The accurate identification of drilling conditions is pivotal for optimizing drilling operations. Detailed condition identification enables refined management of the drilling process, facilitates the detection of non-productive time (NPT), and offers actionable insights to improve drilling performance. In this study, a novel predictive model is presented that integrates the Kepler optimization algorithm (KOA) with random forest (RF) for detailed drilling condition identification. This method enhances the granularity of the operation differentiation and improves the classification accuracy. A systematic comparison of various classification algorithms reveals the superior performance of the RF algorithm in accurately identifying drilling conditions. The incorporation of the KOA allows for the refinement of the hyperparameter randomness and for iterative optimization of the RF model’s initial weights and thresholds, leading to optimal classification results. Case studies demonstrate that three predictive models were trained and evaluated. Validation results show that the KOA-RF model achieves an accuracy of 95.65%. To demonstrate the practical applicability of this methodology in enhancing rig efficiency. The KOA-RF algorithm proposed in this study for detailed drilling condition identification can be programmed into software to enable automated, real-time, and intelligent identification of drilling conditions on-site. Key performance indicator (KPI) analysis can be performed based on the identification results, thereby identifying the sources of NPT, optimizing the drilling process, and improving drilling efficiency. In this study, the KOA-RF model achieves detailed drilling condition identification with high accuracy, providing both theoretical contributions and practical value for real-time drilling optimization.

Aiming at the engineering challenges of low Rate of Penetration (ROP), high rock-breaking energy consumption, and the difficulty in realizing collaborative optimization of efficiency improvement and consumption reduction through traditional drilling parameter regulation in deep interbedded sandstone and mudstone formations of the Southwest Oil and Gas Field, a study on multi-objective optimization of drilling parameters is carried out. First, the traditional Teale Mechanical Specific Energy (MSE) model is modified, with the introduction of the effective energy utilization coefficient of the drill bit and the downhole torque calculation method, to establish an improved MSE model that fits the actual field drilling conditions. Second, a hybrid TCN-LSTM model fused with an additive attention mechanism is constructed to realize high-precision dynamic prediction of ROP. Finally, taking the on-site real-time controllable Weight on Bit (WOB) and rotary speed as decision variables, the synergy–conflict boundary between the dual objectives of ROP maximization and MSE minimization is clarified, a dual-objective coupled optimization model is established, and the genetic algorithm is used to complete the global optimization. Case verification is carried out based on the actual drilling data of 12 deep wells in the study block, and the results show that the Coefficient of Determination R2 of the established ROP prediction model on the test set reaches 0.91, and the prediction accuracy is significantly better than that of BP, CNN and single LSTM models; after optimization, the ROP of the target well interval is increased by 13.1% compared with the on-site actual drilling value, and the improved MSE is reduced by 23.5%, which simultaneously realizes drilling efficiency improvement and rock-breaking energy saving within the safe operation boundary. The stability of the model and the effectiveness of each module are verified through five-fold inter-well cross-validation and module ablation experiments. The research results can provide a theoretical basis and technical support for the precise regulation of drilling parameters in deep formations of the target block.

Deep coalbed methane reservoirs exhibit strong heterogeneity and complex stress environments, making accurate prediction of casing loads during multi-stage hydraulic fracturing challenging. This study developed a coupled prediction framework integrating three-dimensional geomechanical modeling, modified induced stress calculation, and numerical simulation. The geomechanical model was constructed using logging, drilling, rock mechanics, and in situ stress data and validated against fracture monitoring results. The simulated fracture half-length and stimulated area differed from the monitored values by less than 10% and 8%, respectively, with a spatial matching degree exceeding 92%. A modified analytical model was then established by introducing correction coefficients for fracture net pressure and stress propagation. Among the results obtained using five parameter inversion methods, the sparrow search algorithm achieved the highest fitting accuracy, with an (R²) of 0.9371 and an RMSE of 0.3761, yielding (A = 0.7330) and (B = 0.9238). These coefficients indicate an approximately 26.7% reduction in effective net pressure and enhanced attenuation of induced stress in heterogeneous, cleat-developed coal seams. Furthermore, a multi-parameter casing stress model was developed by coupling treatment scale, injection rate, fracture spacing, and stage number. Sensitivity analysis showed that the number of fracturing stages and injection rate were the dominant factors, followed by fracture spacing and treatment scale. The proposed framework quantitatively characterizes casing stress evolution and facilitates casing load assessment under different multi-stage fracturing conditions.