Construction method and calculation method of target damage probability calculation model under fragment strike

Abstract

The invention discloses a construction method and a calculation method of a target damage probability calculation model under fragment strike, belonging to the field of damage assessment, comprising: importing unit node information; dividing the surface of each component of an aircraft into a triangular facet grid in finite element software, numbering the obtained node information and facet information of each component respectively, and importing them into Matlab in a matrix format, and establishing an aircraft facet grid node coordinate system; the damage probability calculation method based on the Kriging proxy model is adopted, and compared with the Monte Carlo method, the number of calculations of fragment intersection, fragment limit penetration speed and fragment residual speed can be effectively reduced, and the calculation efficiency of the target damage probability can be improved; in addition, the invention adopts a U learning function to further improve the modeling efficiency of the Kriging proxy model, and further improve the overall calculation efficiency of the target damage probability; and solves the technical problem of low efficiency of the calculation method of the target damage probability in the prior art.

Description

Technical Field

[0001] The present invention belongs to the field of damage assessment, and relates to a target vulnerability analysis method, in particular to a method for constructing a target damage probability calculation model under a fragment strike and a method for calculating the same. Background Art

[0002] When studying the vulnerability of targets or the damage effectiveness of ammunition on targets, calculating the probability of target damage is a very important task. According to the damage probability calculation formula, based on the law of large numbers, the Monte Carlo method can be used to easily solve the result. Researchers have proposed many methods for solving damage indicators based on the Monte Carlo method. For example, Li Jianming and others gave a method for calculating the damage probability of target components based on the Monte Carlo method earlier; Suga Lujing and others combined fuzzy reasoning with the Monte Carlo method, and calculated the reliability and damage probability of target components based on the finite element analysis of thermal damage of target components; Fan Zhifeng and others used the Monte Carlo method to calculate the number of grenades of a certain type required to damage a solid bunker under certain conditions; Ye Jun and others used the Monte Carlo method to calculate the probability of a fuel-air bomb blocking a bunker, and verified the calculation results using numerical simulation and damage tests.

[0003] Although the Monte Carlo method is a relatively simple method for calculating target damage indicators, its theoretical basis is the law of large numbers. This characteristic determines that it requires a very large number of calculations to ensure the accuracy of the calculation results, resulting in its low calculation efficiency. If finite element numerical simulation calculations are introduced, it will cause unbearable calculation costs. In recent years, in the field of structural reliability analysis, the Kriging proxy model has been widely used in the calculation of structural failure probability, greatly improving the calculation efficiency. At the same time, both the damage probability and the failure probability are calculated by random simulation, which are very similar in mathematical models.

[0004] Therefore, in order to improve the efficiency of target damage probability calculation, a damage probability calculation method based on the Kriging proxy model and the shooting line method is urgently needed, with fragments hitting aircraft as a typical scenario. At the same time, an adaptive learning strategy is introduced to provide a technical approach to reduce the cost of target damage probability calculation. Summary of the invention

[0005] In view of the shortcomings of the prior art, the purpose of the present invention is to provide a method for constructing a model and a calculation method for calculating the probability of target damage under fragment strikes, so as to solve the technical problem of low efficiency of the calculation method of the target damage probability in the prior art.

[0006] In order to solve the above technical problems, the present invention adopts the following technical solutions to achieve the above problems:

[0007] A method for constructing a target damage probability calculation model under fragment strike, the method comprising the following steps:

[0008] Step 1: Import unit node information;

[0009] In the finite element software, the surfaces of each aircraft component are divided into triangular facet meshes, the node information and facet information of each component are numbered respectively, and imported into Matlab in matrix format, and the aircraft facet mesh node coordinate system is established;

[0010] The facet information includes the material density of the component, the material ultimate strength of the component and the thickness of the component;

[0011] Step 2: Set fragmentation parameters;

[0012] Set the fragment's initial coordinates, initial direction vector, density, mass, diameter, and initial velocity v c ;

[0013] Set a preset strike area, and determine the azimuth and elevation angle of the fragments according to the preset strike area;

[0014] Step 3: node coordinate transformation;

[0015] According to the azimuth and pitch angle of the fragment incident determined in step 2, the generated aircraft facet grid node coordinate system is transformed into a shooting line projection coordinate system to obtain the coordinates of the aircraft triangular facet grid in the shooting line projection coordinate system;

[0016] Step 4, generate the shooting line;

[0017] Determine the strike area according to the coordinates of each triangular facet mesh of the aircraft in the shooting line projection coordinate system;

[0018] The Monte Carlo method is used to randomly generate M aiming points in the strike area according to normal distribution, and the initial coordinates of the fragments and the coordinates of the aiming points are connected to generate M shooting lines.

[0019] Step 5, selecting W shooting lines from the M shooting lines obtained in step 4 as a sample set, and judging whether the shooting line i intersects with the triangular facet mesh of the aircraft component j, wherein i is 1, 2, ..., W; j is 1, 2, ..., N;

[0020] If the shooting line i intersects with the triangular surface element grid of component j, it is determined that the fragments hit the triangular surface element grid, and the intersection point is obtained, and the process goes to step 6; otherwise, the shooting line does not intersect with the triangular surface element grid, and it is determined that the fragments do not hit the triangular surface element grid, let i=i+1, and determine i≤W, if yes, continue to determine, otherwise go to step 8;

[0021] Step 6: Calculate the limit penetration velocity V' of the fragments according to the following formula: i,j , m / s;

[0022]

[0023] Where:

[0024] h j represents the thickness of component j, m;

[0025] d represents the fragment diameter, m;

[0026] α ij is the angle between the shooting line i and the normal vector of component j, rad;

[0027] ρ t j Represents the material density of component j, kg / m 3 ;

[0028] ρ p Represents the density of the fragment, kg / m 3 ;

[0029] σ j represents the material ultimate strength of component j, MPa;

[0030] Step 7: Subtract the maximum penetration velocity of the fragments obtained in step 6 from the initial velocity v of the fragments in step 2. c Compare and determine whether the fragments penetrate the aircraft components; if the fragments do not penetrate the aircraft components, then count the hit information, set i=i+1, and return to step 5; if the fragments penetrate the aircraft components, calculate the residual speed of the fragments, set j=j+1, and return to step 5;

[0031] The hit information includes the hit component number and the firing line number;

[0032] The remaining speed of the fragments The calculation formula is as follows:

[0033]

[0034] Where:

[0035] Indicates the residual velocity of the fragments, m / s;

[0036] v c represents the initial velocity, m / s;

[0037] m represents the mass of the fragment, kg;

[0038] Represents the mass of the block produced when the fragment hits component j along the shooting line i, kg;

[0039] Step 8, determining the input variable set and the output variable set;

[0040] The density, mass, diameter and initial velocity v of the fragment obtained in step 2 c The coordinates of the W aiming points obtained in step 4 form W sets of data and use them as the input variable set;

[0041] The hit information obtained in step 7, the limit penetration speed of the fragments obtained in step 6, and the remaining speed of the fragments obtained in step 7 are used as the output variable set;

[0042] Step nine, constructing a calculation model for the probability of target damage under the impact of fragments;

[0043] The hit information in the input variable set and the output variable set obtained in step eight is used as the first training sample set; the limit penetration speed of the fragments in the input variable set and the output variable set obtained in step eight is used as the second training sample set; the residual speed of the fragments in the input variable set and the output variable set obtained in step eight is used as the third training sample set;

[0044] The first training sample set, the second training sample set and the third training sample set are respectively imported into the toolbox of the Kriging proxy model in Matlab, and the Kriging proxy model for calculating the rendezvous and hit of fragments, the Kriging proxy model for calculating the limit penetration speed of fragment penetration and the Kriging proxy model for calculating the residual speed of fragments are respectively obtained, that is, the calculation model of the target damage probability under the impact of fragments is obtained;

[0045] Step 10, training the target damage probability calculation model under the impact of fragments to obtain a trained final target damage probability calculation model under the impact of fragments;

[0046] Respectively taking the shooting lines outside the first training sample set, the second training sample set and the third training sample set as the first candidate training sample set, the second candidate training sample set and the third candidate training sample set;

[0047] The first candidate training sample set, the second candidate training sample set and the third candidate training sample set are respectively input into the fragment rendezvous hit calculation Kriging proxy model, the fragment penetration limit penetration speed calculation Kriging proxy model and the fragment residual speed calculation Kriging proxy model for training, so as to obtain the trained fragment rendezvous hit calculation Kriging proxy model, the fragment penetration limit penetration speed calculation Kriging proxy model and the fragment residual speed calculation Kriging proxy model, that is, to obtain the trained final fragment strike target damage probability calculation model.

[0048] The step five specifically includes the following steps:

[0049] Step 5.1, the parametric equation of any point in the triangular surface element mesh of component j is formula (1);

[0050] P=(1-uv)V1+uV2+vV3

[0051] E1=V2-V1, E2=V3-V1, T=O-V1 (1)

[0052] Where:

[0053] V1, V2, and V3 represent the three vertices of the triangular facet mesh;

[0054] u and v are the weights of V2 and V3 respectively;

[0055] Step 5.2, the parametric equation of any point on the shooting line i is formula (2);

[0056] P=O+tD (2)

[0057] Where:

[0058] O and D represent the initial coordinates and initial direction vector of the shooting line, respectively;

[0059] Step 5.3, by solving formula (1) and formula (2) together, we can obtain formula (3):

[0060]

[0061] Wherein, E1=V2-V1, E2=V3-V1, T=O-V1;

[0062] Step 5.4, determine whether there is a real number solution to formula (3). If there is a real number solution, the shooting line intersects with the triangular surface element grid, and it is determined that the fragments hit the triangular surface element grid, and the intersection point is obtained, and the process proceeds to step 6; otherwise, the shooting line does not intersect with the triangular surface element grid, and it is determined that the fragments do not hit the triangular surface element grid, and the process proceeds to step 5.5;

[0063] Step 5.5, let i=i+1, determine if i≤W, if so, return to step 5.1, otherwise go to step eight.

[0064] The step ten specifically includes the following steps:

[0065] Step 10.1, taking the shooting lines outside the first training sample set, the second training sample set and the third training sample set as the first candidate training sample set, the second candidate training sample set and the third candidate training sample set respectively;

[0066] Step 10.2, respectively inputting the first candidate training sample set, the second candidate training sample set and the third candidate training sample set obtained in step 10.1 into the Kriging proxy model for calculating the fragment intersection hit, the Kriging proxy model for calculating the limit penetration velocity of fragment penetration and the Kriging proxy model for calculating the residual velocity of fragments, to obtain the hit information, the limit penetration velocity of fragment penetration and the residual velocity of fragments, as well as the standard deviations corresponding to the three;

[0067] Step 10.3, the hit information, the limit penetration velocity of fragment penetration, the residual velocity of fragments and the standard deviations corresponding to the three obtained in step 10.2 are respectively substituted into the active learning U function to calculate the U value, and the minimum U value corresponding to the Kriging proxy model for calculating the fragment intersection hit, the Kriging proxy model for calculating the limit penetration velocity of fragment penetration and the Kriging proxy model for calculating the residual velocity of fragments is selected;

[0068] The expression of the active learning U function is:

[0069] σ G(X) Indicates the hit information, the maximum penetration speed of the fragments or the residual speed of the fragments;

[0070] μ G( X) indicates the hit information, the maximum penetration velocity of the fragments or the standard deviation of the residual velocity of the fragments;

[0071] Step 10.4, determine in turn whether the three minimum U values ​​obtained in step 10.3 are greater than the preset convergence value. If the three minimum U values ​​are all greater than the preset convergence value, the target damage probability calculation model under the final fragment strike is obtained; if one or two of the minimum U values ​​are greater than the preset convergence value, the corresponding Kriging proxy model stops training and keeps the corresponding minimum U value unchanged. The Kriging proxy model with the minimum U value less than the preset convergence value enters step 10.5; if the three minimum U values ​​are all less than the preset convergence value, the three Kriging proxy models all enter step 10.5;

[0072] Step 10.5, select the shooting line corresponding to the minimum U value of the Kriging proxy model, and set the aircraft component j = 1;

[0073] Step 10.6, performing intersection calculation on the shooting line selected in step 10.5 and the plane component surface element grid, and judging whether the selected shooting line intersects with the triangle surface element grid of the plane component j, if so, proceeding to step 10.7; otherwise, proceeding to step 10.8; wherein j is 1, 2, ..., N;

[0074] Step 10.7, calculate the maximum penetration velocity of the fragments, and determine whether the fragments penetrate the aircraft components; if so, calculate the residual velocity of the fragments, set j = j + 1, and return to step 10.6; if not, count the hit information and go to step 10.8;

[0075] Step 10.8, the coordinates of the aiming point corresponding to the shooting line selected in step 10.5, as well as the density, mass, diameter and initial velocity v of the corresponding fragments c As input samples, the hit information, the maximum penetration velocity of the fragments, and the residual velocity of the fragments obtained in step 10.7 are used as output samples respectively, and the three sample pairs consisting of the input sample and the output sample are added to the first training sample set, the second training sample set, and the third training sample set respectively to obtain the updated first training sample set, the second training sample set, and the third training sample set;

[0076] Step 10.9, import the updated first training sample set, second training sample set and third training sample set obtained in step 10.8 into the toolbox of Kriging proxy model in Matlab, respectively, obtain the updated Kriging proxy model for fragment intersection and hit calculation, the Kriging proxy model for fragment penetration limit penetration velocity calculation and the Kriging proxy model for fragment residual velocity calculation, respectively, and return to step 10.1.

[0077] A method for calculating the probability of target damage under a fragment strike, based on a method for constructing a model for calculating the probability of target damage under a fragment strike, specifically comprises the following steps:

[0078] S1, inputting the candidate training sample set into the final target damage probability calculation model under the impact of fragments obtained in the construction method of the target damage probability calculation model under the impact of fragments, and obtaining the hit information of each fragment, the maximum penetration speed of the fragment penetration, and the residual speed of the fragment;

[0079] S2, obtain the number of fragments M that hit the aircraft based on the hit information obtained in S1 * , calculate the hit probability P M ;

[0080] P M =M * / M;

[0081] S3, according to the fragment penetration limit speed and fragment residual speed obtained in S1, determine the situation of the fragment penetrating the aircraft parts, and obtain the number of fragments that penetrated the aircraft M # , calculate the damage probability P H ;

[0082] P H =M # / M;

[0083] S4, adjust the fragment diameter, material density, material ultimate strength and initial velocity v c , repeat S1 to S3, calculate the probability of hitting and damaging the aircraft by the fragments, and obtain the curve of the probability of hitting the aircraft parts varying with the fragment parameters and the curve of the probability of damaging the aircraft parts varying with the fragment parameters.

[0084] Compared with the prior art, the present invention has the following beneficial technical effects:

[0085] (I) Compared with the Monte Carlo method, the damage probability calculation method based on the Kriging proxy model adopted by the present invention can effectively reduce the number of calculations of fragment intersection, fragment limit penetration velocity and fragment residual velocity, and improve the calculation efficiency of the target damage probability; in addition, the U learning function adopted by the present invention can further improve the modeling efficiency of the Kriging proxy model and further improve the overall calculation efficiency of the target damage probability; thus, the technical problem of low efficiency of the target damage probability calculation method in the prior art is solved.

[0086] (II) The present invention combines the shooting line method with the Kriging proxy model, avoids the finite element simulation process of fragment penetration, realizes the algorithmization and integration of the calculation of the probability of damage to the target by fragments, and improves the convenience and efficiency of the calculation.

[0087] (III) The present invention combines the limiting penetration velocity of fragment penetration, the residual velocity formula and the firing line method, enriches the physical connotation of the target damage probability calculation model under fragment strike, and provides a basis for the functional damage probability calculation.

[0088] (V) The method for calculating the probability of target damage under fragment strike provided by the present invention can simultaneously calculate the probability of fragment hit and the probability of damage. BRIEF DESCRIPTION OF THE DRAWINGS

[0089] Figure 1 It is a schematic flow chart of the method of the present invention;

[0090] Figure 2 It is a schematic diagram of the shooting line of the present invention;

[0091] Figure 3 is a histogram of the probability of fragments hitting the aircraft calculated by the present invention;

[0092] Figure 4 It is a radar diagram of the probability of damage to the aircraft by fragments calculated by the present invention;

[0093] Figure 5 is a comparison diagram of the damage probability curve calculated by the present invention and the Monte Carlo method;

[0094] The specific contents of the present invention are further explained in detail below in conjunction with embodiments. DETAILED DESCRIPTION

[0095] It should be noted that, unless otherwise specified, all components in the present invention are components known in the art.

[0096] Specific embodiments of the present invention are given below. It should be noted that the present invention is not limited to the following specific embodiments, and all equivalent changes made on the basis of the technical solution of this application fall within the protection scope of the present invention.

[0097] A method for constructing a target damage probability calculation model under fragment strike, the method comprising the following steps:

[0098] Step 1: Import unit node information;

[0099] In the finite element software, the surfaces of each aircraft component are divided into triangular facet meshes, the node information and facet information of each component are numbered respectively, and imported into Matlab in matrix format, and the aircraft facet mesh node coordinate system is established;

[0100] The facet information includes the material density of the component, the material ultimate strength of the component and the thickness of the component;

[0101] Step 2: Set fragmentation parameters;

[0102] Set the fragment's initial coordinates, initial direction vector, density, mass, diameter, and initial velocity v c ;

[0103] Set a preset strike area, and determine the azimuth and elevation angle of the fragments according to the preset strike area;

[0104] Step 3: node coordinate transformation;

[0105] According to the azimuth and pitch angle of the fragment incident determined in step 2, the generated aircraft facet grid node coordinate system is transformed into a shooting line projection coordinate system to obtain the coordinates of the aircraft triangular facet grid in the shooting line projection coordinate system;

[0106] Step 4, generate the shooting line;

[0107] Determine the strike area according to the coordinates of each triangular facet mesh of the aircraft in the shooting line projection coordinate system;

[0108] The Monte Carlo method is used to randomly generate M aiming points in the strike area according to normal distribution, and the initial coordinates of the fragments and the coordinates of the aiming points are connected to generate M shooting lines.

[0109] Step 5, selecting W shooting lines from the M shooting lines obtained in step 4 as a sample set, and judging whether the shooting line i intersects with the triangular facet mesh of the aircraft component j, wherein i is 1, 2, ..., W; j is 1, 2, ..., N;

[0110] If the shooting line i intersects with the triangular surface element grid of component j, it is determined that the fragments hit the triangular surface element grid, and the intersection point is obtained, and the process goes to step 6; otherwise, the shooting line does not intersect with the triangular surface element grid, and it is determined that the fragments do not hit the triangular surface element grid, let i=i+1, and determine i≤W, if yes, continue to determine, otherwise go to step 8;

[0111] Step 6: Calculate the limit penetration velocity V' of the fragments according to the following formula: i,j , m / s;

[0112]

[0113] Where:

[0114] h j represents the thickness of component j, m;

[0115] d represents the fragment diameter, m;

[0116] α ij is the angle between the shooting line i and the normal vector of component j, rad;

[0117] ρ t j Represents the material density of component j, kg / m 3 ;

[0118] ρ p Represents the density of the fragment, kg / m 3 ;

[0119] σ j represents the material ultimate strength of component j, MPa;

[0120] Step 7: Subtract the maximum penetration velocity of the fragments obtained in step 6 from the initial velocity v of the fragments in step 2. c Compare and determine whether the fragments penetrate the aircraft components; if the fragments do not penetrate the aircraft components, then count the hit information, set i=i+1, and return to step 5; if the fragments penetrate the aircraft components, calculate the residual speed of the fragments, set j=j+1, and return to step 5;

[0121] The hit information includes the hit component number and the firing line number;

[0122] The remaining speed of the fragments The calculation formula is as follows:

[0123]

[0124] Where:

[0125] Indicates the residual velocity of the fragments, m / s;

[0126] v c represents the initial velocity, m / s;

[0127] m represents the mass of the fragment, kg;

[0128] Represents the mass of the block produced when the fragment hits component j along the shooting line i, kg;

[0129] Step 8, determining the input variable set and the output variable set;

[0130] The density, mass, diameter and initial velocity v of the fragment obtained in step 2 c The coordinates of the W aiming points in step 5 form W groups of data, which are used as the input variable set;

[0131] The hit information obtained in step 7, the limit penetration speed of the fragments obtained in step 6, and the remaining speed of the fragments obtained in step 7 are used as the output variable set;

[0132] Step nine, constructing a calculation model for the probability of target damage under the impact of fragments;

[0133] The hit information in the input variable set and the output variable set obtained in step eight is used as the first training sample set; the limit penetration speed of the fragments in the input variable set and the output variable set obtained in step eight is used as the second training sample set; the residual speed of the fragments in the input variable set and the output variable set obtained in step eight is used as the third training sample set;

[0134] The first training sample set, the second training sample set and the third training sample set are respectively imported into the toolbox of the Kriging proxy model in Matlab, and the Kriging proxy model for calculating the rendezvous and hit of fragments, the Kriging proxy model for calculating the limit penetration speed of fragment penetration and the Kriging proxy model for calculating the residual speed of fragments are respectively obtained, that is, the calculation model of the target damage probability under the impact of fragments is obtained;

[0135] Step 10, training the target damage probability calculation model under the impact of fragments to obtain a trained final target damage probability calculation model under the impact of fragments;

[0136] Respectively taking the shooting lines outside the first training sample set, the second training sample set and the third training sample set as the first candidate training sample set, the second candidate training sample set and the third candidate training sample set;

[0137] The first candidate training sample set, the second candidate training sample set and the third candidate training sample set are respectively input into the fragment rendezvous hit calculation Kriging proxy model, the fragment penetration limit penetration speed calculation Kriging proxy model and the fragment residual speed calculation Kriging proxy model for training, so as to obtain the trained fragment rendezvous hit calculation Kriging proxy model, the fragment penetration limit penetration speed calculation Kriging proxy model and the fragment residual speed calculation Kriging proxy model, that is, to obtain the trained final fragment strike target damage probability calculation model.

[0138] In the above technical scheme, the damage probability calculation method based on the Kriging proxy model adopted can effectively reduce the number of calculations of fragment intersection, fragment limit penetration velocity and fragment residual velocity, and improve the calculation efficiency of the target damage probability compared with the Monte Carlo method; in addition, the present invention adopts the U learning function to further improve the modeling efficiency of the Kriging proxy model, and further improve the overall calculation efficiency of the target damage probability; it solves the technical problem of low efficiency of the calculation method of the target damage probability in the prior art.

[0139] Combining the shooting line method with the Kriging proxy model avoids the finite element simulation process of fragment penetration, realizes the algorithmization and integration of the calculation of the probability of damage to the target by fragments, and improves the convenience and efficiency of the calculation.

[0140] Combining the limiting penetration speed of fragment penetration, the residual speed formula and the shooting line method enriches the physical connotation of the target damage probability calculation model under fragment strikes and provides a basis for the calculation of functional damage probability.

[0141] In the above technical solution, in step 3, the coordinate conversion steps of the fragments are as follows:

[0142] Assuming the coordinates in the original coordinate system are (x, y, z), the formula for transforming the azimuth from 0 to α is as follows

[0143]

[0144] The formula for transforming the pitch angle from 0 to θ is as follows.

[0145]

[0146] Finally (x2, y2, z2) is the transformed projection coordinate system coordinate.

[0147] In step 5, assume that the aircraft includes N = {N1, N2, ..., N N} parts, generate M shooting lines, and calculate the intersection of each component face element with W of the M shooting lines, that is, a double-layer nested for loop is used, the inner layer is the shooting line loop, and the outer layer is the aircraft component loop. A total of W×N intersection calculations need to be carried out to record the number of shooting lines that intersect and the face element information hit by the shooting lines.

[0148] In step 6, according to the three damage modes of fragment penetration, namely, ricochet, embedment, and penetration, the fragment penetration is used as the damage criterion for aircraft components. According to the parameters such as the mass and velocity of the fragments carried by the shooting line, the material parameters and structural thickness parameters of the surface element where the shooting line intersects, and the angle between the normal vector of the shooting line and the surface element, the limit penetration velocity of the fragment penetration is calculated and compared with the initial velocity v of the fragment shooting line. c Compare and determine whether the fragments have penetrated the aircraft components;

[0149] In step nine, the expression of the Kriging proxy model is as follows

[0150] G(X)=F(X,β)+z(X)=f(X)β+z(X)

[0151] Where X is the design sample point, G(X) is the function to be fitted, f(X) is the polynomial of X, β is the regression coefficient of f(X), F(X,β) represents the average value of the Kriging model, and z(X) is a random distribution.

[0152] The predicted value and predicted variance of the Kriging surrogate model can be expressed as

[0153]

[0154] Where f is the n-dimensional unit column vector, r(X) is the correlation function between the prediction point and the design sample point, R is the correlation coefficient matrix, and Y is the n-dimensional column vector composed of the responses of the design sample points.

[0155] The toolbox used to build the Kriging surrogate model is the DACE toolbox;

[0156] The coordinates of the starting point and direction vector of the shooting line can be adjusted according to the import of targets such as aircraft and research needs.

[0157] The step five specifically includes the following steps:

[0158] Step 5.1, the parametric equation of any point in the triangular surface element mesh of component j is formula (1);

[0159] P=(1-uv)V1+uV2+vV3

[0160] E1=V2-V1, E2=V3-V1, T=O-V1 (1)

[0161] Where:

[0162] V1, V2, and V3 represent the three vertices of the triangular facet mesh;

[0163] u and v are the weights of V2 and V3 respectively;

[0164] Step 5.2, the parametric equation of any point on the shooting line i is formula (2);

[0165] P=O+tD (2)

[0166] Where:

[0167] O and D represent the initial coordinates and initial direction vector of the shooting line, respectively;

[0168] Step 5.3, by solving formula (1) and formula (2) together, we can obtain formula (3):

[0169]

[0170] Wherein, E1=V2-V1, E2=V3-V1, T=O-V1;

[0171] Step 5.4, determine whether there is a real number solution to formula (3). If there is a real number solution, the shooting line intersects with the triangular surface element grid, and it is determined that the fragments hit the triangular surface element grid, and the intersection point is obtained, and the process proceeds to step 6; otherwise, the shooting line does not intersect with the triangular surface element grid, and it is determined that the fragments do not hit the triangular surface element grid, and the process proceeds to step 5.5;

[0172] Step 5.5, let i=i+1, determine if i≤W, if so, return to step 5.1, otherwise go to step eight.

[0173] The step ten specifically includes the following steps:

[0174] Step 10.1, taking the shooting lines outside the first training sample set, the second training sample set and the third training sample set as the first candidate training sample set, the second candidate training sample set and the third candidate training sample set respectively;

[0175] Step 10.2, respectively inputting the first candidate training sample set, the second candidate training sample set and the third candidate training sample set obtained in step 10.1 into the Kriging proxy model for calculating the fragment intersection hit, the Kriging proxy model for calculating the limit penetration velocity of fragment penetration and the Kriging proxy model for calculating the residual velocity of fragments, to obtain the hit information, the limit penetration velocity of fragment penetration and the residual velocity of fragments, as well as the standard deviations corresponding to the three;

[0176] Step 10.3, the hit information, the limit penetration velocity of fragment penetration, the residual velocity of fragments and the standard deviations corresponding to the three obtained in step 10.2 are respectively substituted into the active learning U function to calculate the U value, and the minimum U value corresponding to the Kriging proxy model for calculating the fragment intersection hit, the Kriging proxy model for calculating the limit penetration velocity of fragment penetration and the Kriging proxy model for calculating the residual velocity of fragments is selected;

[0177] The expression of the active learning U function is:

[0178] σ G(X) Indicates the hit information, the maximum penetration speed of the fragments or the residual speed of the fragments;

[0179] μ G(X) Indicates the hit information, the maximum penetration velocity of the fragments or the standard deviation of the residual velocity of the fragments;

[0180] Step 10.4, determine in turn whether the three minimum U values ​​obtained in step 10.3 are greater than the preset convergence value. If the three minimum U values ​​are all greater than the preset convergence value, the target damage probability calculation model under the final fragment strike is obtained; if one or two of the minimum U values ​​are greater than the preset convergence value, the corresponding Kriging proxy model stops training and keeps the corresponding minimum U value unchanged. The Kriging proxy model with the minimum U value less than the preset convergence value enters step 10.5; if the three minimum U values ​​are all less than the preset convergence value, the three Kriging proxy models all enter step 10.5;

[0181] Step 10.5, select the shooting line corresponding to the minimum U value of the Kriging proxy model, and set the aircraft component j = 1;

[0182] Step 10.6, performing intersection calculation on the shooting line selected in step 10.5 and the plane component surface element grid, and judging whether the selected shooting line intersects with the triangle surface element grid of the plane component j, if so, proceeding to step 10.7; otherwise, proceeding to step 10.8; wherein j is 1, 2, ..., N;

[0183] Step 10.7, calculate the maximum penetration velocity of the fragments, and determine whether the fragments penetrate the aircraft components; if so, calculate the residual velocity of the fragments, set j = j + 1, and return to step 10.6; if not, count the hit information and go to step 10.8;

[0184] Step 10.8, the coordinates of the aiming point corresponding to the shooting line selected in step 10.5, as well as the density, mass, diameter and initial velocity v of the corresponding fragments cAs input samples, the hit information, the maximum penetration velocity of the fragments, and the residual velocity of the fragments obtained in step 10.7 are used as output samples respectively, and the three sample pairs consisting of the input sample and the output sample are added to the first training sample set, the second training sample set, and the third training sample set respectively to obtain the updated first training sample set, the second training sample set, and the third training sample set;

[0185] Step 10.9, import the updated first training sample set, second training sample set and third training sample set obtained in step 10.8 into the toolbox of Kriging proxy model in Matlab, respectively, obtain the updated Kriging proxy model for fragment intersection and hit calculation, the Kriging proxy model for fragment penetration limit penetration velocity calculation and the Kriging proxy model for fragment residual velocity calculation, respectively, and return to step 10.1.

[0186] The present invention also provides a method for calculating the probability of target damage under the impact of fragments. The method for constructing a model for calculating the probability of target damage under the impact of fragments specifically includes the following steps:

[0187] S1, inputting the candidate training sample set into the final target damage probability calculation model under the impact of fragments obtained in the construction method of the target damage probability calculation model under the impact of fragments, and obtaining the hit information of each fragment, the maximum penetration speed of the fragment penetration, and the residual speed of the fragment;

[0188] S2, obtain the number of fragments M that hit the aircraft based on the hit information obtained in S1 * , calculate the hit probability P M ;

[0189] P M =M * / M;

[0190] S3, according to the fragment penetration limit speed and fragment residual speed obtained in S1, determine the situation of the fragment penetrating the aircraft parts, and obtain the number of fragments that penetrated the aircraft M # , calculate the damage probability P H ;

[0191] P H =M # / M;

[0192] S4, adjust the fragment diameter, material density, material ultimate strength and initial velocity v c , repeat S1 to S3, calculate the probability of hitting and damaging the aircraft by the fragments, and obtain the curve of the probability of hitting the aircraft parts varying with the fragment parameters and the curve of the probability of damaging the aircraft parts varying with the fragment parameters.

[0193] In the above technical solution, the provided method for calculating the probability of target damage under fragment strike can simultaneously calculate the probability of fragment hit and the probability of damage.

[0194] Example:

[0195] This embodiment provides a method for calculating the probability of target damage under the impact of fragments, where the fragments are Φ10mm tungsten balls and the initial velocity v c is 1300m / s, set the initial coordinates of the fragment to (0,0,0), the initial direction vector to (1,0,0), and the density ρ p 17600kg / m 3 , mass is 9.3g; where M is 1000 and W is 50;

[0196] Here are the relevant parameters of the aircraft shell: h j =20mm, σ j =410MPa;

[0197] Here, the preset convergence value of the Kriging agent model active learning U function is 2;

[0198] According to the aircraft projection and research requirements, 11 fragment attack directions are preset, and the corresponding pitch angles and azimuth angles are shown in the following table

[0199]

[0200] At the same time, the Monte Carlo method is used to calculate the probability of fragments hitting the aircraft and the probability of damage corresponding to 11 working conditions, and the calculation accuracy of the method of the present invention is verified by comparison.

[0201] Figure 3 and Figure 4 The aircraft hit probability and damage probability calculated by the method of the present invention are shown respectively. Figure 5 The comparison between the damage probability result calculated by the method of the present invention and the Monte Carlo calculation result is shown. The method of the present invention only takes 247.38 seconds, while the Monte Carlo method takes 3354.9 seconds. The calculation time of the method of the present invention is much less than that of the Monte Carlo method, and the calculation efficiency is greatly improved. At the same time, the damage probability curves obtained by the two methods are almost the same, which proves the calculation accuracy of the present invention.

Claims

1. A method for constructing a target damage probability calculation model under fragment strike, characterized in that: The method comprises the following steps: Step 1: Import unit node information; In the finite element software, the surfaces of each aircraft component are divided into triangular facet meshes, the node information and facet information of each component are numbered respectively, and imported into Matlab in matrix format, and the aircraft facet mesh node coordinate system is established; The facet information includes the material density of the component, the material ultimate strength of the component and the thickness of the component; Step 2: Set fragmentation parameters; Set the fragment's initial coordinates, initial direction vector, density, mass, diameter, and initial velocity v c ; Set a preset strike area, and determine the azimuth and elevation angle of the fragments according to the preset strike area; Step 3: node coordinate transformation; According to the azimuth and pitch angle of the fragment incident determined in step 2, the generated aircraft facet grid node coordinate system is transformed into a shooting line projection coordinate system to obtain the coordinates of the aircraft triangular facet grid in the shooting line projection coordinate system; Step 4, generate the shooting line; Determine the strike area according to the coordinates of each triangular facet mesh of the aircraft in the shooting line projection coordinate system; The Monte Carlo method is used to randomly generate M aiming points in the strike area according to normal distribution, and the initial coordinates of the fragments and the coordinates of the aiming points are connected to generate M shooting lines. Step 5, selecting W shooting lines from the M shooting lines obtained in step 4 as a sample set, and judging whether the shooting line i intersects with the triangular facet mesh of the aircraft component j, wherein i is 1, 2, ..., W; j is 1, 2, ..., N; If the shooting line i intersects with the triangular surface element grid of component j, it is determined that the fragments hit the triangular surface element grid, and the intersection point is obtained, and the process goes to step 6; otherwise, the shooting line does not intersect with the triangular surface element grid, and it is determined that the fragments do not hit the triangular surface element grid, let i=i+1, and determine i≤W, if yes, continue to determine, otherwise go to step 8; Step 6: Calculate the limit penetration velocity V' of the fragments according to the following formula: i,j , m / s; Where: h j represents the thickness of component j, m; d represents the fragment diameter, m; α ij is the angle between the shooting line i and the normal vector of component j, rad; ρ t j Represents the material density of component j, kg / m 3 ; ρ p Represents the density of the fragment, kg / m 3 ; σ j represents the material ultimate strength of component j, MPa; Step 7: Subtract the maximum penetration velocity of the fragments obtained in step 6 from the initial velocity v of the fragments in step 2. c Compare and determine whether the fragments penetrate the aircraft components; if the fragments do not penetrate the aircraft components, then count the hit information, set i=i+1, and return to step 5; if the fragments penetrate the aircraft components, calculate the residual speed of the fragments, set j=j+1, and return to step 5; The hit information includes the hit component number and the firing line number; The remaining speed of the fragments The calculation formula is as follows: Where: Indicates the residual velocity of the fragments, m / s; v c represents the initial velocity, m / s; m represents the mass of the fragment, kg; Represents the mass of the block produced when the fragment hits component j along the shooting line i, kg; Step 8, determining the input variable set and the output variable set; The density, mass, diameter and initial velocity v of the fragment obtained in step 2 c The coordinates of the W aiming points in step 5 form W groups of data, which are used as the input variable set; The hit information obtained in step 7, the limit penetration speed of the fragments obtained in step 6, and the remaining speed of the fragments obtained in step 7 are used as the output variable set; Step nine, constructing a calculation model for the probability of target damage under the impact of fragments; The hit information in the input variable set and the output variable set obtained in step eight is used as the first training sample set; the limit penetration speed of the fragments in the input variable set and the output variable set obtained in step eight is used as the second training sample set; the residual speed of the fragments in the input variable set and the output variable set obtained in step eight is used as the third training sample set; The first training sample set, the second training sample set and the third training sample set are respectively imported into the toolbox of the Kriging proxy model in Matlab, and the Kriging proxy model for calculating the rendezvous and hit of fragments, the Kriging proxy model for calculating the limit penetration speed of fragment penetration and the Kriging proxy model for calculating the residual speed of fragments are respectively obtained, that is, the calculation model of the target damage probability under the impact of fragments is obtained; Step 10, training the target damage probability calculation model under the impact of fragments to obtain a trained final target damage probability calculation model under the impact of fragments; Respectively taking the shooting lines outside the first training sample set, the second training sample set and the third training sample set as the first candidate training sample set, the second candidate training sample set and the third candidate training sample set; The first candidate training sample set, the second candidate training sample set and the third candidate training sample set are respectively input into the fragment rendezvous hit calculation Kriging proxy model, the fragment penetration limit penetration speed calculation Kriging proxy model and the fragment residual speed calculation Kriging proxy model for training, and the trained fragment rendezvous hit calculation Kriging proxy model is obtained. The Kriging proxy model for calculating the limit penetration speed of fragment penetration and the Kriging proxy model for calculating the residual speed of fragments are used to obtain the trained model for calculating the probability of target damage under the impact of the final fragments.

2. The method for constructing a target damage probability calculation model under fragment strike according to claim 1, characterized in that: The step five specifically includes the following steps: Step 5.1, the parametric equation of any point in the triangular surface element mesh of component j is formula (1); P=(1-uv)V1+uV2+vV3 E1=V2-V1, E2=V3-V1, T=O-V1 (1) Where: V1, V2, and V3 represent the three vertices of the triangular facet mesh; u and v are the weights of V2 and V3 respectively; Step 5.2, the parametric equation of any point on the shooting line i is formula (2); P=O+tD (2) Where: O and D represent the initial coordinates and initial direction vector of the shooting line, respectively; Step 5.3, by solving formula (1) and formula (2) together, we can obtain formula (3): Wherein, E1=V2-V1, E2=V3-V1, T=O-V1; Step 5.4, determine whether there is a real number solution to formula (3). If there is a real number solution, the shooting line intersects with the triangular surface element grid, and it is determined that the fragments hit the triangular surface element grid, and the intersection point is obtained, and the process proceeds to step 6; otherwise, the shooting line does not intersect with the triangular surface element grid, and it is determined that the fragments do not hit the triangular surface element grid, and the process proceeds to step 5.5; Step 5.5, let i=i+1, determine if i≤W, if so, return to step 5.1, otherwise go to step eight.

3. The method for constructing a target damage probability calculation model under fragment strike according to claim 1, characterized in that: The step ten specifically includes the following steps: Step 10.1, taking the shooting lines outside the first training sample set, the second training sample set and the third training sample set as the first candidate training sample set, the second candidate training sample set and the third candidate training sample set respectively; Step 10.2, respectively inputting the first candidate training sample set, the second candidate training sample set and the third candidate training sample set obtained in step 10.1 into the Kriging proxy model for calculating the fragment intersection hit, the Kriging proxy model for calculating the limit penetration velocity of fragment penetration and the Kriging proxy model for calculating the residual velocity of fragments, to obtain the hit information, the limit penetration velocity of fragment penetration and the residual velocity of fragments, as well as the standard deviations corresponding to the three; Step 10.3, the hit information, the limit penetration velocity of fragment penetration, the residual velocity of fragments and the standard deviations corresponding to the three obtained in step 10.2 are respectively substituted into the active learning U function to calculate the U value, and the minimum U value corresponding to the Kriging proxy model for calculating the fragment intersection hit, the Kriging proxy model for calculating the limit penetration velocity of fragment penetration and the Kriging proxy model for calculating the residual velocity of fragments is selected; The expression of the active learning U function is: σ G(X) Indicates the hit information, the maximum penetration speed of the fragments or the residual speed of the fragments; μ G(X) Indicates the hit information, the maximum penetration velocity of the fragments or the standard deviation of the residual velocity of the fragments; Step 10.4, determine in turn whether the three minimum U values ​​obtained in step 10.3 are greater than the preset convergence value. If the three minimum U values ​​are all greater than the preset convergence value, the target damage probability calculation model under the final fragment strike is obtained; if one or two of the minimum U values ​​are greater than the preset convergence value, the corresponding Kriging proxy model stops training and keeps the corresponding minimum U value unchanged. The Kriging proxy model with the minimum U value less than the preset convergence value enters step 10.5; if the three minimum U values ​​are all less than the preset convergence value, the three Kriging proxy models all enter step 10.5; Step 10.5, select the shooting line corresponding to the minimum U value of the Kriging proxy model, and set the aircraft component j = 1; Step 10.6, performing intersection calculation on the shooting line selected in step 10.5 and the plane component surface element grid, and judging whether the selected shooting line intersects with the triangle surface element grid of the plane component j, if so, proceeding to step 10.7; otherwise, proceeding to step 10.8; wherein j is 1, 2, ..., N; Step 10.7, calculate the maximum penetration velocity of the fragments, and determine whether the fragments penetrate the aircraft components; if so, calculate the residual velocity of the fragments, set j = j + 1, and return to step 10.6; if not, count the hit information and go to step 10.8; Step 10.8, the coordinates of the aiming point corresponding to the shooting line selected in step 10.5, as well as the density, mass, diameter and initial velocity v of the corresponding fragments c As input samples, the hit information, the maximum penetration velocity of the fragments, and the residual velocity of the fragments obtained in step 10.7 are used as output samples respectively, and the three sample pairs consisting of the input sample and the output sample are added to the first training sample set, the second training sample set, and the third training sample set respectively to obtain the updated first training sample set, the second training sample set, and the third training sample set; Step 10.9, import the updated first training sample set, second training sample set and third training sample set obtained in step 10.8 into the toolbox of Kriging proxy model in Matlab, respectively, obtain the updated Kriging proxy model for fragment intersection and hit calculation, the Kriging proxy model for fragment penetration limit penetration velocity calculation and the Kriging proxy model for fragment residual velocity calculation, respectively, and return to step 10.

1.

4. A method for calculating the probability of target damage under fragment strike, characterized in that: The method for constructing a target damage probability calculation model under the impact of fragments according to any one of claims 1 to 3 specifically comprises the following steps: S1, inputting the candidate training sample set into the final target damage probability calculation model under the impact of fragments obtained in the construction method of the target damage probability calculation model under the impact of fragments, and obtaining the hit information of each fragment, the maximum penetration speed of the fragment penetration, and the residual speed of the fragment; S2, obtain the number of fragments M that hit the aircraft based on the hit information obtained in S1 * , calculate the hit probability P M ; P M =M * / M; S3, according to the fragment penetration limit speed and fragment residual speed obtained in S1, determine the situation of the fragment penetrating the aircraft parts, and obtain the number of fragments that penetrated the aircraft M # , calculate the damage probability P H ; P H =M # / M; S4, adjust the fragment diameter, material density, material ultimate strength and initial velocity v c , repeat S1 to S3, calculate the probability of hitting and damaging the aircraft by the fragments, and obtain the curve of the probability of hitting the aircraft parts varying with the fragment parameters and the curve of the probability of damaging the aircraft parts varying with the fragment parameters.

Images