A method and system for designing a composite insulation layer of a container battery cabin

Abstract

The present application relates to the technical field of composite insulation layer, specifically to a container battery cabin composite insulation layer design method and system, in the present application, by establishing the initial parameter matrix composed of layer sequence coding, thickness vector, interface thermal resistance, honeycomb sandwich and corrugated geometric parameters and cabin wall heat exchange and radiation coefficient, the heat insulation performance evaluation can obtain accurate baseline under multi-dimensional input condition, through gradient boosting regression tree under the constraint of thermal conductivity, thermal diffusivity and propagation time, difference comparison and iteration screening is carried out, so that the heat insulation combination can gradually approach the optimal solution in the convergence process, avoid the deviation caused by the experience in the parameter selection in the traditional structure design, adopt simulated annealing algorithm to execute difference operation and convergence judgment on honeycomb size, wall thickness, height and corrugated geometric parameters, combined with finite element checking to ensure that the optimization process considers the requirements of thermodynamics and mechanics, so that the geometric parameters and thermal performance are synchronized to converge to stable combination.

Description

Technical Field

[0001] This invention relates to the field of composite insulation layer technology, and in particular to a composite insulation layer design method and system for a containerized battery compartment. Background Technology

[0002] The technology of composite isolation layer aims to construct a safety barrier for battery compartments through the functional combination of multi-layer materials, suppress the spread of thermal runaway of individual battery cells, block the transfer of heat and flame inside and outside the battery compartment, and enhance the impact resistance of the battery compartment through a metal load-bearing layer and a high-strength fiber layer.

[0003] The purpose of a composite isolation layer design method for a containerized battery compartment is to establish a multi-layer composite protection system in the battery compartment structure with flame retardant, heat insulation, impact resistance and corrosion resistance properties. When a battery cell experiences thermal runaway, it can effectively suppress the spread of heat and flame, reduce the risk of failure of adjacent battery cells and the entire battery compartment, and ensure the safety and reliability of the energy storage system under long-term operation and extreme environmental conditions by enhancing the structural strength and durability of the compartment.

[0004] Existing technologies have several drawbacks in practical applications. They rely too heavily on single-layer or simple multilayer structures for thermal insulation and flame retardancy. While material combinations can provide basic protection, heat can still spread rapidly along weak paths under high-intensity thermal shock, resulting in uneven thermal barrier performance. The lack of quantitative screening and optimization of thermal diffusion delay means that the temperature rise diffusion rate may still exceed the safety threshold when a single battery cell experiences thermal runaway, causing damage to adjacent cells. The structural impact resistance relies on the direct load-bearing capacity of the metal and fiber layers, lacking coordinated control of geometric parameters. This results in insufficient energy dissipation efficiency after impact, easily leading to stress concentration in localized areas and reduced durability. The corrosion resistance and long-term operational stability of the cabin mainly depend on the inherent properties of the materials, lacking systematic thickness ratio and parameter optimization, making it difficult to achieve a balance between lightweight and high strength. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the existing technology and to propose a composite isolation layer design method and system for containerized battery compartments.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a composite isolation layer design method for a containerized battery compartment, comprising the following steps:

[0007] S1: Based on the sequence coding, thickness vector, interface thermal resistance, honeycomb interlayer and corrugated geometric parameters, bulkhead heat transfer and radiation coefficients, matrix arrangement and merging are performed to obtain the initial parameter matrix of the thermal insulation structure.

[0008] S2: Based on the initial parameter matrix of the insulation structure, according to the thermal conductivity, thermal diffusivity and propagation time, a gradient boosting regression tree is used to combine the layer sequence and thickness and perform difference comparison to screen out qualified combinations. Then, the results are weighted and iterated repeatedly to obtain the set of predicted combinations for insulation performance.

[0009] S3: Based on the predicted thermal insulation performance combination set, the propagation time, equivalent thermal conductivity and areal density are used as the judgment conditions, and are compared with the threshold in turn. The combination that meets the conditions is retained and the combination that does not meet the conditions is eliminated to form a qualified thermal insulation performance set.

[0010] S4: Based on the set of qualified thermal insulation performance, set the honeycomb size, wall thickness, height and corrugation parameters, and use the simulated annealing algorithm to update them one by one, perform difference calculation and convergence judgment, and obtain the thermo-mechanical coupling geometric parameter set through iterative loop and finite element verification;

[0011] S5: Based on the thermo-coupling geometric parameter set, combined with the thickness information of the bearing layer, core layer, reinforcement layer and coating layer, a ratio combination is generated, compared with the limiting conditions and then filtered. The matching items are sorted to obtain the recommended combination of multi-layer thickness ratios.

[0012] As a further aspect of the present invention, the initial parameter matrix of the thermal insulation structure includes layer sequence coding, thickness vector, interface thermal resistance value, honeycomb interlayer size, corrugation parameter, bulkhead heat transfer coefficient, and radiation coefficient; the thermal insulation performance prediction combination set includes thermal conductivity combination, thermal diffusivity combination, and propagation time combination; the qualified thermal insulation performance set includes propagation time qualified items, equivalent thermal conductivity qualified items, and areal density qualified items; the thermo-coupling geometric parameter set includes honeycomb size parameter, wall thickness parameter, height parameter, and corrugation parameter; and the recommended combination of multilayer thickness ratios includes load-bearing layer thickness ratio, core layer thickness ratio, reinforcement layer thickness ratio, and coating thickness ratio.

[0013] As a further aspect of the present invention, the specific steps for generating the initial parameter matrix of the thermal insulation structure are as follows:

[0014] Based on the sequence coding, thickness vector, interface thermal resistance, honeycomb interlayer and corrugated geometric parameters, bulkhead heat transfer and radiation coefficients, the values ​​of each item are first arranged in a row and column manner, the thickness vector and interface thermal resistance are added item by item and written into the matrix unit, the honeycomb interlayer size and corrugated parameters are merged in sequence, and the bulkhead heat transfer and radiation coefficients are added to the matrix field to generate a parameter combination matrix.

[0015] Based on the parameter combination matrix, different parameter columns are cross-merged to form a unified matrix structure, and the values ​​of multiple columns are arranged in order to form an overall matrix, thus establishing the initial parameter matrix of the thermal insulation structure.

[0016] As a further aspect of the present invention, the specific steps for generating the thermal insulation performance prediction combination set are as follows:

[0017] Based on the initial parameter matrix of the thermal insulation structure, the layer sequence and thickness are arranged in groups, and the thermal conductivity prediction value is obtained by using a gradient boosting regression tree. The difference between each group of prediction values ​​and the thermal conductivity coefficient is calculated and written into a table. The difference is compared with a set threshold and marked as qualified or not, generating a set of thermal conductivity difference values.

[0018] Based on the set of thermal conductivity differences, each combination value is compared with the thermal diffusivity, the difference is recorded and inconsistent combinations are eliminated, and the results that meet the conditions are merged into a new table to obtain the diffusion screening results.

[0019] Based on the diffusion screening results, the propagation time and combined data are weighted and synthesized, the convergence situation is repeatedly compared and the groups that meet the conditions are saved. The qualified items are then grouped together to obtain the thermal insulation performance prediction combination set.

[0020] As a further embodiment of the present invention, the gradient boosting regression tree, based on the initial parameter matrix of the thermal insulation structure, arranges the layer sequence and thickness in groups, uses the combined data as input samples, generates multiple weak regression trees in rounds, constructs a new regression tree based on the residual of the previous round in each round and accumulates and updates the prediction results, obtains the predicted thermal conductivity values ​​of each group, and calculates the difference between the predicted values ​​and the actual thermal conductivity values ​​item by item and writes them into a table, compares the difference values ​​with the set threshold column by column, marks the samples that meet the conditions as qualified, and marks the samples that do not meet the conditions as unqualified, and generates a set of thermal conductivity difference values.

[0021] As a further aspect of the present invention, the specific steps for generating the qualified thermal insulation performance set are as follows:

[0022] Based on the thermal insulation performance prediction combination set, the propagation time value is first compared with the threshold and the combination that does not meet the requirements is screened out. Then, the equivalent thermal conductivity is compared with the threshold and classified. The combination that meets both conditions at the same time is recorded and marked as valid, thus generating a thermal performance screening set.

[0023] Based on the thermal performance screening set, the areal density values ​​are compared with the threshold values ​​one by one. The combinations that meet the conditions are written into the result table, and the combinations that do not meet the conditions are removed, thus forming a qualified thermal insulation performance set.

[0024] As a further aspect of the present invention, the specific steps for generating the thermo-coupled geometric parameter set are as follows:

[0025] Based on the set of qualified thermal insulation performance, the honeycomb size, wall thickness, height and corrugation parameters are set, and each parameter is combined into a matrix entry in sequence to complete numerical initialization and storage. The candidate update and acceptance mechanism of the simulated annealing algorithm is introduced and the current adaptation entry is continuously recorded to generate a geometric parameter matrix.

[0026] Based on the geometric parameter matrix, update the parameter group one by one, compare the difference results with the data of the previous round, record the trend of change and determine the convergence status, and obtain the convergence determination result.

[0027] Based on the convergence determination results, the converged parameter set is sent for verification, the values ​​are checked and unqualified groups are removed, and the groups that pass the verification are archived and organized to obtain the thermo-coupling geometric parameter set.

[0028] As a further aspect of the present invention, the simulated annealing algorithm, based on the qualified thermal insulation performance set, sets the honeycomb size, wall thickness, height, and corrugation parameters, takes the initial parameter set as the starting solution, generates a neighborhood parameter set according to the set temperature, calculates the thermal conductivity and bending stiffness of the neighborhood parameter set and compares it with the current solution, directly replaces the new solution when it is better than the current solution, and accepts and retains it according to the temperature factor probability when the new solution is worse than the current solution. Then, the temperature is gradually reduced and the neighborhood perturbation and acceptance judgment are repeated, continuously recording the accepted adaptive parameter set and updating the adaptive solution, forming a converged geometric parameter matrix and outputting it as a thermo-coupling geometric parameter set.

[0029] As a further aspect of the present invention, the specific steps for generating the recommended combination of multilayer thickness ratios are as follows:

[0030] Based on the thermo-coupling geometric parameter set, the ratio of the thickness of the bearing layer, the core layer, the reinforcement layer, and the coating layer is set. The generated multi-layer thickness values ​​are written into a matrix, and the total thickness of each combination is calculated to obtain a set of thickness ratio combinations.

[0031] Based on the set of thickness ratio combinations, the constraints are compared with the values ​​of each combination, and those that exceed the conditions are eliminated. The combinations that meet the conditions are then sorted in order to obtain the recommended combinations of multi-layer thickness ratios.

[0032] A composite isolation layer design system for a containerized battery compartment, the system being used to execute the aforementioned composite isolation layer design method for a containerized battery compartment, the system comprising:

[0033] Parameter construction module: Based on sequence coding, thickness vector, interface thermal resistance, honeycomb interlayer geometry, corrugated geometry, bulkhead heat transfer and radiation coefficients, the values ​​are arranged and merged column by column to establish the initial matrix of the thermal insulation structure;

[0034] Performance prediction module: Based on the initial matrix of the insulation structure, a gradient boosting regression tree is used to generate thermal conductivity prediction values. The predicted values ​​are compared with the thermal conductivity and written into the difference table to obtain a set of thermal conductivity differences. Then, the thermal diffusivity is compared and unqualified combinations are deleted to obtain the diffusion screening results. Subsequently, the propagation time and combination are weighted and synthesized to determine convergence and save qualified groups, which are summarized into a set of thermal insulation performance prediction combinations.

[0035] Performance screening module: Based on the predicted thermal insulation performance combination set, it compares the propagation time, equivalent thermal conductivity, and areal density item by item, retains qualified combinations and classifies them to form a qualified performance set;

[0036] Configuration optimization module: Based on the qualified performance set, set the honeycomb size, wall thickness, height and corrugation parameters, generate a geometric parameter matrix, update each parameter and compare the differences, record the convergence status, obtain the convergence judgment result, introduce the simulated annealing algorithm for candidate update and acceptance, send it for verification and remove unqualified groups, and obtain the thermo-coupling geometric group.

[0037] Thickness allocation module: Based on the thermo-coupling geometry group, it calls the thicknesses of the bearing layer, core layer, reinforcement layer and coating layer to generate proportional combinations, compares and sorts the constraints item by item, and obtains the recommended thickness ratio group.

[0038] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0039] In this invention, by establishing an initial parameter matrix composed of sequence coding, thickness vector, interface thermal resistance, honeycomb interlayer and corrugated geometric parameters, and bulkhead heat transfer and radiation coefficients, the thermal insulation performance evaluation can obtain an accurate baseline under multi-dimensional input conditions.

[0040] In this invention, gradient boosting regression trees are used to compare and iteratively screen differences under constraints of thermal conductivity, thermal diffusivity and propagation time, so that the thermal insulation combination can gradually approach the optimal solution during the convergence process, avoiding the deviation caused by the reliance on experience in parameter selection in traditional structural design.

[0041] In this invention, a simulated annealing algorithm is used to perform difference calculations and convergence determination on the honeycomb size, wall thickness, height and corrugation geometric parameters. Combined with finite element verification, the optimization process is ensured to take into account both thermodynamic and mechanical requirements, so that the geometric parameters and thermal performance converge to a stable combination simultaneously. The thermo-coupling geometric parameters and the thickness ratio of each layer are integrated to form a sorted recommended combination that meets the constraints, thereby achieving the synergistic optimization of thermal insulation, strength and lightweight performance. Attached Figure Description

[0042] Figure 1 This is a schematic diagram of the workflow of the present invention. Detailed Implementation

[0043] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0044] Example 1

[0045] Please see Figure 1This invention provides a technical solution: a composite insulation layer design method for a containerized battery compartment, comprising the following steps:

[0046] S1: Based on the sequence coding, thickness vector, interface thermal resistance, honeycomb interlayer and corrugated geometric parameters, bulkhead heat transfer and radiation coefficients, matrix arrangement and merging are performed to obtain the initial parameter matrix of the thermal insulation structure.

[0047] S2: Based on the initial parameter matrix of the thermal insulation structure, according to the thermal conductivity, thermal diffusivity and propagation time, a gradient boosting regression tree is used to combine the layer sequence and thickness and perform difference comparison to screen out qualified combinations. Then, the results are weighted and iterated repeatedly to obtain the thermal insulation performance prediction combination set.

[0048] S3: Based on the predicted thermal insulation performance combination set, the propagation time, equivalent thermal conductivity and areal density are used as the judgment conditions. The combination is compared with the threshold in turn. The combination that meets the conditions is retained and the combination that does not meet the conditions is eliminated to form a qualified thermal insulation performance set.

[0049] S4: Based on the qualified thermal insulation performance set, the honeycomb size, wall thickness, height and corrugation parameters are set, and the simulated annealing algorithm is used to update them one by one. The difference calculation and convergence judgment are performed. After iterative loop and finite element verification, the thermo-mechanical coupling geometric parameter set is obtained.

[0050] S5: Based on the thermo-coupled geometric parameter set, combined with the thickness information of the load-bearing layer, core layer, reinforcement layer and coating layer, a ratio combination is generated. After comparison with the constraints, the combinations are filtered, and the matching items are sorted to obtain the recommended combination of multi-layer thickness ratios.

[0051] The initial parameter matrix of the thermal insulation structure includes layer sequence coding, thickness vector, interfacial thermal resistance, honeycomb interlayer size, corrugation parameters, bulkhead heat transfer coefficient, and radiation coefficient. The thermal insulation performance prediction combination set includes thermal conductivity combination, thermal diffusivity combination, and propagation time combination. The qualified thermal insulation performance set includes qualified propagation time, qualified equivalent thermal conductivity, and qualified areal density. The thermo-coupling geometric parameter set includes honeycomb size parameters, wall thickness parameters, height parameters, and corrugation parameters. The recommended combination of multilayer thickness ratios includes the thickness ratio of the load-bearing layer, the core layer, the reinforcement layer, and the coating layer.

[0052] The specific steps for generating the initial parameter matrix of the thermal insulation structure are as follows:

[0053] Based on the sequence coding, thickness vector, interface thermal resistance, honeycomb interlayer and corrugated geometric parameters, bulkhead heat transfer and radiation coefficients, the values ​​of each item are first arranged in a row and column manner, the thickness vector and interface thermal resistance are added item by item and written into the matrix unit, the honeycomb interlayer size and corrugated parameters are merged in sequence, and the bulkhead heat transfer and radiation coefficients are added to the matrix field to generate a parameter combination matrix.

[0054] Based on the parameter combination matrix, different parameter columns are cross-merged to form a unified matrix structure, and the values ​​of multiple columns are arranged into an overall matrix in order to establish the initial parameter matrix of the thermal insulation structure.

[0055] Based on sequence coding, thickness vector, interface thermal resistance, honeycomb interlayer and corrugated geometric parameters, bulkhead heat transfer and radiation coefficients, a matrix filling algorithm is used to establish a 20-row, 20-column matrix and initialize it to zero. The thickness vector and interface thermal resistance are added item by item to generate an array, which is then written into the matrix row and column in sequence. The honeycomb interlayer size and corrugated geometric parameters are merged according to column order and filled into the first ten columns. The bulkhead heat transfer coefficient and radiation coefficient are added to the last column in sequence, and the parameter combination matrix is ​​derived.

[0056] Based on the parameter combination matrix, a matrix merging algorithm is used to cross-combine the values ​​in each column of the parameter matrix to generate a new array, write it into a unified matrix structure in order, and consolidate the values ​​in multiple columns into an overall matrix according to the ascending order rule, thereby deriving the initial parameter matrix of the thermal insulation structure.

[0057] The specific steps for generating the thermal insulation performance prediction set are as follows:

[0058] Based on the initial parameter matrix of the thermal insulation structure, the layer sequence and thickness are arranged in groups, and the thermal conductivity prediction value is obtained by using a gradient boosting regression tree. The difference between each group of prediction values ​​and the thermal conductivity coefficient is calculated and written into a table. The difference is compared with a set threshold and marked as qualified or not, generating a set of thermal conductivity difference values.

[0059] Based on the set of thermal conductivity differences, the combination values ​​are compared with the thermal diffusivity one by one, the differences are recorded and non-matching combinations are eliminated, and the results that meet the conditions are merged into a new table to obtain the diffusion screening results.

[0060] Based on the diffusion screening results, the propagation time and combined data are weighted and synthesized, the convergence situation is repeatedly compared and the groups that meet the conditions are saved. The qualified items are grouped into one set to obtain the thermal insulation performance prediction combination set.

[0061] Based on the initial parameter matrix of the thermal insulation structure, the layer sequence and thickness are arranged in groups. The gradient boosting regression tree algorithm is used for thermal conductivity prediction. The training parameters are set as follows: learning rate 0.05, number of weak learners 200, maximum tree depth 6, minimum number of samples per split 10, and minimum sample weight of leaf nodes 0.1. The input variable is the combination of layer sequence and thickness, and the output variable is the predicted thermal conductivity value. The regression tree is called group by group to iterate and update the residuals and generate prediction results. The predicted value of each group is subtracted from the thermal conductivity coefficient item by item and written into the table field. The difference is recorded in a row and column and the numerical format is retained to three decimal places. The difference is compared with the set threshold of 0.05 item by item. If it is less than or equal to the threshold, it is marked as qualified; if it is greater than the threshold, it is marked as unqualified. A mark field is added to the corresponding row of the table to generate a set of thermal conductivity differences.

[0062] Based on the set of thermal conductivity differences, a numerical comparison and screening method is used to compare the combined values ​​with the thermal diffusivity item by item. First, the calculation formula diffusivity = thermal conductivity / (density × specific heat capacity) is called to calculate according to the input parameters. The difference of each combination is recorded and a column is added to the table to store the comparison results. When the difference exceeds the limit of 0.1, it is directly marked as inconsistent and filtered out from the data. When the difference is within the limit, it is marked as consistent and retained. The records that meet the conditions are merged into a new table and arranged in order in the table fields to obtain the diffusion screening results.

[0063] Based on the diffusion screening results, a weighted synthesis iterative algorithm is used to synthesize the propagation time and combined data with weights. The propagation time weight is set to 0.6 and the combined data weight is set to 0.4. The calculation method is the weighted sum of the propagation time and combined data of each group. The iteration method is to compare the difference between the newly generated data and the previous round of data one by one. If the difference is less than the convergence threshold of 0.001, the iteration stops and is saved. If the difference is greater than the threshold, the iteration continues until the condition is met. The groups that meet the condition are saved and marked as qualified items in the table. Finally, all qualified items are collected into one set to obtain the thermal insulation performance prediction combination set.

[0064] Gradient boosting regression tree, based on the initial parameter matrix of the thermal insulation structure, arranges the layer sequence and thickness in groups, takes the combined data as input samples, generates multiple weak regression trees in rounds, constructs a new regression tree based on the residual of the previous round in each round and accumulates and updates the prediction results, obtains the predicted thermal conductivity values ​​of each group, and calculates the difference between the predicted value and the actual thermal conductivity item by item and writes it into a table. The difference is compared with the set threshold column by column. Samples that meet the conditions are marked as qualified and samples that do not meet the conditions are marked as unqualified, generating a set of thermal conductivity difference values.

[0065] Gradient boosting regression trees, according to the formula:

[0066] ,

[0067] in: Representation group The corrected thermal conductivity, Indicates the first The thickness of the layer, Indicates the first The thermal conductivity of the layer, Indicates the first Layer weight coefficients, Indicates the first Temperature sensitivity coefficient of the layer material Representation group The difference between the test temperature and the reference temperature, Indicates the first Layers and Thermal contact resistance of the layer This represents the disorder penalty coefficient. This represents the penalty coefficient for thickness dispersion. Indicates the degree of disorder in the hierarchical order. Indicator of thickness dispersion Indicates the total number of composite isolation layers;

[0068] Execution process: First, read the group hierarchical sequence And sort them from the outside in to obtain the thickness of each layer. and thermal conductivity Then, based on the training results of the gradient boosting regression tree, the weight coefficients of each layer are calculated. This reflects the importance of each layer in the overall heat conduction process, ensuring that the relative influence of materials is taken into account in the calculation, and also introduces the temperature sensitivity coefficient of the materials. and the difference between the test temperature and the reference temperature By correcting the thermal conductivity of each layer The thermal conductivity after temperature correction is obtained. Furthermore, the thermal contact resistance between adjacent layers needs to be considered. This affects the heat transfer efficiency between layers, and the formula then calculates the disorder index of the sequence. and thickness dispersion index These reflect the stability of the layer sequence and the uniformity of the thickness of each layer, respectively. Excessive disorder and dispersion can negatively impact thermal conductivity, so a penalty needs to be applied to the final result. Finally, penalty coefficients for disorder and thickness dispersion are introduced. and By comprehensively adjusting the thermal conductivity of each layer, the corrected equivalent thermal conductivity was finally obtained. It accurately reflects the actual thermal management effect of the composite insulation layer and is of great significance in the design of containerized battery compartments, used to optimize the thermal insulation and heat dissipation performance of the battery compartment.

[0069] The specific steps for generating a set of qualified thermal insulation performance are as follows:

[0070] Based on the thermal insulation performance prediction combination set, the propagation time value is first compared with the threshold and the combination that does not meet the requirements is screened out. Then, the equivalent thermal conductivity is compared with the threshold and classified. The combination that meets both conditions at the same time is recorded and marked as valid, thus generating a thermal performance screening set.

[0071] Based on the thermal performance screening set, the areal density values ​​are compared with the threshold values ​​in turn. The combinations that meet the conditions are written into the result table, and the combinations that do not meet the conditions are removed, thus forming a qualified thermal insulation performance set.

[0072] Based on the thermal insulation performance prediction combination set, a threshold comparison screening algorithm is adopted to compare the propagation time value with the threshold value one by one. The threshold value is set to 50 seconds. The comparison method is to take the propagation time of each group and subtract it from the threshold value and keep three decimal places. If the difference is less than zero, it is rejected. If the difference is greater than or equal to zero, it is retained and written into the screening table. Then, the equivalent thermal conductivity is compared. The threshold value is set to 0.25 W / m Kelvin. The comparison method is to take the equivalent thermal conductivity of each group and subtract it from the threshold value and write it into a new column. If the value is less than or equal to zero, it is marked as meeting the condition. If the value is greater than zero, it is marked as not meeting the condition. For combinations that meet both the propagation time and equivalent thermal conductivity conditions, a valid label is written in the label field of the table. The results are collected in the result table to generate a thermal performance screening set.

[0073] Based on the thermal performance screening set, a threshold classification algorithm is used to compare the areal density values ​​with a threshold of 15 kg / m². The comparison is performed by subtracting the threshold from the areal density values ​​row by row. If the difference is less than or equal to zero, it is recorded in the result table and labeled as qualified. If the difference is greater than zero, it is removed during the screening process. The combinations that meet the conditions are written into the result table in sequence and uniformly numbered. The combinations that do not meet the conditions are discarded directly during the output process. The tables are then merged and output to form a qualified thermal insulation performance set.

[0074] The specific steps for generating the thermo-coupled geometric parameter set are as follows:

[0075] Based on the set of qualified thermal insulation performance, the honeycomb size, wall thickness, height and corrugation parameters are set, and each parameter is combined into a matrix entry in sequence to complete numerical initialization and storage. The candidate update and acceptance mechanism of the simulated annealing algorithm is introduced and the current adaptation entry is continuously recorded to generate a geometric parameter matrix.

[0076] Based on the geometric parameter matrix, update the parameter group one by one, compare the difference results with the data of the previous round, record the trend of change and determine the convergence status, and obtain the convergence determination result.

[0077] Based on the convergence determination results, the converged parameter set is sent for verification, the values ​​are checked and unqualified groups are removed, and the groups that pass the verification are archived and organized to obtain the thermo-mechanical coupling geometric parameter set.

[0078] Based on a set of qualified thermal insulation performance, the honeycomb size is set to 5 mm, the wall thickness to 0.3 mm, the height to 15 mm, and the corrugation parameter to 0.25. These parameters are sequentially combined into matrix entries and filled in row and column order. A simulated annealing algorithm is used for candidate updating and acceptance. The initial temperature is set to 100, the cooling rate to 0.95, and the maximum number of iterations to 500. Candidates are generated by adding or subtracting a random perturbation of 0.01 to the current entry value. The evaluation function is a weighted sum of the honeycomb size, wall thickness, height, and corrugation parameter, with weights of 0.4, 0.2, 0.3, and 0.1, respectively. If a candidate's result is better than the current entry, it is directly accepted; otherwise, it is accepted with a probability exp(-difference / temperature). Acceptance and rejection are recorded in the matrix storage table. The temperature decreases sequentially at a set rate, iterating until the termination condition is met. The final entries are stored in the matrix field, generating a geometric parameter matrix.

[0079] Based on the geometric parameter matrix, the parameter group is updated one by one using the difference comparison update method. The value of each parameter group is subtracted from the value of the corresponding parameter group in the previous round, and the difference is recorded. The difference is stored in the result column one by one. Each difference is compared with the set convergence threshold of 0.001. If the difference is greater than the threshold, it is marked as non-converged. If the difference is less than or equal to the threshold, it is marked as converged. The marking status of all parameter groups is counted in turn, the change trend is written into the trend table, and the comparison round number is recorded one by one to obtain the convergence judgment result.

[0080] Based on the convergence determination results, the converged parameter sets are sent to the verification stage using a verification method. The verification parameters include upper and lower limits of honeycomb size (3 to 7 mm), upper and lower limits of wall thickness (0.2 to 0.5 mm), upper and lower limits of height (10 to 20 mm), and upper and lower limits of corrugation parameters (0.2 to 0.3 mm). Each parameter set is compared item by item. If it exceeds the range, it is rejected and marked as unqualified. If it is within the range, it is marked as passed. The groups that pass the verification are archived in the result table in sequence and numbered according to order. Finally, the results are sorted and output to obtain the thermo-coupling geometric parameter set.

[0081] The simulated annealing algorithm, based on a set of qualified thermal insulation performance, sets the honeycomb size, wall thickness, height, and corrugation parameters. The initial parameter set is used as the starting solution. A neighborhood parameter set is generated according to the set temperature. The thermal conductivity and bending stiffness of the neighborhood parameter set are calculated and compared with the current solution. When the new solution is better than the current solution, it is directly replaced. When the new solution is worse than the current solution, it is accepted and retained according to the temperature factor probability. Then, the temperature is gradually reduced and the neighborhood perturbation and acceptance judgment are repeated. The accepted adapted parameter sets are continuously recorded and the adapted solutions are updated to form a converged geometric parameter matrix and output as a thermo-mechanical coupling geometric parameter set.

[0082] Simulated annealing algorithm, according to the formula:

[0083] ,

[0084] in: This indicates the improved heat flow. Indicates the thermal conductivity of a material. Represents the thermal conductivity surface area. Indicates the ambient temperature difference. Indicates the temperature correction factor. This represents the shape optimization weight coefficient. Indicates the length of the heat conduction path. This represents the disorder penalty coefficient. Indicates the degree of hierarchical disorder;

[0085] Execution process: First, determine the thermal conductivity of the material. The thermal conductivity surface area is then calculated by taking the properties from the physical property database of the selected composite insulating layer material and correcting it with test results. The area is obtained by superimposing geometric conditions such as honeycomb size, wall thickness, and corrugation parameters, followed by obtaining the ambient temperature difference. The value is determined by the difference between the design operating temperature inside the battery compartment and the external boundary ambient temperature. To account for temperature fluctuations during operation, a temperature correction factor is introduced. The shape optimization weights were determined by fitting the environmental temperature change curve. Based on the geometric configuration of the cellular cells, the heat conduction path length is calculated and normalized. The result is calculated from the equivalent thickness of the isolation layer and the length of the main heat flow path, while also evaluating the sequence disorder index. The disorder penalty coefficient is calculated by comparing the differences between the actual sequence and the standard sequence. The heat flux error was obtained and fixed by fitting it with non-negative least squares on a historical qualified sample set. All parameters were substituted into the formula to calculate the heat flux. and with The geometric parameter matrix is ​​used as a fitness function to generate and accept candidates for combinations of geometric parameters until a geometric parameter matrix that satisfies the design objectives is formed.

[0086] The specific steps for generating recommended combinations of multi-layer thickness ratios are as follows:

[0087] Based on the thermo-coupled geometric parameter set, the ratio of the thickness of the bearing layer, the core layer, the reinforcement layer and the coating layer is set. The generated multi-layer thickness values ​​are written into a matrix and the total thickness of each combination is calculated to obtain the thickness ratio combination set.

[0088] Based on the set of thickness ratio combinations, the constraints are compared with the values ​​of each combination, and those exceeding the conditions are eliminated. The combinations that meet the conditions are sorted in order to obtain the recommended combinations of multi-layer thickness ratios.

[0089] Based on the thermo-coupled geometric parameter set, a matrix combination generation algorithm was used to set the thickness range of the bearing layer to 2 to 5 mm, the core layer to 10 to 20 mm, the reinforcement layer to 1 to 3 mm, and the coating layer to 0.2 to 1 mm. The four thickness ratios were selected at 0.2 intervals to generate multiple combinations. Each combination was written into the matrix row and column in sequence and recorded with a number. The thickness values ​​of each group in the matrix were added one by one to obtain the total thickness and written into the total thickness column. All results were stored in the matrix table in sequence to generate a set of thickness ratio combinations.

[0090] Based on the set of thickness ratio combinations, a threshold comparison sorting algorithm is used to compare the constraints with the values ​​of each combination item by item. The constraints are set as follows: the total thickness does not exceed 25 mm and the lower limit of the single layer thickness is not less than the corresponding minimum value. If the thickness of any layer in the combination exceeds the limit, it is marked as removed and will no longer participate in the sorting. If all values ​​of the combination are within the limit range, it is marked as qualified and written to a new table. The sorting steps are called for all qualified combinations to sort them in ascending order of total thickness. The sorted results are stored and output row by row to obtain the recommended combinations of multi-layer thickness ratios.

[0091] A composite isolation layer design system for a containerized battery compartment, the system being used to execute the aforementioned composite isolation layer design method for a containerized battery compartment, the system comprising:

[0092] Parameter construction module: Based on sequence coding, thickness vector, interface thermal resistance, honeycomb interlayer geometry, corrugated geometry, bulkhead heat transfer and radiation coefficients, the values ​​are arranged and merged column by column to establish the initial matrix of the thermal insulation structure;

[0093] Performance prediction module: Based on the initial matrix of the thermal insulation structure, a gradient boosting regression tree is used to generate thermal conductivity prediction values. The predicted values ​​are compared with the thermal conductivity and written into the difference table to obtain the thermal conductivity difference set. Then, the thermal diffusivity is compared and unqualified combinations are deleted to obtain the diffusion screening results. Subsequently, the propagation time and combination are weighted and synthesized to determine convergence and save qualified groups, which are summarized into a thermal insulation performance prediction combination set.

[0094] Performance screening module: Based on the thermal insulation performance prediction combination set, it compares the propagation time, equivalent thermal conductivity, and areal density item by item, retains qualified combinations and classifies them to form a qualified performance set;

[0095] Configuration optimization module: Based on the qualified performance set, set the cell size, wall thickness, height and corrugation parameters, generate a geometric parameter matrix, update each parameter and compare the difference, record the convergence status, obtain the convergence judgment result, introduce simulated annealing algorithm for candidate update and acceptance, send it for verification and remove unqualified groups, and obtain the thermo-coupling geometric group.

[0096] Thickness allocation module: Based on the thermo-coupled geometry group, it calls the thicknesses of the bearing layer, core layer, reinforcement layer and coating layer to generate proportional combinations, compares and sorts the constraints item by item, and obtains the recommended group of thickness proportions.

[0097] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A composite isolation layer design method for a containerized battery compartment, characterized in that, Includes the following steps: S1: Based on the sequence coding, thickness vector, interface thermal resistance, honeycomb interlayer and corrugated geometric parameters, bulkhead heat transfer and radiation coefficients, matrix arrangement and merging are performed to obtain the initial parameter matrix of the thermal insulation structure. S2: Based on the initial parameter matrix of the insulation structure, according to the thermal conductivity, thermal diffusivity and propagation time, a gradient boosting regression tree is used to combine the layer sequence and thickness and perform difference comparison to screen out qualified combinations. Then, the results are weighted and iterated repeatedly to obtain the set of predicted combinations for insulation performance. S3: Based on the predicted thermal insulation performance combination set, the propagation time, equivalent thermal conductivity and areal density are used as the judgment conditions, and are compared with the threshold in turn. The combination that meets the conditions is retained and the combination that does not meet the conditions is eliminated to form a qualified thermal insulation performance set. S4: Based on the set of qualified thermal insulation performance, set the honeycomb size, wall thickness, height and corrugation parameters, and use the simulated annealing algorithm to update them one by one, perform difference calculation and convergence judgment, and obtain the thermo-mechanical coupling geometric parameter set through iterative loop and finite element verification; S5: Based on the thermo-coupling geometric parameter set, combined with the thickness information of the bearing layer, core layer, reinforcement layer and coating layer, a ratio combination is generated, compared with the limiting conditions and then filtered. The matching items are sorted to obtain the recommended combination of multi-layer thickness ratios.

2. The composite isolation layer design method for containerized battery compartments according to claim 1, characterized in that, The initial parameter matrix of the thermal insulation structure includes layer sequence coding, thickness vector, interfacial thermal resistance value, honeycomb interlayer size, corrugation parameters, bulkhead heat transfer coefficient, and radiation coefficient. The thermal insulation performance prediction combination set includes thermal conductivity combination, thermal diffusivity combination, and propagation time combination. The qualified thermal insulation performance set includes propagation time qualified items, equivalent thermal conductivity qualified items, and areal density qualified items. The thermo-coupling geometric parameter set includes honeycomb size parameters, wall thickness parameters, height parameters, and corrugation parameters. The recommended combination of multilayer thickness ratios includes the load-bearing layer thickness ratio, core layer thickness ratio, reinforcement layer thickness ratio, and coating thickness ratio.

3. The composite isolation layer design method for containerized battery compartments according to claim 1, characterized in that, The specific steps for generating the initial parameter matrix of the thermal insulation structure are as follows: Based on the sequence coding, thickness vector, interface thermal resistance, honeycomb interlayer and corrugated geometric parameters, bulkhead heat transfer and radiation coefficients, the values ​​of each item are first arranged in a row and column manner, the thickness vector and interface thermal resistance are added item by item and written into the matrix unit, the honeycomb interlayer size and corrugated parameters are merged in sequence, and the bulkhead heat transfer and radiation coefficients are added to the matrix field to generate a parameter combination matrix. Based on the parameter combination matrix, different parameter columns are cross-merged to form a unified matrix structure, and the values ​​of multiple columns are arranged in order to form an overall matrix, thus establishing the initial parameter matrix of the thermal insulation structure.

4. The composite isolation layer design method for containerized battery compartments according to claim 1, characterized in that, The specific steps for generating the predicted thermal insulation performance set are as follows: Based on the initial parameter matrix of the thermal insulation structure, the layer sequence and thickness are arranged in groups, and the thermal conductivity prediction value is obtained by using a gradient boosting regression tree. The difference between each group of prediction values ​​and the thermal conductivity coefficient is calculated and written into a table. The difference is compared with a set threshold and marked as qualified or not, generating a set of thermal conductivity difference values. Based on the set of thermal conductivity differences, each combination value is compared with the thermal diffusivity, the difference is recorded and inconsistent combinations are eliminated, and the results that meet the conditions are merged into a new table to obtain the diffusion screening results. Based on the diffusion screening results, the propagation time and combined data are weighted and synthesized, the convergence situation is repeatedly compared and the groups that meet the conditions are saved. The qualified items are then grouped together to obtain the thermal insulation performance prediction combination set.

5. The composite isolation layer design method for containerized battery compartments according to claim 4, characterized in that, The gradient boosting regression tree, based on the initial parameter matrix of the thermal insulation structure, arranges the layer sequence and thickness in groups, uses the combined data as input samples, and generates multiple weak regression trees in rounds. In each round, a new regression tree is constructed based on the residual of the previous round, and the prediction results are accumulated and updated to obtain the predicted thermal conductivity values ​​of each group. The predicted values ​​are subtracted from the actual thermal conductivity values ​​item by item and written into a table. The difference is compared with a set threshold column by column. Samples that meet the conditions are marked as qualified, and those that do not meet the conditions are marked as unqualified, thus generating a set of thermal conductivity difference values.

6. The composite isolation layer design method for containerized battery compartments according to claim 1, characterized in that, The specific steps for generating the set of qualified thermal insulation performance are as follows: Based on the thermal insulation performance prediction combination set, the propagation time value is first compared with the threshold and the combination that does not meet the requirements is screened out. Then, the equivalent thermal conductivity is compared with the threshold and classified. The combination that meets both conditions at the same time is recorded and marked as valid, thus generating a thermal performance screening set. Based on the thermal performance screening set, the areal density values ​​are compared with the threshold values ​​one by one. The combinations that meet the conditions are written into the result table, and the combinations that do not meet the conditions are removed, thus forming a qualified thermal insulation performance set.

7. The composite isolation layer design method for containerized battery compartments according to claim 1, characterized in that, The specific steps for generating the thermo-coupled geometric parameter set are as follows: Based on the set of qualified thermal insulation performance, the honeycomb size, wall thickness, height and corrugation parameters are set, and each parameter is combined into a matrix entry in sequence to complete numerical initialization and storage. The candidate update and acceptance mechanism of the simulated annealing algorithm is introduced and the current adaptation entry is continuously recorded to generate a geometric parameter matrix. Based on the geometric parameter matrix, update the parameter group one by one, compare the difference results with the data of the previous round, record the trend of change and determine the convergence status, and obtain the convergence determination result. Based on the convergence determination results, the converged parameter set is sent for verification, the values ​​are checked and unqualified groups are removed, and the groups that pass the verification are archived and organized to obtain the thermo-coupling geometric parameter set.

8. The composite isolation layer design method for containerized battery compartments according to claim 7, characterized in that, The simulated annealing algorithm, based on the qualified thermal insulation performance set, sets the honeycomb size, wall thickness, height, and corrugation parameters. It uses the initial parameter set as the starting solution, generates a neighborhood parameter set according to the set temperature, calculates the thermal conductivity and bending stiffness of the neighborhood parameter set and compares them with the current solution. When the new solution is better than the current solution, it is directly replaced. When the new solution is worse than the current solution, it is accepted and retained according to the temperature factor probability. Then, the temperature is gradually reduced and the neighborhood perturbation and acceptance judgment are repeated. The accepted adaptive parameter sets are continuously recorded and the adaptive solutions are updated to form a converged geometric parameter matrix and output as a thermo-coupling geometric parameter set.

9. The composite isolation layer design method for containerized battery compartments according to claim 1, characterized in that, The specific steps for generating the recommended combination of multi-layer thickness ratios are as follows: Based on the thermo-coupling geometric parameter set, the ratio of the thickness of the bearing layer, the core layer, the reinforcement layer, and the coating layer is set. The generated multi-layer thickness values ​​are written into a matrix, and the total thickness of each combination is calculated to obtain a set of thickness ratio combinations. Based on the set of thickness ratio combinations, the constraints are compared with the values ​​of each combination, and those that exceed the conditions are eliminated. The combinations that meet the conditions are then sorted in order to obtain the recommended combinations of multi-layer thickness ratios.

10. A composite isolation layer design system for a containerized battery compartment, characterized in that, The composite isolation layer design method for containerized battery compartments according to any one of claims 1-9, wherein the system comprises: Parameter construction module: Based on layer sequence coding, thickness vector, interface thermal resistance, honeycomb interlayer geometry, corrugated geometry, bulkhead heat transfer and radiation coefficient, the values ​​are arranged and merged column by column to establish the initial matrix of the thermal insulation structure; Performance prediction module: Based on the initial matrix of the insulation structure, a gradient boosting regression tree is used to generate thermal conductivity prediction values. The predicted values ​​are compared with the thermal conductivity and written into the difference table to obtain a set of thermal conductivity differences. Then, the thermal diffusivity is compared and unqualified combinations are deleted to obtain the diffusion screening results. Subsequently, the propagation time and combination are weighted and synthesized to determine convergence and save qualified groups, which are summarized into a set of thermal insulation performance prediction combinations. Performance screening module: Based on the predicted thermal insulation performance combination set, it compares the propagation time, equivalent thermal conductivity, and areal density item by item, retains qualified combinations and classifies them to form a qualified performance set; Configuration optimization module: Based on the qualified performance set, set the honeycomb size, wall thickness, height and corrugation parameters, generate a geometric parameter matrix, update each parameter and compare the differences, record the convergence status, obtain the convergence judgment result, introduce the simulated annealing algorithm for candidate update and acceptance, send it for verification and remove unqualified groups, and obtain the thermo-coupling geometric group. Thickness allocation module: Based on the thermo-coupling geometry group, it calls the thicknesses of the bearing layer, core layer, reinforcement layer and coating layer to generate proportional combinations, compares and sorts the constraints item by item, and obtains the recommended thickness ratio group.

Images