The Optimal Gaits of Fish-like Locomotion

In this work, we investigated the optimal gaits for fish-like swimming. Specifically, we develop and use an efficient adjoint-based optimization algorithm to find the periodic gaits that maximize the time-average thrust or propulsive efficiency. The optimization algorithm consists of three components: (1) an immersed boundary DNS solver, which is employed to solve for the flow field and compute the lift and drag on an oscillating 2D plate; (2) an iterative adjoint-based optimizer, formulated to compute the gradient of the objective function with respect to the parameters and obtain the optimal gaits; and (3) a Newton-GMRES solver, which is used to accelerate the calculation of the flowfield associated with the periodically oscillating plate. We will discuss the optimal gaits and associated flows in detail, including the sensitivity information furnished by the adjoint method.