Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Research on the Undulatory Motion Mechanism of Seahorse Based on Dynamic Mesh

Research on the Undulatory Motion Mechanism of Seahorse Based on Dynamic Mesh Hindawi Applied Bionics and Biomechanics Volume 2021, Article ID 2807236, 19 pages https://doi.org/10.1155/2021/2807236 Research Article Research on the Undulatory Motion Mechanism of Seahorse Based on Dynamic Mesh Xinyu Quan , Ximing Zhao , Shijie Zhang , Jie Zhou , Nan Yu , and Xuyan Hou State Key Laboratory of Robotics and System, Harbin Institute of Technology, No. 2 Yikuang Street, Nangang, Harbin City, Heilongjiang Province, China 150080 Correspondence should be addressed to Xuyan Hou; houxuyan@hit.edu.cn Received 2 June 2021; Accepted 25 August 2021; Published 21 September 2021 Academic Editor: Fuhao MO Copyright © 2021 Xinyu Quan et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The seahorse relies on the undulatory motion of the dorsal fin to generate thrust, which makes it possess quite high maneuverability and efficiency, and due to its low volume of the dorsal fin, it is conducive to the study of miniaturization of the driving mechanism. This paper carried out a study on the undulatory motion mechanism of the seahorse’s dorsal fin and proposed a dynamic model of the interaction between the seahorse’s dorsal fin and seawater based on the hydrodynamic properties of seawater and the theory of fluid-structure coupling. A simulation model was established using the Fluent software, and the 3D fluid dynamic mesh was used to study the undulatory motion mechanism of the seahorse’s dorsal fin. The effect of the swing frequency, amplitude, and wavelength of the seahorse’s dorsal fin on its propulsion performance was studied. On this basis, an optimized design method was used to design a bionic seahorse’s dorsal fin undulatory motion mechanism. The paper has important guiding significance for the research and miniaturization of new underwater vehicles. 1. Introduction breeding of seahorse populations in global seas [1–3]; some focus on the evolution of the seahorse, as well as basic bio- logical characteristics and living habits [4, 5]; others focus With the increasing demand for natural resources in mod- ern society, the speed of exploitation of terrestrial resources on the kinematics and dynamics of the seahorse’s dorsal is difficult to meet people’s needs for material life. The ocean, and caudal fins [6–10]. In addition, a high-speed camera sys- which accounts for 71% of the entire surface of the earth, has tem is used to study the undulatory motion of the seahorse’s become a treasure trove of resources for all countries. There dorsal fin [11]; and also, some focus on the study of the physiological structure and the mechanical properties of are not only abundant fishery resources and mineral resources but also sufficient energy resources, such as large the muscle of the seahorse’s dorsal fin [12]. oil fields and combustible ice. Whether it is economic or mil- So far, few studies have been conducted on the undula- itary, the treasure of the ocean is attractive enough, and the tory motion mechanism of the seahorse’s dorsal fin. The rea- rapid development of underwater vehicles will become inev- son is mainly due to the uncertainty of living seahorse’s itable. The seahorse relies on the undulatory motion of the movement, which leads to considerable difficulties in the dorsal fin to generate thrust, which makes it possess quite setup of the experimental device, and this uncontrollable high maneuverability and efficiency, and due to its low vol- movement will also make it difficult for high-speed cameras ume of the dorsal fin, it is conducive to the study of minia- to obtain sufficient illumination and focus. In addition, dif- turization of the driving mechanism. This article has ferences between seahorse’s species, between different gen- carried out research on the undulatory motion mechanism ders, and between different individuals will hinder of the seahorse’s dorsal fin. systematic and reproducible research. There are so many The current research on the seahorse mainly focuses on difficulties in living animal experiments, so many scholars the following aspects: some focus on the distribution and usually manufacture bionic prototypes for mechanism 2 Applied Bionics and Biomechanics Low spines Spines on first dorsal Medium-high Fine lines in neck region trunk ring enlarged Coronet medium coronet with spines cm radiate from eye cm Prominent height with five and curved eye spine spines 1 1 Spins well Snout oen Narrow body developed 2 covered in a striped Prominent net-like 3 sharp pattern of brown spine Snout oen Double cheek lines 4 finely striped spine Double cheek spine DORSAL Spins well 6 developed DORSAL Body oen Regular series 8 covered with of long and short Male Female 8 tiy red dots tail spines Male Female (a) (b) Low coronet Oen with thick cm but with five distinct Prominent rounded skin fronds on head Prominent knob-like eye spine rounded kobs or columnar coronet and neck Dark spots cm Prominent, rounded on face eye and cheek spines 3 Short Rounded Fairly thick snout spines Brood pounch very snout (except prominent in mature males some DORSAL western specimens) DORSAL White spots Proportionally Sometimes quite Male Female surrounded long tail, by a dark ring large, rounded oen with on body tubercles paler stripes Male Female (c) (d) Figure 1: Some types of seahorse’s samples. research. This method can solve the above-mentioned prob- motion of the seahorse’s dorsal fin was analyzed, which pro- lems, but the prototype dorsal finisdifficult to achieve the vide important support for the research on the undulatory swing amplitude and frequency like a living seahorse, and motion mechanism of the seahorse’s dorsal fin. On this the difficulty in manufacturing and control causes the exper- basis, an optimal design method was used to design a imental results to be inaccurate. In addition to physical seahorse-like dorsal fin wave motion mechanism, which experiments, simulation can also be used to study the mech- has important guiding significance for the development of anism of seahorse’s movement. However, due to insufficient new underwater vehicle research and the miniaturization computer computing power and other reasons, early simula- of the vehicle. tions were mainly two-dimensional plane simulations, which were difficult to directly explore the movement mech- 2. Materials and Methods anism. Nowadays, the computing power is greatly improved. Compared with physical experiments, fluid-structure cou- 2.1. Physical Model. The seahorse’s dorsal fin is composed of pling simulation can improve the efficiency of experiments fin rays and fin membrane. The length of adult seahorse’s and reduce the cost of making prototypes. At the same time, dorsal fin is generally between 3 and 25 mm, and the number it can reduce the impact of the uncontrollable motion of the of fin rays is between 10 and 30. Some types of seahorse are living seahorse on the experiment, and it is beneficial to the shown in Figure 1 [14], and the number of fins and dorsal systematic and repeatable research [13]. fin length of some types of seahorse are shown in Table 1. Based on the theory of fluid-structure interaction, this The common perpendicular of the fins is defined as the z paper uses Fluent software to construct a dynamic model -axis, and this direction is called the chord direction. The y of the interaction between the seahorse dorsal fin and seawa- -axis is perpendicular to the z-axis and straight down, and ter. The influence of the different swing frequency, wave- this direction is called the span direction. Finally, according length, and amplitude of the dorsal fin on the undulatory to the right-hand spiral, determine the x-axis direction. The Applied Bionics and Biomechanics 3 Table 1: The number of fins and dorsal fin length of some types of seahorse. Dorsal fin length Species Amount of dorsal fins 0~55~10 10~15 15~20 20~25 Big-belly seahorse 27–28 √ West African seahorse 17–18 √ Narrow-bellied seahorse 17–19 √ Barbour’s seahorse 16–22 √ Bargibanti’s seahorse 13–15 √ Réunion seahorse 16–18 √ Short-snouted seahorse 20–21 √ Giraffe seahorse 19–22 √ Knysna seahorse 16–18 √ Tiger tail seahorse 17–19 √ Crowned seahorse 14 √ Denise’s pygmy seahorse 14 √ Lined seahorse 18–19 √ Fisher’s seahorse 17–18 √ Sea pony 14–17 √ coordinate system is shown in Figure 2. Points on the dorsal fin of the seahorse approximately rotate around the z-axis. For the real seahorse’s dorsal fin swinging, its swing ampli- tude should change with the z-axis position. In order to sim- plify the model, it is considered that the swing amplitude does not change with the z-axis position. The kinematics equation of the dorsal finis r ∈½ 0, l , z ∈½ 0, L , π 2πz θ = + A sin 2πft + , ð1Þ 2 λ Figure 2: Coordinate system. x = r cos ðÞ θ , y = r sin ðÞ θ , The position of the origin of the coordinate system O is given by: where l is the fin length, L is the total length of the dorsal fin, A is the swing amplitude, f is the swing frequency, λ is the swing wavelength, and t is time. a = a + v dt, 0 x First, calculate the coordinates of a large number of points on the dorsal fin in excel. Then, import them to Solid- works to generate a point cloud, as shown in Figure 3(a). b = b + v dt, ð2Þ 0 y Through surface treatment, the physical model of the dorsal fin is obtained as shown in Figure 3(b). c = c + v dt, 0 z 2.2. Fluid-Structure Coupling Dynamics Modelling. In order to obtain the dynamic model of the seahorse’s dorsal fin, the dynamic equation of the seahorse’s dorsal fin needs to where a, b, c is the position of the origin of the coordi- be established first. First, establish two coordinate systems, nate system O in the coordinate system O. a , b , c is the 0 0 0 one of which is fixedly connected to the seahorse’s dorsal initial position of the origin of the coordinate system O in fin and is called the coordinate system O . And the other the coordinate system O. v , v , v is the speed of origin of x y z is an inertial coordinate system which is called the coordi- nate system O. The two coordinate systems coincide at the the coordinate system O along the x-, y-, and z-axes in the coordinate system O. initial moment. 4 Applied Bionics and Biomechanics (a) (b) Figure 3: (a) Point cloud import to Solidworks. (b) Generated physical model of the dorsal fin. The relationship between Euler angle change rate and where ω is the initial angular velocity of the origin of angular velocity is given by: the coordinate system O . The position of any point on the dorsal fin in the coordi- 2 3 2 3 2 3 α 1 tan ϕ sin α tan ϕ cos α ω nate system O is given by: 6 7 6 7 6 7 6 7 6 7 6 7 ′ = ⋅ : ð3Þ ϕ 0 cos α −sin α ω 4 5 4 5 4 5 ! ! r =Ht ∗ r t , ð7Þ ðÞ ðÞ θ 0 sin α/cos ϕ cos α/cos ϕ ω where HðtÞ is the coordinate transformation matrix. r Then, we can get the Euler angle by: ðtÞ is the position of any point on the dorsal fin in the coor- dinate system O . ð ð ð In order to obtain the required force parameter in the α = α + ω dt+ tan ϕ sin α ⋅ ω dt+ tan ϕ cos α ⋅ ω dt, 0 x y z dynamic equation, the N-S (Navier-Stokes) equation under ð ð turbulent flow is used to solve it. For incompressible fluids, ϕ = ϕ + cos α ⋅ ω dt− sin α ⋅ ω dt, the N-S equation turns to: 0 y z ð ð sin α ⋅ ω cos α ⋅ ω y 1 ∂p dv 2 x θ = θ + dt+ dt, f − + v∇ v = , x x cos ϕ cos ϕ ρ ∂x dt ð4Þ dv 1 ∂p ð8Þ f − + v∇ v = , ρ ∂y dt where α, ϕ, θ is the Euler angle of the coordinate system 1 ∂p dv ′ 2 z O . α , ϕ , θ is the initial Euler angle of the coordinate sys- 0 0 0 f − + v∇ v = , z z ρ ∂ dt ′ z tem O . ω , ω , ω is the angular velocity of the coordinate x y z system O . where f , f , and f are the mass force components of x y z The speed in formula (2) is obtained by: unit mass fluid in x, y, and z directions. v , v , and v are x y z ! ! the velocity components of the fluid in the x, y, and z direc- ð ð F M ! ! ! tions. p is relative pressure. v is the kinematic viscosity of the v = v + dt+ × r dt ð5Þ 0 c m I c fluid. The continuity equation for viscous fluid is given by: where v is the initial velocity of the origin of the coor- ∇·ν =0: ð9Þ dinate system O in the coordinate system O. F is the resul- tant external force on the dorsal fin. m is the total mass of There are three existing turbulence numerical simulation the dorsal fin. M is the resultant moment of the dorsal fin. methods: Direct Numerical Simulation (DNS), Reynolds I is the moment of inertia with the center of rotation as c Average Navier-Stokes(RANS), and Large Eddy Simulation the axis of rotation. r is the distance between the origin (LES). RANS is the application of statistical theory of turbu- ′ lence, which is the simulation method commonly used in of the coordinate system O and the center of mass. engineering. Usually based on Boussinesq’s eddy viscosity The angular velocity in formula (4) is solved by hypothesis, the zero equation, one equation, or two equa- ð tions are introduced to close the equation. The zero- ! ! equation model has a common shortcoming, that is, the tur- ω = ω + dt, ð6Þ c bulence viscosity coefficient only depends on the local flow Applied Bionics and Biomechanics 5 Figure 4: Meshing diagram in ICEM. Table 2: Model boundary parameters. parameters, and has nothing to do with the flow elsewhere, which is inconsistent with experimental observations. On Boundary name Boundary type this basis, a one-equation model and two-equation model Above surface Wall were developed. Among the two-equation model, the SST Below surface Wall model has certain accuracy and consumes limited comput- Left surface Wall ing resources and has higher calculation accuracy for near- Right surface Wall wall surfaces compared to the other two-equation models. Therefore, the SST model is used in this project. Front surface Wall Behind surface Wall 2.3. Simulation Model. Use the grid processing software Inlet Velocity-inlet ICEM to mesh the computing space. The area near the fin Outlet Pressure-outlet surface needs to be focused, so the density of mesh nodes near the fin surface increases. The height of the first cell per- pendicular to the body surface is 0.05 mm. This height is Table 3: Fluent simulation preset parameters. chosen to make the y + of most of the cells in contact with the body surface fall within the effective range of the stan- Parameter Prevalue dard wall function. At the same time, considering the calcu- Solver Pressure-based lation speed, take the outer division step of the model as Time Transient 0.1 mm, and the model is shown in Figure 4. Turbulence model SST Import the ICEM file into Fluent, and set the seawater Pressure-velocity coupling PISO parameters according to the seahorse’s living environment, and then, you can simulate the undulatory motion process Transient formulation First-order implicit of the seahorse’s dorsal fin under different conditions. Set Other term spatial discretization First-order upwind the model boundary parameters as shown in Table 2, and Time step size 0.001 s/0.0001 s/0.00003 s the Fluent simulation preset parameters as shown in Table 3. Number of time steps 4000 The undulatory motion of the seahorse’s dorsal finis controlled by using Fluent UDF. In the numerical calcula- pressure area, and the other side is a low-pressure area. tion process, the instantaneous force acting on the dorsal The pressure difference between the two sides results in the fin is obtained by integrating the fin surface pressure and generation of forces in the x and z directions. Since the dor- shear stress. sal fin of this example is composed of two complete sine waves, the forces generated in the x direction cancel each 3. Results and Discussion other out. Take another section as shown in Figure 8 to obtain the flow field pressure distribution of the section, as 3.1. Simulation Process. Import the Fluent calculation results shown in Figure 9. into CFD POST for postprocessing, and then, we can get the It can be seen from Figure 9 that when the dorsal fin ray flow field distribution at any time. Take 1 Hz frequency, swings, due to the pressure difference between the upper and 200 mm wavelength, and π/5 swing amplitude for qualitative lower fins, a force in the negative direction of the y-axis is explanation. The pressure distribution on the surface of the generated. dorsal fin in 0.25 s, 0.5 s, 0.75 s, and 1 s is shown in Figure 5. It can be clearly seen from Figure 5 that one side of the surface of the dorsal fin that pushes the water flow is a 3.2. The Effect of Dorsal Fin Swing Frequency on Propulsion. high-pressure area, and the other side is a low-pressure area. Use Fluent to simulate the five swing frequencies of 1 Hz, Take a section as shown in Figure 6 to obtain the flow 10 Hz, 35 Hz, 50 Hz, and 100 Hz. The wavelengths are all field pressure distribution of the section, as shown in 200 mm, and the swing amplitudes are all π/5. Record the Figure 7. force of the dorsal finin x, y, and z directions, respectively, It can be seen more clearly from Figure 7 that one side of and perform curve fitting in matlab. Since the force scatter the fin along the wave propagation direction is a high- diagram has not stabilized in the first few periods, the third 6 Applied Bionics and Biomechanics Pressure Time = 0.25 (s) Time = 0.50097 (s) contour 1 2.00e+01 1.632e+01 1.263e+01 8.947e+00 5.263e+00 1.579e+00 –2.105e+00 –5.789e+00 –9.474e+00 –1.316e+01 –1.684e+01 Time = 0.750994 (s) Time = 1.0009 (s) –2.053e+01 –2.421e+01 –2.789e+01 –3.158e+01 –3.526e+01 –3.895e+01 –4.263e+01 –4.632e+01 –5.000e+01 (Pa) Figure 5: Surface pressure distribution on the dorsal fin. Figure 6: Section position. Pressure Time = 0.25 (s) Time = 0.50 (s) contour 2 2.00e+01 1.632e+01 1.263e+01 8.947e+00 5.263e+00 1.579e+00 –2.105e+00 Wave direction Wave direction –5.789e+00 –9.474e+00 –1.316e+01 –1.684e+01 Time = 0.75 (s) Time = 1.00 (s) –2.053e+01 –2.421e+01 –2.789e+01 –3.158e+01 –3.526e+01 –3.895e+01 –4.263e+01 –4.632e+01 Wave direction Wave direction –5.000e+01 (Pa) Figure 7: Pressure distribution diagram of cross-sectional flow field. Applied Bionics and Biomechanics 7 Figure 8: Section position. Pressure Time = 0.25 (s) Time = 0.50 (s) contour 3 2.00e+01 1.632e+01 1.263e+01 8.947e+00 5.263e+00 1.579e+00 –2.105e+00 –5.789e+00 –9.474e+00 –1.316e+01 –1.684e+01 Time = 0.75 (s) Time = 1.00 (s) –2.053e+01 –2.421e+01 –2.789e+01 –3.158e+01 –3.526e+01 –3.895e+01 –4.263e+01 –4.632e+01 –5.000e+01 (Pa) Figure 9: Cross-sectional flow field pressure distribution. and fourth period scatter diagrams are selected for fitting. frequency of the dorsal fin, and the phase is also basically the Since the minimum frequency of 1 Hz and the maximum same. From Figure 10(b), the fluctuation amplitude a and frequency of 100 Hz are too far apart, the abscissa is set to frequency f can be better fitted with a quadratic function, time multiplied by the frequency of the corresponding work- and the relationship between the amplitude and frequency ing condition, in order to show the difference of different of the force fluctuation in the x direction of the dorsal fin working conditions more clearly and intuitively. can be obtained by: By using a custom function f ðxÞ = a sin ð2πbx + cÞ + d to curve-fit the force data of the dorsal finatdifferent swing fre- F = −0:02467 × f +8:105 +3:93: ð10Þ ðÞ x amplitude quencies in the x direction, the following results are obtained. It can be clearly seen from Figure 10(a) that as the fre- quency increases, the average force d and the fluctuation However, in Figure 10(c), the average force d does not amplitude a in the x direction both increase. It can be seen change much after 35 Hz. from Table 4 that the approximate sinusoidal frequency of By using a custom function f ðxÞ = a sin ð2πbx + cÞ + d the force in the x direction is basically the same as the swing to curve-fit the force data of the dorsal finatdifferent swing 8 Applied Bionics and Biomechanics –100 –200 –300 2 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4 ⁎ ⁎ Time frequency (s Hz) Frequency 1Hz 50Hz 10Hz 100Hz 35Hz (a) ⁎ ⁎ 0 ⁎ 0 ⁎ –1 –50 –2 –100 –3 –150 –4 –200 –5 ⁎ –250 –6 –7 0 10 20 30 40 50607080 90 100 0 10 20 30 40 50 60 70 80 90 100 Oscillate frequency (Hz) Oscillate frequency (Hz) Data Data Fitting curve Fitting curve (b) (c) Figure 10: (a) x-direction force on the dorsal finatdifferent frequencies. (b) The relationship between amplitude of force and frequency. (c) The relationship between average value of force and frequency. Table 4: x-direction force parameter values of the dorsal finat frequency of the force in the y direction is twice the swing different frequencies. frequency of the dorsal fin, and the phase difference is approximately π/2. From Figure 11(b), the amplitude of Parameter ab c d fluctuation a and frequency f can be better fitted with a qua- Frequency dratic function, and the relationship between the amplitude 1 Hz -0.0274 1.02 -0.1266 -0.0016 of the force fluctuation in the y direction of the dorsal fin 10 Hz -3.038 10.42 -0.2721 -0.1448 and the frequency can be obtained by: 35 Hz -36.51 36.16 -0.0187 -5.132 F = −0:02735 ×ðÞ f +0:5421 +0:2775: ð11Þ 50 Hz -84.58 51.59 -0.1224 -4.175 y amplitude 100 Hz -283.7 104.9 -0.3844 -6.922 From Figure 11(c), the average force d and frequency f can also be better fitted with a quadratic function, and the relationship between the average force in the y direction of frequencies in the y direction, the following results are the dorsal fin and the frequency can be obtained by: obtained. It can be clearly seen from Figure 11(a) that as the fre- F = −0:2353 × f − 0:1594 − 0:3479 ð12Þ ðÞ y average quency increases, the average force d and the fluctuation amplitude a in the y direction both increase significantly. It F = −0:2353 × f − 0:1594 − 0:3479 ð13Þ ðÞ y average can be seen from Table 5 that the approximate sinusoidal Amplitude (N) Force in x–direction (N) Average force (N) Applied Bionics and Biomechanics 9 0 ⁎ –500 –50 –1000 –100 –1500 –150 –2000 –200 –2500 –250 –3000 010 20 30 40 50 60 70 80 90 100 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4.0 Oscillate frequency (Hz) Time frequency (Hz) ⁎ Data Frequency Fitting curve 50Hz 1Hz 100Hz 10Hz 35Hz (a) (b) –500 –1000 –1500 –2000 0 102030405060708090 100 Oscillate frequency (Hz) ⁎ Data Fitting curve (c) Figure 11: (a) y-direction force on the dorsal finatdifferent frequencies. (b) The relationship between amplitude of force and frequency. (c) The relationship between average value of force and frequency. Table 5: y-direction force parameter values of the dorsal finat same. From Figure 12(b), the fluctuation amplitude a and different frequencies. frequency f can be better fitted with a quadratic function, and the relationship between the amplitude of the force fluc- Parameter ab c d tuation in the z direction of the dorsal fin and the frequency Frequency can be obtained by 1 Hz -0.027 2.003 1.714 -0.2349 10 Hz -2.769 20.13 1.600 -23.59 F = −0:05819 ×ðÞ f +0:1248 +0:07337: ð14Þ z amplitude 35 Hz -33.06 73.5 1.919 -285.7 50 Hz -70.69 100.5 1.667 -585 From Figure 12(c), the mean force d and frequency f can 100 Hz -276.1 201.3 1.612 -2346 also be better fitted with a quadratic function, and the rela- tionship between the mean force in the z direction of the dorsal fin and the frequency can be obtained by: By using a custom function f ðxÞ = a sin ð2πbx + cÞ + d to curve-fit the force data of the dorsal finat different swing F = −0:4687 × f +0:01611 +0:386 ð15Þ ðÞ frequencies in the z direction, the following results are z average obtained. It can be clearly seen from Figure 12(a) that as the fre- In summary, the average force and fluctuation amplitude quency increases, the average force d and the fluctuation in each direction increase with the increase of frequency, and amplitude a in the z direction both increase significantly. It the average force and fluctuation amplitude in the y and z can be seen from Table 6 that the approximate sinusoidal directions have a quadratic relationship with frequency. frequency of the force in the z direction is twice the swing The amplitude of the force fluctuation in the x direction also frequency of the dorsal fin, and the phase is basically the has a quadratic relationship with the frequency, but the Force in y-direction Average force (N) Amplitude (N) 10 Applied Bionics and Biomechanics 0 ⁎ –1000 –100 –2000 –200 –3000 –300 –4000 –400 –5000 –500 –600 –6000 0 10 20 30 40 50 60 70 80 90 100 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4.0 Time frequency (s Hz) Oscillate frequency (Hz) ⁎ Data Frequency Fitting 1Hz 50Hz 10Hz 100Hz 35Hz (a) (b) –500 –1000 –1500 –2000 –2500 –3000 –3500 –4000 –4500 0 10 20 30 40 50 60 70 80 90 100 Oscillate frequency (Hz) ⁎ Data Fitting (c) Figure 12: (a) z-direction force on the dorsal finatdifferent frequencies. (b) The relationship between amplitude of force and frequency. (c) The relationship between average value of force and frequency. Table 6: z-direction force parameter values of the dorsal finat increases, the net thrust gradually increases, and it becomes different frequencies. an approximate sine function in the swing period, which verifies the correctness of the simulation in this study. Parameter ab c d Frequency 1 Hz -0.0594 2.004 -0.0411 -0.4454 3.3. The Influence of the Swing Wavelength of the Dorsal Fin 10 Hz -5.995 19.99 0.1282 -46.41 on the Propulsion Force. Since the total length of the dorsal fin is 400 mm, in order to minimize the force fluctuations 35 Hz -71.04 73.5 0.0686 -573.4 in the x direction, an integer number of traveling waves 50 Hz -146.7 100 0.1183 -1173 should be selected for the entire dorsal fin. Therefore, Fluent 100 Hz -583.2 200.4 0.0136 -4688 is used to simulate the three swing frequencies of 400 mm, 200 mm, and 133 mm, respectively. The frequencies are all 10 Hz, and the swing amplitudes are all π/5. Record the force average value of the force in the x direction does not change of the dorsal finin x, y, and z directions, respectively, and much after 35 Hz. In reference [15], the authors presented an perform curve fitting in matlab. Since the force scatter dia- experimental investigation of flexible panels actuated with gram has not stabilized in the first few periods, the third heave oscillations at their leading edge. Results were pre- and fourth period scatter diagrams are selected for fitting. sented from kinematic video analysis, particle image veloci- By using a custom function f ðxÞ = a sin ð2πbx + cÞ + d metry, and direct force measurements. They draw the to curve-fit the force data of the dorsal finatdifferent swing conclusion “The magnitudes of both signals (net thrust wavelengths in the x direction, the following results are and power) increase with heaving frequency, as expected.” obtained. And they gave the net thrust and power curves of different It can be clearly seen from Figure 14 and Table 7 that as frequencies as Figure 13. It can be seen that as the frequency the wavelength increases, the average force d in the x Force in z-direction (N) Average Force (N) Amplitude (N) Applied Bionics and Biomechanics 11 0.5 0.5 –0.5 –0.5 0.3 0.3 0.2 0.2 0.1 𝜏 0.1 0 𝜋/2 3𝜋/2 2𝜋 0 𝜋/2 3𝜋/2 2𝜋 Figure 13: Frequency dependence of phase-averaged net thrust and power for (d). (a, b) Position of the leading edge y . (c, d) Net thrust (c) LE and power (d). Frequencies: ●: f =0:5 Hz; ■: 1.4 Hz; ◊: 2.6 Hz; ×: 3.5 Hz. –10 –20 –30 0.20 0.22 0.24 0.26 0.28 0.30 0.32 0.34 0.36 0.38 0.40 Time (s) Wavelength 400mm 200mm 133mm Figure 14: x-direction force on the dorsal finat different wavelengths. Table 7: x-direction force parameter values of the dorsal finat wavelength on the amplitude a of the force fluctuation in different wavelengths. the x direction is greatly reduced. The approximate sinusoi- dal frequency of the force in the x direction is basically the Parameter ab c d same as the swing frequency of the dorsal fin, and the phase Wavelength changes with wavelength. 133 mm -3.637 10.05 0.3407 -0.0742 By using a custom function f ðxÞ = a sin ð2πbx + cÞ + d 200 mm -3.038 10.42 -0.2721 -0.1448 to curve-fit the force data of the dorsal finatdifferent swing wavelengths in the y direction, the following results are 400 mm -17.4 10.03 -2.1592 -0.7686 obtained. It can be seen from Figure 15 and Table 8 that the aver- age force d in the y direction increases significantly with the direction increases. When the swing wavelength is 400 mm, increase of the swing wavelength, but the fluctuation ampli- reducing the wavelength can significantly suppress the fluc- tude a hardly changes. tuation amplitude a of the force in the x direction. But after By using a custom function f ðxÞ = a sin ð2πbx + cÞ + d the wavelength of 200 mm, the influence of decreasing the to curve-fit the force data of the dorsal finatdifferent swing Force in x-direction (N) (W) vLE/a (W) vLE/a 12 Applied Bionics and Biomechanics –10 –20 –30 –40 –50 –60 0.20 0.22 0.24 0.26 0.28 0.30 0.32 0.34 0.36 0.38 0.40 Time (s) Wavelength 400mm 200mm 133mm Figure 15: y-direction force on the dorsal finatdifferent wavelengths. and z directions, respectively, and perform curve fitting in Table 8: y-direction force parameter values of the dorsal finat matlab. Since the force scatter diagram has not stabilized different wavelengths. in the first few periods, the third and fourth period scatter diagrams are selected for fitting. Parameter ab c d By using a custom function f ðxÞ = a sin ð2πbx + cÞ + d Wavelength to curve-fit the force data of the dorsal finatdifferent swing 133 mm -1.556 20.12 1.5236 -13.85 amplitudes in the x direction, the following results are 200 mm -2.77 20.13 1.5946 -23.59 obtained. 400 mm -2.636 19.97 2.336 -50.11 From Figure 17 and Table 10, it can be seen that the swing amplitude has little effect on the force in the x direc- tion. No matter how the swing amplitude changes, the aver- wavelengths in the z direction, the following results are age force d in the x direction fluctuates near the 0 line, and obtained. the fluctuation amplitude a is relatively small. The approxi- It can be seen from Figure 16 and Table 9 that the mean mate sinusoidal frequency of the force in the x direction is value d of the force in the z direction increases significantly basically the same as the swing frequency of the dorsal fin, with the increase of the swing wavelength, but the fluctua- and the phase is also basically the same. tion amplitude a hardly changes. By using a custom function f ðxÞ = a sin ð2πbx + cÞ + d Based on the analysis of the force in the above three to curve-fit the force data of the dorsal finatdifferent swing directions, it can be seen that when the wavelength is amplitudes in the y direction, the following results are between 400 mm and 200 mm, as the wavelength decreases, obtained. the x direction fluctuation of the dorsal fin is significantly It can be seen from Figure 18 that as the swing amplitude suppressed while below 200 mm, the impact is small. At increases, the average force d and the fluctuation amplitude the same time, as the wavelength increases, the mean value a in the y direction both increase. It can be seen from of the force in the y and z directions increases significantly, Table 11 that the approximate sinusoidal frequency of the but the fluctuation range is almost unchanged. force in the y direction is twice the swing frequency of the dorsal fin, and the phase difference is π/2. 3.4. The Influence of the Swing Amplitude of the Dorsal Fin By using a custom function f ðxÞ = a sin ð2πbx + cÞ + d on the Propulsion Force. Use Fluent to simulate the three to curve-fit the force data of the dorsal finatdifferent swing swing amplitudes, the frequency is 10 Hz, and the wave- amplitude in the z direction, the following results are length is 200 mm. Record the force of the dorsal finin x, y, obtained. Force in y-direction (N) Applied Bionics and Biomechanics 13 –10 –20 –30 –40 –50 –60 –70 –80 0.20 0.22 0.24 0.26 0.28 0.30 0.32 0.34 0.36 0.38 0.40 Time (s) Wavelength 400mm 200mm 133mm Figure 16: z-direction force on the dorsal finatdifferent wavelengths. Table 9: z-direction force parameter values of the dorsal finat different wavelengths. ter while the right one with eccentricity. There is a shaft flat position on the right-side extension to match the middle Parameter ab c d part. An eccentric hole on the left side of the middle part Wavelength is matched with the left part, and an eccentric shaft on the 133 mm -4.735 20 0.07619 -27.15 right side is matched with the next middle part. There is also 200 mm -5.995 19.99 0.1235 -46.41 a flat shaft position, and there is a 45 phase difference 400 mm -4.671 19.93 -0.1782 -69.39 between the left hole and the right shaft, which makes adja- cent fins produce a fixed phase difference. The right part also has an eccentric hole to fit with the middle part, and a con- centric shaft on the right is connected to the motor. When It can be seen from Figure 19 that as the swing amplitude making crankshaft parts, consider hollowing out the middle increases, the average force d in the z direction and the fluc- disc to reduce the moment of inertia. tuation amplitude a both increase. It can be seen from The comprehensive problem of the function mechanism Table 12 that the approximate sinusoidal frequency of the means that the functional relationship between the input force in the z direction is twice the swing frequency of the and output angles corresponding to the rocker and the crank dorsal fin, and the phase is the same. is required to be as close as possible to the given In summary, the average value of the force in each direc- function ψ = f ðφÞ. The comprehensive theory of the mecha- tion increases with the increase of the swing amplitude, but nism proves that the kinematics of the crank-rocker mecha- the influence on the force in the x direction is negligible. nism has nothing to do with the actual length of the rod, but At the same time, the increase of the swing will cause the only with the shape of the mechanism, that is, the relative fluctuation of the force to increase. length between the rods. Let us set the relative length of the crankshaft as l =1. The relative length of the connecting 3.5. Design of Undulatory Motion Mechanism Imitating rod is l . The relative length of the pendulum is l . The rela- Seahorse’s Dorsal Fin. In order to realize that adjacent fins 2 3 tive length of the frame is l . And the initial angle of the oscillate with a fixed phase difference, a crank-rocker mech- crankshaft and the swing lever is φ , ψ , so there are 5 vari- anism driven by a crankshaft is designed, as shown in 0 0 Figure 20. ables. Therefore, at most five sets of corresponding angular positions can be accurately met. If the organization is A crankshaft is composed of a left part, a right part, and a plurality of middle parts, as shown in Figure 21. The left required to best approximate the expected function in more positions, the optimal synthesis method can be used. In the part has two protruding ends, the left of which is in the cen- Force in z-direction (N) 14 Applied Bionics and Biomechanics –5 –10 0.20 0.22 0.24 0.26 0.28 0.30 0.32 0.34 0.36 0.38 0.40 Time (s) Amplitude 36° 30° 15° Figure 17: x-direction force on the dorsal finatdifferent swing amplitudes. Table 10: x-direction force parameter values of the dorsal finat when φ = φ ~ ðφ +2πÞ, the output angle of the rocker is 0 0 different swing amplitudes. as close to the function ψ = ψ + π/5ðsin ðφ − φ − π/2Þ +1Þ 0 0 Parameter as possible. Assuming that the initial position angles of the ab c d Amplitude crank and the joystick φ , ψ correspond to the input and 0 0 π/12 output angles when the joystick is in the right extreme posi- -3.552 9.885 0.3358 -0.01746 tion. In this way, φ , ψ is no longer an independent vari- π/6 0 0 -4.124 9.983 0.4201 -0.1871 able, so there are three relative lever length variables left. π/6 -3.038 10.42 -0.2722 -0.1448 Since the three-dimensional search is more complicated and time-consuming, it is assumed that the relative length of the rack is l =15. Take the relative length of the connect- ing rod l and the relative length of the rocker l as design optimization design of the kinematics of the planar four-bar 2 3 variables. linkage, the objective function is generally established The crank-rocker mechanism is in accordance with the according to the kinematics parameters of the mechanism. corresponding relationship between the input and output For example, the movement realized by a four-bar linkage angles between the driving crank and the driven rocker. mechanism is an input-output angular function derived The independent parameters include the relative length of from the geometric relationship of the mechanism’s move- the rod l /l , l /l , and l /l and the initial angular positions ment, and it is required to have the smallest deviation from 2 1 3 1 4 1 a given function within a certain range of motion. In the of the crank and the rocker φ and ψ . As shown in 0 0 Figure 22, assume that the acute angle between the crank optimization design of linkage mechanism dynamics, it is relatively simple to use the pressure angle and transmission and the rocker and the frame when the pendulum reaches the right limit position is taken as the initial position angle angle in the mechanism as important indicators for the motion analysis and dynamic analysis of the mechanism. φ and ψ . Due to the geometric relationship of the right 0 0 In order to obtain good transmission performance and limit position, the crank and the connecting rod are collin- increase the reliability of the mechanism, it is necessary to ear, so the two initial angles can be determined according select the best mechanism dynamics parameters so that the to the geometric relationship of the initial position. maximum pressure angle is the smallest or the minimum transmission angle is the largest during the movement of 2 2 2 l + l − l − l ðÞ 1 2 3 4 the mechanism. In this project, the crank angle is required ð16Þ ψ = a cos 2ðÞ l + l l to be at any position in a circle, that is, 1 2 4 Force in x-direction (N) Applied Bionics and Biomechanics 15 –5 –10 –15 –20 –25 0.20 0.22 0.24 0.26 0.28 0.30 0.32 0.34 0.36 0.38 0.40 Time (s) Amplitude 36° 30° 15° Figure 18: y-direction force on the dorsal finatdifferent swing amplitudes. Table 11: y-direction force parameter values of the dorsal finat different swing amplitude. min fXðÞ = 〠ðÞ ψ − ψ , ð18Þ i si i=0 Parameter ab c d Amplitude where ψ is the expected output angle and ψ is the i si π/12 -0.5752 20.1 1.7978 -6.333 actual output angle. π/6 -2.142 20.04 1.9166 -20.42 According to the given functional relationship and the corresponding relationship between the two initial angles, π/5 -2.77 20.13 1.5946 -23.59 the output angle expression can be obtained by π π ð19Þ ψ = ψ + sin φ − φ − +1 , i 0 0 5 2 2 2 2 l + l − l − l ðÞ 1 2 3 4 ð17Þ ψ = a cos π i 2l l 3 4 ð20Þ φ = φ + ×ðÞ i =0,1,2,⋯,s i 0 2 s Therefore, the two initial angles can be expressed by where s is the uniform number of discrete points of the other rod lengths and are no longer independent parameters. crank angle φ in the interval φ ~ ðφ +2πÞ and i is the serial 0 0 According to the above assumption, the crank length is the number of each discrete point. The actual output angle unit length l =1. Frame length is l =15. Therefore, the 1 4 expression can be determined according to the geometric length of the connecting rod l and the length of the rocker "# "# relationship of the movement of the mechanism, as shown l x 2 1 in Figure 23. l are selected as design variables X = = . This l x 3 2 π − α − β , 0< φ ⩽ π , turns into a two-dimensional optimization design problem. ðÞ i i i ψ = ð21Þ Taking the least square deviation of the output angle of si π − α + β , π < φ ⩽ 2π : ðÞ i i i the mechanism as the design goal, the given function and the actual function are discretized, and the discrete deviation function is obtained. The sum of the discrete deviation func- Among them, in ΔBDC and ΔABD, applying the law of tions is used as the objective function, as shown in cosines, we can get Force in y-direction (N) 16 Applied Bionics and Biomechanics –10 –20 –30 –40 –50 –60 0.20 0.22 0.24 0.26 0.28 0.30 0.32 0.34 0.36 0.38 0.40 Time (s) Amplitude 36° 30° 15° Figure 19: z-direction force on the dorsal finatdifferent swing amplitude. Table 12: z-direction force parameter values of the dorsal finat 2 2 2 r + x − x i 2 1 different swing amplitudes. > > α = acos , > i 2r x i 2 Parameter ab c d 2 2 r + l − l Amplitude i 4 1 ð22Þ β = acos , > i 2l r π/12 -1.442 19.99 0.1344 -9.263 > 4 i qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi π/6 > -4.833 19.99 0.1575 -33.73 2 2 r = l + l − 2l l cos φ : i 4 1 1 4 π/6 -5.995 19.99 0.1235 -46.41 In order to make the transmission performance of the mechanism better, the minimum transmission angle of the mechanism γ ≥ 45 and the maximum transmission angle min of the mechanism γ ≤ 135 . When the crank and the max frame are collinear, the mechanism has the minimum or maximum transmission angle, as shown in Figure 24. When the mechanism is in these two positions, the law of cosines can be used to obtain by: 2 2 x + x −ðÞ l − l > 4 1 1 2 ° cos γ = ⩽ cos 45 , min 2x x 1 2 ð23Þ > 2 2 x + x −ðÞ l + l > 4 1 ° 1 2 cos γ = ⩾ cos 135 : max 2x x Figure 20: Mechanism assembly drawing. 1 2 Force in z-direction (N) Applied Bionics and Biomechanics 17 (a) (b) (c) Figure 21: Crankshaft components: (a) the left part; (b) the middle part; (c) the right part. Figure 22: Right limit position of the joystick. After sorting, get the constraint equation: According to the condition of the sum of the rod lengths of the crank connecting rod (in the crank-rocker mecha- 2 2 ° 2 nism, the crank is the shortest rod, and the sum of the length g ðÞ X = x + x − 2 cos 45 x x −ðÞ 15 − 1 ⩽ 0, 1 2 1 2 of the shortest rod and the longest rod is not greater than the 2 2 ° g ðÞ X = −x − x + 2 cos 135 x x −ðÞ 15 + 1 ⩽ 0: 1 2 1 2 2 sum of the lengths of the other two rods), the constraint con- ð24Þ ditions are obtained after sorting out: 18 Applied Bionics and Biomechanics C C l l 3 2 r ψ α si i si i β α i i 1 l Φ 4 A 1 β D l i (a) (b) Figure 23: Different ranges of crank input angle correspond to rocker output angle. min 𝛾 110 max Figure 24: Maximum and minimum transmission angle position of the mechanism. 0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 Time(s) g5 (x) = 0 g1 (x) = 0 Data Fitting curve Figure 26: Crank-rocker output angle simulation. g2 (x) = 0 g3 (x) = 0 g5 (x) = 0 In summary, the mathematical model of the optimiza- g6 (x) = 0 g4 (x) = 0 tion problem is 0 5 10 15 20 25 > min fX = min 〠 ψ − ψ  X ∈ D, ðÞ ðÞ i ii i=0 "# "# 1 > l x 2 1 X = = , ð26Þ l x > 3 2 Figure 25: Constraint planning area. 2 2 ° 2 g X = x + x − 2 cos 45 x x − l − l ⩽ 0, > ðÞ ðÞ 1 2 4 1 > 1 1 2 2 2 ° g ðÞ X = −x − x + 2 cos 135 x x −ðÞ l + l ⩽ 0: 2 1 2 1 2 4 1 l ⩾ l , g X = l − x ⩽ 0, ðÞ > 1 1 2 1 3 > Using matlab to optimize the design, the results are as l ⩾ l , g ðÞ X = l − x ⩽ 0, follows: 3 1 4 1 2 Relative length of connecting rod l /l =14:7497. 2 1 l + l ⩾ l + l , g X = l + l − x − x ⩽ 0, ð25Þ ðÞ ðÞ 4 1 1 2 2 3 1 4 5 Relative length of rocker l /l =1:7039. > 3 1 Combined with the overall size of the mechanism, take > l + l ⩾ l + l , g ðÞ X = x − x −ðÞ l − l ⩽ 0, > 3 4 1 2 6 1 2 4 1 the length of the crank l =8 and the length of the frame l : 1 4 l + l ⩾ l + l , g X = −x + x − l − l ⩽ 0: ðÞ ðÞ 1 2 4 1 2 4 1 3 7 = 120, and according to the matlab optimization design results, the connecting rod length and the rocker length can be obtained: Draw the constrained planning area of the optimal design problem, as shown in Figure 25. It can be seen from l =14:7497 × 8 ≈ 118, the figure that the constraint condition of the sum of the ð27Þ l =1:7039 × 8 ≈ 13:63: rod length of the crank is a nonfunctional constraint, and the effective constraint is the constraint condition of the mechanism transmission angle g ðXÞ ⩽ 0 and g ðXÞ ⩽ 0. The kinematics simulation of the crank and rocker 1 2 The area enclosed by them is the feasible region of the two mechanism is performed, and the output angle function is design parameters. shown in Figure 26. Output degree (deg) Applied Bionics and Biomechanics 19 Fitting result: References [1] H. J. Koldewey and K. M. Martin-Smith, “A global review of seahorse aquaculture,” Aquaculture, vol. 302, no. 3-4, fx =35:65 sin 2πx − 1:677 + 102:9: ð28Þ ðÞ ðÞ pp. 131–152, 2010. [2] S. D. Job, H. H. Do, J. J. Meeuwig, and H. J. Hall, “Culturing the oceanic seahorse, _Hippocampus kuda_,” Aquaculture, It can be seen that the output angle function of the vol. 214, no. 1-4, pp. 333–341, 2002. mechanism fits well with the objective function. [3] I. Olivotto, M. A. Avella, G. Sampaolesi, C. C. Piccinetti, P. N. Ruiz, and O. Carnevali, “Breeding and rearing the longsnout 4. Conclusions seahorse _Hippocampus reidi_ : rearing and feeding studies,” Aquaculture, vol. 283, no. 1-4, pp. 92–96, 2008. In this paper, a theoretical model of the interaction between [4] C. M. Woods, “Natural diet of the seahorseHippocampus the seahorse’s dorsal fin and seawater is firstly carried out, abdominalis,” New Zealand Journal of Marine and Freshwater and a fluid-structure coupling model suitable for this subject Research, vol. 36, no. 3, pp. 655–660, 2002. is derived, and then, a simulation model of the seahorse’s [5] Q. Lin, S. Fan, Y. Zhang et al., “The seahorse genome and the dorsal fin and seawater is established based on the fluid- evolution of its specialized morphology,” Nature, vol. 540, structure coupling theory and the physical properties of sea- no. 7633, pp. 395–399, 2016. water. And refer to the Fluent database for parameter match- [6] R. W. Blake, “The swimming of the mandarin fishSynchropus ing. Taking a certain working condition as an example, the picturatus(Callionyiidae: Teleostei),” Journal of the Marine interaction process between the seahorse’sdorsal fin and the Biological Association of the United Kingdom, vol. 59, no. 2, sea is analyzed, and the pressure distribution on the surface pp. 421–428, 1979. of the seahorse’sdorsal fin and the flow field around it during [7] R. W. Blake, “On seahorse locomotion,” Journal of the Marine the undulatory motion of the seahorse’sdorsal fin is obtained Biological Association of the United Kingdom, vol. 56, no. 4, through data postprocessing. Then, using the controlled vari- pp. 939–949, 1976. able method, keeping other variables the same, the influence of [8] J. Lighthill and R. Blake, “Biofluiddynamics of balistiform and the swing frequency, wavelength, and amplitude of the sea- gymnotiform locomotion. Part 1. Biological background, and horse’sdorsal fin on the instantaneous force and average force analysis by elongated-body theory,” Journal of Fluid Mechan- in various directions of the seahorse’sdorsal finduring the ics, vol. 212, no. 1, pp. 183–207, 1990. undulatory motion is studied, and the swing frequency, wave- [9] T. R. Consi, P. A. Seifert, M. S. Triantafyllou, and E. R. Edel- length, and amplitude of the seahorse’sdorsal fin are summa- man, “The dorsal fin engine of the seahorse (Hippocampus rized. For the influence of the instantaneous force and the sp.),” Journal of Morphology, vol. 248, no. 1, pp. 80–97, 2001. average force on the seahorse’sdorsal fin in various directions [10] M. Sfakiotakis, D. M. Laue, and B. C. Davies, “An experimental during undulatory motion, finally, an optimal design method undulating-fin device using the parallel bellows actuator,” in In is used to design a seahorse-like dorsal fin undulatory motion Proceedings 2001 ICRA. IEEE International Conference on mechanism, which has important guiding significance for the Robotics and Automation (Cat. No. 01CH37164), vol. 3, development of new underwater vehicle research and the min- pp. 2356–2362, IEEE, 2001. iaturization of the vehicle. [11] C. M. Breder Jr. and H. E. Edgerton, “An analysis of the loco- motion of the seahorse, Hippocampus, by means of high speed cinematography,” Annals of the New York Academy of Sci- Data Availability ences, vol. 43, no. 4, pp. 145–172, 1942. The data used to support the findings of this study are [12] M. A. Ashley-ross, “Mechanical properties of the dorsal fin included within the article. muscle of seahorse (Hippocampus) and pipefish (Syng- nathus),” Journal of Experimental Zoology, vol. 293, no. 6, pp. 561–577, 2002. Conflicts of Interest [13] W. Wei, X. Quan, H. Cao et al., “Research on the rapid closing The authors declare that there is no conflict of interest jet mechanism of pistol shrimp’s claws based on fluid dynamic grid,” Mathematical Problems in Engineering, vol. 2021, 17 regarding the publication of this paper. pages, 2021. [14] S. A. Lourie, S. J. Foster, E. W. Cooper, and A. C. Vincent, “A Acknowledgments guide to the identification of seahorses,” Project Seahorse and TRAFFIC North America, vol. 114, 2004. This work was financially supported by the National Key R&D [15] D. B. Quinn, G. V. Lauder, and A. J. Smits, “Scaling the propul- Program of China under Grant 2019YFB1309600, the National sive performance of heaving flexible panels,” Journal of Fluid Nature Science Foundation of China (No. 62073229, No. Mechanics, vol. 738, pp. 250–267, 2014. 51902026, and No. U1637207), the Self-Planned Task (No. SKLRS201801B) of State Key Laboratory of Robot Technology and System (HIT), the Qian Xuesen Laboratory of Space Tech- nology Seed Fund (No. QXSZZJJ03-03), and the International Science and Technology Cooperation Programme of China (No. 2014DFR50250). http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Applied Bionics and Biomechanics Hindawi Publishing Corporation

Research on the Undulatory Motion Mechanism of Seahorse Based on Dynamic Mesh

Loading next page...
 
/lp/hindawi-publishing-corporation/research-on-the-undulatory-motion-mechanism-of-seahorse-based-on-yCDTJRmFe0
Publisher
Hindawi Publishing Corporation
Copyright
Copyright © 2021 Xinyu Quan et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
ISSN
1176-2322
eISSN
1754-2103
DOI
10.1155/2021/2807236
Publisher site
See Article on Publisher Site

Abstract

Hindawi Applied Bionics and Biomechanics Volume 2021, Article ID 2807236, 19 pages https://doi.org/10.1155/2021/2807236 Research Article Research on the Undulatory Motion Mechanism of Seahorse Based on Dynamic Mesh Xinyu Quan , Ximing Zhao , Shijie Zhang , Jie Zhou , Nan Yu , and Xuyan Hou State Key Laboratory of Robotics and System, Harbin Institute of Technology, No. 2 Yikuang Street, Nangang, Harbin City, Heilongjiang Province, China 150080 Correspondence should be addressed to Xuyan Hou; houxuyan@hit.edu.cn Received 2 June 2021; Accepted 25 August 2021; Published 21 September 2021 Academic Editor: Fuhao MO Copyright © 2021 Xinyu Quan et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The seahorse relies on the undulatory motion of the dorsal fin to generate thrust, which makes it possess quite high maneuverability and efficiency, and due to its low volume of the dorsal fin, it is conducive to the study of miniaturization of the driving mechanism. This paper carried out a study on the undulatory motion mechanism of the seahorse’s dorsal fin and proposed a dynamic model of the interaction between the seahorse’s dorsal fin and seawater based on the hydrodynamic properties of seawater and the theory of fluid-structure coupling. A simulation model was established using the Fluent software, and the 3D fluid dynamic mesh was used to study the undulatory motion mechanism of the seahorse’s dorsal fin. The effect of the swing frequency, amplitude, and wavelength of the seahorse’s dorsal fin on its propulsion performance was studied. On this basis, an optimized design method was used to design a bionic seahorse’s dorsal fin undulatory motion mechanism. The paper has important guiding significance for the research and miniaturization of new underwater vehicles. 1. Introduction breeding of seahorse populations in global seas [1–3]; some focus on the evolution of the seahorse, as well as basic bio- logical characteristics and living habits [4, 5]; others focus With the increasing demand for natural resources in mod- ern society, the speed of exploitation of terrestrial resources on the kinematics and dynamics of the seahorse’s dorsal is difficult to meet people’s needs for material life. The ocean, and caudal fins [6–10]. In addition, a high-speed camera sys- which accounts for 71% of the entire surface of the earth, has tem is used to study the undulatory motion of the seahorse’s become a treasure trove of resources for all countries. There dorsal fin [11]; and also, some focus on the study of the physiological structure and the mechanical properties of are not only abundant fishery resources and mineral resources but also sufficient energy resources, such as large the muscle of the seahorse’s dorsal fin [12]. oil fields and combustible ice. Whether it is economic or mil- So far, few studies have been conducted on the undula- itary, the treasure of the ocean is attractive enough, and the tory motion mechanism of the seahorse’s dorsal fin. The rea- rapid development of underwater vehicles will become inev- son is mainly due to the uncertainty of living seahorse’s itable. The seahorse relies on the undulatory motion of the movement, which leads to considerable difficulties in the dorsal fin to generate thrust, which makes it possess quite setup of the experimental device, and this uncontrollable high maneuverability and efficiency, and due to its low vol- movement will also make it difficult for high-speed cameras ume of the dorsal fin, it is conducive to the study of minia- to obtain sufficient illumination and focus. In addition, dif- turization of the driving mechanism. This article has ferences between seahorse’s species, between different gen- carried out research on the undulatory motion mechanism ders, and between different individuals will hinder of the seahorse’s dorsal fin. systematic and reproducible research. There are so many The current research on the seahorse mainly focuses on difficulties in living animal experiments, so many scholars the following aspects: some focus on the distribution and usually manufacture bionic prototypes for mechanism 2 Applied Bionics and Biomechanics Low spines Spines on first dorsal Medium-high Fine lines in neck region trunk ring enlarged Coronet medium coronet with spines cm radiate from eye cm Prominent height with five and curved eye spine spines 1 1 Spins well Snout oen Narrow body developed 2 covered in a striped Prominent net-like 3 sharp pattern of brown spine Snout oen Double cheek lines 4 finely striped spine Double cheek spine DORSAL Spins well 6 developed DORSAL Body oen Regular series 8 covered with of long and short Male Female 8 tiy red dots tail spines Male Female (a) (b) Low coronet Oen with thick cm but with five distinct Prominent rounded skin fronds on head Prominent knob-like eye spine rounded kobs or columnar coronet and neck Dark spots cm Prominent, rounded on face eye and cheek spines 3 Short Rounded Fairly thick snout spines Brood pounch very snout (except prominent in mature males some DORSAL western specimens) DORSAL White spots Proportionally Sometimes quite Male Female surrounded long tail, by a dark ring large, rounded oen with on body tubercles paler stripes Male Female (c) (d) Figure 1: Some types of seahorse’s samples. research. This method can solve the above-mentioned prob- motion of the seahorse’s dorsal fin was analyzed, which pro- lems, but the prototype dorsal finisdifficult to achieve the vide important support for the research on the undulatory swing amplitude and frequency like a living seahorse, and motion mechanism of the seahorse’s dorsal fin. On this the difficulty in manufacturing and control causes the exper- basis, an optimal design method was used to design a imental results to be inaccurate. In addition to physical seahorse-like dorsal fin wave motion mechanism, which experiments, simulation can also be used to study the mech- has important guiding significance for the development of anism of seahorse’s movement. However, due to insufficient new underwater vehicle research and the miniaturization computer computing power and other reasons, early simula- of the vehicle. tions were mainly two-dimensional plane simulations, which were difficult to directly explore the movement mech- 2. Materials and Methods anism. Nowadays, the computing power is greatly improved. Compared with physical experiments, fluid-structure cou- 2.1. Physical Model. The seahorse’s dorsal fin is composed of pling simulation can improve the efficiency of experiments fin rays and fin membrane. The length of adult seahorse’s and reduce the cost of making prototypes. At the same time, dorsal fin is generally between 3 and 25 mm, and the number it can reduce the impact of the uncontrollable motion of the of fin rays is between 10 and 30. Some types of seahorse are living seahorse on the experiment, and it is beneficial to the shown in Figure 1 [14], and the number of fins and dorsal systematic and repeatable research [13]. fin length of some types of seahorse are shown in Table 1. Based on the theory of fluid-structure interaction, this The common perpendicular of the fins is defined as the z paper uses Fluent software to construct a dynamic model -axis, and this direction is called the chord direction. The y of the interaction between the seahorse dorsal fin and seawa- -axis is perpendicular to the z-axis and straight down, and ter. The influence of the different swing frequency, wave- this direction is called the span direction. Finally, according length, and amplitude of the dorsal fin on the undulatory to the right-hand spiral, determine the x-axis direction. The Applied Bionics and Biomechanics 3 Table 1: The number of fins and dorsal fin length of some types of seahorse. Dorsal fin length Species Amount of dorsal fins 0~55~10 10~15 15~20 20~25 Big-belly seahorse 27–28 √ West African seahorse 17–18 √ Narrow-bellied seahorse 17–19 √ Barbour’s seahorse 16–22 √ Bargibanti’s seahorse 13–15 √ Réunion seahorse 16–18 √ Short-snouted seahorse 20–21 √ Giraffe seahorse 19–22 √ Knysna seahorse 16–18 √ Tiger tail seahorse 17–19 √ Crowned seahorse 14 √ Denise’s pygmy seahorse 14 √ Lined seahorse 18–19 √ Fisher’s seahorse 17–18 √ Sea pony 14–17 √ coordinate system is shown in Figure 2. Points on the dorsal fin of the seahorse approximately rotate around the z-axis. For the real seahorse’s dorsal fin swinging, its swing ampli- tude should change with the z-axis position. In order to sim- plify the model, it is considered that the swing amplitude does not change with the z-axis position. The kinematics equation of the dorsal finis r ∈½ 0, l , z ∈½ 0, L , π 2πz θ = + A sin 2πft + , ð1Þ 2 λ Figure 2: Coordinate system. x = r cos ðÞ θ , y = r sin ðÞ θ , The position of the origin of the coordinate system O is given by: where l is the fin length, L is the total length of the dorsal fin, A is the swing amplitude, f is the swing frequency, λ is the swing wavelength, and t is time. a = a + v dt, 0 x First, calculate the coordinates of a large number of points on the dorsal fin in excel. Then, import them to Solid- works to generate a point cloud, as shown in Figure 3(a). b = b + v dt, ð2Þ 0 y Through surface treatment, the physical model of the dorsal fin is obtained as shown in Figure 3(b). c = c + v dt, 0 z 2.2. Fluid-Structure Coupling Dynamics Modelling. In order to obtain the dynamic model of the seahorse’s dorsal fin, the dynamic equation of the seahorse’s dorsal fin needs to where a, b, c is the position of the origin of the coordi- be established first. First, establish two coordinate systems, nate system O in the coordinate system O. a , b , c is the 0 0 0 one of which is fixedly connected to the seahorse’s dorsal initial position of the origin of the coordinate system O in fin and is called the coordinate system O . And the other the coordinate system O. v , v , v is the speed of origin of x y z is an inertial coordinate system which is called the coordi- nate system O. The two coordinate systems coincide at the the coordinate system O along the x-, y-, and z-axes in the coordinate system O. initial moment. 4 Applied Bionics and Biomechanics (a) (b) Figure 3: (a) Point cloud import to Solidworks. (b) Generated physical model of the dorsal fin. The relationship between Euler angle change rate and where ω is the initial angular velocity of the origin of angular velocity is given by: the coordinate system O . The position of any point on the dorsal fin in the coordi- 2 3 2 3 2 3 α 1 tan ϕ sin α tan ϕ cos α ω nate system O is given by: 6 7 6 7 6 7 6 7 6 7 6 7 ′ = ⋅ : ð3Þ ϕ 0 cos α −sin α ω 4 5 4 5 4 5 ! ! r =Ht ∗ r t , ð7Þ ðÞ ðÞ θ 0 sin α/cos ϕ cos α/cos ϕ ω where HðtÞ is the coordinate transformation matrix. r Then, we can get the Euler angle by: ðtÞ is the position of any point on the dorsal fin in the coor- dinate system O . ð ð ð In order to obtain the required force parameter in the α = α + ω dt+ tan ϕ sin α ⋅ ω dt+ tan ϕ cos α ⋅ ω dt, 0 x y z dynamic equation, the N-S (Navier-Stokes) equation under ð ð turbulent flow is used to solve it. For incompressible fluids, ϕ = ϕ + cos α ⋅ ω dt− sin α ⋅ ω dt, the N-S equation turns to: 0 y z ð ð sin α ⋅ ω cos α ⋅ ω y 1 ∂p dv 2 x θ = θ + dt+ dt, f − + v∇ v = , x x cos ϕ cos ϕ ρ ∂x dt ð4Þ dv 1 ∂p ð8Þ f − + v∇ v = , ρ ∂y dt where α, ϕ, θ is the Euler angle of the coordinate system 1 ∂p dv ′ 2 z O . α , ϕ , θ is the initial Euler angle of the coordinate sys- 0 0 0 f − + v∇ v = , z z ρ ∂ dt ′ z tem O . ω , ω , ω is the angular velocity of the coordinate x y z system O . where f , f , and f are the mass force components of x y z The speed in formula (2) is obtained by: unit mass fluid in x, y, and z directions. v , v , and v are x y z ! ! the velocity components of the fluid in the x, y, and z direc- ð ð F M ! ! ! tions. p is relative pressure. v is the kinematic viscosity of the v = v + dt+ × r dt ð5Þ 0 c m I c fluid. The continuity equation for viscous fluid is given by: where v is the initial velocity of the origin of the coor- ∇·ν =0: ð9Þ dinate system O in the coordinate system O. F is the resul- tant external force on the dorsal fin. m is the total mass of There are three existing turbulence numerical simulation the dorsal fin. M is the resultant moment of the dorsal fin. methods: Direct Numerical Simulation (DNS), Reynolds I is the moment of inertia with the center of rotation as c Average Navier-Stokes(RANS), and Large Eddy Simulation the axis of rotation. r is the distance between the origin (LES). RANS is the application of statistical theory of turbu- ′ lence, which is the simulation method commonly used in of the coordinate system O and the center of mass. engineering. Usually based on Boussinesq’s eddy viscosity The angular velocity in formula (4) is solved by hypothesis, the zero equation, one equation, or two equa- ð tions are introduced to close the equation. The zero- ! ! equation model has a common shortcoming, that is, the tur- ω = ω + dt, ð6Þ c bulence viscosity coefficient only depends on the local flow Applied Bionics and Biomechanics 5 Figure 4: Meshing diagram in ICEM. Table 2: Model boundary parameters. parameters, and has nothing to do with the flow elsewhere, which is inconsistent with experimental observations. On Boundary name Boundary type this basis, a one-equation model and two-equation model Above surface Wall were developed. Among the two-equation model, the SST Below surface Wall model has certain accuracy and consumes limited comput- Left surface Wall ing resources and has higher calculation accuracy for near- Right surface Wall wall surfaces compared to the other two-equation models. Therefore, the SST model is used in this project. Front surface Wall Behind surface Wall 2.3. Simulation Model. Use the grid processing software Inlet Velocity-inlet ICEM to mesh the computing space. The area near the fin Outlet Pressure-outlet surface needs to be focused, so the density of mesh nodes near the fin surface increases. The height of the first cell per- pendicular to the body surface is 0.05 mm. This height is Table 3: Fluent simulation preset parameters. chosen to make the y + of most of the cells in contact with the body surface fall within the effective range of the stan- Parameter Prevalue dard wall function. At the same time, considering the calcu- Solver Pressure-based lation speed, take the outer division step of the model as Time Transient 0.1 mm, and the model is shown in Figure 4. Turbulence model SST Import the ICEM file into Fluent, and set the seawater Pressure-velocity coupling PISO parameters according to the seahorse’s living environment, and then, you can simulate the undulatory motion process Transient formulation First-order implicit of the seahorse’s dorsal fin under different conditions. Set Other term spatial discretization First-order upwind the model boundary parameters as shown in Table 2, and Time step size 0.001 s/0.0001 s/0.00003 s the Fluent simulation preset parameters as shown in Table 3. Number of time steps 4000 The undulatory motion of the seahorse’s dorsal finis controlled by using Fluent UDF. In the numerical calcula- pressure area, and the other side is a low-pressure area. tion process, the instantaneous force acting on the dorsal The pressure difference between the two sides results in the fin is obtained by integrating the fin surface pressure and generation of forces in the x and z directions. Since the dor- shear stress. sal fin of this example is composed of two complete sine waves, the forces generated in the x direction cancel each 3. Results and Discussion other out. Take another section as shown in Figure 8 to obtain the flow field pressure distribution of the section, as 3.1. Simulation Process. Import the Fluent calculation results shown in Figure 9. into CFD POST for postprocessing, and then, we can get the It can be seen from Figure 9 that when the dorsal fin ray flow field distribution at any time. Take 1 Hz frequency, swings, due to the pressure difference between the upper and 200 mm wavelength, and π/5 swing amplitude for qualitative lower fins, a force in the negative direction of the y-axis is explanation. The pressure distribution on the surface of the generated. dorsal fin in 0.25 s, 0.5 s, 0.75 s, and 1 s is shown in Figure 5. It can be clearly seen from Figure 5 that one side of the surface of the dorsal fin that pushes the water flow is a 3.2. The Effect of Dorsal Fin Swing Frequency on Propulsion. high-pressure area, and the other side is a low-pressure area. Use Fluent to simulate the five swing frequencies of 1 Hz, Take a section as shown in Figure 6 to obtain the flow 10 Hz, 35 Hz, 50 Hz, and 100 Hz. The wavelengths are all field pressure distribution of the section, as shown in 200 mm, and the swing amplitudes are all π/5. Record the Figure 7. force of the dorsal finin x, y, and z directions, respectively, It can be seen more clearly from Figure 7 that one side of and perform curve fitting in matlab. Since the force scatter the fin along the wave propagation direction is a high- diagram has not stabilized in the first few periods, the third 6 Applied Bionics and Biomechanics Pressure Time = 0.25 (s) Time = 0.50097 (s) contour 1 2.00e+01 1.632e+01 1.263e+01 8.947e+00 5.263e+00 1.579e+00 –2.105e+00 –5.789e+00 –9.474e+00 –1.316e+01 –1.684e+01 Time = 0.750994 (s) Time = 1.0009 (s) –2.053e+01 –2.421e+01 –2.789e+01 –3.158e+01 –3.526e+01 –3.895e+01 –4.263e+01 –4.632e+01 –5.000e+01 (Pa) Figure 5: Surface pressure distribution on the dorsal fin. Figure 6: Section position. Pressure Time = 0.25 (s) Time = 0.50 (s) contour 2 2.00e+01 1.632e+01 1.263e+01 8.947e+00 5.263e+00 1.579e+00 –2.105e+00 Wave direction Wave direction –5.789e+00 –9.474e+00 –1.316e+01 –1.684e+01 Time = 0.75 (s) Time = 1.00 (s) –2.053e+01 –2.421e+01 –2.789e+01 –3.158e+01 –3.526e+01 –3.895e+01 –4.263e+01 –4.632e+01 Wave direction Wave direction –5.000e+01 (Pa) Figure 7: Pressure distribution diagram of cross-sectional flow field. Applied Bionics and Biomechanics 7 Figure 8: Section position. Pressure Time = 0.25 (s) Time = 0.50 (s) contour 3 2.00e+01 1.632e+01 1.263e+01 8.947e+00 5.263e+00 1.579e+00 –2.105e+00 –5.789e+00 –9.474e+00 –1.316e+01 –1.684e+01 Time = 0.75 (s) Time = 1.00 (s) –2.053e+01 –2.421e+01 –2.789e+01 –3.158e+01 –3.526e+01 –3.895e+01 –4.263e+01 –4.632e+01 –5.000e+01 (Pa) Figure 9: Cross-sectional flow field pressure distribution. and fourth period scatter diagrams are selected for fitting. frequency of the dorsal fin, and the phase is also basically the Since the minimum frequency of 1 Hz and the maximum same. From Figure 10(b), the fluctuation amplitude a and frequency of 100 Hz are too far apart, the abscissa is set to frequency f can be better fitted with a quadratic function, time multiplied by the frequency of the corresponding work- and the relationship between the amplitude and frequency ing condition, in order to show the difference of different of the force fluctuation in the x direction of the dorsal fin working conditions more clearly and intuitively. can be obtained by: By using a custom function f ðxÞ = a sin ð2πbx + cÞ + d to curve-fit the force data of the dorsal finatdifferent swing fre- F = −0:02467 × f +8:105 +3:93: ð10Þ ðÞ x amplitude quencies in the x direction, the following results are obtained. It can be clearly seen from Figure 10(a) that as the fre- quency increases, the average force d and the fluctuation However, in Figure 10(c), the average force d does not amplitude a in the x direction both increase. It can be seen change much after 35 Hz. from Table 4 that the approximate sinusoidal frequency of By using a custom function f ðxÞ = a sin ð2πbx + cÞ + d the force in the x direction is basically the same as the swing to curve-fit the force data of the dorsal finatdifferent swing 8 Applied Bionics and Biomechanics –100 –200 –300 2 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4 ⁎ ⁎ Time frequency (s Hz) Frequency 1Hz 50Hz 10Hz 100Hz 35Hz (a) ⁎ ⁎ 0 ⁎ 0 ⁎ –1 –50 –2 –100 –3 –150 –4 –200 –5 ⁎ –250 –6 –7 0 10 20 30 40 50607080 90 100 0 10 20 30 40 50 60 70 80 90 100 Oscillate frequency (Hz) Oscillate frequency (Hz) Data Data Fitting curve Fitting curve (b) (c) Figure 10: (a) x-direction force on the dorsal finatdifferent frequencies. (b) The relationship between amplitude of force and frequency. (c) The relationship between average value of force and frequency. Table 4: x-direction force parameter values of the dorsal finat frequency of the force in the y direction is twice the swing different frequencies. frequency of the dorsal fin, and the phase difference is approximately π/2. From Figure 11(b), the amplitude of Parameter ab c d fluctuation a and frequency f can be better fitted with a qua- Frequency dratic function, and the relationship between the amplitude 1 Hz -0.0274 1.02 -0.1266 -0.0016 of the force fluctuation in the y direction of the dorsal fin 10 Hz -3.038 10.42 -0.2721 -0.1448 and the frequency can be obtained by: 35 Hz -36.51 36.16 -0.0187 -5.132 F = −0:02735 ×ðÞ f +0:5421 +0:2775: ð11Þ 50 Hz -84.58 51.59 -0.1224 -4.175 y amplitude 100 Hz -283.7 104.9 -0.3844 -6.922 From Figure 11(c), the average force d and frequency f can also be better fitted with a quadratic function, and the relationship between the average force in the y direction of frequencies in the y direction, the following results are the dorsal fin and the frequency can be obtained by: obtained. It can be clearly seen from Figure 11(a) that as the fre- F = −0:2353 × f − 0:1594 − 0:3479 ð12Þ ðÞ y average quency increases, the average force d and the fluctuation amplitude a in the y direction both increase significantly. It F = −0:2353 × f − 0:1594 − 0:3479 ð13Þ ðÞ y average can be seen from Table 5 that the approximate sinusoidal Amplitude (N) Force in x–direction (N) Average force (N) Applied Bionics and Biomechanics 9 0 ⁎ –500 –50 –1000 –100 –1500 –150 –2000 –200 –2500 –250 –3000 010 20 30 40 50 60 70 80 90 100 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4.0 Oscillate frequency (Hz) Time frequency (Hz) ⁎ Data Frequency Fitting curve 50Hz 1Hz 100Hz 10Hz 35Hz (a) (b) –500 –1000 –1500 –2000 0 102030405060708090 100 Oscillate frequency (Hz) ⁎ Data Fitting curve (c) Figure 11: (a) y-direction force on the dorsal finatdifferent frequencies. (b) The relationship between amplitude of force and frequency. (c) The relationship between average value of force and frequency. Table 5: y-direction force parameter values of the dorsal finat same. From Figure 12(b), the fluctuation amplitude a and different frequencies. frequency f can be better fitted with a quadratic function, and the relationship between the amplitude of the force fluc- Parameter ab c d tuation in the z direction of the dorsal fin and the frequency Frequency can be obtained by 1 Hz -0.027 2.003 1.714 -0.2349 10 Hz -2.769 20.13 1.600 -23.59 F = −0:05819 ×ðÞ f +0:1248 +0:07337: ð14Þ z amplitude 35 Hz -33.06 73.5 1.919 -285.7 50 Hz -70.69 100.5 1.667 -585 From Figure 12(c), the mean force d and frequency f can 100 Hz -276.1 201.3 1.612 -2346 also be better fitted with a quadratic function, and the rela- tionship between the mean force in the z direction of the dorsal fin and the frequency can be obtained by: By using a custom function f ðxÞ = a sin ð2πbx + cÞ + d to curve-fit the force data of the dorsal finat different swing F = −0:4687 × f +0:01611 +0:386 ð15Þ ðÞ frequencies in the z direction, the following results are z average obtained. It can be clearly seen from Figure 12(a) that as the fre- In summary, the average force and fluctuation amplitude quency increases, the average force d and the fluctuation in each direction increase with the increase of frequency, and amplitude a in the z direction both increase significantly. It the average force and fluctuation amplitude in the y and z can be seen from Table 6 that the approximate sinusoidal directions have a quadratic relationship with frequency. frequency of the force in the z direction is twice the swing The amplitude of the force fluctuation in the x direction also frequency of the dorsal fin, and the phase is basically the has a quadratic relationship with the frequency, but the Force in y-direction Average force (N) Amplitude (N) 10 Applied Bionics and Biomechanics 0 ⁎ –1000 –100 –2000 –200 –3000 –300 –4000 –400 –5000 –500 –600 –6000 0 10 20 30 40 50 60 70 80 90 100 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4.0 Time frequency (s Hz) Oscillate frequency (Hz) ⁎ Data Frequency Fitting 1Hz 50Hz 10Hz 100Hz 35Hz (a) (b) –500 –1000 –1500 –2000 –2500 –3000 –3500 –4000 –4500 0 10 20 30 40 50 60 70 80 90 100 Oscillate frequency (Hz) ⁎ Data Fitting (c) Figure 12: (a) z-direction force on the dorsal finatdifferent frequencies. (b) The relationship between amplitude of force and frequency. (c) The relationship between average value of force and frequency. Table 6: z-direction force parameter values of the dorsal finat increases, the net thrust gradually increases, and it becomes different frequencies. an approximate sine function in the swing period, which verifies the correctness of the simulation in this study. Parameter ab c d Frequency 1 Hz -0.0594 2.004 -0.0411 -0.4454 3.3. The Influence of the Swing Wavelength of the Dorsal Fin 10 Hz -5.995 19.99 0.1282 -46.41 on the Propulsion Force. Since the total length of the dorsal fin is 400 mm, in order to minimize the force fluctuations 35 Hz -71.04 73.5 0.0686 -573.4 in the x direction, an integer number of traveling waves 50 Hz -146.7 100 0.1183 -1173 should be selected for the entire dorsal fin. Therefore, Fluent 100 Hz -583.2 200.4 0.0136 -4688 is used to simulate the three swing frequencies of 400 mm, 200 mm, and 133 mm, respectively. The frequencies are all 10 Hz, and the swing amplitudes are all π/5. Record the force average value of the force in the x direction does not change of the dorsal finin x, y, and z directions, respectively, and much after 35 Hz. In reference [15], the authors presented an perform curve fitting in matlab. Since the force scatter dia- experimental investigation of flexible panels actuated with gram has not stabilized in the first few periods, the third heave oscillations at their leading edge. Results were pre- and fourth period scatter diagrams are selected for fitting. sented from kinematic video analysis, particle image veloci- By using a custom function f ðxÞ = a sin ð2πbx + cÞ + d metry, and direct force measurements. They draw the to curve-fit the force data of the dorsal finatdifferent swing conclusion “The magnitudes of both signals (net thrust wavelengths in the x direction, the following results are and power) increase with heaving frequency, as expected.” obtained. And they gave the net thrust and power curves of different It can be clearly seen from Figure 14 and Table 7 that as frequencies as Figure 13. It can be seen that as the frequency the wavelength increases, the average force d in the x Force in z-direction (N) Average Force (N) Amplitude (N) Applied Bionics and Biomechanics 11 0.5 0.5 –0.5 –0.5 0.3 0.3 0.2 0.2 0.1 𝜏 0.1 0 𝜋/2 3𝜋/2 2𝜋 0 𝜋/2 3𝜋/2 2𝜋 Figure 13: Frequency dependence of phase-averaged net thrust and power for (d). (a, b) Position of the leading edge y . (c, d) Net thrust (c) LE and power (d). Frequencies: ●: f =0:5 Hz; ■: 1.4 Hz; ◊: 2.6 Hz; ×: 3.5 Hz. –10 –20 –30 0.20 0.22 0.24 0.26 0.28 0.30 0.32 0.34 0.36 0.38 0.40 Time (s) Wavelength 400mm 200mm 133mm Figure 14: x-direction force on the dorsal finat different wavelengths. Table 7: x-direction force parameter values of the dorsal finat wavelength on the amplitude a of the force fluctuation in different wavelengths. the x direction is greatly reduced. The approximate sinusoi- dal frequency of the force in the x direction is basically the Parameter ab c d same as the swing frequency of the dorsal fin, and the phase Wavelength changes with wavelength. 133 mm -3.637 10.05 0.3407 -0.0742 By using a custom function f ðxÞ = a sin ð2πbx + cÞ + d 200 mm -3.038 10.42 -0.2721 -0.1448 to curve-fit the force data of the dorsal finatdifferent swing wavelengths in the y direction, the following results are 400 mm -17.4 10.03 -2.1592 -0.7686 obtained. It can be seen from Figure 15 and Table 8 that the aver- age force d in the y direction increases significantly with the direction increases. When the swing wavelength is 400 mm, increase of the swing wavelength, but the fluctuation ampli- reducing the wavelength can significantly suppress the fluc- tude a hardly changes. tuation amplitude a of the force in the x direction. But after By using a custom function f ðxÞ = a sin ð2πbx + cÞ + d the wavelength of 200 mm, the influence of decreasing the to curve-fit the force data of the dorsal finatdifferent swing Force in x-direction (N) (W) vLE/a (W) vLE/a 12 Applied Bionics and Biomechanics –10 –20 –30 –40 –50 –60 0.20 0.22 0.24 0.26 0.28 0.30 0.32 0.34 0.36 0.38 0.40 Time (s) Wavelength 400mm 200mm 133mm Figure 15: y-direction force on the dorsal finatdifferent wavelengths. and z directions, respectively, and perform curve fitting in Table 8: y-direction force parameter values of the dorsal finat matlab. Since the force scatter diagram has not stabilized different wavelengths. in the first few periods, the third and fourth period scatter diagrams are selected for fitting. Parameter ab c d By using a custom function f ðxÞ = a sin ð2πbx + cÞ + d Wavelength to curve-fit the force data of the dorsal finatdifferent swing 133 mm -1.556 20.12 1.5236 -13.85 amplitudes in the x direction, the following results are 200 mm -2.77 20.13 1.5946 -23.59 obtained. 400 mm -2.636 19.97 2.336 -50.11 From Figure 17 and Table 10, it can be seen that the swing amplitude has little effect on the force in the x direc- tion. No matter how the swing amplitude changes, the aver- wavelengths in the z direction, the following results are age force d in the x direction fluctuates near the 0 line, and obtained. the fluctuation amplitude a is relatively small. The approxi- It can be seen from Figure 16 and Table 9 that the mean mate sinusoidal frequency of the force in the x direction is value d of the force in the z direction increases significantly basically the same as the swing frequency of the dorsal fin, with the increase of the swing wavelength, but the fluctua- and the phase is also basically the same. tion amplitude a hardly changes. By using a custom function f ðxÞ = a sin ð2πbx + cÞ + d Based on the analysis of the force in the above three to curve-fit the force data of the dorsal finatdifferent swing directions, it can be seen that when the wavelength is amplitudes in the y direction, the following results are between 400 mm and 200 mm, as the wavelength decreases, obtained. the x direction fluctuation of the dorsal fin is significantly It can be seen from Figure 18 that as the swing amplitude suppressed while below 200 mm, the impact is small. At increases, the average force d and the fluctuation amplitude the same time, as the wavelength increases, the mean value a in the y direction both increase. It can be seen from of the force in the y and z directions increases significantly, Table 11 that the approximate sinusoidal frequency of the but the fluctuation range is almost unchanged. force in the y direction is twice the swing frequency of the dorsal fin, and the phase difference is π/2. 3.4. The Influence of the Swing Amplitude of the Dorsal Fin By using a custom function f ðxÞ = a sin ð2πbx + cÞ + d on the Propulsion Force. Use Fluent to simulate the three to curve-fit the force data of the dorsal finatdifferent swing swing amplitudes, the frequency is 10 Hz, and the wave- amplitude in the z direction, the following results are length is 200 mm. Record the force of the dorsal finin x, y, obtained. Force in y-direction (N) Applied Bionics and Biomechanics 13 –10 –20 –30 –40 –50 –60 –70 –80 0.20 0.22 0.24 0.26 0.28 0.30 0.32 0.34 0.36 0.38 0.40 Time (s) Wavelength 400mm 200mm 133mm Figure 16: z-direction force on the dorsal finatdifferent wavelengths. Table 9: z-direction force parameter values of the dorsal finat different wavelengths. ter while the right one with eccentricity. There is a shaft flat position on the right-side extension to match the middle Parameter ab c d part. An eccentric hole on the left side of the middle part Wavelength is matched with the left part, and an eccentric shaft on the 133 mm -4.735 20 0.07619 -27.15 right side is matched with the next middle part. There is also 200 mm -5.995 19.99 0.1235 -46.41 a flat shaft position, and there is a 45 phase difference 400 mm -4.671 19.93 -0.1782 -69.39 between the left hole and the right shaft, which makes adja- cent fins produce a fixed phase difference. The right part also has an eccentric hole to fit with the middle part, and a con- centric shaft on the right is connected to the motor. When It can be seen from Figure 19 that as the swing amplitude making crankshaft parts, consider hollowing out the middle increases, the average force d in the z direction and the fluc- disc to reduce the moment of inertia. tuation amplitude a both increase. It can be seen from The comprehensive problem of the function mechanism Table 12 that the approximate sinusoidal frequency of the means that the functional relationship between the input force in the z direction is twice the swing frequency of the and output angles corresponding to the rocker and the crank dorsal fin, and the phase is the same. is required to be as close as possible to the given In summary, the average value of the force in each direc- function ψ = f ðφÞ. The comprehensive theory of the mecha- tion increases with the increase of the swing amplitude, but nism proves that the kinematics of the crank-rocker mecha- the influence on the force in the x direction is negligible. nism has nothing to do with the actual length of the rod, but At the same time, the increase of the swing will cause the only with the shape of the mechanism, that is, the relative fluctuation of the force to increase. length between the rods. Let us set the relative length of the crankshaft as l =1. The relative length of the connecting 3.5. Design of Undulatory Motion Mechanism Imitating rod is l . The relative length of the pendulum is l . The rela- Seahorse’s Dorsal Fin. In order to realize that adjacent fins 2 3 tive length of the frame is l . And the initial angle of the oscillate with a fixed phase difference, a crank-rocker mech- crankshaft and the swing lever is φ , ψ , so there are 5 vari- anism driven by a crankshaft is designed, as shown in 0 0 Figure 20. ables. Therefore, at most five sets of corresponding angular positions can be accurately met. If the organization is A crankshaft is composed of a left part, a right part, and a plurality of middle parts, as shown in Figure 21. The left required to best approximate the expected function in more positions, the optimal synthesis method can be used. In the part has two protruding ends, the left of which is in the cen- Force in z-direction (N) 14 Applied Bionics and Biomechanics –5 –10 0.20 0.22 0.24 0.26 0.28 0.30 0.32 0.34 0.36 0.38 0.40 Time (s) Amplitude 36° 30° 15° Figure 17: x-direction force on the dorsal finatdifferent swing amplitudes. Table 10: x-direction force parameter values of the dorsal finat when φ = φ ~ ðφ +2πÞ, the output angle of the rocker is 0 0 different swing amplitudes. as close to the function ψ = ψ + π/5ðsin ðφ − φ − π/2Þ +1Þ 0 0 Parameter as possible. Assuming that the initial position angles of the ab c d Amplitude crank and the joystick φ , ψ correspond to the input and 0 0 π/12 output angles when the joystick is in the right extreme posi- -3.552 9.885 0.3358 -0.01746 tion. In this way, φ , ψ is no longer an independent vari- π/6 0 0 -4.124 9.983 0.4201 -0.1871 able, so there are three relative lever length variables left. π/6 -3.038 10.42 -0.2722 -0.1448 Since the three-dimensional search is more complicated and time-consuming, it is assumed that the relative length of the rack is l =15. Take the relative length of the connect- ing rod l and the relative length of the rocker l as design optimization design of the kinematics of the planar four-bar 2 3 variables. linkage, the objective function is generally established The crank-rocker mechanism is in accordance with the according to the kinematics parameters of the mechanism. corresponding relationship between the input and output For example, the movement realized by a four-bar linkage angles between the driving crank and the driven rocker. mechanism is an input-output angular function derived The independent parameters include the relative length of from the geometric relationship of the mechanism’s move- the rod l /l , l /l , and l /l and the initial angular positions ment, and it is required to have the smallest deviation from 2 1 3 1 4 1 a given function within a certain range of motion. In the of the crank and the rocker φ and ψ . As shown in 0 0 Figure 22, assume that the acute angle between the crank optimization design of linkage mechanism dynamics, it is relatively simple to use the pressure angle and transmission and the rocker and the frame when the pendulum reaches the right limit position is taken as the initial position angle angle in the mechanism as important indicators for the motion analysis and dynamic analysis of the mechanism. φ and ψ . Due to the geometric relationship of the right 0 0 In order to obtain good transmission performance and limit position, the crank and the connecting rod are collin- increase the reliability of the mechanism, it is necessary to ear, so the two initial angles can be determined according select the best mechanism dynamics parameters so that the to the geometric relationship of the initial position. maximum pressure angle is the smallest or the minimum transmission angle is the largest during the movement of 2 2 2 l + l − l − l ðÞ 1 2 3 4 the mechanism. In this project, the crank angle is required ð16Þ ψ = a cos 2ðÞ l + l l to be at any position in a circle, that is, 1 2 4 Force in x-direction (N) Applied Bionics and Biomechanics 15 –5 –10 –15 –20 –25 0.20 0.22 0.24 0.26 0.28 0.30 0.32 0.34 0.36 0.38 0.40 Time (s) Amplitude 36° 30° 15° Figure 18: y-direction force on the dorsal finatdifferent swing amplitudes. Table 11: y-direction force parameter values of the dorsal finat different swing amplitude. min fXðÞ = 〠ðÞ ψ − ψ , ð18Þ i si i=0 Parameter ab c d Amplitude where ψ is the expected output angle and ψ is the i si π/12 -0.5752 20.1 1.7978 -6.333 actual output angle. π/6 -2.142 20.04 1.9166 -20.42 According to the given functional relationship and the corresponding relationship between the two initial angles, π/5 -2.77 20.13 1.5946 -23.59 the output angle expression can be obtained by π π ð19Þ ψ = ψ + sin φ − φ − +1 , i 0 0 5 2 2 2 2 l + l − l − l ðÞ 1 2 3 4 ð17Þ ψ = a cos π i 2l l 3 4 ð20Þ φ = φ + ×ðÞ i =0,1,2,⋯,s i 0 2 s Therefore, the two initial angles can be expressed by where s is the uniform number of discrete points of the other rod lengths and are no longer independent parameters. crank angle φ in the interval φ ~ ðφ +2πÞ and i is the serial 0 0 According to the above assumption, the crank length is the number of each discrete point. The actual output angle unit length l =1. Frame length is l =15. Therefore, the 1 4 expression can be determined according to the geometric length of the connecting rod l and the length of the rocker "# "# relationship of the movement of the mechanism, as shown l x 2 1 in Figure 23. l are selected as design variables X = = . This l x 3 2 π − α − β , 0< φ ⩽ π , turns into a two-dimensional optimization design problem. ðÞ i i i ψ = ð21Þ Taking the least square deviation of the output angle of si π − α + β , π < φ ⩽ 2π : ðÞ i i i the mechanism as the design goal, the given function and the actual function are discretized, and the discrete deviation function is obtained. The sum of the discrete deviation func- Among them, in ΔBDC and ΔABD, applying the law of tions is used as the objective function, as shown in cosines, we can get Force in y-direction (N) 16 Applied Bionics and Biomechanics –10 –20 –30 –40 –50 –60 0.20 0.22 0.24 0.26 0.28 0.30 0.32 0.34 0.36 0.38 0.40 Time (s) Amplitude 36° 30° 15° Figure 19: z-direction force on the dorsal finatdifferent swing amplitude. Table 12: z-direction force parameter values of the dorsal finat 2 2 2 r + x − x i 2 1 different swing amplitudes. > > α = acos , > i 2r x i 2 Parameter ab c d 2 2 r + l − l Amplitude i 4 1 ð22Þ β = acos , > i 2l r π/12 -1.442 19.99 0.1344 -9.263 > 4 i qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi π/6 > -4.833 19.99 0.1575 -33.73 2 2 r = l + l − 2l l cos φ : i 4 1 1 4 π/6 -5.995 19.99 0.1235 -46.41 In order to make the transmission performance of the mechanism better, the minimum transmission angle of the mechanism γ ≥ 45 and the maximum transmission angle min of the mechanism γ ≤ 135 . When the crank and the max frame are collinear, the mechanism has the minimum or maximum transmission angle, as shown in Figure 24. When the mechanism is in these two positions, the law of cosines can be used to obtain by: 2 2 x + x −ðÞ l − l > 4 1 1 2 ° cos γ = ⩽ cos 45 , min 2x x 1 2 ð23Þ > 2 2 x + x −ðÞ l + l > 4 1 ° 1 2 cos γ = ⩾ cos 135 : max 2x x Figure 20: Mechanism assembly drawing. 1 2 Force in z-direction (N) Applied Bionics and Biomechanics 17 (a) (b) (c) Figure 21: Crankshaft components: (a) the left part; (b) the middle part; (c) the right part. Figure 22: Right limit position of the joystick. After sorting, get the constraint equation: According to the condition of the sum of the rod lengths of the crank connecting rod (in the crank-rocker mecha- 2 2 ° 2 nism, the crank is the shortest rod, and the sum of the length g ðÞ X = x + x − 2 cos 45 x x −ðÞ 15 − 1 ⩽ 0, 1 2 1 2 of the shortest rod and the longest rod is not greater than the 2 2 ° g ðÞ X = −x − x + 2 cos 135 x x −ðÞ 15 + 1 ⩽ 0: 1 2 1 2 2 sum of the lengths of the other two rods), the constraint con- ð24Þ ditions are obtained after sorting out: 18 Applied Bionics and Biomechanics C C l l 3 2 r ψ α si i si i β α i i 1 l Φ 4 A 1 β D l i (a) (b) Figure 23: Different ranges of crank input angle correspond to rocker output angle. min 𝛾 110 max Figure 24: Maximum and minimum transmission angle position of the mechanism. 0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 Time(s) g5 (x) = 0 g1 (x) = 0 Data Fitting curve Figure 26: Crank-rocker output angle simulation. g2 (x) = 0 g3 (x) = 0 g5 (x) = 0 In summary, the mathematical model of the optimiza- g6 (x) = 0 g4 (x) = 0 tion problem is 0 5 10 15 20 25 > min fX = min 〠 ψ − ψ  X ∈ D, ðÞ ðÞ i ii i=0 "# "# 1 > l x 2 1 X = = , ð26Þ l x > 3 2 Figure 25: Constraint planning area. 2 2 ° 2 g X = x + x − 2 cos 45 x x − l − l ⩽ 0, > ðÞ ðÞ 1 2 4 1 > 1 1 2 2 2 ° g ðÞ X = −x − x + 2 cos 135 x x −ðÞ l + l ⩽ 0: 2 1 2 1 2 4 1 l ⩾ l , g X = l − x ⩽ 0, ðÞ > 1 1 2 1 3 > Using matlab to optimize the design, the results are as l ⩾ l , g ðÞ X = l − x ⩽ 0, follows: 3 1 4 1 2 Relative length of connecting rod l /l =14:7497. 2 1 l + l ⩾ l + l , g X = l + l − x − x ⩽ 0, ð25Þ ðÞ ðÞ 4 1 1 2 2 3 1 4 5 Relative length of rocker l /l =1:7039. > 3 1 Combined with the overall size of the mechanism, take > l + l ⩾ l + l , g ðÞ X = x − x −ðÞ l − l ⩽ 0, > 3 4 1 2 6 1 2 4 1 the length of the crank l =8 and the length of the frame l : 1 4 l + l ⩾ l + l , g X = −x + x − l − l ⩽ 0: ðÞ ðÞ 1 2 4 1 2 4 1 3 7 = 120, and according to the matlab optimization design results, the connecting rod length and the rocker length can be obtained: Draw the constrained planning area of the optimal design problem, as shown in Figure 25. It can be seen from l =14:7497 × 8 ≈ 118, the figure that the constraint condition of the sum of the ð27Þ l =1:7039 × 8 ≈ 13:63: rod length of the crank is a nonfunctional constraint, and the effective constraint is the constraint condition of the mechanism transmission angle g ðXÞ ⩽ 0 and g ðXÞ ⩽ 0. The kinematics simulation of the crank and rocker 1 2 The area enclosed by them is the feasible region of the two mechanism is performed, and the output angle function is design parameters. shown in Figure 26. Output degree (deg) Applied Bionics and Biomechanics 19 Fitting result: References [1] H. J. Koldewey and K. M. Martin-Smith, “A global review of seahorse aquaculture,” Aquaculture, vol. 302, no. 3-4, fx =35:65 sin 2πx − 1:677 + 102:9: ð28Þ ðÞ ðÞ pp. 131–152, 2010. [2] S. D. Job, H. H. Do, J. J. Meeuwig, and H. J. Hall, “Culturing the oceanic seahorse, _Hippocampus kuda_,” Aquaculture, It can be seen that the output angle function of the vol. 214, no. 1-4, pp. 333–341, 2002. mechanism fits well with the objective function. [3] I. Olivotto, M. A. Avella, G. Sampaolesi, C. C. Piccinetti, P. N. Ruiz, and O. Carnevali, “Breeding and rearing the longsnout 4. Conclusions seahorse _Hippocampus reidi_ : rearing and feeding studies,” Aquaculture, vol. 283, no. 1-4, pp. 92–96, 2008. In this paper, a theoretical model of the interaction between [4] C. M. Woods, “Natural diet of the seahorseHippocampus the seahorse’s dorsal fin and seawater is firstly carried out, abdominalis,” New Zealand Journal of Marine and Freshwater and a fluid-structure coupling model suitable for this subject Research, vol. 36, no. 3, pp. 655–660, 2002. is derived, and then, a simulation model of the seahorse’s [5] Q. Lin, S. Fan, Y. Zhang et al., “The seahorse genome and the dorsal fin and seawater is established based on the fluid- evolution of its specialized morphology,” Nature, vol. 540, structure coupling theory and the physical properties of sea- no. 7633, pp. 395–399, 2016. water. And refer to the Fluent database for parameter match- [6] R. W. Blake, “The swimming of the mandarin fishSynchropus ing. Taking a certain working condition as an example, the picturatus(Callionyiidae: Teleostei),” Journal of the Marine interaction process between the seahorse’sdorsal fin and the Biological Association of the United Kingdom, vol. 59, no. 2, sea is analyzed, and the pressure distribution on the surface pp. 421–428, 1979. of the seahorse’sdorsal fin and the flow field around it during [7] R. W. Blake, “On seahorse locomotion,” Journal of the Marine the undulatory motion of the seahorse’sdorsal fin is obtained Biological Association of the United Kingdom, vol. 56, no. 4, through data postprocessing. Then, using the controlled vari- pp. 939–949, 1976. able method, keeping other variables the same, the influence of [8] J. Lighthill and R. Blake, “Biofluiddynamics of balistiform and the swing frequency, wavelength, and amplitude of the sea- gymnotiform locomotion. Part 1. Biological background, and horse’sdorsal fin on the instantaneous force and average force analysis by elongated-body theory,” Journal of Fluid Mechan- in various directions of the seahorse’sdorsal finduring the ics, vol. 212, no. 1, pp. 183–207, 1990. undulatory motion is studied, and the swing frequency, wave- [9] T. R. Consi, P. A. Seifert, M. S. Triantafyllou, and E. R. Edel- length, and amplitude of the seahorse’sdorsal fin are summa- man, “The dorsal fin engine of the seahorse (Hippocampus rized. For the influence of the instantaneous force and the sp.),” Journal of Morphology, vol. 248, no. 1, pp. 80–97, 2001. average force on the seahorse’sdorsal fin in various directions [10] M. Sfakiotakis, D. M. Laue, and B. C. Davies, “An experimental during undulatory motion, finally, an optimal design method undulating-fin device using the parallel bellows actuator,” in In is used to design a seahorse-like dorsal fin undulatory motion Proceedings 2001 ICRA. IEEE International Conference on mechanism, which has important guiding significance for the Robotics and Automation (Cat. No. 01CH37164), vol. 3, development of new underwater vehicle research and the min- pp. 2356–2362, IEEE, 2001. iaturization of the vehicle. [11] C. M. Breder Jr. and H. E. Edgerton, “An analysis of the loco- motion of the seahorse, Hippocampus, by means of high speed cinematography,” Annals of the New York Academy of Sci- Data Availability ences, vol. 43, no. 4, pp. 145–172, 1942. The data used to support the findings of this study are [12] M. A. Ashley-ross, “Mechanical properties of the dorsal fin included within the article. muscle of seahorse (Hippocampus) and pipefish (Syng- nathus),” Journal of Experimental Zoology, vol. 293, no. 6, pp. 561–577, 2002. Conflicts of Interest [13] W. Wei, X. Quan, H. Cao et al., “Research on the rapid closing The authors declare that there is no conflict of interest jet mechanism of pistol shrimp’s claws based on fluid dynamic grid,” Mathematical Problems in Engineering, vol. 2021, 17 regarding the publication of this paper. pages, 2021. [14] S. A. Lourie, S. J. Foster, E. W. Cooper, and A. C. Vincent, “A Acknowledgments guide to the identification of seahorses,” Project Seahorse and TRAFFIC North America, vol. 114, 2004. This work was financially supported by the National Key R&D [15] D. B. Quinn, G. V. Lauder, and A. J. Smits, “Scaling the propul- Program of China under Grant 2019YFB1309600, the National sive performance of heaving flexible panels,” Journal of Fluid Nature Science Foundation of China (No. 62073229, No. Mechanics, vol. 738, pp. 250–267, 2014. 51902026, and No. U1637207), the Self-Planned Task (No. SKLRS201801B) of State Key Laboratory of Robot Technology and System (HIT), the Qian Xuesen Laboratory of Space Tech- nology Seed Fund (No. QXSZZJJ03-03), and the International Science and Technology Cooperation Programme of China (No. 2014DFR50250).

Journal

Applied Bionics and BiomechanicsHindawi Publishing Corporation

Published: Sep 21, 2021

References