Mechanical superstructure anti-penetration performance prediction method and device based on machine learning

Abstract

The invention relates to a mechanical superstructure anti-penetration performance prediction method and device based on machine learning, and is applied to the technical field of impact collision protection. The method comprises the steps of obtaining size data of a mechanical superstructure; and performing performance analysis on the size data based on the prediction model, and generating a prediction result of the anti-penetration performance. Wherein the prediction result comprises the ballistic limit velocity of the mechanical superstructure, the relationship between the ballistic resistance and the time under the ballistic limit velocity and the relationship between the residual ballistic velocity and the time, and the prediction model is obtained through training by utilizing the size sample data and anti-penetration sample data determined based on the size sample data. According to the method, related features of the size data are extracted by utilizing the prediction model, efficient prediction between the anti-penetration performance of the mechanical superstructure and the structure size data is realized by utilizing the dependency relationship between the features learned by utilizing the sample data, and the design period of the mechanical superstructure can be shortened. Therefore, the anti-penetration performance prediction efficiency can be improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of impact collision protection, and particularly relates to a mechanical superstructure anti-penetration performance prediction method and device based on machine learning. BACKGROUND

[0002] With the continuous development of aerospace technology, the performance requirements of structural materials under extreme environmental conditions such as external impact and explosion are getting higher and higher. Anti-penetration performance prediction of structural materials can evaluate the impact resistance of structural materials in advance, so as to optimize the design scheme of structural materials, reduce costs and improve protection efficiency.

[0003] Generally, the anti-penetration performance of structural materials can be evaluated and predicted by experimental and simulation methods. However, the design cycle of the simulation and experiment process is relatively long, which reduces the efficiency of anti-penetration performance prediction. SUMMARY

[0004] The present disclosure is proposed in view of the above problems. The present disclosure provides a mechanical superstructure anti-penetration performance prediction method and device based on machine learning, which is beneficial to improve the efficiency of anti-penetration performance prediction.

[0005] According to one aspect of the present disclosure, a mechanical superstructure anti-penetration performance prediction method based on machine learning is provided, comprising: obtaining size data of a mechanical superstructure; performing performance analysis on the size data based on a prediction model to generate a prediction result of anti-penetration performance, the prediction result comprising a ballistic limit speed of the mechanical superstructure, a relationship between ballistic resistance and time and a relationship between residual velocity and time under the ballistic limit speed, the prediction model being obtained by training size sample data and anti-penetration sample data determined based on the size sample data.

[0006] Optionally, the performance analysis on the size data based on the prediction model to generate the prediction result of anti-penetration performance comprises: performing structure feature extraction on the size data to obtain a structure feature vector carrying a time sequence; performing time sequence feature extraction on the structure feature vector carrying the time sequence to obtain a time sequence feature vector; performing dependency relationship analysis based on the time sequence feature vector and the structure feature vector carrying the time sequence to generate the prediction result of anti-penetration performance.

[0007] Optionally, the obtaining of the size data of the mechanical superstructure comprises: obtaining a structure space range of the mechanical superstructure; The structure space range is sampled based on a multi-dimensional Latin hypercube model to generate a set of the size data.

[0008] Optionally, before the performance analysis of the size data based on the prediction model to generate the prediction result of the penetration resistance performance, the method comprises: The structure space range is sampled based on a multi-dimensional Latin hypercube model to generate a plurality of sets of the size sample data. The penetration resistance sample data is generated based on a penetration resistance simulation model and the size sample data. The size sample data and the corresponding penetration resistance sample data are divided into a training sample set and a test sample set. The training prediction model is trained based on the training sample set to determine a test prediction model. The test prediction model is tested based on the test sample set to obtain the prediction model.

[0009] Optionally, after the penetration resistance sample data is generated based on the penetration resistance simulation model and the size sample data, the method further comprises: The penetration resistance sample data and the size sample data are normalized to obtain penetration resistance normalized sample data and size normalized sample data. The penetration resistance normalized sample data is taken as the penetration resistance sample data, and the size normalized sample data is taken as the size sample data.

[0010] Optionally, before the test prediction model is tested based on the test sample set to obtain the prediction model, the method further comprises: The test prediction model is error-evaluated based on a prediction result generated by the test prediction model and the penetration resistance sample data in the test sample set to determine an error index parameter. In a case where the error index parameter meets an error range, the test prediction model is taken as the prediction model.

[0011] According to another aspect of the present disclosure, a mechanical superstructure penetration resistance performance prediction device based on machine learning is provided, comprising: An acquisition module is configured to acquire size data of a mechanical superstructure. A generation module is configured to perform performance analysis on the size data based on a prediction model to generate a prediction result of penetration resistance performance, the prediction result comprising a ballistic limit speed of the mechanical superstructure, a relationship between a ballistic resistance and time at the ballistic limit speed, and a relationship between a residual bullet speed and time, the prediction model being obtained by training size sample data and penetration resistance sample data determined based on the size sample data.

[0012] According to still another aspect of the present disclosure, an electronic device is provided, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, the processor executing the computer program to implement the above machine learning based mechanical superstructure anti-penetration performance prediction method.

[0013] According to another aspect of the present disclosure, a computer readable storage medium is provided, having stored thereon a computer program, the program being executed by a processor to implement the above machine learning based mechanical superstructure anti-penetration performance prediction method.

[0014] According to still another aspect of the present disclosure, a computer program product is provided, comprising computer readable code, or a non-volatile computer readable storage medium carrying computer readable code, when the computer readable code is executed in a processor of an electronic device, the processor in the electronic device executes to implement the above machine learning based mechanical superstructure anti-penetration performance prediction method.

[0015] In the present disclosure, size data of a mechanical superstructure is acquired; performance analysis is performed on the size data based on a prediction model to generate a prediction result of anti-penetration performance. The prediction result includes a ballistic limit speed of the mechanical superstructure, a relationship between ballistic resistance and time and a relationship between residual bullet speed and time at the ballistic limit speed, and the prediction model is obtained by training size sample data and anti-penetration sample data determined based on the size sample data. The prediction model extracts relevant features of the size data, and through the dependency relationship between the features learned by the sample data, efficient prediction between the anti-penetration performance of the mechanical superstructure and the structure size data is achieved, which can shorten the design cycle of the mechanical superstructure. Therefore, the efficiency of anti-penetration performance prediction can be improved.

[0016] It is to be understood that both the foregoing general description and the following detailed description are exemplary and intended to provide further explanation of the subject technology. BRIEF DESCRIPTION OF DRAWINGS

[0017] The above and other objects, features and advantages of the present disclosure will become more apparent from the following detailed description taken in conjunction with the accompanying drawings, in which like reference characters refer to like elements in the several views. The accompanying drawings are intended to provide a further understanding about the embodiments of the present disclosure and constitute a part of the specification, which serve to explain the present disclosure together with the embodiments of the present disclosure and do not constitute a limitation to the present disclosure. In the drawings, like reference numerals refer to like elements or steps throughout.

[0018] Figure 1 A flowchart of a machine learning based mechanical superstructure anti-penetration performance prediction method provided by the present disclosure.

[0019] Figure 2 A schematic diagram of a bionic spiral sandwich structure provided for the present disclosure.

[0020] Figure 3 A structural schematic diagram of a device for predicting the anti-penetration performance of a mechanical superstructure based on machine learning provided for the present disclosure.

[0021] Figure 4 A hardware block diagram of an electronic device provided for the present disclosure.

[0022] Figure 5 A schematic diagram of a computer program product provided for the present disclosure. DETAILED DESCRIPTION

[0023] In order for those skilled in the art to more clearly understand the technical solutions of the present application, the application scenario of the present application solution will be described first as follows.

[0024] With the continuous development of aerospace technology, the performance requirements of structural materials under extreme environmental conditions such as external impact and explosion are becoming higher and higher. Among them, the application of composite sandwich structures is becoming more and more widespread, especially in structural components with strict requirements for lightweight, high strength and good anti-penetration performance. Traditional composite sandwich structures are gradually difficult to meet the demand for lightweight, and bionic spiral mechanical superstructure is a new structure form. This structure is designed by imitating the spiral stacking structure of biological exoskeleton in nature and combining the excellent performance of composite materials, which is a new type of lightweight high-strength structure. The bionic spiral mechanical superstructure can effectively improve the anti-delamination and anti-penetration performance of the structure with its unique spiral layer panel design, and at the same time meet the demand for lightweight. Among them, penetration refers to the penetration of projectiles, fragments, metal jets and other penetrating bodies into or through the target by relying on their own kinetic energy. And the anti-penetration performance refers to the ability of the structural material to resist penetration or damage when subjected to high-speed impact or penetrating load (such as bullets, shrapnel, armor-piercing bullets, etc.), that is, the impact resistance. The anti-penetration performance prediction of the structural material can evaluate the impact resistance of the structural material in advance, so as to optimize the design scheme of the structural material, reduce the cost and improve the protection efficiency.

[0025] At present, the anti-penetration performance of the structural material can be evaluated and predicted by experimental and simulation methods. However, due to the long design cycle of the simulation and experimental process, the efficiency of the anti-penetration performance prediction will be reduced.

[0026] To solve the above technical problems, the present disclosure provides a mechanical superstructure anti-penetration performance prediction method and device based on machine learning. In the present disclosure, the size data of the mechanical superstructure is obtained; the performance of the size data is analyzed based on a prediction model to generate a prediction result of the anti-penetration performance. The prediction result includes the ballistic limit speed of the mechanical superstructure, the relationship between the ballistic resistance and time at the ballistic limit speed, and the relationship between the residual bullet speed and time, and the prediction model is obtained by training the size sample data and the anti-penetration sample data determined based on the size sample data. The related features of the size data are extracted by using the prediction model, and the dependence relationship between the features learned by using the sample data is used to realize efficient prediction between the anti-penetration performance of the mechanical superstructure and the structure size data, which can shorten the design cycle of the mechanical superstructure. Therefore, the efficiency of the anti-penetration performance prediction can be improved.

[0027] To make the purpose, technical solutions and advantages of the present disclosure more obvious, the example embodiments according to the present disclosure will be described in detail below with reference to the drawings. Obviously, the described embodiments are only part of the embodiments of the present disclosure, not all embodiments of the present disclosure, and it should be understood that the present disclosure is not limited by the example embodiments described here.

[0028] Figure 1 A flowchart of a mechanical superstructure anti-penetration performance prediction method based on machine learning provided by the present disclosure is shown in FIG. 1. As shown in FIG. 1, the method comprises: Figure 1 S101: obtaining size data of the mechanical superstructure.

[0029] Specifically, in the present embodiment, the mechanical superstructure can be a bionic spiral mechanical superstructure, as shown in FIG. 2, Figure 2 Figure 2 which is a schematic diagram of a bionic spiral sandwich structure. In Figure 2 , the 4D printed bionic spiral composite panel 02 can be adhered by the aluminum honeycomb 01. When building the bionic spiral composite panel 02, the panel can be an odd layer panel or an even layer panel. For the even layer panel, the layering sequence can be [... / 3θ / 2θ / θ / 0°] s , which is symmetrically laid according to the layering sequence. For the odd layer panel, the layering sequence can be [... / 3θ / 2θ / θ / 0°) s / 0°], which increases a layer of 0° panel relative to the even layer panel, so that the panel can be symmetrically laid.

[0030] In the present embodiment, the size data includes the spiral angle, the spiral layer number, the height of the core, the side length and the thickness, and the size data can be more accurately used to establish the mapping relationship between the spiral angle, the spiral layer number and the anti-penetration performance of the mechanical superstructure.

[0031] ​​S102: Perform performance analysis on the size data based on the prediction model to generate a prediction result of the penetration resistance performance.

[0032] Specifically, the prediction model is obtained by training the size sample data and the penetration resistance sample data determined based on the size sample data. In this embodiment, the prediction model can quickly extract the structural features of the size data, and generate the prediction result of the penetration resistance performance of the current size data by the relationship between the time series of the penetration resistance performance learned in advance for the sample data and the size data. The prediction result includes the ballistic limit speed of the mechanical superstructure, the relationship between the ballistic resistance and time at the ballistic limit speed, and the relationship between the residual bullet speed and time.

[0033] In this disclosure, size data of a mechanical superstructure is obtained; performance analysis is performed on the size data based on a prediction model to generate a prediction result of the penetration resistance performance. The prediction result includes the ballistic limit speed of the mechanical superstructure, the relationship between the ballistic resistance and time at the ballistic limit speed, and the relationship between the residual bullet speed and time. The prediction model is obtained by training the size sample data and the penetration resistance sample data determined based on the size sample data. The prediction model extracts the relevant features of the size data, and realizes efficient prediction between the penetration resistance performance of the mechanical superstructure and the structural size data by using the dependent relationship between the features learned from the sample data, which can shorten the design cycle of the mechanical superstructure. Therefore, the efficiency of the penetration resistance performance prediction can be improved.

[0034] In one possible embodiment, an exemplary method of performing performance analysis on size data based on a prediction model to generate a prediction result of the penetration resistance performance includes: performing structural feature extraction on the size data to obtain a structural feature vector carrying a time series; performing time series feature extraction on the structural feature vector carrying the time series to obtain a time series feature vector; performing dependent relationship analysis based on the time series feature vector and the structural feature vector carrying the time series to generate the prediction result of the penetration resistance performance.

[0035] Specifically, in this embodiment, the prediction model includes a CNN (Convolutional Neural Network) model, an LSTM (Long Short-Term Memory) model, and an attention mechanism. The prediction model includes an input layer, a first convolutional layer, a second convolutional layer, a Relu activation function layer, a first maximum pooling layer, a second maximum pooling layer, a batch normalization layer, an unfolding layer, an LSTM neural network layer, a smoothing layer, an attention mechanism layer, a Dropout layer, a fully connected layer, and an output layer.

[0036] The convolutional layer in the CNN model extracts structural features from the size data to obtain a structural feature vector carrying time series. The pooling layer further reduces the dimension to retain the most significant features. The feature sequence is sent to the LSTM model, which comprehensively captures the forward and backward dependencies between the time series and the size data. Finally, the attention mechanism dynamically adjusts the feature weight according to the importance of the LSTM output features, so that the model can focus on the most critical part of the penetration process to obtain the prediction result of the penetration resistance performance.

[0037] In one possible embodiment, an exemplary method for obtaining size data of a mechanical superstructure includes: Obtaining a structural space range of the mechanical superstructure.

[0038] Specifically, the obtained structural space range needs to be consistent with the range when training the prediction model, so as to improve the accuracy of the current prediction. In the present embodiment, the structural space range can be the spiral angle θ (5-30°), the number of spiral layers N (10-60 layers), the height H (5-50 mm) of the core body, the side length L (3-10 mm), and the thickness t (0.01-0.08 mm) of the biomimetic spiral mechanical superstructure panel.

[0039] Based on the multi-dimensional Latin hypercube model, the structural space range is sampled and processed to generate a set of size data.

[0040] Specifically, according to the structural space range and the required number of samples, a set of size data is randomly sampled. In the present embodiment, only one set of unique size data is required, so the number of samples is 1. After randomly generating the size data, it can be verified whether the size data is the optimal result. If not, continue to sample and generate, and if so, the current result is taken as the size data.

[0041] In one possible embodiment, before the prediction model is used to analyze the performance of the size data and generate the prediction result of the penetration resistance performance, the method includes: Obtaining a structural space range of the mechanical superstructure; based on a multi-dimensional Latin hypercube model, the structural space range is sampled and processed to generate a plurality of sets of size sample data.

[0042] Specifically, in the present embodiment, a plurality of sets of size sample data are required when training the prediction model, therefore, the number of samples of the multi-dimensional Latin hypercube model can be set to N. According to the structural space range and the required number of samples, a plurality of sets of size sample data are randomly sampled.

[0043] Based on the penetration resistance simulation model and the size sample data, penetration resistance sample data is generated.

[0044] Specifically, in the embodiment, the anti-penetration simulation model can be a finite element simulation model, based on the generated size sample data, the bullet speed is adjusted, the ballistic limit speed of the mechanical superstructure is calculated by dichotomy, the curve of the ballistic resistance changing with time and the curve of the residual bullet speed changing with time at the ballistic limit speed are generated. The ballistic limit speed, the curve of the ballistic resistance changing with time and the curve of the residual bullet speed changing with time are taken as the anti-penetration sample data.

[0045] The size sample data and the corresponding anti-penetration sample data are divided into a training sample set and a test sample set.

[0046] Specifically, the anti-penetration sample data can be smoothed and filtered by an S-G filter, and the sample data is divided into a training sample set and a test sample set.

[0047] In addition, a validation sample set can be further divided, for example, after randomly rearranging the size sample data and the anti-penetration sample data, 70% of the size sample data and the anti-penetration sample data are selected as the training sample set, 15% of the size sample data and the anti-penetration sample data are selected as the validation sample set, and 15% of the size sample data and the anti-penetration sample data are selected as the test sample set.

[0048] The training sample set is used to train the to-be-trained prediction model to determine the to-be-tested prediction model.

[0049] Specifically, the training sample set is input into the to-be-trained prediction model, the to-be-trained prediction model is trained and the model parameters are adjusted, so that the model can learn the dependency relationship and corresponding features between the size sample data and the anti-penetration sample data. When the output result of the model meets the condition, the to-be-tested prediction model can be determined.

[0050] The to-be-tested prediction model is tested based on the test sample set to obtain the prediction model.

[0051] For example: The to-be-tested prediction model is error-evaluated based on the prediction result generated by the to-be-tested prediction model and the anti-penetration sample data in the test sample set, and an error index parameter is determined; in the case that the error index parameter meets the error range, the to-be-tested prediction model is taken as the prediction model.

[0052] Specifically, the prediction result generated by the to-be-tested prediction model is error-calculated, wherein the error index can be root mean square error, mean absolute error, mean absolute percentage error and coefficient of determination. Whether the error index parameter is within the error range is judged to determine whether the current to-be-tested prediction model meets the training requirement. If the error range is met, the to-be-tested prediction model is taken as the prediction model; if the error range is not met, the model training is continued.

[0053] In a possible embodiment, after generating the anti-penetration sample data based on the anti-penetration simulation model and the size sample data, the method further includes: normalizing the anti-penetration sample data and the size sample data to obtain anti-penetration normalized sample data and size normalized sample data; taking the anti-penetration normalized sample data as the anti-penetration sample data, and taking the size normalized sample data as the size sample data.

[0054] Specifically, in order to accelerate the model training speed and improve the generalization energy of the model, the anti-penetration sample data and the size sample data can be normalized to limit the data in the range of [0, 1].

[0055] Figure 3 A structure diagram of a mechanical superstructure anti-penetration performance prediction device based on machine learning is provided for the present disclosure. As shown in the figure, Figure 3 The device 300 includes an acquisition module 310 and a generation module 320.

[0056] The acquisition module 310 is configured to acquire size data of a mechanical superstructure. The generation module 320 is configured to perform performance analysis on the size data based on a prediction model to generate a prediction result of anti-penetration performance, the prediction result including a ballistic limit speed of the mechanical superstructure, a relationship between a ballistic resistance and time and a relationship between a residual speed and time at the ballistic limit speed, and the prediction model being obtained by training size sample data and anti-penetration sample data determined based on the size sample data.

[0057] Optionally, the generation module is configured to: perform structural feature extraction on the size data to obtain a structural feature vector carrying a time sequence; perform time sequence feature extraction on the structural feature vector carrying the time sequence to obtain a time sequence feature vector; perform dependency analysis based on the time sequence feature vector and the structural feature vector carrying the time sequence to generate the prediction result of the anti-penetration performance.

[0058] Optionally, the acquisition module is configured to: acquire a structure space range of the mechanical superstructure; perform sampling processing on the structure space range based on a multi-dimensional Latin hypercube model to generate a set of the size data.

[0059] Optionally, the device further includes: Before the performance analysis on the size data based on the prediction model to generate the prediction result of the anti-penetration performance, the structure space range is sampled based on a multi-dimensional Latin hypercube model to generate a plurality of groups of the size sample data; Based on the anti-penetration simulation model and the size sample data, anti-penetration sample data is generated. The size sample data and the corresponding anti-penetration sample data are divided into a training sample set and a test sample set. Based on the training sample set, a to-be-trained prediction model is trained to determine a to-be-tested prediction model. Based on the test sample set, the to-be-tested prediction model is tested to obtain the prediction model.

[0060] Optionally, the apparatus further comprises: After the anti-penetration sample data is generated based on the anti-penetration simulation model and the size sample data, the anti-penetration sample data and the size sample data are normalized to obtain anti-penetration normalized sample data and size normalized sample data. The anti-penetration normalized sample data is used as the anti-penetration sample data, and the size normalized sample data is used as the size sample data.

[0061] Optionally, the apparatus further comprises: Before the to-be-tested prediction model is tested based on the test sample set to obtain the prediction model, the to-be-tested prediction model is error-evaluated based on a prediction result generated by the to-be-tested prediction model and anti-penetration sample data in the test sample set to determine an error index parameter. In a case where the error index parameter meets an error range, the to-be-tested prediction model is used as the prediction model.

[0062] The present application also provides an electronic device for executing the above-mentioned mechanical superstructure anti-penetration performance prediction method based on machine learning. Please refer to Figure 4 It shows a schematic diagram of an electronic device provided by some embodiments of the present application. As Figure 4 As shown, the electronic device 40 comprises a processor 400, a memory 401, a bus 402 and a communication interface 403, the processor 400, the communication interface 403 and the memory 401 are connected through the bus 402; the memory 401 stores a computer program executable on the processor 400, and the processor 400 executes the computer program to perform the mechanical superstructure anti-penetration performance prediction method based on machine learning provided by any of the preceding embodiments of the present application.

[0063] The memory 401 can include a high-speed random access memory (RAM) and can also include a non-volatile memory such as at least one disk memory. The communication connection between the apparatus network element and at least one other network element is realized through at least one communication interface 403 (which can be wired or wireless), and the Internet, a wide area network, a local network, a metropolitan area network, etc. can be used.

[0064] The bus 402 can be an ISA bus, a PCI bus, an EISA bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. The memory 401 is used to store programs, and the processor 400 executes the programs after receiving execution instructions. The method for predicting the mechanical superstructure anti-penetration performance based on machine learning disclosed in any of the embodiments of the present application can be applied to the processor 400 or implemented by the processor 400.

[0065] The processor 400 can be an integrated circuit chip with signal processing capability. In the implementation process, each step of the above method can be completed by the integrated logic circuit of hardware in the processor 400 or the instruction in the form of software. The processor 400 described above can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a ready-to-program gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. Each method, step and logic block disclosed in the embodiments of the present application can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor. The steps of the method disclosed in combination with the embodiments of the present application can be directly embodied as a hardware code processor for execution, or a combination of hardware and software modules in the code processor for execution. The software module can be located in a random access memory, a flash memory, a read-only memory, a programmable read-only memory, an electrically erasable programmable memory, a register, etc. The storage medium in the art. The storage medium is located in the memory 401, and the processor 400 reads the information in the memory 401, and combines the hardware to complete the steps of the above method.

[0066] The electronic device provided by the embodiments of the present application and the method for predicting the mechanical superstructure anti-penetration performance based on machine learning provided by the embodiments of the present application have the same beneficial effects as the method they use, run or implement.

[0067] The embodiment of the present application further provides a computer readable storage medium corresponding to the method for predicting the anti-penetration performance of a mechanical superstructure based on machine learning, and the computer readable storage medium can be an optical disc, and a computer program is stored on the optical disc. When the computer program is run by a processor, the method for predicting the anti-penetration performance of a mechanical superstructure based on machine learning provided by any of the foregoing embodiments is executed.

[0068] It should be noted that examples of the computer readable storage medium can further include, but are not limited to, a phase change memory (PRAM), a static random access memory (SRAM), a dynamic random access memory (DRAM), other types of random access memory (RAM), a read only memory (ROM), an electrically erasable programmable read only memory (EEPROM), a flash memory or other optical, magnetic storage medium, which will not be described one by one here.

[0069] The computer readable storage medium provided by the above embodiment of the present application and the method for predicting the anti-penetration performance of a mechanical superstructure based on machine learning provided by the embodiment of the present application have the same beneficial effects as the method adopted, run or implemented by the application program stored therein.

[0070] The embodiment of the present application further provides a computer program product 500, as shown in the figure. The computer program product carries a computer program 501, and the program code includes instructions for executing the steps of the method for predicting the anti-penetration performance of a mechanical superstructure based on machine learning described in the above method embodiments. Details can be referred to the above method embodiments, which will not be described here. Figure 5

[0071] The computer program product can be specifically implemented by hardware, software or a combination thereof. In an optional embodiment, the computer program product is specifically embodied as a computer storage medium, and in another optional embodiment, the computer program product is specifically embodied as a software product, such as a software development kit (SDK) and the like.

[0072] The basic principles of the present disclosure are described above in combination with specific embodiments, but it should be noted that the advantages, advantages, effects and the like mentioned in the present disclosure are only examples and are not limited, and these advantages, advantages, effects and the like cannot be considered as the must-have of each embodiment of the present disclosure. In addition, the above disclosed specific details are only for the purpose of example and understanding, and are not limited to the above specific details. The present disclosure is not limited to the above specific details.

[0073] ​The block diagrams of devices, apparatuses, equipment, systems referred to in the present disclosure are merely illustrative examples and are not intended to require or imply that the connection, arrangement, configuration must be as shown in the block diagrams. These devices, apparatuses, equipment, systems can be connected, arranged, configured in any manner as will be appreciated by those skilled in the art. Words such as "include," "contain," "have," etc. are open-ended words that are to be interpreted to mean "including but not limited to," and are to be interpreted not to exclude other items. The words "or" and "and" as used herein are to be interpreted as the word "and / or," and are to be interpreted not to exclude other items. The word "such as" as used herein is to be interpreted as the phrase "such as but not limited to," and is to be interpreted not to exclude other items.

[0074] In addition, as used herein, the "or" as used in the context "at least one of A, B, or C" : means A or B or C or any combination thereof. Further, the phrase "example of" is not meant to be limiting in terms of the examples described. For example, the phrase "example of A, B, or C" means A or B or C, or any combination thereof.

[0075] It is also important to note that the systems and methods of the present disclosure can be embodied in a variety of forms including, but not limited to, a data processor, a computer program product, a computer, one or more tangible computer readable storage devices, one or more computer-implemented methods, information, or a bit of information. Additionally the systems and methods of the present disclosure can be embodied as one or more computers or computer implementations that include one or more processors or one or more memory modules.

[0076] Various changes, modifications and alterations in the teachings and techniques described herein can be made without departing from the teachings that are defined by the appended claims. Further, the scope of the claims of the present disclosure is not limited to the specific aspects described herein. Processes, machines, manufacture, compositions of matter, means, methods, or steps, presently existing or later to be developed that perform substantially the same function or achieve substantially the same result as the corresponding aspects described herein can be utilized. Accordingly, the appended claims include within their scope such processes, machines, manufacture, compositions of matter, means, methods, or steps.

[0077] The above description of the disclosed aspects is meant to be illustrative of the application and not limiting. Various modifications of the aspects will be apparent to those with ordinary skill in the art, and the general principles defined herein can be applied to other aspects without departing from the scope of the present disclosure. Accordingly, the present disclosure is not intended to be limited to the aspects shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0078] The foregoing description has been presented for the purposes of illustration and description. Furthermore, the description is not intended to limit the embodiments of the disclosure to the forms disclosed herein. Although the various example aspects and embodiments have been described herein with regard to particular aspects and embodiments, those skilled in the art will recognize that certain modifications, changes, substitutions, additions and sub-combinations can be made without departing from the spirit of the disclosure.

Claims

1. A method for predicting the penetration resistance of mechanical superstructures based on machine learning, characterized in that, include: Obtain dimensional data of mechanical superstructures; Based on the prediction model, the performance of the dimensional data is analyzed to generate the prediction results of the anti-penetration performance. The prediction results include the ballistic limit velocity of the mechanical superstructure, the relationship between ballistic drag and time at the ballistic limit velocity, and the relationship between the remaining velocity and time. The prediction model is obtained by training using dimensional sample data and anti-penetration sample data determined based on the dimensional sample data.

2. The method according to claim 1, characterized in that, The step of performing performance analysis on the size data based on the prediction model to generate prediction results of penetration resistance includes: Structural features are extracted from the size data to obtain a structural feature vector carrying the time series. Temporal feature extraction is performed on the structural feature vector carrying the time series to obtain a temporal feature vector; Dependency analysis is performed based on the time-series feature vector and the structural feature vector carrying the time series to generate the prediction result of the anti-penetration performance.

3. The method according to claim 1, characterized in that, The acquisition of dimensional data of the mechanical superstructure includes: Obtain the structural space range of the mechanical superstructure; Based on the multidimensional Latin hypercube model, the spatial range of the structure is sampled to generate a set of size data.

4. The method according to claim 1, characterized in that, Before performing performance analysis on the size data based on the prediction model to generate a prediction result of the penetration resistance performance, the method includes: Obtain the structural space range of the mechanical superstructure; Based on the multidimensional Latin hypercube model, the spatial range of the structure is sampled to generate multiple sets of size sample data. Based on the anti-penetration simulation model and the size sample data, anti-penetration sample data is generated; The size sample data and the corresponding anti-penetration sample data are divided into a training sample set and a test sample set; The training sample set is used to train the prediction model to be trained, and the prediction model to be tested is determined. The prediction model to be tested is obtained by testing the test sample set.

5. The method according to claim 4, characterized in that, After generating anti-penetration sample data based on the anti-penetration simulation model and the size sample data, the method further includes: The anti-penetration sample data and the size sample data are normalized to obtain anti-penetration normalized sample data and size normalized sample data. The anti-penetration sample data is used as the anti-penetration sample data, and the size sample data is used as the size sample data.

6. The method according to claim 4, characterized in that, Before testing the prediction model based on the test sample set to obtain the prediction model, the method further includes: Based on the prediction results generated by the prediction model to be tested and the anti-penetration sample data in the test sample set, the error of the prediction model to be tested is evaluated, and the error index parameters are determined. If the error index parameters meet the error range, the prediction model to be tested is used as the prediction model.

7. A machine learning-based device for predicting the penetration resistance of mechanical superstructures, characterized in that, include: The acquisition module is used to acquire the dimensional data of the mechanical superstructure; The generation module is used to perform performance analysis on the size data based on the prediction model and generate prediction results of the anti-penetration performance. The prediction results include the ballistic limit velocity of the mechanical superstructure, the relationship between ballistic drag and time at the ballistic limit velocity, and the relationship between the remaining velocity and time. The prediction model is trained using size sample data and anti-penetration sample data determined based on the size sample data.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, The processor executes the computer program to implement the method as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by a processor to implement the method as described in any one of claims 1-6.

10. A computer program product, characterized in that, Includes computer-readable code, or a non-volatile computer-readable storage medium carrying computer-readable code, wherein when the computer-readable code is run in a processor of an electronic device, the processor in the electronic device performs any one of claims 1-6.

Images