Fishes moving through turbulent flows or in formation are regularly exposed to vortices. Although animals living in fluid environments commonly capture energy from vortices, experimental data on the hydrodynamics and neural control of interactions between fish and vortices are lacking. We used quantitative flow visualization and electromyography to show that trout will adopt a novel mode of locomotion to slalom in between experimentally generated vortices by activating only their anterior axial muscles. Reduced muscle activity during vortex exploitation compared with the activity of fishes engaged in undulatory swimming suggests a decrease in the cost of locomotion and provides a mechanism to understand the patterns of fish distributions in schools and riverine environments.

SUMMARYMost fishes commonly experience unsteady flows and hydrodynamic perturbations during their lifetime. In this study, we provide evidence that rainbow trout Oncorhynchus mykiss voluntarily alter their body kinematics when interacting with vortices present in the environment that are not self-generated. To demonstrate this, we measured axial swimming kinematics in response to changes in known hydrodynamic wake characteristics. We compared trout swimming in the Kármán street behind different diameter cylinders (2.5 and 5 cm) at two flow speeds (2.5 and 4.5 Ls-1, where L is total body length) to trout swimming in the free stream and in the cylinder bow wake. Trout swimming behind cylinders adopt a distinctive, previously undescribed pattern of movement in order to hold station, which we term the Kármán gait. During this gait,body amplitudes and curvatures are much larger than those of trout swimming at an equivalent flow velocity in the absence of a cylinder. Tail-beat frequency is not only lower than might be expected for a trout swimming in the reduced flow behind a cylinder, but also matches the vortex shedding frequency of the cylinder. Therefore, in addition to choosing to be in the slower flow velocity offered behind a cylinder (drafting), trout are also altering their body kinematics to synchronize with the shed vortices (tuning), using a mechanism that may not involve propulsive locomotion. This behavior is most distinctive when cylinder diameter is large relative to fish length. While tuning, trout have a longer body wavelength than the prescribed wake wavelength, indicating that only certain regions of the body may need to be oriented in a consistent manner to the oncoming vortices. Our results suggest that fish can capture energy from vortices generated by the environment to maintain station in downstream flow. Interestingly, trout swimming in front of a cylinder display lower tail-beat amplitudes and body wave speeds than trout subjected to any of the other treatments, implying that the bow wake may be the most energetically favorable region for a fish to hold station near a cylinder.

A dead fish is propelled upstream when its flexible body resonates with oncoming vortices formed in the wake of a bluff cylinder, despite being well outside the suction region of the cylinder. Within this passive propulsion mode, the body of the fish extracts sufficient energy from the oncoming vortices to develop thrust to overcome its own drag. In a similar turbulent wake and at roughly the same distance behind a bluff cylinder, a passively mounted high-aspect-ratio foil is also shown to propel itself upstream employing a similar flow energy extraction mechanism. In this case, mechanical energy is extracted from the flow at the same time that thrust is produced. These results prove experimentally that, under proper conditions, a body can follow at a distance or even catch up to another upstream body without expending any energy of its own. This observation is also significant in the development of low-drag energy harvesting devices, and in the energetics of fish dwelling in flowing water and swimming behind wake-forming obstacles.

Hydrodynamic effects of the relationship between the roll and pitch oscillations in low-aspect-ratio fins, with a laminar section and a rounded leading edge, flapping at transitional to moderately high Reynolds numbers, are considered. The fin is hinged at one end and its roll amplitude is large. Also examined is how this relationship is affected by spanwise twist, which alters the pitch oscillation amplitude and its phase relative to the roll motion. Force, efficiency and surface hot-film-anemometry measurements, and flow visualization are carried out in a tow tank. A fin of an abstracted penguin-wing planform and a NACA 0012 cross-section is used, and the chord Reynolds number varies from 3558 to 150 000 based on total speed. The fin is forced near the natural shedding frequency. Strouhal number and pitch amplitude are directly related when thrust is produced, and efficiency is maximized in narrow combinations of Strouhal number and pitch amplitude when oscillation of the leading-edge stagnation point is minimal. Twist makes the angle of attack uniform along the span and enhances thrust by up to 24 %, while maintaining high efficiency. Only 5 % of the power required to roll is spent to pitch, and yet roll and pitch are directly related. During hovering, dye visualization shows that a diffused leading-edge vortex is produced in rigid fins, which enlarges along the span; however, twist makes the vortex more uniform and the fin in turn requires less power to roll. Low-order phase maps of the measurements of force oscillation versus its derivative are modelled as due to van der Pol oscillators; the higher-order maps show trends in the sub-regimes of the transitional Reynolds number. Fin oscillation imparts a chordwise fluid motion, yielding a Stokes wave in the near-wall vorticity layer. When the roll and pitch oscillations are directly related, the wave is optimized: causing vorticity lift-up as the fin is decelerated at the roll extremity; the potential energy at the stagnation point is converted into kinetic energy; a vortex is produced as the lifted vorticity is wrapped around the leading edge; and free-stream reattachment keeps the vortex trapped. When the twist oscillation is phased along the span, this vortex becomes self-preserving at all amplitudes of twist, indicating the most stable (low-bandwidth) tuned nature.

No abstract