A platform door optimization method, system and storage medium

Abstract

The application provides a platform door optimization method, system and storage medium, the method comprises the following steps: obtaining the aerodynamic pressure change curve of the basic platform door under the train passing through the station working condition according to the pre-trained aerodynamic effect calculation model, converting the basic platform door structure into the density variable distribution through the variable density method after the modal analysis, statics structure strength analysis and fatigue strength analysis of the basic platform door according to the aerodynamic pressure change curve, obtaining the target platform door structure by iteratively updating the density variable distribution through the genetic algorithm with the structure light weight, the stiffness maximization and the fatigue life maximization as the target and according to the preset stress constraint condition, and selecting the target material performance parameters of the target platform door structure in the preset material database through the bayesian optimization algorithm and the finite element analysis result with the lowest cost constraint condition.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of platform door structure, and in particular to a platform door optimization method and system and a storage medium. BACKGROUND

[0002] With the rapid development of science and technology in the railway industry, the running speed and organization of intercity railways, city railways and high-speed railways have changed a lot, and the relatively severe aerodynamic effects generated by high-speed train passing through stations have a significant impact on the surrounding environment. For the safety of platform structures, waiting and on-duty personnel and ground facilities, potential safety risks are brought. By analogy with similar measures in the subway, platform doors are installed.

[0003] In the prior art, due to the low train running speed, the uniformity of train types and station types, and the consistency of aerodynamic effects, the corresponding platform door structure form is constrained by the corresponding standard; for intercity railways, city railways and high-speed railways, due to the variety of train types, marshalling types, station forms and train running speeds, it is difficult to determine the setting position of the platform door and the strength requirement of the platform door structure. SUMMARY

[0004] Therefore, the embodiments of the present application provide a platform door optimization method and system and a storage medium to eliminate or improve one or more defects in the prior art, and solve the problem that it is difficult to realize platform door optimization for different lines in the prior art.

[0005] One aspect of the present application provides a platform door optimization method, which comprises the following steps:

[0006] obtaining an aerodynamic pressure change curve of a basic platform door under train passing station working condition according to a pre-trained aerodynamic effect calculation model;

[0007] performing modal analysis, statics structure strength analysis and fatigue strength analysis on the basic platform door according to the aerodynamic pressure change curve, converting the basic platform door structure into a density variable distribution through a variable density method, and iteratively updating the density variable distribution through a genetic algorithm to obtain a target platform door structure, with the objectives of structure lightweight, maximum stiffness and maximum fatigue life, and according to a pre-set stress constraint condition, and selecting target material performance parameters of the target platform door structure in a pre-set material database through a Bayesian optimization algorithm and finite element analysis results with a lowest cost constraint condition.

[0008] In some embodiments, the pre-training process of the aerodynamic effect calculation model comprises:

[0009] obtaining a training sample set, the training sample set comprising a source data set and a target data set, the source data set comprising subway passing station working conditions and corresponding real values of aerodynamic pressure data, and the target data set comprising railway vehicle passing station working conditions and corresponding real values of aerodynamic pressure data;

[0010] training an initial aerodynamic effect calculation model using the source data set, the initial aerodynamic effect calculation model comprising a convolutional neural network for extracting spatial features of aerodynamic pressure, a long short-term memory network for obtaining time series features of aerodynamic pressure, and a fully connected layer, and outputting predicted values of aerodynamic pressure data under subway passing station working conditions after taking the subway passing station working conditions as input; in the initial aerodynamic effect calculation model trained by the source data set, the convolutional layer weight of the convolutional neural network is retained, and the long short-term memory network and the fully connected layer are changed according to preset configuration parameters to obtain a migration aerodynamic effect calculation model, and the migration aerodynamic effect calculation model is trained by the target data set to output predicted values of aerodynamic pressure data under railway vehicle passing station working conditions after taking the railway vehicle passing station working conditions as input;

[0011] constructing a first loss function according to the predicted values and real values of aerodynamic pressure data under subway passing station working conditions, constructing a second loss function according to the predicted values and real values of aerodynamic pressure data under railway vehicle passing station working conditions, taking minimizing the first loss function and the second loss function as an objective, updating parameters of the initial aerodynamic effect calculation model using the training sample set until convergence is achieved, and obtaining the aerodynamic effect calculation model.

[0012] In some embodiments, the pre-training process of the aerodynamic effect calculation model further comprises:

[0013] setting a monitoring index in the training process and an early stopping condition corresponding to the monitoring index;

[0014] checking the monitoring index after each iteration using a preset early stopping function, and terminating iteration when the early stopping condition is reached.

[0015] In some embodiments, the method further comprises:

[0016] converting the basic platform door structure into a density variable distribution by a variable density method, and constructing a digital twin model of the basic platform door and an observation equation describing the stress data and the aerodynamic pressure data collected by the preset sensor and changes thereof;

[0017] predicting the stress data and the aerodynamic pressure data at the current time based on the stress data, the aerodynamic pressure data, the platform door structure and the material performance parameters at the previous time, and correcting the stress data and the aerodynamic pressure data at the current time according to the observation equation and the Kalman filtering algorithm;

[0018] According to the revised stress data and the aerodynamic pressure data, a structural stress-strain nephogram is calculated and obtained;

[0019] According to the target platform door structure and the target material performance parameter, a basic platform door is adjusted to obtain a target platform door, and the target platform door is visually displayed.

[0020] In some embodiments, the method further comprises:

[0021] For components in the platform door structure that do not meet the fatigue strength requirement, a circular arc transition, a patch, and a bar-anchoring method are used for adjustment.

[0022] In some embodiments, the preset material database is constructed through a Bayesian optimization framework, and the process includes:

[0023] Performance parameter data of various materials are collected, and after cleaning, sorting, and standardization processing of the collected material performance parameters, processed material performance parameters are obtained;

[0024] Through a Bayesian statistical method, the material performance parameters are taken as samples to establish a material performance probability model;

[0025] According to the collected material performance parameters, the material performance probability model is trained, and the parameters of the material performance probability model are continuously adjusted to obtain a target material performance probability model;

[0026] The processed material performance parameters and the target material performance probability model are integrated to obtain the preset material database.

[0027] In another aspect, the present application also provides a platform door optimization system, which is used to execute any of the above-mentioned platform door optimization methods, and the system includes:

[0028] An aerodynamic effect calculation module is used to obtain an aerodynamic pressure change curve of a basic platform door under a train passing station working condition according to a pre-trained aerodynamic effect calculation model;

[0029] A structural strength analysis module includes a modal analysis submodule, a statics structural strength analysis submodule, and a fatigue strength analysis submodule, and is used to perform modal analysis, statics structural strength analysis, and fatigue strength analysis on the basic platform door according to the aerodynamic pressure change curve;

[0030] A parameterized structure optimization module is configured to convert a basic platform door structure into a density variable distribution by a variable density method, to iteratively update the density variable distribution by a genetic algorithm to obtain a target platform door structure with the lightest weight, the largest stiffness and the longest fatigue life, and according to a preset stress constraint condition, and to select target material performance parameters of the target platform door structure in a preset material database by a Bayesian optimization algorithm and a finite element analysis result under a lowest cost constraint condition.

[0031] In some embodiments, the system further comprises:

[0032] An aerodynamic effect calculation result library is configured to store aerodynamic pressure variation curves of metro and railway vehicles under a passing condition, so as to select corresponding aerodynamic pressure variation curves for modal analysis, static structural strength analysis and fatigue strength analysis according to platform door structure design requirements.

[0033] In some embodiments, the parameterized structure optimization module further comprises:

[0034] A component and material data management library is configured to uniformly manage components and materials in the target platform door structure and the target material performance parameters.

[0035] In another aspect, the application also provides a computer readable storage medium having a computer program / instruction stored thereon, the program / instruction being executed by a processor to implement the steps of the method according to any one of the above aspects.

[0036] In the platform door optimization method and system, an aerodynamic effect calculation model combining a convolutional neural network and a long short-term memory network is used to replace fluid mechanics simulation, a transfer technology is used to transfer aerodynamic data of a metro platform door to an intercity / high-speed rail scene as a pre-training model, simulation calculation time is significantly reduced, and generalization ability of the aerodynamic effect calculation model is improved; potential failures or performance degradation are predicted through digital twin simulation and analysis, so that maintenance is performed in advance, downtime and maintenance costs are reduced, and design and testing are performed on a digital twin before actual manufacturing or construction of a product or system, so as to optimize performance and reduce development costs. The updating process of the platform door is constrained by various constraint conditions, and a target platform door structure and target material performance parameters that meet requirements are obtained, so that the target platform door is obtained.

[0037] Additional advantages, objects, and features of the application will be set forth in part by the description that follows, and will in part be apparent to those of ordinary skill in the art upon examination of the following or can be learned from practice of the application. The objects and other advantages of the application can be realized and attained by the structure particularly pointed out in the specification as well as in the appended claims.

[0038] Those skilled in the art will understand that the objects and advantages of the application can be realized and attained by means of the subject-matter as described in the following detailed description and appended claims, and it is therefore contemplated by the inventor to cover any and all adaptations, modifications and equivalents of the subject-matter as described in the following detailed description and appended claims. BRIEF DESCRIPTION OF DRAWINGS

[0039] The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiments of the application and together with the detailed description serve to explain the principles of the application. In the drawings:

[0040] Figure 1 The flow chart of the platform door optimization method according to an embodiment of the application.

[0041] Figure 2 The structure diagram of the platform door optimization system according to an embodiment of the application.

[0042] Figure 3 The structure diagram of the platform door optimization method according to an embodiment of the application. DETAILED DESCRIPTION

[0043] In order to make the objects, technical solutions and advantages of the application clearer, the following will further describe the application with reference to the embodiments and drawings. Herein, the illustrative embodiments of the application and their descriptions are used to explain the application, but not to limit the application.

[0044] It should be noted that, in order to avoid the application being obscured by unnecessary details, only the structures and / or processing steps closely related to the solutions according to the application are shown in the drawings, and other details not closely related to the application are omitted.

[0045] It should be emphasized that the terms "comprise / comprising" when used in this specification are taken to specify the presence of stated features, elements, steps or components but do not preclude the presence or addition of one or more other features, elements, steps, components or groups thereof.

[0046] It should be noted that, if not specifically stated, the term "connected" herein can not only mean direct connection, but also indirect connection with an intermediate.

[0047] In the following, the embodiments of the application will be described with reference to the drawings. In the drawings, the same reference numerals represent the same or similar components, or the same or similar steps.

[0048] In the prior art, due to the low train running speed, the unified type of vehicle and station, and the consistent aerodynamic effect, the corresponding standard is used to constrain the structure form of the platform door. For intercity railway, city district railway and high-speed railway, due to the types of vehicles, marshalling, station form and train running speed, it is difficult to determine the setting position of the platform door and the strength requirement of the platform door structure. The present application provides a platform door optimization method, system and storage medium. After obtaining the aerodynamic pressure change curve of the basic platform door under the train passing station working condition according to the pre-trained aerodynamic effect calculation model, the basic platform door is subjected to modal analysis, statics structure strength analysis and fatigue strength analysis, and then the basic platform door structure is converted into a density variable distribution by the variable density method. The density variable distribution is iteratively updated by the genetic algorithm to obtain the target platform door structure, with the objectives of structure lightweight, maximum stiffness and maximum fatigue life, and according to the preset stress constraint condition. The target material performance parameters of the target platform door structure are selected in the preset material database by the Bayesian optimization algorithm and the finite element analysis result under the lowest cost constraint condition.

[0049] Figure 1 The flowchart of the platform door optimization method of an embodiment of the present application is shown. One aspect of the present application provides a platform door optimization method, which comprises the following steps S101-S102:

[0050] Step S101: obtaining the aerodynamic pressure change curve of the basic platform door under the train passing station working condition according to the pre-trained aerodynamic effect calculation model.

[0051] Step S102: after the modal analysis, statics structure strength analysis and fatigue strength analysis of the basic platform door according to the aerodynamic pressure change curve, the basic platform door structure is converted into a density variable distribution by the variable density method. The density variable distribution is iteratively updated by the genetic algorithm to obtain the target platform door structure, with the objectives of structure lightweight, maximum stiffness and maximum fatigue life, and according to the preset stress constraint condition. The target material performance parameters of the target platform door structure are selected in the preset material database by the Bayesian optimization algorithm and the finite element analysis result under the lowest cost constraint condition.

[0052] In step S101, the platform door of various rail transit lines including but not limited to urban subway, intercity railway, urban railway and high-speed railway is optimized, including but not limited to the safety of the surrounding environment, related people and ground facilities affected by the aerodynamic effect generated by various traffic tools including but not limited to subway and railway vehicles when passing at high speed; the pre-trained aerodynamic effect calculation model obtains the aerodynamic pressure change curve of different rail transit lines when the traffic tool passes through by migration technology, and the aerodynamic pressure change curve is the pressure change curve of the aerodynamic effect; further, according to the platform door structure design requirement, the corresponding aerodynamic pressure change curve is selected from the aerodynamic effect calculation result database for storing the aerodynamic pressure change curve of the subway and railway vehicle passing under the working condition, and modal analysis, statics structure strength analysis and fatigue strength analysis are carried out.

[0053] In some embodiments, the pre-training process of the aerodynamic effect calculation model includes steps S1011-S1013:

[0054] Step S1011: Obtain a training sample set, the training sample set includes a source data set and a target data set, the source data set includes subway passing working condition and its corresponding aerodynamic pressure data true value, and the target data set includes railway vehicle passing working condition and its corresponding aerodynamic pressure data true value.

[0055] Step S1012: Train the initial aerodynamic effect calculation model with the source data set, the initial aerodynamic effect calculation model includes a convolutional neural network for extracting aerodynamic pressure spatial features, a long short-term memory network for obtaining aerodynamic pressure time series features, and a full connection layer, and outputs the aerodynamic pressure data prediction value under the subway passing working condition after taking the subway passing working condition as the input; in the initial aerodynamic effect calculation model trained by the source data set, the convolutional layer weight of the convolutional neural network is retained and the long short-term memory network and the full connection layer are changed according to the preset configuration parameter to obtain the migration aerodynamic effect calculation model, and the migration aerodynamic effect calculation model is trained by the target data set, and outputs the aerodynamic pressure data prediction value under the railway vehicle passing working condition after taking the railway vehicle passing working condition as the input.

[0056] Step S1013: Construct a first loss function according to the aerodynamic pressure data prediction value and the true value under the subway passing working condition, construct a second loss function according to the aerodynamic pressure data prediction value and the true value under the railway vehicle passing working condition, minimize the first loss function and the second loss function as the target, and update the parameters of the initial aerodynamic effect calculation model with the training sample set until convergence is obtained, and obtain the aerodynamic effect calculation model.

[0057] In some embodiments, the pre-training process of the aerodynamic effect calculation model further includes steps S10131-S10132:

[0058] Step S10131: Set the monitoring indicators in the training process and the early stopping conditions corresponding to the monitoring indicators.

[0059] Step S10132: Check the monitoring indicators after each iteration round using a preset early stopping function, and terminate the iteration when the early stopping condition is reached.

[0060] Specifically, the overall aerodynamic pressure data of the historical train passing station working condition includes aerodynamic pressure distribution under different speeds, train types and marshalling conditions. After normalization, denoising and segmentation, the training set, validation set and test set are obtained. After training the initial aerodynamic effect calculation model and the migration aerodynamic effect calculation model through the training set, the model performance is monitored through the validation set and the hyperparameters including learning rate, batch size and network layer are adjusted. The model performance is evaluated using the test set, the prediction error is calculated, and the performance of the aerodynamic effect calculation model under different working conditions is analyzed by comparing the prediction results with the actual results; the loss function adopts the mean square error loss function; the early stopping method is used to prevent overfitting.

[0061] In step S102, the basic platform door structure is the topological form of the basic platform door key components, including but not limited to the door frame and the connecting piece; the natural frequency and mode shape of the platform door structure are obtained through modal analysis, so as to determine the weak position of the structure and check the connection relationship of the assembly; the ability of the basic platform door structure to bear load is obtained through statics analysis to determine whether the strength and stiffness of the structure meet the requirements, and the ability of the platform door structure to resist fatigue failure under cyclic load is obtained through fatigue strength analysis, and the load is the spatial distribution of the force borne by the platform door structure. The process of converting the basic platform door structure into a density variable distribution by the variable density method includes: discretizing the basic platform door structure into multiple units and introducing a density variable for each unit to represent the material filling degree, the value range of the density variable is 0-1, 0 represents no material, and 1 represents full material, then a function relationship between the material properties and the density variable is established and the density variable of each unit is determined according to the optimization target, thereby obtaining the density variable distribution of the platform door structure, and the density variable distribution describes the topological form of the platform door structure. Further, the genetic algorithm adopts a multi-objective genetic algorithm (NSGA-II) to solve the multi-objective optimization problem; the preset stress constraint condition introduces a stress concentration sensitivity factor to actively avoid high stress areas during the iteration process to improve the reliability of the structure. In some embodiments, the method further comprises: adjusting the components in the platform door structure that do not meet the fatigue strength requirement by using circular arc transition, patch and barb planting, and can also be adjusted by changing the stress mode.

[0062] In some embodiments, the preset material database is constructed through a Bayesian optimization framework, and the process includes steps S1021-S1024:

[0063] Step S1021: Collect performance parameter data of various materials, and obtain processed material performance parameters after cleaning, arranging and standardizing the collected material performance parameters.

[0064] Step S1022: Establish a material performance probability model by taking the material performance parameters as samples through a Bayesian statistical method.

[0065] Step S1023: Train the material performance probability model according to the collected material performance parameters, and continuously adjust the parameters of the material performance probability model to obtain a target material performance probability model.

[0066] Step S1024: Integrate the processed material performance parameters and the target material performance probability model to obtain a preset material database.

[0067] Further, a finite element model is established according to the target station door structure and the application scenario, a plurality of material combination schemes are obtained by selecting different materials from the preset material database and combining the materials, and each material combination scheme is input into the finite element model for finite element analysis and calculation. The finite element analysis structure is used to evaluate whether each material combination scheme meets the requirements. A surrogate model of the objective function is constructed based on the results of the Bayesian optimization algorithm, and the material combination schemes are compared and tracked to find the optimal material combination scheme as the target material performance parameter. The preset material database contains material performance parameters of various materials with different tensile strengths and fatigue limits.

[0068] In some embodiments, the method further includes steps S1-S4:

[0069] Step S1: Convert the basic station door structure into a density variable distribution by the variable density method, and construct a digital twin model of the basic station door and an observation equation describing the stress data and pneumatic pressure data collected by the preset sensor and the changes thereof.

[0070] Step S2: Predict the stress data and pneumatic pressure data at the current time based on the stress data, pneumatic pressure data, station door structure and material performance parameters at the previous time, and correct the stress data and pneumatic pressure data at the current time according to the observation equation and Kalman filtering algorithm.

[0071] Step S3: Calculate and obtain a structure stress-strain nephogram according to the corrected stress data and pneumatic pressure data.

[0072] Step S4: Adjust the basic station door according to the target station door structure and the target material performance parameter to obtain a target station door, and visually display the target station door.

[0073] Specifically, the digital twin model is used for data integration, mapping and three-dimensional visualization. Data in the platform door optimization process is input into the digital twin model, and the stress data and pneumatic pressure data collected by the preset sensor are used for accuracy correction, so as to obtain more accurate and specific platform door stress data and pneumatic pressure data, and to construct a target platform door composed of a structure stress-strain cloud map and a visual display of the target platform door structure and the target material performance parameter after iterative update. The structure stress-strain map can be visualized by a preset visualization tool.

[0074] Figure 2 The structural schematic diagram of the platform door optimization system according to an embodiment of the present application is shown. On the other hand, the present application also provides a platform door optimization system. The system is used to execute any of the above platform door optimization methods. The system comprises:

[0075] The pneumatic effect calculation module is used to obtain the pneumatic pressure change curve of the basic platform door under the train passing station working condition according to the pre-trained pneumatic effect calculation model.

[0076] The structure strength analysis module comprises a modal analysis submodule, a statics structure strength analysis submodule and a fatigue strength analysis submodule. The module is used to perform modal analysis, statics structure strength analysis and fatigue strength analysis on the basic platform door according to the pneumatic pressure change curve.

[0077] The parameterized structure optimization module is used to convert the basic platform door structure into a density variable distribution by the variable density method, to obtain the target platform door structure by iteratively updating the density variable distribution with the structure lightweight, stiffness maximization and fatigue life maximization as the target and according to the preset stress constraint condition, and to select the target material performance parameter of the target platform door structure in the preset material database with the lowest cost constraint condition by the Bayesian optimization algorithm and the finite element analysis result.

[0078] In some embodiments, the system further comprises:

[0079] The pneumatic effect calculation result library is used to store the pneumatic pressure change curve under the metro and railway vehicle passing station working condition, so as to select the corresponding pneumatic pressure change curve for modal analysis, statics structure strength analysis and fatigue strength analysis according to the platform door structure design requirement.

[0080] In some embodiments, the parameterized structure optimization module further comprises:

[0081] The component and material data management library is used to uniformly manage the components and materials in the target platform door structure and the target material performance parameter.

[0082] In another aspect, the present application also provides a computer readable storage medium having stored thereon computer programs / instructions which, when executed by a processor, implement the steps of any of the above methods.

[0083] The present application will be described below in conjunction with a specific embodiment:

[0084] The present application proposes a platform door optimization method, system and storage medium, which performs multi-disciplinary structural optimization on platform door designs of different lines, and the theoretical methods of multi-disciplinary structural optimization include a dynamic aerodynamic effect prediction algorithm based on deep learning, a multi-objective topology optimization algorithm, a real-time data-driven dynamic adjustment algorithm and a material-structure collaborative optimization algorithm. The platform door optimization methods include but are not limited to adjusting the structure size, changing the connection relationship, adding a reinforcing structural member and upgrading the material performance. The statistical structure size, the changed connection relationship and the added reinforcing structural member are to optimize the platform door structure, and the upgraded material performance is to optimize the material performance parameters of the platform door.

[0085] The platform door optimization system includes an aerodynamic effect calculation module, a structural strength analysis module and a parameterized structural optimization module, wherein the structural strength analysis module includes a modal analysis submodule, a statics structural strength analysis submodule and a fatigue strength analysis submodule. In the parameterized structural optimization module, a basic platform door part library is established, and the structural optimization is realized by parameterized design and management of the sizes of main structural components; a component database is established to uniformly manage the newly added platform door reinforcing structures.

[0086] Figure 3 The present application proposes a platform door optimization method, system and storage medium, which performs multi-disciplinary structural optimization on platform door designs of different lines, and the theoretical methods of multi-disciplinary structural optimization include a dynamic aerodynamic effect prediction algorithm based on deep learning, a multi-objective topology optimization algorithm, a real-time data-driven dynamic adjustment algorithm and a material-structure collaborative optimization algorithm. The platform door optimization methods include but are not limited to adjusting the structure size, changing the connection relationship, adding a reinforcing structural member and upgrading the material performance. The statistical structure size, the changed connection relationship and the added reinforcing structural member are to optimize the platform door structure, and the upgraded material performance is to optimize the material performance parameters of the platform door.

[0087] 1. Deep learning-based dynamic aerodynamic effect prediction algorithm: a hybrid model combining convolutional neural network (CNN) and long short-term memory network (LSTM) is used to train historical train passing station working conditions aerodynamic pressure data, realize the rapid prediction of aerodynamic pressure distribution under different speed, car type and marshalling conditions; through the transfer learning technology, the aerodynamic data of subway platform door is used as the pre-training model, which is transferred to the intercity / high-speed railway scene, which significantly reduces the simulation calculation time; replace part of the fluid mechanics simulation in the aerodynamic effect calculation module, used for rapid iteration in the preliminary design stage. Collect historical train passing station working conditions aerodynamic pressure data, including aerodynamic pressure distribution under different speed, car type and marshalling conditions; pre-process the data, including normalization, denoising, segmentation and other operations; divide the data into training set, validation set and test set according to the preset proportion; transfer learning data includes source data set and target data set, normalize, segment, and filter noise to ensure feature space consistency. Construct CNN-LSTM hybrid model as the initial aerodynamic pressure calculation model, convolutional neural network is used to extract the spatial features of aerodynamic pressure, such as spatial pattern of pressure distribution; long short-term memory network is used to capture the time series characteristics of aerodynamic pressure, such as the trend of pressure change over time; use transfer learning technology, use the aerodynamic data of subway platform door as pre-training model, transfer to intercity / high-speed scene; transfer learning model retains the convolution layer weight of convolutional neural network, freezes its parameters, modifies long short-term memory network and fully connected layer, and adapts the output dimension of the target task. Train the model using the training set data, optimize the loss function; monitor the model performance through the validation set and adjust the hyperparameters, including but not limited to learning rate, batch size and network layers; further, use early stopping method to prevent overfitting. Use the test set to evaluate the model performance and calculate the prediction error; visualize the comparison between predicted and actual results, analyze the model performance under different working conditions.

[0088] 2. Multi-objective topology optimization algorithm: based on variable density method and multi-objective genetic algorithm, taking structure lightweight, stiffness maximization and fatigue life maximization as optimization objectives, generating optimal topology morphology of platform door key components including but not limited to door frame and connecting piece; introduce stress concentration sensitivity factor, actively avoid high stress area and improve structure reliability in optimization process, generate new type of reinforced structure geometry in parameterized structure optimization module and obtain target platform door structure.

[0089] 3. Real-time data-driven dynamic adjustment algorithm: Kalman filter fuses multi-source sensor data collected by preset sensors through prediction-update cycle, dynamically corrects parameters of digital twin model, including data modeling: defining digital twin model of platform door structure and observation equation describing stress data and aerodynamic pressure data collected by sensors and their changes; conducting real-time data fusion: predicting stress data and aerodynamic pressure data at the current time based on stress data, aerodynamic pressure data, platform door structure and material performance parameters at the previous time, and correcting the predicted value by using sensor observation equation; correcting parameters: updating load distribution according to the corrected aerodynamic pressure data. The digital twin model as a virtual mirror of the platform door cooperates with the Kalman filter to form a "monitoring-simulation-optimization" closed loop, including data mapping: synchronizing sensor data to the digital twin model and data preprocessing; simulation model dynamic updating: injecting stress data and aerodynamic pressure data corrected by Kalman filter, calculating and displaying stress data, aerodynamic pressure data, target platform door structure and target material performance parameters.

[0090] 4. Material-structure collaborative optimization algorithm: based on the Bayesian optimization framework, a material database is constructed, combined with the finite element analysis results, the optimal material combination such as aluminum alloy-composite material hybrid structure is automatically matched; the cost constraint condition is introduced, and under the premise of meeting the mechanical performance, the material scheme with low cost and easy processing is preferentially selected; the parts database supports intelligent recommendation of material upgrading scheme.

[0091] In summary, the application provides a platform door optimization method and system and a storage medium, wherein a basic platform door aerodynamic pressure change curve under a train passing station working condition is obtained according to a pre-trained aerodynamic effect calculation model, the basic platform door is subjected to modal analysis, statics structure strength analysis and fatigue strength analysis according to the aerodynamic pressure change curve, the basic platform door structure is converted into a density variable distribution through a variable density method, the density variable distribution is iteratively updated to obtain a target platform door structure through a genetic algorithm taking structure lightweight, maximum stiffness and maximum fatigue life as targets and according to a preset stress constraint condition, and a target material performance parameter of the target platform door structure is selected in a preset material database through a Bayesian optimization algorithm and a finite element analysis result under a lowest cost constraint condition.

[0092] The embodiment of the application further provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the steps of the foregoing edge computing server deployment method. The computer readable storage medium can be a tangible storage medium, such as a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a floppy disk, a hard disk, a removable storage disk, a CD-ROM, or any other form of storage medium known in the technical field.

[0093] Those skilled in the art will appreciate that the application described herein is susceptible to variations and / or modifications as can be best deduced from the teachings herein. A variety of implementations of the application have been described above. However, one of ordinary skill in the art would understand that the application is not limited to these implementations. Other implementations can easily be derived from the teachings of this application without departing from the spirit and scope of the application as defined by the claims. Accordingly, while the application is susceptible to various modifications and alternative forms, specific embodiments have been shown by way of example in order to elucidate the principles of the application. The application should not be limited to the particular examples disclosed in this application. It should be understood that the application can be practiced with modification and alteration, and that the application be limited only by the scope of the appended claims. The application is not limited to the details given herein, but can be practiced with variations and modifications that are apparent to one of ordinary skill in the art. The application was chosen and described in order to explain the principles of the application and its best mode of practice and the it was not intended that this application be limited to the implementation(s) specifically disclosed or by use of the specific nomenclature. Therefore, the scope of the application should be determined by the appended claims and their legal equivalents rather than by the examples given.

[0094] It is to be understood that the application is not limited to particular configurations, reagents, methods, or materials described herein, as such may vary. The application is also not limited to a particular methodology for teaching or disclosing the application unless required as inherent to that application. The application is not limited to the specific embodiments described herein, but includes any and all implementations of these general descriptions, and ranges including and / or overlapping the described ranges. Embodiments described and / or exemplified herein are understood as being open to variations of substitution and / or addition, including any and all techniques that are available to one of ordinary skill in the art.

[0095] In this disclosure, features that are described and / or exemplified in one embodiment of the application can be used in the same or similar way in one or more other embodiments and / or in combination with or in place of features of other embodiments.

[0096] The above description is intended to be illustrative and not restrictive. Many other changes and modifications can occur to those skilled in the art once advised of the principles of the application. Accordingly, the scope of the application should be determined not with reference to the above description but with reference to the claims appended hereto.

Claims

1. A method of platform door optimization, characterized in that, The method comprises the following steps: obtaining an aerodynamic pressure change curve of a basic platform door under a train passing station working condition according to a pre-trained aerodynamic effect calculation model; performing modal analysis, statics structural strength analysis and fatigue strength analysis on the basic platform door according to the aerodynamic pressure change curve, converting the basic platform door structure into a density variable distribution through a variable density method, and iteratively updating the density variable distribution through a genetic algorithm to obtain a target platform door structure, with the objectives of structural lightweighting, maximum stiffness and maximum fatigue life, and according to a pre-set stress constraint condition, and selecting target material performance parameters of the target platform door structure in a pre-set material database through a Bayesian optimization algorithm and finite element analysis results under a lowest cost constraint condition.

2. The platform door optimization method of claim 1, wherein, The pre-training process of the aerodynamic effect calculation model comprises: obtaining a training sample set, wherein the training sample set comprises a source data set and a target data set, the source data set comprises subway passing station working conditions and corresponding aerodynamic pressure data true values, and the target data set comprises railway vehicle passing station working conditions and corresponding aerodynamic pressure data true values; training an initial aerodynamic effect calculation model using the source data set, wherein the initial aerodynamic effect calculation model comprises a convolutional neural network for extracting aerodynamic pressure spatial features, a long short-term memory network for obtaining aerodynamic pressure time series features, and a full connection layer, and outputs aerodynamic pressure data prediction values under subway passing station working conditions after taking the subway passing station working conditions as input; in the initial aerodynamic effect calculation model trained by the source data set, the convolutional layer weight of the convolutional neural network is retained, and the long short-term memory network and the full connection layer are changed according to pre-set configuration parameters to obtain a transfer aerodynamic effect calculation model, and the transfer aerodynamic effect calculation model is trained by the target data set, and outputs aerodynamic pressure data prediction values under railway vehicle passing station working conditions after taking the railway vehicle passing station working conditions as input; constructing a first loss function according to the aerodynamic pressure data prediction values and true values under the subway passing station working conditions, and constructing a second loss function according to the aerodynamic pressure data prediction values and true values under the railway vehicle passing station working conditions, and updating parameters of the initial aerodynamic effect calculation model using the training sample set until convergence is achieved, with the objective of minimizing the first loss function and the second loss function, to obtain the aerodynamic effect calculation model.

3. The platform door optimization method of claim 2, wherein, The pre-training process of the aerodynamic effect calculation model further comprises: setting a monitoring index in the training process and a corresponding early stopping condition of the monitoring index; checking the monitoring index after each iteration round using a pre-set early stopping function, and terminating iteration when the early stopping condition is reached.

4. The platform door optimization method of claim 1, wherein, The method further comprises: converting the basic platform door structure into a density variable distribution through a variable density method, and constructing a digital twin model of the basic platform door and observation equations describing stress data and aerodynamic pressure data collected by pre-set sensors and changes thereof; predicting the stress data and the aerodynamic pressure data at the current time based on the stress data at the previous time, the aerodynamic pressure data, the platform door structure, and the material performance parameters, and correcting the stress data and the aerodynamic pressure data at the current time according to the observation equation and the Kalman filtering algorithm; calculating and obtaining the structural stress-strain nephogram according to the corrected stress data and the aerodynamic pressure data; adjusting the basic platform door according to the target platform door structure and the target material performance parameters to obtain a target platform door, and visually displaying the target platform door.

5. The platform door optimization method of claim 1, wherein, The method further comprises: adjusting the components in the platform door structure that do not meet the fatigue strength requirement in a manner of circular arc transition, patching, and anchoring.

6. The platform door optimization method of claim 1, wherein, The preset material database is constructed through a Bayesian optimization framework, and the process comprises: collecting performance parameter data of various materials, and obtaining processed material performance parameters after cleaning, sorting, and standardizing the collected material performance parameters; establishing a material performance probability model by taking the material performance parameters as samples through a Bayesian statistical method; training the material performance probability model according to the collected material performance parameters and continuously adjusting parameters of the material performance probability model to obtain a target material performance probability model; integrating the processed material performance parameters and the target material performance probability model to obtain the preset material database.

7. A platform door optimization system characterized in that, The system is used to perform the platform door optimization method according to any one of claims 1 to 6, and the system comprises: an aerodynamic effect calculation module, configured to obtain an aerodynamic pressure change curve of the basic platform door under the train passing station working condition according to a pre-trained aerodynamic effect calculation model; a structural strength analysis module, comprising a modal analysis submodule, a statics structural strength analysis submodule, and a fatigue strength analysis submodule, and configured to perform modal analysis, statics structural strength analysis, and fatigue strength analysis on the basic platform door according to the aerodynamic pressure change curve; a parameterized structure optimization module, configured to convert the basic platform door structure into a density variable distribution through a variable density method, to perform iterative updating on the density variable distribution to obtain a target platform door structure through a genetic algorithm with the objectives of structural lightweighting, maximum stiffness, and maximum fatigue life and according to a preset stress constraint condition, and to select target material performance parameters of the target platform door structure in a preset material database through a Bayesian optimization algorithm and a finite element analysis result with a lowest cost constraint condition.

8. The platform door optimization system of claim 7, wherein, The system further comprises: an aerodynamic effect calculation result library, configured to store the aerodynamic pressure change curves under the train passing station working condition of the subway and the railway vehicle, so as to select corresponding aerodynamic pressure change curves for modal analysis, statics structural strength analysis, and fatigue strength analysis according to the design requirements of the platform door structure.

9. The platform door optimization system of claim 7, wherein, The parameterized structure optimization module further comprises: a component and material data management library, configured to uniformly manage the components and materials in the target platform door structure and the target material performance parameters.

10. A computer readable storage medium having stored thereon computer programs / instructions, characterized in that, The computer program / instruction is executed by the processor to implement the steps of the method according to any one of claims 1 to 6.

Images