Research and Application on Automatic Stratigraphic Sequence Division and Comparison Method Based on Integrated Prediction Error Filter Analysis

The effectiveness of the integrated prediction error filter analysis (INPEFA) technique, which reveals the stratigraphic sequence structure with trend and turning points information by highlighting the cyclic features in the logging curves, is verified in a wide range of domestic and international applications. Despite the significant advantages of this technique in revealing the stratigraphic structure, its application process currently still relies on manual operation, which not only increases the complexity of processing, but also limits the efficiency of analysis. To address these issues, based on the technical principles and application characteristics of INPEFA, supplemented by Fourier transform to extract inflection point information, an automated sequence stratigraphic division process is established. Combined with the derivative dynamic time warping algorithm and dynamic programming method for multi-well sequence stratigraphic correlation, a set of automatic sequence stratigraphic division and correlation methods is proposed. In the practical application of the Qingshankou formation in the Qijia-Gulong sag area of the Songliao basin, the Qingshankou formation is automatically divided into two third-order sequences SQ1 and SQ2 and eight fourth-order sequences SQ11, SQ12, SQ13, SQ14, SQ21, SQ22, SQ23, and SQ24. Research shows that sequences have a significant control on the development of source rocks and sweet spots. The source rocks of the Qingshankou formation are mainly distributed in the transgressive systems tract of SQ1, while the main shale sweet spot Q9 develops in the late transgressive stage of the third-order sequence SQ1, and it is located in the period of stable to slowly falling lake level in the fourth-order sequence SQ13. Q8, which is located in the secondary regression period of SQ12, can be inferred to be a more favorable sweet spot based on its physical properties and the micro-migration of source rocks. The method proposed in this paper significantly improves the accuracy and efficiency of sequence division, and the division results have certain guiding significance for the prediction of source rocks in the Qingshankou formation and the deployment of oil and gas exploration.