Ye Liu

Accurate real-time interpretation of logging-while-drilling (LWD) gamma-ray (GR) logs is important for formation evaluation and drilling decisions, but behind-bit sensing delays formation information. This study evaluates a data-local federated learning framework for fixed-step depth-ahead GR prediction across three oilfields with different feature sets, data volumes, and native measured-depth sampling. The main benchmark uses 300 historical samples to predict the next 50 GR samples with preprocessing fitted only on the corresponding training well. Under uniform client weighting, FedAvg reduced mean MAE from 7.13 to 4.44 gAPI in Volve, from 2.23 to 1.49 gAPI in BH, and from 5.76 to 2.79 gAPI in XJ relative to local training, while substantially reducing across-seed dispersion; the paired mean differences did not reach p  < 0.05. Four strictly causal baselines were weaker than FedAvg on the evaluated holdouts. Additional diagnostics show field-dependent effects of sample-size weighting and gap-aware windowing, while a matched-volume source-composition test does not support attributing the observed gains to cross-field source diversity itself. The XJ 300/50 case contains only five highly overlapping test windows covering 54 unique target GR points. Local inference latency is millisecond-scale, whereas federated model updating is a separate minutes-scale process whose communication tolerance depends on bandwidth and model-refresh interval. Overall, the evidence supports data-local collaborative training for fixed-step depth-ahead GR prediction, with important limitations from irregular sampling, severe client imbalance, domain heterogeneity, and small-sample evaluation.

Particle Swarm Optimization (PSO) is a widely used heuristic algorithm valued for its simplicity and robustness in solving diverse optimization problems. However, its high computational cost often limits large-scale applications. With the rapid development of parallel computing and Graphics Processing Units (GPUs), researchers have increasingly leveraged these technologies to enhance PSO efficiency. This paper introduces a High-Efficiency PSO (HEPSO) algorithm designed for GPU-based architectures. HEPSO improves computational performance through two key strategies: (1) transferring data initialization from the CPU to the GPU to reduce I/O overhead caused by repeated data migration, and (2) incorporating a self-adaptive thread management mechanism to enhance execution efficiency. Experiments conducted on nine benchmark optimization functions demonstrate that HEPSO achieves more than a sixfold speedup compared to conventional GPU-PSO. Moreover, in terms of convergence time, HEPSO requires only about one-third of the runtime in most cases.

Summary: In this study, we investigate the problem of wellbore trajectory optimization under subsurface constraints, including the avoidance of pre-existing wells. Traditional optimization techniques, such as metaheuristic and gradient-based methods, have limited capability in directly handling collision avoidance and typically require iterative recalculations for each new design, resulting in increased computational effort. Reinforcement learning (RL) provides an alternative approach by enabling adaptive policy learning through interaction with the environment. We propose a deep recurrent Q-network (DRQN)-based framework, in which engineering constraints are incorporated into the reward function to support multisegment trajectory planning. The framework accounts for factors such as collision avoidance, dogleg severity (DLS) limits, and target-entry parameters, enabling multiobjective design without explicitly defined optimization equations or repeated replanning. Convolutional and long short-term memory (LSTM) layers are used to support learning across trajectory segments. In offline training, the DRQN model demonstrated more stable convergence than the standard deep Q-network (DQN), which showed inconsistent performance under the same conditions. In online testing scenarios, DQN failed to converge and was unable to generate valid trajectories, whereas DRQN successfully produced feasible solutions. When compared with the particle swarm optimization (PSO) method, DRQN yielded similar trajectory results but required substantially less computational time. These results suggest that the proposed DRQN-based framework is applicable to adaptive and constraint-aware trajectory planning tasks in well construction design.

As the energy industry increasingly turns to unconventional shale reservoirs to meet global demands, the development of advanced predictive models for shale oil production has become imperative. The inherent complexity of shale formations, coupled with the intricacies of hydraulic fracturing, poses significant challenges to efficient resource extraction. Our study leverages a substantial data set from the Ordos Basin to develop an advanced predictive model, integrating 18 parameters that blend static petrophysical attributes and dynamic factors, including hydraulic fracturing parameters and real-time pump pressure data. This holistic approach enables our self-attention (SA) model to accurately forecast future production rates by processing the complex interplay between reservoir characteristics and operational inputs. In testing across three wells, the model achieved average accuracies of 99.28% for daily oil production (DOP) and 99.25% for daily liquid production (DLP) over 20 days, surpassing traditional long short-term memory (LSTM) and gated recurrent unit (GRU) models, proving its efficacy in fractured well production forecasting. Furthermore, using the initial 30 days of production data as input, the model demonstrated its capability to predict DOP and DLP over a one-year period, achieving prediction accuracies of 96.2% for DOP and 99.6% for DLP rates. Our model’s profound implications for the shale industry include establishing a quantifiable link between key factors and production forecasts, guiding the optimization of controllable aspects, and serving as a decision-support tool for more efficient and cost-effective oil recovery.

This study presents the development and validation of a robust semi-supervised learning framework specifically designed for the automated segmentation and classification of sandstone thin section images from the Yanchang Formation in the Ordos Basin. Traditional geological image analysis methods encounter significant challenges due to the labor-intensive and error-prone nature of manual labeling, compounded by the diversity and complexity of rock thin sections. Our approach addresses these challenges by integrating the GL-SLIC algorithm, which combines Gabor filters and Local Binary Patterns for effective superpixel segmentation, laying the groundwork for advanced component identification. The primary innovation of this research is the semi-supervised learning model that utilizes a limited set of manually labeled samples to generate high-confidence pseudo labels, thereby significantly expanding the training dataset. This methodology effectively tackles the critical challenge of insufficient labeled data in geological image analysis, enhancing the model’s generalization capability from minimal initial input. Our framework improves segmentation accuracy by closely aligning superpixels with the intricate boundaries of mineral grains and pores. Additionally, it achieves substantial improvements in classification accuracy across various rock types, reaching up to 96.3% in testing scenarios. This semi-supervised approach represents a significant advancement in computational geology, providing a scalable and efficient solution for detailed petrographic analysis. It not only enhances the accuracy and efficiency of geological interpretations but also supports broader hydrocarbon exploration efforts.