Simulation optimization design method and system for battery rack anti-roll collision

Abstract

The present application relates to the technical field of structure optimization, in particular to a simulation optimization design method and system for battery rack anti-tilting collision, in the present application, the energy absorption values of longitudinal beams and cross beam units are normalized according to volume, and are matched with principal stress peak values one by one, the weight of stiffened plate and support frame is calculated, and the double-target data composed of performance and weight are input into a multi-objective non-dominated sorting genetic algorithm, scheme grade division, crowding degree calculation, cross section size exchange and node rigidity coefficient disturbance are executed, so that the structure realizes weight balance under the premise of maintaining stress performance, low energy absorption units are cyclically removed by using the variable density method, and the cross section is increased and the stress path is adjusted in high stress concentration units, so that the material distribution is concentrated to the high-efficiency energy absorption area, further, the node stress is recorded in time in the collision simulation, and integration and sorting are performed, the geometric position of the high stress node is located, and the continuity of the load bearing path and the impact resistance of the structure are optimized under the driving of quantitative data.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of structural optimization, in particular to a simulation optimization design method and system for battery rack anti-rollover collision. BACKGROUND

[0002] The technical field of structural optimization improves structural design through mathematical models and optimization algorithms to achieve performance optimization, aiming to improve the strength, stiffness, stability and safety of the structure, while reducing weight, cost or other unnecessary resource consumption, ensuring the reliability and efficiency of the structure during use.

[0003] A simulation optimization design method for battery rack anti-rollover collision aims to improve the impact resistance of the battery rack when encountering rollover collision through computer simulation and optimization algorithms. Through simulation analysis, the weak links in the battery rack design are identified, and the structure parameters are adjusted through optimization means to improve the overall anti-collision ability, aiming to ensure the safety of the battery rack in applications such as automobiles and energy storage systems, especially in the event of accidents such as collision or tilting, to effectively prevent battery damage and avoid safety hazards.

[0004] Existing structural optimization relies on single performance improvement under static load assumption, lacks multi-factor quantitative processing of structure performance, weight and material distribution under dynamic collision conditions, and is prone to abnormal deformation of local structure in a specific impact direction. Cross-sectional size and rigidity coefficient of longitudinal beam and cross beam lack fine node displacement and residual deformation analysis, leading to failure to identify local stress hotspots in advance. Weight distribution often aims to reduce overall weight as the main target, without linking weight and unit energy absorption and peak principal stress for evaluation, resulting in low energy absorption efficiency in high weight areas. Structural layout optimization is mostly for overall uniform treatment, without accurately selecting low-efficiency and high-efficiency energy absorption units during material removal and reinforcement, causing uneven resource allocation. Stress monitoring methods mostly rely on maximum value interception, lacking analysis of stress integration and sorting over the entire period, leading to insufficient identification and processing of high-risk nodes. SUMMARY

[0005] The purpose of the present application is to solve the problems existing in the prior art and to provide a simulation optimization design method and system for battery rack anti-rollover collision.

[0006] In order to achieve the above purpose, the present application adopts the following technical scheme: a simulation optimization design method for battery rack anti-rollover collision, comprising the following steps:

[0007] S1: based on the section height of the longitudinal beam, the thickness of the cross beam plate, the combination of the joint rigidity coefficient, input the parameters into the three-dimensional battery rack model containing the stiffened plate, set the node constraint, load the roll angle and impact speed load, calculate the displacement field of the longitudinal beam and cross beam, extract the residual deformation and node displacement, and generate the deformation parameter table;

[0008] S2: based on the deformation parameter table, calculate the energy absorption per unit volume of the longitudinal beam and cross beam, extract the main stress peak value pair, calculate the overall weight combined with the density volume of the stiffened plate support frame, compare the performance and weight classification, use the multi-objective non-dominated sorting genetic algorithm to exchange the section size and adjust the joint rigidity coefficient, and generate the trade-off set;

[0009] S3: based on the trade-off set, construct the finite element layout of the longitudinal beam and cross beam stiffened plate, use the variable density method to calculate the unit energy absorption, remove the units below the average value of the specified multiple, increase the section size of the high energy absorption units at the stress concentration and adjust the stress path, and generate the layout model;

[0010] S4: based on the layout model, record the stress of the longitudinal beam and cross beam node at a set interval in the collision simulation, sort the integral, filter the high stress nodes and locate the geometric position in the stiffened plate support frame model, and generate the stress distribution table;

[0011] S5: based on the stress distribution table, extract the local section size of the corresponding high stress node, the local plate thickness of the stiffened plate, and the connection rigidity adjustment value, increase the section height and plate thickness at the low value position and improve the connection rigidity, and write the modified parameters into the optimized layout model to generate the anti-impact parameter set.

[0012] As a further scheme of the application, the deformation parameter table includes the longitudinal beam deformation, the cross beam deformation, and the node displacement value, the trade-off set includes the performance preferred scheme, the weight preferred scheme, and the performance weight balance scheme, the layout model includes the longitudinal beam structure distribution, the cross beam structure distribution, and the stiffened plate distribution form, and the stress distribution table includes the node number, the stress integral value, and the node geometric position.

[0013] As a further scheme of the application, the specific steps for generating the deformation parameter table are:

[0014] Based on the section height of the longitudinal beam, the thickness of the cross beam plate, and the combination of the joint rigidity coefficient, the longitudinal beam section edge line and the cross beam plate profile are established in the three-dimensional space to establish coordinate constraints and input the section parameter value, the stiffened plate nodes are sequentially connected to the longitudinal beam and cross beam node positions and the input node rigidity coefficient is fixed, the roll angle impact load is applied to the outside of the overall battery rack model and is distributed to the longitudinal beam and cross beam contact area, and a load constraint model is generated;

[0015] Based on the load constraint model, the displacement of each unit node of the longitudinal beam and the transverse beam is calculated in the spatial displacement component under the loaded state, and the change is recorded, the residual deformation of the middle part of the longitudinal beam and the end part of the transverse beam is measured, and the displacement value of the connecting node is matched, the deformation parameters table is generated by arranging the deformation of the longitudinal beam, the deformation of the transverse beam and the displacement value of the node according to the structure component category.

[0016] As a further scheme of the present application, the specific steps for generating the trade-off set are:

[0017] Based on the deformation parameter table, the energy absorption value of each unit of the longitudinal beam and the transverse beam is divided by the unit volume one by one, and the result is recorded, the principal stress peak value of the corresponding unit is extracted and matched with the energy absorption value and stored in the list, and the energy absorption stress pairing table is generated;

[0018] Based on the energy absorption stress pairing table, the density and volume of the stiffened plate and the support frame are multiplied and summed to obtain the overall weight data, and the performance weight ratio list is generated by combining the weight value with the corresponding performance value to generate the performance weight ratio table;

[0019] Based on the performance weight ratio table, the multi-objective non-dominated sorting genetic algorithm is used to divide the non-dominated relationship and measure the crowding degree of the double-objective scheme set, the cross section height of the longitudinal beam and the thickness of the transverse beam between each scheme are exchanged, the node rigidity coefficient is adjusted according to the set disturbance and recorded, and the trade-off set is generated.

[0020] As a further scheme of the present application, the multi-objective non-dominated sorting genetic algorithm, first, each scheme in the performance weight ratio table is regarded as an individual, and the performance index and the weight index are used as the double-target evaluation function, the non-dominated relationship is used to sort all individuals, the individuals which are not inferior to other individuals and are superior in at least one index are divided into the first non-dominated layer, and the sorting is performed layer by layer until all individuals are completed, the crowding distance calculation is performed within each non-dominated layer, each individual is sorted according to the target value size in each target dimension, and the crowding degree is calculated by normalizing the difference between adjacent individual target values to maintain the diversity of the solution, then in the selection operation, the non-dominated layer and the crowding degree are used to select the individual into the cross and mutation stage, in the cross operation, the longitudinal beam cross section height and the transverse beam thickness of different individuals are exchanged and combined, in the mutation operation, the node rigidity coefficient is subjected to random disturbance within a set range, thereby generating a new generation of candidate schemes, finally, the newly generated individuals and the parent generation are combined, and the non-dominated sorting and crowding degree measurement are repeated until the convergence condition is met, and the trade-off set between performance and weight is obtained.

[0021] As a further scheme of the present application, the specific steps for generating the layout model are:

[0022] Based on the trade-off set, the node connection relationship of the stringer-beam stiffened plate in three-dimensional space is constructed and a finite element grid is generated, each unit is assigned a material density and volume value, and a structure grid model is generated;

[0023] Based on the structure grid model, a variable density method is used to calculate the energy absorption value per unit volume of each unit, units with a value less than the average value set multiple are added to the removal list, and units in the stress concentration area are added to the reinforcement list to generate a unit screening list;

[0024] Based on the unit screening list, the units in the removal list are deleted from the model, the cross-sectional size is increased at the corresponding position of the reinforcement list, the stress path of the stringer and beam nodes is adjusted to maintain structural continuous force transmission, and a layout model is generated.

[0025] As a further scheme of the present application, the variable density method first defines the material density of each unit as a continuous variable between 0 and 1, which has a physical meaning of the retention degree of unit material in the structure, then calculates the strain energy of each unit under unit load, and obtains the unit volume energy absorption value by dividing the strain energy by the unit volume, which is used as an index to evaluate the contribution of the unit structure, then sets the target function as the total energy absorption of the global structure under unit mass maximization, introduces the material consumption constraint, establishes an optimization problem with the density variable as the design variable, updates the density of each unit according to the sensitivity analysis result, combines the exponential relationship between density and energy absorption value to gradually approach the adaptive distribution, and identifies the units with a density lower than the average value set multiple threshold as low-contribution units and adds them to the removal list, and identifies the units with a stress value in the concentration interval as high-contribution units and adds them to the reinforcement list, finally performs unit removal and local section enhancement operations according to the lists, and readjusts the stress path of the stringer and beam nodes to ensure continuous load transmission, to obtain the optimized layout model.

[0026] As a further scheme of the present application, the specific steps for generating the stress distribution table are:

[0027] Based on the layout model, the instantaneous stress values of each node of the stringer and beam in the collision simulation are extracted at a set time interval and stored in a data table, the stress values of the same node at all time intervals are accumulated into a single integral value, and the integral value and the node number are stored in one-to-one correspondence to generate a node stress integral table;

[0028] Based on the node stress integral table, the integral values of all nodes are sorted according to the numerical value and the nodes with high integral values are screened out, the node number is located on the geometric model of the stiffened plate and the support frame, and the three-dimensional coordinate information is recorded in the table to generate a stress distribution table.

[0029] As a further scheme of the present application, the specific steps for generating the impact resistance parameter set are:

[0030] Based on the stress distribution table, the local section size data of each high stress node connected in the geometric model, the local plate thickness data of the stiffening plate and the rigidity value of the connected node are obtained, classified according to the node number, and a local parameter list is generated;

[0031] Based on the local parameter list, the local section size at the position in the low value range is increased by a preset height increment, the local plate thickness at the position in the low value range is increased by a preset thickness increment, the connected rigidity value at the position in the low value range is increased by a preset rigidity increment, and all the adjusted data are written into the optimized layout model to generate an anti-impact parameter set.

[0032] A simulation optimization design system for battery rack anti-roll collision is used to perform the simulation optimization design method for battery rack anti-roll collision, and the system comprises:

[0033] The structural load determination module: based on the combination of the longitudinal beam section height, the transverse beam plate thickness and the connection node rigidity coefficient, the longitudinal beam section and the contour, the transverse beam plate thickness and the boundary, and the node rigidity and the contact position of the stiffening plate are matched, the roll angle and the impact speed load are set, the residual deformation of the longitudinal beam section and the end of the transverse beam is calculated, and the structural deformation data table is generated.

[0034] The performance trade-off generation module: based on the structural deformation data table, the energy absorption value and the volume of the longitudinal beam and the transverse beam unit are divided and matched with the principal stress peak value, the density and volume of the stiffening plate and the support frame are calculated, and the weight is accumulated, which is combined with the energy absorption value to form the performance weight ratio, which is input into the multi-objective non-dominated sorting genetic algorithm to perform grade division, crowding degree calculation, section size exchange and node rigidity disturbance to generate a performance weight trade-off set.

[0035] The structural layout optimization module: based on the performance weight trade-off set, the longitudinal beam, the transverse beam and the stiffening plate grid are established, the material density and volume are assigned, the unit volume energy absorption is calculated by the variable density method, the units below the average multiple threshold value are recorded as a removal list, the high stress concentration units are recorded as a strengthening list, the unit is removed and the section is increased to generate a strengthening layout model.

[0036] The high stress node screening module: based on the strengthening layout model, the node stress is recorded and the integral is accumulated, the node number is screened in descending order, and the high stress node distribution table is generated according to the spatial position.

[0037] The anti-impact parameter updating module: based on the high stress node distribution table, the local section height, the plate thickness and the connection rigidity are extracted, the increment is increased at the low value and updated to the strengthening layout model to generate an anti-impact structural parameter set.

[0038] Compared with the prior art, the application has the advantages and positive effects that:

[0039] In the application, the residual deformation of the middle section and the end section is paired with the node displacement value item by item to establish a deformation parameter table, so that the stress and deformation relationship of the initial state of the structure forms a quantifiable basis in the overall space.

[0040] In the application, the energy absorption values of the longitudinal beam and the transverse beam unit are normalized by volume and paired with the principal stress peak value one by one, and the weight of the stiffened plate and the supporting frame is calculated, and the double target data of performance and weight are input into the multi-objective non-dominated sorting genetic algorithm to perform scheme level division, crowding degree calculation, cross section size exchange and node rigidity coefficient disturbance, so that the structure realizes weight balance under the premise of maintaining stress performance.

[0041] In the application, the variable density method is used to remove low energy absorption units and increase the section and adjust the stress path of high stress concentration units, so that the material distribution is concentrated in the high-efficiency energy absorption area, and further in the collision simulation, the node stress is recorded in time and integrated and sorted, the high stress node geometric position is located, and the local section, plate thickness and connection rigidity of the node are incrementally adjusted, so that the load bearing path continuity and impact resistance of the structure are optimized under the driving of quantitative data. BRIEF DESCRIPTION OF DRAWINGS

[0042] Figure 1 It is a working flowchart of the application;

[0043] Figure 2 It is a system flowchart of the application. DETAILED DESCRIPTION

[0044] In order to make the purpose, technical scheme and advantages of the application clearer, the application will be further described in detail below in combination with the drawings and examples. It should be understood that the specific examples described herein are only used to explain the application and do not limit the application.

[0045] Example 1: please refer to Figure 1 The application provides a technical scheme: a simulation optimization design method for battery rack anti-roll collision, comprising the following steps:

[0046] S1: based on the combination of longitudinal beam section height, transverse beam plate thickness and connection node rigidity coefficient, input the parameters into the three-dimensional battery rack model with stiffened plate, set the node constraint, load the roll angle and impact speed load, calculate the longitudinal beam transverse beam displacement field and extract the residual deformation and node displacement, and generate a deformation parameter table;

[0047] S2: Based on the deformation parameter table, calculate the energy absorption per unit volume of longitudinal beam and extract the principal stress peak pairing, calculate the overall weight combined with the density volume of the stiffened plate support frame, compare the performance and weight classification, use multi-objective non-dominated sorting genetic algorithm to exchange cross-sectional size and adjust the node rigidity coefficient, and generate a trade-off set;

[0048] S3: Based on the trade-off set, construct the finite element layout of longitudinal beam and transverse beam stiffened plate, use variable density method to calculate the unit energy absorption, remove the units below the average value of the specified multiple, increase the cross-sectional size of the high energy absorption unit at the stress concentration and adjust the stress path, and generate a layout model;

[0049] S4: Based on the layout model, record the stress of longitudinal beam and transverse beam node at the set interval in the crash simulation, sort the integral value, select the high stress node and locate the geometric position in the stiffened plate support frame model, and generate a stress distribution table;

[0050] S5: Based on the stress distribution table, extract the local cross-sectional size of the corresponding high stress node, the local plate thickness of the stiffened plate, and the connection rigidity adjustment value, increase the cross-sectional height and plate thickness of the low value position and improve the connection rigidity, and write the modified parameters into the optimized layout model to generate an anti-impact parameter set.

[0051] The deformation parameter table includes longitudinal beam deformation, transverse beam deformation, and node displacement value, the trade-off set includes performance optimization scheme, weight optimization scheme, and performance weight balance scheme, the layout model includes longitudinal beam structure distribution, transverse beam structure distribution, and stiffened plate distribution form, the stress distribution table includes node number, stress integral value, and node geometric position, and the anti-impact parameter set includes local cross-sectional size value, local plate thickness value, and local connection rigidity value.

[0052] The specific steps for generating the deformation parameter table are as follows:

[0053] Based on the combination of longitudinal beam cross-sectional height, transverse beam plate thickness, and connection node rigidity coefficient, the longitudinal beam cross-sectional edge line and the transverse beam plate profile are established in the three-dimensional space to establish coordinate constraints and record the cross-sectional parameter value, the stiffened plate nodes are sequentially connected to the longitudinal beam and transverse beam node positions and the input node rigidity coefficient is fixed, the roll angle impact load is applied to the outside of the overall battery rack model and is distributed to the longitudinal beam and transverse beam contact area, and a load constraint model is generated;

[0054] Based on the load constraint model, calculate the node displacement of each unit of longitudinal beam and transverse beam under the loaded state according to the spatial displacement component and record the change, measure the residual deformation of the middle part of longitudinal beam and the end part of transverse beam, and pair with the connection node displacement value, sort by structure component type into longitudinal beam deformation, transverse beam deformation and node displacement value, and generate a deformation parameter table;

[0055] Based on the combination of the height of the longitudinal beam section, the thickness of the cross beam plate, and the rigidity coefficient of the connection joint, a three-dimensional finite element modeling method is adopted. In the modeling software, the Solid186 element type is called and the integration points are set to 2x2x2. The material elastic modulus is input as 2.05x10^5 MPa, the Poisson's ratio is 0.3, and the density is 7.85x10^-6 kg / mm³. The grid subdivision is performed on the edge line of the longitudinal beam section and the profile of the cross beam plate in three-dimensional space. In the global coordinate system, fixed constraint conditions are applied in the X, Y, and Z directions. The section parameter values are entered and the nodes are established with a size precision of 0.01 mm. The stiffened plate nodes are sequentially connected to the corresponding node positions of the longitudinal beam and the cross beam. In the node attribute table, the rigidity coefficient is set to 1200 N / mm and specified as a rigid connection. The impact load with a roll angle of 25° is distributed to the contact area of the longitudinal beam and the cross beam in time sequence. The time integration step is set to 1x10^-6 s. The central difference method is used for load time history distribution. The damping coefficient is set to 0.02. The load constraint model is generated.

[0056] Based on the load constraint model, an explicit dynamic displacement solving method is adopted. In the global coordinate system, the instantaneous load value of each loading point is read according to the time sequence and associated with the node position to calculate the displacement. In the loaded state, the X, Y, and Z direction displacement components of each element node of the longitudinal beam and the cross beam are decomposed and recorded. The absolute value of the middle node displacement of the longitudinal beam is taken as the residual deformation variable. The absolute value of the end node displacement of the cross beam is taken as the residual deformation variable. The residual deformation variables of the longitudinal beam and the cross beam are paired with the node displacement values at the same position one by one. According to the type of the structure component, the data is classified into longitudinal beam deformation, cross beam deformation, and node displacement value. The deformation parameter table is generated.

[0057] The specific steps for generating the trade-off set are as follows:

[0058] Based on the deformation parameter table, the energy absorption value of each element of the longitudinal beam and the cross beam is divided by the element volume one by one and the result is recorded. The principal stress peak value of the corresponding element is extracted and paired with the energy absorption value and stored in the list. The energy absorption stress pairing table is generated.

[0059] Based on the energy absorption stress pairing table, the density and volume of the stiffened plate and the support frame are multiplied and summed to obtain the overall weight data. The weight value is combined with the corresponding performance value to generate the performance weight ratio list. The performance weight ratio table is generated.

[0060] Based on the performance weight ratio table, the multi-objective non-dominated sorting genetic algorithm is used to divide the non-dominated relationship and measure the crowding degree of the double-objective scheme set. The longitudinal beam section height and the cross beam plate thickness are exchanged. The node rigidity coefficient is adjusted by a certain disturbance amount and recorded. The trade-off set is generated.

[0061] Based on the deformation parameter table, the unit energy normalization processing method is adopted, and the absorbed energy value field and the volume value field of each unit of the longitudinal beam and the transverse beam are called in the calculation program, the absorbed energy value is divided by the volume value and recorded in the result vector in floating point format, the principal stress value extraction instruction is called to scan the peak value of the complete stress sequence of each unit and return the numerical result, the principal stress peak value of the longitudinal beam and the transverse beam is indexed and paired with the corresponding unit volume energy absorption value according to the unit number, all paired data is written into a two-dimensional array and sorted by row number, stored in a special list variable, and an energy absorption stress pairing table is generated;

[0062] Based on the energy absorption stress pairing table, the structural weight calculation flow is used to read the density parameters and geometric volume parameters of the stiffened plate and the support frame, multiply them item by item, and record them in double precision format, then call the matrix summation instruction to sum all the products to get the overall weight data, match the weight data with the unit volume energy absorption value in the energy absorption stress pairing table by unit matching method, write each group of combined results into the performance weight ratio vector, arrange the performance weight ratio list in turn, output to the file buffer and organize into the performance weight ratio table;

[0063] Based on the performance weight ratio table, a multi-objective non-dominated sorting genetic algorithm is used, the population size is set to 100, the maximum iteration number is set to 200, the crossover probability is set to 0.9, the mutation probability is set to 0.05, the longitudinal beam energy absorption performance and the structure weight are taken as double target inputs, the scheme set in the performance weight ratio table is imported into the algorithm population, the non-dominated relationship judgment step is executed to divide the population into levels, the crowdedness calculation instruction is called inside each level to generate a crowdedness array according to the difference between the two target values, the level priority and crowdedness priority rules are used to select and reserve individuals, the longitudinal beam section height value and the transverse beam plate thickness value of the reserved individuals are numerically exchanged in the parameter table according to the pairing index, the node rigidity coefficient is randomly generated in the range of the set disturbance amount ±5% and added to the original value to complete the update, the scheme parameters after each update are recorded in the population storage array, and the trade-off set is generated after all iterations are completed.

[0064] The multi-objective non-dominated sorting genetic algorithm first treats each scheme in the performance-weight ratio table as an individual, and uses performance and weight indicators as bi-objective evaluation functions. Based on the non-dominated relationship, all individuals are sorted hierarchically. Individuals that are not inferior to other individuals and are superior in at least one indicator are divided into the first non-dominated layer. This process is repeated layer by layer until all individuals are stratified. Within each non-dominated layer, crowding distance is calculated. Each individual is sorted according to the target value in each objective dimension, and crowding is calculated by normalizing the difference in target values ​​between adjacent individuals to maintain solution diversity. Then, in the selection operation, individuals are selected based on the non-dominated layer and crowding to enter the crossover and mutation stage. In the crossover operation, the longitudinal beam section height and transverse beam plate thickness of different individuals are interchanged and combined. In the mutation operation, random perturbations within a set range are applied to the node stiffness coefficient to generate a new generation of candidate schemes. Finally, the newly generated individuals are merged with the parent generation, and the non-dominated sorting and crowding measurement are repeated until the convergence condition is met, resulting in a set of trade-offs between performance and weight.

[0065] A multi-objective non-dominated sorting genetic algorithm, according to the formula:

[0066] ;

[0067] in: Indicates the first Improved crowding distance value of individual battery rack structure schemes in dual-objective optimization for roll-over collision resistance. This indicates the sequence number of the candidate battery rack structure currently participating in the non-dominated sorting. Indicates the objective function number. Indicates the first The weighting coefficients of each objective. Indicates the height of the longitudinal beam section of the battery rack. This indicates the thickness of the battery rack crossbeam plate. Indicates the stiffness coefficient of the battery rack node. Indicates the buckling sensitivity coefficient of the battery holder material. Indicates the first The first target battery rack The objective function values ​​calculated for the subsequent schemes of the given scheme, under the conditions of longitudinal beam section height, transverse beam plate thickness, nodal stiffness coefficient, and material buckling sensitivity coefficient. Indicates the first The first target battery rack The objective function values ​​calculated for each of the preceding neighboring schemes, given the values ​​of the longitudinal beam section height, transverse beam thickness, nodal stiffness coefficient, and material buckling sensitivity coefficient. This indicates that all candidate battery rack schemes are in the... The maximum objective function value under each objective. representing the minimum objective function value of all candidate battery rack schemes under the first objective;

[0068] The execution process is as follows: firstly, a double-objective optimization model containing performance-weight ratio maximization and material consumption minimization is established and a candidate scheme set is initialized, the initial values and allowable variation ranges of the longitudinal beam section height and the cross beam plate thickness are determined for each candidate scheme, the influence of the parameters on the overall stiffness and mass of the battery rack under the action of the collision load is calculated through the finite element model, then the initial values and perturbation ranges of the node rigidity coefficients are determined and adjusted between adjacent schemes to simulate the influence of different node constraint modes on energy absorption and deformation mode, the values of the material buckling sensitivity coefficients are determined and the parameter reflecting the stability of the material is obtained by comparing the ratio of the critical buckling stress of the material to the design stress, after the above four key parameter settings, the two objective function values of each candidate scheme are calculated, the target values of the next adjacent scheme under the conditions of the longitudinal beam section height, the cross beam plate thickness, the node rigidity coefficient and the material buckling sensitivity coefficient and the target values of the previous adjacent scheme are obtained respectively, the maximum value and the minimum value of the objective function of all schemes are further searched, then the variances of the performance-weight ratio objective and the material consumption objective are calculated according to the normalized target value sequence and the inverse variance ratio method is used to determine the double-objective weight coefficient , the determined weight coefficient and the adjacent scheme difference value of each objective are substituted into the formula to complete the improvement crowded distance calculation of each candidate scheme, according to the crowded distance from large to small, the representative non-dominated solutions are selected and reserved into the next generation search, and finally a trade-off solution set containing the optimized balance of the anti-roll stiffness and the collision energy absorption performance is generated.

[0069] The specific steps of generating the layout model are as follows:

[0070] Based on the trade-off set, the node connection relationship of the longitudinal beam and cross beam stiffened plate in three-dimensional space is constructed and the finite element grid is generated, the material density and volume value of each element are assigned, and the structure grid model is generated;

[0071] Based on the structure grid model, the energy absorption value per unit volume of each element is calculated by using the variable density method, the elements with less than a certain multiple of the average value are added to the removal list according to the number, the elements in the stress concentration area are added to the strengthening list, and the element screening list is generated;

[0072] Based on the unit screening list, the units in the list are removed from the model, the cross-sectional size is increased at the corresponding position of the reinforcement list, the stress path of the longitudinal beam and cross beam node is adjusted to keep the structure continuous force transmission, and the layout model is generated;

[0073] Based on the trade-off set, a three-dimensional finite element grid generation method is adopted, the Solid186 unit type instruction is called in the finite element modeling program, the cell edge length is set to 10mm, the node automatic numbering mode is selected as the sequential incremental mode, the three-dimensional coordinates of the longitudinal beam, cross beam and stiffened plate are input into the grid division module, the node connection and cell surface generation operation is performed, the density value 7.85×10^-6 kg / mm³ and the volume value cubic millimeter calculated according to the unit geometry are input for each unit in the unit attribute definition, saved as a list of node and unit corresponding data, and a structure grid model is generated;

[0074] Based on the structure grid model, a variable density method is adopted, the initial density variable is set to 1.0 in the topology optimization calculation module, the penalty factor is 3, the convergence criterion threshold is 1×10^-4, and the upper limit of the iteration number is set to 150, the unit volume energy absorption value of each unit is calculated and written into the energy absorption array, the average of all unit energy absorption values is multiplied by the set multiple coefficient 1.2 as the threshold, the unit number of the unit below the threshold is written into the removal list file, and the unit number of the unit with stress value greater than 200MPa in the stress result is written into the reinforcement list file, and a unit screening list is generated;

[0075] Based on the unit screening list, a grid structure editing method is used, the batch delete instruction is called in the geometry editing module to delete the corresponding unit according to the removal list, the cross-sectional height parameters of the longitudinal beam and cross beam at the reinforcement position are increased by 5mm, and the connection order of the stress path is updated in the node topology relationship table to ensure the continuity of the structure stress transmission, the adjusted grid data is saved to a new geometry model file, and a layout model is generated.

[0076] The variable density method first defines the material density of each unit as a continuous variable between 0 and 1, which represents the retention degree of the unit material in the structure, then calculates the strain energy of each unit under the action of unit load, and obtains the unit volume energy absorption value by dividing the strain energy by the unit volume, which is used as an index to evaluate the contribution degree of the unit structure, then sets the objective function as the maximization of the total energy absorption of the global structure per unit mass, introduces the material consumption constraint, and establishes the optimization problem with the density variable as the design variable, updates the density of each unit according to the sensitivity analysis results, combines the exponential relationship between density and energy absorption value to gradually approach the adaptive distribution, and identifies the units with density lower than the average value by a certain threshold as low-contribution units and adds them to the removal list, identifies the units with stress values in the concentration interval as high-contribution units and adds them to the reinforcement list, finally executes the unit removal and local section reinforcement operations according to the lists, and adjusts the stress path of the longitudinal beam and cross beam nodes to ensure continuous load transfer, and obtains the optimized layout model;

[0077] The variable density method is as follows:

[0078] ;

[0079] Wherein: represents the improved unit volume energy absorption value, represents the shape complexity correction coefficient, the effective volume of the unit calculated later, represents the geometric volume of the battery rack structure unit, represents the geometric shape complexity correction coefficient of the unit, represents the combined material yield ratio , structural ductility ratio and strain rate sensitivity coefficient corrected stress component of the unit, represents the combined material yield ratio , structural ductility ratio and strain rate sensitivity coefficient corrected strain component of the unit, represents the ratio of the yield strength to the ultimate strength of the battery rack material, represents the ratio of the ultimate strain to the yield strain of the battery rack structure, represents the strain rate sensitivity coefficient of the battery rack material, represents the collision direction coupling coefficient, represents the unit volume element;

[0080] Execution process: first establish the finite element grid model of the battery rack structure and determine the geometric characteristics and node topological relationship of each unit, obtain the geometric volume The ratio of the surface area of the recycling unit to the surface area of a standard cube of the same volume is used to obtain a geometric complexity correction coefficient The corrected effective volume is then calculated to reflect the influence of complex shapes on energy absorption efficiency, and then finite element simulation is performed under different anti-rollover crash conditions and the stress components and strain components of each unit are extracted The yield ratio is determined by combining material performance experimental data to determine the ratio of material yield strength to ultimate strength The ductility ratio is calculated by calculating the ratio of ultimate strain to yield strain The stress-strain curve changes are extracted in the crash simulation at different speeds to determine the strain rate sensitivity coefficient , , The original numerical values are corrected by substituting the stress components and strain components into the stress component and strain component expressions, thereby accurately reflecting the true mechanical behavior of the material under impact load, and then the crash direction coupling coefficient is calculated according to the energy absorption ratio of the battery rack structure in different direction crashes and introduced into the formula as an energy scaling factor, the modified stress-strain product is integrated, and the unit volume is numerically integrated to obtain the total energy absorption of the unit, and finally the total energy absorption is divided by the modified effective volume to obtain the improved unit volume energy absorption value as the core indicator for determining whether the unit enters the removal list and the reinforcement list.

[0081] The specific steps for generating the stress distribution table are as follows:

[0082] Based on the layout model, the instantaneous stress values of each node of the longitudinal beam and the cross beam are extracted at a set time interval during the crash simulation and stored in the data table, the stress values of the same node at all time intervals are accumulated into a single integral value, and the integral value is stored in one-to-one correspondence with the node number, generating a node stress integral table;

[0083] Based on the node stress integral table, the integral values of all nodes are sorted according to the numerical value and the nodes with high integral values are selected, the node numbers are located on the geometric model of the stiffened plate and the support frame, and the three-dimensional coordinate information is recorded in the table, generating a stress distribution table;

[0084] Based on the layout model, an explicit dynamic stress collection method is adopted, the time step is set to 1*10^-6s in the collision simulation calculation program, the central difference format is set for integral method, the node stress extraction instruction is called to read the instantaneous stress value of all nodes of the longitudinal beam and the cross beam at the set time interval, and the output format is set to three column data form, including node number, time, stress value, the collected instantaneous stress value is written into the memory data table in real time, in the data processing module, all stress values of the same node number at all sampling times are added one by one to obtain single integral value and keep double precision data type, the integral value and the corresponding node number are established in one-to-one correspondence according to row, and the corresponding information is stored in the two-dimensional array data structure to generate the node stress integral table;

[0085] Based on the node stress integral table, the sorting and coordinate mapping method is used, the quicksort algorithm is called in the data processing program to sort all node integral values according to the value size, the screening ratio parameter is set to the node number of the first 10%, the corresponding node number is extracted according to the sorting result, the data record consistent with the node number is searched in the geometric model database, the number is associated with the geometric model elements of the stiffened plate and the support frame, the coordinate reading instruction is called to obtain the X, Y, Z three-dimensional space coordinate values of the node, and the node number and the coordinate value are written into the table file in column corresponding mode to generate the stress distribution table.

[0086] The specific steps of generating the impact resistance parameter set are:

[0087] Based on the stress distribution table, the local section size data of each high stress node connected in the geometric model, the local plate thickness data of the stiffened plate and the rigidity value of the connected node are obtained, which are classified according to the node number to generate the local parameter list;

[0088] Based on the local parameter list, the local section size in the low value range is increased by a preset height increment, the local plate thickness in the low value range is increased by a preset thickness increment, the connected rigidity value in the low value range is increased by a preset rigidity increment, and all the adjusted data are written into the optimized layout model to generate the impact resistance parameter set;

[0089] Based on the stress distribution table, a geometric parameter extraction method is adopted, the node retrieval instruction is called in the geometric modeling program to locate the node element in the geometric model according to the high stress node number, the section height and width values are extracted by calling the section attribute reading command of the component section connected with the node, and the two decimal precision is kept, the plate thickness value is extracted by calling the thickness attribute reading command of the stiffened plate element, the corresponding connection rigidity value in the node connection relationship table is extracted from the attribute database as a numerical parameter, the section size, plate thickness and rigidity coefficient are classified according to the node number in the two-dimensional array, and the node serial number identifier is added in the array name list to generate the local parameter list.

[0090] Based on the local parameter list, the parameter threshold comparison and numerical updating method is used, the cross section height increment is set to 5mm, the plate thickness increment is set to 2mm, the rigidity coefficient increment is set to 100N / mm, the cross section size value of each node in the local parameter list is read and compared with the preset height threshold value, when it is found that it is lower than the threshold value, the updating instruction is called to increase the specified increment, the plate thickness data is read and compared with the preset thickness threshold value, when it is lower than the threshold value, the updating instruction is called to increase the specified thickness increment, the rigidity coefficient is read and compared with the preset rigidity threshold value, when it is lower than the threshold value, the updating instruction is called to increase the specified rigidity increment, all the updated data is written into the geometry and attribute data table of the optimized layout model, and the impact resistance parameter set is generated.

[0091] Please refer to Figure 2 A simulation optimization design system for battery rack anti-roll collision, the simulation optimization design system for battery rack anti-roll collision is used to execute the simulation optimization design method for battery rack anti-roll collision, the system comprises:

[0092] The structure load measurement module: based on the combination of longitudinal beam cross section height, transverse beam plate thickness and connecting node rigidity, the longitudinal beam cross section and profile, the transverse beam plate thickness and boundary, the node rigidity and stiffened plate contact position are matched, the roll angle and impact speed load are set, the residual deformation of the middle section of the longitudinal beam and the end of the transverse beam is calculated, and the structure deformation data table is generated;

[0093] The performance trade-off generation module: based on the structure deformation data table, the energy absorption value and volume of the longitudinal beam and transverse beam unit are divided and matched with the principal stress peak value, the density and volume of the stiffened plate and support frame are calculated and added to the weight, and the energy absorption value is combined to form the performance weight ratio, the multi-objective non-dominated sorting genetic algorithm is input, the grade division, crowding degree calculation, cross section size exchange and node rigidity disturbance are executed, and the performance weight trade-off set is generated;

[0094] The structure layout optimization module: based on the performance weight trade-off set, the longitudinal beam, transverse beam and stiffened plate grid are established, the material density and volume are assigned, the energy absorption per unit volume is calculated by the variable density method, the units below the average multiple threshold value are recorded as the removal list, the high stress concentration units are recorded as the strengthening list, the unit removal and cross section height increase are executed, and the enhanced layout model is generated;

[0095] The high stress node screening module: based on the enhanced layout model, the node stress is recorded and the integral is added, the node number is screened in descending order, and the high stress node distribution table is generated according to the spatial position;

[0096] The impact resistance parameter updating module: based on the high stress node distribution table, the local cross section height, plate thickness and connecting rigidity are extracted, the increment is increased at the low value and updated to the enhanced layout model, and the impact resistance structure parameter set is generated.

[0097] The above merely describes the preferred embodiments of the present application, but does not limit the present application in other forms. Any person skilled in the art can make changes or modifications to the above disclosed technical contents into equivalent embodiments with equivalent changes, and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present application without departing from the technical solution content of the present application still falls within the protection scope of the present application.

Claims

1. A simulation optimization design method for battery rack anti-roll collision, characterized in that, The method comprises the following steps: S1: based on the section height of the longitudinal beam, the thickness of the cross beam plate, and the combination of the rigidity coefficient of the connecting node, input parameters into a three-dimensional battery rack model with a stiffened plate, set node constraints, load lateral inclination angle and impact speed load, calculate the displacement field of the longitudinal beam and cross beam, extract residual deformation and node displacement, and generate a deformation parameter table; S2: based on the deformation parameter table, calculate the energy absorption per unit volume of the longitudinal beam and cross beam, extract the main stress peak value pair, calculate the overall weight combined with the density volume of the stiffened plate support frame, compare the performance and weight classification, use a multi-objective non-dominated sorting genetic algorithm to exchange section sizes and adjust the rigidity coefficient of the node, and generate a trade-off set; S3: based on the trade-off set, construct a finite element layout of the longitudinal beam and cross beam stiffened plate, use the variable density method to calculate the unit energy absorption, remove the units below the average value of a specified multiple, increase the section size of the high energy absorption unit at the stress concentration and adjust the stress path, and generate a layout model; S4: based on the layout model, record the stress of the longitudinal beam and cross beam node at a set interval in the collision simulation, sort the accumulated integral, screen high stress nodes and locate the geometric position in the stiffened plate support frame model, and generate a stress distribution table; S5: based on the stress distribution table, extract the local section size of the corresponding high stress node, the local plate thickness of the stiffened plate, and the adjustment value of the connection rigidity, increase the section height and plate thickness at the low value position and improve the connection rigidity, write the modified parameters into the optimized layout model, and generate an anti-impact parameter set.

2. The simulation optimization design method of battery rack anti-roll crash according to claim 1, wherein, The deformation parameter table includes longitudinal beam deformation, cross beam deformation, and node displacement value, the trade-off set includes performance preferred scheme, weight preferred scheme, and performance weight balance scheme, the layout model includes longitudinal beam structure distribution, cross beam structure distribution, and stiffened plate distribution form, the stress distribution table includes node number, stress integral value, and node geometric position, and the anti-impact parameter set includes local section size value, local plate thickness value, and local connection rigidity value.

3. The simulation and optimization design method of battery pack anti-roll crash according to claim 1, wherein, The specific steps for generating the deformation parameter table are: Based on the combination of the section height of the longitudinal beam, the thickness of the cross beam plate, and the rigidity coefficient of the connecting node, establish coordinate constraints of the longitudinal beam section edge line and the cross beam plate profile in three-dimensional space and input the section parameter value, sequentially connect the stiffened plate nodes to the longitudinal beam and cross beam node positions and fix the input node rigidity coefficient, apply the lateral inclination angle impact load to the outside of the overall battery rack model and distribute it to the longitudinal beam and cross beam contact areas, and generate a load constraint model; Based on the load constraint model, calculate the node displacement of each unit of the longitudinal beam and cross beam under the loaded state according to the spatial displacement component and record the changes, determine the residual deformation of the middle part of the longitudinal beam and the end part of the cross beam, and pair it with the connection node displacement value, sort it according to the structure component category into longitudinal beam deformation, cross beam deformation, and node displacement value, and generate a deformation parameter table.

4. The simulation and optimization design method of battery pack anti-roll crash according to claim 1, wherein, The specific steps for generating the trade-off set are: Based on the deformation parameter table, divide the energy absorption value of each unit of the longitudinal beam and cross beam by the unit volume one by one and record the results, extract the main stress peak value corresponding to the unit and pair it with the energy absorption value and store it in a list, and generate an energy absorption stress pairing table; Based on the energy-absorbing stress pairing table, the overall weight data is obtained by multiplying and summing the density and volume of the stiffened plate and the support frame, and the performance weight ratio list is generated by combining the weight value with the corresponding performance value, and the performance weight ratio table is generated; Based on the performance weight ratio table, the non-dominant relationship division and the crowdedness measurement are performed on the double-target scheme set by the multi-objective non-dominant sorting genetic algorithm, the cross-section height of the longitudinal beam and the thickness of the cross beam plate are interchanged, the node rigidity coefficient is adjusted according to the set disturbance and recorded, and the trade-off set is generated.

5. The simulation and optimization design method of battery pack anti-roll crash according to claim 4, wherein, The multi-objective non-dominant sorting genetic algorithm first regards each scheme in the performance weight ratio table as an individual, takes the performance index and the weight index as the double-target evaluation function, performs hierarchical sorting on all individuals according to the non-dominant relationship, divides the individuals that are not inferior to other individuals and are superior in at least one index into the first non-dominant layer, and removes them layer by layer until all individuals are completed, calculates the crowdedness distance in each non-dominant layer, sorts each individual in each target dimension according to the target value, and calculates the crowdedness by normalizing the difference between adjacent individual target values to maintain the diversity of solutions, then selects individuals into the cross and mutation stage according to the non-dominant layer and the crowdedness in the selection operation, interchanges the cross-section height of the longitudinal beam and the thickness of the cross beam plate in the cross operation, and applies random disturbance to the node rigidity coefficient within the set range in the mutation operation, thereby generating a new generation of candidate schemes, and finally the newly generated individuals are combined with the parent generation to repeat the non-dominant sorting and crowdedness measurement until the convergence condition is met, and the trade-off set that balances the performance and weight is obtained.

6. The simulation and optimization design method of battery pack anti-roll crash according to claim 1, wherein, The specific steps of generating the layout model are: Based on the trade-off set, the node connection relationship of the longitudinal beam and the cross beam stiffened plate in the three-dimensional space is constructed, and the finite element grid is generated, each unit is assigned a material density and volume value, and the structure grid model is generated; Based on the structure grid model, the variable density method is used to calculate the energy-absorbing value per unit volume of each unit, units with less than a set multiple of the average value are added to the removal list, units in the stress concentration area are added to the reinforcement list, and the unit screening list is generated; Based on the unit screening list, the units in the removal list are deleted from the model, the cross-sectional size is increased at the corresponding position in the reinforcement list, the stress path of the longitudinal beam and the cross beam node is adjusted to maintain the continuous force transmission of the structure, and the layout model is generated.

7. The simulation and optimization design method of battery pack anti-roll crash according to claim 6, wherein, The variable density method first defines the material density of each unit as a continuous variable between 0 and 1, and the physical meaning is the retention degree of the unit material in the structure, then the strain energy of each unit is calculated under the action of unit load, and the unit volume energy absorption value is obtained by dividing the strain energy by the unit volume, which is used as an index to evaluate the contribution of the unit structure, then the objective function is set as the maximum total energy absorption per unit mass of the global structure, and the material consumption constraint is introduced, the optimization problem is established with the density variable as the design variable, the density of each unit is updated according to the sensitivity analysis result, the exponential relationship between density and energy absorption value is combined to gradually approach the adaptive distribution, and the unit whose density is lower than the average value by a set multiple threshold is identified as a low-contribution unit and added to the removal list, and the unit whose stress value is in the concentrated interval is identified as a high-contribution unit and added to the reinforcement list, finally, the unit removal and local section enhancement operations are performed according to the lists, and the stress path of the longitudinal beam and the cross beam node is adjusted to ensure continuous load transmission, and the optimized layout model is obtained.

8. The simulation and optimization design method of battery pack anti-roll crash according to claim 1, wherein, The specific steps for generating the stress distribution table are as follows: Based on the layout model, the instantaneous stress values of each node of the longitudinal beam and the cross beam are extracted at a set time interval in the collision simulation and stored in a data table, the stress values of the same node at all time intervals are accumulated into a single integral value, and the integral value and the node number are stored in one-to-one correspondence to generate a node stress integral table; Based on the node stress integral table, the integral values of all nodes are sorted according to the numerical value and the nodes with high integral values are selected, the node numbers are located on the geometric model of the stiffened plate and the support frame, and the three-dimensional coordinate information is recorded in the table to generate the stress distribution table.

9. The simulation and optimization design method of battery pack anti-roll crash according to claim 1, wherein, The specific steps for generating the anti-impact parameter set are as follows: Based on the stress distribution table, the local section size data of each high-stress node connected in the geometric model, the local plate thickness data of the stiffened plate, and the rigidity value of the connecting node are obtained, and are classified according to the node number to generate a local parameter list; Based on the local parameter list, the height of the local section size in the low value range is increased by a preset height increment, the thickness of the local plate thickness in the low value range is increased by a preset thickness increment, and the rigidity of the connecting rigidity value in the low value range is increased by a preset rigidity increment, and all the adjusted data are written into the optimized layout model to generate the anti-impact parameter set.

10. A simulation-optimization design system for battery pack anti-roll crash, characterized in that, The simulation optimization design method for battery rack anti-roll collision according to any one of claims 1-9, the system comprises: A structure load determination module: based on the combination of longitudinal beam section height, cross beam plate thickness and connecting node rigidity coefficient, the longitudinal beam section and contour, the cross beam plate thickness and boundary, and the node rigidity and stiffened plate contact position are paired, the roll angle and impact speed load are set, the residual deformation of the middle section of the longitudinal beam and the end of the cross beam is calculated, and the structure deformation data table is generated; Performance tradeoff generation module: based on the structure deformation data table, divide the energy absorption value of longitudinal beam and cross beam unit by volume and pair with the principal stress peak value, calculate the density of stiffened plate and support frame multiplied by volume and add the weight, combine with the energy absorption value to form the performance weight ratio, input the multi-objective non-dominated sorting genetic algorithm, perform rank division, crowding calculation, cross section size exchange and node rigidity disturbance, generate performance weight tradeoff set; Structure layout optimization module: based on the performance weight tradeoff set, establish longitudinal beam, cross beam and stiffened plate grid, assign material density and volume, calculate unit volume energy absorption by variable density method, record units below average multiple threshold value as removal list and high stress concentration units as reinforcement list, perform unit removal and cross section heightening, generate reinforcement layout model; High stress node screening module: based on the reinforcement layout model, record node stress and accumulate integral in time, screen node number in descending order, generate high stress node distribution table corresponding to spatial position; Impact resistance parameter update module: based on the high stress node distribution table, extract local cross section height, plate thickness and connection rigidity, increase increment at low value and update to reinforcement layout model, generate impact resistance structure parameter set.

Images