Building structural models is a foundational step in the exploration of deep subsurface resources, such as oil and gas. However, in some complex surveys, expert cognition of the subsurface geological structures in the area is often incomplete, and the quality of seismic data rapidly deteriorates, which leads to poor structural interpretation and makes structural modeling a time-consuming and laborious task. To address this challenge, a structural modeling method based on human–computer knowledge interaction using knowledge graphs (KGs) is proposed. Initially, a KG of the structural model is established based on the original structural interpretation. Subsequently, it is gradually improved through iterative human–computer interaction to obtain a complete KG. Finally, the KG is used to guide the reconstruction of geological surfaces. In the process of improving the initial KG, humans can provide expertise to computers by editing the KG, and computers can cognize the data through the KG to help humans discover errors or new knowledge in the original structural interpretation. The method was tested on a field dataset and yielded robust and efficient results.

构建结构模型是勘探石油和天然气等深层地下资源的基础步骤。然而，在一些复杂的勘察中，专家对该地区地下地质构造的认知往往不完整，地震资料质量迅速恶化，导致构造解释效果差，构造建模成为一项耗时费力的工作。为了应对这一挑战，提出了一种基于知识图谱（KG）的人机知识交互的结构建模方法。最初，根据原始结构解释建立结构模型的知识图谱。随后通过迭代人机交互逐步完善，得到完整的KG。最后，利用知识图谱指导地质表面的重建。在改进初始KG的过程中，人类可以通过编辑KG向计算机提供专业知识，计算机可以通过KG认知数据，帮助人类发现原始结构解释中的错误或新知识。该方法在现场数据集上进行了测试，并产生了稳健且有效的结果。