| Abstract |
In this study we present first results of improvements to the classical divided bar and needle probe methods for measuring rock thermal properties. Both methods are fairly accurate in determining thermal conductivity using approximate analytical expressions, whereas thermal diffusivity and heat capacity are generally not measured. The improvements we implement in order to measure these properties as well, include, for both methods, a combination of fast numerical finite element forward modelling and a Markov Chain Monte Carlo inversion scheme for estimating parameters. In order to extend the classical divided bar method to measure thermal diffusivity in addition to conductivity, the apparatus is run in a transient mode. Temperatures are measured as a function of time across the stack of rock sample and standards. At the upper surface of the stack, a time varying temperature is imposed. By Monte Carlo inversion of the associated transient data set, thermal conductivity as well as thermal diffusivity and heat capacity may be measured with high accuracy. The needle probe, consisting of a thin cylinder containing a heating wire and a thermistor, is inserted into the rock sample to be measured. The probe and the surrounding sample are heated, and the thermistor measures the temperature increase at the probe centre. We apply a numerical forward model and use the entire temperature response to measure thermal conductivity, with the potential of estimating thermal diffusivity as well. |