Method and device for determining arrangement parameters of a door hinge and lock point, and electronic device

Abstract

The application relates to the technical field of vehicle design and manufacturing, in particular to a method and device for determining vehicle door hinge and lock point arrangement parameters and electronic equipment. The method comprises the following steps: obtaining a target stiffness performance of a vehicle door; inputting the target stiffness performance into a preset neural network model, wherein the preset proxy model is obtained by training based on a corresponding relationship between a hinge-lock point sample group and a vehicle door stiffness performance index; and receiving target hinge position parameters and target lock point position parameters corresponding to the target stiffness performance output by the preset neural network model. Thus, the mapping relationship between the hinge-lock point parameters and the vehicle door stiffness performance is established by pre-training the neural network model, and the target stiffness performance is directly input into the model for reverse solving in the design stage, thereby solving the problems of long design cycle and low performance standard rate caused by repeated trial and error depending on the experience of engineers in the related art, and realizing intelligent design of the vehicle door hinge and lock point arrangement.

Description

Technical Field

[0001] This application relates to the field of vehicle design and manufacturing technology, and in particular to a method, device and electronic device for determining the arrangement parameters of door hinges and locking points. Background Technology

[0002] As an important component of the vehicle body, the structural performance of the door directly affects the safety, comfort, and durability of the entire vehicle. Among them, the arrangement of the hinges and locking points is a key design parameter that determines the stiffness performance of the door, and directly affects the deformation characteristics and reliability of the door under load.

[0003] In related technologies, the determination of the door hinge and locking point layout parameters is usually done by engineers based on experience with historical models and design specifications to initially determine the layout positions of the hinges and locking points. Then, finite element simulation analysis is used to evaluate the performance indicators of the door, such as vertical stiffness, waistline stiffness, and window frame stiffness. If the performance does not meet the standards, the parameters need to be adjusted and the simulation needs to be repeated until the design requirements are met.

[0004] However, this method relies heavily on engineers' experience and involves repeated trial and error, resulting in long design cycles and low performance compliance rates, which urgently need to be addressed. Summary of the Invention

[0005] This application provides a method for determining the layout parameters of door hinges and locking points, in order to solve the problems of long design cycles and low performance compliance rates caused by repeated trial and error based on engineers' experience in related technologies, and realizes intelligent design of door hinges and locking point layout.

[0006] To achieve the above objectives, the first aspect of this application proposes a method for determining the arrangement parameters of a car door hinge and locking points, comprising the following steps:

[0007] Obtain the target stiffness performance of the vehicle door; The target stiffness performance is input into a preset neural network model, which is trained based on the correspondence between the hinge-lock point sample set and the door stiffness performance index. Receive the target hinge position parameters and target locking point position parameters corresponding to the target stiffness performance, output by the preset neural network model.

[0008] According to one embodiment of this application, before inputting the target stiffness performance into a preset neural network model, the method further includes: Obtain hinge position parameters and lock point position parameters from multiple historical finite element models of car doors; Based on the hinge position parameters and the locking point position parameters, multiple hinge-locking point sample groups are determined, and the door stiffness performance index corresponding to the multiple hinge-locking point sample groups is calculated by finite element simulation. Multiple hinge-lock point sample groups are used as input samples, and the door stiffness performance index is used as output samples. A training sample set is constructed based on the input samples and the output samples. Based on a preset partitioning ratio, the training sample set is divided into a training set and a test set. Based on the training set, a preset neural network is trained until a preset training termination condition is reached to obtain an initial neural network model. The initial neural network model is tested using the test set, and the preset neural network model is obtained when the test results meet the preset requirements.

[0009] According to one embodiment of this application, the step of testing the initial neural network model using the test set and obtaining the preset neural network model when the test results meet preset requirements includes: The hinge-lock point sample group in the test set is input into the initial neural network model to obtain the predicted stiffness performance index. Calculate the coefficient of determination and root mean square error between the predicted stiffness performance index and the corresponding door stiffness performance index in the test set; If the determination coefficient is greater than or equal to a first preset threshold and the root mean square error is less than or equal to a second preset threshold, the test result is determined to meet the preset requirements.

[0010] According to one embodiment of this application, the step of calculating the door stiffness performance indicators corresponding to multiple hinge-lock point sample groups through finite element simulation includes: Establish a parametric finite element model of the hinge position and the locking point position; The hinge position parameters and lock point position parameters in each hinge-lock point sample group are written into a preset driving script, and the hinge position and lock point position of the parameterized finite element model are updated based on the preset driving script. Finite element simulation analysis was performed on the parameterized finite element model after each update to calculate the door stiffness performance index corresponding to each hinge-lock point sample group.

[0011] According to one embodiment of this application, determining a plurality of hinge-lock point sample groups based on the hinge position parameters and the locking point position parameters includes: Normal distribution fitting is performed on the hinge position parameters and the locking point position parameters respectively to obtain the mean and standard deviation of the hinge position parameters and the mean and standard deviation of the locking point position parameters; Based on the mean and standard deviation of the hinge position parameters and the mean and standard deviation of the locking point position parameters, the target value range of the hinge position parameters and the target value range of the locking point position parameters are determined. Based on the target value range of the hinge position parameter and the target value range of the locking point position parameter, multiple hinge-locking point sample groups are generated using an optimized Latin hypercube strategy.

[0012] According to the method for determining the layout parameters of door hinges and locking points proposed in this application, by obtaining the target stiffness performance of the door and inputting the target stiffness performance into a preset neural network model, the method can receive the target hinge position parameters and target locking point position parameters corresponding to the target stiffness performance output by the preset neural network model. Thus, by pre-training the neural network model to establish a mapping relationship between hinge locking point parameters and door stiffness performance, and by directly inputting the target stiffness performance into the model for reverse engineering during the design phase, the method solves the problems of long design cycles and low performance compliance rates caused by repeated trial and error based on engineer experience in related technologies, thereby realizing intelligent design of door hinge and locking point layout.

[0013] To achieve the above objectives, a second aspect of this application provides a device for determining the arrangement parameters of a car door hinge and locking point, comprising: The acquisition module is used to acquire the target stiffness performance of the vehicle door. The input module is used to input the target stiffness performance into a preset neural network model, wherein the preset surrogate model is trained based on the correspondence between the hinge-lock point sample group and the door stiffness performance index; The receiving module is used to receive the target hinge position parameters and target locking point position parameters output by the preset neural network model, which correspond to the target stiffness performance.

[0014] According to one embodiment of this application, before inputting the target stiffness performance into a preset neural network model, the input module further includes: The acquisition unit is used to acquire hinge position parameters and lock point position parameters from multiple historical finite element models of car doors. The processing unit is used to determine multiple hinge-lock point sample groups based on the hinge position parameters and the locking point position parameters, and to calculate the door stiffness performance index corresponding to the multiple hinge-lock point sample groups through finite element simulation. A construction unit is used to take multiple hinge-lock point sample groups as input samples and the door stiffness performance index as output samples, and construct a training sample set based on the input samples and the output samples; The training unit is used to divide the training sample set into a training set and a test set based on a preset division ratio, and to train a preset neural network based on the training set until a preset training termination condition is reached to obtain an initial neural network model. The testing unit is used to test the initial neural network model using the test set, and to obtain the preset neural network model when the test results meet the preset requirements.

[0015] According to one embodiment of this application, the test unit is specifically used for: The hinge-lock point sample group in the test set is input into the initial neural network model to obtain the predicted stiffness performance index. Calculate the coefficient of determination and root mean square error between the predicted stiffness performance index and the corresponding door stiffness performance index in the test set; If the determination coefficient is greater than or equal to a first preset threshold and the root mean square error is less than or equal to a second preset threshold, the test result is determined to meet the preset requirements.

[0016] According to one embodiment of this application, the processing unit is specifically used for: Establish a parametric finite element model of the hinge position and the locking point position; The hinge position parameters and lock point position parameters in each hinge-lock point sample group are written into a preset driving script, and the hinge position and lock point position of the parameterized finite element model are updated based on the preset driving script. Finite element simulation analysis was performed on the parameterized finite element model after each update to calculate the door stiffness performance index corresponding to each hinge-lock point sample group.

[0017] According to one embodiment of this application, the processing unit is specifically used for: Based on the mean and standard deviation of the hinge position parameters and the mean and standard deviation of the locking point position parameters, the target value range of the hinge position parameters and the target value range of the locking point position parameters are determined. Based on the target value range of the hinge position parameter and the target value range of the locking point position parameter, multiple hinge-locking point sample groups are generated using an optimized Latin hypercube strategy.

[0018] The device for determining the arrangement parameters of door hinges and locking points according to the embodiments of this application obtains the target stiffness performance of the door and inputs the target stiffness performance into a preset neural network model. It can then receive the target hinge position parameters and target locking point position parameters corresponding to the target stiffness performance output by the preset neural network model. Thus, by pre-training the neural network model to establish a mapping relationship between the hinge locking point parameters and the door stiffness performance, and by directly inputting the target stiffness performance into the model for reverse engineering during the design phase, it solves the problems of long design cycles and low performance compliance rates caused by repeated trial and error based on engineer experience in related technologies, achieving intelligent design of door hinge and locking point arrangements.

[0019] To achieve the above objectives, a third aspect of this application provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and capable of running on the processor, wherein the processor executes the program to implement the method for determining the arrangement parameters of the door hinge and locking point as described in the above embodiments.

[0020] To achieve the above objectives, a fourth aspect of this application provides a computer-readable storage medium having a computer program stored thereon, which is executed by a processor to implement the method for determining the arrangement parameters of the door hinges and locking points as described in the above embodiments.

[0021] To achieve the above objectives, a fifth aspect of this application provides a computer program product comprising a computer program that, when executed by a processor, is used to implement the method for determining the arrangement parameters of the door hinges and locking points as described in the above embodiments.

[0022] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0023] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 This is a flowchart illustrating a method for determining the arrangement parameters of a car door hinge and locking point according to an embodiment of this application. Figure 2 This is a schematic diagram of a car door structure according to an embodiment of this application; Figure 3 This is a schematic diagram of the probability distribution of hinge-lock point parameters according to an embodiment of this application; Figure 4 This is a flowchart of parametric finite element modeling and DOE (Design of Experiments) analysis according to an embodiment of this application; Figure 5 This is a schematic diagram illustrating the construction of a full dataset according to an embodiment of this application; Figure 6 A schematic diagram of the full dataset after supplementing the design feature parameters according to an embodiment of this application; Figure 7 This is a schematic diagram of the training process of a random forest model according to an embodiment of this application; Figure 8 This is a schematic diagram of an intelligent tool software architecture according to an embodiment of this application; Figure 9This is a schematic diagram of the main interface of an intelligent tool according to an embodiment of this application; Figure 10 This is a schematic diagram of the influence coefficients of design parameters according to an embodiment of this application; Figure 11 This is a schematic diagram of the correlation matrix of stiffness performance indicators according to an embodiment of this application; Figure 12 This is a block diagram of a device for determining the arrangement parameters of door hinges and locking points according to an embodiment of this application; Figure 13 This is a schematic diagram of the structure of an electronic device provided according to an embodiment of this application. Detailed Implementation

[0024] The embodiments of this application are described in detail below. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0025] The following describes, with reference to the accompanying drawings, a method, apparatus, and electronic device for determining the arrangement parameters of door hinges and locking points according to embodiments of this application. First, the method for determining the arrangement parameters of door hinges and locking points according to embodiments of this application will be described with reference to the accompanying drawings.

[0026] Figure 1 This is a flowchart of a method for determining the arrangement parameters of a car door hinge and locking point according to an embodiment of this application.

[0027] like Figure 1 As shown, the method for determining the door hinge and locking point arrangement parameters includes the following steps: In step S101, the target stiffness performance of the door is obtained.

[0028] It is understood that, in the embodiments of this application, the target stiffness performance refers to the stiffness index value expected to be achieved by the door design (i.e., welded assembly), which may include at least one of the following: vertical stiffness (used to evaluate the door's load-bearing capacity in the vertical direction), lateral stiffness at the front point of the window frame (used to evaluate the lateral resistance to deformation of the front region of the window frame), lateral stiffness at the rear point of the window frame (used to evaluate the lateral resistance to deformation of the rear region of the window frame), outer point stiffness of the waistline (used to evaluate the local stiffness of the outer region of the door's waistline), and inner point stiffness of the waistline (used to evaluate the local stiffness of the inner region of the door's waistline). These indices reflect the door's resistance to deformation when subjected to loads in different directions and are key parameters for evaluating the structural performance of the door.

[0029] Specifically, in practical engineering applications, designers can set one or more sets of target stiffness performance values ​​for vehicle doors based on the design requirements and performance targets of specific vehicle models. These values ​​serve as input for determining subsequent hinge and locking point layout parameters. Methods for obtaining target stiffness performance include, but are not limited to: reading preset target values ​​from the vehicle performance database, receiving expected values ​​input by designers through a user interface, or automatically generating corresponding target performance values ​​based on vehicle model parameters.

[0030] In step S102, the target stiffness performance is input into a preset neural network model. The preset surrogate model is trained based on the correspondence between the hinge-lock point sample group and the door stiffness performance index.

[0031] It is understood that, in the embodiments of this application, as Figure 2 As shown, the hinge-lock point sample group refers to a set of parameter values ​​consisting of the lower hinge position L1, the upper hinge position L2, and the lock point position L3 of the door. L1 and L2 are the distances from the lowest point of the bottom of the door along the Z-axis of the vehicle coordinate system to the intersection of the mid-surface of the fixed and movable hinge plates and the rotation axis, respectively. L3 is the distance from the lowest point of the bottom of the door along the Z-axis to the lock point. Each set of L1, L2, and L3 constitutes a hinge-lock point sample group, representing one arrangement of hinges and lock points. The door stiffness performance indicators here refer to quantitative indicators used to evaluate the deformation resistance of the door structure (such as vertical stiffness, lateral stiffness at the front point of the window frame, lateral stiffness at the rear point of the window frame, stiffness at the outer point of the waistline, and stiffness at the inner point of the waistline). These indicators are calculated through finite element simulation analysis and reflect the stiffness characteristics of the door under different load conditions.

[0032] Specifically, the target stiffness performance obtained in step S101 is used as input and fed into a preset neural network model. Since the preset neural network model has learned the positive mapping relationship between hinge-lock point parameters and door stiffness performance, when the target stiffness performance of the door is input, the preset neural network model can perform reverse calculation through its inherent mapping mechanism, providing a calculation basis for the subsequent output of the corresponding hinge-lock point parameters.

[0033] It should be noted that since the door stiffness performance includes five aspects: vertical stiffness, front point lateral stiffness of the window frame, rear point lateral stiffness of the window frame, outer point stiffness of the waistline, and inner point stiffness of the waistline, the embodiments of this application can train models for each of the five stiffness indices of the door separately, resulting in five surrogate models. That is, each of the five stiffness indices corresponds to an independent surrogate model, rather than a single model outputting all five indices simultaneously. The preset neural network model in the embodiments of this application does not refer to a single neural network model, but can encompass five surrogate models, corresponding to the vertical stiffness prediction model, the front point lateral stiffness prediction model, the rear point lateral stiffness prediction model, the outer point stiffness prediction model, and the inner point stiffness prediction model, respectively. Each model takes hinge position parameters and lock point position parameters as input and the corresponding single stiffness performance index as output, and is trained using an independent training dataset. In practical applications, when the target stiffness performance is input, the five models calculate independently, and finally, the recommended design parameters are obtained by comprehensively solving the problem through a multi-objective optimization algorithm.

[0034] In step S103, the target hinge position parameters and target locking point position parameters corresponding to the target stiffness performance are received from the preset neural network model output.

[0035] It is understood that, in this embodiment, the target hinge position parameter is the hinge installation position that matches the target stiffness performance, calculated through a preset neural network model. This includes the lower hinge position L1 and the upper hinge position L2, which together determine the spatial arrangement of the hinges on the door. The target lock point position parameter refers to the door lock point installation position that matches the target stiffness performance, calculated through a preset neural network model, namely the door lock point position L3. This parameter determines the spatial arrangement of the lock body on the door.

[0036] In other words, after the target stiffness performance is input into the preset neural network model in step S102, the model performs a reverse calculation of the input target stiffness performance based on the mapping relationship between hinge-locking point parameters and stiffness performance established internally. Since the preset neural network model has learned the forward mapping relationship from "hinge position parameters L1, L2 and locking point position parameter L3" to "door stiffness performance index" during the training phase, when the target stiffness performance is input, the preset neural network model can search within its parameter space for a set of hinge-locking point parameter combinations that allow the actual output performance to approximate the target performance. This search process can be based on iterative optimization using an optimization algorithm, or it can be directly calculated using the model's built-in reverse mapping mechanism. The final output target hinge position parameters include the lower door hinge position L1 and the upper door hinge position L2, and the target locking point position parameter is the door locking point position L3. These three together constitute a complete hinge and locking point arrangement scheme. Designers can directly apply this set of parameters to the 3D digital model design of the door as the final installation positions of the hinges and locking points.

[0037] To facilitate understanding, the following details how to obtain the preset neural network model.

[0038] Optionally, in some embodiments, before inputting the target stiffness performance into the preset neural network model, the method further includes: obtaining hinge position parameters and locking point position parameters from multiple historical finite element models of car doors; determining multiple hinge-locking point sample groups based on the hinge position parameters and locking point position parameters, and calculating the car door stiffness performance index corresponding to the multiple hinge-locking point sample groups through finite element simulation; using the multiple hinge-locking point sample groups as input samples and the car door stiffness performance index as output samples, constructing a training sample set based on the input samples and output samples; dividing the training sample set into a training set and a test set based on a preset partitioning ratio, training the preset neural network based on the training set until a preset training termination condition is reached to obtain an initial neural network model; testing the initial neural network model using the test set, and obtaining the preset neural network model when the test results meet preset requirements.

[0039] It is understood that, in this embodiment, the historical door finite element model refers to a door finite element analysis model that has been established and validated in past vehicle projects (i.e., vehicle projects that have been mass-produced or whose designs have been completed). These models contain complete door geometry, material properties, weld point distribution, hinge and lock point assembly relationships, and other information, and can serve as a data source for extracting hinge position parameters and lock point position parameters. The preset partitioning ratio refers to the proportion of the training sample set divided into a training set and a test set, for example, 80%:20% or 70%:30%, where the training set is used for model training and the test set is used for model validation. The preset training termination condition refers to the criteria for terminating neural network training, which could be reaching the maximum number of iterations, the loss function converging to a threshold, or the validation set accuracy no longer improving. The preset requirements refer to the performance standards that the model needs to meet to pass the test set validation.

[0040] Specifically, during the training of the pre-defined neural network model, multiple historical finite element models of car doors are first acquired. Hinge position parameters (including the lower hinge position L1 and the upper hinge position L2) and locking point position parameters (locking point position L3) are extracted from these models. Based on the distribution characteristics of these historical parameters, multiple hinge-locking point sample sets are generated, each containing a set of values ​​for L1, L2, and L3. Next, for each hinge-locking point sample set, the corresponding door stiffness performance index is calculated using finite element simulation. The simulation calculation must ensure that the model parameters can be accurately driven, and that the calculation results are repeatable and consistent. Then, multiple hinge-locking point sample sets are used as input samples, and the corresponding door stiffness performance indexes are used as output samples to construct a training sample set. Each pair of input and output data in this training sample set completely records the mapping relationship from design parameters to performance indexes, providing a data foundation for subsequent model training. Finally, the training sample set is divided into a training set and a test set according to a pre-defined partitioning ratio. The training set drives the model to learn the mapping relationship between parameters and performance, while the test set evaluates the model's generalization ability. The training set is used to iteratively train a pre-defined neural network until a pre-defined training termination condition is met, such as the loss function value stabilizing and no longer decreasing or reaching a pre-defined maximum number of iterations. At this point, an initial neural network model is obtained. Finally, the initial neural network model is tested using the test set to evaluate its predictive performance on unseen data. When the test results meet pre-defined requirements, the initial neural network model is designated as the pre-defined neural network model and can be used for subsequent back-inverse calculation of target stiffness performance.

[0041] As one possible implementation, in some embodiments, determining multiple hinge-lock point sample groups based on hinge position parameters and lock point position parameters includes: fitting the hinge position parameters and lock point position parameters to a normal distribution to obtain the mean and standard deviation of the hinge position parameters and the mean and standard deviation of the lock point position parameters; determining the target value range of the hinge position parameters and the target value range of the lock point position parameters based on the mean and standard deviation of the hinge position parameters and the lock point position parameters; and generating multiple hinge-lock point sample groups using an optimized Latin hypercube strategy based on the target value range of the hinge position parameters and the target value range of the lock point position parameters.

[0042] It is understood that, in the embodiments of this application, normal distribution fitting refers to performing statistical analysis on sample data to determine whether it conforms to a normal distribution and calculating the mean of the distribution. and standard deviation The process. The target value range refers to the range based on the "3" of the normal distribution. The parameter value range determined by the "principle" is, i.e., [ -3 +3 This interval can cover approximately 99.7% of the possible values, ensuring sample diversity while avoiding extremely unreasonable design parameters. The optimized Latin hypercube strategy refers to an improved experimental design method that can generate more evenly distributed sample points in a multidimensional design space.

[0043] Specifically, firstly, the hinge position parameters and lock point position parameters extracted from the finite element model of historical car doors can be fitted with normal distributions. Taking the lower hinge position L1 of the car door as an example, the mean value of L1 can be calculated by statistically analyzing a large number of historical car models. and standard deviation Similarly, by fitting L2 and L3 respectively, their respective means are obtained. , and , This step transforms discrete historical data into statistically significant distribution characteristics, providing a mathematical basis for determining a reasonable design space. Then, based on the mean and standard deviation obtained from the normal distribution fitting, the target value ranges for each parameter (i.e., hinge position parameters (including the lower door hinge position L1 and the upper door hinge position L2) and lock point position parameters (door lock point position L3)) are determined. The target value range for L1 is […]. -3 +3 This interval can serve as the random sampling space for subsequent data augmentation DOE ensemble optimization; the same applies to L2 and L3. Figure 3As shown. Then, the optimized Latin hypercube strategy can be used to randomly sample within the target value ranges of the L1~L3 parameters to generate multiple hinge-lock point sample groups (taking 500 samples per car door as an example, the data augmentation factor can reach 500 times, significantly expanding the scale of the training samples). The optimized Latin hypercube, as an efficient experimental design method, can generate more evenly distributed sample points in a multidimensional design space. Compared with simple random sampling, the optimized Latin hypercube can avoid sample point clustering or overlap, and more comprehensively cover the entire design space with fewer sample points.

[0044] As one possible approach, in some embodiments, the stiffness performance indicators of the vehicle door corresponding to multiple hinge-lock point sample groups are calculated through finite element simulation, including: establishing a parameterized finite element model of the hinge position and the lock point position; writing the hinge position parameters and lock point position parameters in each hinge-lock point sample group into a preset driving script, updating the hinge position and lock point position of the parameterized finite element model based on the preset driving script; performing finite element simulation analysis on the parameterized finite element model after each update, and calculating the stiffness performance indicators of the vehicle door corresponding to each hinge-lock point sample group.

[0045] It is understood that a parametric finite element model refers to a finite element model whose geometric dimensions, material properties, or assembly positions can be driven by external parameters. In this embodiment, it specifically refers to a car door finite element model where the hinge position and locking point position can be driven by parameters L1, L2, and L3. The preset driving script refers to a pre-written program script used to automatically update the parameters of the parametric finite element model. For example, this script can be written in TCL (Tool Command Language) and can automatically adjust the hinge position and locking point position, while ensuring that the distribution of weld points around the hinge changes accordingly as the hinge position moves.

[0046] Specifically, such as Figure 4As shown, firstly, a parametric finite element model (such as an HM (HyperMesh) finite element model file) can be established to represent the hinge and locking point positions. This model differs from a standard finite element model in that the hinge and locking point positions are not fixed geometric coordinates, but variables controlled by external parameters L1, L2, and L3. When these parameters change, the hinges and locking points in the model automatically move to new positions while maintaining the rationality and consistency of other structural features (such as weld distribution and material properties). Secondly, the L1, L2, and L3 values ​​from the hinge-lock sample group are written into a preset driving script (such as a TCL script). This script converts the parameter values ​​into model update instructions according to predefined syntax rules, enabling the movement of the hinge and locking point positions and ensuring that the weld distribution around the hinge changes accordingly with the hinge position movement. Then, the preset driving script is executed to update the hinge and locking point positions of the parametric finite element model to match the current hinge-lock sample group. This process is fully automated, requiring no manual intervention, ensuring the efficiency and accuracy of batch processing. After the parametric finite element model is updated, simulation analysis is performed on the parametric finite element model in its current state. Calculations are performed according to preset working conditions (such as vertical loading, lateral loading, etc.) to obtain the door stiffness performance index corresponding to the hinge-lock point sample group. Finally, the above process of "parameter writing - script execution - finite element model update - simulation calculation" is repeated for each hinge-lock point sample group. By traversing all hinge-lock point sample groups, the door stiffness performance index corresponding to each hinge-lock point sample group is obtained. This batch simulation process can be automated using a simulation integration platform (such as ISIGHT), integrating elements such as parametric driving scripts, finite element model files, mesh files, and working condition calculation files, significantly improving the efficiency and consistency of data generation.

[0047] Furthermore, after generating multiple hinge-lock point sample groups and obtaining the corresponding door stiffness performance indicators for each sample group through finite element simulation calculation, the door stiffness performance indicators obtained from the simulation calculation can be preprocessed to ensure the integrity and reliability of the data, providing a high-quality data foundation for subsequent model training.

[0048] Specifically, such as Figure 5As shown, in large-scale batch simulations, due to mesh distortion, non-convergence, or computational interruption, stiffness performance indices of some hinge-lock point sample groups may be missing. To address this issue, the K-nearest neighbor algorithm can be used to fill in missing values. The basic principle of this algorithm is: for a sample group with missing values, find the K complete sample groups in the dataset that are closest to its hinge-lock point parameters (L1, L2, L3). Based on the stiffness performance indices of these K sample groups, estimate reasonable values ​​for the missing terms through weighted averaging or majority voting, thereby filling in the missing data. In simulation calculations, due to numerical calculation errors, model setup problems, or abnormal convergence of contacts, the stiffness performance indices of some hinge-lock point sample groups may deviate significantly from the normal range, forming outliers. To address this issue, the Z-score method can be used for outlier detection and processing. The principle of this method is: first, calculate the mean and standard deviation of a certain stiffness performance index for all sample groups; then calculate the deviation of the index from the mean for each sample group (i.e., Z-score = |x- | / When the Z-score exceeds a preset threshold (e.g., 3), the corresponding indicator for that sample group is immediately identified as an outlier. Identified outliers can be removed or replaced with reasonable estimates based on the actual situation.

[0049] After the above preprocessing, each hinge-lock point sample group and its corresponding, padded and processed stiffness performance indicators are integrated to form a complete full dataset (i.e., training sample set). In this dataset, each hinge-lock point sample group has a corresponding set of complete and reliable stiffness performance indicators, providing a high-quality data foundation for subsequent neural network model training.

[0050] As one possible implementation, in some embodiments, an initial neural network model is tested using a test set, and a preset neural network model is obtained when the test results meet preset requirements. This includes: inputting the hinge-lock point sample group from the test set into the initial neural network model to obtain a predicted stiffness performance index; calculating the coefficient of determination and root mean square error between the predicted stiffness performance index and the corresponding door stiffness performance index in the test set; and determining that the test results meet the preset requirements when the coefficient of determination is greater than or equal to a first preset threshold and the root mean square error is less than or equal to a second preset threshold.

[0051] Understandably, the coefficient of determination is R. 2, is a statistical indicator used to measure the goodness of fit of a regression model. Its value ranges from 0 to 1; the closer it is to 1, the stronger the explanatory power of the preset neural network model and the higher its prediction accuracy. Both the first and second preset thresholds can be pre-set by researchers in the field, obtained through a limited number of experiments, or obtained through a limited number of computer simulations; no specific limitations are made here. In this embodiment, the first preset threshold can be set to 0.96, and the second preset threshold can be set to 5%.

[0052] Specifically, firstly, the hinge-lock point sample groups from the test set can be input into the initial neural network model. The predicted stiffness performance index can be obtained through the model's forward computation. Since the hinge-lock point sample groups in the test set were not used during model training, their predictive ability for unknown data can be objectively evaluated. For each hinge-lock point sample group, the model can output a set of predicted values ​​(including predicted results for vertical stiffness, lateral stiffness at the front and rear points of the window frame, stiffness at the outer and inner points of the waistline). Secondly, the coefficient of determination and root mean square error (RMSE) between the predicted stiffness performance index and the corresponding real door stiffness performance index in the test set can be calculated. The coefficient of determination can assess the model's overall explanatory power for data variation; the RMSE directly measures the average deviation between the predicted and actual values, with a smaller value indicating higher prediction accuracy. These two indicators can jointly evaluate the model's predictive performance from different dimensions, avoiding the bias of a single indicator. Then, the calculated coefficient of determination is compared with a first preset threshold (e.g., 0.96), and the root mean square error is compared with a second preset threshold (e.g., 5%). If the coefficient of determination is greater than or equal to the first preset threshold and the root mean square error is less than or equal to the second preset threshold, the test result is considered to meet the preset requirements, and the initial neural network model can be determined as the preset neural network model. If the test result does not meet the preset requirements, the reasons need to be analyzed and corresponding measures taken, such as adjusting the model structure, optimizing hyperparameters, expanding training data, or improving feature engineering. Then, training and testing are repeated until a model that meets the accuracy requirements is obtained.

[0053] For example, after constructing the full dataset (i.e., the training sample set), auxiliary design features can be selected based on design and simulation experience, extracting additional design feature parameters from historical door data. The selection principles can include two aspects: first, classification based on current design requirements, such as vehicle type parameters (sedan, SUV (Sport Utility Vehicle), MPV (Multi-Purpose Vehicle)), door type (front or rear door), whether integrated modular design is used, project code, etc.); second, based on design and simulation experience, parameters with significant impact and distinguishing characteristics on various door stiffness performances are selected. Subsequently, principal component analysis is performed on the selected structural feature parameters, extracting components with a variance contribution rate of over 80%. Based on this, the load characteristics of each design feature parameter are analyzed, core features are selected, and these are added as parameters to the full dataset, resulting in a full dataset with supplemented design feature parameters, such as... Figure 6 As shown. Secondly, the full dataset with supplemented design feature parameters is used as input, and various classic machine learning algorithms (such as linear regression, logistic regression, support vector regression, decision trees, random forests, etc.) are used to train models on five stiffness performance indicators of the car door (i.e., vertical stiffness, lateral stiffness at the front point of the window frame, lateral stiffness at the rear point of the window frame, stiffness at the outer point of the waistline, and stiffness at the inner point of the waistline), resulting in five sets of surrogate models (i.e., pre-set neural network models). After training, the coefficient of determination and root mean square error can be used to evaluate each surrogate model. The evaluation results show that the coefficient of determination of the decision tree and random forest algorithms performs best, with the random forest algorithm showing particularly outstanding performance. From the training process of the random forest algorithm, as shown... Figure 7 As shown, the sample learning and validation curves reveal that as the number of samples increases, the training set score and cross-validation score gradually converge, with the coefficient of determination consistently above 0.99, indicating that the neural network model possesses excellent fitting performance and generalization ability. During the training process of the neural network model, for the best-performing random forest algorithm, the feature importance coefficients of all design parameters (i.e., L1, L2, L3) and design feature parameters to each door stiffness performance index were calculated. The feature importance coefficients of the five surrogate models were merged to form a complete feature importance matrix. This matrix provides a clear view of the importance of each design parameter and design feature parameter to each door stiffness performance index. Based on this analysis, the relationship between the parameters and performance of the neural network model can be interpreted in engineering, providing designers with quantitative design guidance, clarifying the influence weight of each design parameter on door stiffness performance, and thus enabling targeted optimization of hinge and locking point arrangements during the conceptual design phase.

[0054] It is understood that the embodiments of this application employ five independent random forest regression models, corresponding to the prediction of five door stiffness performance indicators. Each model is an independent random forest regressor with the same ensemble learning architecture, but is trained and optimized independently. This ensemble learning architecture can be based on a Bagging (Bootstrap Aggregating) strategy (an ensemble learning strategy whose core idea is to build multiple base learners and combine their prediction results to obtain better generalization performance than a single learner), consisting of 100 CART (Classification and Regression Tree) regression trees (a decision tree algorithm that can handle both classification problems (i.e., CART classification trees) and regression problems (i.e., CART regression trees; in this embodiment, CART regression trees are used to predict continuous stiffness performance indicators) in parallel. A training subset is generated through bootstrapping sampling. The maximum depth of each tree is set to 10, the minimum number of samples required for internal node splits is 5, the minimum number of samples for leaf nodes is 2, the number of features considered in each split is the square root of the number of input features, and the final prediction value is the arithmetic mean of the prediction results of all decision trees. Each model's input features include hinge position parameters L1 and L2, lock point position parameter L3, and structural feature parameters such as vehicle type and door type; the output is the predicted value of the corresponding single stiffness performance index. Before model training, a large amount of sample data obtained through parametric finite element simulation is preprocessed. The K-nearest neighbor algorithm is used to fill missing values, and the Z-score method is used to detect and handle outliers, forming a high-quality full dataset. During training, 5-fold cross-validation and grid search are used to optimize hyperparameters to ensure optimal model performance. Validation on the test set shows that the determination coefficients of the five surrogate models all exceed 0.99, and the root mean square error of prediction is controlled within 5%, demonstrating excellent fitting ability and generalization performance.

[0055] After training and evaluating the surrogate models, it is necessary to develop intelligent design tools based on these models to transform the model capabilities into software applications that can be directly used by engineering designers. The development of intelligent tools mainly includes three levels: functional logic design, software architecture design, and software development and implementation.

[0056] At the functional logic design level, the core function of the intelligent tool revolves around the stiffness performance design requirements of the door hinge and locking point arrangement, and can include two main functional modules: a prediction module and a multi-objective optimization module. The prediction module supports designers in evaluating the performance of specific design schemes: after the user sets the hinge position parameters, locking point position parameters, and structural feature parameters, clicking to trigger prediction allows the tool to calculate and display the predicted values ​​of the five door stiffness performance parameters corresponding to that set in real time, while simultaneously providing performance evaluation limit warnings and intuitively identifying unsatisfactory performance indicators. During the user's adjustment of design parameters, the tool can display the expected trend and amount of change of each parameter in real time until the user clicks to trigger prediction, at which point the dynamic display disappears. The multi-objective optimization module supports automatic optimization of design schemes: the user locks the structural feature parameters, gives the design range of the hinge position parameters and locking point position parameters, and defines an optimization strategy (including maximization, minimization, or target value setting strategies) for each stiffness performance indicator. After clicking to trigger optimization, the tool can perform multi-objective optimization calculations based on a surrogate model, outputting a set of hinge and locking point parameter combinations that optimize or closely approximate the target values ​​for each performance indicator as a recommended design scheme. Both the prediction function logic and the multi-objective optimization function logic mentioned above can be implemented by writing independent function code.

[0057] At the software architecture design level, based on the above functional logic design, such as Figure 8 As shown, the intelligent tool can adopt a modular software architecture, divided into a prediction module and a multi-objective optimization module. In the prediction module, the tool first loads five trained surrogate models (corresponding to five stiffness performance indicators) and implements real-time invocation of the prediction function logic code. When the user inputs specific design parameters (including hinge position parameters and locking point position parameters) and provides structural characteristic parameters, the prediction module can call the corresponding surrogate model according to the function code logic to perform performance prediction and output the prediction results. In the multi-objective optimization module, the tool also loads five surrogate models and implements real-time invocation of the multi-objective optimization function logic code. When the user inputs the specific range of design parameters, locks the structural characteristic parameters, and determines the target strategy (maximization, minimization, or target setpoint) for each stiffness performance indicator, the multi-objective optimization module can perform multi-objective optimization calculations based on the surrogate models, providing the design parameter values ​​corresponding to the optimal solution that satisfies the target strategy, as a recommended design scheme to the user.

[0058] At the software development and implementation level, based on the aforementioned functional logic design and software architecture design, the graphical user interface for the intelligent tools is developed and its functions integrated. The development process can employ agile development methods, continuously improving the relevant functions through iteration. For example... Figure 9As shown, the tool's main interface integrates the entry points for the prediction module and the multi-objective optimization module, facilitating user switching. In addition, the tool provides a separate performance assistance module, presenting two auxiliary functions in a separate dialog box: one is a display of design parameter influence coefficients, which visually shows the influence coefficients of all design-related parameters (including hinge position parameters, locking point position parameters, and structural characteristic parameters) on each stiffness performance index, clearly presenting the positive or negative impact of each parameter on different performance indices and the degree of its influence, such as... Figure 10 As shown; secondly, the performance index correlation matrix is ​​displayed, such as... Figure 11 As shown, the autocorrelation and cross-correlation relationships among various stiffness performance indicators are presented in matrix form, helping designers understand the intrinsic connections between performance indicators and providing a reference for strategy selection in multi-objective optimization. Through the above development, a complete and user-friendly intelligent design tool for door hinges and locking point layouts was finally formed.

[0059] According to the method for determining the layout parameters of door hinges and locking points proposed in this application, by obtaining the target stiffness performance of the door and inputting the target stiffness performance into a preset neural network model, the method can receive the target hinge position parameters and target locking point position parameters corresponding to the target stiffness performance output by the preset neural network model. Thus, by pre-training the neural network model to establish a mapping relationship between hinge locking point parameters and door stiffness performance, and by directly inputting the target stiffness performance into the model for reverse engineering during the design phase, the method solves the problems of long design cycles and low performance compliance rates caused by repeated trial and error based on engineer experience in related technologies, thereby realizing intelligent design of door hinge and locking point layout.

[0060] Next, referring to the accompanying drawings, a device for determining the arrangement parameters of door hinges and locking points according to an embodiment of this application is described.

[0061] Figure 12 This is a block diagram of a device for determining the arrangement parameters of a car door hinge and locking point according to an embodiment of this application.

[0062] like Figure 12 As shown, the device 10 for determining the arrangement parameters of the door hinge and locking point includes: an acquisition module 100, an input module 200, and a receiving module 300.

[0063] Among them, the acquisition module 100 is used to acquire the target stiffness performance of the vehicle door; The input module 200 is used to input the target stiffness performance into a preset neural network model. The preset surrogate model is trained based on the correspondence between the hinge-lock point sample group and the door stiffness performance index. The receiving module 300 is used to receive the target hinge position parameters and target locking point position parameters corresponding to the target stiffness performance, which are output by a preset neural network model.

[0064] Optionally, in some embodiments, before inputting the target stiffness performance into a preset neural network model, the input module 200 further includes: The acquisition unit is used to acquire hinge position parameters and lock point position parameters from multiple historical finite element models of car doors. The processing unit is used to determine multiple hinge-lock point sample groups based on hinge position parameters and lock point position parameters, and to calculate the door stiffness performance index corresponding to the multiple hinge-lock point sample groups through finite element simulation. The construction unit is used to take multiple hinge-lock point sample groups as input samples and the door stiffness performance index as output samples to construct a training sample set based on the input samples and output samples. The training unit is used to divide the training sample set into a training set and a test set based on a preset partitioning ratio. Based on the training set, the preset neural network is trained until the preset training termination condition is reached, and the initial neural network model is obtained. The test unit is used to test the initial neural network model using the test set, and to obtain the preset neural network model when the test results meet the preset requirements.

[0065] Optionally, in some embodiments, the test unit is specifically used for: The hinge-lock point sample group in the test set is input into the initial neural network model to obtain the predicted stiffness performance index. Calculate the coefficient of determination and root mean square error between the predicted stiffness performance index and the corresponding door stiffness performance index in the test set; If the coefficient of determination is greater than or equal to the first preset threshold and the root mean square error is less than or equal to the second preset threshold, the test result is determined to meet the preset requirements.

[0066] Optionally, in some embodiments, the processing unit is specifically used for: Establish a parametric finite element model of the hinge position and the locking point position; Write the hinge position parameters and lock point position parameters in each hinge-lock point sample group into a preset driving script, and update the hinge position and lock point position of the parameterized finite element model based on the preset driving script. Finite element simulation analysis was performed on the parameterized finite element model after each update, and the door stiffness performance index corresponding to each hinge-lock point sample group was calculated.

[0067] Optionally, in some embodiments, the processing unit is specifically used for: Based on the mean and standard deviation of the hinge position parameters and the mean and standard deviation of the locking point position parameters, the target value range of the hinge position parameters and the target value range of the locking point position parameters are determined. Based on the target value ranges of the hinge position parameters and the lock point position parameters, multiple hinge-lock point sample groups are generated using an optimized Latin hypercube strategy.

[0068] It should be noted that the explanation of the aforementioned method embodiment for determining the door hinge and locking point arrangement parameters also applies to the device for determining the door hinge and locking point arrangement parameters in this embodiment, and will not be repeated here.

[0069] The device for determining the arrangement parameters of door hinges and locking points according to the embodiments of this application obtains the target stiffness performance of the door and inputs the target stiffness performance into a preset neural network model. It can then receive the target hinge position parameters and target locking point position parameters corresponding to the target stiffness performance output by the preset neural network model. Thus, by pre-training the neural network model to establish a mapping relationship between the hinge locking point parameters and the door stiffness performance, and by directly inputting the target stiffness performance into the model for reverse engineering during the design phase, it solves the problems of long design cycles and low performance compliance rates caused by repeated trial and error based on engineer experience in related technologies, achieving intelligent design of door hinge and locking point arrangements.

[0070] Figure 13 A schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device may include: The memory 1301, the processor 1302, and the computer program stored on the memory 1301 and executable on the processor 1302.

[0071] When the processor 1302 executes the program, it implements the method for determining the door hinge and locking point arrangement parameters provided in the above embodiments.

[0072] Furthermore, electronic devices also include: Communication interface 1303 is used for communication between memory 1301 and processor 1302.

[0073] The memory 1301 is used to store computer programs that can run on the processor 1302.

[0074] The memory 1301 may include high-speed RAM (Random Access Memory) memory, and may also include non-volatile memory, such as at least one disk storage.

[0075] If the memory 1301, processor 1302, and communication interface 1303 are implemented independently, then the communication interface 1303, memory 1301, and processor 1302 can be interconnected via a bus to complete communication between them. The bus can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 13 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0076] Optionally, in a specific implementation, if the memory 1301, processor 1302, and communication interface 1303 are integrated on a single chip, then the memory 1301, processor 1302, and communication interface 1303 can communicate with each other through an internal interface.

[0077] The processor 1302 may be a CPU (Central Processing Unit), an ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of this application.

[0078] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method for determining the arrangement parameters of the door hinges and locking points as described above.

[0079] This application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the above-described method for determining the arrangement parameters of the door hinges and locking points.

[0080] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0081] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0082] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A method for determining the arrangement parameters of a car door hinge and locking points, characterized in that, Includes the following steps: Obtain the target stiffness performance of the vehicle door; The target stiffness performance is input into a preset neural network model, which is trained based on the correspondence between the hinge-lock point sample set and the door stiffness performance index. Receive the target hinge position parameters and target locking point position parameters corresponding to the target stiffness performance, output by the preset neural network model.

2. The method according to claim 1, characterized in that, Before inputting the target stiffness performance into the preset neural network model, the method further includes: Obtain hinge position parameters and lock point position parameters from multiple historical finite element models of car doors; Based on the hinge position parameters and the locking point position parameters, multiple hinge-locking point sample groups are determined, and the door stiffness performance index corresponding to the multiple hinge-locking point sample groups is calculated by finite element simulation. Multiple hinge-lock point sample groups are used as input samples, and the door stiffness performance index is used as output samples. A training sample set is constructed based on the input samples and the output samples. Based on a preset partitioning ratio, the training sample set is divided into a training set and a test set. Based on the training set, a preset neural network is trained until a preset training termination condition is reached to obtain an initial neural network model. The initial neural network model is tested using the test set, and the preset neural network model is obtained when the test results meet the preset requirements.

3. The method according to claim 2, characterized in that, The step of testing the initial neural network model using the test set and obtaining the preset neural network model when the test results meet preset requirements includes: The hinge-lock point sample group in the test set is input into the initial neural network model to obtain the predicted stiffness performance index. Calculate the coefficient of determination and root mean square error between the predicted stiffness performance index and the corresponding door stiffness performance index in the test set; If the determination coefficient is greater than or equal to a first preset threshold and the root mean square error is less than or equal to a second preset threshold, the test result is determined to meet the preset requirements.

4. The method according to claim 2, characterized in that, The calculation of door stiffness performance indicators corresponding to multiple hinge-lock point sample groups through finite element simulation includes: Establish a parametric finite element model of the hinge position and the locking point position; The hinge position parameters and lock point position parameters in each hinge-lock point sample group are written into a preset driving script, and the hinge position and lock point position of the parameterized finite element model are updated based on the preset driving script. Finite element simulation analysis was performed on the parameterized finite element model after each update to calculate the door stiffness performance index corresponding to each hinge-lock point sample group.

5. The method according to claim 2, characterized in that, The determination of multiple hinge-lock point sample groups based on the hinge position parameters and the locking point position parameters includes: Normal distribution fitting is performed on the hinge position parameters and the locking point position parameters respectively to obtain the mean and standard deviation of the hinge position parameters and the mean and standard deviation of the locking point position parameters; Based on the mean and standard deviation of the hinge position parameters and the mean and standard deviation of the locking point position parameters, the target value range of the hinge position parameters and the target value range of the locking point position parameters are determined. Based on the target value range of the hinge position parameter and the target value range of the locking point position parameter, multiple hinge-locking point sample groups are generated using an optimized Latin hypercube strategy.

6. A device for determining the arrangement parameters of a car door hinge and locking point, characterized in that, include: The acquisition module is used to acquire the target stiffness performance of the vehicle door. The input module is used to input the target stiffness performance into a preset neural network model, wherein the preset surrogate model is trained based on the correspondence between the hinge-lock point sample group and the door stiffness performance index; The receiving module is used to receive the target hinge position parameters and target locking point position parameters output by the preset neural network model, which correspond to the target stiffness performance.

7. The apparatus according to claim 6, characterized in that, Before inputting the target stiffness performance into a preset neural network model, the input module further includes: The acquisition unit is used to acquire hinge position parameters and lock point position parameters from multiple historical finite element models of car doors. The processing unit is used to determine multiple hinge-lock point sample groups based on the hinge position parameters and the locking point position parameters, and to calculate the door stiffness performance index corresponding to the multiple hinge-lock point sample groups through finite element simulation. A construction unit is used to take multiple hinge-lock point sample groups as input samples and the door stiffness performance index as output samples, and construct a training sample set based on the input samples and the output samples; The training unit is used to divide the training sample set into a training set and a test set based on a preset division ratio, and to train a preset neural network based on the training set until a preset training termination condition is reached to obtain an initial neural network model. The testing unit is used to test the initial neural network model using the test set, and to obtain the preset neural network model when the test results meet the preset requirements.

8. An electronic device, characterized in that, include: The memory, the processor, and the computer program stored in the memory and capable of running on the processor, the processor executing the program to implement the method for determining the door hinge and locking point arrangement parameters as described in any one of claims 1-5.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to implement the method for determining the door hinge and locking point arrangement parameters as described in any one of claims 1-5.

10. A computer program product, characterized in that, The system includes a computer program, which, when executed by a processor, is used to implement the method for determining the door hinge and locking point arrangement parameters as described in any one of claims 1-5.

Images