Method for detecting integrity of a warhead shell after penetration based on simulation

Abstract

The application relates to a method for detecting the integrity of a post-penetration warhead shell based on simulation simulation, and belongs to the technical field of image recognition for simulation result data preprocessing and mining in a machine learning process. A topographic map under a side view and a bottom view of a simulated warhead after penetration is extracted; the extracted topographic map is subjected to grayscale and binarization processing in sequence, and then a damage area is identified and a damage area size is counted; based on the counted damage area size, the damage of the warhead shell structure is counted and identified. The application provides a method for automatically detecting the integrity of a warhead shell structure based on a finite element simulation method, can greatly improve the time and energy cost of manual viewing and distinguishing, is suitable for large-sample penetration result prediction simulation work, and can explore the main control factors of the incomplete warhead structure, and in a certain sense, guide material research and preparation.

Description

Technical Field

[0001] This invention relates to a method for detecting the integrity of a warhead shell after penetration based on simulation, belonging to the technical field of image recognition for simulation result data preprocessing and mining in machine learning. Background Technology

[0002] Semi-armor-piercing projectiles are characterized by their large chamber size and high explosive charge, combining armor-piercing capability with explosive power. They are widely used in weapon systems such as missiles, bombs, rockets, and artillery shells. The warhead, as the terminal damage system that directly destroys the target, mainly consists of a casing, explosive charge, and fuse. After penetrating the target using its own kinetic energy, the projectile is detonated by the fuse, generating a strong shock wave, overpressure, and high-speed fragments that damage the target. The semi-armor-piercing projectile achieves maximum destructive power only after penetrating the target's armor; therefore, the design of the semi-armor-piercing warhead must ensure that the casing does not break during penetration. In military research, evaluating the penetration effectiveness of semi-armor-piercing projectiles is a key topic. Due to limitations in target testing cycles, costs, and safety, quantitatively evaluating the service performance of the warhead casing using the finite element method has become the primary research approach.

[0003] In recent years, machine learning and data science have become research hotspots. By training on numerous samples, they can identify high-dimensional functional relationships between inputs and outputs within a data sample that are not easily found. For example, machine learning algorithms can establish a quantitative relationship between material properties and warhead service performance. Furthermore, by using quantitative indicators of the high-speed penetration performance of the warhead shell, the main controlling factors of the material properties affecting its penetration performance can be determined, guiding subsequent material design. The penetration performance indicators of a semi-armor-piercing warhead shell can only be discussed if the structure remains intact. Therefore, it is necessary to first determine the integrity of the simulated projectile structure, i.e., that the shell structure does not break, preventing leakage of the internal explosive charge. However, current finite element simulation results cannot directly reflect whether the shell structure is damaged. The integrity of the shell structure is mainly judged by visual observation, but this requires extensive manual adjustment of the model's perspective, which is cumbersome and cannot achieve automatic filtering of massive simulation data. Summary of the Invention

[0004] In response to the current situation where the integrity of the warhead shell after penetration is mainly determined by visual observation, this invention provides a method for detecting the integrity of the warhead shell after penetration based on simulation. This method can effectively identify and process the penetration simulation results data, extract a dataset of warhead structural integrity that can be used for subsequent machine learning, and automatically identify the shell damage mode, thus filling the gap in the automatic identification of the integrity of the projectile shell structure based on finite element simulation.

[0005] The objective of this invention is achieved through the following technical solutions.

[0006] A method for detecting the integrity of a warhead's outer shell after penetration based on simulation, the method comprising the following steps:

[0007] (i) Extract the morphology images of the warhead simulation projectile after penetration from the side and bottom views;

[0008] (ii) The extracted morphology image is processed by grayscale and binarization in sequence, and then the damaged area is identified and the area of ​​the damaged area is statistically analyzed.

[0009] (iii) Based on the statistical size of the damaged area, the damage status of the warhead shell structure is statistically identified.

[0010] Further, in step (i), LS-DYNA software is used to conduct multiple sets of simulated projectile penetration target plate simulation studies on the warhead shell material, obtain penetration simulation result files under different material parameters, and import the file into lsprepost post-processing software, set the number of states, set different viewing angles, and obtain the morphology images of the side and bottom views after penetration.

[0011] Furthermore, obtaining penetration simulation result files under different material parameters usually refers to obtaining penetration simulation result files under different dynamic strength and dynamic plastic failure strain conditions.

[0012] Furthermore, in the lsprepost post-processing software, setting the display mode to no grid and no shadow can display the broken parts of the projectile shell after penetration as clearly as possible, reducing the impact of light and shadow color changes and grid lines on the subsequent automatic identification of broken parts.

[0013] Further, in step (ii), the OpenCV module in Python is used to read the topography image extracted in step (i). Each pixel in the image is traversed and stored as a list [height, width, channel], where height represents the image height, width represents the image width, and channel represents the image's channels, which are in BGR format. Then, grayscale conversion is performed, typically using a weighted average method to convert the original image's three BGR channels into a single channel, resulting in the converted grayscale image. The grayscale image obtained after grayscale processing is then subjected to fixed-threshold image segmentation, i.e., a fixed threshold is set for the entire image. Each pixel's value is compared to a threshold. If it's less than the threshold, the pixel value is changed to a fixed value (usually 0). If it's greater than or equal to the threshold, the pixel value is changed to another fixed value (usually 255). Then, the cv2.threshold function is used for binarization to obtain a binary image. The findcontours function, based on OpenCV's contour algorithm, identifies the edge contour pixels of the binary image after binarization. The contouraArea function, a built-in algorithm for calculating closed regions in OpenCV, is used to calculate the area within the contour, thereby displaying broken parts and calculating the area of ​​damaged regions within the contour.

[0014] Furthermore, in the actual simulation process, the damage to the warhead casing can be divided into the following four situations: (1) the casing structure remains intact; (2) only the head of the casing is damaged; (3) only the shoulder of the casing is damaged; (4) both the shoulder and the head of the casing are damaged.

[0015] In step (iii), if the area of ​​the damaged region counted in step (ii) is greater than zero, it indicates that the shell structure is damaged; otherwise, it indicates that the shell structure remains intact. Specifically, the following judgment situations apply: if the area of ​​the damaged region is greater than zero in both the side view and the bottom view of the simulated projectile, it indicates that the head and shoulder of the shell are damaged; otherwise, the projectile structure remains intact. If the area of ​​the damaged region in the side view of the simulated projectile is greater than zero while the area of ​​the damaged region in the bottom view is zero, it indicates that only the shoulder of the shell is damaged; otherwise, it indicates that only the head of the shell is damaged.

[0016] Beneficial effects:

[0017] (1) This invention provides a method for automatically detecting the structural integrity of a warhead shell based on the finite element simulation method, which can greatly improve the time and effort cost of manual inspection and judgment, and is suitable for large-sample penetration result prediction simulation work.

[0018] (2) In actual machine learning processes, it is usually necessary to clean the acquired data to improve the accuracy and efficiency of subsequent data mining. This invention uses Python to batch identify the bottom and side views of the simulated projectile after penetrating the target plate, which can automatically determine the shell damage and screen the simulation examples without damage. The examples without shell damage can be used as a high-quality dataset for subsequent machine learning to further explore the quantitative relationship between different material properties and warhead service performance (i.e., destructive power). This method also achieves efficient cleaning of a large amount of simulation result data, solves the problem of data cleaning in the process of finite element simulation of penetration problems based on machine learning, improves the accuracy of data mining, and greatly saves R&D costs.

[0019] (3) The method described in this invention can identify different damage modes after the simulated projectile penetrates the target plate. Combined with different material parameters, it can explore the main controlling factors of the incomplete projectile structure, predict the materials with better strength-plasticity matching for shell production, and guide material research and development and preparation to a certain extent.

[0020] (4) When the present invention uses lsprepost post-processing software to extract images, the display mode is set to a state of no grid and no shadow, which can display the broken parts of the projectile shell after penetration as clearly as possible, and reduce the influence of light and shadow color changes and grid lines on the subsequent automatic identification of broken parts. Attached Figure Description

[0021] Figure 1 The images show the deformation and failure morphology of the simulated projectile structure obtained under different dynamic strengths and dynamic plasticities in Example 1. In Figure a, the shell structure is intact; Figure b, the head of the shell is damaged; Figure c, the shoulder of the shell is damaged; and Figure d, the shoulder and head of the shell are damaged simultaneously.

[0022] Figure 2 This is a flowchart of the process for processing the topographic image extracted after the simulated projectile penetrates, as shown in Example 1.

[0023] Figure 3 This is a logic flowchart of Example 1 for detecting the integrity of the warhead shell after penetration based on simulation.

[0024] Figure 4 This is a pie chart showing the statistical results of the warhead shell damage after penetration, based on simulation detection in Example 1. Detailed Implementation

[0025] The present invention will be further described below with reference to specific embodiments.

[0026] Example 1

[0027] Step 1: For the titanium alloy material used in the warhead shell, the Johnson-Cook constitutive model is often used. The dynamic strength of the material is set to vary in the range of 1200-1700 MPa and the dynamic plastic failure strain is set to vary in the range of 0.10-0.35. LS-DYNA software is used to conduct 100 sets of simulated projectile penetration target plate simulation studies on the warhead shell material. The result file d3plot under different material parameters is obtained and imported into lsprepost post-processing software. The number of states is set, different viewpoints are set, and the display mode is set to no mesh and no shadow. The morphology images of the side and bottom views after penetration are obtained.

[0028] Step two: The OpenCV module in Python is used to read the side and bottom topographic images obtained in step (i). Each pixel is iterated through and stored as a list [height, width, channel], where height represents the image height, width represents the image width, and channel represents the image channels in BGR format. Then, a weighted average method (i.e., using the cv2.COLOR_RGB2GRAY function) is used to convert the original BGR three-channel image into a single channel to complete the grayscale conversion, resulting in a grayscale image. The obtained grayscale image is then subjected to fixed-threshold image segmentation. In this implementation, a threshold is set... The value segmentation parameter is set to 200. Pixels with a value less than 200 are all set to a fixed value of 0, while pixels with a value greater than or equal to 200 are all set to another fixed value of 255. Binarization is performed using the `cv2.threshold` function to obtain a binary image. The `findcontours` function of OpenCV is used to identify the edge contour pixels in the obtained binary image. Then, the `contouraArea` function, a built-in OpenCV algorithm for calculating closed regions, is used to calculate the area within the contour, thereby displaying broken parts and calculating the area of ​​damaged regions within the contour. The flowchart for processing the extracted morphology image is detailed below. Figure 2 ;

[0029] Step 3: In the actual simulation process, the damage to the warhead casing can be divided into the following four situations: (1) The casing structure remains intact, such as Figure 1 As shown in a; (2) Only the head of the shell is damaged, such as Figure 1 As shown in b; (3) Only the shoulder of the shell is damaged, such as Figure 1 As shown in c; (4) Both the shoulder and head of the shell are damaged, as shown in c. Figure 1 As shown in d;

[0030] If the damaged area counted in step two is greater than zero, it indicates that the shell structure is damaged; otherwise, it indicates that the shell structure remains intact. Specifically, the following judgment scenarios apply: If the damaged area is greater than zero in both the side and bottom views of the simulated projectile, it indicates that both the head and shoulder of the shell are damaged; otherwise, the projectile structure remains intact. If the damaged area is greater than zero in the side view but zero in the bottom view, it indicates that only the shoulder of the shell is damaged; otherwise, it indicates that only the head of the shell is damaged. The specific judgment flowchart is as follows: Figure 3 As shown;

[0031] Step four: The results of automatic identification of the projectile's structural integrity using the selected 100 sets of simulation data are as follows: Figure 4 As shown in the figure, 50% of the projectiles suffered simultaneous damage to both the shoulder and head, 33% had intact projectile structures, 16% had only shoulder damage, and 1% had only head damage. This result indicates that the shoulder is a weak point in the projectile structure, and the main damage to the projectile structure occurs in the shoulder.

[0032] In addition, the complete projectile structure can serve as a high-quality dataset for subsequent machine learning, further exploring the quantitative relationship between different material properties and warhead service performance (i.e., destructive power). The datasets of different projectile shell failure modes can be used to explore the main material parameters that lead to different failure modes of the projectile shell, achieving efficient cleaning and identification of a large amount of finite element simulation data.

[0033] In summary, the above are merely preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for detecting the integrity of a warhead shell after penetration based on simulation, characterized in that: The method comprises the following steps: (i) extracting the topography graph of the warhead simulation bomb under the side view and bottom view after penetration; (ii) sequentially performing gray scale and binary processing on the extracted topography graph, and then identifying the damaged area and counting the area of the damaged area; The opencv module in the Python program is used to read the topography graph extracted in step (i). By traversing each pixel point of the image, it is stored as a list [height, width, channel], where height represents the height of the picture, width represents the width of the picture, and channel represents the channel of the picture. The original image BGR three channels are converted into a single channel by using the weighted average method, and the corresponding gray scale conversion gray scale graph is obtained. The gray scale graph is subjected to fixed threshold image segmentation, that is, a fixed threshold is set, and the pixel value of each pixel of the entire picture is compared with the threshold. If it is less than the threshold, the pixel value is changed to a fixed value. If it is greater than or equal to the threshold, the pixel value is changed to another fixed value. Then the function cv2.threshold function is used for binaryzation, and the corresponding binary graph is obtained. Based on the opencv contour algorithm findcontours function, the edge contour pixel points of the binary graph are identified, and the area inside the contour is counted by using the built-in calculation closed area algorithm contouraArea function. Accordingly, the display of the damaged part and the counting of the area of the damaged area inside the contour are realized. (iii) Based on the size of the counted damaged area, the damage of the warhead shell structure is counted and identified. In step (i), LS-DYNA software is used to simulate the penetration of the target plate by a plurality of simulation bombs, and the penetration simulation result files under different material parameters are obtained. The penetration simulation result files are imported into the lsprepost post-processing software, the state number is set, and different views are set to obtain the topography graph under the side view and bottom view after penetration. In the lsprepost post-processing software, the display mode is set to meshless and shadowless state.

2. The method for detecting the integrity of the warhead shell after penetration based on the simulation analog according to claim 1, characterized in that: Obtaining the penetration simulation result files under different material parameters means obtaining the penetration simulation result files under different dynamic strength and dynamic plastic failure strain conditions.

3. The method for detecting the integrity of the warhead shell after penetration based on the simulation analog according to claim 2, characterized in that: In step (iii), if the damaged area in the side view and the top view of the simulation bomb is greater than zero, it means that the head and shoulder of the shell body are damaged, otherwise the shell body structure remains intact; 4. The method for detecting the integrity of the warhead shell after penetration based on the simulation analog according to claim 2, characterized in that: If the damaged area in the side view of the simulation bomb is greater than zero and the damaged area in the top view is zero, it means that only the shoulder of the shell body is damaged, otherwise it means that only the head of the shell body is damaged.

5. The method for detecting the integrity of the warhead shell after penetration based on the simulation analog according to claim 1, characterized in that: ​ ​

Images