Physics-informed machine learning modelling of hydrokinetic energy harvester

Abstract Hydrokinetic energy converters are environmentally friendly and hold significant potential for bridging the energy gap in remote villages where mainstream electrical grid infrastructure is unavailable. However, the design of these energy conversion systems is complicated by the Multiphysics nature of fluid–structure interactions. Bluff bodies, such as cylinders commonly used in these converters, often exhibit oscillatory fluid–structure interactions known as vortex-induced vibrations (VIV). Traditional low-order models can capture the basic physics of these systems but often struggle to account for nonlinearities and real-world complexities. On the other hand, purely data-driven models may suffer from overfitting or require large volumes of data to perform effectively. While physics-informed neural networks (PINNs) have been applied to VIV using fully embedded Navier–Stokes this research proposes a computationally lightweight hybrid physics-informed machine learning (PIML) framework that augments a low-order structural model with a data driven neural correction term, trained exclusively on real towing-tank experimental data at low flow speed applicable to hydrokinetic energy converters in remote communities. The model achieved an 80% reduction in mean squared error compared to the physics-only baseline showing effective compensation for unmodelled nonlinear wake effects without needing full flow-field simulation.