Annular pressure buildup (APB) is caused by heating of the trapped drilling fluids (during production), which may lead to burst/collapse of the casing or axial ballooning, especially in subsea high-pressure/high-temperature wells. The objective of this paper is to apply machine-learning (ML) tools to increase precision of the APB estimation, and thereby improve the fluid and casing design for APB mitigation in a given well.
The APB estimation methods in literature involve theoretical and computational tools that accommodate two separate effects: volumetric expansion [pressure/volume/temperature (PVT) response] of the annulus drilling fluids and circumferential expansion (and corresponding mechanical equilibrium) of the well casings. In the present work, ML algorithms were used to accurately model “fluid density = f(T, P)” based on the experimental PVT data of a given fluid at a range of (T, P) conditions. Sensitivity analysis was performed to demonstrate improvement in precision of APB estimation (for different subsea well scenarios using different fluids) using the ML-basedmodels.
This study demonstrates that, in several subsea scenarios, a relatively small error in the experimental fluid PVT data can lead to significant variation in APB estimation. The ML-based models for “density = f(T, P)” for the fluids ensure that the cumulative error during the modeling process is minimized. The use of certain ML-based density models was shown to improve the precision of APB estimation by several hundred psi. This advantage of the ML-based density models could be used to improve the safety factors for APB mitigation, and accordingly, the work may be used to better handle the APB issue in the subsea high-pressure/high-temperature wells.