Abstract:Abstract Core reservoir management studies involving sensitivity, history matching, predictive forecast, optimization and uncertainty tasks require hundreds of numerical simulations. To enhance reservoir descriptions; more data acquisition, computational power and complex models are engineered. This creates an unstainable time consuming, computationally expensive cyclical process to which the industry could not yet realize the true potential of such full-field reservoir models [1]. In line with the upcoming fourth paradigm shift involving data driven science, this study aims to use data mining and artificial intelligence to investigate the conceptual framework, development and feasibility of data-driven approach in Artificial Neural Network based Grid-block Surrogate Reservoir Model (GSRM) Smart Proxy. The Smart Proxy model is applied for unconventional shale gas reservoirs with dynamic property response prediction output of grid-block pressure and gas saturation. The GSRM Smart Proxy was developed based on data mining and artificial intelligence technique using VBA programming language in Excel and MATLAB Neural Network Toolbox. Validation of the Smart Proxy model is conducted against commercial simulator unconventional benchmark model in Computer Modelling Group (CMG). Data gathering of the benchmark numerical shale gas model was first conducted followed by the sensitivity analysis of parameters. Most sensitive parameters were identified and brought forward to the design of experiment task to generate a total of 58 datasets covering the solution space. The datasets were run, and simulator output results of grid gas saturation were extracted and deconvoluted using Excel VBA program. The scenarios database is then utilized in MATLAB Neural Network Toolbox to construct GSRM Smart Proxy through data science tasks such as feature engineering, neural net architecture design, model training, validation and testing. Results of the developed GSRM Smart Proxy test predictions of grid-block pressure and grid-block gas saturation neural network models were compared to have a sufficiently acceptable match accuracy and validated to the benchmark numerical reservoir model solution with R2of 0.947 and R2of 0.934 respectively, also with MSE of 2.52E-03 and MSE of 2.55E-03 respectively for the test sets. GSRM outputs are found to be obtained with minimal computation power and runtime (seconds).

A data-driven smart proxy model for a comprehensive reservoir simulation

Data-driven Simulation of Reservoir Dynamics Based on Erroneous Inputs from Fast Supply Streamline Data

Artificial Neural Network Surrogate Reservoir Proxy Modeling for Generating Pressure and Gas Saturation at the Grid–Block Level in Shale Gas Reservoirs

Data-Driven Proxy Modeling of Water Front Propagation in Porous Media

Accelerating Physics-Based Simulations Using Neural Network Proxies: An Application in Oil Reservoir Modeling

Physics-Aware Deep-Learning-Based Proxy Reservoir Simulation Model Equipped With State and Well Output Prediction

Accelerating Physics-Based Simulations Using End-to-End Neural Network Proxies: An Application in Oil Reservoir Modeling

Neural Operator-Based Proxy for Reservoir Simulations Considering Varying Well Settings, Locations, and Permeability Fields

The prediction of reservoir production based proxy model considering spatial data and vector data

A deep learning-based convolutional spatiotemporal network proxy model for reservoir production prediction

Structure of an Artificial-Intelligence-assisted Reservoir Characterization and Field Development Protocol

Simulating Dynamics of Pressure and Fluid Saturation at Grid-Scale by a Deep Learning-Based Surrogate Reservoir Modeling Based on a Fast-Supply Hybrid Database and Developing Preliminary Insights for Future Gas Hydrate Exploitations in China

Novel Self-Adaptive Shale Gas Production Proxy Model and Its Practical Application

Applying proxy-modeling tools to estimate the effective injection factor for an oil field

The Development of Artificial-neural-network-based Universal Proxies to Study Steam Assisted Gravity Drainage (SAGD) and Cyclic Steam Stimulation (CSS) Processes

A Convolutional Neural Network-Based Proxy Model for Field Production Prediction and History Matching

A Physics-Based Data-Driven Numerical Model For Reservoir History Matching And Prediction With A Field Application

Artificial Intelligence Assisted Hydraulic Fracturing Design in Shale Gas Reservoir

A Large-Scale Comparison of Artificial Intelligence and Data Mining (AI&DM) Techniques in Simulating Reservoir Releases over the Upper Colorado Region

A Surrogate Model for Numerical Reservoir Simulation of CO2 Flooding and Storage Based on Deep Learning

Optimization of Well Control During Gas Flooding Using the Deep-LSTM-Based Proxy Model: A Case Study in the Baoshaceng Reservoir, Tarim, China