Aug 2026· Journal of Physics, Conference Series· Vol 3290· 0 citations· 2 references
Physics
Abstract
With the continuous development of drilling technology, accurately predicting mechanical penetration rates is particularly important for improving operational efficiency and reducing costs. Existing methods often struggle to provide reliable predictions when faced with complex geological conditions and variable drilling environments because they primarily rely on traditional models and fail to adequately consider various influencing factors and their nonlinear relationships. To ad-dress these issues, this paper proposes a mechanical penetration rate prediction model based on committee machines. This model effectively captures the variability characteristics of mechanical penetration rates by integrating multiple expert models while employing wavelet filtering methods to denoise the data to enhance data quality. In the application case, this paper collects relevant drilling parameter data based on a vertical well in a specific block. The evaluation of the model shows that it performs excellently in key indicators such as mean square error, coefficient of determination, root mean square error, and mean absolute error, particularly demonstrating a high predictive capability and stability by explaining 97.19% of data variability. The advantage of the constructed model lies in its strong ensemble learning ability, which not only enhances the prediction accuracy of mechanical penetration rates but also helps to deepen the understanding of the dynamic changes in the drilling process, providing effective support for subsequent drilling optimization and resource development.
Against the backdrop of continuously growing global energy demand, the rate of penetration (ROP) serves as a key indicator for measuring drilling efficiency. Accurate ROP prediction is significant for optimizing drilling parameters and reducing project costs. In recent years, machine learning (ML) has been widely applied in ROP prediction research due to its advantages in handling high-dimensional data and nonlinear modeling. This paper systematically reviews and comparatively analyses the research progress of machine learning in ROP prediction. Existing studies are classified into three categories: purely data-driven models, hybrid models integrating physical mechanisms, and intelligent optimisation and decision-support methods. Cross-study comparisons reveal that pure data-driven models generally achieve high prediction accuracy when sufficient data is available. However, multiple studies note their limitations in generalization capability and physical consistency under complex geological conditions. Hybrid models integrating physical mechanisms demonstrate better robustness and interpretability, but their higher computational complexity constrains potential for real-time applications. Intelligent optimization and decision support methods show promise in multi-objective collaborative optimization, though challenges in stability and real-time performance remain. Based on this review, four main research directions are summarized: deep integration of physical constraints and intelligent optimization algorithms, development of decision support systems for real-time drilling, advancement of interdisciplinary hybrid modeling methods, and application of efficient computing and edge intelligence technologies. These directions stem from recurring technical bottlenecks and research trends in recent literature, providing guidance for future research and engineering practice.
Yulin Ma, T. Cao, Zhou Du· Journal of Petroleum Explora...· 0 citations
Reliable forecasting of drilling rate of penetration (ROP) remains a technical challenge due to its complex dependence on several operational, geological, and directional conditions that are highly nonlinear and well-specific. These difficulties are amplified in deviated wells, where traditional empirical relations often fail to capture the combined effects of lithological variability and directional changes. This study develops a machine learning (ML) model for ROP prediction using multi-well field data and evaluates candidate ML algorithms/models not only on numerical accuracy but also on their ability to reproduce reasonable trend responses to changes in key drilling parameters. The objective is to establish a validation approach that integrates statistical performance, cross-well generalization, and sensitivity behavior consistent with drilling mechanics.
Field data from 18 wells were used to develop the model, while two additional wells were reserved for independent validation. A total of 91,140 drilling datasets, comprising 24 parameters/features, were collected. After preprocessing and selective outlier screening, 85,695 records were retained for model training and testing. Feature selection resulted in nine highly influential parameters: True Vertical Depth (TVD), Weight on Bit (WOB), rotational speed (Rotation), mud flow rates (FLOW), mud density, shale content (Shale), Sandstone content (Sandstone), inclination (Inc), and dogleg severity (DLS). Eleven supervised ML algorithms were evaluated, representing instance-based, tree-based, boosting-based, neural-network, and stacking-ensemble model families. Model performance was assessed using feature-importance ratings and statistical metrics, including train/test R2, MSE, RMSE, and MAE, which were used to select the best-performing model. Finally, sensitivity analysis was conducted to evaluate the selected model's ability to reproduce physically meaningful ROP responses.
Model comparison across the training and testing splits demonstrated strong overall predictive performance, with training R2 values ranging from 0.91 to 0.99 and testing R2 values ranging from 0.88 to 0.94. Among the evaluated models, the Gradient Boosting (GB) model provided the best overall performance, with training R2 = 0.9940, testing R2 = 0.9399, RMSE = 36.88, and MAE = 21.52. The GB model also showed strong agreement with the averaged feature-importance ranking and reproduced physically meaningful ROP sensitivity trends for the dominantly influential drilling parameters. SHAP and feature-importance analyses confirmed that TVD, FLOW, WOB, shale content, and rotation were the most influential variables controlling ROP. External validation on two unseen wells further demonstrated strong generalization, with R2 = 0.9694 for Well-19 and R2 = 0.96 for Well-20.
This study presents a physically informed framework for evaluating and selecting data-driven ML models for deviated wells. The approach combines conventional accuracy metrics with feature-importance consistency, reasonable trend prediction in sensitivity analysis, and strong agreement between model predictions and unseen-well data. Through this integrated evaluation, a model is identified that not only reproduces historical data accurately but also yields correct responses under a diverse range of varying input parameters. The proposed methodology establishes a reproducible basis for developing more reliable ROP forecasting tools for complex well trajectories.
Nayem Ahmed, Ramadan Ahmed, V. Soriano et al.· SPE/IADC Asia Pacific Drilli...· 0 citations
This study presents the development and validation of a machine learning model for predicting the residual life of bronze bearing liners used in the universal spindle of a 1680 rolling mill. The model is based on the Random Forest ensemble algorithm and implemented in the RStudio environment using real industrial obtained from 36 bearing replacement events. The input feature set includes operating time, replacement frequency, failure probability density, and replacement count, enabling the identification of nonlinear relationships between technological parameters and component degradation. Model performance was evaluated using mean squared error (MSE) and mean absolute error (MAE), achieving an MSE of 0.125 and an MAE of 0.32 days on the test data set, with cross‐validation confirming model stability (MSE = 0.130 ± 0.012). Sensitivity analysis demonstrated robustness to input data variations typical for industrial environments. Comparative analysis with linear regression showed significantly lower predictive accuracy of conventional statistical approaches. The obtained results confirm the feasibility of applying ensemble machine learning methods for reliable residual life prediction under small‐sample industrial conditions. The proposed approach enables data‐driven maintenance planning, reduces the risk of unplanned downtime, and supports the implementation of intelligent condition monitoring systems aligned with Industry 4.0 principles.
Accurate estimation of coal pillar strength is essential for ensuring safety and operational efficiency in underground mining. Current assessment methods often face limitations in addressing time-dependent failure mechanisms, geological discontinuities, and dynamic loading conditions. This paper identifies future research directions aimed at enhancing the reliability of pillar strength evaluations. One important focus is the development of advanced numerical models that account for time-dependent behaviours such as creep and fatigue. Incorporating multi-scale modelling techniques, which connect micro-scale material responses to macro-scale structural performance, may lead to more precise predictions. The integration of real-time monitoring systems measuring stress, deformation, and environmental factors into predictive models can enable continuous assessment and proactive management. When combined with machine learning algorithms analysing large datasets and recognizing patterns, these approaches can optimize predictive accuracy and maintenance strategies. Improved geological characterization using advanced mapping technologies, such as geophysical surveys, is critical to account for weak planes, fractures, and faults. Additionally, long-term field monitoring and laboratory experiments are necessary to validate and refine models. Establishing standardized regulatory guidelines will help ensure consistency, particularly in challenging mining environments. Collaboration between academia and industry is essential to drive innovation and develop robust, reliable methods for coal pillar strength estimation.
Abhishek Kumar Singh, S. Ram· Journal of Sustainable Minin...· 0 citations
This study proposes a data-driven surrogate modeling framework for predicting
solidification time and mold thermal stress during low-pressure die casting
(LPDC) of aluminum alloy wheels. The methodology employed an optimal Latin
hypercube design (OLHD) to sample key parameters including cooling channel
geometry and process conditions. A sequential simulation methodology combining
ProCAST and Abaqus was implemented to generate a comprehensive dataset of
solidification times and thermal stress distributions. Based on this dataset,
surrogate models were developed using Support Vector Regression, Kriging, and
Polynomial Response Surface Methodology, with their hyperparameters
automatically tuned through Bayesian Optimization (BO). The optimized models
were rigorously evaluated using four statistical metrics: Coefficient of
Determination (R2), Mean Squared Error (MSE), Mean Absolute Error (MAE), and Root
Mean Squared Error (RMSE). The evaluation results show that the BO–SVR model
demonstrated superior prediction accuracy for both output responses and
exhibited exceptional nonlinear fitting capability. This work establishes an
effective modeling approach for simultaneous quality and efficiency optimization
in wheel manufacturing.
Fan Fuhao, Yunlang Zhan, Zhenfei Zhan et al.· SAE technical paper series· 0 citations
Precisely predicting the forces acting on the disc cutters employed in tunnel boring machines (TBMs)—namely, the normal force (FN) and rolling force (FR)—is critical for optimizing their excavation performance and service life. Existing models for cutter performance prediction are primarily based on static or average penetration (P), rendering them inadequate for operating conditions where the penetration varies during excavation. Consequently, these models have limited accuracy and engineering applicability under complex, heterogeneous geological conditions. To address this challenge, this paper proposes a novel predictive model capable of adapting to variable penetration to more accurately predict the disc-cutter performance. Based on indoor linear cutting rock test data from 21 distinct rock samples and combined with ABAQUS numerical simulations, the coupled mechanisms between TBM disc-cutter cutting parameters were systematically analyzed. Subsequently, a modified dual-force prediction model for FN and FR was developed, incorporating a dimensionless correction factor κ to enhance the accuracy of the ratio FR/FN. Validation results showed that compared with eight conventional predictive models, the proposed framework achieved a 44.31% reduction in the mean relative error under low-to-medium-strength rock conditions, demonstrating its superior reliability in predicting the strength within these specific geological regimes. This research provides theoretical and practical insights for accurately predicting disc-cutter forces under complex geological conditions, thus improving excavation efficiency and durability optimization of TBMs.
Yizhe Peng, Hailong Zhang, Long Cheng et al.· Canadian geotechnical journa...· 0 citations