Abstract:Horizontal in situ stress prediction plays an important role in petroleum exploration, such as wellbore stability analysis and hydraulic fracturing design. However, owing to the complexity, uncertainty, and heterogeneity of deep formations, a strong non-linear relationship exists between the different logging data, making it difficult to express their spatial correlation using an analytical model and a traditional neural network model. Therefore, based on the logging data of horizontal wells X-1 and X-2 in the Sichuan Basin, a hybrid neural network was proposed to predict horizontal in situ stresses. This hybrid neural network integrated the features of a convolutional neural network (CNN), bidirectional long short-term memory neural network (BiLSTM), and the attention mechanism, and is called the CNN-BiLSTM-attention hybrid neural network. Logging data of the horizontal X-2 well were utilized to verify this model. The results indicate that the CNN-BiLSTM-attention hybrid neural network can accurately predict horizontal in situ stresses. The mean absolute percentage errors of the minimum and maximum horizontal principal stresses were 0.48 and 0.39%, respectively. Additionally, the predictions of FCNN, CNN, BiLSTM, and CNN-BiLSTM-attention were compared to show that the proposed model was much better than the others because the CNN can effectively extract the morphological characteristics of logging curves and the BiLSTM model can better reflect their correlation information. Therefore, the CNN-BiLSTM-attention hybrid neural network model is recommended for predicting horizontal in situ stresses. Article Highlights The mechanical earth model of horizontal well was introduced for shale gas reservoirs. A new CNN-BiLSTM-attention hybrid neural network was proposed for horizontal in situ stress prediction. The prediction accuracy of CNN-BiLSTM-attention hybrid neural network is better than that of FCNN, CNN, and BiLSTM.

Automatic Well Test Interpretation Method for Circular Reservoirs with Changing Wellbore Storage Using One-Dimensional Convolutional Neural Network

Automatic Well Test Interpretation Based on Convolutional Neural Network for a Radial Composite Reservoir

Automatic Reservoir Model Identification Method Based on Convolutional Neural Network

Well test inversion method for fractured horizontal wells based on deep neural network

Automatic prediction of shear wave velocity using convolutional neural networks for different reservoirs in Ordos Basin

A Convolutional Neural Network Approach for Stratigraphic Interface Detection

The Connectivity Evaluation among Wells in Reservoir Utilizing Machine Learning Methods

Forecasting Gas Well Classification Based on a Two-Dimensional Convolutional Neural Network Deep Learning Model

Application of Inter-Well Connectivity Analysis with a Data-Driven Method in the SAGD Development of Heavy Oil Reservoirs

Optimizing seismic-based reservoir property prediction: a synthetic data-driven approach using convolutional neural networks and transfer learning with real data integration

Shale gas geological "sweet spot" parameter prediction method and its application based on convolutional neural network

Reservoir Prediction Based on Closed-Loop CNN and Virtual Well-Logging Labels.

A deep reinforcement learning (DRL) based approach for well-testing interpretation to evaluate reservoir parameters

Horizontal in situ stresses prediction using a CNN-BiLSTM-attention hybrid neural network

Reservoir characterization reimagined: a hybrid neural network approach for direct three-dimensional petrophysical property characterization

Residual Convolutional Neural Network for Lithology Classification: A Case Study of an Iranian Gas Field

Automation distinguishing oil from water bearing reservoirs based on real coding genetic neural network

Porosity Prediction in Tight Sandstone Reservoirs Based on a One–dimensional Convolutional Neural Network–gated Recurrent Unit Model

High-resolution Seismic Impedance Inversion Integrating the Closed-Loop Convolutional Neural Network and Geostatistics: an Application to the Thin Interbedded Reservoir

A method for well log data generation based on a spatio-temporal neural network

Simulating the Behavior of Reservoirs with Convolutional and Recurrent Neural Networks