19th Asia-Pacific Vibration Conference (APVC2021)

188 / 2021-03-31 20:12:31

Study of Drag Force of Cone-shaped Drogue in Fluid Based on Machine Learning

Cone-shaped drogue; Drag force; Vibration; Data processing; Virtual simulation

   The cone-shaped drogue is used to stabilize the seismic array in marine geological surveying. Circular truncated cone with both circular sections opened is one potential choice. To design a suitable cone-shaped drogue, the towed drogue is modeled by the co-simulation of RecurDyn and Particleworks. The accuracy of the virtual model is verify by physical experiment which is conducted in one water tank. The simulation results indicated that the drag force is mainly affected by the towing velocity and the geometry of the drogue. Therefore, the large radius, small radius, length of cone and the towing speed are chosen as the design variables in design study. Due to the limitation of experiments, most of the sampling data can be obtained through virtual simulation. Since vibration exists in the drag force, both the mean value and the vibration amplitude of the drag force is analyzed by using machine learning method. Finally, a surrogate model reflecting the relations among the drogue size, towing speed, and the mean value and vibration amplitude of the drag force are established by machine learning method.