Automobile door inner plate structure size optimization method and device, equipment and storage medium

Abstract

The invention provides an automobile door inner plate structure size optimization method and device, equipment and a storage medium, and belongs to the technical field of structure size optimizing.The method comprises the steps that multiple structure sizes of an automobile door inner plate serve as variables, lightweight of the automobile door inner plate serves as a target, a target function is constructed, and the structure size of the automobile door inner plate is obtained according to a preset structure stress threshold value; determining constraint conditions of the target function; determining a plurality of variable combinations according to the value ranges of the plurality of structure sizes; inputting each variable combination into a pre-constructed proxy model to obtain an automobile door inner plate structure stress corresponding to each variable combination; solving the target function under constraint conditions by adopting a genetic algorithm according to the structural stress of the automobile door inner plate to obtain an optimal variable combination, and determining the optimal structural size of the automobile door inner plate; the proxy model is obtained by training based on a sample variable combination and an automobile door inner plate structure stress label of the sample variable combination. The efficiency of optimizing the structural size of the automobile door inner plate can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of structural dimension optimization, and in particular to a method, device, equipment and storage medium for optimizing the structural dimension of an automobile door inner panel. Background Art

[0002] The structural dimension optimization of the inner door panel of an automobile is a complex project. While pursuing the goal of lightweighting, it is necessary to comprehensively consider multiple performance requirements. At present, in the process of optimizing the structural dimension of the inner door panel of an automobile, the performance of the inner door panel of an automobile with different structural dimensions is mainly evaluated by simulation calculation. Each design optimization of the inner door panel structure of an automobile requires a large amount of performance calculation, which further guides the direction of the next design optimization, resulting in a significant increase in the time cost of the calculation, which seriously affects the efficiency of the structural dimension optimization. Summary of the invention

[0003] The present invention provides a method, device, equipment and storage medium for optimizing the structural dimensions of an automobile door inner panel, so as to solve the defect in the prior art that the performance of automobile door inner panels of different structural dimensions is evaluated by simulation calculation, the time cost of calculation is greatly increased, and the efficiency of structural dimension optimization is seriously affected.

[0004] In a first aspect, the present invention provides a method for optimizing the structural dimensions of an automobile door inner panel, comprising: Taking multiple structural dimensions of the inner door panel of the automobile as variables and lightweighting of the inner door panel of the automobile as a goal, an objective function is constructed, and according to a preset structural stress threshold, a constraint condition of the objective function is determined; Determining multiple variable combinations according to the value ranges of the multiple structural dimensions; Input each variable combination into a pre-built proxy model to obtain the automobile door inner panel structural stress corresponding to each variable combination output by the proxy model; Using a genetic algorithm, according to the structural stress of the automobile door inner panel corresponding to each variable combination, solving the objective function under the constraint condition to obtain an optimal variable combination, and determining the optimal structural size of the automobile door inner panel based on the optimal variable combination; The proxy model is obtained by training based on a sample variable combination corresponding to multiple structural dimensions of a sample automobile door inner panel and a structural stress label of the automobile door inner panel of the sample variable combination.

[0005] In some embodiments, the genetic algorithm is used to solve the objective function under the constraint conditions according to the automobile door inner panel structural stress corresponding to each variable combination to obtain the optimal variable combination, including: Encoding the multiple variable combinations to obtain an initial population, wherein each individual of the initial population represents each variable combination; According to the automobile door inner panel structural stress corresponding to each variable combination and the constraint condition, a plurality of individuals in the primary population are screened to obtain a screened primary population; According to the objective function, calculating the fitness value of each individual in the screened primary population; According to the fitness value of each individual, determining the best individual from the screened primary population; Determine whether the fitness value of the optimal individual meets the iteration termination condition. If so, determine that the variable combination corresponding to the optimal individual is the optimal variable combination. Otherwise, perform crossover and mutation based on the optimal individual to obtain a new generation population, and start a cycle by screening multiple individuals in the new generation population according to the constraint conditions until it is determined that the fitness value of the new optimal individual meets the iteration termination condition, and obtain the optimal variable combination based on the new optimal individual.

[0006] In some embodiments, screening multiple individuals in the primary population according to the automobile door inner panel structural stress corresponding to each variable combination and the constraint condition includes: Determine whether the automobile door inner panel structural stress corresponding to each variable combination is less than or equal to the preset structural stress threshold; if so, retain the individual corresponding to each variable combination; otherwise, eliminate the individual corresponding to each variable combination.

[0007] In some embodiments, the variable combination includes the hole size and the reinforcement rib size of the automobile door inner panel; the constraint condition includes that the structural stress of the automobile door inner panel is less than or equal to the preset structural stress threshold; and the value of the objective function is the weight of the automobile door inner panel.

[0008] In some embodiments, the proxy model includes a convolutional neural network, which includes: multiple convolutional layers based on a hard attention mechanism, multiple pooling layers, an SE residual block layer, and a fully connected layer.

[0009] In some embodiments, the process of determining the proxy model includes: Construct the initial sample variable combination according to the preset value range of each variable; Amplifying the initial sample variable combination to obtain multiple sample variable combinations; Determining a stress label of the automobile door inner panel structure for the plurality of sample variable combinations; Constructing an initial proxy model; The initial proxy model is trained using the multiple sample variable combinations as training samples and the automobile door inner panel structural stress labels of the multiple sample variable combinations as sample labels. After the training is completed, the proxy model is obtained.

[0010] In some embodiments, the training the initial proxy model comprises: Inputting each sample variable combination into the initial proxy model to obtain a predicted value of the automobile door inner panel structural stress corresponding to each sample variable combination output by the initial proxy model; Calculate the loss function value according to the predicted value of the automobile door inner panel structural stress corresponding to each sample variable combination and the automobile door inner panel structural stress label of each sample variable combination; Based on the loss function value, the parameters of the initial proxy model are iteratively optimized.

[0011] In a second aspect, the present invention further provides a device for optimizing the structural dimensions of an automobile door inner panel, comprising: A construction unit, used to construct an objective function with multiple structural dimensions of the automobile door inner panel as variables and lightweighting of the automobile door inner panel as a goal, and determine the constraint conditions of the objective function according to a preset structural stress threshold; A determination unit, used to determine a plurality of variable combinations according to the value ranges of the plurality of structural dimensions; A prediction unit, used for inputting each variable combination into a pre-built proxy model to obtain the automobile door inner panel structural stress corresponding to each variable combination output by the proxy model; A solving unit, configured to solve the objective function under the constraint conditions by using a genetic algorithm according to the structural stress of the automobile door inner panel corresponding to each variable combination, to obtain an optimal variable combination, and to determine an optimal structural size of the automobile door inner panel based on the optimal variable combination; The proxy model is obtained by training based on a sample variable combination corresponding to multiple structural dimensions of a sample automobile door inner panel and a structural stress label of the automobile door inner panel of the sample variable combination.

[0012] In a third aspect, the present invention further provides an electronic device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for optimizing the structural dimensions of an automobile door inner panel as described above is implemented.

[0013] In a fourth aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described methods for optimizing the structural dimensions of an automobile door inner panel.

[0014] The method, device, equipment and storage medium for optimizing the structural dimensions of an automobile door inner panel provided by the present invention use multiple structural dimensions of the automobile door inner panel as variables and lightweighting of the automobile door inner panel as a goal to construct an objective function, determine the constraints of the objective function, obtain multiple variable combinations, and use a pre-constructed proxy model to calculate the structural stress of the automobile door inner panel corresponding to each variable combination, thereby improving the efficiency of performance calculation; a genetic algorithm is used to solve the objective function under the constraints to obtain the optimal variable combination, thereby being able to quickly determine the optimal structural dimensions of the automobile door inner panel, thereby improving the efficiency of optimizing the structural dimensions of the automobile door inner panel. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0016] Figure 1 It is a schematic flow chart of a method for optimizing the structural dimensions of an automobile door inner panel provided by an embodiment of the present invention; Figure 2 is a flowchart of a process for determining a proxy model provided by an embodiment of the present invention; Figure 3 is a schematic structural diagram of a device for optimizing the structural dimensions of an automobile door inner panel provided by an embodiment of the present invention; Figure 4 It is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] Figure 1 The following is a flow chart of a method for optimizing the structural dimensions of an automobile door inner panel provided by an embodiment of the present invention. Figure 1 As shown, a method for optimizing the structural dimensions of an automobile door inner panel is provided, comprising the following steps: step 110, step 120, step 130 and step 140. The method flow steps are only a possible implementation of the present invention.

[0019] Step 110, constructing an objective function with multiple structural dimensions of the automobile door inner panel as variables and lightweighting of the automobile door inner panel as a goal, and determining the constraint conditions of the objective function according to a preset structural stress threshold.

[0020] Optionally, the multiple structural dimensions of the automobile door inner panel include dimensions of a series of accessories such as holes, reinforcing ribs, glass guides, lifters, and speakers.

[0021] Optionally, the preset structural stress threshold is determined according to the material properties of the inner panel of the automobile door.

[0022] In some embodiments, the variable combination includes the hole size and the reinforcement rib size of the automobile door inner panel; the constraint condition includes that the structural stress of the automobile door inner panel is less than or equal to a preset structural stress threshold; and the value of the objective function is the weight of the automobile door inner panel.

[0023] Optionally, the hole size may be the size of a ventilation hole, a weight reduction hole, or a mounting hole.

[0024] Optionally, the hole size is the diameter of the hole, and the rib size is the length, width and height of the rib.

[0025] Step 120: Determine multiple variable combinations according to the value ranges of multiple structural dimensions.

[0026] Optionally, the value ranges of multiple structural dimensions of the automobile door inner panel are determined based on user input or empirical data, such as the value range of the hole diameter, and the value range of the length, width, and height of the reinforcement rib.

[0027] Optionally, within the value ranges of multiple structural dimensions, multiple variables are randomly sampled and combined to obtain multiple variable combinations.

[0028] Step 130: Input each variable combination into a pre-built proxy model to obtain the automobile door inner panel structural stress corresponding to each variable combination output by the proxy model.

[0029] Among them, the proxy model is trained based on the sample variable combination corresponding to multiple structural dimensions of the sample automobile door inner panel and the automobile door inner panel structural stress label of the sample variable combination.

[0030] Optionally, for different door inner panel structures, multiple target domain problems are constructed according to the source domain problem corresponding to the objective function. According to the multiple target domain problems, the structural dimensions of the multiple target domains are used as new variables to obtain multiple new sample variable combinations, and the automobile door inner panel structural stress labels of the multiple new sample variable combinations are determined. The multiple new sample variable combinations are used as new training samples, and the automobile door inner panel structural stress labels of the multiple new sample variable combinations are used as new sample labels to train the proxy model. After the training is completed, a new proxy model is obtained.

[0031] In some embodiments, the proxy model includes a convolutional neural network, which includes: multiple convolutional layers based on a hard attention mechanism, multiple pooling layers, SE residual block layers, and a fully connected layer.

[0032] Optionally, the convolutional neural network is a deep convolutional neural network.

[0033] Among them, the hard attention mechanism is an attention mechanism used to select information focus in deep learning. The main characteristics of the hard attention mechanism are its selectivity and discontinuity, that is, the hard attention mechanism only focuses on certain specific parts when processing data, instead of weighting all parts.

[0034] Among them, the SE (Squeeze-and-Excitation) residual block is a residual network structure combined with the SE module, which is used to improve the performance of deep convolutional neural networks. The SE module enhances the feature representation capability by adaptively recalibrating the channels.

[0035] Step 140: Using a genetic algorithm, according to the structural stress of the automobile door inner panel corresponding to each variable combination, solving the objective function under constraint conditions, obtaining the optimal variable combination, and determining the optimal structural dimensions of the automobile door inner panel based on the optimal variable combination.

[0036] Optionally, the optimal combination of variables includes an optimal hole size and an optimal rib size.

[0037] In an embodiment of the present invention, by taking multiple structural dimensions of an automobile door inner panel as variables and taking lightweighting of the automobile door inner panel as a goal, an objective function is constructed, and the constraints of the objective function are determined, multiple variable combinations are obtained, and a pre-constructed proxy model is used to calculate the structural stress of the automobile door inner panel corresponding to each variable combination, thereby improving the efficiency of performance calculation; a genetic algorithm is used to solve the objective function under the constraints to obtain the optimal variable combination, so that the optimal structural dimensions of the automobile door inner panel can be quickly determined, thereby improving the efficiency of optimizing the structural dimensions of the automobile door inner panel.

[0038] In some embodiments, a genetic algorithm is used to solve the objective function under constraints according to the structural stress of the inner door panel of the automobile corresponding to each variable combination to obtain the optimal variable combination, including: Encode multiple variable combinations to obtain the initial population, where each individual in the initial population represents each variable combination; According to the structural stress and constraint conditions of the inner door panel of the automobile corresponding to each variable combination, multiple individuals in the initial population are screened to obtain the screened initial population; According to the objective function, calculate the fitness value of each individual in the initial population after screening; According to the fitness value of each individual, the best individual is determined from the screened primary population; Determine whether the fitness value of the optimal individual meets the iteration termination condition. If so, determine the variable combination corresponding to the optimal individual as the optimal variable combination. Otherwise, perform crossover and mutation based on the optimal individual to obtain a new generation of population. Start the cycle by screening multiple individuals in the new generation of population according to the constraints until it is determined that the fitness value of the new optimal individual meets the iteration termination condition, and obtain the optimal variable combination based on the new optimal individual.

[0039] Optionally, the parameters of the genetic algorithm are set, for example, the population size (ie, the total number of individuals in the population) is 200, the mutation probability is 0.005, the crossover probability is 0.8, and the threshold of the number of genetic iterations is 100 times.

[0040] Among them, the fitness value is the objective function value, which is used to measure the quality of the individual.

[0041] It can be understood that by using genetic algorithms to optimize the structural dimensions of the inner door panel of an automobile under constraints, not only the efficiency of the optimization design can be improved, but also the quality and reliability of the optimization design solution can be improved.

[0042] In some embodiments, multiple individuals in the initial population are screened according to the automobile door inner panel structural stress and constraint conditions corresponding to each variable combination, including: It is determined whether the structural stress of the inner panel of the automobile door corresponding to each variable combination is less than or equal to the preset structural stress threshold. If so, the individual corresponding to each variable combination is retained; otherwise, the individual corresponding to each variable combination is eliminated.

[0043] It can be understood that by judging whether the structural stress of the inner door panel of the automobile corresponding to each variable combination is less than or equal to the preset structural stress threshold, multiple individuals in the initial population are screened, which can eliminate the structural dimensions that do not meet the performance requirements and narrow the search space, which helps to improve the efficiency of the genetic algorithm, reduce the consumption of computing resources, and improve the reliability of structural dimension optimization.

[0044] Figure 2 A flow chart of the process of determining the proxy model provided by the embodiment of the present invention is as follows: Figure 2 As shown, in some embodiments, the process of determining the proxy model includes: Step 210: construct an initial sample variable combination according to the preset value range of each variable; Step 220: amplify the initial sample variable combination to obtain multiple sample variable combinations; Step 230, determining the automobile door inner panel structural stress labels of multiple sample variable combinations; Step 240: construct an initial proxy model; Step 250: Using a plurality of sample variable combinations as training samples and a plurality of sample variable combinations of automobile door inner panel structural stress labels as sample labels, an initial proxy model is trained. After the training is completed, a proxy model is obtained.

[0045] Optionally, the sample variable combination includes a sample hole size and a sample reinforcement rib size.

[0046] Optionally, an initial proxy model is constructed based on the initial sample variable combination.

[0047] Optionally, the initial proxy model includes an initial convolutional neural network, which includes: multiple initial convolutional layers based on a hard attention mechanism, multiple initial pooling layers, an initial SE residual block layer, and an initial fully connected layer.

[0048] It can be understood that by amplifying the initial sample variable combination, multiple sample variable combinations similar to the initial sample variable combination can be obtained. By using the amplified multiple sample variable combinations as training samples and the automobile door inner panel structural stress labels of the multiple sample variable combinations as sample labels to train the initial proxy model, not only the efficiency of proxy model training can be improved, but also the computational accuracy and adaptability of the proxy model can be improved.

[0049] In some embodiments, training the initial proxy model includes: Input each sample variable combination into the initial proxy model to obtain the predicted value of the automobile door inner panel structural stress corresponding to each sample variable combination output by the initial proxy model; Calculate the loss function value according to the predicted value of the automobile door inner panel structural stress corresponding to each sample variable combination and the automobile door inner panel structural stress label of each sample variable combination; Based on the loss function value, the parameters of the initial proxy model are iteratively optimized.

[0050] Optionally, the loss function is calculated as follows: ; in, represents the loss function value, For the The predicted value of the structural stress of the inner door panel of the automobile corresponding to the sample variable combination, is the total number of sample variable combinations, For the The structural stress labels of the inner door panels of automobiles for the sample variable combinations.

[0051] The following is a description of the vehicle door inner panel structure size optimization device provided by an embodiment of the present invention. The vehicle door inner panel structure size optimization device described below and the vehicle door inner panel structure size optimization method described above can be referenced to each other.

[0052] Figure 3 The schematic diagram of the structure of the vehicle door inner panel structure size optimization device provided by the embodiment of the present invention is as follows: Figure 3 As shown, the automobile door inner panel structure size optimization device 300 includes: A construction unit 310 is used to construct an objective function with multiple structural dimensions of the automobile door inner panel as variables and lightweighting of the automobile door inner panel as a goal, and determine the constraint conditions of the objective function according to a preset structural stress threshold; A determination unit 320, configured to determine a plurality of variable combinations according to a plurality of value ranges of structural dimensions; The prediction unit 330 is used to input each variable combination into a pre-built proxy model to obtain the automobile door inner panel structural stress corresponding to each variable combination output by the proxy model; A solving unit 340 is used to solve the objective function under constraint conditions by using a genetic algorithm according to the structural stress of the automobile door inner panel corresponding to each variable combination, obtain an optimal variable combination, and determine the optimal structural size of the automobile door inner panel based on the optimal variable combination; Among them, the proxy model is trained based on the sample variable combination corresponding to multiple structural dimensions of the sample automobile door inner panel and the automobile door inner panel structural stress label of the sample variable combination.

[0053] Optionally, a genetic algorithm is used to solve the objective function under constraints according to the structural stress of the inner door panel of the automobile corresponding to each variable combination to obtain the optimal variable combination, including: Encode multiple variable combinations to obtain the initial population, where each individual in the initial population represents each variable combination; According to the structural stress and constraint conditions of the inner door panel of the automobile corresponding to each variable combination, multiple individuals in the initial population are screened to obtain the screened initial population; According to the objective function, calculate the fitness value of each individual in the initial population after screening; According to the fitness value of each individual, the best individual is determined from the screened primary population; Determine whether the fitness value of the optimal individual meets the iteration termination condition. If so, determine the variable combination corresponding to the optimal individual as the optimal variable combination. Otherwise, perform crossover and mutation based on the optimal individual to obtain a new generation of population. Start the cycle by screening multiple individuals in the new generation of population according to the constraints until it is determined that the fitness value of the new optimal individual meets the iteration termination condition, and obtain the optimal variable combination based on the new optimal individual.

[0054] Optionally, multiple individuals in the initial population are screened according to the automobile door inner panel structural stress and constraint conditions corresponding to each variable combination, including: It is determined whether the structural stress of the inner panel of the automobile door corresponding to each variable combination is less than or equal to the preset structural stress threshold. If so, the individual corresponding to each variable combination is retained; otherwise, the individual corresponding to each variable combination is eliminated.

[0055] Optionally, the variable combination includes the hole size and the reinforcement rib size of the automobile door inner panel; the constraint condition includes that the structural stress of the automobile door inner panel is less than or equal to a preset structural stress threshold; and the value of the objective function is the weight of the automobile door inner panel.

[0056] Optionally, the proxy model includes a convolutional neural network, which includes: multiple convolutional layers based on a hard attention mechanism, multiple pooling layers, an SE residual block layer, and a fully connected layer.

[0057] Optionally, the process of determining the proxy model includes: Construct the initial sample variable combination according to the preset value range of each variable; Expand the initial sample variable combination to obtain multiple sample variable combinations; Determine the structural stress labels of the inner door panels of automobiles for multiple sample variable combinations; Constructing an initial proxy model; The initial proxy model is trained with multiple sample variable combinations as training samples and the automobile door inner panel structural stress labels of multiple sample variable combinations as sample labels. After the training is completed, the proxy model is obtained.

[0058] Optionally, training the initial proxy model includes: Input each sample variable combination into the initial proxy model to obtain the predicted value of the automobile door inner panel structural stress corresponding to each sample variable combination output by the initial proxy model; Calculate the loss function value according to the predicted value of the automobile door inner panel structural stress corresponding to each sample variable combination and the automobile door inner panel structural stress label of each sample variable combination; Based on the loss function value, the parameters of the initial proxy model are iteratively optimized.

[0059] It should be noted here that the automobile door inner panel structure size optimization device provided in the embodiment of the present invention can implement all the method steps implemented in the above-mentioned automobile door inner panel structure size optimization method embodiment, and can achieve the same technical effect. The parts and beneficial effects of this embodiment that are the same as the method embodiment will not be described in detail here.

[0060] Figure 4A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention, such as Figure 4 As shown, the electronic device may include: a processor (processor) 410 , a communication interface (Communications Interface) 420 , a memory (memory) 430 and a communication bus 440 , wherein the processor 410 , the communication interface 420 , and the memory 430 communicate with each other through the communication bus 440 . The processor 410 can call the logic instructions in the memory 430 to execute a method for optimizing the structural dimensions of an automobile door inner panel, the method comprising: taking multiple structural dimensions of the automobile door inner panel as variables and taking lightweighting of the automobile door inner panel as a goal, constructing an objective function, and determining the constraints of the objective function according to a preset structural stress threshold; determining multiple variable combinations according to the value ranges of the multiple structural dimensions; inputting each variable combination into a pre-constructed proxy model to obtain the structural stress of the automobile door inner panel corresponding to each variable combination output by the proxy model; using a genetic algorithm, according to the structural stress of the automobile door inner panel corresponding to each variable combination, solving the objective function under the constraints to obtain the optimal variable combination, and determining the optimal structural dimensions of the automobile door inner panel based on the optimal variable combination; wherein the proxy model is trained based on sample variable combinations corresponding to multiple structural dimensions of the sample automobile door inner panel and the automobile door inner panel structural stress labels of the sample variable combinations.

[0061] In addition, the logic instructions in the above-mentioned memory 430 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0062] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a method for optimizing the structural dimensions of an automobile door inner panel provided by the above-mentioned methods, the method comprising: taking multiple structural dimensions of an automobile door inner panel as variables and taking lightweighting of the automobile door inner panel as a goal, constructing an objective function, and determining constraints of the objective function according to a preset structural stress threshold; determining multiple variable combinations according to the value ranges of the multiple structural dimensions; inputting each variable combination into a pre-constructed proxy model to obtain the structural stress of the automobile door inner panel corresponding to each variable combination output by the proxy model; using a genetic algorithm, according to the structural stress of the automobile door inner panel corresponding to each variable combination, solving the objective function under constraints to obtain an optimal variable combination, and determining the optimal structural dimensions of the automobile door inner panel based on the optimal variable combination; wherein the proxy model is trained based on sample variable combinations corresponding to multiple structural dimensions of the sample automobile door inner panel and the automobile door inner panel structural stress labels of the sample variable combinations.

[0063] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0064] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0065] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for optimizing the structural dimensions of an automobile door inner panel, characterized in that: include: Taking multiple structural dimensions of the inner door panel of the automobile as variables and lightweighting of the inner door panel of the automobile as a goal, an objective function is constructed, and according to a preset structural stress threshold, a constraint condition of the objective function is determined; Determining multiple variable combinations according to the value ranges of the multiple structural dimensions; Input each variable combination into a pre-built proxy model to obtain the automobile door inner panel structural stress corresponding to each variable combination output by the proxy model; Using a genetic algorithm, according to the structural stress of the automobile door inner panel corresponding to each variable combination, solving the objective function under the constraint condition to obtain an optimal variable combination, and determining the optimal structural size of the automobile door inner panel based on the optimal variable combination; The proxy model is obtained by training based on a sample variable combination corresponding to multiple structural dimensions of a sample automobile door inner panel and a structural stress label of the automobile door inner panel of the sample variable combination.

2. The method for optimizing the structural dimensions of an automobile door inner panel according to claim 1, characterized in that: The genetic algorithm is used to solve the objective function under the constraint conditions according to the structural stress of the inner door panel of the automobile corresponding to each variable combination to obtain the optimal variable combination, including: Encoding the multiple variable combinations to obtain an initial population, wherein each individual of the initial population represents each variable combination; According to the automobile door inner panel structural stress corresponding to each variable combination and the constraint condition, a plurality of individuals in the primary population are screened to obtain a screened primary population; According to the objective function, calculating the fitness value of each individual in the screened primary population; According to the fitness value of each individual, determining the best individual from the screened primary population; Determine whether the fitness value of the optimal individual meets the iteration termination condition. If so, determine that the variable combination corresponding to the optimal individual is the optimal variable combination. Otherwise, perform crossover and mutation based on the optimal individual to obtain a new generation population, and start a cycle by screening multiple individuals in the new generation population according to the constraint conditions until it is determined that the fitness value of the new optimal individual meets the iteration termination condition, and obtain the optimal variable combination based on the new optimal individual.

3. The method for optimizing the structural dimensions of an automobile door inner panel according to claim 2, characterized in that: The screening of multiple individuals in the primary population according to the automobile door inner panel structural stress corresponding to each variable combination and the constraint condition includes: Determine whether the automobile door inner panel structural stress corresponding to each variable combination is less than or equal to the preset structural stress threshold; if so, retain the individual corresponding to each variable combination; otherwise, eliminate the individual corresponding to each variable combination.

4. The method for optimizing the structural dimensions of an automobile door inner panel according to any one of claims 1 to 3, characterized in that: The variable combination includes the hole size and the reinforcement rib size of the automobile door inner panel; the constraint condition includes that the structural stress of the automobile door inner panel is less than or equal to the preset structural stress threshold; and the value of the objective function is the weight of the automobile door inner panel.

5. The method for optimizing the structural dimensions of an automobile door inner panel according to claim 1, characterized in that: The proxy model includes a convolutional neural network, which includes: multiple convolutional layers based on a hard attention mechanism, multiple pooling layers, an SE residual block layer and a fully connected layer.

6. The method for optimizing the structural dimensions of an automobile door inner panel according to claim 1, characterized in that: The process of determining the proxy model includes: Construct the initial sample variable combination according to the preset value range of each variable; Amplifying the initial sample variable combination to obtain multiple sample variable combinations; Determining a stress label of the automobile door inner panel structure for the plurality of sample variable combinations; Constructing an initial proxy model; The initial proxy model is trained using the multiple sample variable combinations as training samples and the automobile door inner panel structural stress labels of the multiple sample variable combinations as sample labels. After the training is completed, the proxy model is obtained.

7. The method for optimizing the structural dimensions of an automobile door inner panel according to claim 6, characterized in that: The training of the initial proxy model comprises: Inputting each sample variable combination into the initial proxy model to obtain a predicted value of the automobile door inner panel structural stress corresponding to each sample variable combination output by the initial proxy model; Calculate the loss function value according to the predicted value of the automobile door inner panel structural stress corresponding to each sample variable combination and the automobile door inner panel structural stress label of each sample variable combination; Based on the loss function value, the parameters of the initial proxy model are iteratively optimized.

8. A device for optimizing the structural dimensions of an automobile door inner panel, characterized in that: include: A construction unit, used to construct an objective function with multiple structural dimensions of the automobile door inner panel as variables and lightweighting of the automobile door inner panel as a goal, and determine the constraint conditions of the objective function according to a preset structural stress threshold; A determination unit, used to determine a plurality of variable combinations according to the value ranges of the plurality of structural dimensions; A prediction unit, used for inputting each variable combination into a pre-built proxy model to obtain the automobile door inner panel structural stress corresponding to each variable combination output by the proxy model; A solving unit, configured to solve the objective function under the constraint conditions by using a genetic algorithm according to the structural stress of the automobile door inner panel corresponding to each variable combination, to obtain an optimal variable combination, and to determine an optimal structural size of the automobile door inner panel based on the optimal variable combination; The proxy model is obtained by training based on a sample variable combination corresponding to multiple structural dimensions of a sample automobile door inner panel and a structural stress label of the automobile door inner panel of the sample variable combination.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the method for optimizing the structural dimensions of an automobile door inner panel as claimed in any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for optimizing the structural dimensions of an automobile door inner panel as claimed in any one of claims 1 to 7 is implemented.

Images