To address the limitations of traditional methods in modeling complex nonlinear relationships in horizontal in situ stress prediction for shale reservoirs, this study proposes an integrated framework that combines well logging interpretation with machine learning to accurately predict horizontal in situ stress in shale reservoirs. Based on the logging data from five wells in the Luzhou Block of the Sichuan Basin (16,000 samples), Recursive Feature Elimination (RF-RFE) was used to identify nine key factors, including Stoneley wave slowness and caliper, from 30 feature parameters. Bayesian optimization was employed to fine-tune the hyperparameters of the XGBoost model globally. Results indicate that the XGBoost model performs optimally in predicting maximum horizontal principal stress (SHmax) and minimum horizontal principal stress (SHmin). It achieves R² values of 0.978 and 0.959, respectively, on the test set. The error metrics (MAE, MSE, RMSE) of the XGBoost model are significantly lower than those of SVM and Random Forest, demonstrating its precise capture of the nonlinear relationships between logging parameters and in situ stress. This framework enhances the model’s adaptability to complex geological conditions through multi-well data training and eliminating redundant features, providing a reliable tool for hydraulic fracturing design and wellbore stability assessment in shale gas development.

Shale gas is a critical energy resource, and estimating its ultimate recoverable reserves (EUR) is a key indicator for evaluating the development potential and effectiveness of gas wells. To address the challenges in accurately predicting shale gas EUR, this study analyzed production data from 200 wells in the CN block. Sixteen potential factors influencing EUR were considered, and key geological, engineering, and production factors were identified using Spearman correlation analysis and mutual information methods to exclude highly linearly correlated variables. An attention mechanism was introduced to weight input features prior to model training, enhancing the interpretability of feature contributions. The hyperparameters of the model were optimized using the Rabbit Optimization Algorithm (ROA), and 10-fold cross-validation was employed to improve the stability and reliability of model evaluation, mitigating overfitting and bias. The performance of four machine learning models was compared, and the optimal model was selected. The results indicated that the ROA-CatBoost-AM model exhibited superior performance in both fitting accuracy and prediction effectiveness. This model was subsequently applied for EUR prediction and for identifying the primary factors controlling productivity, providing effective guidance for development practices. The dominant factors and production forecasts determined by the model offer valuable references for optimizing block development strategies.

After 5 different plans of development, the Weiyuan block is struggling to maintain stable production and urgently needs a new method to improve the recovery rate. Practice at home and abroad shows that three-dimensional development can effectively improve the degree of vertical shale gas utilization. Therefore, this paper first analyzes the distribution of reservoirs in the Weiyuan block and calculates the distribution of remaining reserves through the methods of fine description of gas reservoirs and establishment of a three-dimensional geological model. The results show that the recovery degree of the platform in the main production area of Weiyuan is 35~60%, and the average remaining reserves of the wells are (1.2~3) × 10⁸ m³, and there is three-dimensional development potential in the local area. Secondly, through the establishment of a fracture and rock mechanics model, the target layer system for three-dimensional development of the Weiyuan block is preliminary selected from five aspects, namely, remaining reserves, reservoir characteristics, fracture density, mechanical characteristics, and the comparison of the height of the fracture network. Finally, in order to check the correctness of the target layer for three-dimensional development and evaluate the effect of the three-dimensional development, four scenarios for the lower branch of the H8 platform are designed and numerically simulated using Petrel RE software. The results of the numerical simulation show that Option 2 is the most effective, which can increase the production of the platform by 1.76 × 10⁸ m³, and the recovery rate grows from 16.4% to 31.0%, with an overall increase of 14.6%. The relevant understanding can provide a reference for the three-dimensional development strategy of the Weiyuan shale gas block.

The primary horizontal well network in the Weiyuan block is vertically deployed in the Long1₁¹ sub-layer, and it is difficult to fully utilize the 43.9–53.6 m thick Wufeng formation-Long1₁ sub-member. To enhance the reserves utilization degree, this study proposes a method for evaluating the vertical utilized degree of shale gas reserves based on the calibrated estimated ultimate recovery (EUR) iterative utilized thickness and a quantitative evaluation method of shale gas reservoir productivity based on reservoir classification. The reserves utilization degree and vertical utilization thickness of the primary well network in the Weiyuan block are evaluated, the distribution characteristics of remaining reserves are clarified, and the three-dimensional development potential is analyzed. The results show that: (1) The horizontal well hydraulic fracture network of the primary well network in the production area mainly utilized the Wufeng formation-Long1₁³ sub-layer, and the utilized thickness is between 3.49 and 19.40 m, with an average of 13.79 m. Vertically, only 30% of the reservoir is utilized. (2) The well-control reserve recovery of the primary well network in the production area is 75%, the average EUR of half-branch wells is 0.87 × 10⁸ m³, the average well-control geological reserves of half-branch wells are 1.16 × 10⁸ m³, and the total well-controlled geological reserves in the whole area are 906.7 × 10⁸ m³. (3) The total remaining geological reserves are 1690.5 × 10⁸ m³, the average remaining reserves abundance is 4.31 × 10⁸ m³/km², the average thickness of unutilized reserves is 26.16 m, and the high-value areas are all located in the eastern part of the production area. (4) Under the condition of the same thickness, the ratio of gas production capacity of Type I and Type II reservoirs is 3.7:1. (5) According to the favorable area screening principle, the favorable area is screened out to be 116.7 km², and the total geological reserves are 700.7 × 10⁸ m³. This study provides a new development idea for improving shale gas production in the Weiyuan block.