Application of SVM algorithm in fluid prediction of volcanic reservoirs in Nanpu Sag, Bohai Bay Basin
Volcanic rock reservoirs are affected by many factors such as lithofacies, lithology, and reservoir space types, and fluid identification is difficult, which is one of the difficulties in well logging interpretation. It is urgent to establish a convenient and quick identification method. For this re...
Saved in:
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | zho |
| Published: |
Editorial Department of Petroleum Reservoir Evaluation and Development
2023-04-01
|
| Series: | Youqicang pingjia yu kaifa |
| Subjects: | |
| Online Access: | https://red.magtech.org.cn/fileup/2095-1426/PDF/1682496939069-139812177.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Volcanic rock reservoirs are affected by many factors such as lithofacies, lithology, and reservoir space types, and fluid identification is difficult, which is one of the difficulties in well logging interpretation. It is urgent to establish a convenient and quick identification method. For this reason, the SVM（Support Vector Machine） algorithm of machine learning is used to predict the fluids of unknown reservoirs for the volcanic rock reservoirs in the Nanpu Sag of the Bohai Bay Basin. The research shows that: ① Comprehensive application of core, well logging, mud logging and other data to optimize fluid sensitive characteristic parameters, single information sensitive parameters are acoustic time difference, compensation density, resistivity, multi-information fusion parameters are natural gamma relative value, total hydrocarbon Ratio, hydrocarbon gas density index, hydrocarbon gas humidity index, the above seven parameters participate in the model establishment; ②Using the SVM algorithm for volcanic fluid prediction, the reservoir fluid is divided into three types: oil layer, oil-water layer and water layer. Sensitive parameters of well logging and mud logging are selected, and a reliable sample library is trained. The correct judgment rate of the prediction library reaches 90 %. The prediction application of SVM algorithm shows that it has low calculation complexity and strong generalization ability, which can quickly identify the fluid properties of volcanic rocks and provide a reliable basis for the analysis of oil and gas accumulation rules and the production and development of geological reserves. |
|---|---|
| ISSN: | 2095-1426 |