This study investigates wellbore stability in ultra-deep extended-reach wells (ERWs) in the East China Sea, where perforated pipes (a type of screen completion) are commonly used to support wellbore walls and prevent collapse. Cost constraints sometimes lead to the omission of this support, yet significant wellbore collapse is rarely observed. The instability is primarily attributed to variations in production pressure differences. A predictive model for critical pressure difference was developed based on immersion experiments and single-triaxial rock mechanics tests. The results from immersion tests revealed that, in water-bearing strata, the critical pressure difference decreased significantly, drop-ping by 20.07% after two days of rock core immersion and by 28.35% after seven days. Key factors influencing this difference, such as well inclination, rock cohesion, internal friction angle, Poisson’s ratio, and Biot coefficient, were identified. As production continues, pore pressure depletion reduces this difference, particularly when pore pressure falls below 23.5 MPa, leading to wellbore instability. On-site validation in three ultra-deep ERWs showed that the model’s predictions aligned well with actual conditions, with a confidence interval analysis further validating the model’s accuracy. The proposed model provides valuable guidance for future ultra-deep well development in the East China Sea.

As oil and gas exploration advances, tubular strings in extended-reach wells are becoming increasingly susceptible to buckling, self-locking, and other failure modes due to their relatively low weight and limited structural stiffness. Existing studies have mainly focused on running feasibility and associated mechanical responses, whereas systematic investigations of the ultimate axial load capacity of deployed tubular strings remain limited. To address this research gap, the present study investigates the ultimate axial load capacity of tubular strings in deepwater extended-reach wells, explicitly accounting for the constitutive behavior of the tubular material. Numerical simulations were performed under both tensile and compressive loading conditions for different well trajectories, dogleg severities, material grades, and wellbore diameters, and the effects of these factors on the ultimate axial load capacity were systematically evaluated. The numerical predictions at the initial yielding stage were compared with available analytical solutions, and good agreement was obtained, indicating that the proposed model is reliable for initial yield analysis. The results show that well trajectory, dogleg severity, material grade, and wellbore diameter significantly influence the ultimate axial load capacity of tubular strings. The findings of this study provide useful guidance for structural integrity assessment, design optimization, and operating-parameter selection for downhole tubular strings in deepwater and ultra-deepwater oil and gas developments, thereby supporting the safe and efficient completion of extended-reach wells.

Extended-reach wells contain long, highly inclined intervals in which drill-string/wellbore contact and cumulative friction evolve continuously with depth, making torque prediction strongly nonstationary. This study develops a multi-source torque-prediction framework using 5977 field samples from a deepwater extended-reach well in the East China Sea. To eliminate the optimistic bias caused by randomly mixing adjacent depth samples, all records were ordered by measured depth; preprocessing and feature screening were fitted only on historical data; hyperparameters were selected with forward-chaining TimeSeriesSplit; and model performance was assessed using both a fixed deep-depth stress test and expanding-window rolling-origin prediction. Training-only correlation screening and variance inflation factor analysis reduced multicollinearity before comparing ExtraTrees, XGBoost, LightGBM, BPNN, and time-aware ensemble regression. Static extrapolation to the deepest 15% of the well revealed severe distribution shift and poor long-range generalization. In contrast, 50 m rolling-forward updating produced reliable prediction: ExtraTrees achieved an MAE of 1.121, RMSE of 1.643, MAPE of 2.750%, R² of 0.907, and Pearson correlation of 0.964, while LightGBM achieved the lowest MAE of 1.090. ExtraTrees R² decreased from 0.927 at a 25 m horizon to 0.711 at 200 m, demonstrating progressive concept drift with prediction distance. SHAP analysis identified measured depth as the dominant predictive proxy, whereas WOB, flow rate, and RPM were more relevant to operational intervention. The revised framework therefore emphasizes leakage-free validation, adaptive updating, and separation of predictive from controllable factors rather than random-split interpolation accuracy.