A Method for Predicting Different Types of Natural Fractures in Tight Sandstone Based on the Secondary Rescaled Range Analysis of Logging Curves: A Case Study From the Chang 7 Member in Huaqing Oilfield, Ordos Basin, China

Currently, there are various methods for predicting natural fractures using logging data, however these methods are primarily for predicting the number and location of fractures. This is making it difficult to determine fracture types. This paper introduces the R/S-FD method, and combined with the natural fracture development pattern in the study area, secondary R/S analysis was introduced to construct the Secondary R/S-FD method. This method overcomes the limitations of traditional R/S-FD methods that can only predict the location of fractures and cannot predict the type of fractures. After eliminating systematic errors, the prediction accuracy of the Secondary R/S-FD method for bedding fractures and high-angle fractures reaches 73% and 74%, respectively. By analyzing the fracture development characteristics of 23 wells in the study area, the research provided insights into the development characteristics of bedding fractures and high-angle fractures in oil layers within the region. The secondary R/S-FD method is a precise, fast, and cost-effective approach for predicting the development characteristics of different types of natural fractures. The next step involves leveraging a large number of fracture prediction cases as the data foundation, based on big data analysis and machine learning techniques, to establish a correlation between the F value and fracture type and number to enabling more accurate predictions of the types and quantities of natural fractures.