Study On The Vortex Interaction And The Intelligent Wake Sensing And Control Of Tandem Flapping Foils

In nature,fish swimming in tandem and geese flying in tandem have excellent hydrodynamic performance.In recent years,more and more attention has been paid to the formation of biomimetic robots that mimic the tandem propulsion of fish and birds.Wake interaction and control are the key mechanical problems.In order to find out the mechanical characteristics and laws of wake interaction in tandem swimming and flight,and to support the research of bionic robot tandem formation,this paper reasonably simplified the wake interaction of formation swimming into the problem of tandem flapping foils,carried out the research on fluid dynamics,and designed the wake interaction sensing,estimation and optimization control system based on this,so as to enhance the propulsion performance and coordination ability of the tandem formation.The existing studies have not yet proved the common mechanical characteristics and laws of wake interaction of tandem flapping foils under different wake modes,and have not effectively realized the active sensing and online optimal control of wake interaction.In this paper,the characteristics and change rules of wake interaction of tandem flapping foils are studied through experiments and numerical simulations.Combined with bionic flow field sensing and artificial intelligence,a real-time wake interaction sensing,estimation and control system is established,which realizes the optimization control of wake interaction,the enhancement of propulsion performance and the tracking of lateral wake.Firstly,the experimental study on the wake interaction characteristics of tandem flapping foils is carried out systematically.An experimental system for testing the hydrodynamic performance of tandem flapping foils is built,which realizes accurate control of the motion trajectory and accurate measurement of instantaneous forces.The effects of Strouhal number St,effective angle of attack α0 and motion phase difference △φ on propulsion performance were systematically tested.It is found that,under the specific motion parameter combination,the thrust coefficient and propulsion efficiency of the tandem downstream foil can reach 3.62 times and 1.76 times of that of a single foil,respectively..Combined with fluorescence dye based flow visualization and three-dimensional CFD simulation,the wake interaction process of typical working conditions is studied.Through the comparative analysis of wake interaction process,instantaneous force and wake structure,the mapping relationship between wake interaction process and hydrodynamic performance of the downstream foil is found,and the mechanism of wake interaction for enhancing hydrodynamic performance of the downstream foil is explored.Secondly,a parametric model is established to describe the wake interaction law of tandem flapping foils.In order to unify the influence of different motion parameters of tandem flapping foils on the wake interaction,based on the spatiotemporal periodic analysis of the wake interaction process of tandem flapping foils,and considering the acceleration and deceleration effects of wake,the wake-foil phase Φ is introduced to describe the wake interaction process parametrically.The wake-foil phase Φ unifies the effects of vortex advection velocity Uv,flow direction distance between two foils Lx and flapping phase difference △φ.Combined with hydrodynamic measurement and flow visualization,the feasibility of accurately predicting the wake interaction process and the propulsion performance of the downstream foil using a single parameter of wake-foil phase Φ is verified,which provides great convenience for the parametric estimation and control of wake interaction.Thirdly,based on the principle of pressure sensing of fish lateral line and artificial intelligence,a wake interaction sensing and intelligent estimation model is established.The sensing and estimation of wake interaction process is the core component of wake interaction closed-loop control.Inspired by the pressure gradient based flow sensing of the lateral line of fish,a flow field sensing model based on the pressure difference sensing is designed.Based on the study of the wake interaction law of tandem flapping foils,the mechanical mapping relationship between the downstream foil surface pressure difference perception and the wake interaction process(wake-foil phase Φ)is established,which provides a reasonable input and learning label for supervised intelligent learning.With pressure difference sensing as input and wake-foil phase Φ as output,an intelligent wake interaction estimation model based on neural networks is established.It is found that compared with fully connected neural networks(ANN),one-dimensional convolutional neural networks(1D CNN)has fewer parameters and higher estimation accuracy for wake interaction process.Finally,an intelligent estimation model of wake interaction mode based on 1D CNN is established,which provides support for on-line optimization control of wake interaction mode.Fourth,a real-time sensing and optimization control system for wake interaction of tandem foils is built to enhance the thrust and efficiency of the downstream foils online.Aiming at the real-time sensing and control of wake interaction of tandem flapping foils,a real-time sensing,estimation and control system of wake interaction of tandem flapping foils is established by combining the parametric description of wake interaction law of tandem flapping foils and the intelligent estimation model of wake interaction,so as to enhance the average thrust and propulsion efficiency of downstream flapping foils.The establishment of the control system is divided into two stages:offline and online.In the off-line stage,the flow field pressure difference sensing data sets under different wake-foil phases are collected through the flume experiment,which are divided into training set,verification set and test set.The 1D CNN model learned from the training set has good estimation ability for the verification set and the test set.In the online phase,the trained 1D CNN model is used as the online estimation model of wake interaction,the closed-loop control system is designed,and the wake interaction mode with the maximum average thrust Φ|Ct max is used as the expected wake interaction mode.The synchronous flow visualization and hydrodynamic measurement show that the thrust and efficiency of the downstream foil are significantly improved through the wake interaction optimization control.Fifthly,the perceptual tracking and phase control of the downstream foil to the lateral offset wake of the upstream foil are realized.On the basis of the established wake-foil phase estimation model,the estimation of the lateral offset distance of upstream is added,the pressure sensing data under different lateral offset conditions are collected,and the 1D CNN model is trained to synchronously estimate the lateral offset of the wake Ly and the wake-foil phase Φ.The optimal combination of different neural network structures and pressure difference sensing design is studied.The ability of using the pressure difference sensing signal to estimate the frequency of the upstream wake is analyzed.The lateral wake tracking and phase control system of tandem flapping foil is designed.Based on the perceived pressure difference signal,the rear foil simultaneously estimates the wake lateral offset Ly and the wake-foil phase Φ through 1D CNN model,and realizes the tracking of the wake vortex and the optimization control of the wake-foil phase through closed-loop control.The innovations of this paper are as follows:(1)The wake-foil interaction mode and mechanical mechanism for wake interaction to enhance the propulsion efficiency of the downstream foil are proved:in the reverse Karman vortex street,the downstream foil passes through the vortex cores,which improves the average thrust lift ratio of the downstream foil;In the wake of 2P(+2S),the downstream foil passes through the vortex pair,reducing the synchronization of the lift and heave speed.(2)An modified parameterized description model of wake interaction of tandem foils is proposed,which corrects the effect of wake acceleration and deceleration,providing a parameterized,accurate and easy to calculate label for describing the wake interaction process,and supports the parameterized estimation and control of wake interaction.(3)Combined with bionic differential pressure sensing and one-dimensional convolutional neural network algorithm,an intelligent estimation model of wake interaction state is established.For the first time,online synchronous estimation and optimal control of wake-foil phase and wake lateral offset under different wake structures are realized,effectively enhancing the propulsion performance of downstream foil and the ability of formation of tandem foils.