Artificial intelligence-assisted station keeping for improved drillship operations
Year: 2025
Type: article
Abstract: The rising global demand for oil has driven the offshore industry toward deep and ultra-deepwater exploration. Drillships are critical in these operations due to their high mobility and adaptability to challenging environments. Station-keeping is paramount for safe operations, as drifting beyond thresholds can result in severe economic losses and environmental disasters. This study presents a novel approach to drillship station-keeping by leveraging artificial intelligence (AI) to locally control the dynamic positioning (DP) system, thereby eliminating reliance on global positioning systems or internet-based systems. A numerical model of a drillship was developed, and simulations across multiple sea states generated a comprehensive database to train an AI controller. The system focuses on key degrees of freedom: surge, sway, and yaw. Positional changes detected by the onboard inertial navigation system are analyzed to calculate displacement, representing the vessel’s response to external forces. The trained AI matches these responses to database entries, calculates the required thrust force, and applies it through DP thrusters to restore the vessel’s position. The results showed that the AI controller achieves high precision in station-keeping across various sea states, confirming its robustness and reliability. The key novelty of this method lies in its onboard, localized control system, which enhances operational independence and safety by eliminating external dependencies while significantly reducing the risk of positional loss in ultra-deepwater environments. By combining advanced numerical simulations with AI tools, this study introduces an innovative, safer, and more efficient solution for maintaining drillship stability in demanding marine conditions.
DOI: 10.36922/ijocta.8524
Language: English
Subfield: Ocean Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4408682053