Using neural networks to predict thermal conductivity from geophysical well logs

Abstract

International audienceWe present a new approach, based on neural networks, to predict the thermal conductivity of sedimentary rocks from a set of geophysical well logs. This method is calibrated on Ocean Drilling Program (ODP) data, which provide several thousands of conductivity measurements combined with five geophysical well logs (sonic, density, neutron porosity, resistivity and gamma ray). This data set is used to train multilayer perceptrons (MLP) and to find an empirical relationship between well logs (MLP inputs) and thermal conductivity (MLP output). Validation tests suggest that MLP provide better estimates of thermal conductivity (within ?15 per cent confidence level) than classical linear models, and still give satisfying results with sets of only four well logs if neutron porosity is included. In two ODP sites (863B and 1109D), MLP's predictions are compared to conventional 'mixing' methods. Although this latter technique gives reliable results provided that rocks description is precise enough, the MLP is more straightforward, does not need any extra parameter and makes predictions in good agreement with the experimental trends. This method will be useful in the estimation of heat flow from data acquired in scientific and industrial boreholes

Evaluation of common mixing models for calculating bulk thermal conductivity of sedimentary rocks: Correction charts and new conversion equations

“…If core samples are not available, indirect methods are used to calculate TC from petrophysical properties, including porosity, a parameter provided through well logging (e.g., Balling et al, 1981;Goss and Combs, 1976;Goutorbe et al, 2006;Hartmann et al, 2005). Another indirect approach of TC determination uses the abundance and composition of the rock-forming minerals and the porosity as a multi-component system (e.g., Brailsford and Major, 1964;Brigaud et al, 1990;Demongodin et al, 1991, Vasseur et al, 1995.…”

Heat Flow and Thermal Structure of the Lithosphere

“…that are routinely logged (Goutorbe et al, 2006). In deep wells, the gradient can thus be estimated with a precision of 10-15%.…”

Heat flow of the eastern Canadian rifted continental margin revisited

“…This makes this region very exciting for geodynamic studies and the interpretation of these data very critical. We recently developed a new method based on neural network to estimate thermal conductivity from geophysical well logs with results that appear far better than the traditional methods (Goutorbe et al. , 2006).…”