Navigation body tail beat force prediction model construction method based on physical test system
By building a direct-flight high-frequency swing physical test device, obtaining the tail-beat characteristic parameters of the supercavitating vehicle, and constructing a data-driven tail-beat force prediction model, the problem of deviation between the tail-beat force prediction model and the actual scene was solved, achieving more accurate tail-beat force prediction and optimized motion control.
Patent Information
- Application Number
- CN202510737670.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-04
- Publication Date
- 2025-09-19
AI Technical Summary
In the existing technology, the tail-beat force prediction model of the supercavitating vehicle deviates from the actual physical scene, resulting in poor motion control effect and difficulty in accurately predicting the tail-beat force.
A direct-flight high-frequency swing physical test device was built, and prototype motion tests were carried out on a scaled model of a supercavitating vehicle to obtain the tail beat characteristic parameters. A tail beat force prediction model was constructed using a data-driven method. The tail beat force test results were obtained using a data acquisition device, and a fully connected network was constructed to predict the tail beat force.
It improves the accuracy of tail-beat force prediction, optimizes the motion control strategy of the vehicle, reduces the cost of prototype testing, and has high practical value.
Smart Images
Figure CN120671354A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of supercavitating vehicles, and in particular to a method for constructing a vehicle tail-beat force prediction model based on a physical test system. Background Art
[0002] In order to reduce the influence of water resistance during the movement of underwater high-speed vehicles, a cavitator is generally installed at the head of the vehicle to form cavitation bubbles that wrap around the vehicle during the high-speed operation of the vehicle, which is conducive to achieving the speed increase and resistance reduction of the vehicle. This type of underwater high-speed vehicle is generally also called a supercavitating vehicle.
[0003] During high-speed navigation, a supercavitating vehicle, encased in a cavitation bubble, is unable to withstand the lifting force of the water and, therefore, naturally falls under the influence of gravity. Cavitation is generated by the cavitator at the vehicle's head, so the tail of the vehicle first contacts the water during each natural fall, a phenomenon known as tail slap. During this periodic tail slap, the tail surface interacts with the water medium outside the cavitation bubble, which has a complex geometric interface, generating a tail slap force. Due to the complex flow between the tail surface and the cavitation bubble during the tail slap, this tail slap force exhibits strong nonlinear characteristics. It changes rapidly and at high frequency during navigation and significantly affects the vehicle's motion and attitude, posing a significant challenge to the motion control of supercavitating vehicles.
[0004] Through in-depth research on tail-slap forces, a tail-slap force prediction model is constructed to predict the nonlinear tail-slap force. This force is then added as an external disturbance to the closed-loop motion control process of the supercavitating vehicle. On this basis, the motion control strategy of the supercavitating vehicle is continuously optimized, which is beneficial to improving the motion stability of the supercavitating vehicle. Currently, a common practice is to conduct prototype tests on a scaled-down model of the supercavitating vehicle to obtain relevant experimental data. The data is then substituted into semi-empirical theoretical formula models such as the Paryshev model and the Hassan model to modify the parameters and construct a tail-slap force prediction model. However, this theoretical formula model deviates from the actual physical scenario, resulting in low prediction accuracy of the tail-slap force prediction model, which affects the motion control effect of the supercavitating vehicle. Summary of the Invention
[0005] In response to the above-mentioned problems and technical needs, this application proposes a method for constructing a tail-beat force prediction model for a vehicle based on a physical test system. The technical solution of this application is as follows:
[0006] A method for constructing a tail-beat force prediction model for a vehicle based on a physical test system, the method comprising:
[0007] A direct-flight high-frequency oscillation physical test device was constructed. The direct-flight high-frequency oscillation physical test device included a main controller, a traveling mechanism, a oscillating mechanism, a tail slapping water model, and a test pool. The tail slapping water model was a local equivalent model of the tail surface of a scaled-down supercavitating vehicle model, and a data acquisition device was built into the tail slapping water model. The tail slapping water model was rotatably connected to the oscillating mechanism, which was connected to the traveling mechanism. The main controller was electrically connected to the traveling mechanism, the oscillating mechanism, and the data acquisition device. The traveling mechanism drove the tail slapping water model to move via the oscillating mechanism, which drove the tail slapping water model to rotate relative to the horizontal plane. The tail slapping water model was located above the liquid surface of the test pool.
[0008] Conduct prototype motion tests on a scaled-down supercavitating vehicle model to obtain the tailbeat characteristic parameters of the tailbeat phenomenon that occurs during the prototype motion test of the scaled-down supercavitating vehicle model.
[0009] Based on tailbeat characteristic parameters, multiple sets of test parameters were constructed. The relative position of the liquid surface in the test pool and the tailbeat water model was adjusted according to each set of test parameters. The motion of the walking and swinging mechanisms was controlled by a main controller. Tailbeat physical simulation tests were conducted using a direct-flight high-frequency swing physical test device. The tailbeat force test results under the current test parameters were obtained using the data acquisition device built into the tailbeat water model.
[0010] The tailbeat force prediction model was constructed in a data-driven manner based on the tailbeat force test results under each group of test parameters.
[0011] The beneficial technical effects of this application are:
[0012] The present application discloses a method for constructing a tail-slap force prediction model for a vehicle based on a physical test system. This method constructs a direct-flight high-frequency swing physical test device for prototype motion tests of a scaled-down supercavitation vehicle model. Using this direct-flight high-frequency swing physical test device, based on the test results of the prototype motion test, a tail-slap physical simulation test can be quickly carried out to obtain tail-slap force test results. The tail-slap force test results obtained from the actual physical test can then be used to construct a tail-slap force prediction model based on a data-driven approach. The tail-slap force prediction model constructed in this way is closer to the actual physical scene than the theoretical formula model, and can therefore more accurately predict the tail-slap force. The control strategy verified by using this tail-slap force prediction model during the high-speed operation of an actual underwater vehicle can achieve better motion posture effects. At the same time, it can reversely guide the overall design of the vehicle, achieving this through forced motion simulating free motion. Furthermore, it can also be applied to the study of the tail-slap mechanics of bionic fish or the high-frequency swing of vector nozzles, thus having high practical value.
[0013] This method uses a constructed direct-flight high-frequency swing physical test device to conduct physical experiments and obtain data. Physical experiments conducted using the direct-flight high-frequency swing physical test device are easier to implement than prototype tests. Therefore, this method is conducive to reducing the number of prototype tests, thereby enabling rapid iteration and reducing the high cost of prototype tests. The direct-flight high-frequency swing physical test device combined with the test method provided in this application can also be used to deeply study the formation mechanism of the tail slap force of the navigation body. BRIEF DESCRIPTION OF THE DRAWINGS
[0014] Figure 1 This is a flowchart of a method for constructing a tail-slap force prediction model for a navigation body in one embodiment of the present application.
[0015] Figure 2 It is a structural diagram of a direct-flight high-frequency oscillation physical test device constructed in one embodiment of the present application.
[0016] Figure 3 yes Figure 2 A partial structural diagram of the direct flight high-frequency oscillation physical test device from one perspective.
[0017] Figure 4 yes Figure 2 A partial structural diagram of the direct-flight high-frequency oscillation physical test device from another perspective.
[0018] Figure 5 It is a schematic diagram of the rotation range of the tail-slapping water model driven by the swing mechanism.
[0019] Figure 6 This is an electrical structure diagram of a direct flight high-frequency oscillation physical test device constructed in one embodiment of the present application.
[0020] Figure 7 It is a schematic diagram of the state in which the liquid surface in the test pool and the tail water model are in the first relative position. DETAILED DESCRIPTION
[0021] The specific implementation of this application will be further described below with reference to the accompanying drawings.
[0022] This application discloses a method for constructing a tail-beat force prediction model for a navigation body based on a physical test system. Please refer to Figure 1 As shown in the flowchart, the method for constructing the tail-beat force prediction model of the vehicle includes the following steps:
[0023] Step 110: Build a direct flight high-frequency oscillation physical test device.
[0024] Please refer to Figure 2 As shown in the device structure diagram, the direct flight high-frequency swing physical test device includes a main controller, a walking mechanism 1, a swing mechanism 2, a tail-slapping water model 3, and a test pool 4, and actually also includes an external power supply. Figure 2 The main controller and external power supply are not shown.
[0025] When studying the tail-slap force mechanism of a supercavitating vehicle, a scaled-down model of the supercavitating vehicle is usually constructed and then studied on the scaled-down model. This application builds on this foundation by constructing a direct-flight, high-frequency oscillation physical test apparatus for the scaled-down supercavitating vehicle model. Therefore, the tail-slap water model in this application is a local equivalent model of the tail surface of the scaled-down supercavitating vehicle model.
[0026] The tail slap water model is rotatably connected to the swing mechanism 2, and the swing mechanism 2 is connected to the walking mechanism 1. The walking mechanism 1 drives the tail slap water model 3 to move via the swing mechanism 2, and the swing mechanism 2 drives the tail slap water model 3 to rotate relative to the horizontal plane.
[0027] Please refer to Figure 3 and Figure 4 An enlarged view of a partial structure of the traveling mechanism 1 shows that the traveling mechanism 1 includes a linear motor track 101 and a linear motor slide 102 positioned within the linear motor track 101. The linear motor track 101 spans over the test water tank 4. The motor driver within the traveling mechanism 1 drives the linear motor slide 102 within the linear motor track 101. The use of linear motors in the traveling mechanism 1 enables greater acceleration and deceleration, maximizing the distance traveled in the uniform speed segment and maximizing the maximum achievable speed.
[0028] The swing mechanism 2 includes a rotary motor 201, which is fixed to the linear motor slide 102. The output shaft of the rotary motor 201 is connected to the axis of one side of the swing wheel 202. The other side of the swing wheel 202 is provided with an eccentric shaft 203. The eccentric shaft 203 is movably arranged in the keyway of the crank 204. The tail water model 3 is assembled at the bottom of the crank 204. A movably connected ear shaft is provided in the ear shaft hole at the bottom of the crank 204. Support members 205 are fixed at both ends of the ear shaft. One end of the support member 205 is fixed to the bottom of the linear motor slide 102. The output shaft of the rotary motor 201 drives the swing wheel 202 to rotate. The swing wheel 202 drives the eccentric shaft 203 to rotate in the keyway of the crank 204, and then drives the crank 204 to rotate around the ear shaft, so that the crank 204 drives the tail water model 3 to rotate relative to the horizontal plane with the ear shaft as the rotation axis.
[0029] The main controller is electrically connected to the walking mechanism 1 and the swing mechanism 2, including a motor driver electrically connected to the walking mechanism 1 to control the movement of the walking mechanism 1, and a motor driver electrically connected to the swing mechanism 2 to control the movement of the swing mechanism 2. The horizontal movement speed of the tail water model 3 can be adjusted by adjusting the sliding speed of the linear motor slide 102 within the linear motor track 101. The rotation angle α of the tail water model 3 relative to the horizontal plane is related to the rotation motor angle of the rotary motor 201, and the rotation speed of the tail water model 3 can be adjusted by adjusting the rotational angular velocity of the rotary motor 201.
[0030] During the rotation of the rotating motor 201, the tail water model 3 rotates within a predetermined angle range relative to the horizontal plane. The angle between the axis of the tail water model 3 and the horizontal plane is the rotation angle α of the tail water model 3 relative to the horizontal plane. When the axis of the tail water model 3 is below the horizontal plane, the rotation angle α is positive, and when the axis of the tail water model 3 is above the horizontal plane, the rotation angle α is negative. After completing the structural design, the extreme rotation position of the tail water model 3 can be pre-calibrated to determine the value range of the rotation angle α. The value of β can be pre-calibrated. Please refer to Figure 5 It shows a schematic structural diagram of the tail-slapping water model 3 at the extreme rotation positions on both sides.
[0031] The tail slap water model 3 also has a built-in data acquisition device. The main controller is electrically connected to the data acquisition device in the tail slap water model 3 to obtain data collected by the data acquisition device. The data acquisition device built into the tail slap water model includes a ring balance, an acceleration sensor, and an angular velocity sensor.
[0032] The direct flight high frequency swing physical test device also includes a test collection box 5, such as Figure 2-Figure 4 As shown, the test collection box 5 is located above the liquid surface and is usually fixed on the linear motor slide 102. The data acquisition device built into the tail water model 3 is connected to the test collection box 5 via the acquisition power cable 6, and the acquisition power cable 6 is connected to the tail water model 3 and the test collection box 5 respectively using watertight connectors. Figure 6 As shown in the electrical structure diagram, the test collection box 5 includes a power module. The power module in the test collection box 5 supplies power to the data acquisition device in the tail water model 3 via the acquisition power supply cable 6, and the data acquisition device in the tail water model 3 feeds back data to the test collection box 5 via the acquisition power supply cable 6, and uploads the data to the main controller via the test collection box 5.
[0033] The built-in data acquisition device of the tail slap water model 3 includes a ring balance, an acceleration sensor and an angular velocity sensor. The ring balance is used to measure the force and torque parameters of the tail slap water model 3, the acceleration sensor is used to measure the acceleration of the tail slap water model 3, and the angular velocity sensor is used to measure the angular velocity of the tail slap water model 3. Figure 6 As shown in the electrical structure diagram, the test acquisition box 5 also includes a voltage amplifier and an analog input module. The built-in annular balance of the tail water model 3 is connected to the voltage amplifier in the test acquisition box 5 via the acquisition power cable 6 and then connected to the analog input module via the signal isolator. The voltage amplifier amplifies the weak mV signal of the annular balance to a voltage signal of 0-5V, and the signal isolator is used to filter out interference in the power frequency signal. The acceleration sensor is an integrated sensor, which is directly connected to the test acquisition box 5 via the acquisition power cable 6 and connected to the analog input module via the signal isolator. The angular velocity sensor is also an integrated sensor, which is connected to the test acquisition box 5 via the acquisition power cable 6 and connected to the analog input module via the signal isolator.
[0034] The communication lines for the motor drivers in travel mechanism 1, swing mechanism 2, and the analog input module in test and acquisition box 5 are connected to the main controller via the EtherCAT bus. The main controller forms the master station, the analog input module in test and acquisition box 5 forms a slave station, the motor drivers in travel mechanism 1, and the motor drivers in swing mechanism 2 also form slave stations. These stations are connected via the EtherCAT bus, achieving a control and data acquisition cycle of less than 1ms, resulting in a fast response time.
[0035] Connect the power cables of the motor drivers in travel mechanism 1, swing mechanism 2, the power module in test and acquisition box 5, and the main controller to external power supplies. Connect the power cables of the motor drivers in travel mechanism 1 and swing mechanism 2 to a 220V power supply, and the power cables of the power module in test and acquisition box 5 and the main controller to a 24V power supply.
[0036] The tail-slapping water model 3 is positioned above the liquid level of the test pool 4, and the relative position of the liquid level of the test pool 4 and the tail-slapping water model 3 is adjustable. Typically, the height of the tail-slapping water model 3 is fixed, and the relative position of the liquid level and the tail-slapping water model 3 is adjusted by adjusting the liquid level of the test pool 4. Furthermore, the test pool 4 is typically made of a transparent material, such as acrylic, to facilitate external observation and high-speed video recording of the movement of the tail-slapping water model 3.
[0037] Step 120, conduct a prototype motion test on the scaled-down model of the supercavitating vehicle to obtain the tail beat characteristic parameters of the tail beat phenomenon that occurs during the prototype motion test of the scaled-down model of the supercavitating vehicle. The tail beat characteristic parameters can be obtained through inertial element measurement and camera image information. This part can be implemented according to the existing practice of the prototype motion test of the scaled-down model of the supercavitating vehicle, and will not be repeated here.
[0038] The tailbeat characteristic parameters obtained include the tailbeat angle θ0 of the tailbeat phenomenon that occurs during the prototype motion test of the supercavitating vehicle scale model, the angular velocity ω0 when the tailbeat phenomenon occurs, and the navigation speed ν0 of the supercavitating vehicle scale model.
[0039] Step 130: construct multiple groups of test parameters based on the tail beat characteristic parameters.
[0040] The tailbeat characteristic parameters are used as typical values to determine the value range, including the tailbeat angle value range [θ0-Δθ1, θ0+Δθ2], the angular velocity value range [ω0-Δω1, ω0+Δω2], and the navigation speed value range [υ0-Δυ1, υ0+Δυ2]. Δθ1 and Δθ2 are angle differences, Δω1 and Δω2 are angular velocity differences, and Δυ1 and Δυ2 are speed differences. These parameters can be customized.
[0041] Then, multiple sets of experimental parameters were constructed by combining parameters for different values of the tailbeat angle θ within the range of [θ0-Δθ1, θ0+Δθ2], different values of the angular velocity ω within the range of [ω0-Δω1, ω0+Δω2], and different values of the sailing speed υ within the range of [υ0-Δυ1, υ0+Δυ2]. Each set of experimental parameters constructed in this way includes a set of values for the tailbeat angle θ, angular velocity ω, and sailing speed υ.
[0042] For example, in one example, the tail angle range [0.9θ0, 1.1θ0] is incremented by 0.05θ0, resulting in five different values: 0.9θ0, 0.95θ0, θ0, 1.05θ0, and 1.1θ0. The angular velocity range [0.9ω0, 1.1ω0] is incremented by 0.05ω0, resulting in five different values: 0.9ω0, 0.95ω0, ω0, 1.05ω0, and 1.1ω0. The angular velocity range [0.9υ0, 1.1υ0] is incremented by 0.05υ0, resulting in five different values: 0.9υ0, 0.95υ0, υ0, 1.05υ0, and 1.1υ0. In this way, 125 different sets of experimental parameters can be constructed and combined, for example, one set of experimental parameters is (0.95θ0, 1.05ω0, 1.05υ0).
[0043] In addition, the extreme rotation position of the tail-slap water model 3 is adjusted in advance during the structural design so that the tail-slap angle in any set of test parameters is This can ensure that the tail-slapping water model 3 continues to rotate after contacting the liquid surface in the test water pool 4 instead of just making point contact.
[0044] Step 140: Adjust the relative position of the liquid surface in the test pool and the tail slap water model according to each set of test parameters and control the movement process of the walking mechanism and the swing mechanism through the main controller, so as to use the direct-flight high-frequency swing physical test device to carry out the tail slap physical simulation test, and obtain the tail slap force test results under the current test parameters through the built-in data acquisition device of the tail slap water model.
[0045] When conducting tail-beat physical simulation tests using a direct-flight high-frequency swing physical test device according to a set of test parameters, two test processes were carried out, including:
[0046] (1) First test process
[0047] First, adjust the liquid surface in the test pool 4 and the tail slap water model 3 to a first relative position, so that when the tail slap water model 3 contacts the liquid surface in the test pool 4 during the rotation relative to the horizontal plane, the angle between the axis of the tail slap water model 3 and the horizontal plane (i.e., the rotation angle α) is the tail slap angle θ in the test parameters, as shown in FIG. Figure 7 shown.
[0048] Then, the main controller controls the motion process of the walking mechanism 1 and the swing mechanism 2 according to the navigation speed υ and the angular velocity ω in the test parameters, and obtains the first test data through the built-in data acquisition device of the wake water model 3.
[0049] Controlling the motion process of the walking mechanism 1 and the swing mechanism 2 according to the test parameters includes:
[0050] The navigation speed υ in the test parameters is used as the uniform speed of the walking mechanism 1 to control the movement process of the walking mechanism 1, and appropriate acceleration and deceleration sections are set. The rotational angular velocity of the swing mechanism 2 is controlled according to the angular velocity ω in the test parameters.
[0051] In order to ensure the validity of data acquisition, during the test, when it is determined that the tail water model 3 has reached a stable motion state, the first test data is acquired through the built-in data acquisition device. Specifically:
[0052] Keep the walking mechanism 1 stationary, first control the swing mechanism 2 to drive the tail water model 3 to rotate at an angular velocity The tail slapping water model 3 rotates and detects the real-time angular velocity of the tail slapping water model through the built-in angular velocity sensor of the tail slapping water model 3.
[0053] When it is determined that the real-time angular velocity of the tail-slapping water model 3 reaches the angular velocity ω, the walking mechanism 1 is controlled to start accelerating until it reaches the sailing speed υ and then maintains uniform motion, and the real-time acceleration of the tail-slapping water model 3 is detected by the acceleration sensor built into the tail-slapping water model 3.
[0054] When it is determined that the tail lap water model 3 has reached a uniform motion state based on the real-time acceleration of the tail lap water model 3, it can be determined that the tail lap water model 3 has reached a stable motion state. At this time, the built-in circular balance of the tail lap water model 3 is used to measure the force of the tail lap water model 3 as the rotation angle α under the current test parameters (υ, ω, θ) to obtain the first test data F1 (υ, ω, θ, α).
[0055] During the test, the swing mechanism 2 drives the tail water model 3 to rotate. When the rotation angle of the tail water model 3 is When the tail-slapping water model 3 is within the range, it is in the air and not in contact with the liquid surface, and is not affected by the tail-slapping force. When it is in the range and in contact with the liquid, it is affected by the tail force, as well as other interfering forces.
[0056] (2) Second test process
[0057] Adjust the liquid level in the test water pool 4 and the tail water model 3 to a second relative position so that the tail water model 3 never contacts the liquid level in the test water pool during the rotation relative to the horizontal plane. This can be achieved by lowering the liquid level in the test water pool.
[0058] Similarly, the motion process of the walking mechanism and the swing mechanism is controlled by the main controller according to the navigation speed υ and angular velocity ω in the test parameters, and the second test data is obtained by the data acquisition device built into the tail water model. The method for controlling the motion process of the walking mechanism 1 and the swing mechanism 2 is the same as that in the first test process, and will not be repeated here. Similarly, when the tail water model 3 is in a stable motion state, the force of the tail water model 3 is measured by the annular balance built into the tail water model 3 under the current test parameters (υ, ω, θ) as the rotation angle α obtains the second test data F2 (υ, ω, θ, α).
[0059] The difference between the second test process and the first test process is that in this test process, the swing mechanism 2 drives the tail water model 3 to rotate, and the tail water model 3 rotates at an angle of When it is within the range, it is always in the air, not in contact with the liquid surface and not affected by the tail force.
[0060] (3) Obtain the tail-beat force test results under the current test parameters (υ, ω, θ).
[0061] After completing two tests, the first test data and the second test data under the current test parameters (υ, ω, θ) can be obtained. The first test data includes the force F1 (υ, ω, θ, α) of the tail slap water model 3 under any rotation angle α under the current test parameters (υ, ω, θ), which includes the tail slap force. The second test data includes the force F2 (υ, ω, θ, α) of the tail slap water model 3 under any rotation angle α under the current test parameters (υ, ω, θ), which does not include the tail slap force.
[0062] Then, the tail slap force test result under the current test parameters (υ, ω, θ) can be obtained according to the first test data and the second test data, including: using the force F1 (υ, ω, θ, α) of the tail slap water model under any rotation angle α under the current test parameters (υ, ω, θ) in the first test data, minus the force F2 (υ, ω, θ, α) of the tail slap water model under the same rotation angle α under the current test parameters (υ, ω, θ) in the second test data, to obtain the tail slap force F of the tail slap water model at the current rotation angle α under the current test parameters (υ, ω, θ). y (υ, ω, θ, α) = F1(υ, ω, θ, α) - F2(υ, ω, θ, α). This can remove the influence of other interfering forces and extract the tail slap force exerted on the tail slap water model 3 when it contacts the liquid.
[0063] Since the tail water model 3 is only in the rotation angle When it is in the range, it is in contact with the liquid and receives the tail force, so the combined rotation angle Tail force F within the range y (υ, ω, θ, α) to obtain the tail beat force test results under the current test parameters.
[0064] By conducting physical tests on each group of test parameters according to the above method, the tail beat force test results under each group of test parameters can be obtained.
[0065] Step 150 , synthesizing the tailbeat force test results under each set of test parameters and constructing a tailbeat force prediction model in a data-driven manner.
[0066] As mentioned above, the tail slap force test results under each set of test parameters (υ, ω, θ) actually include the forces under multiple different rotation angles α under the test parameters (υ, ω, θ). Multiple sampling can be performed within the range to obtain multiple data samples, thereby constructing a larger number of data samples, such as the tail beat force test results under each set of test parameters. If 100 samplings are performed within the range, there are 100 data samples under each set of test parameters, and a total of 12,500 data samples under all 125 sets of test parameters.
[0067] A fully connected network was constructed using Keras. The tailbeat angle θ, angular velocity ω, navigation speed υ, and rotation angle α in each set of tailbeat force test results were used as inputs to the fully connected network. The tailbeat force F at the rotation angle α included in the tailbeat force test results of the current test parameters (υ, ω, θ) was calculated. y (υ, ω, θ, α) is the output of the fully connected network. The tailbeat force test results under each set of test parameters are used to train the constructed fully connected network to obtain the tailbeat force prediction model.
[0068] The resulting tail-beat force prediction model is then applied to a supercavitating vehicle motion control model simulation to continuously optimize the control strategy. This optimized control strategy is then updated to the supercavitating vehicle's external structure and control program. Prototype motion tests are then conducted on a scaled-down supercavitating vehicle model to verify the optimized control strategy. If the control objectives are still not achieved, the above process is repeated.
[0069] The above description is only a preferred embodiment of the present application, and the present application is not limited to the above embodiments. It is understood that other improvements and variations directly derived or imagined by those skilled in the art without departing from the spirit and concept of the present application should be considered to be included in the scope of protection of the present application.
Claims
1. A method for constructing a tail-beat force prediction model for a vehicle based on a physical test system, characterized in that: The method for constructing a tail-beat force prediction model of a vehicle comprises: A direct-flight high-frequency oscillation physical test device is constructed, comprising a main controller, a traveling mechanism, a swinging mechanism, a tail slapping water model, and a test pool. The tail slapping water model is a local equivalent model of the tail surface of a scaled-down supercavitating vehicle model, and a data acquisition device is built into the tail slapping water model. The tail slapping water model is rotatably connected to the swinging mechanism, which is connected to the traveling mechanism. The main controller is electrically connected to the traveling mechanism, the swinging mechanism, and the data acquisition device. The traveling mechanism drives the tail slapping water model to move via the swinging mechanism, and the swinging mechanism drives the tail slapping water model to rotate relative to the horizontal plane. The tail slapping water model is located above the liquid surface of the test pool. Conduct prototype motion tests on a scaled-down supercavitating vehicle model to obtain the tailbeat characteristic parameters of the tailbeat phenomenon that occurs during the prototype motion test of the scaled-down supercavitating vehicle model. Based on the tailbeat characteristic parameters, multiple sets of test parameters are constructed. The relative position of the liquid surface in the test pool and the tailbeat water model is adjusted according to each set of test parameters. The motion process of the walking mechanism and the swing mechanism is controlled by a main controller. A tailbeat physical simulation test is carried out using the direct-flight high-frequency swing physical test device. The tailbeat force test results under the current test parameters are obtained through the built-in data acquisition device of the tailbeat water model. The tailbeat force prediction model was constructed in a data-driven manner based on the tailbeat force test results under each group of test parameters.
2. The method for constructing a tail-beat force prediction model for a vehicle according to claim 1, wherein: The tailbeat characteristic parameters include the tailbeat angle θ0 of the tailbeat phenomenon that occurs during the prototype motion test of the supercavitating vehicle scale model, the angular velocity ω0 when the tailbeat phenomenon occurs, and the navigation speed v0 of the supercavitating vehicle scale model.
3. The method for constructing a tail-beat force prediction model for a vehicle according to claim 2, wherein: Based on the tail beat characteristic parameters, multiple groups of test parameters are constructed, including: The tail beat angle θ with different values within the tail beat angle value range [θ0-Δθ1, θ0+Δθ2], the angular velocity ω with different values within the angular velocity value range [ω0-Δω1, ω0+Δω2], and the sailing speed v with different values within the sailing speed value range [v0-Δv1, v0+Δv2] are combined to construct multiple groups of experimental parameters, each group of experimental parameters includes the tail beat angle θ, the angular velocity ω and the sailing speed v; Among them, Δθ1 and Δθ2 are both angle differences, Δω1 and Δω2 are both angular velocity differences, and Δv1 and Δv2 are both velocity differences.
4. The method for constructing a tail-beat force prediction model for a vehicle according to claim 3, wherein: The tail-beat physical simulation test is carried out using the direct flight high-frequency swing physical test device under each set of test parameters, including for each set of test parameters: Adjusting the liquid surface in the test water tank and the tail-slap water model to a first relative position so that the angle between the axis of the tail-slap water model and the horizontal plane when it contacts the liquid surface in the test water tank during the rotation relative to the horizontal plane is a tail-slap angle θ in the test parameters; controlling the motion of the walking mechanism and the swing mechanism by a main controller according to the sailing speed v and angular velocity ω in the test parameters; and acquiring first test data by a data acquisition device built into the tail-slap water model; Adjusting the liquid surface in the test water tank and the tail-slap water model to a second relative position so that the tail-slap water model does not contact the liquid surface in the test water tank during rotation relative to the horizontal plane, controlling the motion of the traveling mechanism and the swinging mechanism according to the test parameters of the sailing speed v and the angular velocity ω by a main controller, and acquiring second test data by a data acquisition device built into the tail-slap water model; The tail beat force test result under the current test parameters is obtained according to the first test data and the second test data.
5. The method for constructing a tail-beat force prediction model for a vehicle according to claim 4, wherein: Acquiring the first test data and the second test data includes: When the liquid surface and the tail-slapping water model are at the first relative position in the test pool, the sailing speed v in the test parameters is used as the uniform speed of the walking mechanism to control the motion process of the walking mechanism. The rotational angular velocity of the swing mechanism is controlled according to the angular velocity ω in the test parameters. and obtaining first test data through a built-in data acquisition device when the tail-slapping water model is in a stable motion state; When the liquid surface and the tail-slapping water model are at the second relative position in the test pool, the sailing speed v in the test parameters is used as the uniform speed of the walking mechanism to control the motion process of the walking mechanism, and the rotational angular velocity of the swing mechanism is controlled according to the angular velocity ω in the test parameters. and obtaining second test data through a built-in data acquisition device in a stable motion state of the wake water model; Among them, the swing mechanism drives the tail water model to rotate at an angle The rotation angle α of the tail-slapping water model is the angle between the axis of the tail-slapping water model and the horizontal plane, and the rotation angle is positive when it is below the horizontal plane.
6. The method for constructing a tail-beat force prediction model for a vehicle according to claim 5, wherein: The built-in data acquisition device of the tail-slapping water model includes a ring balance, an acceleration sensor, and an angular velocity sensor. The data obtained for each test set include: The swing mechanism is controlled to drive the tail-slapping water model to rotate, and the real-time angular velocity of the tail-slapping water model is detected by an angular velocity sensor; When it is determined that the real-time angular velocity of the tail-slapping water model reaches the angular velocity ω, the walking mechanism is controlled to start accelerating until it reaches the sailing speed v and then maintains a constant speed, and the real-time acceleration of the tail-slapping water model is detected by the acceleration sensor; When the tail lap water model reaches a uniform motion state according to the real-time acceleration of the tail lap water model, it is determined that the tail lap water model reaches a stable motion state, and the force exerted on the tail lap water model at the current test parameters (v, ω, θ) as the rotation angle α is measured by a circular balance to obtain the test data of the liquid surface in the test pool and the tail lap water model in the current relative position.
7. The method for constructing a tail-beat force prediction model for a vehicle according to claim 6, wherein: The tail force test results under the current test parameters obtained based on the first test data and the second test data are: The force F1(v, ω, θ, α) of the tail slap water model at any rotation angle α under the current test parameters (v, ω, θ) in the first test data is subtracted from the force F2(v, ω, θ, α) of the tail slap water model at the same rotation angle α under the current test parameters (v, ω, θ) in the second test data to obtain the tail slap force F at the current rotation angle α under the current test parameters (v, ω, θ). y (v, ω, θ, α), combined rotation angle Tail force F within the range y (v, ω, θ, α) to obtain the tail beat force test results under the current test parameters.
8. The method for constructing a tail-beat force prediction model for a vehicle according to claim 7, wherein: The tailbeat force prediction model is constructed in a data-driven manner based on the tailbeat force test results under various test parameters, including: A fully connected network was constructed using Keras. The tailbeat angle θ, angular velocity ω, navigation speed v, and rotation angle α in each set of tailbeat force test results were used as inputs to the fully connected network. The tailbeat force F at the rotation angle α included in the tailbeat force test results of the current test parameters (v, ω, θ) was calculated. y (v, ω, θ, α) is the output of the fully connected network. The tailbeat force test results under each set of test parameters are used to train the constructed fully connected network to obtain the tailbeat force prediction model.
9. The method for constructing a tail-beat force prediction model for a vehicle according to claim 1, wherein: The constructed direct-flight high-frequency oscillation physical test device also includes a test collection box, which contains a voltage amplifier, a power module and an analog input module. The data acquisition device built into the tail-slapping water model is connected to the test collection box via a collection power supply cable, and the collection power supply cable is connected to the tail-slapping water model and the test collection box respectively by watertight connectors; the power module in the test collection box supplies power to the data acquisition device via the collection power supply cable, and the data acquisition device built into the tail-slapping water model includes a ring balance, an acceleration sensor and an angular velocity sensor, and the ring balance is connected to the voltage amplifier in the test collection box via the collection power supply cable and then connected to the analog input module via a signal isolator. The acceleration sensor is connected to the test collection box via the collection power supply cable and connected to the analog input module via a signal isolator. The angular velocity sensor is connected to the test collection box via the collection power supply cable and connected to the analog input module via a signal isolator; The communication lines of the motor driver in the walking mechanism, the communication lines of the motor driver in the swing mechanism, and the communication lines connected to the analog input module of the test acquisition box are connected to the main controller through the EtherCAT bus respectively; The power lines of the motor driver in the walking mechanism, the motor driver in the swing mechanism, the power lines connected to the power module of the test collection box and the main controller are respectively connected to the external power supply.
10. The method for constructing a tail-beat force prediction model for a vehicle according to claim 9, wherein: The walking mechanism includes a linear motor track and a linear motor slide placed in the linear motor track. The linear motor track is arranged across the test water pool. The motor driver in the walking mechanism drives the linear motor slide to move in the linear motor track. The swing mechanism includes a rotating motor, which is fixed on the linear motor slide. The output shaft of the rotating motor is connected to the axis of the swing wheel. An eccentric shaft is provided on the other side of the swing wheel. The eccentric shaft is movably provided in the keyway of the crank, and the tail water model is assembled at the bottom of the crank; a movably connected ear shaft is provided in the ear shaft hole of the bottom of the crank, and support members are fixed at both ends of the ear shaft, and one end of the support member is fixed to the bottom of the linear motor slide; the output shaft of the rotating motor drives the swing wheel to rotate, and the swing wheel drives the eccentric shaft to rotate in the keyway of the crank, and then drives the crank to rotate along the ear shaft, and the crank drives the tail water model to rotate relative to the horizontal plane.