An Efficient Hydraulic Fracturing Fracture Network Connectivity Evaluation Method Based on Point Cloud Pattern Recognition
X. Liu, Z. Chen, Xiaoyang Zhang, G. Y. Zhong, S. M. Li · 2024
ABSTRACT: Due to the lack of methods to reconstruct the fracture network after fracturing, it is difficult to evaluate the connectivity of the fracture network. Inspired by the similarity between microseismic data and point cloud data, we have proposed a fracture network reconstruction model based on point cloud pattern recognition algorithms. The model is integrated by the RANASC fracture orientation identification method, the Alpha shape fracture shape identification method and the DBSCAN fracture denoising method. Besides, we have also developed a fracture connectivity index calculation method based on topology theory. Through the application example, it is found that our model could efficiently obtain the fracture network with excellent tightness and fitness with microseismic events after fracturing. The fracture connectivity map could intuitively and quantitatively represent the connectivity relationship between wellbore, hydraulic fractures, and natural fractures. The proposed model in this paper would provide the essential foundation for shale hydraulic fracturing effectiveness evaluation and operation optimization. 1. INTRODUCTION Shale gas reservoirs have the characteristics of ultra-low permeability and ultra-low porosity, which require multi-stage hydraulic fracturing for production enhancement. Due to the heterogeneity of shale reservoirs and the complex distribution of natural fractures and stress fields, it is difficult to predict and characterize the complex fracture network distribution and connectivity after hydraulic fracturing (Liu et al., 2021). Currently, there is a lack of efficient and effective methods for characterizing complex fracture network connectivity. According to the different ideas for obtaining complex fracture networks, the current methods can be divided into the fracture network simulation method and the fracture network reconstruction method. The fracture network simulation method simplifies the intricate intersecting patterns within the fracture network into the fundamental connectivity relationship between two fractures. After delineating the distribution of natural fractures, the propagation path of hydraulic fractures is numerically simulated based on the specific hydraulic-natural fracture interaction criterion. The microseismic data obtained after hydraulic fracturing only serves to verify and calibrate the numerical simulated fracture model. The commonly used numerical methods mainly include the finite element method (Wangen, 2011), extended finite element method (Shi et al., 2016), displacement discontinuity method (Wu and Olson, 2015), discrete element method (Zhang et al., 2019), and recently developed phase-field simulation method (Fan et al., 2022). The accuracy of fracture network simulation methods severely depends on the precise description of natural fractures and rock mechanical properties in the reservoir. Since it is difficult to obtain an accurate description of natural fractures in actual engineering applications, the accuracy and reliability of the complex fracture network obtained by fracture simulation methods are largely limited.