Research on Inter-Well Connectivity of Water-Flooding Reservoir: Temporal Neural Network Based on Graph Structure

Yulong Zhao,Huilin Li,Xingjie Zeng,Feng Ge,Liehui Zhang,Linsheng Wang,Bo Liao,Qingyu Xiao
DOI: https://doi.org/10.1016/j.geoen.2024.213221
2024-01-01
Geoenergy Science and Engineering
Abstract:Inter-well connectivity plays a crucial role in the development and production of water-flooding reservoirs, involving fluid transfer and mutual influence between different wells, and playing an important role in optimizing oilfield development and production. The traditional study of inter-well connectivity is mainly based on static information, while ignoring the dynamic changes over time and the graph structure of the well network. To address this issue, this paper proposes a temporal neural network based on graph structure (GSTNN). It combines the advantages of traditional neural networks and graph structures, fully considering the temporal information of the well network. A series of well network temporal feature information is used to predict inter-well connectivity within the framework of traditional neural networks, and optimize the model based on the graph structure of the well network. Experiments with GSTNN in different basic models and different graph sizes show that the method proposed in this paper is consistent with the actual geological characteristics of oil reservoirs. Compared with traditional neural network methods, GSTNN achieves more accurate results and shows good performance in less time in production prediction and inter-well connectivity reflection.
What problem does this paper attempt to address?