Predicting the transient hydrodynamic loads on submarine hulls in unsteady maneuvers
University of New Brunswick
Submarine maneuvering analysis requires accurate and computationally efficient methods to predict the hydrodynamic loading on a vehicle. During maneuvers in a viscous fluid, this loading will have unsteady contributions from both the added mass of the surrounding fluid, and from motion history effects due to the shedding of vorticity into the wake. Despite this, the majority of models for submarine hydrodynamics assume quasi-steady loading, which neglects the motion history effect. This assumption is primarily made due to our limited understanding of unsteady submarine hydrodynamics. This work presents an analysis of the flow around unappended, maneuvering submarine hulls. This was accomplished using Computational Fluid Dynamics simulations of the flow around the hull. These simulations were validated by comparison against experimental data for an Unmanned Underwater Vehicle of similar geometry to a submarine hull. A range of maneuvers were simulated to explore the effect of unsteady motion. These maneuvers were restricted to planar motions without rotation, and included steady translations, impulsive accelerations, sustained accelerations, and sinusoidally oscillating maneuvers. The hulls studied range in slenderness from 7.5 < l/D < 9.5, with a Reynolds number of Re = 3.14e6 and reduced frequencies from k = f l/2U = 0.01 to k = 0.3. Motion history effects were seen to vary in significance with reduced frequency, and have a maximum impact at moderate values of 0.12 < k < 0.2. This analysis was conducted using both classical techniques and modern vorticity based methods. It was demonstrated that the vorticity based analysis allows for the decomposition of unsteady forces without the need for external data. A slender body approximation of the vorticity based description was developed and demonstrated to have good accuracy for all cases. An existing quasi-steady force estimation model was extended to include unsteady motion. This model used indicial function theory and the vortex-based force description. The unsteady model improved predictions relative to the quasi-steady assumption for reduced frequencies above k ≈ 0.1, but significantly increased the computational cost.