Abstract:Understanding water saturation levels in tight gas carbonate reservoirs is vital for optimizing hydrocarbon production and mitigating challenges such as reduced permeability due to water saturation (Sw) and pore throat blockages, given its critical role in managing capillary pressure in water drive mechanisms reservoirs. Traditional sediment characterization methods such as core analysis, are often costly, invasive, and lack comprehensive spatial information. In recent years, several classical machine learning models have been developed to address these shortcomings. Traditional machine learning methods utilized in reservoir characterization encounter various challenges, including the ability to capture intricate relationships, potential overfitting, and handling extensive, multi-dimensional datasets. Moreover, these methods often face difficulties in dealing with temporal dependencies and subtle patterns within geological formations, particularly evident in heterogeneous carbonate reservoirs. Consequently, despite technological advancements, enhancing the reliability, interpretability, and applicability of predictive models remains imperative for effectively characterizing tight gas carbonate reservoirs. This study employs a novel data-driven strategy to prediction of water saturation in tight gas reservoir powered by three recurrent neural network type deep/shallow learning algorithms—Gated Recurrent Unit (GRU), Recurrent Neural Networks (RNN), Long Short-Term Memory (LSTM), Support Vector Machine (SVM), K-nearest neighbor (KNN) and Decision tree (DT)—customized to accurately forecast sequential sedimentary structure data. These models, optimized using Adam's optimizer algorithm, demonstrated impressive performance in predicting water saturation levels using conventional petrophysical data. Particularly, the GRU model stood out, achieving remarkable accuracy (an R-squared value of 0.9973) with minimal errors (RMSE of 0.0198) compared to LSTM, RNN, SVM, KNN and, DT algorithms, thus showcasing its proficiency in processing extensive datasets and effectively identifying patterns. By achieving unprecedented accuracy levels, this study not only enhances the understanding of sediment properties and fluid saturation dynamics but also offers practical implications for reservoir management and hydrocarbon exploration in complex geological settings. These insights pave the way for more reliable and efficient decision-making processes, thereby advancing the forefront of reservoir engineering and petroleum geoscience.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to accurately predict water saturation (\(SW\)) in tight - gas carbonate reservoirs in the Bohai Bay Basin. Specifically, traditional methods such as core analysis, although reliable, are costly, invasive and lack comprehensive spatial information. In addition, traditional machine - learning methods face many challenges when dealing with complex geological data, including difficulty in capturing subtle relationships, possible over - fitting and limited ability to handle multi - dimensional data sets. These problems are particularly evident when dealing with heterogeneous carbonate reservoirs. To overcome these challenges, this study adopts a data - driven strategy based on deep - learning and shallow - learning algorithms to predict water saturation. Three types of deep - learning algorithms of recurrent neural networks - Gated Recurrent Unit (GRU), Recurrent Neural Network (RNN) and Long - Short - Term Memory Network (LSTM), as well as three classic machine - learning algorithms - Support Vector Machine (SVM), K - Nearest Neighbor (KNN) and Decision Tree (DT) are used in the study. These models are optimized by the Adam optimizer and use conventional geophysical data to predict water - saturation levels. In particular, the GRU model has shown remarkable accuracy (the \(R^2\) value is 0.9973) and low error (the Root Mean Square Error RMSE is 0.0198). Compared with other algorithms such as LSTM, RNN, SVM, KNN and DT, it shows its advantages in handling large - scale data sets and effectively identifying patterns. By achieving an unprecedented level of accuracy, this study not only enhances the understanding of sediment characteristics and fluid - saturation dynamics, but also provides practical application value for reservoir management and oil - gas exploration in complex geological environments. These insights contribute to a more reliable and efficient decision - making process, thus promoting the development of reservoir engineering and petroleum geoscience.

Deep learning algorithm-enabled sediment characterization techniques to determination of water saturation for tight gas carbonate reservoirs in Bohai Bay Basin, China

Deep learning for predicting porosity in ultra-deep fractured vuggy reservoirs from the Shunbei oilfield in Tarim Basin, China

Method and practice of deep favorable shale reservoirs prediction based on machine learning

Deep carbonate reservoir characterisation using multi-seismic attributes via machine learning with physical constraints

Integrated Carbonate Lithofacies Modeling Based on the Deep Learning and Seismic Inversion and Its Application

Deep carbonate reservoir characterization using multi seismic attributes:A comparison of unsupervised machine learning approaches

Data Driven Modeling and Prediction for Reservoir Characterization Using Seismic Attribute Analyses and Big Data Analytics

Quantitative characterization of shale gas reservoir properties based on BiLSTM with attention mechanism

A prediction method for favorable hydrocarbon accumulation areas in carbonate oil reservoirs based on machine learning and driven by formation water chemical characteristic data

The State-of-the-Art Techniques of Hydrocarbon Detection and Its Application in Ultra-Deep Carbonate Reservoir Characterization in the Sichuan Basin, China

Approaches of Combining Machine Learning with NMR-Based Pore Structure Characterization for Reservoir Evaluation

Data-driven lithofacies prediction in complex tight sandstone reservoirs: a supervised workflow integrating clustering and classification models

New Deep Learning Network (Deep Residual Shrinkage Network) Is Applied for Lithology Identification To Search for the Reservoir of CO 2 Geological Storage

High‐resolution reservoir prediction method based on data‐driven and model‐based approaches

A Carbonate Reservoir Prediction Method Based on Deep Learning and Multiparameter Joint Inversion

Integrating deep learning and logging data analytics for lithofacies classification and 3D modeling of tight sandstone reservoirs

A new method of predicting the saturation pressure of oil reservoir and its application

Prediction of reservoir saturation field in high water cut stage by bore-ground electromagnetic method based on machine learning

Quantitative Analysis of the Main Controlling Factors of Oil Saturation Variation

Identification of reservoir types in deep carbonates based on mixed-kernel machine learning using geophysical logging data

3D seismic intelligent prediction of fault-controlled fractured-vuggy reservoirs in carbonate reservoirs based on deep learning method