Time series modeling of riser vibration modes using KAN-GRU algorithm

Riser is an important component in offshore oil and gas production. Under the excitation of long-term marine fluid, the riser may suffer fatigue damage. Therefore, monitoring and predicting the dynamic behavior of the riser is crucial to prevent potential failures. This paper proposed a KAN-GRU model to capture the complex excitations and long-term dependencies contained in the riser vibration mode. We collected the acceleration data of the riser from the offshore oil platform and used this data to verify the effectiveness and accuracy of the proposed method in monitoring the dynamic behavior of the riser.