AUTHOR=Li Jing , Gao Guozhong 

TITLE=Digital construction of geophysical well logging curves using the LSTM deep-learning network

JOURNAL=Frontiers in Earth Science

VOLUME=Volume 10 - 2022

YEAR=2023

URL=https://www.frontiersin.org/journals/earth-science/articles/10.3389/feart.2022.1041807

DOI=10.3389/feart.2022.1041807

ISSN=2296-6463

ABSTRACT=Frequently, a complete well logging suit is needed but not available or some are missing. Mudstone section is prone to wellbore collapse, which frequently causes distortion in well logs.  In many cases, some well logging curves were never measured, but required for petrophysical or other analysis. However, re-logging is expensive and difficult to achieve, while manual construction of the missing well logging curves is costly and low in accuracy.  The rapid technical evolvement of deep learning algorithms makes it possible to realize the digital construction of missing well logging curves with high precision in an automatic manner.  In this paper, a workflow for the digital construction of wells logging curves based on the long short-term memory (LSTM) network is proposed. The LSTM network is chosen because it has the advantage of avoiding the vanishing gradient problem existing in the traditional Recurrent Neural Networks (RNN). Additionally, it can process sequential data. When it is used in the construction of missing well logging curves, it not only considers the relationship between each logging curve, but also considers the influence of the data at the previous depth on those at the following depth.  This is validated by the exercises on construction of acoustic, neutron porosity and resistivity logging curves using the LSTM network, which effectively achieve high-precision construction of missing well logging curves.  Exercises show that the LSTM network is much more superior than the RNN in the digital construction of well logging curves in terms of accuracy, efficiency and reliability.