Identification of Diagenetic Facies Logging of Tight Oil Reservoirs Based on Deep Learning—A Case Study in the Permian Lucaogou Formation of the Jimsar Sag, Junggar Basin

Abstract

As a typical tight oil reservoir in a lake basin, the Permian Lucaogou Formation of the Jimsar Sag in the Junggar Basin has great potential for exploration and development. However, at present, there are few studies on the identification of the diagenetic facies of tight oil reservoir logging in the study area, and the control effect of diagenesis on tight oil reservoirs is not clear. The present work investigates the diagenesis and diagenetic facies logging of the study area, making full use of core data, thin sections, and logs, among other data, in order to understand the reservoir characteristics of the Permian Lucaogou Formation in the Jimsar Sag. The results show that the Lucaogou Formation has undergone diagenetic activity such as compaction, carbonate cementation, quartz cementation, and clay mineral infilling and dissolution. The diagenetic facies are classified according to mineral and diagenetic type, namely, tightly compacted facies, carbonate-cemented facies, clay mineral-filling facies, quartz-cemented facies, and dissolution facies. The GR, RT, AC, DEN, and CNL logging curves were selected, among others, and the convolutional neural network was introduced to construct a diagenetic facies logging recognition model. The diagenetic facies of a single well was divided and identified, and the predicted diagenetic facies types were compared with thin sections and SEM images of the corresponding depths. Prediction results had a high coincidence rate, which indicates that the model is of a certain significance to accurately identify the diagenetic facies of tight oil reservoirs. Assessing the physical properties of the studied reservoirs, dissolution facies are the dominant diagenetic facies in the study area and are also the preferred sequence for exploration—to find dominant reservoirs in the following stage.