Battery pack system mechanical safety prediction method

Abstract

The invention discloses a battery pack system mechanical safety prediction method, which comprises the following steps: 1) acquiring n groups of battery pack system mechanical responses and deformation information of each battery monomer under different extrusion speeds, extrusion positions and battery pack system core component thicknesses, and constructing a battery pack system mechanical safety prediction sample set; 2) training the machine learning model by using the battery pack system mechanical safety prediction sample set to obtain a battery pack mechanical safety performance prediction model; and 3) utilizing the battery pack mechanical safety performance prediction model to establish battery pack system mechanical response and deformation information of each battery monomer under different extrusion speeds, extrusion positions and battery pack system core component thicknesses, and establishing a fitting curved surface between the core component and the battery pack system, module or monomer mechanical information. When the vehicle is extruded, the mechanical safety performance of the battery pack system can be efficiently predicted in time.

Description

Technical Field

[0001] The present invention relates to the field of electric vehicle safety and vehicle battery pack system design, and in particular to a battery pack system mechanical safety prediction method. Background Art

[0002] With the rapid development of my country's automobile industry, the number of electric vehicles has ranked first in the world. As a key core component of electric vehicles, the battery pack system plays a vital role in power supply. However, given the harsh road environment and increasingly complex traffic environment, different extrusion conditions (such as vehicle collisions, battery pack scratches, obstacle impacts, etc.) will cause immeasurable damage to the battery pack system, and in severe cases may even cause safety accidents such as fires and explosions, greatly affecting the driving safety and traffic safety of electric vehicles. In addition, if the battery pack system under extrusion conditions is not subjected to stress analysis, it is impossible to evaluate the reliability of the battery pack system after micro-extrusion, which will bring safety hazards to the continued use of the battery pack and vehicle driving in the future.

[0003] Mechanical safety prediction of battery pack systems can timely identify potential mechanical structural defects and weak links. For example, the battery pack's tolerance under different extrusion conditions can be determined through evaluation to avoid damage to the battery pack due to vehicle collisions, scratches, or impacts with obstacles, thereby reducing the risk of serious accidents such as fire and explosion. It ensures that the battery pack maintains stable structural integrity under various complex driving conditions, providing strong protection for the safe operation of electric vehicles.

[0004] When an electric vehicle has an accident, good mechanical safety performance of the battery pack can reduce the harm of the battery system to passengers. If the battery pack can remain stable during a collision without rupture or leakage, the risk of electric shock and burns to passengers can be reduced. At the same time, a reliable battery pack structure can also help maintain the overall stability of the vehicle and reduce the probability of secondary accidents.

[0005] Therefore, it is necessary to propose a reliable mechanical safety detection scheme for battery pack systems. Summary of the invention

[0006] The object of the present invention is to provide a method for predicting mechanical safety of a battery pack system, comprising the following steps:

[0007] 1) Establish a complete finite element model of the battery pack system and a multi-body dynamics model of different SOC battery pack system modules;

[0008] 2) Set the extrusion speed, extrusion position and thickness of the core components of the battery pack system, and carry out extrusion simulation analysis of the battery pack system, and use the finite element model to obtain the mechanical response of the battery pack system under different extrusion speeds, extrusion positions and core component thicknesses;

[0009] 3) Import the obtained module deformation into the multi-body dynamics model, and obtain the deformation information of each battery cell through dynamic simulation;

[0010] 4) Repeat step 2) to step 3) n times to obtain the mechanical response of the battery pack system and the deformation information of each battery cell under n different extrusion speeds, extrusion positions and thicknesses of the core components of the battery pack system, and construct a sample set of mechanical safety prediction of the battery pack system;

[0011] 5) Use the battery pack system mechanical safety prediction sample set to train the machine learning model to obtain the battery pack mechanical safety performance prediction model;

[0012] 6) Use the battery pack mechanical safety performance prediction model to build the mechanical response of the battery pack system and the deformation information of each battery cell under different extrusion speeds, extrusion positions and thicknesses of the core components of the battery pack system.

[0013] Furthermore, in step 1), the steps of establishing a complete finite element model of the battery pack system and a multi-body dynamics model of different SOC battery pack system modules include:

[0014] 1.1) Establish a finite element model of the battery pack shell;

[0015] 1.2) Establish a finite element model of the battery module;

[0016] 1.3) According to the actual connection relationship of the battery pack system module, the battery pack shell finite element model and the battery module finite element model are integrated to establish a complete battery pack system finite element model including the contact connection relationship.

[0017] 1.4) A multi-body dynamics model of battery pack system modules with different SOCs is constructed by treating the battery and the filling glue as spring-damping structures; among them, the stiffness and damping parameters are used to characterize the battery SOC.

[0018] Further, in step 1.1), the step of establishing a finite element model of the battery pack shell includes: defining the battery pack system shell parameters in a finite element commercial software according to the size, structure and material parameters of the battery pack system shell, thereby establishing a finite element model of the battery pack shell.

[0019] Furthermore, the battery pack system shell parameters include unit type, size parameters, thickness parameters, and material parameters.

[0020] Further, in step 1.2), the step of establishing a finite element model of the battery module includes:

[0021] 1.2.1) Establishing a geometric model of the battery module according to the size parameters of the battery module;

[0022] 1.2.2) Homogenize the battery module materials, define the module material parameters, and establish the finite element model of the battery module.

[0023] Furthermore, the core components include a long bracket, a lifting ear, a bottom shell, a lower supporting beam, upper and lower connecting brackets and an upper bracket that are not in the existing battery pack system.

[0024] Furthermore, the battery pack mechanical safety performance prediction model takes the extrusion speed, extrusion position and thickness of the core components of the battery pack system as input, and takes the mechanical response of the battery pack system and the deformation information of each battery cell as output.

[0025] Furthermore, the hyperparameters of the battery pack mechanical safety performance prediction model are determined through random experiments; the hyperparameters include learning rate and number of iterations.

[0026] Furthermore, machine learning models include but are not limited to deep neural network models, particle swarm optimization models, condor optimization-extreme learning machines, and random forest models.

[0027] Furthermore, the mean square error loss function of the machine learning model is As shown below:

[0028]

[0029] In the formula, y represents the output value and true value of the machine learning model respectively; n is the number of samples.

[0030] The technical effect of the present invention is unquestionable. The present invention establishes a complete finite element model of the battery pack system based on the real structure of the battery pack system, obtains the mechanical information of the battery pack system and the module based on the finite element extrusion analysis, builds a multi-body dynamics model of the module through an equivalent model, uses the maximum deformation of the module obtained by finite element analysis, and obtains the deformation information of the monomer through dynamic simulation, thereby completing the calculation of the maximum stress, maximum deformation, maximum acceleration and maximum deformation of the battery pack system and the module under different extrusion speeds, different extrusion positions, different SOCs and different core component thicknesses, and combines machine learning methods to use limited data to efficiently predict the mechanical response of the system under all extrusion conditions and component thickness combinations. And build a fitting surface between the core components and the mechanical information of the battery pack system, module or monomer to guide R&D personnel in design.

[0031] The present invention combines finite element modeling and analysis of the battery pack system, multi-body dynamics modeling and analysis, and machine learning methods to efficiently predict the mechanical safety of the battery pack system under different extrusion speeds, different extrusion positions, different SOCs, and different core component thickness combinations. When the vehicle is extruded, the present invention can timely and efficiently predict the mechanical safety performance of the battery pack system, which can provide a reference for the timely response of the occupants and ensure the safety of passengers' lives and property. In addition, the present invention reveals the impact of changes in the thickness of core components on the mechanical safety performance of the battery pack system, which can provide a reference for researchers' design. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 A flowchart of a method for predicting mechanical safety of an efficient battery pack system;

[0033] Figure 2 It is the finite element model and multi-body dynamics equivalent diagram of the battery pack system;

[0034] Figure 3 A comparison chart of the prediction accuracy of different machine learning models;

[0035] Figure 4 It is a fitted surface diagram of the thickness of the bottom shell and upper cover of the battery pack system and the maximum deformation of the module. DETAILED DESCRIPTION

[0036] The present invention is further described below in conjunction with the embodiments, but it should not be understood that the above subject matter of the present invention is limited to the following embodiments. Without departing from the above technical ideas of the present invention, various substitutions and changes are made according to the common technical knowledge and customary means in the art, which should all be included in the protection scope of the present invention.

[0037] Embodiment 1:

[0038] See also Figures 1 to 4 , a battery pack system mechanical safety prediction method, comprising the following steps:

[0039] 1) Establish a complete finite element model of the battery pack system and a multi-body dynamics model of different SOC battery pack system modules;

[0040] 2) Set the extrusion speed, extrusion position and thickness of the core components of the battery pack system, and carry out extrusion simulation analysis of the battery pack system, and use the finite element model to obtain the mechanical response of the battery pack system under different extrusion speeds, extrusion positions and core component thicknesses;

[0041] 3) Import the obtained module deformation into the multi-body dynamics model, and obtain the deformation information of each battery cell through dynamic simulation;

[0042] 4) Repeat step 2) to step 3) n times to obtain the mechanical response of the battery pack system and the deformation information of each battery cell under n different extrusion speeds, extrusion positions and thicknesses of the core components of the battery pack system, and construct a sample set of mechanical safety prediction of the battery pack system;

[0043] 5) Use the battery pack system mechanical safety prediction sample set to train the machine learning model to obtain the battery pack mechanical safety performance prediction model;

[0044] 6) Use the battery pack mechanical safety performance prediction model to build the mechanical response of the battery pack system and the deformation information of each battery cell under different extrusion speeds, extrusion positions and thicknesses of the core components of the battery pack system, and build a fitting surface between the core components and the battery pack system, module or cell mechanical information based on this information.

[0045] In step 1), the steps of establishing a complete finite element model of the battery pack system and a multi-body dynamics model of different SOC battery pack system modules include:

[0046] 1.1) Establish a finite element model of the battery pack shell;

[0047] 1.2) Establish a finite element model of the battery module;

[0048] 1.3) According to the actual connection relationship of the battery pack system module, the battery pack shell finite element model and the battery module finite element model are integrated to establish a complete battery pack system finite element model including the contact connection relationship.

[0049] 1.4) A multi-body dynamics model of battery pack system modules with different SOCs is constructed by treating the battery and the filling glue as spring-damping structures; among them, the stiffness and damping parameters are used to characterize the battery SOC.

[0050] In step 1.1), the step of establishing a finite element model of the battery pack shell includes: defining the battery pack system shell parameters in a finite element commercial software according to the size, structure and material parameters of the battery pack system shell, thereby establishing a finite element model of the battery pack shell.

[0051] The battery pack system shell parameters include unit type, size parameters, thickness parameters, and material parameters.

[0052] In step 1.2), the steps of establishing the battery module finite element model include:

[0053] 1.2.1) Establishing a geometric model of the battery module according to the size parameters of the battery module;

[0054] 1.2.2) Homogenize the battery module materials, define the module material parameters, and establish the finite element model of the battery module.

[0055] The core components include but are not limited to the long bracket, lifting ears, bottom shell, lower supporting beam, upper and lower connecting brackets and upper bracket of the existing battery pack system.

[0056] The battery pack mechanical safety performance prediction model takes the extrusion speed, extrusion position and thickness of the core components of the battery pack system as input, and takes the mechanical response of the battery pack system and the deformation information of each battery cell as output.

[0057] The hyperparameters of the battery pack mechanical safety performance prediction model are determined through random experiments; the hyperparameters include learning rate and number of iterations.

[0058] Machine learning models include but are not limited to deep neural network models, particle swarm optimization models, condor optimization-extreme learning machines, and random forest models.

[0059] Mean Squared Error Loss Function for Machine Learning Models As shown below:

[0060]

[0061] In the formula, y represents the output value and true value of the machine learning model respectively; n is the number of samples.

[0062] Embodiment 2:

[0063] A method for predicting mechanical safety of a battery pack system comprises the following steps:

[0064] 1) Establish a complete finite element model of the battery pack system and a multi-body dynamics model of different SOC battery pack system modules;

[0065] 2) Set the extrusion speed, extrusion position and thickness of the core components of the battery pack system, and carry out extrusion simulation analysis of the battery pack system, and use the finite element model to obtain the mechanical response of the battery pack system under different extrusion speeds, extrusion positions and core component thicknesses;

[0066] 3) Import the obtained module deformation into the multi-body dynamics model, and obtain the deformation information of each battery cell through dynamic simulation;

[0067] 4) Repeat step 2) to step 3) n times to obtain the mechanical response of the battery pack system and the deformation information of each battery cell under n different extrusion speeds, extrusion positions and thicknesses of the core components of the battery pack system, and construct a sample set of mechanical safety prediction of the battery pack system;

[0068] 5) Use the battery pack system mechanical safety prediction sample set to train the machine learning model to obtain the battery pack mechanical safety performance prediction model;

[0069] 6) Use the battery pack mechanical safety performance prediction model to build the mechanical response of the battery pack system and the deformation information of each battery cell under different extrusion speeds, extrusion positions and thicknesses of the core components of the battery pack system, and build a fitting surface between the core components and the mechanical information of the battery pack system, module or cell.

[0070] Embodiment 3:

[0071] A method for predicting mechanical safety of a battery pack system, the technical content of which is the same as that of Example 2, further, in step 1), the steps of establishing a complete finite element model of the battery pack system and a multi-body dynamics model of different SOC battery pack system modules include:

[0072] 1.1) Establish a finite element model of the battery pack shell;

[0073] 1.2) Establish a finite element model of the battery module;

[0074] 1.3) According to the actual connection relationship of the battery pack system module, the battery pack shell finite element model and the battery module finite element model are integrated to establish a complete battery pack system finite element model including the contact connection relationship.

[0075] 1.4) A multi-body dynamics model of battery pack system modules with different SOCs is constructed by treating the battery and the filling glue as spring-damping structures; among them, the stiffness and damping parameters are used to characterize the battery SOC.

[0076] Embodiment 4:

[0077] A method for predicting mechanical safety of a battery pack system, the technical content of which is the same as any one of Examples 2-3, and further, in step 1.1), the step of establishing a finite element model of a battery pack shell includes: defining the battery pack system shell parameters in a finite element commercial software according to the size, structure and material parameters of the battery pack system shell, thereby establishing a finite element model of the battery pack shell.

[0078] Embodiment 5:

[0079] A method for predicting mechanical safety of a battery pack system, the technical content of which is the same as any one of Examples 2-4, and further, the shell parameters of the battery pack system include unit type, size parameters, thickness parameters, and material parameters.

[0080] Embodiment 6:

[0081] A method for predicting mechanical safety of a battery pack system, the technical content of which is the same as any one of Embodiments 2-5, further, in step 1.2), the step of establishing a finite element model of a battery module comprises:

[0082] 1.2.1) Establishing a geometric model of the battery module according to the size parameters of the battery module;

[0083] 1.2.2) Homogenize the battery module materials, define the module material parameters, and establish the finite element model of the battery module.

[0084] Embodiment 7:

[0085] A method for predicting mechanical safety of a battery pack system, the technical content of which is the same as any one of Examples 2-6, and further, the core components include a long bracket, a lifting lug, a bottom shell, a lower supporting beam, upper and lower connecting brackets and an upper bracket of the battery pack system.

[0086] Embodiment 8:

[0087] A method for predicting the mechanical safety of a battery pack system, the technical content of which is the same as any one of Examples 2-7. Furthermore, the battery pack mechanical safety performance prediction model takes the extrusion speed, extrusion position and thickness of the core components of the battery pack system as inputs, and takes the mechanical response of the battery pack system and the deformation information of each battery cell as outputs.

[0088] Embodiment 9:

[0089] A method for predicting the mechanical safety of a battery pack system, the technical content of which is the same as any one of Examples 2-8, and further, the hyperparameters of a battery pack mechanical safety performance prediction model are determined by random experiments; the hyperparameters include a learning rate and the number of iterations.

[0090] Embodiment 10:

[0091] A method for predicting mechanical safety of a battery pack system, the technical content of which is the same as any one of Examples 2-9, and further, the machine learning model includes but is not limited to a deep neural network model, a particle swarm optimization model, a vulture optimization-extreme learning machine, and a random forest model.

[0092] Embodiment 11:

[0093] A method for predicting mechanical safety of a battery pack system, the technical content of which is the same as any one of Embodiments 2-10, and further, the mean square error loss function of the machine learning model As shown below:

[0094]

[0095] In the formula, y represents the output value and true value of the machine learning model respectively; n is the number of samples.

[0096] Embodiment 12:

[0097] An efficient battery pack system mechanical safety prediction method includes the following steps:

[0098] 1) Establish a complete finite element model of the battery pack system.

[0099] The steps to build a complete finite element model of the battery pack system include:

[0100] 1.1) Establish a finite element model of the battery pack shell;

[0101] 1.2) Establish a finite element model of the battery module;

[0102] 1.3) Based on the actual connection relationship of the battery pack system module, a complete battery pack system finite element model including the contact connection relationship is established.

[0103] 1.4) The multi-body dynamics model of the module is constructed by equating the battery and the filling glue to a spring damping structure. Different battery SOCs correspond to different stiffness damping parameters.

[0104] 2) Select different extrusion speeds, extrusion positions, and thicknesses of core components of the battery pack system to carry out extrusion simulation analysis of the battery pack system, and use the finite element model to obtain the mechanical response of the battery pack system under different extrusion speeds, extrusion positions, and thicknesses of core components;

[0105] 3) The obtained module deformation is imported into the multi-body dynamics model, and the deformation information of each battery cell is obtained through dynamic simulation.

[0106] The main steps are:

[0107] 3.1) After obtaining the mechanical information of the battery pack system using finite element analysis, the deformation of the module is input into the multi-body dynamics model;

[0108] 3.2) The deformation information of the monomer can be quickly obtained through dynamic simulation;

[0109] 3.3) Combined with the finite element analysis results, the maximum stress, maximum deformation, maximum acceleration and maximum deformation of the battery pack system and module at different extrusion positions, different extrusion speeds and different SOCs can be obtained to obtain a data set.

[0110] 4) Use the extrusion speed, extrusion position, SOC, thickness of the core components of the battery pack system, and the mechanical information data set of the battery pack system, module or monomer to establish a machine learning model. When establishing the model, divide the data set into a training set and a test set according to a certain ratio. For example, if 500 sets of data are obtained, 400 sets of data (80%, but not limited to) and 100 sets of data (20%, but not limited to) can be extracted from the 500 sets of data as the training set and test set of the machine learning model. The ultimate goal is to establish the optimal mechanical safety prediction model for the battery pack system.

[0111] 5) Use the established machine learning model to accurately predict the mechanical safety performance of the battery pack system under different extrusion conditions and different thicknesses of the core components of the battery pack system;

[0112] 6) Build fitting surfaces between core components and battery pack systems, modules or monomer mechanical information, and guide R&D personnel in design.

[0113] Embodiment 13:

[0114] A method for predicting mechanical safety of a battery pack system comprises the following steps:

[0115] 1) Establish a complete finite element model of the battery pack system and a multi-body dynamics model of different SOC battery pack system modules;

[0116] 2) Select different extrusion speeds, extrusion positions, and thicknesses of core components of the battery pack system to carry out extrusion simulation analysis of the battery pack system, and use the finite element model to obtain the mechanical response of the battery pack system under different extrusion speeds, extrusion positions, and thicknesses of core components;

[0117] 3) Import the obtained module deformation into the multi-body dynamics model, and obtain the deformation information of each battery cell through dynamic simulation;

[0118] 4) Establish a machine learning model using the extrusion speed, extrusion position, SOC, thickness of core components of the battery pack system, and mechanical information data sets of the battery pack system, module or monomer;

[0119] 5) Use the established machine learning model to accurately predict the mechanical safety performance of the battery pack system under different extrusion conditions;

[0120] 6) Build fitting surfaces between core components and battery pack systems, modules or monomer mechanical information, and guide R&D personnel in design.

[0121] The steps to build a complete finite element model of the battery pack system include:

[0122] 1.1) Establish a finite element model of the battery pack shell;

[0123] 1.2) Establish a finite element model of the battery module;

[0124] 1.3) Based on the actual connection relationship of the battery pack system module, a complete battery pack system finite element model including the contact connection relationship is established.

[0125] 1.4) The multi-body dynamics model of the module is constructed by equating the battery and the filling glue to a spring damping structure. Different battery SOCs correspond to different stiffness damping parameters.

[0126] The working condition analyzed is the extrusion of the battery pack system (the loading method refers to the national standard GB38031-2020).

[0127] After obtaining the mechanical information of the battery pack system through finite element analysis, the deformation of the module is input into the multi-body dynamics model. The deformation information of the single body can be quickly obtained through dynamic simulation. Combined with the results of finite element analysis, the maximum stress, maximum deformation, maximum acceleration and maximum deformation of the battery pack system and module under different extrusion positions, different extrusion speeds and different SOC can be obtained.

[0128] A machine learning model is established using the extrusion speed, extrusion position, SOC, thickness of core components of the battery pack system, and mechanical information data sets of the battery pack system, module or monomer, revealing the complex nonlinear relationship between the extrusion speed, extrusion position, SOC, thickness of core components of the battery pack system, and mechanical information of the battery pack system, module or monomer.

[0129] The established machine learning model takes the extrusion speed, extrusion position, SOC, and thickness parameters of the core components of the battery pack system as input, and the mechanical information parameters of the battery pack system, module or monomer as output. In addition, other hyperparameters in the model, such as learning rate, number of iterations, etc., can be determined through random experiments (but not limited to), and the ultimate goal is to find the optimal machine learning parameters and establish the optimal mechanical safety prediction model of the battery pack system.

[0130] Embodiment 14:

[0131] See also Figures 1 to 4 , an efficient battery pack system mechanical safety prediction method, comprising the following steps:

[0132] The purpose of the present invention is to provide an efficient method for predicting mechanical safety of a battery pack system, comprising the following steps:

[0133] 1. Establish a complete finite element model of the battery pack system, and build a multi-body dynamic model of the module by equating the battery and the filling glue to a spring damping structure. Different battery SOCs correspond to different stiffness and damping parameters.

[0134] The steps to build a complete finite element model of the battery pack system include:

[0135] 1.1) According to the size, structure and material parameters of the battery pack system shell of a certain electric vehicle, define the unit type, size parameters, thickness parameters, material parameters, etc. in the finite element commercial software, and establish the finite element model of the battery pack shell;

[0136] 1.2) Establish a geometric model of the battery module based on the size parameters of the battery module, homogenize the battery module material, define the module material parameters, and establish a finite element model of the battery module;

[0137] 1.3) According to the actual connection relationship of the battery pack system module, the contact connection relationships such as welding and friction between the different components of the battery pack system and the battery module are established in the finite element commercial software, and the battery pack shell finite element model and the battery module finite element model are coupled to establish a complete battery pack system finite element model including the contact connection relationship.

[0138] Figure 2 It is the finite element model and multi-body dynamics equivalent diagram of the battery pack system.

[0139] 2. Select different extrusion speeds, extrusion positions, and thicknesses of core components of the battery pack system to carry out extrusion simulation analysis of the battery pack system, and use the finite element model to obtain the mechanical response of the battery pack system under different extrusion speeds, extrusion positions, and thicknesses of core components. The complete finite element model of the battery pack system is universal and can be used for complex mechanical analysis of different finite element commercial software (such as LS-DYNA or ABAQUS):

[0140] 3. After using finite element analysis to obtain the mechanical information of the battery pack system, the deformation of the module is input into the multi-body dynamics model. The multi-body dynamics model is universal and can be used in different commercial dynamics software (such as ADAMS). The deformation information of the single body can be quickly obtained through dynamic simulation. Combined with the results of finite element analysis, the maximum stress, maximum deformation, maximum acceleration and maximum deformation of the battery pack system and module under different extrusion positions, different extrusion speeds, different SOCs and combinations of the bottom shell and upper cover thickness of the battery pack system can be obtained to form a data set.

[0141] 4. Use the extrusion speed, extrusion position, SOC, thickness of the core components of the battery pack system, and the mechanical information data set of the battery pack system, module or monomer to establish a machine learning model. When establishing the model, divide the data set into a training set and a test set according to a certain ratio. For example, if 500 sets of data are obtained, 400 sets of data (80%, but not limited to) and 100 sets of data (20%, but not limited to) can be extracted from the 500 sets of data as the training set and test set of the machine learning model. The ultimate goal is to establish the optimal mechanical safety prediction model for the battery pack system.

[0142] 5. Use the established machine learning model to accurately predict the mechanical safety performance of the battery pack system under different extrusion conditions and different thicknesses of the core components of the battery pack system;

[0143] Figure 3 A comparison chart of the prediction accuracy of different machine learning models.

[0144] 6. Build fitting surfaces between core components and battery pack systems, modules or monomer mechanical information, and guide R&D personnel in design.

[0145] Figure 4 It is a fitted surface diagram of the thickness of the bottom shell and upper cover of the battery pack system and the maximum deformation of the module.

[0146] Experimental results:

[0147] exist Figure 3 Four machine learning algorithms, namely deep neural network (DNN), particle swarm optimization-radial basis function (PSO-RBF), condor optimization-extreme learning machine (BES-ELM) and random forest (RF), were used. 2 As the evaluation index of model accuracy, four machine learning algorithms were used to predict the maximum deformation of the module, the maximum Mises stress of the module, and the maximum deformation of the monomer. From the results, it can be seen that the prediction accuracy of the four models is above 0.9, and the DNN model has the highest prediction accuracy, which proves that the use of machine learning algorithms can make high-precision predictions on the mechanical safety performance of the battery pack system.

[0148] exist Figure 4 In the figure, the fitting relationship between the bottom shell thickness (t_bottom), the upper cover thickness (t_upper) and the maximum deformation of the module (Max deformation-module) shows that the reduction of the bottom shell thickness of the battery pack system and the reduction of the upper cover thickness of the battery pack system can help reduce the mechanical response of the battery pack system when it is squeezed and improve the safety performance of the battery pack system.

[0149] In summary, this embodiment combines finite element modeling and analysis of the battery pack system, multi-body dynamics modeling and analysis, and machine learning methods to efficiently and accurately predict the mechanical safety of the battery pack system under different extrusion speeds, different extrusion positions, different SOCs, and different core component thickness combinations. When the vehicle is squeezed, timely and efficient prediction of the mechanical safety performance of the battery pack system can provide a reference for the timely response of the occupants and ensure the safety of the passengers' lives and property. In addition, the impact of changes in the thickness of core components on the mechanical safety performance of the battery pack system is revealed, which can provide a reference for researchers' design.

Claims

1. A method for predicting mechanical safety of a battery pack system, characterized in that: The following steps are involved: 1) Establish a complete finite element model of the battery pack system and a multi-body dynamics model of different SOC battery pack system modules; 2) Set the extrusion speed, extrusion position and thickness of the core components of the battery pack system, and carry out extrusion simulation analysis of the battery pack system, and use the finite element model to obtain the mechanical response of the battery pack system under different extrusion speeds, extrusion positions and core component thicknesses; 3) Import the obtained module deformation into the multi-body dynamics model, and obtain the deformation information of each battery cell through dynamic simulation; 4) Repeat step 2) to step 3) n times to obtain the mechanical response of the battery pack system and the deformation information of each battery cell under n different extrusion speeds, extrusion positions and thicknesses of the core components of the battery pack system, and construct a sample set of mechanical safety prediction of the battery pack system; 5) Use the battery pack system mechanical safety prediction sample set to train the machine learning model to obtain the battery pack mechanical safety performance prediction model. 6) Use the battery pack mechanical safety performance prediction model to build the mechanical response of the battery pack system and the deformation information of each battery cell under different extrusion speeds, extrusion positions and thicknesses of the core components of the battery pack system.

2. According to claim 1, an efficient battery pack system mechanical safety prediction method is characterized in that: In step 1), the steps of establishing a complete finite element model of the battery pack system and a multi-body dynamics model of different SOC battery pack system modules include: 1.1) Establish a finite element model of the battery pack shell; 1.2) Establish a finite element model of the battery module; 1.3) According to the actual connection relationship of the battery pack system module, the battery pack shell finite element model and the battery module finite element model are integrated to establish a complete battery pack system finite element model including the contact connection relationship. 1.4) A multi-body dynamics model of battery pack system modules with different SOCs is constructed by treating the battery and the filling glue as spring-damping structures; among them, the stiffness and damping parameters are used to characterize the battery SOC.

3. According to claim 2, an efficient battery pack system mechanical safety prediction method is characterized in that: In step 1.1), the step of establishing a finite element model of the battery pack shell includes: defining the battery pack system shell parameters in a finite element commercial software according to the size, structure and material parameters of the battery pack system shell, thereby establishing a finite element model of the battery pack shell.

4. According to claim 3, an efficient battery pack system mechanical safety prediction method is characterized in that: The battery pack system shell parameters include unit type, size parameters, thickness parameters, and material parameters.

5. According to claim 2, an efficient battery pack system mechanical safety prediction method is characterized in that: In step 1.2), the steps of establishing the battery module finite element model include: 1.2.1) Establishing a geometric model of the battery module according to the size parameters of the battery module; 1.2.2) Homogenize the battery module materials, define the module material parameters, and establish the finite element model of the battery module.

6. According to claim 1, an efficient battery pack system mechanical safety prediction method is characterized in that: The core components include but are not limited to a long bracket, a lifting lug, a bottom shell, a lower supporting beam, upper and lower connecting brackets and an upper bracket of the battery pack system.

7. The efficient mechanical safety prediction method of a battery pack system according to claim 1, characterized in that: The battery pack mechanical safety performance prediction model takes the extrusion speed, extrusion position and thickness of the core components of the battery pack system as input, and takes the mechanical response of the battery pack system and the deformation information of each battery cell as output.

8. The efficient battery pack system mechanical safety prediction method according to claim 1 is characterized in that: The hyperparameters of the battery pack mechanical safety performance prediction model are determined through random experiments; the hyperparameters include learning rate and number of iterations.

9. The efficient battery pack system mechanical safety prediction method according to claim 1, characterized in that: Machine learning models include but are not limited to deep neural network models, particle swarm optimization models, condor optimization-extreme learning machines, and random forest models.

10. An efficient battery pack system mechanical safety prediction method according to claim 9, characterized in that: Mean Squared Error Loss Function for Machine Learning Models As shown below: In the formula, y represents the output value and true value of the machine learning model respectively; n is the number of samples.

Images