雪弗龙公司

River bifurcations control water and sediment distribution in fluvial systems, but the physical mechanisms governing sediment partitioning remain poorly validated in natural rivers. We present a comprehensive field test of nodal point relations using radio frequency identification (RFID) tracking of 376 gravel clasts through a meandering river bifurcation in Montana, USA. The study bifurcation features key characteristics for testing existing theory: upstream channel curvature, a 40-cm bed ramp at the shallower bifurcate entrance, and unequal branch geometry. During the 2017 flood season, we recovered 202 (out of 376) clasts transported through the bifurcation, which divided nearly equally between bifurcate arms. Statistical analysis reveals that sediment partitioning is primarily controlled by upstream transverse position and longitudinal deployment distance, while grain size, shape, and mass show no significant influence. Calibrated Delft3D modeling combined with theoretical nodal point relations demonstrates that for curved bifurcations, helical flow and non-uniform incoming sediment push sediment in opposite directions. The nodal point relation that accounts for non-uniform incoming sediment distribution due to curvature best predicts the observed sediment delivery to the northern arm. Our results provide critical field validation showing that curvature-induced sorting begins well upstream of the bifurcation node, and the two curvature effects (helical flow deflection and non-uniform incoming sediment distribution) must both be included for accurate predictions. These findings advance our understanding of bifurcation mechanics and provide guidance for improving theoretical models and river restoration designs.

As drilling activity intensifies across the oil and gas industry, there is a growing need to enhance the efficiency, reliability, and safety of drilling operations. Traditional methods are often constrained by high costs, non-productive time (NPT), and operational risks. These challenges are exacerbated by inconsistencies in real-time decision-making and uncertainties in subsurface rock properties. Existing uncertainty modeling approaches, such as analog models, geostatistical simulations, and Bayesian expert systems, offer limited scope and rely heavily on operator expertise and simplistic forward modeling techniques. This paper introduces DRUID (deep reinforcement learning used to improve drilling), a real-time optimal control agent that automates and optimizes drilling parameters using deep reinforcement learning. DRUID employs proximal policy optimization (PPO) to model subsurface conditions and dynamically recommend key control parameters, including rate of penetration (ROP), weight on bit (WOB), rotations per minute (RPM), and bit inclination. By integrating expert domain knowledge with static reservoir data, DRUID supports geosteerers in making informed decisions while accounting for geological variability. A comparative study with a value-function approximation method, deep Q-network (DQN), demonstrates that the PPO-based agent achieves superior drilling performance, benefiting from trust-region updates that ensure stable training and controlled exploration across the state–action space. To further understand the agent’s learning behavior, a reward sensitivity analysis was conducted to evaluate the impact of reward shaping on convergence and final policy quality. The analysis revealed that the structure and scaling of intermediate and terminal rewards significantly influence the agent’s ability to balance exploration and exploitation, underscoring the importance of carefully designed reward functions in reinforcement learning-based control systems. The effectiveness of DRUID is further validated through a real-world case study using the Volve field dataset. Applied to the F-12 well, DRUID achieved a simulated rotary drilling time of 38.7 h, closely approximating the robustly optimized benchmark of 36.7 h sand significantly shorter than the actual field drilling time of 139.2 h. These results highlight DRUID’s ability to generalize from synthetic training environments to real-world data, capturing the complex relationships between rock properties and drilling dynamics. By combining reinforcement learning with expert-informed modeling and real-time optimization, DRUID represents a significant advancement in intelligent drilling automation. Its deployment at rig sites and drilling support centers offers a scalable, data-driven solution for reducing NPT, improving operational consistency, and enabling end-to-end automation across both rotary and non-rotary drilling phases.

We describe a method of full waveform inversion (FWI) for simultaneously resolving velocity and anisotropy using multiple independent monochromatic frequency-space solutions to the wave equation. Operating with more than one monochromatic solution combines benefits of both frequency and time domain approaches without a substantial increase in computational complexity. By computing many frequency-space solutions during a single time-space propagation, we address scaling issues related to large sparse matrices for three spatial dimensions in the frequency domain. We use a domain-specific language (DSL) to symbolically represent the mathematics of seismic modeling and inversion and automatically generate high performance kernels that run on any modern CPU or GPU hardware. In this abstract, we first describe frequency domain FWI and compare time domain and monochromatic approaches. Then, we introduce our polychromatic approach and describe multiparameter inversion for velocity and anisotropy, including our self-adjoint pseudo-acoustic tilted transversely isotropic (TTI) propagators. We highlight the advantages of using a DSL to symbolically abstract the physics of modeling and inversion and decouple it from the computer science implementation. Finally, we summarize our experience from practical field application.

This paper presents a comparison of dark fiber single-component surface distributed acoustic sensing (S-DAS) data with multi-component ocean bottom node (OBN) data acquired simultaneously in deep-water Gulf of America. S-DAS shows broader bandwidth and improved low-frequency response at significantly denser spatial sampling but exhibits a unique "butterfly-wing" amplitude pattern likely due to near-surface scattering. The S-DAS data were appropriately processed and prepared for imaging. The S-DAS PP image is of high quality in the shallow section but declined with depth, especially subsalt.