Lithology identification from well log curves via neural networks with additional geological constraint

Jiang，Chunbi1,3; Zhang，Dongxiao1,2; Chen，Shifeng3

2021-06-24

10.1190/geo2020-0676.1

GEOPHYSICS   影响因子和分区

0016-8033

1942-2156

Lithology identification is of great importance in reservoir characterization. Recently, many researchers have applied machine learning techniques to solve lithology identification problems from well log curves, and their works show three main characteristics. First, most works predict lithofacies using features measured during logging, while very few consider adding stratigraphic sequence information that is available prior to drilling to solve this problem. Second, most studies predict lithofacies using measured properties of one depth point, while few take the influence of the neighboring formation into account. Third, due to a lack of publicly available interpreted well log data, previous research has concentrated on applying different algorithms on their private dataset, making it impossible to perform a comparison. We propose a machine learning framework to solve the lithology classification problem from well log curves by incorporating an additional geological constraint. The constraint is a stratigraphic unit, and we use it as an additional feature. We evaluated three types of recurrent neural networks (RNNs), bidirectional long short-term memory (Bi-LSTM), bidirectional gated recurrent unit (Bi-GRU), GRU-based encoder-decoder architecture with attention (ABi-GRU), and two types of one-dimensional convolutional networks (1D CNNs), temporal convolutional networks (TCN) and multi-scale residual networks (MsRNet) on a publicly available dataset from the North Sea. Both RNN-based networks and 1D CNN-based networks can process sequential data, enabling the model to have access to information from neighboring formations when performing lithofacies prediction at a particular depth. Our experiments show that geological constraint improves the performance of the models significantly, and that the overall performance of RNN-based networks is better and more consistent.

Artificial intelligence Lithology Machine learning Well-log interpretation

EI ; SCI

英语

通讯

Geochemistry & Geophysics

Geochemistry & Geophysics

WOS:000711980200016

SOC EXPLORATION GEOPHYSICISTS

20212610570115

Convolution ; Convolutional neural networks ; Drilling machines (machine tools) ; Forecasting ; Lithology ; Logging while drilling ; Machine learning ; Stratigraphy ; Turing machines

Geology:481.1 ; Machine Tools, General:603.1 ; Information Theory and Signal Processing:716.1 ; Computer Theory, Includes Formal Logic, Automata Theory, Switching Theory, Programming Theory:721.1

GEOSCIENCES

2-s2.0-85108675832

Scopus

1.Intelligent Energy Laboratory,Frontier Research Center,Peng Cheng Laboratory,China
2.School of Environmental Science and Engineering,Southern University of Science and Technology,China
3.Shenzhen Institutes of Advanced Technology,Chinese Academy of Sciences,China

Jiang，Chunbi,Zhang，Dongxiao,Chen，Shifeng. Lithology identification from well log curves via neural networks with additional geological constraint[J]. GEOPHYSICS,2021,86(5).

Jiang，Chunbi,Zhang，Dongxiao,&Chen，Shifeng.(2021).Lithology identification from well log curves via neural networks with additional geological constraint.GEOPHYSICS,86(5).

Jiang，Chunbi,et al."Lithology identification from well log curves via neural networks with additional geological constraint".GEOPHYSICS 86.5(2021).