Application of BiLSTM in Lithology Identification of Beach-Bar Sand Reservoir
The tight beach-bar sand reservoir in the study area has rich petroleum reserves and high exploration and development potential. However, it is characterized by deep burial, thin single-layer thickness, ultra-low permeability, complex pore structure, and extremely low natural productivity of the sin...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | zho |
| Published: |
Editorial Office of Well Logging Technology
2023-06-01
|
| Series: | Cejing jishu |
| Subjects: | |
| Online Access: | https://www.cnpcwlt.com/#/digest?ArticleID=5493 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The tight beach-bar sand reservoir in the study area has rich petroleum reserves and high exploration and development potential. However, it is characterized by deep burial, thin single-layer thickness, ultra-low permeability, complex pore structure, and extremely low natural productivity of the single well and it is difficult to classify the reservoir and identify the lithology. Based on the characteristic of time sequence logging data, a bi-directional long short-term memory neural network (BiLSTM) lithology identification model is constructed. The Random Forest method is used to conduct feature selection for conventional logging parameters and other parameters. The selected parameters are used as input variables to train the BiLSTM model. The model is applied to validate the well data of the test set, and the results showed that the accuracy of the model is 0.86, achieving good application results. This proves that BiLSTM model is suitable for the lithology identification of the beach-bar sand reservoir. |
|---|---|
| ISSN: | 1004-1338 |