Leading-Edge Vortices Predict Efficient Arrangements in Fish Schools

Schooling fish can enhance propulsion by exploiting flow structures shed by their neighbors, yet determining hydrodynamically beneficial configurations remains a significant challenge. We present a physics-informed model based on leading-edge vortex (LEV) dynamics that accurately predicts thrust performance in carangiform fish schools. This model uses wake velocity fields from direct numerical simulations (DNS) of the leading swimmer(s) and couples them with the kinematics of a trailing swimmer to compute thrust changes, without requiring additional high-fidelity simulations. The model demonstrates strong agreement with DNS results and reveals the effects of relative spacing, phase difference, tailbeat amplitude, and Reynolds number on thrust enhancement. Results show that coordinated positioning and kinematics can yield up to 20% gains in thrust, and the model scales efficiently to multiple swimmers. These findings provide new insight into the fluid dynamics underlying collective swimming and offer guidance for the design of efficient bio-inspired multi-body propulsion systems.