Penetration performance index prediction method and device
Through a hierarchical machine learning model, using Pearson's correlation coefficient to screen parameters and combining random forests and multi-layer perceptrons for prediction, the problem of scarcity and low efficiency of infiltration performance research in the existing technology is solved, and material infiltration is achieved Fast, accurate evaluation and efficient prediction of performance.
Patent Information
- Application Number
- CN202410940574.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-15
- Publication Date
- 2024-11-08
AI Technical Summary
In the prior art, the research on infiltration performance is relatively rare, and it depends on costly and time-consuming experimental tests, making it difficult to achieve rapid and accurate evaluation of material infiltration performance, especially in scenarios where highly nonlinear and complex dependencies are handled. ,Existing machine learning research rarely involves simultaneous prediction problems with multiple indicators.
Using a hierarchical machine learning model, important parameters are selected by calculating the Pearson correlation coefficient between input parameters, and classification and regression prediction is performed using random forest algorithms and multi-layer perceptrons to ensure the prediction optimization of each indicator and combined with precision Regulate computing resources to improve processing efficiency.
It realizes rapid and accurate evaluation of material intrusion performance, improves prediction accuracy and processing efficiency, adapts to in-depth data analysis in complex scenarios, reduces the computational burden and reduces the risk of inaccurate prediction results.
Smart Images

Figure CN118916792A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of material engineering technology, and in particular to a method and device for predicting penetration performance indicators. Background Art
[0002] The study of the penetration performance of materials plays a vital role. Penetration performance, including penetration velocity, whether the target is penetrated, and the integrity of the projectile, is a key parameter for evaluating protective equipment such as armored vehicles, protective clothing, and bulletproof materials. Its performance indicators can directly affect the survivability and protective effect of the equipment in an actual combat environment. In the prior art, research on penetration performance is relatively scarce, and most of the relevant data and research results are not known to the public. In addition, the complexity of penetration events and their high requirements for accurate physical models further limit the possibility of extensive research. Traditional research methods mainly rely on costly and time-consuming experimental tests, which often need to be carried out under strictly controlled conditions to ensure the accuracy and repeatability of the data.
[0003] In recent years, machine learning and deep learning technologies have shown great potential in solving complex problems in many fields, but their application in the field of penetration performance prediction is still in its early stages. In addition, existing machine learning research rarely involves the problem of simultaneously predicting multiple indicators of penetration performance, especially in scenarios that require handling highly nonlinear and complex dependencies. Summary of the invention
[0004] The purpose of this application is to provide a method and device for predicting penetration performance indicators in order to overcome the existing technical defects, which can ensure that the prediction of each indicator is optimized to meet its specific analysis needs. In addition, by precisely controlling computing resources, the processing efficiency can be greatly improved, ultimately achieving rapid and accurate evaluation of material penetration performance.
[0005] The purpose of this application is achieved through the following technical solutions:
[0006] In a first aspect, the present application proposes a method for predicting penetration performance indicators, which is applied to a hierarchical machine learning model, and the method includes:
[0007] Calculate the Pearson correlation coefficient between each pair of input parameters;
[0008] Filter the input parameters according to the Pearson correlation coefficient to obtain filtered input parameters;
[0009] The random forest algorithm and multi-layer perceptron are used to perform classification prediction and regression prediction on the screened input parameters to obtain the penetration performance.
[0010] In a possible implementation, the input parameters include:
[0011] Firing speed, Vickers hardness;
[0012] Yield strength, tensile strength, and total elongation at break in quasi-static tensile tests;
[0013] Specified non-proportional compressive strength and compressive strength for quasi-static compression tests;
[0014] Dynamic yield strength, tensile strength and dynamic total elongation at break from dynamic tensile tests;
[0015] Dynamic yield strength and compressive strength at dynamic compression of 3000s-1;
[0016] Dynamic yield strength and compressive strength at dynamic compression of 5000s-1.
[0017] In a possible implementation, the penetration performance includes penetration speed, penetration determination result and projectile integrity, the random forest algorithm includes a first random forest algorithm and a second random forest algorithm, and the steps of using the random forest algorithm and the multi-layer perceptron to respectively perform classification prediction and regression prediction on the screened input parameters to obtain the penetration performance include:
[0018] The screened input parameters are input into the first random forest algorithm for classification prediction to obtain the penetration determination result;
[0019] When the penetration is judged as penetration according to the penetration judgment result, the screened input parameters are input into the multi-layer perceptron for regression prediction to obtain a non-zero penetration velocity, and the screened input parameters are input into the second random forest algorithm for classification prediction to obtain the projectile integrity. The projectile integrity corresponds to: projectile fragmentation or projectile integrity;
[0020] When it is determined that the projectile has not penetrated based on the penetration determination result, the penetration velocity is output as 0 and the projectile integrity is output as projectile fragmentation.
[0021] In a possible implementation, the step of calculating the Pearson correlation coefficient between each pair of input parameters includes:
[0022] Calculate the covariance and standard deviation between each pair of input parameters;
[0023] Dividing the covariance by the standard deviation gives the Pearson correlation coefficient.
[0024] In a possible implementation, the step of filtering the input parameters according to the Pearson correlation coefficient to obtain filtered input parameters includes:
[0025] The absolute values of the Pearson correlation coefficients are sorted in order, and multiple input parameters are selected as screened input parameters.
[0026] In a second aspect, the present application proposes a penetration performance index prediction device, the device comprising:
[0027] A calculation module for calculating the Pearson correlation coefficient between each pair of input parameters;
[0028] A screening module, used for screening input parameters according to the Pearson correlation coefficient to obtain screened input parameters;
[0029] The generation module is used to use the random forest algorithm and the multi-layer perceptron to perform classification prediction and regression prediction on the screened input parameters to obtain the penetration performance.
[0030] In a possible implementation, the input parameters include:
[0031] firing speed and Vickers hardness;
[0032] Yield strength, tensile strength, and total elongation at break in quasi-static tensile tests;
[0033] Specified non-proportional compressive strength and compressive strength for quasi-static compression tests;
[0034] Dynamic yield strength, tensile strength and dynamic total elongation at break from dynamic tensile tests;
[0035] Dynamic yield strength and compressive strength at dynamic compression of 3000s-1;
[0036] Dynamic yield strength and compressive strength at dynamic compression of 5000s-1.
[0037] In a possible implementation, the generating module includes:
[0038] A classification prediction submodule is used to input the screened input parameters into the first random forest algorithm for classification prediction to obtain a penetration determination result;
[0039] A generation submodule is used for inputting the screened input parameters into a multi-layer perceptron for regression prediction to obtain a non-zero penetration velocity when the penetration judgment result is judged as penetration, and inputting the screened input parameters into a second random forest algorithm for classification prediction to obtain the projectile integrity, and the projectile integrity corresponds to the result of: projectile fragmentation or projectile integrity;
[0040] When it is determined that the projectile has not penetrated based on the penetration determination result, the penetration velocity is output as 0 and the projectile integrity is output as projectile fragmentation.
[0041] In a possible implementation, the computing module is specifically configured to:
[0042] Calculate the covariance and standard deviation between each pair of input parameters;
[0043] Dividing the covariance by the standard deviation gives the Pearson correlation coefficient.
[0044] In a possible implementation, the screening module is specifically used to:
[0045] The absolute values of the Pearson correlation coefficients are sorted in order, and multiple input parameters are selected as screened input parameters.
[0046] The above-mentioned main scheme of the present application and its further options can be freely combined to form multiple schemes, all of which are schemes that can be adopted and claimed for protection in the present application; and in the present application, (non-conflicting options) options and other options can also be freely combined. After understanding the scheme of the present application, those skilled in the art can understand that there are multiple combinations based on the prior art and common knowledge, all of which are technical schemes to be protected by the present application, and they are not exhaustively listed here.
[0047] The present application discloses a method and device for predicting penetration performance indicators. First, the Pearson correlation coefficient between each pair of input parameters is calculated. Secondly, the input parameters are screened according to the Pearson correlation coefficient to obtain screened input parameters. Finally, a random forest algorithm and a multi-layer perceptron are used to perform classification prediction and regression prediction on the screened input parameters to obtain penetration performance, thereby completing the prediction of the penetration performance indicators. This ensures that the prediction of each indicator is optimized to meet its specific analysis requirements. In addition, the processing efficiency can be greatly improved by precisely controlling the computing resources, thereby ultimately achieving rapid and accurate evaluation of the material penetration performance. BRIEF DESCRIPTION OF THE DRAWINGS
[0048] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.
[0049] Figure 1 A flow chart of a method for predicting a penetration performance index proposed in an embodiment of the present application is shown.
[0050] Figure 2 A schematic diagram of the screened input parameters proposed in the embodiment of the present application is shown.
[0051] Figure 3 A schematic diagram of a hierarchical machine learning model proposed in an embodiment of the present application is shown.
[0052] Figure 4 A prediction schematic diagram of the random forest algorithm and the multi-layer perceptron proposed in the embodiment of the present application is shown. DETAILED DESCRIPTION
[0053] The following describes the embodiments of the present application through specific examples, and those skilled in the art can easily understand other advantages and effects of the present application from the contents disclosed in this specification. The present application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present application. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.
[0054] Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making any creative work shall fall within the scope of protection of this application.
[0055] In the prior art, research on penetration performance is relatively rare, and most of the relevant data and research results are not known to the public. In addition, the complexity of penetration events and their high requirements for accurate physical models further limit the possibility of extensive research. Traditional research methods mainly rely on costly and time-consuming experimental tests, which often need to be carried out under strictly controlled conditions to ensure the accuracy and repeatability of the data.
[0056] In recent years, machine learning and deep learning technologies have shown great potential in solving complex problems in many fields, but their application in the field of penetration performance prediction is still in its early stages. In addition, existing machine learning research rarely involves the problem of simultaneously predicting multiple indicators of penetration performance, especially in scenarios that require handling highly nonlinear and complex dependencies.
[0057] Therefore, in order to solve the above-mentioned problems, the embodiments of the present application propose a method and device for predicting penetration performance indicators. By predicting the penetration performance indicators, it can ensure that the prediction of each indicator is optimized to meet its specific analysis needs. In addition, by precisely controlling the computing resources, the processing efficiency can be greatly improved, ultimately achieving rapid and accurate evaluation of the material penetration performance.
[0058] Please refer to Figure 1 , Figure 1 A flow chart of a penetration performance index prediction method proposed in an embodiment of the present application is shown. The method is applied to a hierarchical machine learning model. The model uses a random forest algorithm (RF) and a multi-layer perceptron (MLP). At the same time, a hierarchical prediction method is adopted to gradually predict the penetration performance index, ensuring that the prediction of each index is optimized to meet its specific analysis needs. The method includes:
[0059] Step S100: Calculate the Pearson correlation coefficient between each pair of input parameters.
[0060] The input parameters include:
[0061] firing speed and Vickers hardness;
[0062] Yield strength, tensile strength, and total elongation at break in quasi-static tensile tests;
[0063] Specified non-proportional compressive strength and compressive strength for quasi-static compression tests;
[0064] Dynamic yield strength, tensile strength and dynamic total elongation at break from dynamic tensile tests;
[0065] Dynamic yield strength and compressive strength at dynamic compression of 3000s-1;
[0066] Dynamic yield strength and compressive strength at dynamic compression of 5000s-1.
[0067] The penetration performance prediction of materials is based on a variety of mechanical performance parameters obtained through tests and simulations and the bullet firing speed. The input parameters are optimized and screened. The more input parameters there are, the more complex the model and the more computational burden it will cause.
[0068] In order to improve the efficiency and effectiveness of the model, the embodiment of the present application evaluates the linear correlation between the parameters by calculating the Pearson correlation coefficient between each pair of input parameters. Pearson correlation analysis is a method for evaluating the strength and direction of the linear relationship between two quantitative variables. The calculation result is the Pearson correlation coefficient, which ranges from -1 to +1. The coefficient value +1 indicates a perfect positive correlation, -1 indicates a perfect negative correlation, and 0 indicates no linear correlation.
[0069] It is worth noting that in the embodiment of the present application, the Pearson correlation analysis can be implemented by the Pandas library and the NumPy library in Python to obtain the Pearson correlation coefficient.
[0070] Step S100 includes:
[0071] Calculate the covariance and standard deviation between each pair of input parameters;
[0072] Dividing the covariance by the standard deviation gives the Pearson correlation coefficient.
[0073] In order to identify the most influential input parameters for the model to predict mechanical properties, the covariance and standard deviation between each pair of input parameters were calculated, and then the covariance and standard deviation were divided to obtain the Pearson correlation coefficient.
[0074] Step S200: filter the input parameters according to the Pearson correlation coefficient to obtain filtered input parameters.
[0075] The absolute values of the Pearson correlation coefficients are sorted in order, and multiple input parameters are selected as screened input parameters.
[0076] In one possible embodiment, Figure 2 A schematic diagram of the screened input parameters proposed in the embodiment of the present application is shown. After the Pearson correlation coefficient is obtained based on the 14 input parameters, the top 9 parameters with the highest correlation coefficient values are selected as the input of the final model based on the absolute value of the correlation coefficient.
[0077] Step S300: using the random forest algorithm and the multi-layer perceptron to perform classification prediction and regression prediction on the screened input parameters to obtain the penetration performance.
[0078] The penetration performance includes penetration speed, penetration judgment result and projectile integrity. The penetration judgment result indicates whether the target is penetrated, and its output is in binary form 0 or 1, 0 represents no penetration, 1 represents penetration. The projectile integrity indicates whether the projectile is intact, and its output is also in binary form 0 or 1, 0 represents projectile fragmentation, 1 represents projectile integrity. The predicted output of the penetration speed is a specific value, which can be predicted by a regression model.
[0079] The random forest algorithm can effectively handle the interaction between variables when processing classification problems, and has good anti-overfitting ability. The random forest algorithm proposed in this application includes a first random forest algorithm and a second random forest algorithm. Step S300 includes:
[0080] The screened input parameters are input into the first random forest algorithm for classification prediction to obtain the penetration determination result;
[0081] When the penetration is judged as penetration according to the penetration judgment result, the screened input parameters are input into the multi-layer perceptron for regression prediction to obtain a non-zero penetration velocity, and the screened input parameters are input into the second random forest algorithm for classification prediction to obtain the projectile integrity. The projectile integrity corresponds to: projectile fragmentation or projectile integrity;
[0082] When it is determined that the projectile has not penetrated based on the penetration determination result, the penetration velocity is output as 0 and the projectile integrity is output as projectile fragmentation.
[0083] Since there are also dependencies between penetration performance, there is a direct correlation between the penetration judgment result and the penetration speed. If the target is not penetrated, the penetration speed is zero. Similarly, the penetration judgment result is also closely related to whether the projectile remains intact; usually, if the target is not penetrated, the projectile will often appear in a broken state. These dependencies reflect the interaction and influence between different indicators. Therefore, based on Figure 1 Please refer to Figure 3 , Figure 3 A schematic diagram of a hierarchical machine learning model proposed in an embodiment of the present application is shown. The model adopts a step-by-step hierarchical prediction method, which is mainly divided into two layers. The first layer is a first random forest algorithm, and the second layer is a combination of a multilayer perceptron and a second random forest algorithm.
[0084] based on Figure 3 Please refer to Figure 4 , Figure 4 The prediction schematic diagram of the random forest algorithm and the multi-layer perceptron proposed in the embodiment of the present application is shown. The first layer of the model designs a random forest classifier to execute the first random forest algorithm for classification prediction to obtain a penetration judgment result to determine whether the target is penetrated. The classifier processes the input parameters and outputs a binary result, where 0 indicates that the target is not penetrated and 1 indicates that the target is penetrated.
[0085] The second layer of the model designs a random forest classifier and a multi-layer perceptron. When the target is penetrated (output is 1), the multi-layer perceptron is activated to perform regression prediction of the penetration speed. The multi-layer structure of the multi-layer perceptron can capture the complex nonlinear relationship in the input data, thereby accurately predicting the penetration speed. The second random forest algorithm is used to perform classification prediction of the projectile integrity.
[0086] In addition, when the penetration judgment result is judged as no penetration (output is 0), the penetration speed is automatically set to 0 because the speed does not exist due to no penetration. The projectile state is determined to be broken, and the model does not need to perform further machine learning calculations at this time, avoiding unnecessary resource consumption. This strategy improves the efficiency and response speed of the model when dealing with no penetration events, and enhances the accuracy of the judgment of penetration speed and projectile integrity.
[0087] The embodiment of the present application combines two different types of machine learning algorithms (random forest algorithm and multi-layer perceptron) to solve different prediction tasks of penetration performance, which not only improves the accuracy of prediction of each performance indicator, but also optimizes the performance of the overall model.
[0088] Compared with the prior art, the embodiments of the present application have the following beneficial effects:
[0089] First, by adopting a hierarchical model design, not only the accuracy of the prediction is improved, but also the processing efficiency is greatly improved by precisely controlling computing resources, thus achieving a rapid and accurate evaluation of the material penetration performance.
[0090] Second, a multi-layer perceptron is used for regression analysis. The multi-layer structure of the multi-layer perceptron enables it to capture and learn complex nonlinear relationships and improve the accuracy of predictions.
[0091] Third, the hierarchical machine learning model used in the embodiment of the present application not only considers the balance between computational efficiency and accuracy, so as to enable quick response in simple scenarios and provide in-depth data analysis in complex scenarios, but also significantly improves the accuracy of predictions. This strategy not only reduces the computational burden by avoiding unnecessary complex calculations in simple situations, but also reduces the risk of inaccurate prediction results.
[0092] A possible implementation of a penetration performance index prediction device is provided below, which is used to execute the various execution steps and corresponding technical effects of the penetration performance index prediction method shown in the above embodiments and possible implementations. The penetration performance index prediction device includes:
[0093] A calculation module for calculating the Pearson correlation coefficient between each pair of input parameters;
[0094] A screening module, used for screening input parameters according to the Pearson correlation coefficient to obtain screened input parameters;
[0095] The generation module is used to use the random forest algorithm and the multi-layer perceptron to perform classification prediction and regression prediction on the screened input parameters to obtain the penetration performance.
[0096] In a possible implementation, the input parameters include:
[0097] firing speed and Vickers hardness;
[0098] Yield strength, tensile strength, and total elongation at break in quasi-static tensile tests;
[0099] Specified non-proportional compressive strength and compressive strength for quasi-static compression tests;
[0100] Dynamic yield strength, tensile strength and dynamic total elongation at break from dynamic tensile tests;
[0101] Dynamic yield strength and compressive strength at dynamic compression of 3000s-1;
[0102] Dynamic yield strength and compressive strength at dynamic compression of 5000s-1.
[0103] In a possible implementation, the generating module includes:
[0104] A classification prediction submodule is used to input the screened input parameters into the first random forest algorithm for classification prediction to obtain a penetration determination result;
[0105] A generation submodule is used for inputting the screened input parameters into a multi-layer perceptron for regression prediction to obtain a non-zero penetration velocity when the penetration judgment result is judged as penetration, and inputting the screened input parameters into a second random forest algorithm for classification prediction to obtain the projectile integrity, and the projectile integrity corresponds to the result of: projectile fragmentation or projectile integrity;
[0106] When it is determined that the projectile has not penetrated based on the penetration determination result, the penetration velocity is output as 0 and the projectile integrity is output as projectile fragmentation.
[0107] In a possible implementation, the computing module is specifically configured to:
[0108] Calculate the covariance and standard deviation between each pair of input parameters;
[0109] Dividing the covariance by the standard deviation gives the Pearson correlation coefficient.
[0110] In a possible implementation, the screening module is specifically used to:
[0111] The absolute values of the Pearson correlation coefficients are sorted in order, and multiple input parameters are selected as screened input parameters.
[0112] Therefore, the present application discloses a method and device for predicting a penetration performance index. First, the Pearson correlation coefficient between each pair of input parameters is calculated. Secondly, the input parameters are screened according to the Pearson correlation coefficient to obtain screened input parameters. Finally, the random forest algorithm and the multi-layer perceptron are used to perform classification prediction and regression prediction on the screened input parameters to obtain the penetration performance. The prediction of the penetration performance index is completed, which can ensure that the prediction of each indicator is optimized to meet its specific analysis needs. In addition, the processing efficiency can be greatly improved by precisely controlling the computing resources, thereby finally realizing the rapid and accurate evaluation of the material penetration performance.
[0113] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the protection scope of the present application.
Claims
1. A penetration performance index prediction method, characterized in that: The method is applied to a hierarchical machine learning model, and the method comprises: Calculate the Pearson correlation coefficient between each pair of input parameters; Filter the input parameters according to the Pearson correlation coefficient to obtain filtered input parameters; The random forest algorithm and multi-layer perceptron are used to perform classification prediction and regression prediction on the screened input parameters to obtain the penetration performance.
2. The penetration performance index prediction method according to claim 1, characterized in that: The input parameters include: Firing speed, Vickers hardness; Yield strength, tensile strength, and total elongation at break in quasi-static tensile tests; Specified non-proportional compressive strength and compressive strength for quasi-static compression tests; Dynamic yield strength, tensile strength and dynamic total elongation at break from dynamic tensile tests; Dynamic yield strength and compressive strength at dynamic compression of 3000s-1; Dynamic yield strength and compressive strength at dynamic compression of 5000s-1.
3. The penetration performance index prediction method according to claim 1, characterized in that: The penetration performance includes penetration speed, penetration judgment result and projectile integrity. The random forest algorithm includes the first random forest algorithm and the second random forest algorithm. The steps of using the random forest algorithm and the multi-layer perceptron to perform classification prediction and regression prediction on the screened input parameters to obtain the penetration performance include: The screened input parameters are input into the first random forest algorithm for classification prediction to obtain the penetration determination result; When the penetration is judged as penetration according to the penetration judgment result, the screened input parameters are input into the multi-layer perceptron for regression prediction to obtain a non-zero penetration velocity, and the screened input parameters are input into the second random forest algorithm for classification prediction to obtain the projectile integrity. The projectile integrity corresponds to: projectile fragmentation or projectile integrity; When it is determined that the projectile has not penetrated based on the penetration determination result, the penetration velocity is output as 0 and the projectile integrity is output as projectile fragmentation.
4. The penetration performance index prediction method according to claim 1, characterized in that: The steps to calculate the Pearson correlation coefficient between each pair of input parameters include: Calculate the covariance and standard deviation between each pair of input parameters; Dividing the covariance by the standard deviation gives the Pearson correlation coefficient.
5. The penetration performance index prediction method according to claim 1, characterized in that: The step of filtering the input parameters according to the Pearson correlation coefficient to obtain the filtered input parameters includes: The absolute values of the Pearson correlation coefficients are sorted in order, and multiple input parameters are selected as screened input parameters.
6. A penetration performance index prediction device, characterized in that: The device comprises: A calculation module for calculating the Pearson correlation coefficient between each pair of input parameters; A screening module, used for screening input parameters according to the Pearson correlation coefficient to obtain screened input parameters; The generation module is used to use the random forest algorithm and the multi-layer perceptron to perform classification prediction and regression prediction on the screened input parameters to obtain the penetration performance.
7. The penetration performance index prediction device according to claim 6, characterized in that: The input parameters include: Firing speed and Vickers hardness; Yield strength, tensile strength, and total elongation at break in quasi-static tensile tests; Specified non-proportional compressive strength and compressive strength for quasi-static compression tests; Dynamic yield strength, tensile strength and dynamic total elongation at break from dynamic tensile tests; Dynamic yield strength and compressive strength at dynamic compression of 3000s-1; Dynamic yield strength and compressive strength at dynamic compression of 5000s-1.
8. The penetration performance index prediction device according to claim 6, characterized in that: The generating module comprises: A classification prediction submodule is used to input the screened input parameters into the first random forest algorithm for classification prediction to obtain a penetration determination result; A generation submodule is used for inputting the screened input parameters into a multi-layer perceptron for regression prediction to obtain a non-zero penetration velocity when the penetration judgment result is judged as penetration, and inputting the screened input parameters into a second random forest algorithm for classification prediction to obtain the projectile integrity, and the projectile integrity corresponds to the result of: projectile fragmentation or projectile integrity; When it is determined that the projectile has not penetrated based on the penetration determination result, the penetration velocity is output as 0 and the projectile integrity is output as projectile fragmentation.
9. The penetration performance index prediction device according to claim 6, characterized in that: Computing module, specifically used for: Calculate the covariance and standard deviation between each pair of input parameters; Dividing the covariance by the standard deviation gives the Pearson correlation coefficient.
10. The penetration performance index prediction device according to claim 6, characterized in that: The screening module is specifically used for: The absolute values of the Pearson correlation coefficients are sorted in order, and multiple input parameters are selected as screened input parameters.
