Automatic Preidentification of Fault Structural Traps from Graph View

Jing Li,Ting Xu,Wenjun Lv,Yu Kang,Hongwei Han

DOI: https://doi.org/10.1109/lgrs.2022.3184522

IF: 5.343

2022-01-01

IEEE Geoscience and Remote Sensing Letters

Abstract:Fault structural traps (FSTs) are one of the main gathering places for oil and gas, so their identification on the structure map is of great importance and thus a fundamental work in exploration. In the society of exploration geophysics, such a work is accomplished by experienced experts manually, which motivates us to propose a more effective method to identify the traps automatically. From the graph view, we propose a workflow to realize the automatic preidentification of FSTs preliminarily according to the subsurface faults distribution. The automatic preidentification strategy includes three main components: fault information acquisition, traps coarse identification, and rock body correction. First, the fault information acquisition is a fundamental work to obtain the connectivities between faults. Second, in the traps coarse identification, an undirected graph (undigraph) model is established to digitally describe the relationship between faults. Meanwhile, the alpha shape (AS) algorithm and depth first search (DFS) algorithm are used to add auxiliary lines for connected subgraphs and find basic loops of the undigraph, respectively. Finally, a correction algorithm is designed to find the traps formed by blocky rock bodies and their connected faults. The effectiveness of the FSTs preidentification (i.e., annotating all potential FSTs on the structure map) method is verified by illustrative examples in an oilfield, which are almost consistent with the expertise.