Intelligent design method, system and storage medium of energy dissipator based on neural network

Abstract

The present invention provides an intelligent design method for energy absorbers based on neural networks, including: S1. numerical model calibration; S2. establishment of a data set; S3. training of a PSO-BP neural network; S4. establishment of an energy absorber structural design model. The present invention uses numerical simulation to analyze parameters that affect the mechanical properties of the energy absorber, combines the PSO-BP neural network to establish a prediction model from the structural size of the energy absorber to the mechanical properties, and then establishes an intelligent design model for energy absorbers based on inverse parameter optimization based on the PSO algorithm and target mechanical properties. The present invention mainly solves the problems of complex calculations, low calculation efficiency and high experimental costs in existing research methods.

Description

Technical Field

[0001] The present invention relates to the technical field of energy consumers, and in particular to an intelligent design method, system and storage medium of energy consumers based on a neural network. Background Art

[0002] Flexible protection net system is often used in geological disaster protection projects. The system consists of flexible mesh, steel wire rope, steel column and energy absorber, and is widely used to protect against serious slope geological disasters such as rockfall and debris flow. The performance of the energy absorber largely determines the flexibility of the flexible protection net and its ability to dissipate impact energy. At present, the commonly used energy dissipation parts of the system are pressure relief rings. In the calculation model of the flexible protection net, the energy absorber often adopts a simplified equivalent mechanical model. Based on the calculation model of simplified complex conditions, it is extremely important to quickly obtain a high-precision equivalent model and design an energy absorber with a specific structural size on demand for the accurate calculation and design of the flexible protection net.

[0003] Existing research on the mechanical properties of energy absorbers focuses on numerical simulation and mechanical testing. The numerical simulation method involves highly nonlinear characteristics of materials, geometry and contact, which is difficult to calculate and inefficient. The test cost is high, and energy absorbers of the same batch model also have differences in structural size and performance. It is often impossible to obtain sufficiently accurate mechanical properties through a single test. Choosing an energy absorber that is too strong will make the energy dissipation performance of the flexible protection net excessive, resulting in unnecessary costs. Choosing an energy absorber that is too weak will make the flexible protection net interception ineffective. At present, there is a lack of effective basis for selecting the size of the energy absorber according to demand.

[0004] Therefore, based on the above problems, it is urgent to propose a new technical means to quickly, conveniently and accurately calculate the mechanical properties and design structure dimensions of energy absorbers. Summary of the invention

[0005] The present invention provides an intelligent design method, system and storage medium for energy absorbers based on neural networks, uses numerical simulation to analyze parameters that affect the mechanical properties of energy absorbers, combines PSO-BP neural networks to establish a prediction model from the structural dimensions of energy absorbers to mechanical properties, and then establishes an intelligent design model for energy absorbers that optimizes inverse parameters based on the PSO algorithm and target mechanical properties. The present invention mainly solves the problems of complex calculations, low calculation efficiency and high experimental costs in existing research methods, and provides a fast, convenient and high-precision calculation and design method.

[0006] The first aspect of the embodiments of this specification discloses an intelligent design method for energy consumption device based on a neural network, comprising:

[0007] S1. Numerical model calibration:

[0008] Conduct mechanical tests on energy dissipators, debug and establish numerical models that match the test structure;

[0009] S2. Create a dataset:

[0010] Based on the debugged numerical model, the parameters of the energy dissipator structure size are analyzed, the structural size parameters and mechanical performance parameters are obtained to form a data set, the data is preprocessed, and the data set is divided into a training set and a test set;

[0011] S3. Training PSO-BP neural network:

[0012] Initialize the PSO algorithm hyperparameters, define the fitness function, use the training set to train the PSO-BP neural network, use the test set to test the trained PSO-BP neural network and calculate the fitness value, minimize the fitness value, and extract the optimal initial weight and threshold;

[0013] S4. Establish the energy dissipator structure design model:

[0014] The trained PSO-BP neural network is imported, and an intelligent design model of energy consumer is established based on the PSO-BP neural network and the PSO inverse parameter optimization algorithm.

[0015] In some embodiments of the present specification, structural dimensions are designed based on an intelligent design model of an energy absorber, and mechanical tests are performed on actual products of the structural dimensions to obtain experimental mechanical properties; the experimental mechanical properties are compared with the target mechanical properties to verify the effectiveness of the intelligent design model of the energy absorber.

[0016] In some embodiments of the present specification, the energy absorber is a pressure reducing ring, a rod-type energy absorber, a plate-type energy absorber, a double U-shaped energy absorber or a piston rod-type energy absorber.

[0017] In some embodiments of the present specification, when performing parameter analysis to establish a data set, the structural dimensions of the energy absorber are used as input parameters, and the mechanical performance index of the energy absorber is used as output parameters.

[0018] In some embodiments of the present specification, when the energy absorber is a pressure reducing ring, the structural dimensions of the energy absorber are the ring diameter, tube diameter, tube wall thickness and aluminum sleeve length of the pressure reducing ring, and the mechanical performance indicators of the energy absorber are the starting load, maximum load and energy consumption of the pressure reducing ring.

[0019] In some embodiments of the present specification, the data preprocessing is normalization processing, and the normalization processing includes minimum-maximum normalization, zero-mean normalization and decimal calibration normalization.

[0020] In some embodiments of the present specification, when dividing a data set, a division method is selected according to the amount of data and the difficulty of training, and the division methods include a leave-out method, a cross-validation method, a leave-one-out method, and a bootstrap method.

[0021] In some embodiments of this specification, the PSO algorithm neural network weights and threshold position and speed update functions are:

[0022] ;

[0023] ;

[0024] Where: It is a particle i exist t The speed of the moment; It is a particle i exist t The location at the moment; is the inertia weight; is the learning factor; is a random number in the interval [0,1]; Particles i The best position encountered so far and the best positions encountered so far by all particles.

[0025] A second aspect of the embodiments of this specification discloses a neural network-based intelligent design system for energy consumers, which is used to implement any of the neural network-based intelligent design methods for energy consumers described above;

[0026] The energy consumption device intelligent design system based on neural network includes:

[0027] Numerical model calibration module, used to conduct mechanical tests on energy dissipators, debug and establish a numerical model that matches the test structure;

[0028] The data set establishment module is used to perform parameter analysis of the energy dissipator structure size based on the debugged numerical model, obtain the structure size parameters and mechanical performance parameters to form a data set, perform data preprocessing, and divide the data set into a training set and a test set;

[0029] PSO-BP neural network training module, used to initialize PSO algorithm hyperparameters, define fitness function, train PSO-BP neural network using training set, test trained PSO-BP neural network using test set and calculate fitness value, minimize fitness value, and extract optimal initial weight and threshold;

[0030] The energy absorber structure design model building module is used to import the trained PSO-BP neural network and build the energy absorber intelligent design model based on the PSO-BP neural network and the PSO inverse parameter optimization algorithm.

[0031] A third aspect of the embodiments of this specification discloses a computer-readable storage medium, wherein the storage medium stores computer instructions. When a computer reads the computer instructions, the computer executes any one of the above-mentioned intelligent design methods for energy consumers based on a neural network.

[0032] The embodiments of this specification can at least achieve the following beneficial effects:

[0033] (1) This invention proposes for the first time an intelligent design method for energy consumers based on neural networks, which can be used to quickly train intelligent models for forward prediction and reverse design of energy consumers, simplify the calculation and design process, lower the threshold for calculation and design, and greatly improve efficiency.

[0034] (2) The influencing parameters of the mechanical properties of the decompression ring were clarified, and a method for extracting the training features of the energy absorber was proposed.

[0035] (3) Based on the method combining numerical simulation and neural network, the mechanical properties of the energy absorber can be calculated with sufficient accuracy and the reasonable structural dimensions can be designed, providing effective data support for the subsequent calculation and design of the energy absorber.

[0036] The energy consumption device intelligent design method based on neural network of the present invention has clear logic, reasonable simplification and high operability. Compared with the traditional calculation and design process, the training neural network is more efficient, fast and operable, and is more suitable for engineering design. The present invention has substantial progress and is suitable for wide promotion and application. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0038] Figure 1 It is a schematic diagram of the intelligent design method of energy consumer based on neural network involved in the present invention.

[0039] Figure 2 It is a schematic diagram of the mechanical test and energy dissipator geometric model involved in the present invention.

[0040] Figure 3 It is a schematic diagram of the inversion numerical simulation force-displacement curve involved in the present invention.

[0041] Figure 4 A schematic diagram of the calculation conditions and results of some data sets involved in the present invention.

[0042] Figure 5a is a schematic diagram showing the comparison between the starting load prediction result and the actual value involved in the present invention.

[0043] Figure 5 b is a schematic diagram comparing the maximum load prediction result involved in the present invention with the actual value.

[0044] Figure 5 c is a schematic diagram comparing the energy consumption prediction result involved in the present invention with the actual value.

[0045] Figure 6 It is a schematic diagram of a fitness value variation curve when training a neural network involved in the present invention.

[0046] Figure 7 It is a schematic diagram comparing the design structure size test results and target performance involved in the present invention. Implementation

[0047] In the following, only some exemplary embodiments are briefly described. As those skilled in the art will appreciate, the described embodiments may be modified in various ways without departing from the spirit or scope of the embodiments of the present invention. Therefore, the drawings and descriptions are considered to be exemplary and non-restrictive in nature.

[0048] The disclosure below provides many different embodiments or examples to implement different structures of the embodiments of the present invention. In order to simplify the disclosure of the embodiments of the present invention, the parts and settings of specific examples are described below. Of course, they are only examples, and the purpose is not to limit the embodiments of the present invention. In addition, the embodiments of the present invention can repeat reference numbers and / or reference letters in different examples, and this repetition is for the purpose of simplification and clarity, which itself does not indicate the relationship between the various embodiments and / or settings discussed.

[0049] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0050] like Figure 1 As shown, the first aspect of the embodiment of this specification discloses an intelligent design method for energy consumption device based on neural network, including:

[0051] S1. Numerical model calibration:

[0052] Conduct mechanical tests on energy dissipators, debug and establish numerical models that match the test structure;

[0053] Furthermore, in order to improve the accuracy of calculation and design, the accuracy should be increased as much as possible when calibrating the numerical model to ensure that the accumulated errors of forward prediction and reverse design are small enough.

[0054] S2. Create a dataset:

[0055] Based on the debugged numerical model, the parameters of the energy dissipator structure size are analyzed, the structural size parameters and mechanical performance parameters are obtained to form a data set, the data is preprocessed, and the data set is divided into a training set and a test set;

[0056] S3. Training PSO-BP neural network:

[0057] Initialize the PSO algorithm hyperparameters, define the fitness function, use the training set to train the PSO-BP neural network, use the test set to test the trained PSO-BP neural network and calculate the fitness value, minimize the fitness value, and extract the optimal initial weight and threshold;

[0058] S4. Establish the energy dissipator structure design model:

[0059] The trained PSO-BP neural network is imported, and an intelligent design model of energy consumer is established based on the PSO-BP neural network and the PSO inverse parameter optimization algorithm.

[0060] In some embodiments, the structural dimensions are designed based on the intelligent design model of the energy absorber, and mechanical tests are conducted on actual products of the structural dimensions to obtain experimental mechanical properties; the experimental mechanical properties are compared with the target mechanical properties to verify the effectiveness of the intelligent design model of the energy absorber.

[0061] In some embodiments, the energy absorber includes but is not limited to a pressure reducing ring, a rod type energy absorber, a plate type energy absorber, a U-shaped energy absorber and a piston rod type energy absorber.

[0062] In some embodiments, in order to establish a more effective data set, the parameters of the structural dimensions that have the greatest impact on the mechanical properties of the energy absorber should be selected during parameter analysis. Taking the pressure reducing ring as an example, key dimensions such as the ring diameter, tube diameter, tube wall thickness, and aluminum sleeve length are selected as input parameters X; the starting load, maximum load, and energy consumption that are of greatest concern in terms of its mechanical properties are selected as output parameters Y.

[0063] In some embodiments, in order to improve the training accuracy of the model, the sample data needs to be normalized before training, including but not limited to minimum-maximum normalization, zero-mean normalization, and decimal calibration normalization. When dividing the data set, select an appropriate division method based on the amount of data and the difficulty of training, including but not limited to the hold-out method, cross-validation method, leave-one-out method, and bootstrap method.

[0064] In some embodiments, using a training set to train a neural network requires the initial definition of the number of nodes in the input layer, hidden layer, and output layer of the neural network, the learning factor of the PSO optimization algorithm, the population size, the number of population updates, and other key hyperparameters.

[0065] In some embodiments, in order to obtain the optimal network structure, before using the training set to train the neural network, it is necessary to debug key hyperparameters such as the number of hidden layer nodes of the neural network and the learning factor, population size, and number of population updates in the PSO optimization algorithm.

[0066] In some embodiments, the PSO algorithm neural network weights and threshold position and velocity update functions are:

[0067] ;

[0068] ;

[0069] Where: It is a particle i exist t The speed of the moment; It is a particle i exist t The location at the moment; is the inertia weight; is the learning factor; is a random number in the interval [0,1]; Particles i The best position encountered so far and the best positions encountered so far by all particles.

[0070] In some embodiments, in order to train the model more efficiently, the selection of the fitness function should comprehensively consider the complexity and computational efficiency of the model, or you can design the function that best suits your own data set and network structure.

[0071] A second aspect of the embodiments of this specification discloses a neural network-based intelligent design system for energy consumers, which is used to implement any of the neural network-based intelligent design methods for energy consumers described above;

[0072] The energy consumption device intelligent design system based on neural network includes:

[0073] Numerical model calibration module, used to conduct mechanical tests on energy dissipators, debug and establish a numerical model that matches the test structure;

[0074] The data set establishment module is used to perform parameter analysis of the energy dissipator structure size based on the debugged numerical model, obtain the structure size parameters and mechanical performance parameters to form a data set, perform data preprocessing, and divide the data set into a training set and a test set;

[0075] PSO-BP neural network training module, used to initialize PSO algorithm hyperparameters, define fitness function, train PSO-BP neural network using training set, test trained PSO-BP neural network using test set and calculate fitness value, minimize fitness value, and extract optimal initial weight and threshold;

[0076] The energy absorber structure design model building module is used to import the trained PSO-BP neural network and build the energy absorber intelligent design model based on the PSO-BP neural network and the PSO inverse parameter optimization algorithm.

[0077] A third aspect of the embodiments of this specification discloses a computer-readable storage medium, wherein the storage medium stores computer instructions. When a computer reads the computer instructions, the computer executes any one of the above-mentioned intelligent design methods for energy consumers based on a neural network.

[0078] The technical concept of the present invention is as follows:

[0079] The method includes the following key steps: conducting mechanical tests on energy absorbers; establishing an accurate numerical model and conducting parameter analysis; taking a decompression ring as an example, extracting important structural dimensions of the energy absorber including ring diameter, tube diameter, tube wall thickness and aluminum sleeve length as input parameters of sample data; extracting the starting load, maximum load and energy consumption of the decompression ring stretching process as output parameters of sample data; preprocessing the input parameters and output parameters of the data set, and dividing the training set and the test set according to a certain ratio; constructing a PSO-BP neural network, inputting the training set into the PSO-BP neural network for training; using the test set to perform error analysis on the trained PSO-BP neural network; establishing an intelligent design model of energy absorbers based on the required mechanical properties of energy absorbers and a PSO inverse parameter optimization algorithm; carrying out actual product mechanical tests on energy absorbers designed according to specific mechanical performance requirements, and comparing them with numerical simulations and required performance to verify the effectiveness of the design model. The intelligent design method described in the present invention can quickly calculate the mechanical performance index through the structural dimensions of the energy absorber, and design a reasonable structural dimension with given mechanical performance indexes. This method can effectively replace complex and tedious mechanical tests and numerical simulations to analyze the mechanical properties and design the structural dimensions of the energy absorber.

[0080] like Figure 1 As shown, the design process is as follows:

[0081] Step 1: Numerical model calibration

[0082] Carry out mechanical tests on energy absorbers. Figure 2 The figure is a schematic diagram of the mechanical test and geometric model of the decompression ring of the present invention. The test adopts a 150t range vertical tensile testing machine, the lower end of the decompression ring wire rope is connected to the shackle and fixed on the steel beam anchored to the ground, and the upper end is connected to the shackle and fixed on the tensile arm of the testing machine. The tensile force and displacement data are recorded every 0.2s during the stretching process to obtain the force-displacement curve of the test process.

[0083] The numerical model of inversion analysis was established, and the finite element method was selected using LS-DYNA calculation software to establish the calculation model. The calculation results were made consistent with the test results by debugging the calculation model. Figure 3 Force-displacement curve calculated by inversion.

[0084] Step 2: Create a dataset

[0085] The calibrated numerical model is subjected to a large number of parameter analyses. The parameters with the largest structural dimensions that affect the mechanical properties of the energy absorber are selected. Taking the pressure relief ring as an example, the key dimensions such as the ring diameter, tube diameter, tube wall thickness and aluminum sleeve length are selected as the input parameter X; the starting load, maximum load and energy consumption that are most concerned about its mechanical properties are selected as the output parameter Y. In each calculation condition, X includes 4 structural dimension parameters and Y includes 3 mechanical performance parameters. The data set contains the calculation results of 256 conditions. Figure 4 The results of the first 30 calculation cases out of 256 calculation cases are shown.

[0086] After the data set is established, it is normalized.

[0087] In this embodiment, the data set is divided into a training set and a test set using a holdout method at a ratio of 85% and 15%, that is, the training set contains 217 groups of working condition calculation results, and the test set contains the remaining 39 groups of working condition calculation results.

[0088] Step 3: Training PSO-BP neural network

[0089] Initially define the BP neural network structure, the number of neurons in the hidden layer is 6, the learning rate is 0.01, and the number of training rounds is 1000. Initialize the PSO optimization algorithm hyperparameters, define the learning factor as 2, the population size as 5, and the number of population updates as 100.

[0090] The PSO optimization algorithm neural network weights and threshold position and speed update functions are calculated by the following formula:

[0091] ;

[0092] ;

[0093] Where: It is a particle i exist t The speed of the moment; It is a particle i exist t The location at the moment; is the inertia weight; is the learning factor; is a random number in the interval [0,1]; Particlesi The best position encountered so far and the best positions encountered so far by all particles.

[0094] Figure 5 a. Figure 5 b and Figure 5 c shows the error distribution of the trained neural network tested using the test set; Figure 5 a is the comparison chart between the predicted result and the actual value of the starting load. Figure 5 b is the comparison chart between the maximum load prediction result and the actual value. Figure 5 c is a comparison chart of energy consumption prediction results and actual values. The errors of the three output results are all less than 10%. The fitness value iteration is as follows: Figure 6 As shown, the training results are good and meet the engineering requirements.

[0095] Step 4: Create a structural design model

[0096] Use PSO iterative optimization algorithm to establish a structural design model based on target performance;

[0097] Take the specific target mechanical properties as an example: starting force 40kN, maximum force 85kN and energy consumption 54kJ. The reverse design model selects and outputs 4 different structural size combinations, as shown in the following table:

[0098] ;

[0099] Taking the design structure size combination of No. 1 as an example, two groups of mechanical tests and numerical simulations of actual products were carried out. The structural dimensions were ring diameter 450mm, tube diameter 42mm, tube wall thickness 3.5mm and aluminum sleeve length 80mm. Figure 7 shown.

[0100] The above-described embodiments are used to illustrate the present invention, not to limit the present invention, so changes in the exemplified values ​​or replacement of equivalent elements should still fall within the scope of the present invention.

[0101] From the above detailed description, it can be understood by those skilled in the art that the present invention can indeed achieve the aforementioned objectives and is in compliance with the provisions of the Patent Law.

[0102] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they are aware of the basic creative concepts. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention. The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. It should be pointed out that any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

[0103] It should be noted that the above description of the relevant process is only for example and explanation, and does not limit the scope of application of this specification. For those skilled in the art, various modifications and changes can be made to the process under the guidance of this specification. However, these modifications and changes are still within the scope of this specification.

[0104] The basic concepts have been described above. Obviously, for those of ordinary skill in the art who have read this application, the above invention disclosure is only for example and does not constitute a limitation of this application. Although not explicitly stated here, those of ordinary skill in the art may make various modifications, improvements and amendments to this application. Such modifications, improvements and amendments are suggested in this application, so such modifications, improvements and amendments still belong to the spirit and scope of the exemplary embodiments of this application.

[0105] At the same time, the present application uses specific words to describe the embodiments of the present application. For example, "one embodiment", "an embodiment", and / or "some embodiments" refer to a certain feature, structure or characteristic related to at least one embodiment of the present application. Therefore, it should be emphasized and noted that "one embodiment" or "an embodiment" or "an alternative embodiment" mentioned twice or more in different positions in this specification does not necessarily refer to the same embodiment. In addition, some features, structures or characteristics in one or more embodiments of the present application can be appropriately combined.

[0106] In addition, it will be understood by those skilled in the art that various aspects of the present application may be illustrated and described by a number of patentable categories or situations, including any new and useful process, machine, product or combination of substances, or any new and useful improvement thereof. Therefore, various aspects of the present application may be implemented entirely by hardware, entirely by software (including firmware, resident software, microcode, etc.), or by a combination of hardware and software. The above hardware or software may be referred to as a "unit", "module" or "system". In addition, various aspects of the present application may take the form of a computer program product embodied in one or more computer-readable media, wherein computer-readable program code is contained therein.

[0107] The computer program code required for the operation of each part of the present application can be written in any one or more programming languages, including object-oriented programming languages ​​such as Java, Scala, Smalltalk, Eiffel, JADE, Emerald, C++, C#, VB.NET, Python, etc., conventional procedural programming languages ​​such as C programming language, VisualBasic, Fortran2103, Perl, COBOL2102, PHP, ABAP, dynamic programming languages ​​such as Python, Ruby and Groovy, or other programming languages. The program code can be run entirely on the user's computer, or run on the user's computer as an independent software package, or run partially on the user's computer and partially on the remote computer, or run entirely on the remote computer or server. In the latter case, the remote computer can be connected to the user's computer through any network form, such as a local area network (LAN) or a wide area network (WAN), or connected to an external computer (e.g., via the Internet), or in a cloud computing environment, or used as a service such as software as a service (SaaS).

[0108] In addition, unless explicitly stated in the claims, the order of the processing elements and sequences described in this application, the use of alphanumeric characters, or the use of other names are not intended to limit the order of the processes and methods of this application. Although the above disclosure discusses some invention embodiments that are currently considered useful through various examples, it should be understood that such details are only for illustrative purposes, and the attached claims are not limited to the disclosed embodiments. On the contrary, the claims are intended to cover all modifications and equivalent combinations that are consistent with the essence and scope of the embodiments of this application. For example, although the implementation of the various components described above can be embodied in a hardware device, it can also be implemented as a pure software solution, for example, installation on an existing server or mobile device.

[0109] Similarly, it should be noted that in order to simplify the description disclosed in this application and thus help understand one or more embodiments of the invention, in the above description of the embodiments of the application, multiple features are sometimes combined into one embodiment, figure or description thereof. However, this method of the application should not be interpreted as reflecting the intention that the claimed object requires more features than those explicitly stated in each claim. On the contrary, the subject of the invention should have fewer features than the above single embodiment.

Claims

1. An intelligent design method for energy dissipators based on neural networks, characterized in that: include: S1. Numerical model calibration: Conduct mechanical tests on energy dissipators, debug and establish numerical models that match the test structure; The test used a 150t range vertical tensile testing machine. The lower end of the pressure relief ring wire rope was connected to the shackle and fixed on the steel beam anchored to the ground. The upper end of the shackle was fixed to the tensile arm of the testing machine. The tension and displacement data were recorded every 0.2s during the stretching process to obtain the force-displacement curve of the test process. An inverse analysis numerical model was established, and the finite element method was selected using the LS-DYNA calculation software to establish the calculation model. The calculation model was debugged to make the calculation results consistent with the test results. S2. Create a dataset: Based on the debugged numerical model, the parameters of the energy dissipator structure size are analyzed, the structural size parameters and mechanical performance parameters are obtained to form a data set, the data is preprocessed, and the data set is divided into a training set and a test set; When performing parameter analysis to establish a data set, the structural dimensions of the energy absorber are used as input parameters, and the mechanical performance indicators of the energy absorber are used as output parameters; when the energy absorber is a pressure reducing ring, the structural dimensions of the energy absorber are the ring diameter, tube diameter, tube wall thickness and aluminum sleeve length of the pressure reducing ring, and the mechanical performance indicators of the energy absorber are the starting load, maximum load and energy consumption of the pressure reducing ring; The data set contains calculation results for 256 working conditions; After the data set is established, it is normalized; the data set is divided into a training set and a test set according to the ratio of 85% and 15%, that is, the training set contains 217 groups of working condition calculation results, and the test set contains the remaining 39 groups of working condition calculation results; S3. Training PSO-BP neural network: Initialize the PSO algorithm hyperparameters, define the fitness function, use the training set to train the PSO-BP neural network, use the test set to test the trained PSO-BP neural network and calculate the fitness value, minimize the fitness value, and extract the optimal initial weight and threshold; Initially define the BP neural network structure, the number of neurons in the hidden layer is 6, the learning rate is 0.01, and the number of training rounds is 1000; initialize the PSO optimization algorithm hyperparameters, define the learning factor as 2, the population size as 5, and the number of population updates as 100; S4. Establish the energy dissipator structure design model: Import the trained PSO-BP neural network, and establish the intelligent design model of energy consumer based on the PSO-BP neural network and PSO inverse parameter optimization algorithm; Based on the intelligent design model of the energy absorber, the structural dimensions are designed, and mechanical tests of the actual product of the structural dimensions are carried out to obtain the experimental mechanical properties; the experimental mechanical properties are compared with the target mechanical properties to verify the effectiveness of the intelligent design model of the energy absorber; The energy absorber is a pressure reducing ring, a rod-type energy absorber, a plate-type energy absorber, a U-shaped energy absorber or a piston rod-type energy absorber; The data preprocessing is normalization processing, and the normalization processing includes minimum-maximum normalization, zero-mean normalization and decimal calibration normalization; When dividing the data set, a division method is selected according to the size of the data and the difficulty of training, and the division methods include the leave-out method, the cross-validation method, the leave-one-out method, and the bootstrap method; PSO algorithm neural network weights and threshold position and speed update function: v i (t+1)=ω×v i (t)+c1×rand1×(pbest i -x i (t))+ c2×rand2×(gbest-x i (t)); x i (t+1)=x i (t)+v i (t+1); Where: v i (t) is the velocity of particle i at time t; x i (t) is the position of particle i at time t; ω is the inertia weight; c1 and c2 are learning factors; rand1 and rand2 are random numbers in the interval [0,1]; gbest i and pbest i are the best position encountered by particle i so far and the best position encountered by all particles so far, respectively.

2. The intelligent design system of energy consumption device based on neural network is characterized by: Used to implement the intelligent design method of energy consumption device based on neural network as described in claim 1; The energy consumption device intelligent design system based on neural network includes: Numerical model calibration module, used to conduct mechanical tests on energy dissipators, debug and establish a numerical model that matches the test structure; The data set establishment module is used to perform parameter analysis of the energy dissipator structure size based on the debugged numerical model, obtain the structure size parameters and mechanical performance parameters to form a data set, perform data preprocessing, and divide the data set into a training set and a test set; PSO-BP neural network training module, used to initialize PSO algorithm hyperparameters, define fitness function, train PSO-BP neural network using training set, test trained PSO-BP neural network using test set and calculate fitness value, minimize fitness value, and extract optimal initial weight and threshold; The energy absorber structure design model building module is used to import the trained PSO-BP neural network and build the energy absorber intelligent design model based on the PSO-BP neural network and the PSO inverse parameter optimization algorithm.

3. A computer-readable storage medium, characterized in that: The storage medium stores computer instructions. When a computer reads the computer instructions, the computer executes the intelligent design method of energy consumer based on neural network as claimed in claim 1.

Images