Xinjiang Petroleum Geology ›› 2024, Vol. 45 ›› Issue (5): 595-603.doi: 10.7657/XJPG20240512

• APPLICATION OF TECHNOLOGY • Previous Articles     Next Articles

Shale Lithology Identification Based on Improved Random Forest Algorithm:A Case of Lucaogou Formation in Junggar Basin

QIN Zhijun1,2(), CAO Yingchang1, FENG Cheng3   

  1. 1. School of Geosciences, China University of Petroleum (East China), Qingdao, Shandong 266580, China
    2. Research Institute of Exploration and Development, Xinjiang Oilfield Company, PetroChina, Karamay, Xinjiang 834000, China
    3. College of Petroleum, Karamay Campus, China University of Petroleum (Beijing), Xinjiang 834000, China
  • Received:2024-07-24 Revised:2024-08-12 Online:2024-10-01 Published:2024-10-09

Abstract:

In the application of reservoir lithology identification, the efficiency, accuracy and effective information integration ability of machine learning algorithm have been fully verified, especially in unconventional reservoirs with strong heterogeneity such as shale. Based on the optimal selection of parameters such as natural gamma, T2 geometric mean, structural index, skeleton density index, density, and deep lateral resistivity, and using a random forest algorithm combined with recursive feature elimination (RF-RFE), major lithologies of the shale reservoirs in the Middle Permian Lucaogou formation in the Junggar basin were identified. Lithology prediction was conducted on the same dataset using conventional RF and support vector machine (SVM) algorithms, and the results were compared with those obtained from thin-section identifications. It is found that RF-RFE yields better results with only half of the logging parameters, and the parameters defined by optimal selection help reduce the algorithm’s running time. Thus, the use of RF-RFE algorithm can realize optimal selection of characteristic logging parameters, more accurate identification of shale lithology, and reduction of running time. The algorithm provides a new approach for complex lithology identification and multi-parameter selection.

Key words: random forest algorithm, recursive feature elimination, optimal selection, Middle Permian, Lucaogou formation, shale reservoir, lithology identification

CLC Number: