. The temperature data were inverted using iTOUGH2-EOS3 to estimate 12 unknown parameters: four values of the thermal conductivity, four values of the specific heat, and four values of the heat exchange coefficients of the semi-insulated cells. The uncertainty of the estimated parameters were reasonable, i.e., the standard deviations were within a few percents of the parameter value. Estimates of the thermal properties of benchmark samples (quartz QU1) and 31 rock samples from different volcanogenic rocks (rhyolite tuffs and lavas Triassic and Quaternary age) were obtained.