Train basic resistance prediction methods, devices, trains, electronic equipment, storage media, and software products

Abstract

This application provides a method, device, train, electronic device, storage medium, and program product for predicting basic resistance of a train, applicable to the field of rail transit technology. The method includes: normalizing a set of train design parameters to obtain a preprocessed set of train design parameters; selecting multiple trained basic resistance prediction models from a set of trained basic resistance prediction models to obtain multiple candidate basic resistance prediction models, and using these candidate models to process the preprocessed set of train design parameters to obtain multiple basic resistance prediction sequences, wherein the multiple candidate basic resistance prediction models are multiple basic resistance prediction models with different model structures; fusing the multiple basic resistance prediction sequences to obtain an integrated basic resistance prediction result; performing quadratic polynomial fitting on the integrated basic resistance prediction result, and using the fitting result to predict the basic resistance of the train.

Description

Technical Field

[0001] This application relates to the field of rail transit technology, and more specifically to a method, device, train, electronic equipment, storage medium, and program product for predicting basic resistance of a train. Background Technology

[0002] The basic resistance of a train is a core input parameter for traction calculations, energy consumption assessments, and the design of traction and braking systems. Its accurate prediction is a prerequisite for achieving optimal train design.

[0003] In realizing the concept of this application, the inventors discovered at least the following problems in the related technologies: some existing solutions cannot establish a quantitative mapping relationship between train design parameters and basic train resistance, resulting in low accuracy in predicting basic train resistance; some existing solutions rely on computational fluid dynamics simulation, resulting in long prediction time for basic train resistance, which is difficult to meet the needs of rapid iterative train design; some existing solutions require measurement of basic train resistance after prototype manufacturing is completed, which, although highly accurate, has obvious lag, prolongs the train development cycle, and increases train development costs. Summary of the Invention

[0004] In view of this, this application provides a method, apparatus, train, electronic device, storage medium, and program product for predicting basic resistance of a train, to at least solve one of the existing problems.

[0005] One aspect of this application provides a method for predicting basic resistance of a train, comprising: normalizing a set of train design parameters to obtain a preprocessed set of train design parameters; selecting multiple trained basic resistance prediction models from a set of trained basic resistance prediction models to obtain multiple candidate basic resistance prediction models, and using the multiple candidate basic resistance prediction models to process the preprocessed set of train design parameters to obtain multiple basic resistance prediction sequences, wherein the multiple candidate basic resistance prediction models are multiple basic resistance prediction models with different model structures; fusing the multiple basic resistance prediction sequences to obtain an integrated basic resistance prediction result; performing quadratic polynomial fitting on the integrated basic resistance prediction result, and using the fitting result to predict the basic resistance of the train.

[0006] According to an embodiment of this application, the above-mentioned normalization processing of the train design parameter set to obtain a preprocessed train design parameter set includes: determining the maximum and minimum values ​​of the parameters in the train design parameter set; and normalizing each parameter in the train design parameter set using the maximum and minimum values ​​to obtain the preprocessed train design parameter set.

[0007] According to an embodiment of this application, the above-mentioned selection of multiple trained basic resistance prediction models from the set of trained basic resistance prediction models to obtain multiple candidate basic resistance prediction models includes: selecting multiple trained basic resistance prediction models from the set of trained basic resistance prediction models whose mean square error on the target training dataset meets a first threshold to obtain a first type of candidate basic resistance prediction models; and selecting multiple trained basic resistance prediction models from the set of trained basic resistance prediction models whose mean square error on the target validation dataset meets a second threshold to obtain a second type of candidate basic resistance prediction models.

[0008] According to an embodiment of this application, the basic resistance prediction model set trained as described above is obtained through the following operations: The historical train design dataset is preprocessed, and the preprocessed historical train design dataset is randomly partitioned to obtain a target training dataset and a target validation dataset; multiple first-class basic resistance prediction models with different hyperparameters are set based on a first-class model structure, multiple second-class basic resistance prediction models with different hyperparameters are set based on a second-class model structure, and multiple third-class basic resistance prediction models with different hyperparameters are set based on a third-class model structure; the multiple first-class, multiple second-class, and multiple third-class basic resistance prediction models are iteratively trained independently based on mean square error using the target training dataset until preset training conditions are met, resulting in a pre-trained basic resistance prediction model set; the pre-trained basic resistance prediction model set is validated based on mean square error using the target validation dataset until preset validation conditions are met, resulting in a trained basic resistance prediction model set.

[0009] According to embodiments of this application, the first type of model structure includes a gradient boosting decision tree structure or a support vector machine structure; the second type of model structure includes a generalized additive regression structure or a random forest structure; and the third type of model structure includes a regression decision tree structure or a neural network structure.

[0010] According to embodiments of this application, the above-mentioned fusion processing of multiple basic resistance prediction sequences to obtain a basic resistance integrated prediction result includes: calculating the average value of multiple basic resistance prediction sequences to obtain a basic resistance integrated prediction result; or performing weighted calculation on multiple basic resistance prediction sequences to obtain a basic resistance integrated prediction result.

[0011] According to embodiments of this application, the above-described process of performing quadratic polynomial fitting on the integrated prediction results of basic resistance and using the fitting results to predict the basic resistance of the train includes:

[0012] A quadratic polynomial is determined that corresponds to the basic resistance and speed of the train, which conforms to physical constraints. The undetermined coefficients of the quadratic polynomial are fitted using the train's basic resistance value and the corresponding speed value from the integrated prediction results of the basic resistance, based on minimizing the sum of squared residuals, to obtain the fitted quadratic polynomial. The fitted quadratic polynomial is then used to predict the basic resistance of the train, and the prediction result of the basic resistance of the train is obtained.

[0013] According to the embodiments of this application, the above-mentioned train design parameters include the cross-sectional area of ​​the train body, the length of the train body, the width of the train body, the height of the train body, the aspect ratio of the train head, the number of equipment on the train roof, the total weight of the train, the diameter of the train wheels, the train transmission efficiency, the number of axles of the train motor car, the number of axles of the train trailer car, and the train speed sequence.

[0014] Another aspect of this application provides a train basic resistance prediction device, comprising: a data normalization module for normalizing a train design parameter set to obtain a preprocessed train design parameter set; a multi-model prediction module for selecting multiple trained basic resistance prediction models from a set of trained basic resistance prediction models to obtain multiple candidate basic resistance prediction models, and calling the multiple candidate basic resistance prediction models to process the preprocessed train design parameter set to obtain multiple basic resistance prediction sequences, wherein the multiple candidate basic resistance prediction models are multiple basic resistance prediction models with different model structures; a data fusion module for fusing the multiple basic resistance prediction sequences to obtain an integrated basic resistance prediction result; and a fitting prediction module for performing quadratic polynomial fitting on the integrated basic resistance prediction result and using the fitting result to predict the train's basic resistance.

[0015] Another aspect of this application provides a train including the train basic resistance prediction device as described above.

[0016] Another aspect of this application provides an electronic device comprising: one or more processors; and a memory for storing one or more programs, wherein, when the one or more programs are executed by the one or more processors, the one or more processors cause the one or more processors to perform the method as described above.

[0017] Another aspect of this application provides a computer-readable storage medium storing computer-executable instructions that, when executed, are used to implement the method described above.

[0018] Another aspect of this application provides a computer program product comprising computer-executable instructions which, when executed, are used to implement the method described above.

[0019] According to the embodiments of this application, by employing the technical means of normalizing and preprocessing train design parameters, screening different structural candidate models for multi-sequence prediction, and combining the results of multiple models with quadratic polynomial fitting, the technical problems of low estimation accuracy, inability to establish a quantitative mapping relationship between train design parameters and basic resistance, long computational fluid dynamics simulation time, and lag in actual vehicle coasting tests are at least partially overcome by the existing schemes. This achieves the technical effects of improving the prediction accuracy of train basic resistance, establishing a quantitative correlation between design parameters and basic resistance, balancing prediction efficiency and engineering adaptability, accurately predicting in advance without relying on prototype vehicle manufacturing, and guiding train optimization design. Attached Figure Description

[0020] The above and other objects, features and advantages of this application will become clearer from the following description of embodiments with reference to the accompanying drawings, in which:

[0021] Figure 1 An exemplary system architecture diagram is shown that can be applied to the train basic resistance prediction method according to an embodiment of this application.

[0022] Figure 2 A flowchart of a train basic resistance prediction method according to an embodiment of this application is shown.

[0023] Figure 3 A diagram illustrating the candidate process of a basic resistance prediction model according to an embodiment of this application is shown.

[0024] Figure 4 A flowchart for predicting basic resistance of a train according to another embodiment of this application is shown.

[0025] Figure 5 A block diagram of a train basic resistance prediction device according to an embodiment of this application is shown.

[0026] Figure 6 A block diagram of an electronic device suitable for implementing a train basic resistance prediction method according to an embodiment of this application is shown. Detailed Implementation

[0027] The embodiments of this application will now be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of this application. In the following detailed description, numerous specific details are set forth to provide a thorough understanding of the embodiments of this application for ease of explanation. However, it will be apparent that one or more embodiments may be implemented without these specific details. Furthermore, descriptions of well-known structures and technologies are omitted in the following description to avoid unnecessarily obscuring the concepts of this application.

[0028] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of this application. The terms “comprising,” “including,” etc., as used herein indicate the presence of the stated features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0029] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein are to be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0030] When using expressions such as "at least one of A, B and C", they should generally be interpreted in accordance with the meaning that is commonly understood by those skilled in the art (e.g., "a system having at least one of A, B and C" should include, but is not limited to, a system having A alone, a system having B alone, a system having C alone, a system having A and B, a system having A and C, a system having B and C, and / or a system having A, B and C, etc.).

[0031] Existing methods for predicting (or calculating, measuring) basic train resistance have the following problems:

[0032] Existing estimation schemes based on empirical formulas. This scheme uses a general empirical formula derived from regression analysis of a large amount of experimental data for estimation. The general form of the empirical formula for the basic resistance of a train is: ,in For speed, , , These are empirical coefficients related to train type. However, existing solutions have the following problems: first, they have low accuracy and cannot reflect specific design characteristics such as new streamlined head shapes and lightweight materials; second, they cannot establish a quantitative mapping relationship between design parameters (such as car body cross-sectional area, length, and train weight) and basic resistance, making it difficult to guide optimization design.

[0033] Based on the existing approach that combines aerodynamic drag simulation with empirical formulas, the basic drag formula is simply decomposed into a mechanical drag term. ) and aerodynamic drag term ( Among them, the second-order coefficients of the aerodynamic drag term. According to CFD simulation calculations, This approach, based on traditional empirical formulas, makes simple proportional adjustments according to weight. However, this type of approach has the following problems: First, it severs the coupling relationship between A, B, and C, limiting the overall accuracy; second, CFD (Computational Fluid Dynamics) simulations are time-consuming.

[0034] Based on the actual vehicle coasting test scheme after the prototype vehicle is manufactured, the basic resistance-speed curve is directly measured through track testing after the first prototype vehicle is manufactured. Although this type of scheme has high accuracy, it suffers from serious lag.

[0035] To address at least one of the existing problems, this application provides a method, device, train, electronic equipment, storage medium, and program product for predicting basic train resistance. The technical solution provided in this application constructs a method for predicting basic train resistance that integrates data-driven approaches and physical laws. By mining the inherent correlation between historical train design parameters and measured resistance data, and employing a multi-model ensemble learning strategy, the basic resistance performance of new train designs can be quantitatively predicted during the train design phase. In particular, by introducing constraints based on physical laws, the prediction results are ensured to be both accurate and consistent with engineering realities.

[0036] This application provides a crucial early basic resistance performance evaluation tool for train design, effectively overcoming the problems of existing methods that rely on actual vehicle testing, have strong lag, and are time-consuming. It is of great significance for shortening the R&D cycle and reducing development costs.

[0037] Figure 1 An exemplary system architecture diagram is shown that can be applied to the train basic resistance prediction method according to an embodiment of this application.

[0038] It is important to note that Figure 1 The examples shown are merely examples of system architectures that can be applied to the embodiments of this application, in order to help those skilled in the art understand the technical content of this application, but do not mean that the embodiments of this application cannot be used in other devices, systems, environments or scenarios.

[0039] like Figure 1 As shown, the system architecture 100 according to this embodiment may include a first terminal device 101, a second terminal device 102, a third terminal device 103, a network 104, and a server 105. The network 104 serves as a medium for providing communication links between the first terminal device 101, the second terminal device 102, the third terminal device 103, and the server 105. The network 104 may include various connection types, such as wired and / or wireless communication links, etc.

[0040] Users can use the first terminal device 101, the second terminal device 102, and the third terminal device 103 to interact with the server 105 via the network 104 to receive or send messages, etc. Various communication client applications can be installed on the first terminal device 101, the second terminal device 102, and the third terminal device 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, and / or social media platform software, etc. (for example only).

[0041] The first terminal device 101, the second terminal device 102, and the third terminal device 103 can be various electronic devices with displays and support web browsing, including but not limited to smartphones, tablets, laptops, and desktop computers.

[0042] Server 105 can be a server that provides various services, such as a backend management server that supports websites browsed by users using the first terminal device 101, the second terminal device 102, and the third terminal device 103 (this is just an example). The backend management server can analyze and process data such as received user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal devices.

[0043] It should be noted that the train basic resistance prediction method provided in this application embodiment can generally be executed by server 105. Correspondingly, the train basic resistance prediction device provided in this application embodiment can generally be located in server 105. The train basic resistance prediction method provided in this application embodiment can also be executed by a server or server cluster that is different from server 105 and capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103, and / or server 105. Correspondingly, the train basic resistance prediction device provided in this application embodiment can also be located in a server or server cluster that is different from server 105 and capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103, and / or server 105. Alternatively, the train basic resistance prediction method provided in this application embodiment can also be executed by the first terminal device 101, the second terminal device 102, or the third terminal device 103, or it can be executed by other terminal devices different from the first terminal device 101, the second terminal device 102, or the third terminal device 103. Accordingly, the train basic resistance prediction device provided in this application embodiment can also be set in the first terminal device 101, the second terminal device 102 or the third terminal device 103, or in other terminal devices different from the first terminal device 101, the second terminal device 102 or the third terminal device 103.

[0044] For example, the data to be processed may originally be stored in any one of the first terminal device 101, the second terminal device 102, or the third terminal device 103 (e.g., the first terminal device 101, but not limited thereto), or it may be stored on an external storage device and imported into the first terminal device 101. Then, the first terminal device 101 may execute the train basic resistance prediction method provided in the embodiments of this application locally, or send the data to be processed to other terminal devices, servers, or server clusters, and have the other terminal devices, servers, or server clusters that receive the data to be processed execute the train basic resistance prediction method provided in the embodiments of this application.

[0045] It should be understood that Figure 1 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0046] Figure 2 A flowchart of a train basic resistance prediction method according to an embodiment of this application is shown.

[0047] like Figure 2 As shown, the basic resistance prediction method for this train includes operations S210~S240.

[0048] In operation S210, the train design parameter set is normalized to obtain the preprocessed train design parameter set.

[0049] Other processing methods can be used for the preprocessing of the train design parameter set. For example, if the distribution of the train design parameter set data does not conform to a uniform distribution, the standard deviation standardization method can be used to process the train design parameter set.

[0050] In operation S220, multiple trained basic resistance prediction models are selected from the set of trained basic resistance prediction models to obtain multiple candidate basic resistance prediction models. These candidate basic resistance prediction models are then used to process the preprocessed train design parameter set to obtain multiple basic resistance prediction sequences. Among these, the multiple candidate basic resistance prediction models are multiple basic resistance prediction models with different model structures.

[0051] The multiple candidate basic resistance prediction models are different models with different structures, indicating that the basic resistance prediction model used in this application has different model structures, model learning methods, and model parameters. By integrating multiple different types of models to predict the basic resistance of the train, the prediction accuracy of the basic resistance can be greatly improved, and a reliable data foundation can be provided for subsequent quadratic polynomial fitting that conforms to physical constraints.

[0052] In operation S230, multiple basic resistance prediction sequences are fused to obtain the basic resistance integrated prediction result.

[0053] The basic resistance integrated prediction results include multiple sets of basic resistance values ​​and the corresponding velocity values.

[0054] In operation S240, the basic resistance integrated prediction results are subjected to quadratic polynomial fitting, and the fitting results are used to predict the basic resistance of the train.

[0055] The coefficients of the quadratic polynomial between basic resistance and speed are fitted using the integrated prediction results of basic resistance to obtain the fitting curve between basic resistance and speed. Then, the basic resistance of the train at any speed is predicted (or calculated) using the fitting curve.

[0056] According to the embodiments of this application, by employing the technical means of normalizing and preprocessing train design parameters, screening different structural candidate models for multi-sequence prediction, and combining the results of multiple models with quadratic polynomial fitting, the technical problems of low estimation accuracy, inability to establish a quantitative mapping relationship between train design parameters and basic resistance, long computational fluid dynamics simulation time, and lag in actual vehicle coasting tests are at least partially overcome by the existing schemes. This achieves the technical effects of improving the prediction accuracy of train basic resistance, establishing a quantitative correlation between design parameters and basic resistance, balancing prediction efficiency and engineering adaptability, accurately predicting in advance without relying on prototype vehicle manufacturing, and guiding train optimization design.

[0057] According to an embodiment of this application, the above-mentioned normalization processing of the train design parameter set to obtain a preprocessed train design parameter set includes: determining the maximum and minimum values ​​of the parameters in the train design parameter set; and normalizing each parameter in the train design parameter set using the maximum and minimum values ​​to obtain the preprocessed train design parameter set.

[0058] The following detailed explanation illustrates the process of normalizing the train design parameter set provided in this application through specific implementation methods.

[0059] When the train design parameters in the train design parameter set conform to a uniform distribution, the maximum and minimum values ​​of the parameters in the train design parameter set are determined (when the train design parameters are multi-dimensional parameters, the minimum and maximum values ​​on each feature dimension are calculated). The Min-Max (minimum value - maximum value) normalization method is used to linearly map the feature data to the [0,1] interval to eliminate the influence of dimensions. The normalization is shown in formula (1):

[0060] (1).

[0061] in, These are the original design parameters (or the original characteristic values ​​of the design parameters); and These are the minimum and maximum values ​​of the lumped design parameters (or design parameter characteristics) of the train design parameters; It is a very small positive number used to prevent the denominator from being zero.

[0062] When the train design parameters in the train design parameter set do not conform to a uniform distribution, the standard deviation standardization method can be used on the train design parameters in the train design parameter set, as shown in formula (2):

[0063] (2).

[0064] in This refers to the average value of the train design parameters (or characteristics of the train design parameters). This represents the standard deviation of train design parameters (or characteristics of train design parameters). This standard deviation standardization method can also transform data into a distribution with a mean of 0 and a standard deviation of 1.

[0065] According to an embodiment of this application, the above-mentioned selection of multiple trained basic resistance prediction models from the set of trained basic resistance prediction models to obtain multiple candidate basic resistance prediction models includes: selecting multiple trained basic resistance prediction models from the set of trained basic resistance prediction models whose mean square error on the target training dataset meets a first threshold to obtain a first type of candidate basic resistance prediction models; and selecting multiple trained basic resistance prediction models from the set of trained basic resistance prediction models whose mean square error on the target validation dataset meets a second threshold to obtain a second type of candidate basic resistance prediction models.

[0066] The above embodiments of this application employ a technical means of selecting first-class and second-class candidate models from a set of trained basic resistance prediction models, based on the mean square error of the target training dataset (meeting a first threshold) and the mean square error of the target validation dataset (meeting a second threshold), thereby determining multiple candidate basic resistance prediction models. This approach at least partially overcomes the technical problems of low estimation accuracy of existing empirical formulas and the tendency for overfitting / underfitting and insufficient generalization ability in single-model predictions. It achieves the effect of selecting candidate models with high fitting accuracy, excellent generalization performance, and strong adaptability, laying a high-quality foundation for subsequent multi-model integrated prediction, further improving the accuracy, stability, and engineering reliability of train basic resistance prediction, while avoiding invalid models from participating in the calculation and improving prediction efficiency.

[0067] The following describes specific implementation methods in conjunction with... Figure 3 The candidate process for the basic resistance prediction model provided in the above embodiments of this application will be described in further detail.

[0068] Figure 3 A diagram illustrating the candidate process of a basic resistance prediction model according to an embodiment of this application is shown.

[0069] The model with the smallest MSE (Mean Squared Error) value on the training dataset is selected. This type of model represents the best-fit capability achievable for the current task.

[0070] Select the model with the smallest MSE value on the validation dataset. This type of model usually has the strongest generalization ability and helps prevent overfitting.

[0071] The trained basic resistance models are screened using the MSE during the training and validation phases to obtain candidate basic resistance prediction models. Taking a three-class model structure as an example, two models are selected from each class, resulting in a library of three classes × two candidate basic resistance prediction models per class. Figure 3 As shown.

[0072] According to an embodiment of this application, the basic resistance prediction model set trained as described above is obtained through the following operations: The historical train design dataset is preprocessed, and the preprocessed historical train design dataset is randomly partitioned to obtain a target training dataset and a target validation dataset; multiple first-class basic resistance prediction models with different hyperparameters are set based on a first-class model structure, multiple second-class basic resistance prediction models with different hyperparameters are set based on a second-class model structure, and multiple third-class basic resistance prediction models with different hyperparameters are set based on a third-class model structure; the multiple first-class, multiple second-class, and multiple third-class basic resistance prediction models are iteratively trained independently based on mean square error using the target training dataset until preset training conditions are met, resulting in a pre-trained basic resistance prediction model set; the pre-trained basic resistance prediction model set is validated based on mean square error using the target validation dataset until preset validation conditions are met, resulting in a trained basic resistance prediction model set.

[0073] According to embodiments of this application, the first type of model structure includes a gradient boosting decision tree structure or a support vector machine structure; the second type of model structure includes a generalized additive regression structure or a random forest structure; and the third type of model structure includes a regression decision tree structure or a neural network structure.

[0074] The embodiments described above in this application employ a technical approach that involves preprocessing historical train design data and randomly dividing it into target training / validation datasets. Simultaneously, it constructs multiple basic resistance prediction models with varying hyperparameters across different structures, including gradient boosting decision trees, support vector machines, and generalized additive regression. Then, it performs iterative independent training based on mean squared error using the target training dataset and validation based on mean squared error using the target validation dataset, ultimately obtaining a set of trained basic resistance prediction models. This approach at least partially overcomes the technical problems of weak generalization ability, limited fitting accuracy, and insufficient adaptability of single-structure / hyperparameter models. It enriches the diversity and adaptability of the model set, ensuring high fitting accuracy and strong generalization ability, thus providing a solid foundation for subsequent selection of high-quality candidate models and improving the accuracy and stability of multi-model ensemble prediction.

[0075] The following detailed description of the process for obtaining the basic resistance prediction model set trained according to the above embodiments of this application will be provided in further detail through specific implementation methods.

[0076] The train historical design dataset consists of two parts: train design parameters and train basic resistance labels. The train basic resistance labels represent the basic resistance values ​​at each speed point measured in actual train resistance tests.

[0077] The historical train design dataset is preprocessed using a data preprocessing algorithm (such as Min-Max normalization). The preprocessed dataset is then randomly divided: the complete historical train design dataset is randomly divided into training and validation datasets according to a preset ratio (e.g., 8:2) (or using a fixed-interval sampling strategy). A fixed random seed ensures the repeatability of the data partitioning and avoids the impact of data partitioning differences on model training performance. The purpose of randomly partitioning the dataset is to evaluate the generalization ability of the trained model.

[0078] During the model training phase, for example, three different types of machine learning models are trained using a preprocessed training dataset: gradient boosting tree model, generalized additive model, and regression decision tree model. For each type of model, different combinations of hyperparameters are set, and N independent training iterations are performed to obtain 3*N (N is a positive integer) base models. For each base model obtained from the above training, its MSE is calculated using a validation dataset to quantify the generalization performance of each base model.

[0079] This application employs three model structures: gradient boosting tree, generalized additive regression, and regression decision tree. However, it should be noted that the purpose of this application is multi-model ensemble, not specifically limited to the aforementioned three model structures. Based on the embodiments or implementations disclosed in this application, other machine learning models, such as support vector machines, random forests, and neural networks, can be used. The types of models integrated in the embodiments or implementations of this application are not limited to three; they can be two, four, or more. As long as multiple models with differences are integrated, they fall within the scope of this application. Any regression model capable of nonlinearly mapping design parameters to resistance can be used as an ensemble model to achieve the purpose of complementary model advantages.

[0080] 2) Alternative Model Selection Strategies: This invention retains two versions for each model class: one with the minimum validation set MSE and the other with the minimum validation set MSE. An alternative could be to simplify this by retaining only one optimal version for each model class; alternatively, a performance threshold could be set (e.g., MSE below a certain value), retaining all models exceeding that threshold, thus dynamically determining the size of the candidate model library. The core objective is to select several high-performing instances from a model family; the specific selection criteria can be adjusted according to actual circumstances.

[0081] According to embodiments of this application, the above-mentioned fusion processing of multiple basic resistance prediction sequences to obtain a basic resistance integrated prediction result includes: calculating the average value of multiple basic resistance prediction sequences to obtain a basic resistance integrated prediction result; or performing weighted calculation on multiple basic resistance prediction sequences to obtain a basic resistance integrated prediction result.

[0082] The embodiments described above in this application achieve fusion by averaging or weighting multiple basic resistance prediction sequences, which effectively weakens the inherent errors and overfitting defects of a single candidate model and avoids the problem of insufficient generalization ability of a single model. At the same time, integrating the advantages of multiple models improves the stability and accuracy of resistance prediction results, provides a reliable foundation for subsequent physical law-constrained fitting, and balances prediction efficiency and engineering credibility.

[0083] The following detailed description of the process for obtaining the basic resistance integrated prediction results provided in the embodiments of this application will be further explained through specific implementation methods.

[0084] Taking six candidate basic resistance prediction models as an example, the resulting six basic resistance prediction sequences are... , ,..., According to formula (3):

[0085] (3).

[0086] Calculate the arithmetic mean of these basic resistance prediction sequences as the integrated basic resistance prediction result; or assign weights to each candidate model based on its performance on the validation dataset (e.g., the inverse of MSE), with higher weights for better performance, and perform a weighted average. According to formula (4):

[0087] (4).

[0088] in .

[0089] Both simple averaging and weighted averaging can achieve model ensemble, with the aim of combining the outputs of multiple models to obtain more stable and accurate predictions.

[0090] According to an embodiment of this application, the above-mentioned process of performing quadratic polynomial fitting on the integrated prediction results of basic resistance and using the fitting results to predict the basic resistance of the train includes: determining a quadratic polynomial between the basic resistance of the train and the speed that meets physical constraints; performing a fitting process on the undetermined coefficients of the quadratic polynomial based on minimizing the sum of squared residuals using the basic resistance value of the train and the corresponding speed value in the integrated prediction results of basic resistance, to obtain the fitted quadratic polynomial; and using the fitted quadratic polynomial to predict the basic resistance of the train, to obtain the prediction result of the basic resistance of the train.

[0091] The quadratic polynomial fitting process provided in the above embodiments of this application will be further described in detail below through specific implementation methods.

[0092] The basic resistance integrated prediction results were fitted with a quadratic polynomial. , As the basic resistance, Let A, B, and C be the speed, and A, B, and C be the fitting coefficients. This form is consistent with the classical physical formula for the basic resistance of a train, ensuring that the final output resistance-speed curve conforms to recognized physical laws and improving the engineering reliability of the prediction results. After fitting, the predicted basic resistance value at any speed point and the quadratic polynomial fitting coefficients A, B, and C are output, supporting relevant designers in evaluating the basic resistance or optimizing and adjusting design parameters.

[0093] According to the embodiments of this application, the above-mentioned train design parameters include the cross-sectional area of ​​the train body, the length of the train body, the width of the train body, the height of the train body, the aspect ratio of the train head, the number of equipment on the train roof, the total weight of the train, the diameter of the train wheels, the train transmission efficiency, the number of axles of the train motor car, the number of axles of the train trailer car, and the train speed sequence.

[0094] The above train speed sequence refers to the train speed sequence obtained by selecting train speeds from 0 to the maximum operating speed of the train at certain intervals.

[0095] It should be noted that the above train design parameters are exemplary and can be added to or deleted from.

[0096] The following describes another specific embodiment in conjunction with the appendix. Figure 4 The method for predicting basic train resistance provided in this application is described in further detail.

[0097] Figure 4 A flowchart for predicting basic resistance of a train according to another embodiment of this application is shown.

[0098] like Figure 4 As shown, the technical solution provided in this application integrates multiple complementary machine learning models and introduces physical constraints to construct an integrated prediction scheme that can learn the complex mapping relationship between "design parameters and basic resistance" from historical data. This application overcomes the lag of traditional solutions that rely on real-vehicle testing: it eliminates the need to manufacture physical prototypes, and predictions can be made using only parameters available during the design phase (such as slenderness ratio and total weight), enabling early design assessment. It shifts the assessment of basic resistance from later stages to the early design phase, effectively shortening the development cycle and reducing cost increases due to later design modifications. It also addresses the problems of low accuracy and non-physical fluctuations in single models: the integration of multiple models combines the advantages of different algorithms, and the quadratic polynomial fitting constraint ensures that the results conform to the physical laws of resistance (i.e.,...). This ensures that the prediction results conform to the actual engineering situation and avoids non-physical fluctuations that may be generated by pure data models; it solves the problem of no quantitative mapping between design parameters and resistance: it establishes a direct correlation between design parameters and resistance, supports the prediction of resistance changes after parameter adjustment, and assists in design optimization.

[0099] Figure 5 A block diagram of a train basic resistance prediction device according to an embodiment of this application is shown.

[0100] like Figure 5 As shown, the train basic resistance prediction device 500 includes a data normalization module 510, a multi-model prediction module 520, a data fusion module 530, and a fitting prediction module 540.

[0101] The data normalization module 510 is used to normalize the train design parameter set to obtain a preprocessed train design parameter set.

[0102] The multi-model prediction module 520 is used to select multiple trained basic resistance prediction models from the set of trained basic resistance prediction models to obtain multiple candidate basic resistance prediction models, and to call multiple candidate basic resistance prediction models to process the preprocessed train design parameter set to obtain multiple basic resistance prediction sequences. Among them, the multiple candidate basic resistance prediction models are multiple basic resistance prediction models with different model structures.

[0103] The data fusion module 530 is used to fuse multiple basic resistance prediction sequences to obtain the basic resistance integrated prediction result.

[0104] The fitting and prediction module 540 is used to perform quadratic polynomial fitting on the integrated prediction results of the basic resistance, and to use the fitting results to predict the basic resistance of the train.

[0105] Any one or more of the modules, submodules, units, and subunits according to the embodiments of this application, or at least part of the functions of any one or more of them, can be implemented in one module. Any one or more of the modules, submodules, units, and subunits according to the embodiments of this application can be implemented by dividing them into multiple modules. Any one or more of the modules, submodules, units, and subunits according to the embodiments of this application can be at least partially implemented as hardware circuits, such as field-programmable gate arrays (FPGAs), programmable logic arrays (PLAs), systems-on-a-chip, systems-on-a-substrate, systems-on-package, application-specific integrated circuits (ASICs), or implemented by hardware or firmware in any other reasonable manner by integrating or packaging circuits, or implemented in any one of software, hardware, and firmware, or in a suitable combination of any of these. Alternatively, one or more of the modules, submodules, units, and subunits according to the embodiments of this application can be at least partially implemented as computer program modules, which, when run, can perform corresponding functions.

[0106] For example, any plurality of the data normalization module 510, multi-model prediction module 520, data fusion module 530, and fitting prediction module 540 can be combined into one module / unit / subunit, or any one of these modules / units / subunits can be split into multiple modules / units / subunits. Alternatively, at least part of the functionality of one or more of these modules / units / subunits can be combined with at least part of the functionality of other modules / units / subunits and implemented in one module / unit / subunit. According to embodiments of this application, at least one of the data normalization module 510, multi-model prediction module 520, data fusion module 530, and fitting prediction module 540 can be at least partially implemented as hardware circuitry, such as a field-programmable gate array (FPGA), a programmable logic array (PLA), a system-on-a-chip, a system-on-a-substrate, a system-on-package, an application-specific integrated circuit (ASIC), or any other reasonable means of integrating or packaging the circuitry, or implemented in software, hardware, or firmware, or in any suitable combination of any of these three implementation methods. Alternatively, at least one of the data normalization module 510, the multi-model prediction module 520, the data fusion module 530, and the fitting prediction module 540 may be implemented at least partially as a computer program module, which can perform the corresponding function when the computer program module is run.

[0107] It should be noted that the train basic resistance prediction device part in the embodiments of this application corresponds to the train basic resistance prediction method part in the embodiments of this application. The description of the train basic resistance prediction device part is specifically referred to in the train basic resistance prediction method part, and will not be repeated here.

[0108] Another aspect of this application provides a train including the aforementioned train basic resistance prediction device.

[0109] Figure 6 A block diagram of an electronic device suitable for implementing a train basic resistance prediction method according to an embodiment of this application is shown.

[0110] Figure 6 The electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0111] like Figure 6As shown, an electronic device 600 according to an embodiment of this application includes a processor 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage portion 608 into a random access memory (RAM) 603. The processor 601 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or an associated chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 601 may also include onboard memory for caching purposes. The processor 601 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of this application.

[0112] RAM 603 stores various programs and data required for the operation of electronic device 600. Processor 601, ROM 602, and RAM 603 are interconnected via bus 604. Processor 601 executes various operations of the method flow according to embodiments of this application by executing programs in ROM 602 and / or RAM 603. It should be noted that the programs may also be stored in one or more memories other than ROM 602 and RAM 603. Processor 601 may also execute various operations of the method flow according to embodiments of this application by executing programs stored in said one or more memories.

[0113] According to embodiments of this application, the electronic device 600 may further include an input / output (I / O) interface 605, which is also connected to a bus 604. The electronic device 600 may also include one or more of the following components connected to the input / output (I / O) interface 605: an input section 606 including a keyboard, mouse, etc.; an output section 607 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 608 including a hard disk, etc.; and a communication section 609 including a network interface card such as a LAN card, modem, etc. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to the input / output (I / O) interface 605 as needed. A removable medium 611, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on the drive 610 as needed so that computer programs read from it can be installed into the storage section 608 as needed.

[0114] According to embodiments of this application, the method flow according to embodiments of this application can be implemented as a computer software program. For example, embodiments of this application include a computer program product comprising a computer program carried on a computer-readable storage medium, the computer program containing program code for performing the methods shown in the flowchart. In such embodiments, the computer program can be downloaded and installed from a network via communication section 609, and / or installed from removable medium 611. When the computer program is executed by processor 601, it performs the functions defined in the system of embodiments of this application. According to embodiments of this application, the systems, devices, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0115] This application also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or it may exist independently and not assembled into the device / apparatus / system. The computer-readable storage medium carries one or more programs, which, when executed, implement the method according to the embodiments of this application.

[0116] According to embodiments of this application, the computer-readable storage medium can be a non-volatile computer-readable storage medium. Examples include, but are not limited to: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this application, the computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0117] For example, according to embodiments of this application, a computer-readable storage medium may include the ROM 602 and / or RAM 603 described above and / or one or more memories other than ROM 602 and RAM 603.

[0118] Embodiments of this application also include a computer program product comprising a computer program containing program code for performing the methods provided in the embodiments of this application. When the computer program product is run on an electronic device, the program code is used to enable the electronic device to implement the train basic resistance prediction method provided in the embodiments of this application.

[0119] When the computer program is executed by the processor 601, it performs the functions defined in the system / apparatus of this application embodiment. According to the embodiments of this application, the systems, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0120] In one embodiment, the computer program may rely on a tangible storage medium such as an optical storage device or a magnetic storage device. In another embodiment, the computer program may also be transmitted and distributed in the form of signals over a network medium, and downloaded and installed via the communication section 609, and / or installed from the removable medium 611. The program code contained in the computer program can be transmitted using any suitable network medium, including but not limited to: wireless, wired, etc., or any suitable combination thereof.

[0121] According to embodiments of this application, program code for executing the computer programs provided in the embodiments of this application can be written in any combination of one or more programming languages. Specifically, these computational programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, languages ​​such as Java, C++, Python, "C", or similar programming languages. The program code can be executed entirely on the user's computing device, partially on the user's device, partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0122] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions. Those skilled in the art will understand that the features described in the various embodiments of this application can be combined and / or combined in various ways, even if such combinations are not explicitly described in this application. In particular, without departing from the spirit and teachings of this application, the features described in the various embodiments of this application can be combined and / or combined in various ways. All such combinations and / or combinations fall within the scope of this application.

[0123] The embodiments of this application have been described above. However, these embodiments are merely illustrative and not intended to limit the scope of this application. Although various embodiments have been described above, this does not mean that the measures in the various embodiments cannot be used advantageously in combination. Without departing from the scope of this application, those skilled in the art can make various substitutions and modifications, all of which should fall within the scope of this application.

Claims

1. A method for predicting the basic resistance of a train, characterized in that, The method includes: The train design parameter set is normalized to obtain the preprocessed train design parameter set. Multiple trained basic resistance prediction models are selected from the set of trained basic resistance prediction models to obtain multiple candidate basic resistance prediction models. The multiple candidate basic resistance prediction models are then used to process the preprocessed train design parameter set to obtain multiple basic resistance prediction sequences. The multiple candidate basic resistance prediction models are multiple basic resistance prediction models with different model structures. The multiple basic resistance prediction sequences are fused together to obtain the basic resistance integrated prediction result. The basic resistance integrated prediction results are subjected to quadratic polynomial fitting, and the fitting results are used to predict the basic resistance of the train.

2. The method according to claim 1, characterized in that, The train design parameter set is normalized to obtain the preprocessed train design parameter set, which includes: Determine the maximum and minimum values ​​of the parameters in the train design parameter set; The maximum and minimum values ​​of the parameters are used to normalize each parameter in the train design parameter set to obtain the preprocessed train design parameter set.

3. The method according to claim 1, characterized in that, Multiple trained basic resistance prediction models were selected from the set of trained basic resistance prediction models, resulting in several candidate basic resistance prediction models, including: From the set of trained basic resistance prediction models, multiple trained basic resistance prediction models whose mean square error on the target training dataset meets the first threshold are selected to obtain the first type of candidate basic resistance prediction models. From the set of trained basic resistance prediction models, multiple trained basic resistance prediction models whose mean square error meets the second threshold on the target validation dataset are selected to obtain the second type of candidate basic resistance prediction models.

4. The method according to claim 3, characterized in that, The trained set of basic resistance prediction models is obtained through the following operations: The historical train design dataset is preprocessed, and the preprocessed historical train design dataset is randomly divided to obtain the target training dataset and the target validation dataset. Based on the first type of model structure, multiple first type basic resistance prediction models with different hyperparameters are set up; based on the second type of model structure, multiple second type basic resistance prediction models with different hyperparameters are set up; and based on the third type of model structure, multiple third type basic resistance prediction models with different hyperparameters are set up. Using the target training dataset, multiple first-class basic resistance prediction models, multiple second-class basic resistance prediction models, and multiple third-class basic resistance prediction models are iteratively trained independently based on mean square error until the preset training conditions are met, thus obtaining a pre-trained set of basic resistance prediction models. The pre-trained basic resistance prediction model set is validated based on mean square error using the target validation dataset until the preset validation conditions are met, thus obtaining the trained basic resistance prediction model set.

5. The method according to claim 4, characterized in that, The first type of model structure includes gradient boosting decision tree structure or support vector machine structure; The second type of model structure includes generalized additive regression structure or random forest structure; The third type of model structure includes regression decision tree structure or neural network structure.

6. The method according to claim 1, characterized in that, The integrated prediction results of the basic resistance are obtained by fusing multiple basic resistance prediction sequences. The average value of multiple basic resistance prediction sequences is calculated to obtain the integrated prediction result of the basic resistance; or The basic resistance prediction sequences are weighted and calculated to obtain the integrated prediction result of the basic resistance.

7. The method according to claim 1, characterized in that, The basic resistance integrated prediction results are subjected to quadratic polynomial fitting, and the fitting results are used to predict the basic resistance of the train, including: Determine the quadratic polynomial between the basic resistance and speed of a train that conforms to physical constraints; The undetermined coefficients of the quadratic polynomial are fitted using the train's basic resistance value and the corresponding speed value from the integrated prediction results of the basic resistance, based on minimizing the sum of squared residuals, to obtain the fitted quadratic polynomial. The basic resistance of the train is predicted using the fitted quadratic polynomial, and the prediction result of the basic resistance of the train is obtained.

8. The method according to any one of claims 1 to 7, characterized in that, The train design parameters include the cross-sectional area of ​​the train body, the length of the train body, the width of the train body, the height of the train body, the slenderness ratio of the train head, the number of equipment on the train roof, the total weight of the train, the diameter of the train wheels, the train transmission efficiency, the number of axles of the train motor car, the number of axles of the train trailer car, and the train speed sequence.

9. A train basic resistance prediction device, characterized in that, The device includes: The data normalization module is used to normalize the train design parameter set to obtain a preprocessed train design parameter set. The multi-model prediction module is used to select multiple trained basic resistance prediction models from the set of trained basic resistance prediction models to obtain multiple candidate basic resistance prediction models, and to call the multiple candidate basic resistance prediction models to process the preprocessed train design parameter set to obtain multiple basic resistance prediction sequences. The multiple candidate basic resistance prediction models are multiple basic resistance prediction models with different model structures. The data fusion module is used to fuse multiple basic resistance prediction sequences to obtain the basic resistance integrated prediction result; The fitting and prediction module is used to perform quadratic polynomial fitting on the integrated prediction results of the basic resistance, and to use the fitting results to predict the basic resistance of the train.

10. A train, comprising the train basic resistance prediction device as described in claim 9.

11. An electronic device, comprising: One or more processors; Memory, used to store one or more programs. Wherein, when the one or more programs are executed by the one or more processors, the one or more processors implement the method of any one of claims 1 to 8.

12. A computer-readable storage medium having stored thereon executable instructions that, when executed by a processor, cause the processor to perform the method of any one of claims 1 to 8.

13. A computer program product comprising a computer program that, when executed by a processor, implements the method according to any one of claims 1-8.

Images