CN116384258A - Hub-tire integrated wheel impact dynamics simulation method - Google Patents
Hub-tire integrated wheel impact dynamics simulation method Download PDFInfo
- Publication number
- CN116384258A CN116384258A CN202310615295.2A CN202310615295A CN116384258A CN 116384258 A CN116384258 A CN 116384258A CN 202310615295 A CN202310615295 A CN 202310615295A CN 116384258 A CN116384258 A CN 116384258A
- Authority
- CN
- China
- Prior art keywords
- equation
- hub
- model
- simulation
- tire
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000004088 simulation Methods 0.000 title claims abstract description 133
- 238000000034 method Methods 0.000 title claims abstract description 82
- 239000000463 material Substances 0.000 claims abstract description 129
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 57
- 230000002068 genetic effect Effects 0.000 claims abstract description 57
- 238000009863 impact test Methods 0.000 claims abstract description 44
- 238000013528 artificial neural network Methods 0.000 claims abstract description 39
- 238000004364 calculation method Methods 0.000 claims abstract description 33
- 238000010801 machine learning Methods 0.000 claims abstract description 23
- 238000012360 testing method Methods 0.000 claims description 43
- 230000006870 function Effects 0.000 claims description 42
- 238000005070 sampling Methods 0.000 claims description 27
- 230000008569 process Effects 0.000 claims description 24
- 238000007781 pre-processing Methods 0.000 claims description 18
- 239000004033 plastic Substances 0.000 claims description 17
- 238000005457 optimization Methods 0.000 claims description 15
- 238000005094 computer simulation Methods 0.000 claims description 14
- 230000014509 gene expression Effects 0.000 claims description 14
- 238000012549 training Methods 0.000 claims description 13
- 238000005728 strengthening Methods 0.000 claims description 8
- 230000035772 mutation Effects 0.000 claims description 7
- 238000012804 iterative process Methods 0.000 claims description 4
- 230000008676 import Effects 0.000 abstract 1
- 229910001234 light alloy Inorganic materials 0.000 description 11
- 238000004458 analytical method Methods 0.000 description 6
- 230000003068 static effect Effects 0.000 description 6
- 230000004913 activation Effects 0.000 description 5
- 238000011160 research Methods 0.000 description 5
- 230000001133 acceleration Effects 0.000 description 4
- 239000002131 composite material Substances 0.000 description 4
- 230000007547 defect Effects 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 238000013507 mapping Methods 0.000 description 4
- 238000009864 tensile test Methods 0.000 description 4
- 229910000831 Steel Inorganic materials 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 3
- 238000013461 design Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 230000010354 integration Effects 0.000 description 3
- 239000011159 matrix material Substances 0.000 description 3
- 230000035945 sensitivity Effects 0.000 description 3
- 239000010959 steel Substances 0.000 description 3
- 238000012356 Product development Methods 0.000 description 2
- 239000011324 bead Substances 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 150000001875 compounds Chemical class 0.000 description 2
- 230000001010 compromised effect Effects 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 238000006073 displacement reaction Methods 0.000 description 2
- 238000007620 mathematical function Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 201000004569 Blindness Diseases 0.000 description 1
- 238000012897 Levenberg–Marquardt algorithm Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 239000003795 chemical substances by application Substances 0.000 description 1
- 230000002860 competitive effect Effects 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 239000013013 elastic material Substances 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000013213 extrapolation Methods 0.000 description 1
- 239000000945 filler Substances 0.000 description 1
- 230000005484 gravity Effects 0.000 description 1
- 230000000977 initiatory effect Effects 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000007670 refining Methods 0.000 description 1
- 230000010076 replication Effects 0.000 description 1
- 238000012916 structural analysis Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/11—Complex mathematical operations for solving equations, e.g. nonlinear equations, general mathematical optimization problems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/15—Vehicle, aircraft or watercraft design
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/17—Mechanical parametric or variational design
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/23—Design optimisation, verification or simulation using finite element methods [FEM] or finite difference methods [FDM]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/048—Activation functions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/086—Learning methods using evolutionary algorithms, e.g. genetic algorithms or genetic programming
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/06—Multi-objective optimisation, e.g. Pareto optimisation using simulated annealing [SA], ant colony algorithms or genetic algorithms [GA]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/14—Force analysis or force optimisation, e.g. static or dynamic forces
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Geometry (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Computer Hardware Design (AREA)
- Data Mining & Analysis (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Automation & Control Theory (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Operations Research (AREA)
- Databases & Information Systems (AREA)
- Aviation & Aerospace Engineering (AREA)
- Medical Informatics (AREA)
- Algebra (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Physiology (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a hub-tire integrated wheel impact dynamics simulation method, and belongs to the field of machine learning and cloud computing. According to the invention, a basic equation obtained by combining field knowledge and priori knowledge is used as an initial solution of the optimizing part, and an equation learning neural network is introduced to automatically explore a solution space of a constitutive equation of a material to generate more fitting terms which are not involved in the basic equation, so that the constitutive equations of the two materials are added into an initial population of a genetic algorithm, the optimizing efficiency of the optimal solution of a constitutive model of a hub material can be ensured, and meanwhile, the problem of sinking into a local optimal solution is avoided. According to the invention, the hub-tire integrated wheel impact dynamics simulation is further arranged on the cloud service platform, so that a user is allowed to call and integrate resources on the cloud platform. The user only needs to connect the cloud service platform, uploads or online imports the hub model and simulation parameters which need to be subjected to impact test simulation, and can carry out simulation calculation according to the steps, so that the simulation difficulty is greatly reduced.
Description
Technical Field
The invention belongs to the field of machine learning and cloud computing, and particularly relates to a method for realizing hub-tire integrated simulation through a neural network.
Background
The hub is used as a key bearing structure in the running process of the automobile, and the stability, safety and economy of the automobile are directly influenced by the light weight degree and the mechanical property of the hub. In order to meet the market demand and the service performance of the hub, the cost is saved, the product competitiveness is improved, and the high-strength lightweight hub becomes the development target of the industry. The impact resistance is one of important indexes for measuring the strength of the hub. In order to simulate the situation that the Wheels transversely strike the curb or pass through obstacles such as broken stones in the running process of an automobile, the capability of the Wheels against lateral impact and longitudinal impact is considered, the Wheels are verified and controlled in quality, different national wheel impact test standards are given out, the national common standards are GB/T15704-2012, the Road vehicle light alloy wheel impact test method, QC/T991-2015, the passenger vehicle light alloy wheel 90-degree impact test method, the national common standards are ISO 7141:2022 Road Vehicles-Light alloy Wheels-Lateral impact test, SAE J175_202107 Wheels-Lateral Impact Test Procedure-Road Vehicles and the like, and the test principle is that a special-shaped impact hammer is adopted to impact the Wheels at different positions in different angles to test the impact resistance of the Wheels.
The impact test of the wheel needs to be carried out on a specific impact tester, and the test object must be a representative finished wheel which can be used for a vehicle through a complete process route, and if the impact performance of the wheel is not qualified, a great deal of die opening cost and processing cost are lost. Therefore, in order to reduce the test cost and shorten the development period, it is necessary to perform dynamic simulation on the impact test of the wheel in the early stage of product development. The accurate and efficient wheel impact dynamics simulation can quickly feed back the mechanical properties of the hub under the condition of not performing a test, and according to the simulation output result, a product development engineer can quickly perform structural analysis and optimization of the hub until the impact properties of the hub meet the requirements.
The traditional wheel impact simulation method generally simplifies the dynamic problem into the statics problem, converts the impact force of the impact hammer into static load to act on the hub, frequently ignores the tire model, compensates the energy absorbed by the rubber tire by subtracting the energy of the impact hammer part, and simplifies the simulation, so that the simulation efficiency is improved, but the problem of overlarge error is caused. With the deep research, more and more researches take a tire model into consideration, a wheel impact dynamics simulation model of hub-tire coupling is established, and dynamics calculation is performed by adopting a display analysis method, so that the actual simulation accuracy and simulation precision of simulation are improved. The existing wheel impact dynamics simulation method solves the problem of larger simulation error of the traditional method, has the defects of narrow application range, complex operation, long calculation time and high cost, can not meet the convenient and efficient use requirements of hub developers, needs to self-maintain software and hardware facilities, and has poor user experience when needing to be equipped with tires of different specifications and models and analyzing wheel impact tests of different angles. Cloud computing is scalable and convenient network access, and computer resources are provided according to the computer environment scale required by a user by using a virtualization technology, so that the sharing of the computer resources is realized. However, how to transfer the traditional wheel impact dynamics simulation to the cloud end brings new technical problems to be solved.
Disclosure of Invention
The invention aims to solve the defects of larger error and complicated operation of wheel impact dynamics simulation in the prior art and provides a wheel impact dynamics simulation method for hub-tire integration.
The specific technical scheme adopted by the invention is as follows:
a hub-tire integrated wheel impact dynamics simulation method, comprising:
s1, sampling material mechanics test data of a hub to obtain a sample data set;
s2, training the equation learning neural network for a plurality of times by using the sample data set, and fitting a constitutive equation of the material by the equation learning neural network to obtain a plurality of machine learning fitting equations; simultaneously, fitting an existing material constitutive equation obtained according to field knowledge and priori knowledge by using the sample data set to obtain a basic equation;
s3, each basic equation, the machine learning fitting equation and the randomly generated random expression are respectively encoded into a binary tree form consisting of a terminator and an operator; taking all binary trees obtained by coding as an initial population of a genetic algorithm, performing initial cross operation, and updating individuals obtained after the initial cross operation into the original initial population;
S4, carrying out multi-objective optimization on population iteration through a genetic algorithm based on the updated initial population and an fitness function constructed by combining equation fitting degree and equation complexity, and periodically adding a basic equation fitted by the sample data set into the latest offspring population to guide the evolution direction of the genetic algorithm in the iteration process, and outputting an optimal solution after the genetic algorithm is iterated to a final condition to serve as a hub material constitutive optimal equation;
s5, the hub model and the tire model of the target wheel, the wheel impact test standard and simulation parameters including the optimal equation of the hub material structure are led into a dynamic simulation system, finite element simulation is carried out by calling a solver, and a simulation result of the target wheel under impact is output.
Preferably, the dynamics simulation system is mounted on a cloud service platform, and a standard library, a model library and a software resource library are built in the cloud service platform;
different wheel impact test standards for users to select are built in the standard library;
the model library comprises one or more of a hub model sub-library, a tire model sub-library, a wheel assembly model sub-library, a bench model sub-library and a material model sub-library, and the model in each sub-library is used for providing a direct calling function when a user does not upload a corresponding model;
And one or more dynamic simulation software for user call is built in the software resource library.
Preferably, the cloud service platform is further provided with a preprocessing module, wherein the preprocessing module comprises a user data sub-module and a model generation sub-module; the user data sub-module is used for uploading or selecting a hub model and a tire model of a target wheel, a wheel impact test standard and simulation parameters by a user, wherein a hub material constitutive optimal equation in the simulation parameters is fitted in an online or offline mode; the model generation sub-module is used for generating a calculation file required by the dynamic simulation system according to the data in the user data sub-module.
Preferably, the cloud service platform is further provided with a computing module, the computing module comprises a preprocessing sub-module and a solving sub-module, the preprocessing sub-module is used for allocating corresponding cloud computing resources according to the simulation calculation amount of the dynamic simulation system, and the solving sub-module is used for carrying out finite element simulation through a solver according to the cloud computing resources allocated by the preprocessing sub-module.
Preferably, a result output module is further provided in the cloud service platform, and is configured to store relevant data of the dynamic simulation in a cloud storage form in the cloud, and output a simulation result according to an output form set by a user.
Preferably, the existing material constitutive equations used to fit to form the base equation include a Johnson-Cook yield model, a Cowper-Symonds yield model, a Swift hardening model, and a Voce hardening model determined from knowledge in the field, and a plurality of custom material constitutive equations determined from prior knowledge.
Preferably, the sample data set is divided into a first sample data set and a second sample data set; the basic equation serving as the initial population and the machine learning fitting equation are trained or fitted by the first sample data set, and the basic equation added with the offspring population in the iteration process is fitted by the second sample data set;
the sampling range of the first sample data set in the material mechanical test data is near a yield platform or 0.2% equivalent plastic strain, and a necking stage and a yield strengthening stage, wherein the sampling interval of the yield strengthening stage is larger than that of other stages;
the sampling range of the second sample data set in the material mechanics test data is near the yield platform or 0.2% equivalent plastic strain, and a necking stage; and the sampling intervals of the second sample data set are each smaller than the sampling intervals of the first sample data set.
Preferably, before the iterative process is executed, the genetic algorithm needs to preset initialization parameters, including a terminal symbol set, a function symbol set, a genetic operator, the maximum evolution times, the population number and the coding form; wherein the terminals in the terminal set include all of the independent variables and real constants involved in the base equation and the machine learning fit equation; the function symbol set comprises all basic equations and operators involved in the machine learning fit equation; the coding form is a binary tree structure; and in the genetic operators, the crossover operator only carries out crossover operation on the structure containing the independent variable in the binary tree, and the mutation operator is used for carrying out local search on the real constant of the new individual and simultaneously changing the operator.
Preferably, the fitness function includes an average absolute relative error, a root mean square error, a fitness coefficient and an equation complexity, wherein the fitness coefficient is a difference value obtained by subtracting the decision coefficient from 1, and the equation complexity is a sum of respective complexities of all operators in the equation; the genetic algorithm comprehensively performs multi-objective optimization in a decision space based on four fitness functions.
Preferably, the optimal solution of the genetic algorithm is selected from the one compromise solution with the smallest fitting coefficient in the pareto non-dominant solution set.
Compared with the prior art, the invention has the following beneficial effects:
1) The invention takes a basic equation obtained by combining field knowledge and priori knowledge as an initial solution of optimizing, introduces an equation learning neural network to autonomously explore a solution space of a constitutive equation of a material, and generates more fitting terms which are not related to the basic equation. By setting the two types of constitutive equations of the materials to construct an initial population of a genetic algorithm, the optimization efficiency of an optimal solution of a constitutive model of the hub material can be ensured, and meanwhile, the problem of sinking into a local optimal solution is avoided.
2) According to the invention, the hub-tire integrated wheel impact dynamics simulation is further arranged on the cloud service platform, and the user is allowed to dynamically integrate distributed data on resources on the cloud platform through cloud computing. The user does not need to install a complex dynamic simulation software module, does not need to configure the running environment, a server, a memory and other hardware facilities by himself, and can perform simulation calculation according to steps by importing a hub model and related simulation parameters which need to perform impact test simulation only by connecting a cloud service platform, so that the calculation resources can be allocated according to needs and efficiently utilized.
3) According to the invention, users can rapidly develop, analyze and research specific problems and requirements, high-efficiency and accurate simulation calculation of the wheel impact dynamics is realized, the defects of complex parameter setting, large workload, time consumption in calculation and the like of the existing wheel impact dynamics simulation method are overcome, and the wheel impact dynamics simulation analysis can be developed more efficiently by directly calling related resources on a cloud platform for first-line engineers such as hub manufacturers and the like. And the invention can realize the serial simulation calculation of the wheel impact dynamics with different models and different requirements by virtue of a complete resource library and model library on the cloud service platform and strong calculation force, thereby helping enterprises to formulate corresponding wheel impact performance evaluation standards.
Drawings
FIG. 1 is a schematic illustration of the steps of a method for simulating the impact dynamics of a hub-tire integrated wheel;
FIG. 2 is a fitting flow chart of the constitutive optimal equation for hub materials;
fig. 3 is a schematic diagram of a hub-tire integrated wheel impact dynamics simulation system based on a cloud service platform.
Detailed Description
In order that the above objects, features and advantages of the invention will be readily understood, a more particular description of the invention will be rendered by reference to the appended drawings. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention. The present invention may be embodied in many other forms than described herein and similarly modified by those skilled in the art without departing from the spirit of the invention, whereby the invention is not limited to the specific embodiments disclosed below. The technical features of the embodiments of the invention can be combined correspondingly on the premise of no mutual conflict.
In the description of the present invention, it should be understood that the terms "first" and "second" are used solely for the purpose of distinguishing between the descriptions and not necessarily for the purpose of indicating or implying a relative importance or implicitly indicating the number of features indicated. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include at least one such feature.
The invention provides a method for simulating the impact dynamics of a wheel hub-tire integrated wheel, which can simulate the impact dynamics of the wheel hub-tire integrated wheel. The wheel hub-tire integrated wheel refers to a wheel formed by integrally assembling a wheel hub and a tire, namely the wheel hub and the tire in an assembled state are considered at the same time when simulation is carried out in the invention, so that the impact form of the wheel hub-tire integrated wheel is closer to the actual running condition of the tire. The specific implementation manner of the above-mentioned hub-tire integrated wheel impact dynamics simulation method is described in detail below.
As shown in FIG. 1, in a preferred implementation manner of the present invention, the above-mentioned method for simulating the impact dynamics of a hub-tire integrated wheel comprises the following steps S1-S5. The following describes the specific implementation procedure.
S1, sampling material mechanics test data of the hub to obtain a sample data set.
It should be noted that the data of the wheel hub material mechanical test can be obtained according to the material mechanical test scheme in the conventional impact test method. In general, the adopted material mechanical test is a uniaxial tensile test under different tensile rates, and the test raw data is subjected to a certain conversion formula to obtain corresponding data 
(the three parameters respectively represent flow stress, equivalent plastic strain and equivalent plastic strain rate), and the conversion formula is common knowledge in the field and will not be described again.
The sample data set obtained by sampling is used for fitting a material constitutive equation of the hub. When the material mechanics test data are obtained, a corresponding sampling mode can be selected based on the data characteristics, and a non-uniform sampling method is preferably adopted in the embodiment of the invention.
S2, training an Equation learning (EQL) neural network for a plurality of times by using the sample data set, and obtaining a plurality of machine learning fit equations by using the Equation learning neural network to fit a constitutive Equation of a material; and simultaneously, fitting the existing material constitutive equation obtained according to the field knowledge and the priori knowledge by using the sample data set to obtain a basic equation.
It should be noted that the constitutive equations of the material in the present invention include two types, in which the machine learning fit equation is obtained by fitting an equation learning neural network, and the basic equation is obtained by fitting an existing constitutive equation of the material. However, the machine learning fitting equation and the basic equation are equations after regression fitting, and parameters to be fitted have specific fitting values. In an embodiment of the present invention, existing material constitutive equations used to fit to form the base equation include some classical material constitutive equations in the field determined from knowledge of the field, including but not limited to Johnson-Cook yield model, cowper-Symonds yield model, swift hardening model, and Voce hardening model, and may additionally include a plurality of custom material constitutive equations determined from prior knowledge. The custom material constitutive equation is a possible custom material constitutive equation deduced from the prior knowledge on the trend of the test curve, and the parameters of the equations are unknown but the expression form is known. The prior knowledge can be related literature, previous experience or an equation which is self-fitted and optimized according to experimental data, namely the self-defined material constitutive equation can be an equation which is proposed in the related literature, an equation which is corrected according to experience or an equation which is self-constructed. The form of the specific self-defined material constitutive equation is not limited, and theoretically, the richer and more accurate the self-defined material constitutive equation is, the more initial information guidance can be provided for the follow-up genetic algorithm optimization.
The wheel impact test is a dynamic mechanical problem, relates to strain strengthening and strain rate hardening effects of materials, is often accompanied by larger plastic deformation and fracture failure behaviors, and in order to better describe the stress-strain relationship of the wheel hub materials, the wheel impact dynamics simulation accurately reflects the local deformation and the occurrence of cracks or fractures of the wheel hub in the actual impact test, and is very important to accurately describe the material constitutive model of the wheel hub (if the fracture condition of failure needs to be simulated, the failure model needs to be added). The reason for setting the constitutive equations of the two materials in the invention is to ensure the optimizing efficiency of the optimal solution of the constitutive model of the hub material and avoid sinking into the local optimal solution at the same time. Because the basic equation is obtained according to the field knowledge and the priori knowledge, the fitting performance of the basic equation is verified, and the constitutive equation of the materials can be used as an initial solution of optimizing, so that the optimizing efficiency is ensured. However, the basic equations are completely adopted, the possibility of variation in the process of genetic algorithm optimization is insufficient, and fitting terms which are not found in some basic equations are difficult to introduce, so that the invention further introduces the equation learning EQL neural network to autonomously explore the solution space of the fitting equations, and more fitting terms which are not related to the basic equations are generated.
The equation learning neural network adopted in the invention is a neural network which can be used for fitting equations, and the implementation mode belongs to the prior art, and can be seen in the prior art document: martius G, lambert C H. Extrapolation and learning equations [ J ]. 2016 ], EQL neural networks have been widely reported in other literature, except for this document. This will be briefly described below to facilitate an understanding of the principles of the neural network.
The EQL neural network is essentially a full-connection-layer neural network, which consists of an input layer, two hidden layers and an output layer, and the forward reasoning process can be expressed as follows:
in the method, in the process of the invention,
weight matrix of 1 st and 2 nd full connection layer respectively, +.>
Is a weight matrix from layer 2 to output layer, < >>
Is input data, namely equivalent plastic strain and equivalent plastic strain rate (+)>
) In matrix form,/->
Is the output of the neural network, i.e., a material constitutive equation in the form of a specific function. The EQL neural network differs from the traditional fully connected neural network in that the activation function +.>
. Specifically, common to traditional machine learning +.>
Equal function is different, activation function of EQL neural network +.>
Alternative mathematical functions and operator custom activation functions are employed to obtain the final specific function output. Where alternative mathematical functions and operators include, but are not limited to, units, squares, cubic, exponents, logarithms, multiplications, and the like. Allowing replication of the activation function in each layer, i.e. in +. >
The same activation function may be used by multiple components of the system in order to reduce the sensitivity of the system to random initialization, creating a smoother optimization environment.
In addition, the EQL neural network is introduced into the loss function during training
Regularization terms ensure potential sparsity of the neural network, fitting the neural network to a simple solution as much as possible.
In the method, in the process of the invention,
is a weight of->
Is a constant value, and is set to 0.01.
The mean square error loss function that ultimately introduces the regularization term is represented as follows:
Random initialization due to EQL neural network system pairThe method has certain sensitivity, different results are obtained in each training, and local minima are easy to fall into, so that N times of training are carried out to obtain N machine learning fit equations of EQL fitting
The machine learning fitting equations can be used as an initial population together with other basic equations to be optimized through an evolutionary algorithm, so that the fine correction effect of the equations is achieved, test data can be better fitted, material characteristics can be accurately described, and simulation accuracy is improved.
S3, each basic equation, the machine learning fitting equation and the randomly generated random expression are respectively encoded into a binary tree form consisting of a terminator and an operator; and taking all binary trees obtained by coding as an initial population of a genetic algorithm, performing initial crossover operation, and updating individuals obtained after the initial crossover operation into the original initial population.
The terminal includes an argument related to a constitutive equation of the material (e.g 
) And real constants, operators are the operational coincidences involved in the equation (e.g. +.>
) The terminators and operators combine to obtain an expression of the material constitutive equation. The specific way of encoding the equations into a binary tree form belongs to the prior art and will not be described in detail.
It should be noted that the genetic algorithm adopted in the invention can be selected according to actual situations, and the NSGA-III genetic algorithm is preferably adopted, because the NSGA-III genetic algorithm can be better combined with the EQL neural network to carry out composite regression so as to fit the constitutive equation of the material.
It should be noted that, the random expression in the initial population before the initial crossover operation is performed may be generated by genetic algorithm recursively and randomly through an operator. Compared with the traditional genetic algorithm, the invention is characterized in that an initial crossover operation is specially carried out before the genetic algorithm is executed, population individuals which are obtained by the initial crossover operation and are different from the original equation are added into the original initial population, and the population individuals are used as a new initial population which participates in the iteration of the genetic algorithm together with the original basic equation, the machine learning fit equation and the binary tree of the random expression. The purpose of this is that because these equations in the original initial population are all fitted, directly acting as the initial population easily results in the genetic algorithm being directly trapped in the locally optimal solution, and the globally optimal solution cannot be found.
S4, carrying out multi-objective optimization on population iteration through a genetic algorithm based on the updated initial population and an fitness function constructed by combining equation fitting degree and equation complexity, and periodically adding a basic equation fitted through the sample data set into the latest offspring population to guide the evolution direction of the genetic algorithm in the iteration process, and outputting an optimal solution after the genetic algorithm is iterated to a final condition to serve as a hub material constitutive optimal equation.
It should be noted that, before the iterative process is executed, the genetic algorithm needs to preset initialization parameters, and specific parameters can be processed according to relevant parameter settings of the genetic algorithm. In the embodiment of the invention, since the optimizing object of the genetic algorithm is a material constitutive equation, the initializing parameters which need to be set comprise a terminal symbol set and a function symbol set, a genetic operator, a maximum evolution number, a population number and a coding form. Wherein the terminals in the terminal set comprise all the independent variables and real constants involved in the basic equation and the machine learning fit equation; the function symbol set comprises all basic equations and operators related in the machine learning fitting equation; the coding form is a binary tree structure; in the genetic operators, the crossover operator only carries out crossover operation on the structure containing the independent variable in the binary tree, and the mutation operator is used for carrying out local search on the real constant of a new individual and simultaneously changing the operator; the maximum evolution times are used for controlling the maximum iteration times of the algorithm, so that the situation that dead loops are involved when convergence is impossible is avoided; the population number should be greater than the original initial population number, so that a certain space is reserved for the initial cross operation, and population diversity is enriched.
It should be noted that the specific function form and type of the fitness function may be selected according to actual needs. In embodiments of the present invention, the fitness function may include an average absolute relative error, a root mean square error, a fitness coefficient, and an equation complexity, where the fitness coefficient is 1 minus the decision coefficient R 2 The obtained difference value, the equation complexity is the sum of the respective complexity of all operators in the equation, and different operators can set corresponding complexity values according to the actual operation complexity; the genetic algorithm comprehensively performs multi-objective optimization in a decision space based on four fitness functions. Under the multi-objective optimization framework, the optimal solution of the final genetic algorithm selects one compromised solution with the smallest fitting coefficient in the pareto non-dominated solution set, and the solving process of the whole compromised solution can be seen in fig. 2.
In addition, it should be noted that the sample data set obtained by sampling in the invention is used for fitting the material constitutive equation of the hub, and in theory, the sample data set only needs to be capable of accurately fitting the material constitutive equation, and the specific sampling interval and the sample number of the sample data set can not be limited. However, since the sample data sets are required to be fitted in S2 and S4, in an embodiment of the present invention, the sampled sample data set may be divided into two different data sets, which are respectively denoted as a first sample data set and a second sample data set, in consideration of the difference in functional requirements of the fitting of constitutive equations of different materials on the sample data. The first sample data set is sampled in the material mechanics test data set above in a range around the yield plateau or 0.2% equivalent plastic strain, and the necking and yield strengthening phases, with the yield strengthening phases being sampled at a greater interval than the other phases. And the second sample data set has a sampling range in the material mechanics test data set above that is near the yield plateau or 0.2% equivalent plastic strain, and a necking stage; and the sampling intervals of the second sample data set are smaller than the sampling intervals of the first sample data set. Based on the first sample data set and the second sample data set, the basic equation serving as the initial population and the machine learning fitting equation are trained or fitted by the first sample data set, and the basic equation added into the child population in the iteration process is fitted by the second sample data set.
S5, the hub model and the tire model of the target wheel, the wheel impact test standard and simulation parameters including the hub material constitutive optimal equation are imported into a dynamics simulation system, finite element simulation is carried out by calling a solver, and a simulation result of the target wheel under impact is output.
It should be noted that the simulation parameters described above should include simulation parameters of other necessary settings in the dynamics simulation system in addition to the hub material constitutive optimal equation. In an embodiment of the present invention, simulation parameters include, but are not limited to, material data, wheel impact dynamics model structural parameters, load and constraints, grid data, solvers, pre-output data settings, and the like.
In addition, the solver used in the present invention may be selected based on the actual dynamics simulation system, including but not limited to a finite element software solver such as ANSYS, ABAQUS, LS-DYNA.
In the dynamic simulation process, the hub model and the tire model in the target wheel are integrally assembled and then simulated, so that the simulation result is consistent with the actual working condition of the wheel. The hub model and the tire model can be designed in advance in the three-dimensional modeling software and then imported, and if the existing three-dimensional model file exists, the existing three-dimensional model file can be directly called. In addition, the wheel impact test standard comprises a plurality of wheel impact test standards in different countries and regions, and one of the wheel impact test standards is selected according to actual needs.
Therefore, according to the hub-tire integrated wheel impact dynamics simulation method described in S1-S5, a composite regression fitting material constitutive equation is carried out by combining a genetic algorithm with an EQL neural network, so that a user can rapidly develop, analyze and research specific problems and requirements, and efficient and accurate serial simulation calculation of the wheel impact dynamics is realized. In order to better demonstrate the specific implementation form of the present invention, the specific implementation cases of the hub-tire integrated wheel impact dynamics simulation method described in S1 to S5 are demonstrated by the following two embodiments.
Example 1
In the embodiment, a NSGA-III genetic algorithm is combined with an EQL neural network to perform composite regression fitting on a hub material constitutive optimal equation, and hub-tire integrated wheel impact dynamics simulation is performed based on the hub material constitutive optimal equation. The hub-tire integrated wheel impact dynamics simulation method comprises the following specific steps:
step 1, initial sample collection:
and carrying out a material mechanics test on the hub in advance, and collecting material mechanics test data of the hub. A non-uniform sampling method based on data characteristics is adopted: dense sampling with a sampling interval of 10 is carried out at the yield platform or the vicinity of 0.2% equivalent plastic strain of the mechanical test data of the material and in the necking stage, so as to better capture the yield characteristic and the strength limit of the material, sparse sampling with a sampling interval of 100 is carried out in the yield strengthening stage, and a first sample data set with reasonable capacity is obtained by converting the sampled original test data; meanwhile, local sample collection with a sampling interval of 5 is carried out at the yield platform or the vicinity of 0.2% plastic strain of the material mechanical test data and in the necking stage, and the local sample collection is similarly converted to be used as a second sample data set. The material mechanical test in the embodiment is a uniaxial tensile test of the hub at different tensile rates, and the test raw data is obtained through a conversion formula 
Data.
Step 2, algorithm initialization setting:
terminal and function symbol sets, genetic operators, maximum evolution times, population numbers, coding forms, etc. are set. The initial evolution times are set to be 0, the maximum evolution times are set to be M, and the population number is set to be N. The terminator includes the independent variable related to the constitutive equation of the material
) And real constants, the function set includes some operators (++>
) The terminators and operators are combined to obtain an expression. The division operation has a corresponding protection mechanism, and a tiny amount is added in a denominator to ensure the operation meaning. The encoding form is a binary tree structure. The crossover operator only operates the structure containing the independent variable in the binary tree, the mutation operator firstly carries out local search on the constant value of the new individual, and secondly changes the operator, thus enriching the diversity. The crossover probability was 0.8 and the mutation probability was 0.05.
Step 3, generating an initial population:
in this embodiment, the multiple regression method refers to a method of combining numerical regression with symbolic regression. Specifically, numerical regression is performed on equations of some models existing in the field and possible custom equations deduced from test curve trends according to priori knowledge to obtain a material constitutive equation with specific expression with known parameters, an equation obtained based on equation learning (EQL) neural network training is added, the individuals and the expression generated by the genetic algorithm through operator recursion random are combined to form an initial population, the direction of genetic algorithm symbol regression fitting is guided through the equation obtained by the numerical regression and the equation obtained by the neural network training, multi-objective optimization in the symbol regression process is realized by using NSGA-III algorithm to obtain a pareto solution set of the material constitutive equation, and finally a decision coefficient is selected 
The compromise solution closest to 1 is used as a final solution, and the effect of carrying out fine correction on the existing material constitutive equation so as to better fit test data and predict material performance can be achieved.
In this example, common material flow models and dynamic mechanical constitutive models include the Johnson-Cook yield model, the Cowper-Symonds yield model, the Swift hardening model, and the Voce hardening model. The impact test of the wheel is carried out at room temperature, and the temperature rise in the impact process is very small, so that the influence of the temperature on the material performance can be ignored. Ignoring the temperature term, the material constitutive equation of the four models described above and some 4 custom material constitutive equations inferred from prior knowledge (the last four of the eight equations below) are shown below:
in the method, in the process of the invention,
the flow stress is represented, and the flow stress of different models corresponding to different subscripts is output of an equation to be fitted;
Respectively representing equivalent plastic strain and equivalent plastic strain rate, which are input of an equation to be fitted;
Separate tableShowing the equivalent plastic strain corresponding to the static yield strength and the static yield strength point, wherein the two physical quantities can be directly obtained through material mechanical test data; / >
The strain rate is a self-defined value;
And the parameters to be fitted are the constitutive equation of the material.
In this embodiment, the four classical material constitutive equations defined above and the four custom material constitutive equations are used as the basis equations, the first sample data set is used to perform numerical regression fit on the basis equations by using a least squares method, and specifically, an Lsqcurvefit function and a Levenberg-marquardt algorithm in Matlab are used to obtain eight material constitutive equations with known parameters
. In addition, since the EQL neural network system has a certain sensitivity to random initialization, different results are obtained in each training and is easy to fall into local minima, in this embodiment, the EQL neural network is trained 20 times on the first sample data set to obtain 20 EQL fitting equations->
And the method is optimized together with other initial populations through an evolutionary algorithm, so that the effect of refining and correcting the equation is achieved, and the equation is better fitted with test data. Therefore will->
、
Generating a series of random expressions ++by recursive random genetic operators>
All are encoded into binary tree form to form initial population, the number of initial population is less than preset initial population number N so as to make a certain empty space And subsequent initial cross operation is performed between the two steps, so that the initial population diversity is enriched. The equations obtained through numerical regression and neural network training can guide the direction of the symbolic regression fitting of the genetic algorithm, and avoid blindness of the genetic algorithm in the symbolic regression process. When the traditional genetic algorithm is fitted, the initial population only has randomly generated expressions, and the defects of difficult convergence and blind evolution exist.
Step 4, initial crossover operation:
after the initial population is generated, an initial crossing operation is needed to be performed in advance before iteration of the genetic algorithm to generate new population individuals, and the purpose of crossing in advance is to prevent the excessive competitive advantage of individuals generated by numerical regression from directly eliminating random expressions generated randomly by the genetic algorithm, so that the diversity of the population is ensured through the advanced crossing operation. Therefore, in this embodiment, the parent population before the initial crossover needs to be copied to the next generation, and the parent population and the individuals generated after the crossover form a child population together, and this child population replaces the original initial population, and is used as the actual initial population in the iterative process of the genetic algorithm, and the initial population number is finally N.
Step 5, calculating individual fitness:
a fitness function is determined and fitness of each individual in the population is calculated. Using average absolute relative error 
) Root mean square error ()>
) Determining coefficients ()>
) Complexity of equation (+)>
) The individual fitness is comprehensively evaluated, and the purpose is to obtain a material constitutive equation with the form as simple as possible and good fitness. The equation complexity formula is obtained by comprehensively considering the independent variable, constant and the occurrence times of different operators. Will->
The complexity of the four operators is weighted three times as high as that of other operators, and the exponential complexity of the equation is further limited. The functions are as follows:
in the method, in the process of the invention,
for fitting the predicted value of the resulting constitutive equation of the material,/->
For the sample data values collected for the actual test,
for the capacity of the first sample dataset, +.>
Representing the number of all nodes of the equation encoded binary tree in the longest path from the root node to the leaf node, the>
Representing +.>
Operator node number->
The highest power of the exponential term contained in the equation is represented.
in the method, in the process of the invention,
for decision vector, ++>
The decision space is the population formed by each material constitutive equation individual; / >
The space where the vector is located is the target space. The goal of the optimization is to obtain a minimum target space.
Step 6, selection, crossover and mutation:
and when the genetic algorithm iterates, merging the parent population and the offspring population, carrying out rapid non-dominant sorting and sorting based on reference points on the merged population, selecting N individuals from the sorted population, carrying out crossover and mutation operation on the individuals, adding all obtained new individuals into the offspring population, and recalculating the fitness.
Step 7, carrying out offspring population capacity expansion once every 50 times of evolution:
and carrying out numerical regression fitting on the second sample data set by adopting a basic equation, adding the second sample data set into the offspring population, and further guiding the direction of genetic algorithm symbolic regression fitting. The purpose is to make the final material constitutive equation better fit locally to the material yield characteristics and strength limits.
Step 8, termination judgment:
judging whether the maximum evolution times are reached or other termination conditions are met, if yes, ending the algorithm, outputting a Pareto non-dominant solution set, otherwise, returning to the step 3, and continuing to execute the algorithm.
Step 9, optimal solution selection
Selection on pareto non-dominant solution set
The smallest solution is taken as a compromise solution of multi-objective composite regression fitting to obtain a final hub material constitutive optimal equation ++ >
。
Step 9, dynamics simulation
The method comprises the steps of importing simulation parameters including a hub model and a tire model of a target wheel to be simulated, a wheel impact test standard and a self-constructed optimal equation of the hub material into a dynamics simulation system, carrying out finite element simulation by calling a solver, and outputting a simulation result of the target wheel under impact.
In this embodiment, the wheel impact test criteria include, but are not limited to, GB/T15704-2012 "Road vehicle light alloy wheel impact test method", QC/T991-2015 "passenger vehicle light alloy wheel 90 degree impact test method", ISO 71412022 Road Vehicles-Light alloy Wheels-Lateral impact test, SAE J175_202107 Wheels-Lateral Impact Test Procedure-Road Vehicles, etc., and the user can select the corresponding reference criteria according to his/her own wheel impact simulation requirements. The dynamic simulation system can be realized by adopting finite element software with strong dynamic simulation functions such as ANSYS, ABAQUS, LS-DYNA of a local end, and the corresponding solver comprises but is not limited to a ANSYS, ABAQUS, LS-DYNA and other finite element software solver. The simulation parameters to be input include material data of the tire and the hub, structural parameters of a wheel impact dynamics model, load and constraint conditions, grid data, solvers, pre-output data settings and the like.
Therefore, when the wheel impact dynamics simulation of the hub-tire integrated wheel is performed, a user needs to input a large amount of data and set a large amount of parameters, such as simulation parameters of tire specification types, impact angle parameters and the like, a three-dimensional model of the hub and the tire, a material model of the hub and the tire need to be manually set or constructed on a local computer by self, and a process of fitting the optimal equation of the hub material by adopting the NSGA-III genetic algorithm and combining with the EQL neural network is also required to use a large amount of calculation resources, so that a large amount of time and effort are required to be consumed. Moreover, it is particularly noted that users with wheel impact dynamics simulation requirements are mostly hub manufacturers and hub design engineers who generally cannot directly obtain the material data of the tire, and thus it is often difficult to complete the wheel impact dynamics simulation of the wheel-tire integration of the present invention when this part of the data is missing. Therefore, in another embodiment of the present invention, based on the hub-tire integrated wheel impact dynamics simulation method in embodiment 1, cloud end is further performed, and a corresponding dynamics simulation system is built on a cloud platform, so as to implement a hub-tire integrated wheel impact dynamics simulation method that can be used conveniently and quickly and has low cost. When the wheel impact dynamic simulation system is in actual use, a user only needs to call corresponding modules according to actual demands, the wheel impact dynamic simulation system is convenient and fast to self-define such as wheel assembly and impact angle parameters of different specifications, the test working condition is accurately simulated, the user can rapidly develop, analyze and research specific problems and demands, and the wheel impact dynamic simulation system is capable of realizing efficient and accurate simulation calculation of wheel impact dynamics of wheel hub-tire integration. An implementation form of the above-described cloud platform-based hub-tire integrated wheel impact dynamics simulation method is shown below by example 2.
Example 2
In this embodiment, the hub-tire integrated wheel impact dynamics simulation system is built on a cloud service platform. As shown in fig. 3, the modules on the cloud service platform include a preprocessing module, a calculation module, a result output module, a standard library, a model library and a software resource library. The preprocessing module comprises a user data sub-module and a model generation sub-module; the computing module comprises a preprocessing sub-module and a solving sub-module; the result output module comprises a storage sub-module and an analysis sub-module. The user carries out the step flow of the wheel hub-tire integrated wheel impact dynamics simulation through the cloud service platform as follows:
1): the user inputs personal or enterprise account information to log in the cloud service platform, the cloud end sends user information to the system through an agent deployment mode, and after identity authentication and authority configuration, the user enters the hub-tire integrated wheel impact dynamics simulation system of the platform to use corresponding cloud computing service.
2): the method comprises the steps of entering a preprocessing module of a hub-tire integrated wheel impact dynamics simulation system, inputting a reference wheel impact test standard in a user data sub-module, importing a hub model, uploading relevant simulation parameters, automatically establishing the hub-tire integrated impact dynamics simulation model by the system through a self-programming program of a model generation sub-module, and generating a calculation file.
3): and entering a calculation module, preprocessing the calculation file in a preprocessing sub-module, analyzing to obtain the relation between the CPU core number, the memory resource, the calculation time and the cost required by calculation, giving different calculation service schemes, and selecting corresponding calculation service by a user according to own requirements.
4): the cloud service platform is automatically deployed to the back-end cloud server through corresponding codes or instructions according to the task data and command streams generated in the steps, corresponding solver resources are called, and a user can realize cloud computing of the hub-tire integrated impact dynamics simulation by clicking and running in the solving submodule.
5): the simulation calculation is completed, the simulation calculation enters a result output module, the whole calculation file, calculation process data and result data are stored in a storage Chu Zi module in a cloud storage mode, and a user can access and download the simulation calculation at any time or enter an analysis sub-module to perform online data processing and analysis.
The cloud service platform solves the confidentiality problem of data and computation of users by deploying a Trusted Execution Environment (TEE) in the cloud. Specific implementation forms of each functional module in the cloud service platform are described in detail below.
In the cloud service platform, an intelligent fitting system and a stretching simulation system are integrated in the user data sub-module, wherein: the intelligent fitting system provides a compound regression method, wherein regression algorithms such as a least square method, a genetic algorithm, an EQL neural network and the like are integrated in the intelligent fitting system, and a hub and tire material constitutive equation is fitted through the compound regression method; and the tensile simulation system automatically completes the tensile simulation of the sample through the constitutive equation of the hub material obtained by the intelligent fitting system, obtains the data such as equivalent fracture strain, stress triaxial degree and the like which are difficult to directly obtain by an actual tensile test, and returns the data to the intelligent fitting system to fit the material failure model of the hub.
In the cloud service platform described above, the standard library provides a plurality of wheel impact test standards from different countries and regions. Among the wheel impact test criteria include, but are not limited to, GB/T15704-2012 method of Road vehicle light alloy wheel impact test, QC/T991-2015 method of passenger vehicle light alloy wheel 90 degree impact test, ISO 71412022 Road Vehicles-Light alloy Wheels-Lateral impact test, SAE J175_202107 Wheels-Lateral Impact Test Procedure-Road Vehicles, and the like. The user can select corresponding reference standards according to the self wheel impact simulation requirements, and corresponding standard codes are input into the pre-processing module of the cloud service platform.
In the cloud service platform, the model library comprises, but is not limited to, a hub model library, a tire model library, a wheel assembly model library, a rack model library and a material model library. The hub model library and the tire model library comprise models which are voluntarily and openly shared by different users and models which are purchased by the platform from the users; the wheel assembly model library comprises a plurality of hub-tire integrated models which correspond to the specification models and have good assembly relations; the rack model library comprises a wheel impact test rack model (comprising a wheel supporting seat and a punch hammer) which is pre-established by a cloud service platform according to each wheel impact test standard; the material model library comprises, but is not limited to, data such as density, young's modulus, poisson's ratio and the like of common materials, and special material models and model parameters such as Mooney-Rivlin, yeoh, johnson-Cook and the like.
In the cloud service platform, the software resource library provides finite element software services with powerful dynamic simulation functions, including, but not limited to, ANSYS, ABAQUS, LS-DYNA and the like. The user does not need to download and install software by himself, and can enjoy the latest cloud software solver resource as long as the user logs in the cloud service platform.
In the cloud service platform, the hub model in the step 2) is a hub 3D model file designed in advance in three-dimensional modeling software by a user.
In the cloud service platform, the simulation parameters related in the step 2) include, but are not limited to, material data, structural parameters of a wheel impact dynamics model, load and constraint conditions, grid data, solvers, pre-output data settings, and the like. Wherein the material data includes, but is not limited to, density, young's modulus, poisson's ratio, mechanical property test data, selected material model, etc. of the hub material and the tire component materials; the structural parameters of the wheel impact dynamics model comprise the shape of a punch hammer, the falling height, the impact angle and the like; the impact angle is the angle between the axis of the wheel and the falling direction of the impact hammer, namely the angle between the axis of the wheel and the vertical direction; load and constraints include, but are not limited to, tire inflation pressure, ram mass, local gravitational acceleration, hub-to-tire contact pattern and coefficient of friction, hub mounting plate bolt torque, ram-to-impacted wheel contact pattern and coefficient of friction, and the like; grid data includes, but is not limited to, grid division, grid size, etc.; solvers include, but are not limited to, finite element software solvers such as ANSYS, ABAQUS, LS-DYNA; the pre-output data settings include, but are not limited to, stress, strain, acceleration, displacement, energy, etc.
It should be noted that, in the above simulation parameters, a part of data (for example, the shape of the impact hammer, the mass of the impact hammer, the falling height, the impact angle, etc.) has a specified value in the relevant wheel impact test standard, after the standard is selected, the part of data is automatically generated, after the user uploads the simulation parameters, the cloud system automatically analyzes the rationality and the conflict of all the data, generates a data analysis report, the user modifies the data according to the personal needs and determines whether to cover the conflict data, finally completes the data confirmation, and automatically completes the fitting of the material data and the establishment of the impact dynamics simulation model by the cloud service platform.
It should be noted that in step 2), the user may upload specific simulation parameters of the user, or may use model library data provided by the hub-tire integrated wheel impact dynamics simulation system of the cloud service platform. For example, when uploading material data, users with simulation requirements for the impact dynamics of the wheel are mostly hub manufacturers and hub design engineers, who generally cannot directly obtain the material data of the tire. Most users only need to upload the mechanical test data of the hub model and the hub material sample which are developed by themselves, the tire selects and calls the tire model with the corresponding specification model in the tire model library to finish the expected simulation analysis of the wheel impact dynamics, the cloud service platform integrates an intelligent fitting system, and the material constitutive equation and the failure equation for simulation are automatically fitted according to the material constitutive model and the failure model selected by the users and the test data provided by the users. The test data includes, but is not limited to, uniaxial tensile test data, uniaxial compressive test data, pure shear test data, of a smooth round bar or flat bar sample of material at different tensile rates. For users who want to simply and rapidly analyze whether the impact resistance of the hub preliminarily meets the requirement without pursuing extremely high simulation precision, even only upload the hub model, select a reference impact test standard, all simulation parameters can be obtained from a model library, and the system automatically establishes the hub-tire integrated impact dynamics simulation model through a self-programming program of a model generation sub-module, so that the modeling time of finite element simulation is greatly saved, the simulation efficiency is improved, and the method has guiding significance for the development and design of the hub.
For example, in this embodiment, after a user enters a preprocessing module of the hub-tire integrated wheel impact dynamics simulation system, a reference wheel impact test standard code number GB/T15704-2012 can be input into a user data sub-module, a hub three-dimensional model pre-designed by the user is imported, relevant simulation parameters are uploaded, and the system automatically builds the hub-tire integrated wheel impact dynamics simulation model through a self-programming program of a model generation sub-module and generates a calculation file. GB/T15704-2012 "impact test method for light alloy wheels of road vehicles" can evaluate the performance of the light alloy wheels in axial (transverse) impact curbs. The impact angle specified by the standard is
. After the standard code number GB/T15704-2012 is input, the cloud service platform automatically matches an impact test bench model of a bench model library, and generates relevant simulation parameters: the impact hammer is in a cuboid shape, and edges are rounded by 5mm; impact angle->
The method comprises the steps of carrying out a first treatment on the surface of the The falling height is 230mm; the tire inflation pressure was 200kPa. In addition, the simulation parameters that the user needs to upload are as follows:
material data: density, young's modulus, poisson's ratio of hub material, material mechanics test data of hub test specimen, material constitutive model selected for fitting
Load and constraint conditions: hammer mass (calculated by the following formula, the maximum static load of the hub is 800 kg)
Wherein: m-mass of the impact hammer in kilograms (kg);
w-user specified maximum static load of the hub in kilograms (kg).
Hub mounting plate bolt torque 110 N.m; local gravity acceleration of 9.79m/s 2 The method comprises the steps of carrying out a first treatment on the surface of the The contact modes of the impact hammer and the tire are friction contact, the contact friction coefficient of the tire and the wheel hub is 0.5, the contact friction coefficient of the impact hammer and the tire is 0.5, and the contact friction coefficient of the impact hammer and the wheel hub is 0.2.
Grid data: the hub is difficult to divide hexahedral grids due to complex shape and structure, and tetrahedral grids are directly divided into 5mm grids in a free division mode; the tire grid adopts any Lagrangian Euler method, and the size of the grid is 6mm; the impact hammer adopts a hexahedral grid divided by mapping, and the size of the grid is 10mm; the wheel supporting seat adopts a free division form, and the mesh size is 10mm.
The solver: select ABAQUS/Explicit solver
Pre-output data settings include, but are not limited to, mises stress, equivalent plastic strain, equivalent strain energy density, displacement, acceleration, stiffness degradation, damage initiation criteria, state, etc
The model library data used are as follows:
and inputting the tire specification model into the user data submodule to obtain a corresponding tire model from the tire model library.
The tire model consists of a tread, a sidewall, a tire shoulder, a tread base, a steel wire belt layer, a tire body cord layer, an inner liner, a bead filler, a rim liner and a tire bead wire, wherein the steel wire belt layer and the tire body cord layer are used as embedding layers to be embedded into a rubber tire. The rubber material adopts an incompressible Mooney-Rivlin super-elastic material model (the model expression is as follows), and the embedded layer adopts a common steel material, so that the density, young modulus and Poisson's ratio material properties are endowed.
Wherein: w is a function of the strain energy,
is the mechanical property constant of the material, I 1 ，I 2 The first, second, substantially invariant to the cauchy-gurin deformation tensor.
In addition, the method can also provide an automatic grid dividing function, and conveniently and rapidly divide high-quality discrete grids for the hub-tire integrated wheel impact dynamics simulation model. Preferably, the automatic mesh dividing function can adopt a free division mode to directly divide tetrahedral meshes for the whole wheel impact dynamics model; preferably, the automatic grid dividing function can also adopt a mapping dividing mode, and a mapping grid is generated by selecting proper unit attributes and a mapping method; preferably, the automatic grid dividing function can also adopt any Lagrange Euler method, and grid redrawing is realized in a form of mobile nodes in the calculation process, so that the method is suitable for a simulation scene with larger deformation of rubber tires.
The above embodiment is only a preferred embodiment of the present invention, but it is not intended to limit the present invention. Various changes and modifications may be made by one of ordinary skill in the pertinent art without departing from the spirit and scope of the present invention. Therefore, all the technical schemes obtained by adopting the equivalent substitution or equivalent transformation are within the protection scope of the invention.
Claims (10)
1. A method of hub-tire integrated wheel impact dynamics simulation comprising:
s1, sampling material mechanics test data of a hub to obtain a sample data set;
s2, training the equation learning neural network for a plurality of times by using the sample data set, and fitting a constitutive equation of the material by the equation learning neural network to obtain a plurality of machine learning fitting equations; simultaneously, fitting an existing material constitutive equation obtained according to field knowledge and priori knowledge by using the sample data set to obtain a basic equation;
s3, each basic equation, the machine learning fitting equation and the randomly generated random expression are respectively encoded into a binary tree form consisting of a terminator and an operator; taking all binary trees obtained by coding as an initial population of a genetic algorithm, performing initial cross operation, and updating individuals obtained after the initial cross operation into the original initial population;
S4, carrying out multi-objective optimization on population iteration through a genetic algorithm based on the updated initial population and an fitness function constructed by combining equation fitting degree and equation complexity, and periodically adding a basic equation fitted by the sample data set into the latest offspring population to guide the evolution direction of the genetic algorithm in the iteration process, and outputting an optimal solution after the genetic algorithm is iterated to a final condition to serve as a hub material constitutive optimal equation;
s5, the hub model and the tire model of the target wheel, the wheel impact test standard and simulation parameters including the optimal equation of the hub material structure are led into a dynamic simulation system, finite element simulation is carried out by calling a solver, and a simulation result of the target wheel under impact is output.
2. The hub-tire integrated wheel impact dynamics simulation method according to claim 1, wherein the dynamics simulation system is carried on a cloud service platform, and a standard library, a model library and a software resource library are built in the cloud service platform;
different wheel impact test standards for users to select are built in the standard library;
the model library comprises one or more of a hub model sub-library, a tire model sub-library, a wheel assembly model sub-library, a bench model sub-library and a material model sub-library, and the model in each sub-library is used for providing a direct calling function when a user does not upload a corresponding model;
And one or more dynamic simulation software for user call is built in the software resource library.
3. The hub-tire integrated wheel impact dynamics simulation method according to claim 2, wherein a preprocessing module is further arranged in the cloud service platform, and the preprocessing module comprises a user data sub-module and a model generation sub-module; the user data sub-module is used for uploading or selecting a hub model and a tire model of a target wheel, a wheel impact test standard and simulation parameters by a user, wherein a hub material constitutive optimal equation in the simulation parameters is fitted in an online or offline mode; the model generation sub-module is used for generating a calculation file required by the dynamic simulation system according to the data in the user data sub-module.
4. The hub-tire integrated wheel impact dynamics simulation method according to claim 3, wherein a calculation module is further arranged in the cloud service platform, the calculation module comprises a preprocessing sub-module and a solving sub-module, the preprocessing sub-module is used for allocating corresponding cloud computing resources according to simulation calculation amount of the dynamics simulation system, and the solving sub-module is used for carrying out finite element simulation through a solver according to the cloud computing resources allocated by the preprocessing sub-module.
5. The hub-tire integrated wheel impact dynamics simulation method according to claim 4, wherein a result output module is further arranged in the cloud service platform, and is used for storing relevant data of dynamics simulation in a cloud storage mode in the cloud, and outputting simulation results according to an output mode set by a user.
6. The hub-tire integrated wheel impact dynamics simulation method according to claim 1, wherein the existing material constitutive equations used to fit to form the base equation include a Johnson-Cook yield model, a Cowper-Symonds yield model, a Swift hardening model, and a Voce hardening model determined from domain knowledge, and a plurality of custom material constitutive equations determined from prior knowledge.
7. The hub-tire integrated wheel impact dynamics simulation method according to claim 1, wherein the sample data set is divided into a first sample data set and a second sample data set; the basic equation serving as the initial population and the machine learning fitting equation are trained or fitted by the first sample data set, and the basic equation added with the offspring population in the iteration process is fitted by the second sample data set;
The sampling range of the first sample data set in the material mechanical test data is near a yield platform or 0.2% equivalent plastic strain, and a necking stage and a yield strengthening stage, wherein the sampling interval of the yield strengthening stage is larger than that of other stages;
the sampling range of the second sample data set in the material mechanics test data is near the yield platform or 0.2% equivalent plastic strain, and a necking stage; and the sampling intervals of the second sample data set are each smaller than the sampling intervals of the first sample data set.
8. The hub-tire integrated wheel impact dynamics simulation method according to claim 1, wherein the genetic algorithm needs to preset initialization parameters before performing the iterative process, including a terminal symbol set and a function symbol set, a genetic operator, a maximum evolution number, a population number and a coding form; wherein the terminals in the terminal set include all of the independent variables and real constants involved in the base equation and the machine learning fit equation; the function symbol set comprises all basic equations and operators involved in the machine learning fit equation; the coding form is a binary tree structure; and in the genetic operators, the crossover operator only carries out crossover operation on the structure containing the independent variable in the binary tree, and the mutation operator is used for carrying out local search on the real constant of the new individual and simultaneously changing the operator.
9. The hub-tire integrated wheel impact dynamics simulation method according to claim 1, wherein the fitness function comprises an average absolute relative error, a root mean square error, a fitness coefficient and an equation complexity, wherein the fitness coefficient is a difference obtained by subtracting the decision coefficient from 1, and the equation complexity is a sum of respective complexities of all operators in the equation; the genetic algorithm comprehensively performs multi-objective optimization in a decision space based on four fitness functions.
10. The hub-tire integrated wheel impact dynamics simulation method according to claim 1, wherein the optimal solution of the genetic algorithm selects one compromise solution with the smallest coefficient of fit among pareto non-dominant solutions.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310615295.2A CN116384258B (en) | 2023-05-29 | 2023-05-29 | Hub-tire integrated wheel impact dynamics simulation method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310615295.2A CN116384258B (en) | 2023-05-29 | 2023-05-29 | Hub-tire integrated wheel impact dynamics simulation method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116384258A true CN116384258A (en) | 2023-07-04 |
| CN116384258B CN116384258B (en) | 2023-09-29 |
Family
ID=86969734
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310615295.2A Active CN116384258B (en) | 2023-05-29 | 2023-05-29 | Hub-tire integrated wheel impact dynamics simulation method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116384258B (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020077795A1 (en) * | 2000-09-21 | 2002-06-20 | Woods Joseph Thomas | System, method and storage medium for predicting impact performance of thermoplastic |
| JP2010256319A (en) * | 2009-03-31 | 2010-11-11 | Hitachi Metals Ltd | Method of simulating impact properties of wheel with tire |
| US20130289953A1 (en) * | 2012-01-24 | 2013-10-31 | The University Of Akron | Self-optimizing, inverse analysis method for parameter identification of nonlinear material constitutive models |
| CN107145663A (en) * | 2017-05-04 | 2017-09-08 | 吉林大学 | Wheel multi-objective optimization design of power method |
| US20200394278A1 (en) * | 2019-06-14 | 2020-12-17 | Abraham Varon-Weinryb | Hybrid Finite Element and Artificial Neural Network Method and System for Safety Optimization of Vehicles |
| CN113420485A (en) * | 2021-07-19 | 2021-09-21 | 扬州大学 | Non-pneumatic tire transient impact characteristic prediction method |
| CN113764056A (en) * | 2021-09-06 | 2021-12-07 | 北京理工大学重庆创新中心 | Method for obtaining high-precision hardening model parameters of material under multiple strain rates |
| CN114528739A (en) * | 2022-02-28 | 2022-05-24 | 重庆长安汽车股份有限公司 | Simulation method for automobile hub fracture failure |
| CN115510553A (en) * | 2022-08-09 | 2022-12-23 | 岚图汽车科技有限公司 | Modeling method and device for predicting tire failure in small offset collision |
-
2023
- 2023-05-29 CN CN202310615295.2A patent/CN116384258B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020077795A1 (en) * | 2000-09-21 | 2002-06-20 | Woods Joseph Thomas | System, method and storage medium for predicting impact performance of thermoplastic |
| JP2010256319A (en) * | 2009-03-31 | 2010-11-11 | Hitachi Metals Ltd | Method of simulating impact properties of wheel with tire |
| US20130289953A1 (en) * | 2012-01-24 | 2013-10-31 | The University Of Akron | Self-optimizing, inverse analysis method for parameter identification of nonlinear material constitutive models |
| CN107145663A (en) * | 2017-05-04 | 2017-09-08 | 吉林大学 | Wheel multi-objective optimization design of power method |
| US20200394278A1 (en) * | 2019-06-14 | 2020-12-17 | Abraham Varon-Weinryb | Hybrid Finite Element and Artificial Neural Network Method and System for Safety Optimization of Vehicles |
| CN113420485A (en) * | 2021-07-19 | 2021-09-21 | 扬州大学 | Non-pneumatic tire transient impact characteristic prediction method |
| CN113764056A (en) * | 2021-09-06 | 2021-12-07 | 北京理工大学重庆创新中心 | Method for obtaining high-precision hardening model parameters of material under multiple strain rates |
| CN114528739A (en) * | 2022-02-28 | 2022-05-24 | 重庆长安汽车股份有限公司 | Simulation method for automobile hub fracture failure |
| CN115510553A (en) * | 2022-08-09 | 2022-12-23 | 岚图汽车科技有限公司 | Modeling method and device for predicting tire failure in small offset collision |
Non-Patent Citations (4)
| Title |
|---|
| M.G. DICK; B.M. WILSON: "The effect of sustained wheel impacts on tapered roller bearing cages", PROCEEDINGS OF THE 2006 IEEE/ASME JOINT RAIL CONFERENCE * |
| 尹冀;朱平;章斯亮;: "考虑应变率效应的钢制车轮冲击仿真与试验", 上海交通大学学报, no. 06 * |
| 赵成果;范平清;: "胎压监测装置中橡胶材料本构模型的确定及力学研究", 轻工机械, no. 05 * |
| 黄文;赵涵;陆宏;刘智;肖冰冰;秦泗吉;: "基于单轴拉伸试验、模拟及优化方法的材料本构识别", 塑性工程学报, no. 06 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116384258B (en) | 2023-09-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2022083009A1 (en) | Customized product performance prediction method based on heterogeneous data error compensation fusion | |
| WO2021175058A1 (en) | Neural network architecture search method and apparatus, device and medium | |
| CN104978612A (en) | Distributed big data system risk predicating method based on AHP-RBF | |
| CN111275172A (en) | Feedforward neural network structure searching method based on search space optimization | |
| US20220122103A1 (en) | Customized product performance prediction method based on heterogeneous data difference compensation fusion | |
| CN110533150A (en) | Self -adaptive and reuse system and method based on Support vector regression model | |
| CN109815541B (en) | Method and device for dividing rail transit vehicle product parts and modules and electronic equipment | |
| CN111340493A (en) | Multi-dimensional distributed abnormal transaction behavior detection method | |
| CN110727758B (en) | Public opinion analysis method and system based on multi-length text vector splicing | |
| CN110210654A (en) | Product model designing system and method | |
| CN105512755A (en) | Decomposition-based multi-objective distribution estimation optimization method | |
| CN110705080A (en) | Precise assembling method and device for digital twin model | |
| CN114936701A (en) | Real-time monitoring method and device for comprehensive energy consumption and terminal equipment | |
| CN116384258B (en) | Hub-tire integrated wheel impact dynamics simulation method | |
| CN113743461B (en) | Unmanned aerial vehicle cluster health degree assessment method and device | |
| CN111639388B (en) | Method and system for simulating parameters of elastic element of automobile | |
| CN116976220A (en) | Global agent optimization-based acceleration test method for automatic driving system | |
| CN111105127B (en) | Modular product design evaluation method based on data driving | |
| CN115293056B (en) | Modelica model-oriented multi-objective optimization method | |
| CN112241811B (en) | Layered hybrid performance prediction method for customized product in Internet +' environment | |
| CN113919080A (en) | Machine learning-based quick evaluation method for mechanical property of automobile engine hood | |
| CN111523647B (en) | Network model training method and device, feature selection model, method and device | |
| CN114742036A (en) | Combined model compression method and system for pre-training language model | |
| Wang et al. | An improved system for efficient shape optimization of vehicle aerodynamics with “noisy” computations | |
| CN113837440A (en) | Blasting effect prediction method and device, electronic equipment and medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |