Tandem hydrofoil motion system and motion control method

Abstract

The present application relates to the technical field of water current power generation, and discloses a series water wing movement system and a movement control method, the series water wing movement system comprising two oscillating water wing power generation devices, each oscillating water wing power generation device comprising a support frame, a generator, a first guide rail, a sliding block, a water wing, a first motor, a second motor and a first transmission mechanism, the series water wing movement system further comprising a base, a second guide rail, a third motor and a second transmission mechanism. The movement control method comprises the following steps: 1) detecting the current water flow velocity in real time; 2) inputting the water flow velocity into a water flow environment prediction model and a water flow parameter adjustment model respectively; 3) outputting short-term water flow velocity variation trend and long-term water flow velocity variation trend; 4) determining the current control strategy; 5) outputting the optimal water wing movement parameter combination; and 6) adjusting the water wing movement parameter combination to be optimal. The present application can actively adjust multiple movement parameters of the water wing, and has the advantages of strong adaptability, high energy capture efficiency and the like.

Description

Technical Field

[0001] This invention relates to the field of hydroelectric power generation technology, and in particular to a series hydrofoil motion system and motion control method. Background Technology

[0002] With the growth of global energy demand and increasing environmental pressure, the development of clean and renewable hydrodynamic energy (such as tidal energy and river energy) has become a research hotspot in the international energy field. Among the many hydrodynamic energy capture technologies, oscillating hydrofoil power generation devices are gradually showing superior application prospects compared to traditional rotary turbines due to their good adaptability to low-flow-velocity environments, low operating noise, and friendliness to aquatic organisms. In particular, the series oscillating hydrofoil system, by arranging two hydrofoils in the front and rear directions along the water flow direction, utilizes the wake vortex generated by the front hydrofoil to interact with the rear hydrofoil, which can break through the Betz limit of a single hydrofoil and achieve higher energy capture efficiency in a limited space.

[0003] However, despite the theoretically high energy conversion potential of tandem hydrofoils, existing tandem hydrofoil power generation systems still have the following significant drawbacks and shortcomings in practical engineering applications:

[0004] 1) Most existing hydrofoil control technologies focus only on controlling a single degree of freedom (usually lateral movement) of the hydrofoil, and often employ passive spring-damping systems. This passive approach relies on the impact force of the water flow to drive hydrofoil oscillation. In this case, the lateral movement of the hydrofoil is mainly maintained by the physical characteristics of the mechanical structure. Due to the nonlinearity and time-varying nature of the water flow force, passive systems struggle to maintain optimal motion amplitude and phase difference across the entire flow velocity range. Especially at low flow velocities, passive systems often fail to start oscillating due to insufficient starting torque; while at high flow velocities, excessive oscillation can easily lead to structural fatigue damage. The lack of active real-time adjustment of lateral velocity, lateral acceleration, angular velocity, and angular acceleration prevents the device from achieving true dynamic optimization.

[0005] 2) To achieve complex series motion, existing technologies often employ complex linkage mechanisms or hydraulic transmission systems. These systems not only suffer from high mechanical losses, reducing the overall water-to-electricity conversion efficiency, but are also prone to problems such as seal failure, corrosion, and silt abrasion in harsh underwater environments. Furthermore, traditional complex transmission mechanisms often occupy a large amount of radial space, which not only interferes with the flow field distribution around the hydrofoil (i.e., produces a significant blocking effect), but also increases the load on the supporting structure, posing dual challenges of cost and maintenance when deploying large-scale clusters.

[0006] 3) Existing control logic often treats water flow velocity, hydrofoil spacing, hydrofoil stiffness, and hydrofoil damping as isolated variables. For example, some existing technologies adjust the angle only based on real-time flow velocity, neglecting the compensating effect of flow velocity prediction on the lag of control commands. In series logic, the adjustment of the front hydrofoil immediately changes the inflow conditions of the rear hydrofoil. This strongly coupled nonlinear characteristic requires the control system to have extremely high collaborative processing capabilities. However, existing technologies lack an integrated module that can deeply bind water flow environment prediction with multi-degree-of-freedom execution, resulting in a slow response speed when facing sudden changes in flow velocity (such as water flow pulsations caused by swells or gusts), failing to achieve anticipatory adjustment, and thus causing fluctuations in energy harvesting efficiency.

[0007] 4) Traditional tandem hydrofoil devices often employ fixed mechanical connection structures, meaning the relative positions of the front and rear hydrofoils are often immutable after installation. However, fluid dynamics studies show that the evolution of the wake vortex (such as the Karman vortex street) generated by the front hydrofoil is closely related to flow velocity. Under different flow velocities, the optimal vortex-hydrofoil interaction distance shifts significantly. Existing fixed-spacing designs mean the device can only maintain high efficiency within a very narrow "design flow velocity" range; once the environmental flow velocity fluctuates, the rear hydrofoil often misses the optimal vortex-catching phase and may even enter turbulent flow, leading to a sharp drop in energy capture efficiency. This rigid spatial layout severely reduces the energy capture efficiency of the device in natural water bodies (such as those with tidal variations or seasonal runoff changes). Summary of the Invention

[0008] In view of the above-mentioned defects or improvement needs of the existing technology, the present invention provides a tandem hydrofoil motion system and motion control method, which can actively adjust multiple motion parameters of the hydrofoil and has the advantages of strong adaptability to complex flow velocity conditions and high energy capture efficiency in variable flow velocity environments.

[0009] To achieve the above objectives, according to one aspect of the present invention, a tandem hydrofoil motion system is provided, comprising two oscillating hydrofoil power generation devices arranged sequentially along the water flow direction. Each of the oscillating hydrofoil power generation devices includes a support frame, a generator, a first guide rail, a slider, and a hydrofoil. The generator and the first guide rail are both mounted on the support frame, and the slider is slidably mounted on the first guide rail. A connecting shaft is mounted on the upper end of the hydrofoil, and the connecting shaft is rotatably mounted on the slider and connected to the generator.

[0010] Each of the aforementioned oscillating hydrofoil power generation devices further includes a first motor, a second motor, and a first transmission mechanism. The first motor is mounted on the slider, and the output shaft of the first motor is connected to the connecting shaft to adjust the angular velocity and angular acceleration of the hydrofoil. The second motor and the first transmission mechanism are respectively mounted on the support frame, and the output shaft of the second motor is connected to the slider through the first transmission mechanism to adjust the lateral speed and lateral acceleration of the hydrofoil.

[0011] The tandem hydrofoil motion system also includes a base, a second guide rail, a third motor, and a second transmission mechanism. Multiple second guide rails are respectively installed on the base and are parallel to each other. The second guide rails are perpendicular to the first guide rails. Each second guide rail is slidably connected to each support frame. The third motor and the second transmission mechanism are both installed on the base. The third motor is connected to the support frame through the second transmission mechanism to adjust the distance between the two support frames in the direction parallel to the second guide rail, thereby adjusting the distance between the two hydrofoils in the direction parallel to the second guide rail.

[0012] The first motor, the second motor, and the third motor are electrically connected to the motion control module.

[0013] The tandem hydrofoil motion system also includes a water flow velocity sensor, which is electrically connected to the motion control module. The motion control module controls the rotation of the first motor, the second motor, and the third motor based on the water flow velocity feedback from the water flow velocity sensor.

[0014] Preferably, both the first transmission mechanism and the second transmission mechanism are lead screw mechanisms.

[0015] Preferably, the motion control module has a water flow parameter adjustment model, which adopts a residual multilayer perceptron architecture and is trained and generated based on previous actual test data, and is used to generate the optimal combination of motion parameters for the hydrofoil according to the current water flow velocity;

[0016] The preliminary experimental data included water flow velocity, the combination of motion parameters of the hydrofoil, and the energy capture efficiency of the hydrofoil. The combination of motion parameters of the hydrofoil included the lateral stiffness K of the hydrofoil. h Rotational stiffness K θ Lateral damping D h Rotational damping D θ And the distance L between the two hydrofoils along the direction perpendicular to the first guide rail.

[0017] Preferably, the flow parameter adjustment model includes an input layer, a residual block stacking layer, and an output layer, and:

[0018] The input layer is used to acquire the flow velocity data detected by the water flow velocity sensor in real time, and to map it to a high-dimensional feature space of a preset dimension composed of hidden layer nodes through linear transformation.

[0019] The residual block stacking layer consists of a preset number of serially stacked residual blocks. The residual connection mechanism performs nonlinear transformation on the water flow velocity characteristics in the high-dimensional feature space to establish the mapping relationship between water flow velocity fluctuations and the combination of motion parameters of the hydrofoil.

[0020] The output layer is used to linearly combine the high-dimensional feature vectors of the preset dimension using a fully connected weight matrix, mapping them from the feature space to a preliminary parameter space composed of the motion parameters of the hydrofoil, and performing numerical range restriction processing on the mapped preliminary parameter space to ensure that the output value conforms to the adjustment range of the first motor, the second motor and the third motor. Then, the dimensionless predicted value is restored to the actual motion parameter value with physical units, and the optimal hydrofoil motion parameter combination corresponding to the current flow velocity is output.

[0021] Preferably, the motion control module includes a water flow environment prediction model, and the water flow environment prediction model adopts a bidirectional long short-term memory neural network structure;

[0022] The water flow environment prediction model comprises an input layer, a hidden layer, a fully connected layer, and an output layer connected in sequence, wherein:

[0023] The input layer is used to receive water flow velocity data monitored by the water flow velocity sensor in real time.

[0024] The hidden layer is used to extract the depth time-series features of the water flow velocity data through the forward network and the backward network, and to concatenate the time-series feature vectors obtained by bidirectional calculation to generate a high-dimensional time-series feature vector characterizing the evolution law of the flow field.

[0025] The fully connected layer is used to receive the spliced ​​temporal feature vector output by the hidden layer, and to perform feature extraction and nonlinear compression on the spliced ​​temporal feature vector, so as to realize the projection of temporal information from the high-dimensional feature space to the prediction target space.

[0026] The output layer is used to perform multi-scale regression calculation on the spliced ​​time-series feature vector after feature extraction and nonlinear compression of the fully connected layer, convert the spliced ​​time-series feature vector into a value that conforms to physical dimensions, and output the predicted short-term and long-term water flow velocity change trends respectively.

[0027] Preferably, after acquiring a water flow velocity data stream composed of multiple water flow velocity data, the water flow environment prediction model uses Fourier transform to preprocess the water flow velocity data stream to perform feature reduction on the water flow velocity data stream, thereby reducing the impact of the uncertainty of the velocity data on the prediction results.

[0028] The motion control module provides a combination of motion parameters for the hydrofoil based on the short-term water flow velocity change trend and the water flow parameter adjustment model, in order to cope with sudden changes in water flow velocity in the short term.

[0029] The motion control module selects one of a variety of preset control strategies to execute based on the long-term trend of water flow environment changes, in order to cope with long-term changes in water flow velocity environment.

[0030] Among them, the duration of short-term events is no greater than the preset duration T1, and the duration of long-term events is no less than the preset duration T2.

[0031] Preferably, the multiple control strategies include a low-speed stable strategy, a low-speed unstable strategy, a high-speed stable strategy, and a high-speed unstable strategy, and:

[0032] Low-speed stability means that the flow velocity is not greater than the preset velocity V. 预设 And the flow field is stable;

[0033] Low-speed instability refers to a flow velocity that is not greater than the preset velocity V. 预设 Furthermore, the flow field is not stable;

[0034] High-speed stability refers to a flow velocity greater than the preset velocity V. 预设 And the flow field is stable;

[0035] High-speed instability refers to a flow velocity greater than the preset velocity V. 预设 Furthermore, the flow field is not stable.

[0036] Preferably, the operation of each hydrofoil satisfies the following conditions:

[0037] ;

[0038] ;

[0039] ;

[0040] in,

[0041] F represents the lift generated by the hydrofoil;

[0042] M is the water flow torque generated by the hydrofoil and acting on the connecting shaft;

[0043] S is the static moment of the hydrofoil;

[0044] m is the total mass of the lateral movement component, which includes a hydrofoil and components that move laterally along with the hydrofoil;

[0045] I θ Let be the moment of inertia of the hydrofoil;

[0046] , , These are the lateral displacement, lateral velocity, and lateral acceleration of the hydrofoil, respectively. , and These are the rotation angle, angular velocity, and angular acceleration of the hydrofoil, respectively.

[0047] K h The lateral motion stiffness of the hydrofoil;

[0048] K θ The rotational stiffness of the hydrofoil;

[0049] D h Damping for the lateral movement of the hydrofoil;

[0050] D θ Damping for the rotational motion of the hydrofoil;

[0051] x θ This is the distance from the hydrofoil's center of mass to the connecting shaft;

[0052] m θ The total mass of the hydrofoil, connecting shaft, and first motor.

[0053] According to another aspect of the present invention, a motion control method for the aforementioned tandem hydrofoil motion system is also provided, comprising the following steps:

[0054] 1) The water flow velocity sensor detects the current water flow velocity in real time and transmits the water flow velocity to the motion control module;

[0055] 2) The motion control module inputs the water flow velocity into the water flow environment prediction model and the water flow parameter adjustment model, respectively;

[0056] 3) The water flow environment prediction model outputs short-term and long-term water flow velocity change trends based on the water flow velocity data received within a set time period.

[0057] 4) The motion control module determines the current control strategy based on the long-term water flow velocity change trend;

[0058] 5) The water flow parameter adjustment model outputs the optimal combination of hydrofoil motion parameters corresponding to the current water flow velocity environment based on the current water flow velocity and the short-term water flow velocity change trend;

[0059] 6) The motion control module controls the operation of the first motor, the second motor and the third motor based on the current control strategy and the optimal combination of hydrofoil motion parameters, thereby adjusting the combination of hydrofoil motion parameters to the optimal.

[0060] In summary, compared with the prior art, the above-described technical solutions conceived by this invention can achieve the following beneficial effects:

[0061] 1) The tandem hydrofoil motion system of this invention constructs a dual hydrofoil working system with flow field coupling characteristics by sequentially arranging two oscillating hydrofoil power generation devices in the water flow direction. This tandem technical solution can fully utilize the wake flow field generated by the front oscillating hydrofoil power generation device during its work process at the fluid dynamics level. The oscillating hydrofoil power generation device performs controlled forced oscillations in the water flow, and the hydrofoil motion parameters are optimized in real time through the active power intervention of the first and second motors. Through the sequential arrangement of the front and rear oscillating hydrofoil power generation devices, the rear hydrofoil can enter the shedding vortex region generated by the front hydrofoil. This structural layout allows the rear hydrofoil not only to capture kinetic energy in the main flow channel but also to achieve deep kinetic energy extraction through secondary intervention and capture of the wake vortex of the front hydrofoil. Within a limited water projection area, the tandem layout significantly improves the energy capture power density per unit flow cross-section, giving the entire tandem hydrofoil motion system a clear systemic advantage in terms of space utilization and overall power generation efficiency.

[0062] 2) In the tandem hydrofoil motion system of this invention, the output shaft of the first motor is connected to the connecting shaft, realizing active power intervention for the rotational degree of freedom of the hydrofoil. The first motor provides precise dynamic control of the hydrofoil's angle of attack. During the hydrofoil's power generation process, the angle of the hydrofoil relative to the incoming flow directly determines the magnitude of the lift. By actively adjusting the angular velocity and angular acceleration of the connecting shaft by the first motor, the system can compensate or correct the hydrofoil's pitch angle in real time at any moment within each oscillation cycle, based on the instantaneous flow velocity and lateral position.

[0063] This active control enables rapid switching of the hydrofoil angle at the stroke inversion point, reducing energy loss during hydrodynamic reversal. By controlling angular acceleration, the hydrofoil can be maintained within the optimal angle of attack range throughout the entire power stroke, effectively suppressing dynamic stall caused by excessive angle of attack, while also avoiding insufficient lift due to insufficient angle of attack. This active intervention in angular velocity and angular acceleration transforms the hydrofoil's rotational motion from a traditional passive response to active dynamic optimization, enhancing the hydrofoil's lift maintenance capability in complex flow fields.

[0064] 3) In the tandem hydrofoil motion system of the present invention, the second motor is mounted on the support frame and connected to the slider through the first transmission mechanism. This gives the system the ability to control the lateral velocity and acceleration of the hydrofoil, achieving precise shaping of the lateral motion trajectory. During the water flow energy capture process, matching the lateral amplitude and velocity of the hydrofoil with the load of the generator is crucial.

[0065] The second motor can actively provide auxiliary thrust or resistance to the slider, thereby adjusting the lateral speed of the slider and the hydrofoil. In low-flow-rate environments, the second motor can provide initial motion compensation, assisting the system in overcoming static friction torque and initial fluid resistance to achieve rapid start-up. In high-flow-rate or variable-flow-rate environments, by adjusting the lateral acceleration, the smoothness of the slider's movement on the first guide rail can be optimized, mitigating instantaneous impact loads on the mechanical structure. Furthermore, the coordinated operation of the second motor and the first motor enables a high degree of phase matching between the lateral and rotational motions, thereby maximizing the work area within a single cycle.

[0066] 4) The tandem hydrofoil motion system of this invention introduces a base, a second guide rail, a third motor, and a second transmission mechanism. The second guide rail is perpendicular to the first guide rail, allowing the system to adjust the distance between the two hydrofoils along a direction perpendicular to the first guide rail. This is a key technological improvement for the characteristics of tandem flow fields. As the wake vortex (such as a Karman vortex street) generated by the leading hydrofoil propagates downstream, its physical form and energy center change spatially with varying flow velocity. By driving the support frame to slide on the second guide rail via the third motor, the system can adjust the relative distance between the two oscillating hydrofoil power generation devices according to real-time operating conditions, ensuring that the trailing hydrofoil is always at the optimal capture phase point of the trailing vortex of the leading hydrofoil. This dynamic optimization of spatial position solves the problem of a sharp drop in efficiency under variable flow velocity conditions for fixed-gap structures. By precisely adjusting the distance, positive constructive interference between the trailing hydrofoil and the trailing vortex of the leading hydrofoil can be induced, further enhancing the torque output of the trailing hydrofoil using vortex-induced lift, thereby achieving consistent performance of the entire tandem system across the entire flow velocity range.

[0067] 5) The tandem hydrofoil motion system of the present invention, through the introduction of three motors and the integration of motion control modules, constructs a fully active control dynamic platform, which greatly enhances the device's adaptability to unsteady water flow environments in nature.

[0068] In natural water bodies, the direction and magnitude of flow velocity often fluctuate randomly. This invention does not rely on a single physical structural characteristic (such as fixed spring stiffness), but rather dynamically simulates optimal dynamic damping and stiffness characteristics through electronic control. This high degree of programmability and flexibility allows the system to automatically switch to the optimal motion combination based on real-time monitored flow changes. Even under extreme conditions of extremely low or high flow velocities, the system can maintain a stable oscillation frequency and safe operating amplitude through the intervention of the motor. This provides a crucial mechanical and control foundation for the large-scale commercial application of hydroelectric power generation devices, ensuring higher operational reliability and cumulative power generation throughout their entire lifecycle. Attached Figure Description

[0069] Figure 1 This is a three-dimensional schematic diagram of the present invention;

[0070] Figure 2 This is a perspective view of one of the oscillating hydrofoil power generation devices of the present invention;

[0071] Figure 3 This is a flowchart of the control method of the present invention;

[0072] Figure 4 This is a flow velocity-time variation curve obtained by the water flow velocity sensor according to the present invention.

[0073] In all the accompanying drawings, the same reference numerals denote the same technical features, specifically:

[0074] 100. Oscillating hydrofoil generator; 101. Support frame; 102. Generator; 103. First guide rail; 104. Slider; 105. Hydrofoil; 106. First motor; 107. Second motor; 108. First transmission mechanism; 109. Connecting shaft; 200. Base; 300. Second guide rail; 400. Third motor; 500. Motion control module; 600. Water flow velocity sensor. Detailed Implementation

[0075] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0076] Referring to the accompanying drawings, the tandem hydrofoil motion system includes two oscillating hydrofoil power generation devices 100 arranged sequentially along the water flow direction. Each oscillating hydrofoil power generation device 100 includes a support frame 101, a generator 102, a first guide rail 103, a slider 104, and a hydrofoil 105. The generator 102 and the first guide rail 103 are both mounted on the support frame 101. The slider 104 is slidably mounted on the first guide rail 103. A connecting shaft 109 is mounted on the upper end of the hydrofoil 105. The connecting shaft 109 is rotatably mounted on the slider 104 and connected to the generator 102.

[0077] Each of the aforementioned oscillating hydrofoil power generation devices 100 further includes a first motor 106, a second motor 107, and a first transmission mechanism 108. The first motor 106 is mounted on the slider 104, and the first motor 106 can be directly or indirectly mounted on the slider 104. The output shaft of the first motor 106 is connected to the connecting shaft 109 to adjust the angular velocity and angular acceleration of the hydrofoil 105. The second motor 107 and the first transmission mechanism 108 are respectively mounted on the support frame 101. The output shaft of the second motor 107 is connected to the slider 104 through the first transmission mechanism 108 to adjust the lateral speed and lateral acceleration of the hydrofoil 105.

[0078] Since generator 102 and first motor 106 are respectively connected to connecting shaft 109, if one of generator 102 and first motor 106 is directly connected to connecting shaft 109, the other can be indirectly connected to connecting shaft 109 through an intermediate transmission device, or they can be indirectly connected to connecting shaft 109 through intermediate transmission devices respectively. For example, connecting shaft 109 can be mounted on sliding frame through bearings. Sliding frame is fixedly connected to slider 104. First motor 106 and synchronous belt of synchronous belt drive mechanism are mounted on sliding frame. Synchronous belt rotates when sliding frame moves laterally, driving synchronous belt pulley to rotate. Synchronous belt pulley of synchronous belt drive mechanism is connected to the rotating shaft of generator 102, which can drive generator 102 to rotate to generate electricity.

[0079] The tandem hydrofoil motion system further includes a base 200, a second guide rail 300, a third motor 400, and a second transmission mechanism. Multiple second guide rails 300 are respectively mounted on the base 200 and are parallel to each other. The second guide rails 300 are perpendicular to the first guide rail 103. Each second guide rail 300 is slidably connected to each support frame 101. The third motor 400 and the second transmission mechanism are both mounted on the base 200. The third motor 400 is connected to the support frame 101 through the second transmission mechanism to adjust the distance between the two support frames 101 in a direction parallel to the second guide rail 300, thereby adjusting the distance between the two hydrofoils 105 in a direction parallel to the second guide rail 300.

[0080] The first motor 106, the second motor 107, and the third motor 400 are electrically connected to the motion control module 500. The motion control module 500 can be mounted on the base 200.

[0081] The tandem hydrofoil motion system also includes a water flow velocity sensor 600, which is electrically connected to the motion control module 500. The motion control module 500 controls the rotation of the first motor 106, the second motor 107, and the third motor 400 based on the water flow velocity feedback from the water flow velocity sensor 600. The water flow velocity sensor 600 can be an existing Doppler current meter, electromagnetic current meter, acoustic Doppler current profiler, ultrasonic time-of-flight current meter, etc.

[0082] This invention adds a water flow velocity sensor 600 to the tandem hydrofoil motion system and electrically connects it to the motion control module 500, thus establishing a real-time external environment perception dimension for the entire system. This transforms the tandem hydrofoil motion system from a passive system relying solely on physical structural responses into an intelligent system capable of actively acquiring dynamic flow field parameters. The water flow velocity sensor 600 can monitor the instantaneous flow velocity of the current water flow at high frequency and transmit this flow velocity data to the motion control module 500 in real time. This real-time acquisition of information provides objective and accurate data support for subsequent control commands, ensuring that the system can perceive changes in the water flow environment at the first moment, laying an informatics foundation for achieving efficient energy harvesting.

[0083] The motion control module 500 controls the rotation of the first motor 106 based on the water flow velocity feedback from the water flow velocity sensor 600. Since the first motor 106 directly adjusts the angular velocity and angular acceleration of the hydrofoil 105, the system can dynamically compensate for the pitch angle (i.e., angle of attack) of the hydrofoil 105 according to real-time flow velocity fluctuations. Under different flow velocities, the pitch motion pattern required for the hydrofoil 105 to obtain optimal lift varies significantly. Through feedback control, the motion control module 500 can precisely drive the first motor 106, ensuring that the hydrofoil 105 maintains its optimal rotational state matching the real-time flow velocity at each stage of its oscillation stroke. This effectively avoids hydrofoil stall or a decrease in lift coefficient due to sudden changes in flow velocity, thereby maintaining the stability of the rotational degree of freedom during energy conversion.

[0084] The second motor 107 is connected to the slider 104 via the first transmission mechanism 108, and is responsible for adjusting the lateral speed and acceleration of the hydrofoil 105. When the water flow velocity sensor 600 detects an increase or decrease in flow velocity, the motion control module 500 can adjust the output torque and speed of the second motor 107 in real time, thereby changing the lateral movement dynamic characteristics of the hydrofoil 105 on the first guide rail 103. This dynamic adjustment achieves a real-time balance between the inertial force, damping force, and fluid lift of the hydrofoil 105 during lateral movement, ensuring that the oscillation frequency and amplitude of the hydrofoil 105 are always within the optimal range that matches the current kinetic energy density of the flow field, significantly improving the conversion rate of mechanical energy extracted from the fluid by the oscillating hydrofoil power generation device 100.

[0085] The system can automatically drive the third motor 400 to adjust the position of the support frame 101 on the second guide rail 300 through the second transmission mechanism based on the real-time flow velocity provided by the water flow velocity sensor 600. Since the evolution distance of the wake vortex (such as the karman vortex street) generated by the foreplane is proportional to the flow velocity, this spacing adjustment based on flow velocity feedback ensures that the rearplane can always be accurately positioned within the energy convergence area of ​​the foreplane's detached vortex. This adaptive optimization of the spatial layout allows the rear hydrofoil 105 to effectively utilize the induced flow field generated by the front hydrofoil 105, maintaining efficient "vortex-hydrofoil" interaction even in variable flow velocity environments, thus improving the overall energy capture capability of the entire series system.

[0086] This invention integrates the first motor 106, the second motor 107, and the third motor 400 into a closed-loop control framework based on flow velocity feedback, achieving deep coordination of multiple motion parameters. Through unified scheduling by the motion control module 500, the system no longer adjusts the state of a single motor in isolation, but rather coordinates and adjusts angular velocity, angular acceleration, lateral velocity, lateral acceleration, and the spacing between the hydrofoils 105 as a whole motion combination. The system exhibits strong dynamic response and anti-interference capabilities when facing water flow pulsations caused by swells and gusts common in natural water bodies. Through a pre-set control strategy, the system can offset the adverse impacts of flow velocity fluctuations through the rapid coordinated action of the three motors, ensuring that the tandem hydrofoil motion system maintains a long-term, stable, and efficient operating state even in complex and variable flow velocity environments.

[0087] When water flows over the hydrofoil 105, under the action of fluid-structure interaction, the front and rear wings generate periodic rotational motion and reciprocate lateral motion in the direction perpendicular to the water flow. The lateral displacement and rotation angle of the hydrofoil 105 change with time in an approximately sinusoidal manner, and its motion is controlled by multiple motion parameters (such as stiffness K, damping D, etc.).

[0088] This invention achieves significant technical effects in terms of deep energy capture in the flow field, precise shaping of motion parameters, dynamic matching of vortex phase, optimized distribution of system load, and multi-variable coordinated control by organically combining a series layout, multi-motor active drive, multi-stage guide rail support, and integrated control module. It greatly improves the dynamic performance and energy conversion efficiency of the series hydrofoil motion system.

[0089] The motion control module 500 controls the motion parameters of the hydrofoil 105 in two degrees of freedom (lateral and rotational): 1) For the rotational degree of freedom (i.e., the hydrofoil 105 rotates around the connecting shaft 109), the first motor 106 adjusts the angular velocity, angular acceleration, etc., by changing the torque input to the connecting shaft 109 on the hydrofoil 105, thereby changing the rotational motion parameters; 2) For the lateral degree of freedom (the lateral direction is parallel to the length direction of the first guide rail 103, the lateral movement of the hydrofoil 105 is the hydrofoil 105 following the slider 104 along the first guide rail 103), (Translation of the length direction of track 103), the second motor 107, through its first transmission mechanism, changes the torque into a lateral force input to the connecting shaft 109 (the direction of the lateral force is collinear with the lateral movement direction of the hydrofoil 105). By increasing or decreasing the lateral force, the speed and acceleration of the hydrofoil 105 in lateral movement are adjusted, thereby changing the lateral movement parameters; 3) In view of the self-coupling characteristics in the fully passive motion process of the hydrofoil 105, a joint and coordinated control of the rotating motor and the lateral component is adopted to incorporate the self-coupling characteristics of the hydrofoil 105 in the motion process into the control law. The motion control module 500 incorporates the coupling effect in the motion of the hydrofoil 105 into the control method. By changing the input of the rotating first motor 106 and the second motor 107, the optimal torque and force are output to the connecting shaft 109 on the hydrofoil 105 and the slider 104 connected to the connecting shaft 109, thereby adjusting the speed and acceleration (angle and angular acceleration) of the hydrofoil 105 in motion, and achieving the effect of adjusting the combination of motion parameters of the hydrofoil 105 to the optimal level.

[0090] Furthermore, the operation of each hydrofoil 105 satisfies the following conditions (the self-coupling characteristics of the hydrofoil 105):

[0091] ;

[0092] ;

[0093] ;

[0094] in,

[0095] F represents the lift generated by the hydrofoil 105;

[0096] M is the water flow torque generated by the hydrofoil 105 acting on the connecting shaft 109;

[0097] S is the static moment of the hydrofoil, 105.

[0098] m is the total mass of the lateral movement component, that is, the total mass of the components that participate in the lateral reciprocating motion during the movement of the hydrofoil 105. The lateral movement component includes the hydrofoil 105 and the components that reciprocate laterally with the hydrofoil 105. The components that reciprocate laterally with the hydrofoil 105 include a connecting shaft, a slider 104, a sliding frame mounted on the slider, and a first motor 106 mounted on the sliding frame. If a synchronous belt for driving the generator 102 to rotate and generate electricity is also mounted on the sliding frame, the lateral movement component also includes the synchronous belt.

[0099] I θ Let be the moment of inertia of hydrofoil 105, which is a known value;

[0100] , , These are the lateral displacement, lateral velocity, and lateral acceleration of hydrofoil 105. , and These are the rotation angle, angular velocity, and angular acceleration of hydrofoil 105, respectively.

[0101] K h The lateral motion stiffness of hydrofoil 105;

[0102] K θ The rotational stiffness of hydrofoil 105;

[0103] D h Damping for the lateral motion of hydrofoil 105;

[0104] D θ Damping for the rotational motion of hydrofoil 105;

[0105] x θ The distance from the center of mass of hydrofoil 105 to connecting shaft 109 is a known value;

[0106] m θ The total mass of the rotating module that participates in the rotation of the hydrofoil 105 is defined as follows: the rotating module includes the hydrofoil 105, a connecting shaft, and a first motor 106.

[0107] This invention establishes a dynamic model of the operation of each hydrofoil 105 in a tandem hydrofoil motion system through a set of mathematical equations, and accurately quantifies the balance relationship between fluid dynamics and mechanical system motion.

[0108] The characteristics of lift balance are as follows:

[0109] ;

[0110] The above formula accurately characterizes the logical relationship between the water flow lift F generated by the hydrofoil 105 and the force components of the mechanical system. The function of this water flow lift F is that it deeply correlates the lateral acceleration, lateral velocity, and lateral displacement of the hydrofoil 105. This allows the system to accurately calculate the motion state of the hydrofoil 105 on the first guide rail 103 under specific flow field loads, providing a precise predictive basis at the dynamic level for achieving efficient energy harvesting.

[0111] The torque balance equations are characterized as follows:

[0112] ;

[0113] The rotational response is caused by the water flow torque M generated by the hydrofoil 105 acting on the connecting shaft 109. This is achieved by introducing the rotational inertia I of the hydrofoil 105. θ The function of this equation is to quantify the inertial drag and energy conversion process when the hydrofoil 105 rotates around the connecting shaft 109. This provides a clear control target for the first motor 106 to actively adjust the angular velocity and angular acceleration of the hydrofoil 105, ensuring dynamic adaptation between the rotation angle and the flow direction.

[0114] This invention also introduces the static moment S of the hydrofoil 105 and its definition: The static moment S of hydrofoil 105 is the core term connecting lateral and rotational motion. It characterizes the self-coupling properties of hydrofoil 105 during its motion; that is, changes in angular acceleration directly affect the force balance in the lateral direction, and vice versa. This is achieved by defining x... θ and m θ The system establishes the magnitude of the self-coupling effect, which allows the motion control module 500 to utilize the physical characteristics of the hydrofoil 105 itself (rather than relying entirely on external power) to achieve phase coordination of lateral movement and rotation. This physical parameter-based coupling control enables the tandem hydrofoil motion system to exchange energy more smoothly during reciprocating oscillations, improving the overall energy harvesting efficiency of the system.

[0115] All parameters listed in this invention refer to the physical gripper that is intelligently controlled by the motion control module 500. Although the system involves fixed mechanical parameters, through the active intervention of the first motor 106 and the second motor 107, the system can simulate different lateral motion stiffness K of the hydrofoil 105. h The rotational stiffness K of hydrofoil 105 θ And the lateral motion damping D of hydrofoil 105 h The rotational motion damping D of the hydrofoil 105 θIt provides a standardized parameter interface for the optimal hydrofoil motion parameter combination generated by the flow parameter adjustment model. Based on these equations, the motion control module 500 can transform the abstract optimization objective into specific torque commands for the motor. For example, when the system detects a change in flow velocity, it can use the equations to calculate the compensation force required to maintain the optimal hydrofoil motion parameter combination, and then control the operation of the motor to adjust the hydrofoil 105 to the optimal operating state.

[0116] Furthermore, both the first transmission mechanism 108 and the second transmission mechanism are lead screw mechanisms.

[0117] Since the second motor 107 and the third motor 400 output angular displacement, the rotational angle of the motors can be converted into a very small linear displacement through the threaded transmission of the lead screw mechanism. When controlling the lateral speed and acceleration of the slider 104, the lead screw mechanism can provide micron-level positioning accuracy. This high-resolution execution ensures that the optimized parameter commands issued by the motion control module 500 can be executed precisely, resulting in a smoother reciprocating trajectory of the hydrofoil 105 perpendicular to the water flow direction that conforms to the theoretically calculated optimal curve.

[0118] The lead screw mechanism (especially when used as a precision transmission component) exhibits high mechanical transmission efficiency in the tandem hydrofoil motion system. The second motor 107 is connected to the slider 104 via the first transmission mechanism 108, and the third motor 400 is connected to the support frame 101 via the second transmission mechanism. Due to the low coefficient of friction and uniform force distribution of the lead screw mechanism, the torque output by the motor can be converted into driving load with extremely low energy loss. This is crucial for the power generation system, as it minimizes parasitic power consumption within the system, thereby indirectly improving the "water-to-electricity" conversion efficiency of the entire tandem hydrofoil motion system. During long-term underwater operation, this low-loss characteristic also manifests as low heat generation, helping to maintain the lubrication performance and service life of mechanical components.

[0119] As a transmission element, the lead screw mechanism has excellent motion reversibility. Combined with the closed-loop control logic of the motion control module 500, it enables real-time bidirectional adjustment of the lateral speed, lateral acceleration, and the spacing between the hydrofoils 105. Simultaneously, based on the physical characteristics of the lead screw thread, when the system is powered off or in a non-operating state, this mechanism provides a certain degree of static self-locking or high-friction damping to prevent the hydrofoils 105 or support frame 101 from slipping uncontrollably due to their own weight or water flow disturbances. This provides a basic mechanical protection mechanism for the tandem hydrofoil motion system, enhancing the overall operational safety and structural integrity of the device in complex and variable natural aquatic environments.

[0120] Furthermore, the motion control module 500 has a water flow parameter adjustment model, which adopts a residual multilayer perceptron (ResMLP) architecture and is trained and generated based on previous actual test data. This model is used to generate the optimal combination of motion parameters for the hydrofoil 105 according to the current water flow velocity.

[0121] The preliminary experimental data included water flow velocity, the combination of motion parameters of hydrofoil 105, and the energy capture efficiency of hydrofoil 105. The combination of motion parameters of hydrofoil 105 included the lateral stiffness K of hydrofoil 105. h Rotational stiffness K θ Lateral damping D h Rotational damping D θ And the distance L between the two hydrofoils 105 along the direction perpendicular to the first guide rail 103.

[0122] Lateral stiffness K h The recovery capability of hydrofoil 105 in the vertical flow direction is determined by its rotational stiffness K. θ Drag is adjusted by varying the pitch angle of the hydrofoil 105; lateral damping D h Dissipate kinetic energy during lateral movement to prevent oscillation overload; rotational damping D θ Optimize the smoothness of the hydrofoil 105 during rotation; adjust the spatial phase of the front and rear wings of the two hydrofoils 105 along the distance L perpendicular to the first guide rail 103 to achieve optimal "vortex-wing" interaction.

[0123] The motion control module 500 employs a residual multilayer perceptron architecture for its flow parameter adjustment model. This leverages the cross-layer connections inherent in the residual multilayer perceptron architecture to effectively address the vanishing gradient problem during deep neural network training, thereby constructing a deeply nonlinear function mapping relationship. Through training based on prior experimental data, this flow parameter adjustment model generates a multidimensional flow characteristic vector representing the implicit correlation between flow velocity and the hydrofoil 105 response. The role of this multidimensional flow characteristic vector is that it transforms flow velocity information into a high-dimensional mathematical expression containing the evolutionary laws of the flow field, capturing subtle fluid dynamic features that are difficult to describe using traditional linear control methods, thus providing a deep logical basis for subsequent precise parameter adjustment.

[0124] This flow parameter adjustment model was generated based on prior experimental data, which includes water flow velocity, the combination of motion parameters for hydrofoil 105, and the energy harvesting efficiency of hydrofoil 105. This enables the tandem hydrofoil motion system to self-learn and self-optimize based on historical experience. By using the energy harvesting efficiency of hydrofoil 105 as the core dimension of the training objective function, the flow parameter adjustment model can identify which combinations of motion parameters for hydrofoil 105 can achieve optimal energy conversion at specific water flow velocities. This data-driven decision-making mechanism allows the system to derive the theoretically optimal operating scheme through real-time input of current flow velocity data during actual operation, ensuring that the system's work capacity remains at a high level in various dynamic flow fields.

[0125] This invention clarifies that the combination of motion parameters for the hydrofoil 105 includes the lateral stiffness K of the hydrofoil 105. h Rotational stiffness K θ Lateral damping D h Rotational damping D θ And the distance L between the two hydrofoils 105 along the direction perpendicular to the first guide rail 103. This achieves comprehensive integrated control of kinematic and structural parameters: 1) Optimized coordination of stiffness and damping: The flow parameter adjustment model can simultaneously provide the lateral stiffness K of the hydrofoil 105. h Rotational stiffness K θ Lateral damping D h Rotational damping D θ This allows the force characteristics of the hydrofoil 105 during reciprocating oscillation to be highly compatible with the fluid lift, which is beneficial for maintaining stable amplitude and frequency. 2) Coupling adjustment of spatial and dynamic parameters: The distance L between the two hydrofoils 105 along the direction perpendicular to the first guide rail 103 is included in the output of the water flow parameter adjustment model. Its function is to achieve deep coupling between the "vortex-wing" interaction distance and the motion characteristics of the hydrofoil 105 itself. The water flow parameter adjustment model can calculate the propagation position of the detached vortex of the forewing at a specific flow velocity, and accordingly adjust the distance L between the two hydrofoils 105 along the direction perpendicular to the first guide rail 103 and the damping and stiffness state of the hydrofoils 105, so that the rear wing can intercept the energy of the forewing wake vortex with optimal dynamic attitude.

[0126] By abstracting and summarizing a large amount of previous actual experimental data through a residual multilayer perceptron architecture, the flow parameter adjustment model can generalize to handle complex flow conditions. When the flow environment fluctuates, the flow parameter adjustment model can quickly perform nonlinear calculations through its internal fully connected layers and residual blocks, outputting a combination of motion parameters for the hydrofoil 105 that matches the current instantaneous flow velocity. This rapid adjustment means that the system no longer relies on fixed mechanical characteristics, but achieves dynamic reconstruction of the operating state through a digital model. This not only improves the system's versatility under different river or sea conditions, but also enables the tandem hydrofoil motion system to maintain the continuity and accuracy of parameter adjustment when facing unsteady flow fields, achieving efficient operation of the device throughout its entire life cycle.

[0127] A residual multilayer perceptron architecture is used to establish a flow parameter regulation model, which has low computational overhead during the inference phase. This feature enables the motion control module 500 to achieve real-time online parameter updates on the embedded control device. Compared with complex fluid numerical simulation methods, this flow parameter regulation model based on a deep learning architecture can complete the mapping transformation from flow velocity input to the combined output of motion parameters of the hydrofoil 105 in a very short time, meeting the stringent real-time control requirements of underwater power generation equipment and providing an algorithmic foundation for the active, rapid, and precise regulation of the hydrofoil 105's motion parameters.

[0128] Furthermore, the water flow parameter adjustment model includes an input layer, a residual block stacking layer, and an output layer, and:

[0129] The input layer is used to acquire the flow velocity data detected by the water flow velocity sensor 600 in real time, and to map it to a high-dimensional feature space of a preset dimension composed of hidden layer nodes through linear transformation.

[0130] The residual block stacking layer consists of a preset number of serially stacked residual blocks. The residual connection mechanism performs nonlinear transformation on the water flow velocity characteristics in the high-dimensional feature space to establish the mapping relationship between water flow velocity fluctuations and the combination of motion parameters of the hydrofoil.

[0131] The output layer is used to linearly combine the high-dimensional feature vectors of the preset dimension using a fully connected weight matrix, mapping them from the feature space to a preliminary parameter space composed of the motion parameter combination of the hydrofoil 105, and performing numerical range restriction processing on the mapped preliminary parameter space to ensure that the output value conforms to the adjustment range of the first motor, the second motor and the third motor. Then, the dimensionless predicted value is restored to the actual motion parameter value with physical units, and the optimal hydrofoil motion parameter combination corresponding to the current flow velocity is output.

[0132] The input layer of this invention features the technical characteristic of acquiring flow velocity data detected by a water velocity sensor 600 in real time and mapping it to a high-dimensional feature space composed of hidden layer nodes. At the technical execution level, this design addresses the problem of insufficient information density in flow velocity data when representing complex water flow environments. By mapping the flow velocity data to a high-dimensional feature space, the system can provide a broader mathematical representation dimension for subsequent nonlinear processing, allowing minute flow velocity fluctuations to be transformed into significant feature differences in the high-dimensional space. This information enhancement provides a refined input foundation for the water flow parameter adjustment model to identify complex turbulence and pulsations in the water flow, improving the system's sensitivity and analytical capabilities to environmental inputs.

[0133] The residual block stacking layer of this invention serves as the main hidden layer, performing nonlinear processing on the flow velocity characteristics through multiple cascaded residual blocks. Its core technological role lies in optimizing gradient flow during deep learning. Due to the use of cascaded stacked residual blocks, the flow parameter adjustment model can maintain the stability of parameter transmission through cross-layer connection mechanisms, effectively mitigating the degradation tendency of deep networks during training. In the nonlinear processing, this architecture establishes the extremely complex mapping relationship between flow velocity and the motion parameters of the hydrofoil 105. Fluid-structure interaction in hydrodynamics typically exhibits highly nonlinear characteristics. Through layer-by-layer iteration and feature fusion of multiple residual blocks, the flow parameter adjustment model can abstract and remove noise interference in the flow field layer by layer, accurately extracting the core flow characteristics affecting energy harvesting efficiency. This deep feature processing capability ensures that the system can still derive control commands that conform to the laws of fluid mechanics when facing highly dynamic aquatic environments.

[0134] The output layer of this invention maps the processed high-dimensional features back to a parameter dimension composed of the motion parameters of the hydrofoil 105, and outputs the optimal combination of hydrofoil motion parameters corresponding to the current flow velocity. The technical advantage of this step is that it achieves a precise conversion from abstract features to specific mechanical actuators. Through reverse mapping, the output layer solves the optimal control logic calculated by the hidden layer into the specific lateral stiffness K of the hydrofoil 105. h Rotational stiffness K θ Lateral damping D h Rotational damping D θ The system also includes physical control quantities such as the distance L between the two hydrofoils 105 along the direction perpendicular to the first guide rail 103. This mapping ensures that the output parameter combination has an optimal solution at the mathematical level and high operability at the mechanical execution level. Based on this, the motion control module 500 can directly drive the first motor 106, the second motor 107, and the third motor 400 to coordinate their actions, so that the dynamic characteristics of the hydrofoils 105 are highly matched with the fluid characteristics.

[0135] The hierarchical architecture of "input layer - residual block stacking layer - output layer" constructed in this invention improves the real-time response speed and adaptability of the tandem hydrofoil motion system at the overall operational level. This layered processing mechanism decomposes complex computational tasks into efficiently executable tensor operations, reducing the system's latency from perception to execution. In practical applications, this architecture manifests itself in the following way: when the flow velocity data undergoes abrupt changes, the flow parameter adjustment model can quickly complete the forward inference process through its internally preset layer weights, updating the optimal hydrofoil motion parameter combination in real time. This end-to-end control based on a deep learning structure enables the device to maintain high energy capture performance under different geographical environments or seasonal flow velocity changes through adaptive parameter adjustment of the flow parameter adjustment model, thereby ensuring the technical robustness of the tandem hydrofoil motion system across the entire operating range.

[0136] Specifically, after manually labeling and normalizing the experimental data, a dataset, validation set, and test set were created. The hyperparameters in the flow parameter regulation model were set as follows: input dimension 1 (flow velocity), output dimension 5 (lateral stiffness K). h Rotational stiffness K θ Lateral damping D h and rotational damping D θ The model uses a hydrofoil spacing of 105 L, with a hidden layer dimension of 128 and 5 residual blocks. Dropout regularization is used to prevent overfitting. The flow parameter adjustment model consists of three layers: an input layer, a stacked residual block layer, and an output layer. The input layer maps the flow velocity data to high-dimensional features, the stacked residual blocks act as a fully connected layer, and the output layer maps the results back to the parameter dimension. The model uses a regression task (MSE) loss function and the Adam optimization algorithm as the optimizer. Only the optimal model weights from the validation set are saved during training. The model performance is evaluated using test set data, and the results are input into the numerical model. The numerical simulation results are compared with the physical model results to verify the model's actual performance.

[0137] This flow parameter adjustment model allows the use of real-time flow velocity parameters monitored by a flow velocity sensor or predicted flow velocity from a flow parameter prediction model as input. Based on this model, the optimal combination of motion parameters for the hydrofoil 105 under the current flow velocity conditions can be quickly obtained. The motion control module 500 inputs the obtained optimal combination of motion parameters to achieve autonomous adjustment of the motion parameters of the hydrofoil 105, ensuring that the hydrofoil 105 maintains optimal energy-harvesting motion parameters and improving the adaptability of the hydrofoil 105 to complex flow velocity environments and its overall energy-harvesting efficiency.

[0138] Furthermore, the motion control module 500 includes a water flow environment prediction model, and the water flow environment prediction model adopts a bidirectional long short-term memory neural network structure.

[0139] The water flow environment prediction model comprises an input layer, a hidden layer, a fully connected layer, and an output layer connected in sequence, wherein:

[0140] The input layer is used to receive water flow velocity data monitored by the water flow velocity sensor 600 in real time.

[0141] The hidden layer is used to extract the depth time-series features of the water flow velocity data through the forward network and the backward network, and to concatenate the time-series feature vectors obtained by bidirectional calculation to generate a high-dimensional time-series feature vector characterizing the evolution law of the flow field.

[0142] The fully connected layer is used to receive the spliced ​​temporal feature vector output by the hidden layer, and to perform feature extraction and nonlinear compression on the spliced ​​temporal feature vector, so as to realize the projection of temporal information from the high-dimensional feature space to the prediction target space.

[0143] The output layer is used to perform multi-scale regression calculation on the spliced ​​time-series feature vector after feature extraction and nonlinear compression of the fully connected layer, convert the spliced ​​time-series feature vector into a value that conforms to physical dimensions, and output the predicted short-term and long-term water flow velocity change trends respectively.

[0144] The motion control module 500 of this invention integrates a water flow environment prediction model based on a bidirectional long short-term memory neural network structure. This feature provides the logical foundation for the tandem hydrofoil motion system to transition from "real-time feedback control" to "predictive feedforward control." By performing bidirectional calculations on the evolution trend of the water flow environment over a future time period, the system can overcome the lag limitations of sensor monitoring. This proactive response allows the motion control module 500 to anticipate the pulsation of the flow field or the trend of flow velocity evolution, thereby issuing motor commands in advance. This ensures that the dynamic parameters of the hydrofoil 105 are precisely aligned with the actual flow field state in the time dimension, improving the smoothness of system control in unsteady flow fields.

[0145] The bidirectional long short-term memory neural network structure employed in the water flow environment prediction model plays a crucial role in the high-fidelity simulation of this complex nonlinear dynamic system. The hidden layers in the model extract temporal features of flow velocity data through forward and backward networks. Compared to unidirectional temporal processing, this bidirectional computation simultaneously captures the impact of historical flow field conditions on the current state and the logical constraints of potential future trends on the current state. This bidirectional feature extraction identifies complex disturbances with periodic or regular patterns in the water flow and transforms these abstract temporal features into quantitative predictions of the evolution trend of the water flow environment. For tandem hydrofoil motion systems, this means a more accurate prediction of the diffusion path and energy distribution of the foreplane wake vortex in the flow field, providing a data model-level guarantee for achieving high-precision "vortex-foil" interaction.

[0146] The water flow environment prediction model of this invention comprises an input layer, a hidden layer, a fully connected layer, and an output layer connected in sequence. The input layer serves as the data entry point, ensuring that the water flow velocity data obtained in real-time from the water flow velocity sensor 600 can be imported into the algorithm kernel in a lossless and high-frequency manner. The hidden layer, as the computational core, performs deep feature mining, utilizing the nonlinear mapping capabilities of the forward and backward networks to handle temporal fluctuations. The fully connected layer performs feature integration and spatial projection, transforming abstract temporal vectors into physically meaningful predictive variables. The output layer is responsible for the final decision transformation, converting complex computational results into structured data that can be directly used by the control system.

[0147] The output layer can simultaneously output both predicted short-term and long-term flow velocity trends, a feature crucial for hierarchical control strategies. The short-term trend prediction serves the following purposes: for sudden flow changes on the order of seconds or sub-seconds, the system can quickly fine-tune the angular velocity or lateral velocity of the hydrofoil 105 to cope with the load impact of instantaneous swells, ensuring the immediate safety and energy harvesting efficiency of the mechanical structure. The long-term trend prediction serves the following purposes: for flow velocity evolution over hourly or longer timeframes (such as tidal processes), the system can adjust the global state of the support structure or switch long-term operating strategies accordingly, avoiding frequent, small-amplitude ineffective adjustments and reducing motor fatigue and power consumption. This dual-scale prediction enables the tandem hydrofoil motion system to achieve a dynamic balance between precise micro-management and global scheduling.

[0148] The flow environment prediction model based on a bidirectional long short-term memory neural network structure exhibits strong generalization ability after training with measured data. In actual operation, even when facing complex turbulent flow with a certain degree of randomness, the flow environment prediction model can still extract statistically significant evolution patterns from the chaotic raw flow velocity data through its internal weight allocation. This technology ensures that the tandem hydrofoil motion system maintains the consistency and reliability of prediction results under different natural water conditions, enhancing the technical robustness of the entire power generation unit under all-weather operating conditions.

[0149] In water flow environment prediction models, Bi-directional Long-Short-Term Memory (Bi-LSTM) neural networks simulate complex nonlinear dynamic systems of water flow by deeply mining historical flow velocity data. In simple terms, its working mechanism can be broken down into the following key steps: 1) Data preprocessing: Frequency domain "feature reduction" Before the flow velocity data enters the neural network, the system first uses Fourier Transform (FT) to preprocess the flow velocity data stream. By converting the time-series signal to the frequency domain, the dominant frequency features are extracted and feature reduction is performed. This effectively reduces the impact of random noise and uncertainty in the flow velocity data on the prediction results, ensuring the "purity" of the input data. 2) Bi-directional temporal feature extraction. Traditional LSTM can only deduce the future from the past, while Bi-LSTM performs bi-directional computation through forward and backward networks in the hidden layers. The forward network extracts the evolution of flow velocity from the past to the future along the time axis. The backward network extracts temporal features in reverse, using subsequent information to perform logical verification of the current state. The final training result is a concatenation of the computational results from these two directions. 3) Simulating Nonlinear Dynamic Systems: To improve the robustness of the prediction, this water flow environment prediction model establishes a new FRB (Fourier Reduced Basis) model, introducing Dropout and L2 regularization. These mechanisms effectively prevent overfitting in the water flow environment prediction model and improve the network's generalization performance in different water areas (such as the Songhua River channel), enabling it to handle the extreme flow velocity range of 0.6 m / s to 1.4 m / s. 4. The hierarchical output of the prediction results is integrated through a fully connected layer. The output layer transforms the computational results into prediction trends at two different scales.

[0150] The power of Bi-LSTM lies in its ability to not only "see" data, but also to "understand" the temporal context of the flow field. Through the synergy of forward and backward directions, it can capture the deep temporal characteristics of flow velocity changes, thus providing a scientific basis for the advance adjustment of the hydrofoil 105 in complex and ever-changing natural waters.

[0151] The flow parameter prediction model of this invention employs a bidirectional long-short-term memory (Bi-LSTM) neural network to simulate the nonlinear dynamic system of the flow environment. Dropout and L2 regularization are used to improve the network's generalization performance, establishing a novel FRB (Fourier Reduced Basis) model. To mitigate the impact of uncertain data on the flow environment, Fourier transform (FT) is used to preprocess the flow data. The training data for this network uses measured flow data from the Songhua River channel, with an average flow velocity maintained between 1.0 and 1.2 m / s and a limiting velocity range of 0.6 to 1.4 m / s, meeting the operating flow velocity requirements of the hydrofoil 105.

[0152] To address the challenges of complex nonlinear dynamic systems in water flow environment forecasting, and considering the bidirectional nature of this task, a Bi-LSTM model was employed to simulate the nonlinear dynamic system of the water flow environment. The Bi-LSTM network was trained by concatenating the training results of forward and backward LSTM networks. Dropout and L2 regularization were introduced to prevent overfitting in the water flow environment prediction model. Therefore, the constructed FRB model consists of four layers: the first layer is the input layer, using real-time flow velocity parameters as input; the second layer is the Bi-LSTM layer; the third layer is a fully connected layer; and the fourth layer is the output layer. The loss function used was the Mini-Search Engine (MSE), and the Adam optimization algorithm was selected for model training. The model's performance was evaluated and analyzed based on measured water flow data.

[0153] The function of this flow prediction model is to take the real-time flow velocity parameters monitored by the flow velocity sensor as input, and provide short-term and long-term flow velocity change trends. Based on the short-term flow environment prediction, combined with the flow parameter adjustment model, a short-term control method is given to cope with sudden changes in flow velocity within a short period of time. Based on the long-term flow environment change trend, the motion control module 500 selects one of four pre-set control strategies (low-speed stable, low-speed unstable, high-speed stable, and high-speed unstable) as the current long-term motion control strategy for the hydrofoil 105. Through this flow environment prediction model, the long-term, efficient, and stable operation of the hydrofoil 105 in complex flow velocity environments is achieved.

[0154] Furthermore, after acquiring a water flow velocity data stream composed of multiple water flow velocity data, the water flow environment prediction model uses Fourier transform to preprocess the water flow velocity data stream in order to reduce the impact of flow velocity data uncertainty on the prediction results by performing feature reduction on the water flow velocity data stream.

[0155] The motion control module 500 provides a combination of motion parameters for the hydrofoil 105 based on the short-term water flow velocity change trend and the water flow parameter adjustment model, in order to cope with sudden changes in water flow velocity in the short term.

[0156] The motion control module 500 selects one of a variety of preset control strategies to execute based on the long-term trend of water flow environment changes, in order to cope with long-term changes in water flow velocity environment.

[0157] Among them, the duration of short-term events is no greater than the preset duration T1, and the duration of long-term events is no less than the preset duration T2.

[0158] Fourier transform preprocessing maps the chaotic, randomly fluctuating raw flow velocity signal in the time domain to the frequency domain for analysis. Through Fourier transform, the system can identify the dominant frequency component in the flow field, thus achieving feature reduction. This processing method effectively filters out high-frequency random noise and irregular disturbances in the flow field, reducing the impact of flow velocity data uncertainty on the prediction results. This provides a high-quality input benchmark for subsequent calculations of the bidirectional long short-term memory neural network structure, ensuring the accuracy of the predicted short-term and long-term flow velocity trends, and enabling the system's decisions to be based on a realistic and stable flow field evolution logic.

[0159] The motion control module 500, based on the short-term flow velocity variation trend and combined with the flow parameter adjustment model, provides the motion parameter combination for the hydrofoil 105, demonstrating the deep coupling of "prediction-adjustment". Since the flow parameter adjustment model employs a residual multilayer perceptron architecture, it can quickly calculate the optimal parameter combination. By introducing the short-term flow velocity variation trend (duration not exceeding a preset duration T1), the system possesses feedforward control capabilities, capable of handling sudden changes in flow velocity within a short period. This means that before instantaneous pulsations or significant jumps in flow velocity occur, the motion control module 500 has pre-adjusted the rotational speed of the first motor 106, the lateral movement characteristics of the second motor 107, and the driving spacing of the third motor 400, preventing lag in the mechanical system's response under abrupt flow field changes and maintaining the system's energy harvesting efficiency and operational stability during transient processes.

[0160] The motion control module 500 of this invention selects one of several preset control strategies based on the long-term trend of water flow environment changes. Its technical function is to achieve macroscopic control of the long-term operating state of the tandem hydrofoil motion system. For long-term water flow velocity environment changes lasting no less than a preset duration T2, the system no longer performs high-frequency micro-parameter jitter, but adapts to the evolution of the external environment through strategy-level switching. This hierarchical control means that the system can automatically and rationally switch between low-speed stable, low-speed unstable, high-speed stable, and high-speed unstable strategies according to the long-term evolution trend of tidal cycles or seasonal runoff. This reduces mechanical fatigue and reactive power loss caused by frequent motor adjustments, enabling the device to maintain an operating mode most suitable for the current environmental characteristics during long-term operation, achieving global operational economy and safety.

[0161] By setting preset durations T1 and T2, this invention constructs a hierarchical control timescale system. The technical advantage of this dual-timescale architecture lies in dividing the control task into fast-response parameter adjustment and slow-switching strategy selection. For short-timescale disturbances, the system reacts quickly by changing the motion parameter combination of the hydrofoil 105; for long-timescale evolutions, it reshapes the baseline through strategy switching. This hierarchical control avoids coupling confusion of control logic at different timescales, ensuring that the tandem hydrofoil motion system can maintain the convergence and stability of the control logic when facing the multi-scale time-varying characteristics of natural water bodies, thus guaranteeing the device's high adaptability and continuous energy harvesting capability in complex and variable flow environments.

[0162] Furthermore, multiple control strategies are included, such as low-speed stable strategy, low-speed unstable strategy, high-speed stable strategy, and high-speed unstable strategy, and:

[0163] Low-speed stability means that the flow velocity is not greater than the preset velocity V. 预设 And the flow field is stable;

[0164] Low-speed instability refers to a flow velocity that is not greater than the preset velocity V. 预设 Furthermore, the flow field is not stable;

[0165] High-speed stability refers to a flow velocity greater than the preset velocity V. 预设 And the flow field is stable;

[0166] High-speed instability refers to a flow velocity greater than the preset velocity V. 预设 Furthermore, the flow field is not stable.

[0167] The timescale for short-term trends is within a preset duration T1, and its application scenario is to cope with sudden changes in flow velocity (such as swells) and adjust the motion parameters of the hydrofoil 105 in real time. The timescale for long-term trends is beyond the preset duration T2, and its application scenario is to select macro-control strategies (such as high-speed stable / low-speed stable, high-speed unstable / low-speed unstable strategies).

[0168] This invention constructs a comprehensive operating condition matrix for a tandem hydrofoil motion system by defining low-speed stable, low-speed unstable, high-speed stable, and high-speed unstable strategies. This technical feature abstracts the continuously changing flow field state in natural water bodies into four discrete and representative control scenarios. This refined scenario division allows the motion control module 500 to selectively invoke preset dynamic logic. Instead of using a single control algorithm to handle all operating conditions, the system achieves a precise profile of the hydrodynamic environment based on the combination of flow velocity and flow field stability, providing a decision-making basis for subsequently outputting differentiated optimal hydrofoil motion parameter combinations.

[0169] This invention introduces a preset speed V 预设 This serves as a critical point for switching control logic. In the low-speed range (flow rate not exceeding the preset speed V)... 预设 The system identifies and enters a low-speed stable strategy or a low-speed unstable strategy, which focuses on improving the vibration sensitivity of the hydrofoil 105. By adjusting the first motor 106 and the second motor 107, the angle of attack response of the hydrofoil 105 in a low kinetic energy flow field is optimized, thereby ensuring the continuous work capacity of the device in a low flow velocity environment.

[0170] In the high-speed range (flow velocity greater than the preset velocity V) 预设 The system switches to a high-speed stable strategy or a high-speed unstable strategy. This focuses on load control and safety protection of the mechanical structure. By adjusting the combination of motion parameters, the system can limit the oscillation amplitude of the hydrofoil 105 under excess flow velocity, reduce the instantaneous impact on the first guide rail 103 and the connecting shaft 109, and thus extend the mechanical life of the oscillating hydrofoil power generation device 100.

[0171] This invention further refines the control logic by determining whether the flow field is stable or unstable. When the flow field is stable, the system tends to maintain the constancy of the optimal hydrofoil motion parameter combination, which reduces frequent motor fine-tuning and lowers the system's parasitic power consumption. When the flow field is unstable (i.e., turbulent or pulsating flow exists), the system's invoked unstable strategies (such as low-speed or high-speed unstable strategies) can enhance the system's disturbance rejection capability.

[0172] The motion control module 500 can dynamically adjust the angular velocity and angular acceleration of the hydrofoil 105, as well as the lateral velocity and lateral acceleration controlled by the second motor 107, according to the pulsating characteristics of the flow field. This adaptive adjustment based on the turbulence of the flow field establishes a dynamic balance between the motion trajectory of the hydrofoil 105 and the disturbed flow field, ensuring the continuous operation of the system under harsh flow conditions.

[0173] These strategies of the present invention are selected based on long-term trends in water flow environment changes. The four specific strategies defined in this invention serve as a set of instructions for macroscopic scheduling by the motion control module 500.

[0174] Once the flow environment prediction model outputs the long-term flow velocity change trend, the system can predict which strategy range the future flow field will enter (e.g., transitioning from low speed to high speed, or evolving from stable to unstable). This macroscopic switching effect based on a preset strategy avoids oscillations in the control logic under different flow states, achieving a smooth transition of control over a long timescale. It ensures that the tandem hydrofoil motion system can always maintain logical synchronization with the large-scale flow field environment during long-term operation, enhancing the system's adaptability to the periodic evolution of natural water bodies.

[0175] This invention provides a multi-dimensional search space for the motion control module 500 by dually determining the flow velocity and flow characteristics. This strategy classification scheme enhances the robustness of the algorithm. Even in complex mixed flow conditions, by comparing the flow velocity with the preset speed V... 预设 Based on the relationship between the flow field and the stability of the flow field, the system can easily retrieve the most suitable operating mode from the preset library. The role of this logical architecture is that it not only provides macroscopic constraint boundaries for the flow parameter adjustment model, but also provides alternative safe exit or degraded operation schemes for the tandem hydrofoil motion system when facing sudden extreme conditions, greatly improving the engineering practicality and operational robustness of the entire system.

[0176] According to another aspect of the present invention, a motion control method for the aforementioned tandem hydrofoil motion system is also provided, comprising the following steps:

[0177] 1) The water flow velocity sensor 600 detects the current water flow velocity in real time and transmits the water flow velocity to the motion control module 500;

[0178] 2) The motion control module 500 inputs the water flow velocity into the water flow environment prediction model and the water flow parameter adjustment model, respectively;

[0179] 3) The water flow environment prediction model outputs short-term and long-term water flow velocity change trends based on the water flow velocity data received within a set time period.

[0180] 4) The motion control module 500 determines the current control strategy based on the long-term water flow velocity change trend;

[0181] 5) The water flow parameter adjustment model outputs the optimal combination of hydrofoil motion parameters corresponding to the current water flow velocity environment based on the current water flow velocity and the short-term water flow velocity change trend;

[0182] 6) The motion control module 500 controls the operation of the first motor 106, the second motor 107 and the third motor 400 based on the current control strategy and the optimal combination of hydrofoil motion parameters, thereby adjusting the combination of hydrofoil motion parameters to the optimal.

[0183] In the motion control method described in this invention, the first step involves a water flow velocity sensor 600 detecting the current water flow velocity in real time and transmitting it to the motion control module 500. This technical solution establishes a high-frequency digital twin entry point for the entire tandem hydrofoil motion system. By converting the uncontrollable natural flow field into a measurable water flow velocity data stream, the system achieves instantaneous capture of external kinetic energy density. This real-time detection mechanism establishes a reliable benchmark for the control logic, ensuring that all subsequent predictions and adjustments are based on objective data of the current flow state, guaranteeing both the timeliness of information acquisition and the fidelity of the data.

[0184] Step 2) specifies that the motion control module 500 inputs the water flow velocity transmitted by the water flow velocity sensor 600 into the water flow environment prediction model and the water flow parameter adjustment model, respectively. The technical value of this feature lies in achieving parallel processing of breadth and depth in data processing: 1) Time-dimensional look-ahead processing: After the water flow velocity is input into the water flow environment prediction model, the system uses a bidirectional long short-term memory neural network structure to extract temporal features and extrapolate the future flow field. 2) Control-dimensional static mapping: Data is synchronously input into the water flow parameter adjustment model based on a residual multilayer perceptron architecture, establishing the spatial mapping relationship between flow velocity and motion parameters.

[0185] In step 3), the water flow environment prediction model receives water flow velocity data within a set time period and outputs short-term and long-term water flow velocity change trends. This step establishes a heterogeneous timescale system: 1) Macro-control timescale: By outputting long-term water flow velocity change trends (duration not less than the preset duration T2), it provides global operational guidance for the system, which is beneficial for maintaining long-term operational stability. 2) Micro-response timescale: By outputting short-term water flow velocity change trends (duration not greater than the preset duration T1), it identifies instantaneous disturbances in the flow field, providing a basis for refined motor operation. It can output once every set time interval, which is customizable; for example, outputting short-term water flow velocity change trends every 5 seconds and long-term water flow velocity change trends every minute.

[0186] In step 4), the motion control module 500 determines the current control strategy based on the long-term trend of water flow velocity changes. The system has preset multiple modes, including a low-speed stable strategy, a low-speed unstable strategy, a high-speed stable strategy, and a high-speed unstable strategy. This approach allows the system to pre-select a suitable logical framework based on the long-term evolution of the flow field (such as the trend of rising or falling flow velocity caused by tides). Under the high-speed stable strategy, the system tends to protect structural safety and suppress amplitude; while under the low-speed unstable strategy, algorithms are used to improve the vibration response. This hierarchical switching of strategies establishes the system's operating benchmark over a large time span, avoiding macroscopic control logic oscillations caused by short-term disturbances.

[0187] In step 5), the flow parameter adjustment model receives the current flow velocity and the short-term flow velocity change trend, and outputs the optimal hydrofoil motion parameter combination corresponding to the current flow velocity environment. The technical gains of this scheme are: 1) Optimal projection of high-dimensional parameter space: The optimal hydrofoil motion parameter combination output by the flow parameter adjustment model covers the lateral motion stiffness K of the hydrofoil 105. h The rotational stiffness K of hydrofoil 105 θ The lateral motion damping D of hydrofoil 105 h The rotational motion damping D of hydrofoil 105 θ And several key variables, such as the distance L between the two hydrofoils 105 along the direction perpendicular to the first guide rail 103. 2) Proactive dynamic compensation: Combining the short-term trend of water flow velocity changes, the output of the water flow parameter adjustment model is no longer a lagging static parameter, but a dynamic parameter combination with forward compensation properties. This allows the hydrofoil 105 to adjust its dynamic characteristics to the optimal state in advance when actually dealing with sudden changes in flow velocity, greatly enhancing its ability to intercept the energy of pulsating flow fields.

[0188] In step 6), the motion control module 500 controls the operation of the first motor 106, the second motor 107, and the third motor 400 based on the current control strategy and the optimal combination of hydrofoil motion parameters. The effects of this multi-motor coordinated execution scheme are as follows: 1) Precise shaping of rotational dynamics: By adjusting the angular velocity and angular acceleration of the hydrofoil 105 through the first motor 106, the optimal angle of attack for energy capture is maintained throughout the oscillation stroke of the hydrofoil 105. 2) Dynamic optimization of lateral trajectory: By adjusting the lateral velocity and lateral acceleration of the hydrofoil 105 through the second motor 107, the lateral motion is made to achieve phase matching with the fluid lift, thereby achieving higher energy capture efficiency. 3) Spatial adaptation of spacing: By adjusting the spacing between the two hydrofoils 105 through the third motor 400, the rear hydrofoil 105 can accurately fall into the high-energy region of the wake vortex generated by the front hydrofoil 105, achieving constructive intervention of the "vortex-hydrofoil" interaction.

[0189] The synchronous and coordinated operation of the three motors transforms the optimal combination of hydrofoil motion parameters output by the digital model into a real mechanical motion state, enabling the tandem hydrofoil motion system to achieve the optimal dynamic distribution at the physical execution level.

[0190] Through this multi-scale predictive output, the tandem hydrofoil motion system can take graded countermeasures according to flow field fluctuations of different frequencies, achieving deep coverage of variable flow velocity environments.

[0191] The control method of this invention achieves a complete logical closed loop from environmental monitoring, timing prediction, strategy selection to motor execution. This is achieved by considering different flow velocity ranges and flow characteristics (such as preset velocity V). 预设 The system automatically identifies and switches strategies (including flow field smoothness) to ensure normal operation under complex flow velocity conditions. Experimental data shows that compared to schemes lacking this control method, the tandem hydrofoil motion system of this invention exhibits a significant gain in average energy harvesting efficiency, typically reaching around 44.6%, directly stemming from the high dynamic fit between parameter combinations and flow field conditions. Strategy scheduling based on long-term water flow velocity trends reduces motor mechanical losses under ineffective operating conditions. Combined with limitations on parameters such as lateral acceleration, it effectively maintains structural integrity and ensures long-term stable operation of the device in complex hydraulic environments.

[0192] In summary, the control method of this invention, through six progressive steps, constructs a highly intelligent motion control method for a tandem hydrofoil system, encompassing multiple dimensions such as signal processing purity, predictive logic multidimensionality, macroscopic strategy selection, microscopic parameter adjustment precision, and strong synergy in mechanical execution. This method not only establishes the device's energy harvesting advantage under varying flow velocities but also significantly improves the technological maturity and operational stability of the entire power generation system in practical engineering applications through the deep integration of algorithms and mechanics.

[0193] This dual-track parallel processing method enables the tandem hydrofoil motion system to simultaneously predict future evolution trends and calculate the current optimal attitude when processing a single input data, thereby improving the efficiency of information flow utilization and the systematic nature of decision-making.

[0194] This invention employs a system design combining a water flow velocity sensor, a control module, and a motion parameter control system. During operation, it can identify and adjust the energy-harvesting motion parameters of the hydrofoil 105 in real time according to different water flow velocity environments, achieving optimal self-adjustment of motion parameters for the current flow velocity. This ensures the device operates normally under complex flow velocity conditions, thereby improving its adaptability. Simultaneously, based on the water flow parameter prediction model within the control module, it can predict short-term (within 10 seconds) and long-term (not less than 1 hour) trends in water flow conditions, enabling advance changes in the control parameters of the hydrofoil 105, further enhancing its adaptability to complex flow velocity conditions. Prototype observation tests under actual water flow conditions verified that, compared to the uncontrolled method, the average daily normal operating time using the control system of this invention is 23.4 hours, a 25.8% improvement over the 18.6 hours achieved in the uncontrolled group; indicating that this invention exhibits better adaptability under complex flow velocity conditions.

[0195] This invention employs a design combining a water flow velocity sensor, a motion control module 500, and motion parameter control. It can automatically adjust the energy harvesting parameters of the hydrofoil 105 based on different external flow velocities. According to the water flow parameter adjustment model within the control module, using the currently detected flow velocity as input, the optimal combination of motion parameters adapted to the current flow velocity is obtained and input to the motion parameter control system to adjust the motion parameters of the hydrofoil 105 to the optimal state, thereby achieving a higher overall energy harvesting efficiency. Prototype observation tests under actual water flow conditions verified that, compared to the uncontrolled method, the average energy harvesting efficiency of the control system of this invention is 44.6%, which is 24.9% higher than the 35.7% of the uncontrolled group; indicating that the overall energy harvesting efficiency of this invention is higher.

[0196] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A tandem hydrofoil motion system, comprising two oscillating hydrofoil power generation devices arranged sequentially along the water flow direction, each of the oscillating hydrofoil power generation devices comprising a support frame, a generator, a first guide rail, a slider, and a hydrofoil, wherein the generator and the first guide rail are both mounted on the support frame, the slider is slidably mounted on the first guide rail, and a connecting shaft is mounted on the upper end of the hydrofoil, the connecting shaft being rotatably mounted on the slider and connected to the generator, characterized in that: Each of the aforementioned oscillating hydrofoil power generation devices further includes a first motor, a second motor, and a first transmission mechanism. The first motor is mounted on the slider, and the output shaft of the first motor is connected to the connecting shaft to adjust the angular velocity and angular acceleration of the hydrofoil. The second motor and the first transmission mechanism are respectively mounted on the support frame, and the output shaft of the second motor is connected to the slider through the first transmission mechanism to adjust the lateral speed and lateral acceleration of the hydrofoil. The tandem hydrofoil motion system also includes a base, a second guide rail, a third motor, and a second transmission mechanism. Multiple second guide rails are respectively installed on the base and are parallel to each other. The second guide rails are perpendicular to the first guide rails. Each second guide rail is slidably connected to each support frame. The third motor and the second transmission mechanism are both installed on the base. The third motor is connected to the support frame through the second transmission mechanism to adjust the distance between the two support frames in the direction parallel to the second guide rail, thereby adjusting the distance between the two hydrofoils in the direction parallel to the second guide rail. The first motor, the second motor, and the third motor are electrically connected to the motion control module. The tandem hydrofoil motion system also includes a water flow velocity sensor, which is electrically connected to the motion control module. The motion control module controls the rotation of the first motor, the second motor, and the third motor based on the water flow velocity feedback from the water flow velocity sensor. Each hydrofoil operates under the following conditions: ； ； ； in, F The lift generated by the hydrofoil; M The water flow torque generated by the hydrofoil acts on the connecting shaft; S For the hydrofoil's static moment; m The total mass of the lateral movement component, which includes a hydrofoil and components that move laterally along with the hydrofoil; I θ Let be the moment of inertia of the hydrofoil; , , These are the lateral displacement, lateral velocity, and lateral acceleration of the hydrofoil, respectively. , and These are the rotation angle, angular velocity, and angular acceleration of the hydrofoil, respectively. K h The lateral motion stiffness of the hydrofoil; K θ The rotational stiffness of the hydrofoil; D h Damping for the lateral movement of the hydrofoil; D θ Damping for the rotational motion of the hydrofoil; x θ This is the distance from the hydrofoil's center of mass to the connecting shaft; m θ The total mass of the rotating module involved in the hydrofoil rotation is defined as follows: the rotating module includes the hydrofoil, a connecting shaft, and a first motor.

2. The tandem hydrofoil motion system according to claim 1, characterized in that: Both the first transmission mechanism and the second transmission mechanism are lead screw mechanisms.

3. The tandem hydrofoil motion system according to claim 1, characterized in that: The motion control module has a water flow parameter adjustment model, which adopts a residual multilayer perceptron architecture and is trained and generated based on previous actual test data. This model is used to generate the optimal combination of motion parameters for the hydrofoil according to the current water flow velocity. The preliminary experimental data included water flow velocity, hydrofoil motion parameter combinations, and hydrofoil energy capture efficiency. The hydrofoil motion parameter combinations included the hydrofoil's lateral stiffness K. h Rotational stiffness K θ Lateral damping D h Rotational damping D θ And the distance L between the two hydrofoils along the direction perpendicular to the first guide rail.

4. The tandem hydrofoil motion system according to claim 3, characterized in that: The water flow parameter adjustment model includes an input layer, a residual block stacking layer, and an output layer, and: The input layer is used to acquire the flow velocity data detected by the water flow velocity sensor in real time, and to map it to a high-dimensional feature space of a preset dimension composed of hidden layer nodes through linear transformation. The residual block stacking layer consists of a preset number of serially stacked residual blocks. The residual connection mechanism performs nonlinear transformation on the water flow velocity characteristics in the high-dimensional feature space to establish the mapping relationship between water flow velocity fluctuations and the combination of motion parameters of the hydrofoil. The output layer is used to linearly combine the high-dimensional feature vectors of the preset dimension using a fully connected weight matrix, mapping them from the feature space to a preliminary parameter space composed of the motion parameters of the hydrofoil, and performing numerical range restriction processing on the mapped preliminary parameter space to ensure that the output value conforms to the adjustment range of the first motor, the second motor and the third motor. Then, the dimensionless predicted value is restored to the actual motion parameter value with physical units, and the optimal hydrofoil motion parameter combination corresponding to the current flow velocity is output.

5. The tandem hydrofoil motion system according to claim 3, characterized in that: The motion control module includes a water flow environment prediction model, and the water flow environment prediction model adopts a bidirectional long short-term memory neural network structure. The water flow environment prediction model comprises an input layer, a hidden layer, a fully connected layer, and an output layer connected in sequence, wherein: The input layer is used to receive water flow velocity data monitored by the water flow velocity sensor in real time. The hidden layer is used to extract the depth time-series features of the water flow velocity data through the forward network and the backward network, and to concatenate the time-series feature vectors obtained by bidirectional calculation to generate a high-dimensional time-series feature vector characterizing the evolution law of the flow field. The fully connected layer is used to receive the spliced ​​temporal feature vector output by the hidden layer, and to perform feature extraction and nonlinear compression on the spliced ​​temporal feature vector, so as to realize the projection of temporal information from the high-dimensional feature space to the prediction target space. The output layer is used to perform multi-scale regression calculation on the spliced ​​time-series feature vector after feature extraction and nonlinear compression of the fully connected layer, convert the spliced ​​time-series feature vector into a value that conforms to physical dimensions, and output the predicted short-term and long-term water flow velocity change trends respectively.

6. The tandem hydrofoil motion system according to claim 5, characterized in that: After acquiring a water flow velocity data stream composed of multiple water flow velocity data, the water flow environment prediction model uses Fourier transform to preprocess the water flow velocity data stream in order to reduce the impact of the uncertainty of the flow velocity data on the prediction results by performing feature reduction on the water flow velocity data stream. The motion control module provides a combination of motion parameters for the hydrofoil based on the short-term water flow velocity change trend and the water flow parameter adjustment model, in order to cope with sudden changes in water flow velocity in the short term. The motion control module selects one of a variety of preset control strategies to execute based on the long-term trend of water flow environment changes, in order to cope with long-term changes in water flow velocity environment. Among them, the duration of short-term events is no greater than the preset duration T1, and the duration of long-term events is no less than the preset duration T2.

7. The tandem hydrofoil motion system according to claim 6, characterized in that: Multiple control strategies are available, including low-speed stable strategy, low-speed unstable strategy, high-speed stable strategy, and high-speed unstable strategy, and: Low-speed stability means that the flow velocity is not greater than the preset velocity V. 预设 And the flow field is stable; Low-speed instability refers to a flow velocity that is not greater than the preset velocity V. 预设 Furthermore, the flow field is not stable; High-speed stability refers to a flow velocity greater than the preset velocity V. 预设 And the flow field is stable; High-speed instability refers to a flow velocity greater than the preset velocity V. 预设 Furthermore, the flow field is not stable.

8. The motion control method for the tandem hydrofoil motion system according to any one of claims 5 to 7, characterized in that: Includes the following steps: 1) The water flow velocity sensor detects the current water flow velocity in real time and transmits the water flow velocity to the motion control module; 2) The motion control module inputs the water flow velocity into the water flow environment prediction model and the water flow parameter adjustment model, respectively; 3) The water flow environment prediction model outputs short-term and long-term water flow velocity change trends based on the water flow velocity data received within a set time period. 4) The motion control module determines the current control strategy based on the long-term water flow velocity change trend; 5) The water flow parameter adjustment model outputs the optimal combination of hydrofoil motion parameters corresponding to the current water flow velocity environment based on the current water flow velocity and the short-term water flow velocity change trend; 6) The motion control module controls the operation of the first motor, the second motor and the third motor based on the current control strategy and the optimal combination of hydrofoil motion parameters, thereby adjusting the combination of hydrofoil motion parameters to the optimal.

Images