Predicting Thermal Conductivity of Rocks From

A correlation is developed to predict thermal conductivity of drill cores from the L o s Azufres geothermal field. Only andesites are included as they are predominant. Thermal conductivity of geothermal rocks is in general scarce and its determination is not simple. Almost all published correlations were developed for sedimentary rocks. Typically, for igneous rocks, chemical or mineral analyses are used for estimating conductivity by using some type of additive rule. This requires specialized analytical techniques and the procedure may not be sufficiently accurate if, for instance, a chemical analysis is to be changed into a mineral analysis. Thus a simple and accurate estimation method would be useful for engineering purposes. The present correlation predicts thermal conductivity from a knowledge of bulk density and total porosity, properties which provide basic rock characterization and are easy to measure. They may be determined from drill cores or cuttings, and the procedures represent a real advantage given the cost and low availability of cores. The multivariate correlation proposed is a quadratic polynomial and represents a useful tool to estimate thermal conductivity of igneous rocks since data on this property is very limited. For porosities between 0% and 25%, thermal conductivity is estimated with a maximum deviation of 22% and a residual mean square deviation of 4.623-3 in terms of the loglo (kPb) variable. The data were determined as part of a project which includes physical, thermal and mechanical properties of drill cores from L o s Azufres. For the correlation, sixteen determinations of thermal conductivity, bulk density and total porosity are included. The conductivity data represent the first determinations ever made on these rocks. NOMENCLATURE Ai coefficients of proposed model F formation factor k thermal conductivity s standard deviation Y loglo (kpb! pb bulk density ps solids density et total porosity