Attitude calculation method and device for high-speed movement of aspheric penetration body in liquid, electronic equipment and storage medium

Abstract

The invention discloses an attitude calculation method and device for high-speed motion of an aspheric penetration body in liquid, electronic equipment and a storage medium, and relates to the technical field of high-speed motion tests.The method comprises the steps that resistance coefficient data sets of the aspheric penetration body under different working conditions are obtained through finite element simulation, and the resistance coefficient data sets are calculated; constructing a training data set based on the resistance coefficient data sets under the different working conditions; constructing a resistance coefficient model of the aspheric penetration body, and training the resistance coefficient model by using the training data set to obtain a trained resistance coefficient model; establishing a motion control equation when the aspherical penetration body impacts the liquid filling box body; and inputting the working condition information of the target aspheric penetration body into the trained resistance coefficient model to obtain the resistance coefficient of the target aspheric penetration body, and calculating the real-time motion attitude data of the target aspheric penetration body in the liquid by using the resistance coefficient of the target aspheric penetration body and the motion control equation.

Description

Technical Field

[0001] The present invention relates to the technical field of high-speed motion testing, and in particular to a posture calculation method, device, electronic equipment and storage medium for a non-spherical penetrator moving at high speed in a liquid. Background Art

[0002] Engineering problems such as the movement of high-speed projectiles in liquids and the movement of torpedoes in water essentially involve the problem of the fluid motion posture stability of high-speed penetrators. The posture stability of the penetrator in the process of entering the liquid is a complex physical phenomenon. High-speed penetrators that can maintain a stable posture in the air often find it difficult to ensure their own attitude angle stability when moving in liquids, and are prone to rolling. In addition to spherical penetrators, flat penetrators, and specially designed slender penetrators that can maintain underwater posture stability, penetrators in shapes such as cylinders, hemispheres, and conical heads are difficult to have good posture stability in the presence of initial attitude angle disturbances. It is well known that the amount of resistance a penetrator encounters when moving in a fluid depends on the shape of the penetrator, the size of the object, the speed of movement, and its attitude angle. The fluctuation of the attitude angle during the penetrator's entry into the water will cause its drag coefficient C to fluctuate. d The change will cause the real-time resistance of the penetrator to change. However, there is no mature calculation method for accurately predicting the motion trajectory of non-spherical penetrators. Summary of the invention

[0003] The present invention provides a posture calculation method, device, electronic device and storage medium for a non-spherical penetrator moving at high speed in a liquid, so as to solve the technical problem of how to accurately predict the motion trajectory of a non-spherical penetrator moving at high speed in a liquid.

[0004] The embodiment of the present invention provides a method for calculating the posture of a non-spherical penetrator moving at high speed in a liquid, comprising:

[0005] Obtaining a drag coefficient data set of a non-spherical penetrator under different working conditions through finite element simulation, and constructing a training data set based on the drag coefficient data set under different working conditions;

[0006] Constructing a drag coefficient model of a non-spherical penetrator, and training the drag coefficient model using the training data set to obtain a trained drag coefficient model;

[0007] Establish the motion control equation when the non-spherical penetrator hits the liquid-filled box;

[0008] The drag coefficient of the target non-spherical penetrator is obtained by inputting the working condition information of the target non-spherical penetrator into the trained drag coefficient model, and the real-time motion posture data of the target non-spherical penetrator in the liquid is calculated using the drag coefficient of the target non-spherical penetrator and the motion control equation.

[0009] Preferably, obtaining the drag coefficient data set of the non-spherical penetrator under different working conditions through finite element simulation includes:

[0010] Extract characteristic data under different working conditions from the finite element simulation results; wherein the characteristic data include the mass m of the non-spherical penetrator, the real-time velocity v of the non-spherical penetrator, the time t, the liquid density ρ, and the projection area S of the non-spherical penetrator on the plane perpendicular to v;

[0011] According to the motion equation of the non-spherical penetrator in the liquid and the characteristic data under different working conditions, a data set of resistance coefficient of the non-spherical penetrator under different working conditions is obtained;

[0012] Preferably, constructing a training data set based on the drag coefficient data set under different working conditions includes:

[0013] According to the yaw angle ψ and pitch angle θ in the working condition information of each working condition, the sideslip angle β and the angle of attack α of each working condition are obtained;

[0014] By calculating the resistance coefficient C of each working condition d , v, β, α and S are processed one by one to obtain multiple d , v, β, α and S are the training data sets.

[0015] Preferably, the resistance coefficient model is a BPNN (Back propagation neural network) structure, and the BPNN structure includes an input layer, a hidden layer and an output layer in sequence; wherein the input layer is v, β, α and S respectively; the output layer is C d ; The hidden layer is a single hidden layer, and the number of nodes in the hidden layer is n;

[0016] Among them, the number of nodes in the hidden layer is:

[0017]

[0018] In the formula, n1 is the number of input layer nodes; n2 is the number of output layer nodes; b is a constant ranging from 1 to 10; and n is a positive integer.

[0019] Preferably, the motion control equation includes:

[0020]

[0021] Where r is the radius of the non-spherical penetrator; L is the length of the non-spherical penetrator; and t is the movement time of the non-spherical penetrator in the liquid.

[0022] Preferably, using the drag coefficient of the target non-spherical penetrator and the motion control equation, calculating the real-time motion posture data of the target non-spherical penetrator in the liquid includes:

[0023] By inputting the resistance coefficient of the target non-spherical penetrator and the working condition information of the target non-spherical penetrator into the motion control equation, the real-time motion posture data of the target non-spherical penetrator in the liquid is obtained.

[0024] The embodiment of the present invention provides a posture calculation device for a non-spherical penetrator moving at high speed in a liquid, comprising:

[0025] Constructing a training data set module, for obtaining a drag coefficient data set of a non-spherical penetrator under different working conditions through finite element simulation, and constructing a training data set based on the drag coefficient data set under different working conditions;

[0026] A model building and training module is used to build a drag coefficient model of a non-spherical penetrator and train the drag coefficient model using the training data set to obtain a trained drag coefficient model;

[0027] Establishing a motion control equation module, which is used to establish the motion control equation when a non-spherical penetrator hits a liquid-filled box;

[0028] The posture data calculation module is used to obtain the drag coefficient of the target non-spherical penetrator by inputting the working condition information of the target non-spherical penetrator into the trained drag coefficient model, and calculate the real-time motion posture data of the target non-spherical penetrator in the liquid using the drag coefficient of the target non-spherical penetrator and the motion control equation.

[0029] Preferably, the resistance coefficient model is a back propagation neural network BPNN structure, and the BPNN structure includes an input layer, a hidden layer and an output layer in sequence; wherein the input layer is v, β, α and S respectively; the output layer is C d ; The hidden layer is a single hidden layer, and the number of nodes in the hidden layer is n;

[0030] Among them, the number of nodes in the hidden layer is:

[0031]

[0032] In the formula, n1 is the number of input layer nodes; n2 is the number of output layer nodes; b is a constant ranging from 1 to 10; and n is a positive integer.

[0033] An embodiment of the present invention provides an electronic device, comprising: a memory; a processor; and a computer program; wherein the computer program is stored in the memory and is configured to be executed by the processor to implement the steps of the above-mentioned method for calculating the posture of a non-spherical penetrator moving at high speed in a liquid.

[0034] An embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon; the computer program is executed by a processor to implement the steps of the above-mentioned method for calculating the posture of a non-spherical penetrator moving at high speed in a liquid.

[0035] The beneficial effect of the present invention is that, by combining finite element simulation and artificial intelligence algorithm and other methods, a method for constructing a resistance coefficient model of a non-spherical penetrator is given, the motion control equation of the non-spherical penetrator is derived, and the real-time motion posture solution problem of the non-spherical penetrator moving in a liquid is converted into a solution problem of a group of differential equations, and the tumbling phenomenon of the non-spherical penetrator moving in a liquid is efficiently simulated, and quantitative results can be obtained based on the group of differential equations. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 It is a flow chart of a method for calculating the posture of a non-spherical penetrator moving at high speed in a liquid provided by the present invention;

[0037] Figure 2 (a)-(c) are schematic diagrams of the relative positions of the EAS (Earth-fixed Axis System) and the BAS (Body Axis System), the relative positions of the AAS (Air-path Axis System) and the BAS, and the three Euler angles of the projectile provided by the present invention;

[0038] Figure 3 It is the fluid-solid coupling finite element model provided by the present invention;

[0039] Figure 4 (a)-(c) are schematic diagrams of load analysis of the projectile provided by the present invention in the EAS, BAS-1, and BAS three-dimensional coordinate systems, respectively;

[0040] Figure 5 (a)-(b) are respectively the parameters of β, α, v, and C for training the neural network and verifying the prediction results of the neural network under different initial yaw angles ψ0 and initial pitch angles θ0 provided by the present invention. d Schematic diagram of the relationship between;

[0041] Figure 6 It is a schematic diagram of the BPNN structure provided by the present invention;

[0042] Figure 7 (a)-(f) are respectively the C obtained by finite element simulation and neural network prediction under different working conditions provided by the present invention. d A comparative diagram of

[0043] Figure 8 (a)-(d) are schematic diagrams of probability density function diagrams of the neural network model of the training group, validation group, test group and all data provided by the present invention. DETAILED DESCRIPTION

[0044] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention. In the subsequent description, the use of suffixes such as "module", "component" or "unit" used to represent elements is only to facilitate the description of the present invention and has no specific meaning in itself. Therefore, "module", "component" or "unit" can be used in a mixed manner.

[0045] The present invention needs to accurately predict the change of the attitude angle of the penetrator when it rolls, and clarifies the relationship between the attitude angle of the non-spherical penetrator and its motion resistance coefficient C in the fluid. d The relationship between the drag coefficient model and the motion control equation of the non-spherical penetrator is established. By establishing the drag coefficient model and the motion control equation of the non-spherical penetrator, the real-time motion posture solution problem of the non-spherical penetrator moving in the liquid is converted into the solution problem of a group of differential equations. The tumbling phenomenon of the non-spherical penetrator moving in the liquid is efficiently simulated, which is of great significance to promote its military weapon application.

[0046] Figure 1 The flowchart of the method for calculating the posture of a non-spherical penetrator moving at high speed in a liquid provided by the present invention is as follows: Figure 1 As shown, it includes: step S1: obtaining the resistance coefficient data set of the non-spherical penetrator under different working conditions through finite element simulation, and constructing a training data set based on the resistance coefficient data set under different working conditions; step S2: constructing a resistance coefficient model of the non-spherical penetrator, and using the training data set to train the resistance coefficient model to obtain a trained resistance coefficient model; step S3: establishing a motion control equation when the non-spherical penetrator hits a liquid-filled box; step S4: obtaining the resistance coefficient of the target non-spherical penetrator by inputting the working condition information of the target non-spherical penetrator into the trained resistance coefficient model, and using the resistance coefficient of the target non-spherical penetrator and the motion control equation to calculate the real-time motion posture data of the target non-spherical penetrator in the liquid.

[0047] In an embodiment of the present invention, obtaining a resistance coefficient data set of a non-spherical penetrator under different working conditions through finite element simulation includes: extracting characteristic data under different working conditions from the finite element simulation results; wherein the characteristic data includes the mass m of the non-spherical penetrator, the real-time velocity v of the non-spherical penetrator, the time t, the liquid density ρ, and the projection area S of the non-spherical penetrator on a plane perpendicular to v; according to the motion equation of the non-spherical penetrator in the liquid and the characteristic data under different working conditions, obtaining a resistance coefficient data set of the non-spherical penetrator under different working conditions;

[0048] In an embodiment of the present invention, constructing a training data set based on the drag coefficient data set under different working conditions includes: obtaining the sideslip angle β and the angle of attack α of each working condition according to the yaw angle ψ and the pitch angle θ in the working condition information of each working condition; d , v, β, α and S are processed one by one to obtain multiple d , v, β, α and S are the training data sets.

[0049] In the embodiment of the present invention, the resistance coefficient model is a back propagation neural network BPNN structure, which includes an input layer, a hidden layer and an output layer in sequence; wherein the input layer is v, β, α and S respectively; the output layer is C d ; The hidden layer is a single hidden layer, and the number of nodes in the hidden layer is n; wherein the number of nodes in the hidden layer is:

[0050]

[0051] In the formula, n1 is the number of input layer nodes; n2 is the number of output layer nodes; b is a constant ranging from 1 to 10; and n is a positive integer.

[0052] In an embodiment of the present invention, the motion control equation includes:

[0053]

[0054] Where r is the radius of the non-spherical penetrator; L is the length of the non-spherical penetrator; and t is the movement time of the non-spherical penetrator in the liquid.

[0055] In an embodiment of the present invention, the real-time motion posture data of the target non-spherical penetrator in the liquid is calculated using the drag coefficient of the target non-spherical penetrator and the motion control equation, including: obtaining the real-time motion posture data of the target non-spherical penetrator in the liquid by inputting the drag coefficient of the target non-spherical penetrator and the working condition information of the target non-spherical penetrator into the motion control equation.

[0056] An embodiment of the present invention provides a posture calculation device for a non-spherical penetrator moving at high speed in a liquid, comprising: a training data set construction module, used to obtain a drag coefficient data set of the non-spherical penetrator under different working conditions through finite element simulation, and to construct a training data set based on the drag coefficient data set under different working conditions; a model construction and training module, used to construct a drag coefficient model of the non-spherical penetrator, and to train the drag coefficient model using the training data set to obtain a trained drag coefficient model; a motion control equation establishment module, used to establish a motion control equation when the non-spherical penetrator collides with a liquid-filled box; a posture data calculation module, used to obtain the drag coefficient of the target non-spherical penetrator by inputting the working condition information of the target non-spherical penetrator into the trained drag coefficient model, and to calculate the real-time motion posture data of the target non-spherical penetrator in the liquid using the drag coefficient of the target non-spherical penetrator and the motion control equation.

[0057] In the embodiment of the present invention, the drag coefficient model is a back propagation neural network BPNN structure, and the BPNN structure includes an input layer, a hidden layer and an output layer in sequence; wherein the input layer is v, β, α and S respectively; the output layer is C d ; The hidden layer is a single hidden layer, and the number of nodes in the hidden layer is n; wherein the number of nodes in the hidden layer is:

[0058]

[0059] In the formula, n1 is the number of input layer nodes; n2 is the number of output layer nodes; b is a constant, ranging from 1 to 10; n is a positive integer. For example, when n1 = 4; n2 = 1; b = 1, n is approximately 3.23, and the value of n is a positive integer, that is, n = 4.

[0060] An embodiment of the present invention provides an electronic device, comprising: a memory; a processor; and a computer program; wherein the computer program is stored in the memory and is configured to be executed by the processor to implement the steps of the above-mentioned method for calculating the posture of a non-spherical penetrator moving at high speed in a liquid.

[0061] An embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon; the computer program is executed by a processor to implement the steps of the above-mentioned method for calculating the posture of a non-spherical penetrator moving at high speed in a liquid.

[0062] Combine the following Figure 2-Figure 8 The present invention is described in detail, and specifically comprises the following steps:

[0063] S101, obtaining a resistance coefficient data set of a non-spherical penetrator through finite element simulation;

[0064] Firstly, the present invention provides the definition of relevant angles and the conversion relationship between coordinate systems.

[0065] The initial attitude angle of the projectile when it hits the fuel tank is designed as a variable. Euler angles are used to describe the attitude angle of the projectile, that is, the angular relationship between the projectile's ground fixed coordinate system EAS and the body coordinate system BAS. Euler angles consist of three angles: roll angle, yaw angle and pitch angle, respectively. ψ, θ represent. Therefore, the coordinates Represents the state of Euler angles.

[0066] Figure 2 EAS (black), AAS (yellow) and BAS (green) of the projectile. Figure 2 In (a), the blue color represents the coordinate system of EAS after it is rotated around the z-axis, the orange color represents the coordinate system after it is rotated around the y1-axis, and the green color represents the coordinate system after it is rotated around the x2-axis. Figure 2 The orange color in (b) represents the AAS around y a The green color represents the coordinate system after the axis rotation. The orange color represents the coordinate system after the axis rotation around the x3 axis. Figure 2 In (c), two symmetrical initial attitude angle conditions are designed. At this time, the projectile does not deflect when moving in the air. In EAS, the Euler angles of these two working conditions are (0, 0, 0) and (0, 0, -90), respectively.

[0067] Figure 2 (a) Provides the relative position of EAS and BAS. In EAS, the zero coordinate is the geometric center of the projectile; the Z axis is along the trajectory and points in the positive direction; the Y axis is upward along the vertical line, and the X axis, Z axis, and Y axis form a Cartesian coordinate system; Yes b The angle between the y axis and the z axis; b When the positive semi-axis of is on the left side of the plumb bob, is a positive value; ψ is z b The angle between the projection of the axis on the horizontal plane and the z-axis; when z b When the positive semi-axis of the axis is projected to the right of the z-axis, ψ is positive; θ is z b The angle between the axis and the horizontal plane; when z b When the positive semi-axis of the axis lies above the horizontal plane through the origin, θ is positive. Since the projectile used is cylindrical and axisymmetric, attitude equivalence can be considered. Therefore, can be neglected, and the range of ψ and θ is -90° to 0°.

[0068] Figure 2 (b) provides the relative position of AAS and BAS. a Axis and z bThe axis is located in the longitudinal symmetry plane of the projectile. Two angles can describe the angular relationship between the airflow coordinate systems AAS and BAS. They are the sideslip angle β and the angle of attack α. β is the angle between the real-time velocity v of the projectile and the longitudinal symmetry plane of the projectile. b It is positive when the component of the axis is positive. α is the difference between the projection of v on the longitudinal symmetry plane of the projectile and z b The angle between the axes. b When the component of the axis is positive, it is positive. Similar to ψ and θ, in the present invention, β ranges from -90° to 0°, and α ranges from 0° to 90°. The velocity transformation relationship of the projectile in each coordinate system can be expressed as follows:

[0069]

[0070] Where, L z ,L y , and L x is the transformation matrix of the coordinate system rotating around the three axes, and γ is the rotation angle. The transformation relationship between the coordinate systems is given by formulas (4) and (5):

[0071]

[0072] In the formula, v xb ,v yb , and v zb is the velocity of the projectile along the three axes of the BAS. x ,v y , and v z is the velocity of the projectile along the three axes of the EAS.

[0073]

[0074] Where v xa ,v ya , and v za The velocity of the projectile along the three axes of the AAS.

[0075] The drag coefficient data set using the back propagation neural network BPNN structure in the present invention comes from the finite element simulation results. The finite element simulation uses a total of 75 groups of working conditions to obtain the drag coefficient data set of the non-spherical penetrator.

[0076] The S-ALE (Structured Arbitrary Lagrange-Euler) algorithm in LS-DYNAmmp R12.1 was used to numerically simulate the fluid-solid coupling phenomenon of the penetrator impacting the fuel tank. The finite element model is as follows: Figure 3The finite element model mainly includes five parts: projectile, tank, rivet, kerosene and air. The material of the penetrator is 45# steel. The fuel tank is made of 6 riveted 2A12-T4 aluminum alloy plates. The rivets connecting the front and rear plates and the side plates of the fuel tank are made of 5052 aluminum alloy. The rivets connecting the side plates are made of 304L stainless steel.

[0077] In the finite element model, the keyword *CONSTRAINED_LAGRANGE_IN_SOLID is used to set the interaction between fluid and solid, thereby realizing fluid-solid coupling calculation. Each plate of the box uses the Johnson-Cook (i.e. J–C) constitutive model, JC failure model, and Gruneisen state equation. Projectiles and rivets are described by the Plastic Kinematic constitutive model. Air and oil are described by the Null model. The state equations used are the Gruneisen state equation and the Linear Polynomial state equation. The parameters of each component are shown in Tables 1 to 3 below.

[0078] Table 1: Constitutive model parameters used in finite element analysis

[0079]

[0080] Table 2: Failure model parameters used in finite element analysis

[0081]

[0082] Table 3: State equation parameters used for finite element analysis

[0083]

[0084] In the simulation, by changing the initial pitch angle θ0 and initial yaw angle ψ0 of the projectile, the influence of the projectile attitude angle on the hole shape of the fuel tank rear plate was systematically explored. A total of 75 sets of simulation conditions were designed, as shown in Table 4.

[0085] Table 4: Simulation conditions

[0086]

[0087] First, from the simulation results, we can extract the known data such as m, v, t, ρ, S, etc. corresponding to each working condition information, and then calculate the C corresponding to each working condition information through the motion equation of the projectile in kerosene (8). d Among them, the information of each working condition is shown in Table 4 above.

[0088] Secondly, ψ and θ can be calculated by extracting the data in the simulation, and then β and α can be calculated.

[0089] The present invention provides a method for using BPNN to d Prediction. According to the control equation, the attitude of the projectile at the time of launch can be calculated, and the distribution of λ in the ψ0–θ0 space can be obtained. From the simulation results, known data such as m, v, t, ρ, S, etc. can be extracted, and then C can be calculated through formula (8): d The C of the projectile changes in each working condition was calculated by this method. d , and plotted it on θ, v and C d In the four-dimensional space, C d The value of is color mapped, such as Figure 5 As shown, under the condition of ψ0=0°, θ0=0°, C d is significantly greater than other working conditions. At the same time, as v decreases, C d Also gradually decreased. Figure 5 (a) is the training set data, Figure 5 (b) is the validation set data. The model training and validation used a total of 75 working conditions and 11,250 sets of data, with a speed range of 400 m·s -1 –1600m·s -1 .

[0090] S102. Construct a resistance coefficient model of a non-spherical penetrator through an artificial intelligence algorithm;

[0091] Figure 6 The BPNN structure is provided, which mainly includes input layer, hidden layer and output layer. The input layer has 4 sets of data, namely v, β, α, S. The output layer has only 1 set of data, namely C d For the hidden layer, a single hidden layer is used, and the number of nodes in the hidden layer is n. The selection of n will directly affect the prediction accuracy of the model. The present invention compares the C obtained by finite element simulation. d The C is predicted by the drag coefficient model (BPNN for short) using different n. d By comparing the correlation coefficient R value, we can find that as the number of nodes increases, the accuracy of BPNN prediction results is also improving. However, when n reaches 7, the number of nodes increases again, but the R value does not improve significantly, and even decreases. When n = 7, the C predicted by BPNN is d Compared with the C obtained by finite element simulation d The R value is closest to 1 and is greater than 0.94, which indicates that the prediction effect is the best, indicating that the BPNN with n=7 can accurately predict C d , so it is used as the drag coefficient model.

[0092] exist Figure 7 middle, Figure 7(a) ψ0 = 0°, θ0 = -90° and the initial velocity of the non-spherical penetrator v0 = 1600 m·s -1 ; Figure 7 (b) ψ0 = -30°, θ0 = -45° and v0 = 1600 m·s -1 ; Figure 7 (c) ψ0 = -45°, θ0 = 0° and v0 = 1300 m·s -1 ; Figure 7 (d) ψ0 = -45°, θ0 = -60° and v0 = 1300 m·s -1 ; Figure 7 (e) ψ0 = -60°, θ0 = -30° and v0 = 1200 m·s -1 ; Figure 7 (f) is ψ0 = -90°, θ0 = -60° and v0 = 1200 m·s -1 .

[0093] During the training process, BPNN divides the training set data into three parts: training group, validation group and test group. Figure 8 The probability density diagram of the BPNN prediction results when n = 7 is given. The horizontal axis in the figure is calculated using finite elements, and the vertical axis is the prediction of BPNN in normalized space. The correlation coefficient R value of each set of data is higher than 0.94, indicating that there is a strong correlation between the output and the target, verifying the accuracy of the BPNN prediction results. Figure 8 (a) is the probability density diagram of the training group; Figure 8 (b) is the probability density diagram of the validation group; Figure 8 (c) is the probability density diagram of the test group; Figure 8 (d) is the probability density diagram of all data.

[0094] S3. Based on Newton's law, establish the motion control equation of the non-spherical penetrating body;

[0095] The projectile decelerates and rotates in the fluid. The tumbling of the projectile in kerosene is due to the uneven force. Figure 4 The three-dimensional loading analysis of a projectile in kerosene is given. Figure 4 (a) is EAS; Figure 4 (b) is BAS-1 (rotated around the y-axis of EAS); Figure 4 (c) is BAS (rotated around the x1 axis of BAS-1). When the projectile moves in the oil, the curvature of the trajectory is very small and can be ignored. AAS coincides with EAS. When the projectile rotates ψ around the y axis of EAS, EAS becomes BAS-1; when the projectile continues to rotate θ around the x1 axis of BAS-1, BAS-1 becomes BAS.

[0096] In EAS, the projectile is subjected to balanced forces in the x-axis and y-axis directions. The equation of motion in the z-axis direction is:

[0097]

[0098] In the formula, F xb ,F yb , and F zb is the force of the projectile in the three axes of the BAS; F x ,F y and F z It is the force of the projectile in the three axis directions in the EAS.

[0099] According to the law of conservation of momentum, the equation of motion of the projectile in kerosene is:

[0100]

[0101] (F zb -F xb )l=IJ ψ , (9)

[0102] (F zb -F yb )l=IJ θ , (10)

[0103] Where l is the radius of rotation, I is the polar moment of inertia, and J ψ is the angular acceleration of ψ, J θ is the angular acceleration of θ.

[0104] Combining formula (6)-formula (10) and simplifying the result, we get:

[0105]

[0106] C d =C d (v,β,α,s), (12)

[0107] Where r = D / 2.C d There is a functional relationship between v, β, α and S, which can be obtained through BPNN prediction. Formulas (11) and (12) together define the motion control equation when the projectile hits the liquid-filled box.

[0108] S104, calculating the real-time motion posture of the non-spherical penetrator by numerical calculation method;

[0109] Given any different initial values ​​of the working conditions (i.e., v0, ψ0, θ0), the formulas (11) and (12) are numerically solved using the 4th-5th order Runge-Kutta algorithm to solve the real-time motion posture of the non-spherical penetrator.

[0110] Example

[0111] Assume that the target cylindrical penetrator moves at high speed in water. The penetrator is made of 45# steel, with a bottom radius r of 1 cm, a length L of 1 cm, and a density ρ of 8.75 g / cm 3 , so the mass m of the penetrator is 6.165g, and the density ρ of water is 0.775g / cm 3 , the initial velocity (v0) of the penetrator is 1300 m / s, and the initial attitude angle is (ψ0 = -60°, θ0 = -30°);

[0112] Because |β|=|ψ|, |α|=|θ|, so β0=-60°, α0=30°; S is related to β and α, so S can be further calculated;

[0113] S is related to the shape of the penetrator. For example, if the target non-spherical penetrator is a cylinder, the formula for calculating S is:

[0114]

[0115] That is, S = 1.2415 cm 2

[0116] Since v0, β0, α0, and S are known, the initial C of the target penetrator can be obtained by inputting v0, β0, α0, and S into the trained drag coefficient model. d It is 0.5088.

[0117] The real-time posture of the target penetrator can be obtained by solving formula (20) using the 4-5th order Runge-Kutta algorithm:

[0118]

[0119] v is v0, ψ is ψ0, and θ is θ0; if we want to know the posture of the penetrator when it moves in the liquid for 170μs, the calculation end time is 170, and one step is calculated every 1μs, so there are 170 steps in total. d , we can obtain v, ψ, θ, S, and C for the next 1μs d , then at the final moment ψ=-80.17°, θ=-34.37°.

[0120] The preferred embodiments of the present invention are described above with reference to the accompanying drawings, but the scope of the present invention is not limited thereby. Any modification, equivalent substitution and improvement made by a person skilled in the art without departing from the scope and essence of the present invention shall be within the scope of the present invention.

Claims

1. A method for calculating the posture of a non-spherical penetrator moving at high speed in a liquid, characterized in that: include: Obtaining a drag coefficient data set of a non-spherical penetrator under different working conditions through finite element simulation, and constructing a training data set based on the drag coefficient data set under different working conditions; Constructing a drag coefficient model of a non-spherical penetrator, and training the drag coefficient model using the training data set to obtain a trained drag coefficient model; Establish the motion control equation when the non-spherical penetrator hits the liquid-filled box; The drag coefficient of the target non-spherical penetrator is obtained by inputting the working condition information of the target non-spherical penetrator into the trained drag coefficient model, and the real-time motion posture data of the target non-spherical penetrator in the liquid is calculated using the drag coefficient of the target non-spherical penetrator and the motion control equation.

2. The method according to claim 1, characterized in that: Through finite element simulation, the drag coefficient data sets of non-spherical penetrators under different working conditions are obtained, including: Extract characteristic data under different working conditions from the finite element simulation results; wherein the characteristic data include the mass m of the non-spherical penetrator, the real-time velocity v of the non-spherical penetrator, the time t, the liquid density ρ, and the projection area S of the non-spherical penetrator on the plane perpendicular to v; According to the motion equation of the non-spherical penetrator in the liquid and the characteristic data under different working conditions, a data set of the resistance coefficient of the non-spherical penetrator under different working conditions is obtained.

3. The method according to claim 2, characterized in that Constructing a training data set based on the drag coefficient data set under different working conditions includes: According to the yaw angle ψ and pitch angle θ in the working condition information of each working condition, the sideslip angle β and the angle of attack α of each working condition are obtained; By calculating the resistance coefficient C of each working condition d , v, β, α and S are processed one by one to obtain multiple d , v, β, α and S are the training data sets.

4. The method according to claim 3, characterized in that The drag coefficient model is a back propagation neural network BPNN structure, which includes an input layer, a hidden layer and an output layer in sequence; wherein the input layer is v, β, α and S respectively; the output layer is C d ; The hidden layer is a single hidden layer, and the number of nodes in the hidden layer is n; Among them, the number of nodes in the hidden layer is: In the formula, n1 is the number of input layer nodes; n2 is the number of output layer nodes; b is a constant ranging from 1 to 10; and n is a positive integer.

5. The method according to claim 2, characterized in that: The motion control equations include: Where r is the radius of the non-spherical penetrator; L is the length of the non-spherical penetrator; and t is the movement time of the non-spherical penetrator in the liquid.

6. The method according to claim 5, characterized in that Calculating the real-time motion posture data of the target non-spherical penetrator in the liquid by using the drag coefficient of the target non-spherical penetrator and the motion control equation includes: By inputting the resistance coefficient of the target non-spherical penetrator and the working condition information of the target non-spherical penetrator into the motion control equation, the real-time motion posture data of the target non-spherical penetrator in the liquid is obtained.

7. A posture calculation device for a non-spherical penetrator moving at high speed in a liquid, characterized in that: include: Constructing a training data set module, for obtaining a drag coefficient data set of a non-spherical penetrator under different working conditions through finite element simulation, and constructing a training data set based on the drag coefficient data set under different working conditions; A model building and training module is used to build a drag coefficient model of a non-spherical penetrator and train the drag coefficient model using the training data set to obtain a trained drag coefficient model; Establishing a motion control equation module, which is used to establish the motion control equation when a non-spherical penetrator hits a liquid-filled box; The posture data calculation module is used to obtain the drag coefficient of the target non-spherical penetrator by inputting the working condition information of the target non-spherical penetrator into the trained drag coefficient model, and calculate the real-time motion posture data of the target non-spherical penetrator in the liquid using the drag coefficient of the target non-spherical penetrator and the motion control equation.

8. The device according to claim 7, characterized in that The drag coefficient model is a back propagation neural network BPNN structure, which includes an input layer, a hidden layer and an output layer in sequence; wherein the input layer is v, β, α and S respectively; the output layer is C d ; The hidden layer is a single hidden layer, and the number of nodes in the hidden layer is n; Among them, the number of nodes in the hidden layer is: In the formula, n1 is the number of input layer nodes; n2 is the number of output layer nodes; b is a constant ranging from 1 to 10; and n is a positive integer.

9. An electronic device, characterized in that: include: Memory; processor; and a computer program; wherein the computer program is stored in the memory and is configured to be executed by the processor to implement the method according to any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that: A computer program is stored thereon; the computer program is executed by a processor to implement the method according to any one of claims 1 to 6.

Images