Abstract
Stimulated reservoir volume (SRV) is the most critical index to represent the hydraulically fractured volume after stimulation. The existing models using microseismic events usually lack a sufficient unified basis on the geo-metric definition, leading to significant differences in SRV calculation and interpretation. This paper has generally conducted the classification and the evaluation for the existing models and the newly proposed model. According to three typical grid structures, we firstly divided SRV models into three categories: boundary-based, voxel-based, and density-based. Three typical SRV calculation methods are also proposed based on the alpha shape method, the voxel discretization, and the octree decomposition. Then, we have developed an integrated workflow based on synthetic events to evaluate their performance quantitatively under different fracture configurations and noise ratios. Through a large number of simulated calculations, it is indicated that the density-based method outperforms the other two methods in accuracy, adaptability, and anti-noise ability for complex fracture networks with different configurations. Besides, the engineering application has also its practicability and superiority in generating more refined grid structures. This paper attempts to summarize and evaluate the existing methods and proposes new methods based on the classification. It would provide helpful guidance for developing more superior SRV calculation methods.
NSTL
Elsevier