There have been numerous efforts exploring the application of machine learning (ML) techniques for field-scale automated interpretation for well log data. A critical prerequisite for automatic interpretation via computational means is to ensure that the log characteristics are reasonably consistent across multiple wells. Manually correcting logs for consistency is laborious, subjective, and error prone. For some logs, such as gamma ray and neutron porosity, systematic inconsistencies or errors can be caused by borehole effects as well as miscalibration. Biased or consistently inaccurate data in the logs can confound ML approaches into learning erroneous relationships, which leads to inaccurate lithology prediction, reservoir estimation, and incorrect formation markers, etc. To overcome such difficulties, we have developed a deep learning method to provide petrophysicists with a set of consistent logs through an automated workflow. Presently, the corrections we target are systematic shifts or errors on the common logs, especially gamma ray and neutron logs, and to a lesser extent, local errors due to washouts. This workflow can be separated into two steps. The first step represents a semiautomated approach for selecting wells to be used as training and validation; this approach employs statistical analysis to detect and segregate wells with similar log distributions. The second step is the core process of this workflow. It samples intervals across multiple logs identified by the first step and trains a convolutional neural network (CNN) with a U-Net architecture to identify and correct systematic errors such as shifts, gains, random noises, and small local disturbances. The training process is self-supervised and does not require any human labels. This self-supervised deep learning methodology is capable of automatically discovering unique implicit features and contextually applying the relevant log correction. The proposed method has been applied to multiple oil fields around the world. Field tests were successfully conducted in two scenarios. The first scenario aims to correct for synthetic noise and artifacts added to field data when triple-combo logs (gamma ray, density, neutron, and resistivity logs) were available—in this scenario the tests were targeted to correct systematic errors to the gamma ray logs or to both gamma ray and neutron porosity logs simultaneously. The second scenario aims to correct for original field noise in gamma ray logs and neutron porosity logs when quad-combo logs (gamma ray, density, neutron, resistivity, and compressional slowness logs) are available.
Contents of this paper were reviewed by the Technical Committee of the 16 th International Congress of the Brazilian Geophysical Society and do not necessarily represent any position of the SBGf, its officers or members. Electronic reproduction or storage of any part of this paper for commercial purposes without the written consent of the Brazilian Geophysical Society is prohibited.
We present a work flow for joint inversion of sonic flexural-wave dispersion data and array-induction resistivity data acquired in a vertical well. The work flow estimates a pixel-based radial distribution of water saturation and porosity extending several feet into the formation at each log depth. Radial changes in saturation and porosity are caused by mud-filtrate invasion and mechanical damage, respectively. The flexural-wave and array-induction data have similar multiple investigation depths extending several feet into the formation. Furthermore, flexural-wave data are sensitive to porosity but have weak sensitivity to saturation, whereas induction data are sensitive to both porosity and saturation. Thus, integration of these data in a joint inversion can help to characterize the formation beyond the altered zone and reduce uncertainty of the interpretation. The work flow is validated on synthetic data for several scenarios of near-wellbore alteration. The work flow is then applied to field data from an offshore well drilled with oil-based mud in a gas-bearing clastic formation. The results are compared with traditional interpretation and core analysis, demonstrating an efficient and accurate inversion-based work flow that can complement traditional formation evaluation in challenging conditions.
We present an automatic inversion method for data acquired in a vertical well by the sonic, induction, and density borehole logging tools. The method is designed for oil-bearing or gas-bearing formations drilled with oil-base mud. The inversion accounts for challenging scenarios where the tool sensors are affected by filtrate invasion, gas phase, complex mineralogy and mechanical damage. The output consists of porosity and radial distributions of fluid saturations and pore shape extending several feet from the wellbore. The output is robust, accurate, and consistent with radial investigation depths of all the tools. The formation model assumed in the inversion has homogeneous porosity and radially varying pore shape, oil, gas, and water saturation. Radial changes in fluid saturation and pore shape are caused by filtrate invasion and mechanical damage respectively. The data for different tools are simulated from sonic and electromagnetic forward solvers, linked to the formation model through a saturation-resistivity transform and an effective medium rock physics model. The inversion estimates formation properties such that the simulated data match the measured data. For the first time, sonic data for both dipole flexural wave and monopole compressional-headwave are included. These data are sensitive to porosity and pore shape effects, and the compressional-headwave additionally provides sensitivity to gas saturation in soft formations. The inversion was tested on synthetic data and applied to two field data sets for gas-bearing formations. The results are visualized as 2D images with radial distribution of formation properties at each log depth. The images characterize radial depth of filtrate invasion and mechanical damage, which can guide completion and production decisions. The inversion also provides far-field saturation and porosity. The far-field properties are in overall good agreement with core data and traditional interpretation, with differences from traditional interpretation in key intervals. Quality controls enable checking validity of models assumed in the inversion. The results demonstrate an efficient inversion framework for guiding formation evaluation decisions in challenging scenarios.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.