| Abstract | The block lifting operation is a critical process in the construction of ships and offshore structures, where precise, safe control of large cranes is essential. However, ensuring stability during this operation is challenging due to the combined effects of multiple factors, including variations in block geometry and weight, lug positions, and environmental disturbances such as wind and waves. In particular, the control problem becomes even more complex in gantry and offshore floating cranes due to the dynamic coupling among components, including various types of wire ropes, trolleys, hooks, and booms. Traditional control methods, including PID (Proportional-Integral-Derivative) and model-based controllers, often fail to fully capture these nonlinear and coupled dynamics, leading to unsafe control. To address these challenges, we proposed a control method based on safe reinforcement learning. The proposed method simultaneously optimizes trolley motion and wire-rope hoisting by analyzing the positions, velocities, and accelerations of both the crane and the block. Integrating safety constraints into the reinforcement learning allows the controller to prevent excessive wire tension and unstable block oscillation while accurately positioning the block at the target location. The effectiveness of the proposed method was validated through a multibody dynamics-based simulation. Compared to traditional control methods, the proposed method achieved superior performance in suppressing block oscillation, enhancing control stability, and improving operational efficiency. Furthermore, simulations under realistic environmental conditions demonstrated the successful execution of complex operations, including block lifting and turnover. These results indicate that the proposed control method offers a promising direction toward intelligent and reliable automation of crane operations in shipyards. |
|---|---|
| Publication Date | 2026-06-10 |
Do-Hyeok Ahn, Myung-Il Roh, In-Chang Yeo, Hye-Won Lee, Seung-Ho Ham, "A Control Method Based on Safe Reinforcement Learning for Cranes in Shipyards", Proceedings of OMAE(International Conference on Ocean, Offshore and Arctic Engineering) 2026, Tokyo, Japan, 2026.06.07-12