Structural design method and device for amorphous alloy shaped charge liner and medium

Abstract

The invention discloses a structural design method and device for an amorphous alloy shaped charge liner and a medium, and relates to the technical field of shaped charge liners. Obtaining a plurality of initial values of each structural design parameter of the amorphous alloy shaped charge liner; according to the multiple initial values of each structural design parameter, orthogonal experiment design is carried out, and multiple structural design schemes of the amorphous alloy shaped charge liner are obtained; numerical simulation is conducted on the amorphous alloy shaped charge liner under each structural design scheme, and at least one jet flow forming index of each structural design scheme is obtained; according to the multiple structural design schemes and the corresponding at least one jet flow forming index, a neural network model is constructed; and optimizing according to the neural network model to obtain an optimal structural design scheme of the amorphous alloy shaped charge liner. According to the method, the optimal structural design scheme of the amorphous alloy shaped charge liner can be determined, and the jet flow forming effect of the amorphous alloy shaped charge liner is improved.

Description

Technical Field

[0001] The invention relates to the technical field of medicine liner, and in particular to a structural design method, device and medium of an amorphous alloy medicine liner. Background Art

[0002] The liner material is an important factor in improving the penetration performance of shaped charge warheads. In recent years, the rapidly developed Zr amorphous material has shown an energy release effect under impact conditions and has a high density. It can be used in the liner to promote the formation of a stable metal jet, which can effectively improve its armor-piercing ability and has a good application prospect in the defense industry.

[0003] In the prior art, a new type of charge liner structure is designed with W-Zr based amorphous alloy, which has good aftereffect and perforation performance. The energy release characteristics of Zr based amorphous material are used to improve the hole expansion effect of the perforating bullet, and the released energy can achieve the purpose of cleaning the hole.

[0004] However, the amorphous alloy-based liner structures in the prior art are all constructed based on historical experience, and the jet forming effect of the amorphous alloy liner constructed in this way is poor. Summary of the invention

[0005] Based on this, it is necessary to provide a structural design method, device and medium for an amorphous alloy liner to address the above technical problems. This method can determine the optimal structural design scheme of the amorphous alloy liner and improve the jet forming effect of the amorphous alloy liner.

[0006] The present invention adopts the following technical solutions:

[0007] The present invention provides a method for designing the structure of an amorphous alloy liner, comprising:

[0008] Obtaining multiple initial values ​​of each structural design parameter of the amorphous alloy liner;

[0009] According to multiple initial values ​​of each structural design parameter, orthogonal experimental design is carried out to obtain multiple structural design schemes of amorphous alloy liner;

[0010] Numerical simulations are performed on the amorphous alloy liner under each structural design scheme to obtain at least one jet forming index of each structural design scheme;

[0011] Constructing a neural network model according to a plurality of structural design schemes and at least one corresponding jet forming index;

[0012] The optimal structural design of the amorphous alloy liner is obtained by optimizing the model of the neural network.

[0013] Preferably, a neural network model is constructed according to a plurality of structural design schemes and at least one corresponding jet forming index, including:

[0014] For any jet forming index, the values ​​of the jet forming index under each structural design scheme are normalized respectively;

[0015] Under each structural design scheme, the normalized values ​​of each jet forming index are weighted according to the weight of each jet forming index to obtain the comprehensive evaluation index under each structural design scheme;

[0016] A neural network model is constructed based on multiple structural design schemes and corresponding comprehensive evaluation indexes.

[0017] Preferably, the jet forming index includes jet length, jet head velocity and jet total energy; the comprehensive evaluation index is calculated as follows:

[0018]

[0019] Among them, W represents the comprehensive evaluation index, W a represents the jet length, represents the weight of the jet length, W b represents the jet head velocity, represents the weight of the jet head velocity, W c represents the total energy of the jet, Represents the weight of the total energy of the jet.

[0020] Preferably, a neural network model is constructed according to a plurality of structural design schemes and corresponding comprehensive evaluation indexes, including:

[0021] For any structural design parameter, the initial value of the structural design parameter under multiple structural design schemes is fitted with the corresponding comprehensive evaluation index to obtain a fitting function under the structural design parameter; the fitting function represents the corresponding relationship between the structural design parameter and the comprehensive evaluation index;

[0022] Determine the training samples through the fitting function under each structural design parameter;

[0023] Train the neural network model through training samples.

[0024] Preferably, the training samples are determined by fitting functions under various structural design parameters, including:

[0025] For any structural design parameter, according to the fitting function under the structural design parameter, multiple sample values ​​of the structural design parameter and the corresponding comprehensive evaluation index are obtained;

[0026] According to the multiple sample values ​​of each structural design parameter, multiple sample structural design schemes are determined, and the average value of the comprehensive evaluation index corresponding to each sample value under each sample structural design scheme is determined as the comprehensive evaluation index of each sample structural design scheme;

[0027] A variety of sample structure design schemes and corresponding comprehensive evaluation indexes are determined as training samples.

[0028] Preferably, the optimal structural design scheme of the amorphous alloy liner is obtained by optimizing according to the neural network model, including:

[0029] The neural network model is used as the objective function of the genetic algorithm, and multiple structural design parameters are used as the variables of the genetic algorithm. The structural design parameters of the amorphous alloy charge liner are optimized through the genetic algorithm, and the structural design parameter value corresponding to the largest objective function value in the iterative process of the genetic algorithm is determined as the optimal structural design scheme.

[0030] Optionally, numerical simulation is performed on the amorphous alloy liner under each structural design scheme to obtain at least one jet forming index of each structural design scheme, including:

[0031] For any structural design scheme, in the shaped charge finite element model, boundary conditions are added in the air domain and the Euler mesh is divided. Then, the amorphous alloy liner is numerically simulated under the structural design scheme through the numerical simulation algorithm.

[0032] Optionally, the material of the amorphous alloy liner is W skeleton / Zr-based amorphous alloy, and the amorphous alloy liner is a conical liner.

[0033] The present invention provides a structural design device for an amorphous alloy liner, comprising:

[0034] An acquisition module, used to acquire multiple initial values ​​of each structural design parameter of the amorphous alloy liner;

[0035] A design module is used to perform orthogonal experimental design based on multiple initial values ​​of each structural design parameter to obtain multiple structural design schemes of amorphous alloy liner;

[0036] A simulation module is used to numerically simulate the amorphous alloy liner under each structural design scheme to obtain at least one jet forming index of each structural design scheme;

[0037] A construction module, used to construct a neural network model according to a plurality of structural design schemes and at least one corresponding jet forming index;

[0038] The optimization module is used to perform optimization based on the neural network model to obtain the optimal structural design solution for the amorphous alloy charge liner.

[0039] The present invention provides a computer-readable storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a processor, the structural design method of the amorphous alloy liner is implemented.

[0040] The present invention provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the structural design method of the amorphous alloy liner when executing the program.

[0041] At least one of the above technical solutions adopted by the present invention can achieve the following beneficial effects:

[0042] In the present invention, by obtaining multiple initial values ​​of each structural design parameter of the amorphous alloy liner and performing orthogonal experimental design, the horizontal combination of each factor can be evenly dispersed and representative, covering various possible structural design schemes, thereby comprehensively examining the influence of different parameter combinations on the performance of the liner, and providing a basis for finding the optimal solution; the jet forming index can quantitatively reflect the jet forming effect of the liner, and the advantages and disadvantages of jet forming under different structural design schemes can be intuitively understood, and then a neural network model is constructed according to a variety of structural design schemes and corresponding jet forming indicators, which can automatically mine the complex relationship between the structural design parameters and the jet forming indicators from a large amount of data without clarifying the specific mathematical expression of this relationship in advance, thereby fitting the data more accurately, providing a reliable model basis for optimization, and finally using the constructed neural network model for optimization, the optimal structural design scheme can be quickly searched in the entire design parameter space. This optimization method based on the neural network model can avoid local optimal solutions and find the true global optimal structural design scheme, thereby maximizing the jet forming effect of the amorphous alloy liner. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0044] Figure 1 A schematic diagram of a structural design method of an amorphous alloy liner provided by the present invention;

[0045] Figure 2 A schematic diagram of a shaped charge structure provided by the present invention;

[0046] Figure 3 A schematic diagram of a shaped charge finite element model provided by the present invention;

[0047] Figure 4A schematic diagram of the structure of a neural network provided by the present invention;

[0048] Figure 5 A neural network prediction error graph provided by the present invention;

[0049] Figure 6 A linear regression diagram of a prediction result obtained by a neural network and an actual fitting result provided by the present invention;

[0050] Figure 7 A graph showing changes in mean square error of a training set, a validation set, a test set and the whole set with the number of training times provided by the present invention;

[0051] Figure 8 A schematic diagram of a structural design device for an amorphous alloy liner provided by the present invention;

[0052] Fig. 9 A schematic diagram of a computer device for implementing a structural design method of an amorphous alloy liner provided by the present invention. DETAILED DESCRIPTION

[0053] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the specific embodiments of the present invention and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0054] In the prior art, a lot of research has been conducted on charge liner materials. For example, ballistic tests and numerical simulations have been used to describe the energy release process of Zr-based amorphous alloy fragments, and the reaction mechanism of Zr-based amorphous alloy fragments has been well explained. A new type of shaped charge structure has been designed using W-Zr-based amorphous alloy. This structure has good aftereffect and perforation performance, and the energy release characteristics of Zr-based amorphous materials are used to improve the hole expansion effect of perforating bullets. Tungsten is a crystalline metal with extremely excellent mechanical properties. It can be used as a reinforcing phase and combined with Zr-based amorphous alloys to obtain excellent composite materials. A W skeleton / Zr-based amorphous alloy composite material with a porosity of 30% was prepared by a hydraulic process, and its compression performance was studied. The test results show that the maximum compressive strength of the composite material reaches 2764Mpa, and the plastic strain reaches 39.4%, which is much higher than that of Zr-based amorphous alloy and tungsten skeleton; the mechanical properties of W skeleton / Zr-based amorphous alloy composite materials prepared by hydraulic process and infiltration casting are compared. The results show that the composite material prepared by hydraulic process has higher strength, while the cast composite material has better plasticity.

[0055] However, due to the late development of W skeleton / Zr-based amorphous composite materials, their application in warheads is still relatively rare.

[0056] Based on this, the present invention uses W skeleton / Zr-based amorphous alloy composite material as the liner material, studies its jet forming and target penetration characteristics, carries out numerical simulation and experimental research, and optimizes the liner structure to achieve the purpose of improving the liner's jet forming and target penetration capabilities.

[0057] The technical solutions provided by various embodiments of the present invention are described in detail below in conjunction with the accompanying drawings.

[0058] Figure 1 The present invention is a schematic diagram of a structural design method of an amorphous alloy liner, which specifically includes the following steps:

[0059] S101, obtaining multiple initial values ​​of each structural design parameter of the amorphous alloy liner.

[0060] The liner is a key component of the shaped charge. The material selection and structural design of the liner have an important influence on the jet formation and penetration capability of the shaped charge.

[0061] Among them, the material of the amorphous alloy liner is W skeleton / Zr-based amorphous alloy, and the amorphous alloy liner is a conical liner; the present invention adopts a new energetic material W skeleton / Zr-based amorphous alloy as the liner material, such as Figure 2 As shown, Figure 2 It is a shaped charge structure, which adopts a conical charge liner, the wall thickness δ of the charge liner is 1mm, the diameter D1 is 48.59mm, and the height H1 of the charge liner is 35mm. The charge diameter D2 is 50mm, the charge height H2 is 60mm, the charge liner cone angle is 60°, and the inner arc radius R is 8mm.

[0062] The present invention selects three structural design parameters of the liner, namely the cone angle, wall thickness and inner arc radius, as factors affecting the jet forming of the shaped charge structure, and analyzes the influence of different combinations of these three structural design parameters on the jet forming.

[0063] Therefore, multiple initial values ​​of each structural design parameter can be obtained first, and the initial values ​​can be determined based on historical experience, as shown in Table 1.

[0064] Table 1

[0065]

[0066]

[0067] S102, performing orthogonal experimental design based on multiple initial values ​​of each structural design parameter to obtain multiple structural design schemes of the amorphous alloy liner.

[0068] Take the example of three structural design parameters, each of which has four values, which is equivalent to three factors and four levels. Therefore, L can be selected. 16 (43) Orthogonal table, which includes 16 experimental combinations, each combination corresponds to a different factor matching level, that is, there are 16 structural design schemes, each structural design scheme corresponds to a different combination of structural design parameter values.

[0069] S103, numerically simulating the amorphous alloy liner under each structural design scheme to obtain at least one jet forming index for each structural design scheme.

[0070] Among them, in the numerical simulation of shaped charge jet, material models such as explosive, liner, air domain and target plate are mainly used. The AUTODYN-2D material library contains material models for selection. The material models and parameters of the liner and target plate are shown in Table 2.

[0071] The air domain can be regarded as an ideal polytropic gas and described by the ideal gas state equation, which is as follows:

[0072] p=(γ-1)ρe (1)

[0073] Where p is the pressure, γ is the adiabatic index 1.4, and e is the initial energy 2.068×10 5 J / kg, ρ is the density 1.225×10 -3 g / cm 3 .

[0074] In the simulation process, the JWL state equation is used to describe the TNT explosive material model and the expansion work process of the detonation products. Its expression is:

[0075]

[0076] Wherein, p is the pressure of the detonation product; V is the relative specific volume of the detonation product; A, B, R1, R2, ω and E0 are the parameters to be input, and the specific parameters are shown in Table 3 below. Table 3 is the parameters of the projectile-target Johnson-Cook model.

[0077] The warhead and target plate materials adopt the Johnson-Cook strength-failure model, which is widely used in the fields of mechanical processing, explosion and high-speed impact, and consider the strain hardening effect of the material and the influence of strain rate and temperature on strength and plasticity. The stress-strain relationship is expanded as follows:

[0078]

[0079] Among them, σ eff , ε p are equivalent stress and equivalent plastic strain respectively; A, B, n, C, and m are all unknown parameters determined by experiments.

[0080] Table 2

[0081]

[0082] Table 3

[0083]

[0084] In one embodiment, numerical simulation is performed on the amorphous alloy liner under each structural design scheme to obtain at least one jet forming index of each structural design scheme, including:

[0085] For any structural design scheme, in the shaped charge finite element model, boundary conditions are added in the air domain and the Euler mesh is divided. Then, the amorphous alloy liner is numerically simulated under the structural design scheme through the numerical simulation algorithm.

[0086] Specifically, the shaped charge finite element model is as follows: Figure 3 As shown in the figure, the detonation mode is initiated by the center of the bottom of the explosive. In order to eliminate the boundary effect, the "Flow-out" boundary condition is added to the boundary of the air domain. The Euler grid is divided into 0.3mm×0.3mm sizes. The shaped charge and the liner are both based on the Euler algorithm suitable for large deformation simulation. The target part uses the Lagrange algorithm. The jet forming and penetration of the target plate process uses the fluid-solid coupling algorithm.

[0087] Therefore, the amorphous alloy liner under each structural design scheme was numerically simulated by the shaped charge finite element model to obtain at least one jet forming index for each structural design scheme.

[0088] The orthogonal design method can obtain the importance ranking of each factor and the comprehensive optimization of multiple factors. The present invention selects jet length, jet head speed, and total jet energy as the performance indicators of jet forming. The influence of each indicator on jet forming is different, and the proportion is also different. The three structural design parameters of the cone angle, wall thickness, and inner arc radius of the liner are selected as factors affecting the jet forming of the shaped charge structure. Based on the influence of different combinations of these three structural design parameters on jet forming, orthogonal test analysis is performed to obtain a variety of structural design schemes, and then the amorphous alloy liner under each structural design scheme is numerically simulated to obtain at least one jet forming index of each structural design scheme, as shown in Table 4, which shows the values ​​of the jet forming index under each structural design scheme when t=50μs during the jet process.

[0089] Table 4

[0090]

[0091] S104, constructing a neural network model according to the multiple structural design schemes and the corresponding at least one jet forming index.

[0092] Among them, a neural network model is constructed according to a plurality of structural design schemes and at least one corresponding jet forming index, including: for any jet forming index, the values ​​of the jet forming index under each structural design scheme are normalized respectively; under each structural design scheme, according to the weight of each jet forming index, the values ​​of the normalized jet forming index are weighted to obtain a comprehensive evaluation index under each structural design scheme; a neural network model is constructed according to a plurality of structural design schemes and the corresponding comprehensive evaluation index.

[0093] The jet forming index can be normalized according to formula (4) to obtain the normalized jet forming index.

[0094]

[0095] Among them, Y * represents the value of the normalized jet forming index, y represents the value of the jet forming index to be normalized, and y min Represents the minimum value of all values ​​in the jet forming index, y max It represents the maximum value of all the values ​​in the jet forming index.

[0096] The jet forming index includes jet length, jet head velocity and jet total energy; the calculation method of the comprehensive evaluation index is:

[0097]

[0098] Among them, W represents the comprehensive evaluation index, W a represents the jet length, represents the weight of the jet length, W b represents the jet head velocity, represents the weight of the jet head velocity, W c represents the total energy of the jet, Represents the weight of the total energy of the jet.

[0099] To obtain a good jet, the jet needs to have a large jet length. The jet length is the most important factor affecting the jet forming effect, followed by the jet head speed, while the total jet energy has a relatively small impact. Based on this, different weights are assigned to these indicators, namely Then W = 1.5W a +Wb +0.5W c .

[0100] The values ​​corresponding to the jet forming indicators in Table 4 above are normalized to obtain a data normalization analysis table, as shown in Table 5.

[0101] Table 5

[0102] <![CDATA[W a ]]> <![CDATA[W b ]]> <![CDATA[W c ]]> W 1 1.000 1.000 0.338 2.669 2 0.688 0.569 0.171 1.687 3 0.324 0.385 1.000 1.371 4 0.014 0.200 0.000 0.221 5 0.815 0.902 0.379 2.314 6 0.628 0.629 0.167 1.655 7 0.243 0.234 0.091 0.644 8 0.123 0.000 0.054 0.212 9 0.637 0.852 0.207 1.911 10 0.342 0.441 0.086 0.997 11 0.209 0.295 0.082 0.650 12 0.062 0.089 0.867 0.616 13 0.411 0.731 0.110 1.403 14 0.307 0.484 0.100 0.995 15 0.154 0.233 0.071 0.499 16 0.000 0.048 0.044 0.070

[0103] By analyzing the normalized data of these three factors (structural design parameters) obtained under different schemes of orthogonal test, the main and secondary influencing factors are determined. 16 It can be observed that at each level of A0, the four levels of B0 and C0 are all present, and each level only appears once. Therefore, it can be concluded that in the test combination with only A1, A2, A3, and A4, the effects of other variables on the comprehensive evaluation index W remain unchanged, and then W is obtained. A1 , W A2 , W A3 , W A4 The change of directly reflects the effect of the change of factor A0. From this, the numerical change of the jet forming performance index based on factor A0 can be calculated. According to this logic, the influence of factors B0 and C0 on W when they change can also be calculated. The specific calculation data is shown in Table 6, which is the calculation results of each factor.

[0104] Table 6

[0105] <![CDATA[Factor A0]]> <![CDATA[Factor B0]]> <![CDATA[Factor C0]]> 1.333 1.261 1.487 2.074 1.279 1.206 0.791 1.222 1.043 0.280 0.816 0.742

[0106] Calculate the range R of factors A0, B0, and C0 in the table respectively A =1.794, R B =0.463, R C =0.745. According to the calculation results, for this numerical simulation, the cone angle and inner arc radius of the liner are the main factors affecting the jet morphology, and the liner wall thickness is a secondary factor. The optimal combination is A2, B2, and C1.

[0107] It should be pointed out that the above optimal combination is based on the established factor level, not the optimal liner structural parameters within the parameter range. Therefore, the present invention adopts BP neural network and genetic algorithm to optimize parameters. In order to enhance the prediction accuracy of the neural network, the numerical simulation results based on the orthogonal design experimental scheme are used to fit the relationship curve and generate data samples to meet the learning of the neural network.

[0108] Specifically, a neural network model is constructed according to a variety of structural design schemes and corresponding comprehensive evaluation indexes, including: for any structural design parameter, the initial value of the structural design parameter under a variety of structural design schemes is fitted with the corresponding comprehensive evaluation index to obtain a fitting function under the structural design parameter; the fitting function represents the corresponding relationship between the structural design parameter and the comprehensive evaluation index; the training samples are determined through the fitting functions under each structural design parameter; and the neural network model is trained through the training samples.

[0109] For any structural design parameter, the comprehensive evaluation index W is fitted with the value point of the structural design parameter to obtain the fitting function of the structural design parameter: Among them, x i Indicates the value of the structural design parameter. Specifically, taking the cone angle as an example, at each level of the cone angle, the four levels of wall thickness and inner arc radius are all displayed, and each level only appears once. Therefore, it can be concluded that in the test combination where only the four levels of cone angle A1, A2, A3, and A4 are involved, the effects of other variables on the injection molding index remain unchanged, and the corresponding comprehensive evaluation index W is obtained. A1 , W A2 , W A3 , W A4 The change of directly reflects the effect of the change of the cone angle. Therefore, for any value of the cone angle, the average value of multiple comprehensive evaluation indexes corresponding to the cone angle value can be used as the target comprehensive evaluation index of the cone angle at this value to participate in the fitting process of the fitting parameters. In this way, the fitting function of the cone angle is obtained by fitting each value of the cone angle and the target comprehensive evaluation index at each value. The process of determining the fitting function of the wall thickness and the inner arc radius is similar to that of the cone angle, and this embodiment will not be repeated here.

[0110] In an exemplary embodiment, based on the above data, the P value and error corresponding to each factor (structural design parameter) can be obtained, as shown in Table 7.

[0111] Table 7

[0112] <![CDATA[A0]]> <![CDATA[B0]]> <![CDATA[C0]]> <![CDATA[P1]]> 582.31 341.25 452.65 <![CDATA[P2]]> -325.854 -632.174 -832.84 <![CDATA[P3]]> 1471.931 981.23 923.65 <![CDATA[P4]]> 252.17 784.59 426.81 SSE <![CDATA[6.548×10 -23 ]]> 2.193E-22 9.217E-24

[0113] In an exemplary embodiment, training samples are determined by fitting functions under various structural design parameters, including: for any structural design parameter, according to the fitting function under the structural design parameter, multiple sample values ​​of the structural design parameter and the corresponding comprehensive evaluation index are obtained; according to the multiple sample values ​​of each structural design parameter, multiple sample structural design schemes are determined, and the average value of the comprehensive evaluation index corresponding to each sample value under each sample structural design scheme is determined as the comprehensive evaluation index of each sample structural design scheme; and the multiple sample structural design schemes and the corresponding comprehensive evaluation indexes are determined as training samples.

[0114] First, multiple sample values ​​of each structural design parameter are obtained, and then each sample value is substituted into the fitting parameter of the corresponding structural design parameter to obtain a comprehensive evaluation index corresponding to each sample value. According to the multiple sample values ​​of each structural design parameter, multiple sample structural design schemes are determined, including: randomly combining the multiple sample values ​​of each structural design parameter to obtain multiple sample structural design schemes, each sample structural design scheme includes a sample value of each structural design parameter.

[0115] The training samples are randomly divided into training sets and test sets for back-propagation (BP) neural network training and testing to obtain a neural network model; specifically, the network input parameters are a variety of structural design parameters, and the network output parameters are comprehensive evaluation indexes. For example, the network input parameters are the liner cone angle α, the liner wall thickness δ, and the liner inner arc radius R. The hidden layer nodes of the neural network can be set to 10, the grid structure is 3-10-1, and the structure diagram is as follows: Figure 4 Shown

[0116] like Figure 5 As shown in Figure 1, the error histogram is designed to intuitively present the distribution of errors in the training and evaluation process of the neural network. By dividing the error data into different intervals (i.e., the "bins" of the histogram) and counting the number of data points in each interval, the prediction results of the test set data obtained by the neural network are compared with the actual fitting results. Most of the data are in the range of -0.02 to 0.02, with a small error. The prediction accuracy of the neural network for the training samples fluctuates little. The obtained linear regression graph is shown in Figure 1. Figure 6 shown. Figure 7 It indicates the changing trend of the mean square error (MSE) of the training set, validation set, test set and the overall sample as the number of training rounds increases during the training process, where the MSE parameter value is 0.00078. Figure 5-Figure 7 It can be seen that the trained neural network model has a high prediction accuracy for the comprehensive evaluation index of jet forming, and can more accurately reflect the relationship between the structural parameters of the liner and the jet forming performance.

[0117] S105, performing optimization according to the neural network model to obtain the optimal structural design scheme of the amorphous alloy charge liner.

[0118] In an exemplary embodiment, an optimization is performed according to a neural network model to obtain an optimal structural design scheme for an amorphous alloy liner, including: using the neural network model as an objective function of a genetic algorithm, and multiple structural design parameters as variables of the genetic algorithm, optimizing the structural design parameters of the amorphous alloy liner through a genetic algorithm, and determining the structural design parameter value corresponding to the maximum objective function value during the genetic algorithm iteration process as the optimal structural design scheme.

[0119] The trained neural network model is saved as a function as the objective function in the genetic algorithm. The predicted value output by the neural network model is the objective function value of the genetic algorithm, that is, the comprehensive index of jet forming is used as the individual fitness in the genetic algorithm. The genetic algorithm is opened and input into this function. The number of structural design parameters is used as the variable of the genetic algorithm. The value range of each structural design parameter is set. The crossover probability and mutation probability are set, as well as the population size and number of iterations of the genetic algorithm are set. The initial value of the structural design parameter is set. The genetic algorithm is run. When the maximum number of iterations is reached, the optimization process is terminated. The value combination of the structural design parameters corresponding to the maximum fitness value is determined as the optimal structural design scheme of the liner. The liner is designed according to the optimal structural design scheme.

[0120] Specifically, the number of variables is set to 3, and the range of the structural parameters of the liner is set to [55, 70], [0.7, 1.6], and [6, 9] respectively. The crossover probability is set to 0.5, the mutation probability is set to 0.01, the genetic algorithm population size is set to 25, and the genetic cycle algebra is set to 100. When the predetermined maximum genetic algebra is reached, the optimization process is terminated, and the parameter combination corresponding to the optimal fitness value is the most ideal liner structural parameter obtained. As the number of iterations increases, the optimal fitness value gradually increases. When the iteration exceeds 80 times, the fitness value tends to be stable and almost no longer changes significantly. The optimal fitness value obtained by the genetic algorithm is 2.186, and the optimal liner structural parameter combination is 61.29° for the liner cone angle, 1.57mm for the wall thickness, and 7.15mm for the inner arc radius. For the convenience of developing the actual object, the optimized parameters of the liner cone angle, wall thickness, and inner arc radius are approximately 61°, 1.6mm, and 7.2mm.

[0121] When applying the structural design method of the amorphous alloy liner provided by the present invention, it is not necessary to Figure 1 The steps are executed in the order shown. The specific execution order of the steps can be determined according to needs, and the present invention does not limit this.

[0122] The above is a structural design method of an amorphous alloy liner provided by one or more embodiments of the present invention. Based on the same idea, the present invention also provides a corresponding structural design device of an amorphous alloy liner, such as Figure 8shown.

[0123] Figure 8 A schematic diagram of a structural design device for an amorphous alloy liner provided by the present invention, the device 800 comprises:

[0124] An acquisition module 801 is used to acquire a plurality of initial values ​​of each structural design parameter of the amorphous alloy liner;

[0125] The design module 802 is used to perform orthogonal experimental design according to multiple initial values ​​of each structural design parameter to obtain multiple structural design schemes of the amorphous alloy liner;

[0126] The simulation module 803 is used to perform numerical simulation on the amorphous alloy liner under each structural design scheme to obtain at least one jet forming index of each structural design scheme;

[0127] A construction module 804 is used to construct a neural network model according to a plurality of structural design schemes and at least one corresponding jet forming index;

[0128] The optimization module 805 is used to perform optimization according to the neural network model to obtain the optimal structural design solution of the amorphous alloy liner.

[0129] The specific definition of the structural design device of the amorphous alloy liner can be found in the definition of the structural design method of the amorphous alloy liner mentioned above, which will not be repeated here. Each module in the above-mentioned structural design device of the amorphous alloy liner can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0130] The present invention also provides a computer-readable storage medium, which stores a computer program, which can be used to execute the above Figure 1 The present invention provides a structural design method for an amorphous alloy liner.

[0131] The present invention also provides Fig. 9 The structural diagram of the computer device shown in FIG. Fig. 9 As shown in the figure, at the hardware level, the computer device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory, and may also include other hardware required for the business. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to achieve the above Figure 1 The present invention provides a structural design method for an amorphous alloy liner.

[0132] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided by the present invention can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0133] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of the present invention.

Claims

1. A structural design method for an amorphous alloy liner, characterized in that: include: Obtaining multiple initial values ​​of each structural design parameter of the amorphous alloy liner; According to multiple initial values ​​of each structural design parameter, orthogonal experimental design is carried out to obtain multiple structural design schemes of amorphous alloy liner; Numerical simulations are performed on the amorphous alloy liner under each structural design scheme to obtain at least one jet forming index of each structural design scheme; Constructing a neural network model according to the plurality of structural design schemes and the corresponding at least one jet forming index; The optimization is performed according to the neural network model to obtain the optimal structural design scheme of the amorphous alloy liner.

2. The method according to claim 1, characterized in that The step of constructing a neural network model according to the plurality of structural design schemes and the corresponding at least one jet forming index comprises: For any jet forming index, the values ​​of the jet forming index under each structural design scheme are normalized respectively; Under each structural design scheme, the normalized values ​​of each jet forming index are weighted according to the weight of each jet forming index to obtain the comprehensive evaluation index under each structural design scheme; A neural network model is constructed based on the multiple structural design schemes and the corresponding comprehensive evaluation indexes.

3. The method according to claim 2, characterized in that The jet forming index includes jet length, jet head velocity and jet total energy; the calculation method of the comprehensive evaluation index is: Among them, W represents the comprehensive evaluation index, W a represents the jet length, represents the weight of the jet length, W b represents the jet head velocity, represents the weight of the jet head velocity, W c represents the total energy of the jet, Represents the weight of the total energy of the jet.

4. The method according to claim 2, characterized in that: The step of constructing a neural network model according to the plurality of structural design schemes and the corresponding comprehensive evaluation indexes includes: For any structural design parameter, the initial value of the structural design parameter under the multiple structural design schemes is fitted with the corresponding comprehensive evaluation index to obtain a fitting function under the structural design parameter; the fitting function represents the corresponding relationship between the structural design parameter and the comprehensive evaluation index; Determining training samples through fitting functions under each of the structural design parameters; The neural network model is trained through training samples.

5. The method according to claim 4, characterized in that The step of determining the training samples by fitting functions under the structural design parameters comprises: For any structural design parameter, according to the fitting function under the structural design parameter, multiple sample values ​​of the structural design parameter and corresponding comprehensive evaluation index are obtained; According to the multiple sample values ​​of each structural design parameter, multiple sample structural design schemes are determined, and the average value of the comprehensive evaluation index corresponding to each sample value under each sample structural design scheme is determined as the comprehensive evaluation index of each sample structural design scheme; The multiple sample structure design schemes and the corresponding comprehensive evaluation indexes are determined as the training samples.

6. The method according to claim 1, characterized in that The method of performing optimization according to the neural network model to obtain the optimal structural design scheme of the amorphous alloy liner includes: The neural network model is used as the objective function of the genetic algorithm, and multiple structural design parameters are used as variables of the genetic algorithm. The structural design parameters of the amorphous alloy charge liner are optimized through the genetic algorithm, and the structural design parameter value corresponding to the maximum objective function value in the iterative process of the genetic algorithm is determined as the optimal structural design solution.

7. The method according to claim 1, characterized in that The method of numerically simulating the amorphous alloy liner under each structural design scheme to obtain at least one jet forming index of each structural design scheme includes: For any structural design scheme, in the shaped charge finite element model, boundary conditions are added in the air domain and the Euler grid is divided, and then the amorphous alloy liner is numerically simulated under the structural design scheme through a numerical simulation algorithm.

8. The method according to claim 1, characterized in that The material of the amorphous alloy liner is W skeleton / Zr-based amorphous alloy, and the amorphous alloy liner is a conical liner.

9. A structural design device for an amorphous alloy liner, characterized in that: include: An acquisition module, used to acquire multiple initial values ​​of each structural design parameter of the amorphous alloy liner; A design module is used to perform orthogonal experimental design based on multiple initial values ​​of each structural design parameter to obtain multiple structural design schemes of amorphous alloy liner; A simulation module is used to numerically simulate the amorphous alloy liner under each structural design scheme to obtain at least one jet forming index of each structural design scheme; A construction module, used to construct a neural network model according to the multiple structural design schemes and the corresponding at least one jet forming index; The optimization module is used to perform optimization according to the neural network model to obtain the optimal structural design scheme of the amorphous alloy charge liner.

10. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 8 is implemented.

Images