Prediction Model of Kinetic Energy Conversion of Tandem Dual-Oscillator Based on Flow-Induced Vibration Experiment

Flow-induced vibration energy harvesting devices typically use an elastically supported body immersed in an oncoming flow to convert the sea and river current's hydrokinetic energy into electrical energy.The proportion of energy the device collects is greatly influenced by parameters such as the water flow velocity,spacing between device components,structure size,and damping coefficient.For parameter optimization and performance predictions of flow-induced vibration energy harvesting devices,we train a model of the power harvesting efficiency under different damping ratios,stiffnesses,spacing ratios,and reduced velocities based on experimental data.To improve the prediction accuracy,a feedforward network structure is optimized by using the topological evolutionary algorithm and a radial basis function network.Comparative analysis reveals that the radial basis function network model provides the best agreement with the experimental results and realizes accurate predictions of the power harvested by a dual-oscillator system in the vortex-induced vibration,transition region,and galloping.The prediction results show that the model's maximum power harvesting efficiency occurs in the vortex-induced vibration.The efficiency increases and then decreases with increasing stiffness and reduced velocity in this phase;an increase in the spacing ratio causes the effi-ciency to decrease and then increase;finally,increasing the damping ratio enhances the efficiency.The device achieves maximum power harvesting efficiency at a reduced velocity of U_(r)=4.11.The proposed model effectively predicts the maximum efficiency and the corresponding damping ratio and stiffness of the vortex-induced vibration and galloping,providing a new method for predicting tandem dual-oscillator hydrodynamic power conversion in flow-induced vibration.