Inversion of 1D frequency- and time-domain electromagnetic data with convolutional neural networks

Inversion of electromagnetic data finds applications in many areas of geophysics. The inverse problem is commonly solved with either deterministic optimization methods (such as the nonlinear conjugate gradient or Gauss-Newton) which are prone to getting trapped in a local minimum, or probabilistic methods which are very computationally demanding. A recently emerging alternative is to employ deep neural networks for predicting subsurface model properties from measured data. This approach is entirely data-driven, does not employ traditional gradient-based techniques and provides a guess to the model instantaneously. In this study, we apply deep convolutional neural networks for 1D inversion of marine frequency-domain controlled-source electromagnetic (CSEM) data as well as onshore time-domain electromagnetic (TEM) data. Our approach yields accurate results both on synthetic and real data and provides them instantaneously. Using several networks and combining their outputs from various training epochs can also provide insights into the uncertainty distribution, which is found to be higher in the regions where resistivity anomalies are present. The proposed method opens up possibilities to estimate the subsurface resistivity distribution in exploration scenarios in real time.

[1]  Amir Adler,et al.  Deep-learning tomography , 2018 .

[2]  Davood Moghadas,et al.  1D joint multi‐offset inversion of time‐domain marine controlled source electromagnetic data , 2015 .

[3]  Sergey Ioffe,et al.  Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift , 2015, ICML.

[4]  Pejman Tahmasebi,et al.  Segmentation of digital rock images using deep convolutional autoencoder networks , 2019, Comput. Geosci..

[5]  G. W. Hohmann,et al.  4. Electromagnetic Theory for Geophysical Applications , 1987 .

[6]  Sergey Fomel,et al.  FaultSeg3D: Using synthetic data sets to train an end-to-end convolutional neural network for 3D seismic fault segmentation , 2019, GEOPHYSICS.

[7]  Evgeny Burnaev,et al.  Driving Digital Rock towards Machine Learning: predicting permeability with Gradient Boosting and Deep Neural Networks , 2018, Comput. Geosci..

[8]  David Pardo,et al.  Fast 1D inversion of logging-while-drilling resistivity measurements for improved estimation of formation resistivity in high-angle and horizontal wells , 2015 .

[9]  David L. Alumbaugh,et al.  Robust and accelerated Bayesian inversion of marine controlled-source electromagnetic data using parallel tempering , 2013 .

[10]  B. Minsley.  A trans-dimensional Bayesian Markov chain Monte Carlo algorithm for model assessment using frequency-domain electromagnetic data , 2011 .

[11]  The electromagnetic response of a horizontal electric dipole buried in a multi‐layered earth , 2018 .

[12]  R. Streich,et al.  Sensitivity of controlled-source electromagnetic fields in planarly layered media , 2011 .

[13]  J. Byun,et al.  Imaging subsurface resistivity structure from airborne electromagnetic induction data using deep neural network , 2020, Exploration Geophysics.

[14]  Alan Richardson,et al.  Seismic Full-Waveform Inversion Using Deep Learning Tools and Techniques , 2018, 1801.07232.

[15]  Jian Sun,et al.  Deep Residual Learning for Image Recognition , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[16]  Mary M. Poulton,et al.  Location of subsurface targets in geophysical data using neural networks , 1992 .

[17]  Xishuang Dong,et al.  A scalable deep learning platform for identifying geologic features from seismic attributes , 2017 .

[18]  M. Glinsky,et al.  Resolution and uncertainty in 1D CSEM inversion: A Bayesian approach and open-source implementation , 2010 .

[19]  Esben Auken,et al.  Piecewise 1D laterally constrained inversion of resistivity data , 2005 .

[20]  Nitish Srivastava,et al.  Dropout: a simple way to prevent neural networks from overfitting , 2014, J. Mach. Learn. Res..

[21]  B. Ursin,et al.  Electromagnetic fields in planarly layered anisotropic media , 2007 .

[22]  Vladimir Puzyrev,et al.  Deep learning electromagnetic inversion with convolutional neural networks , 2018, Geophysical Journal International.

[23]  Tomaso Poggio,et al.  Automated fault detection without seismic processing , 2017 .

[24]  V. Calo,et al.  Interpretation of deep directional resistivity measurements acquired in high-angle and horizontal wells using 3-D inversion , 2018 .

[25]  Jong Chul Ye,et al.  Deep Convolutional Framelets: A General Deep Learning Framework for Inverse Problems , 2017, SIAM J. Imaging Sci..

[26]  Timothy Dozat,et al.  Incorporating Nesterov Momentum into Adam , 2016 .

[27]  Yuan Yu,et al.  TensorFlow: A system for large-scale machine learning , 2016, OSDI.

[28]  Tao Zhao,et al.  Seismic facies classification using different deep convolutional neural networks , 2018, SEG Technical Program Expanded Abstracts 2018.

[29]  M. Jegen,et al.  On mapping seafloor mineral deposits with central loop transient electromagnetics , 2012 .

[30]  Martin J. Blunt,et al.  Stochastic Reconstruction of an Oolitic Limestone by Generative Adversarial Networks , 2017, Transport in Porous Media.

[31]  Albert Tarantola,et al.  Neural networks and inversion of seismic data , 1994 .

[32]  Ying Wang,et al.  Seismic Impedance Inversion Using Fully Convolutional Residual Network and Transfer Learning , 2020, IEEE Geoscience and Remote Sensing Letters.

[33]  Douglas W. Oldenburg,et al.  Simultaneous 1D inversion of loop loop electromagnetic data for magnetic susceptibility and electrical conductivity , 2003 .

[34]  Andrew Zisserman,et al.  Very Deep Convolutional Networks for Large-Scale Image Recognition , 2014, ICLR.

[35]  J. Byun,et al.  Resistivity Imaging from Electromagnetic Data Using a Fully Convolutional Network , 2019 .

[36]  Charles Blundell,et al.  Simple and Scalable Predictive Uncertainty Estimation using Deep Ensembles , 2016, NIPS.

[37]  D. Colombo,et al.  Deep learning-based EM inversion for reservoir monitoring , 2019, SEG Technical Program Expanded Abstracts 2019.

[38]  Mingliang Liu,et al.  Accelerating geostatistical seismic inversion using TensorFlow: A heterogeneous distributed deep learning framework , 2019, Comput. Geosci..

[39]  Hu Li,et al.  Bayesian Inversion of Logging-While-Drilling Extra-Deep Directional Resistivity Measurements Using Parallel Tempering Markov Chain Monte Carlo Sampling , 2019, IEEE Transactions on Geoscience and Remote Sensing.

[40]  Zoubin Ghahramani,et al.  Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning , 2015, ICML.

[41]  Chester J. Weiss,et al.  Mapping thin resistors and hydrocarbons with marine EM methods: Insights from 1D modeling , 2006 .

[42]  Kerry Key,et al.  1D inversion of multicomponent, multifrequency marine CSEM data: Methodology and synthetic studies for resolving thin resistive layers , 2009 .

[43]  G. P. Oliveira,et al.  Automatic classification of hydrocarbon "leads" in seismic images through artificial and convolutional neural networks , 2019, Comput. Geosci..