Affiliation:
1. School of Mining & Petroleum Engineering, Department of Civil & Environmental Engineering University of Alberta, Edmonton, Alberta, Canada.
Abstract
Abstract
One of the core assumptions of most deep learning-based data-driven models is that samples are independent. However, this assumption poses a key challenge in production forecasting - performance is influenced by well interference and reservoir connectivity. Most shale gas wells are hydraulically fractured and exist in complex fracture systems, and the neighbouring well characteristics should also be considered when constructing data-driven forecast models. Researchers have explored using the Graph Convolutional Network (GCN) to address this issue by incorporating neighbouring well characteristics into production forecasting models. However, applying GCN to field-scale studies is problematic, as it requires training on a full batch, leading to gigantic cache allocation. Additionally, the transductive nature of GCN poses challenges for direct generalization to unseen nodes. To overcome these limitations, we adopt the Graph Sampling and Aggregation (GraphSAGE) network architecture, which allows training large graphs with mini-batches and generalizing predictions for previously unseen nodes. By cooperating with the Gated Recurrent Unit (GRU) network, the proposed Spatial-Temporal (ST)- GraphSAGE model can capture cross-time relationships between the target and the neighbouring wells and generate promising prediction time series for the target wells, even if they are newly drilled wells.
The data set is based on field data corresponding to 2,240 Montney shale gas wells and consists of formation properties, fracture parameters, production history and operational data. The algorithm aggregates the first-hop information to the target node for each timestep. The Encoder-Decoder (ED) architecture is employed to generate forecasts for the subsequent three-year production rate by using the one-year production history of the wells. The trained model enables the evaluation of production predictions for newly developed wells at any location. We evaluate the model's performance using P10, P50, and P90 of the test dataset's Root Mean Square Error (RMSE). Our method preserves the topological characteristics of wells and generalizes the prediction to unseen nodes while significantly reducing training complexity, making it applicable to larger oil/gas fields. By incorporating information from adjacent wells and integrating spatial-temporal data, our ST-GraphSAGE model outperforms the traditional GRU-ED model and shows enhanced interpretability.
Reference28 articles.
1. Tensorflow: a system for large-scale machine learning;Abadi;In Osdi,2016
2. Analysis of fracture interference-Coupling of flow and geomechanical computations with discrete fracture modeling using MRST;Chen;Journal of Petroleum Science and Engineering,2022
3. Chiang, W.L., Liu, X., Si, S., Li, Y., Bengio, S., & Hsieh, C.J.
2019, July. Cluster-gcn: An efficient algorithm for training deep and large graph convolutional networks. In Proceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining (pp. 257–266). https://doi.org/10.1145/3292500.3330925.
4. Learning phrase representations using RNN encoder-decoder for statistical machine translation;Cho,2014
5. Convolutional neural networks on graphs with fast localized spectral filtering;Defferrard;Advances in neural information processing systems,2016
Cited by
1 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献