Railway vehicle-track-bridge system dynamic response prediction method and system, and medium

Abstract

Disclosed in the present invention are a railway vehicle-track-bridge system dynamic response prediction method and system, and a medium. The method comprises: inputting vehicle speed samples and track irregularity samples into a vehicle-track-bridge system coupled random distribution physical model to obtain corresponding bridge dynamic responses and effective loads of the vehicle-track-bridge system, and extracting a global stiffness matrix of the vehicle-track-bridge system; constructing a training sample set; constructing a fitness function considering the effective loads, and then on the basis of the training sample set, using a genetic algorithm to optimize parameters of a BP neural network prediction model and training same to obtain a bridge dynamic response prediction model; and using the bridge dynamic response prediction model to carry out bridge dynamic response prediction. By introducing effective loads in a vehicle-track-bridge system into a fitness function in a genetic algorithm, a neural network model and a vehicle-track-bridge physical model are organically combined, thereby improving the prediction precision.

Description

A railway vehicle-track-bridge system dynamic response prediction method, system and medium Technical Field

[0001] The present invention relates to the technical field of dynamic response of high-speed railway bridges, and in particular to a method, system and medium for predicting the dynamic response of a railway vehicle-track-bridge system. Background Art

[0002] Vehicle loads are among the most significant and common external loads encountered during the service life of a bridge. Whether it's a highway or railway bridge, the structure is expected to withstand vehicle loads, which contribute the majority of static and dynamic load effects on the bridge. As rail transport speeds and weights continue to increase, the dynamic interaction between train vehicles and bridges becomes increasingly significant, leading to increasing dynamic responses in bridges. Understanding these vehicle and bridge responses is crucial for further evaluating performance related to structural behavior and vehicle operational stability. Technical issues

[0003] Currently, bridge response is primarily determined by constructing coupled stochastic analysis physical models of vehicle-track-bridge systems. These models accurately simulate the dynamic response of the vehicle-track-bridge system with high accuracy. However, they suffer from high computational complexity and poor real-time dynamic response prediction. With the continuous advancement of artificial intelligence technology, artificial neural networks (ANNs) have been applied to structural dynamic analysis. Predictive approaches have also been developed, including a CNN-LSTM-based high-speed railway seismic response prediction method and a method for predicting vehicle-bridge dynamic response using FFNN and LSTM. However, current neural network prediction models applied to railway vehicle-track-bridge systems cannot be consistently integrated with the vehicle-track-bridge physical model. Instead, they are limited to a simple mapping of input data to output data, resulting in significant limitations. This has prevented substantial improvements in prediction accuracy. Technical Solutions

[0004] The present invention provides a method, system and medium for predicting the dynamic response of a railway vehicle-track-bridge system, so as to solve the problem that the neural network prediction model currently used in the railway vehicle-track-bridge system cannot be combined with the vehicle-track-bridge physical model, resulting in the inability to substantially improve the prediction accuracy.

[0005] In a first aspect, a method for predicting the dynamic response of a railway vehicle-track-bridge system is provided, comprising:

[0006] S1: Obtain a sample set containing vehicle speed and track irregularity;

[0007] S2: Input the samples in the sample set into the vehicle-track-bridge system coupled random distribution physical model in sequence to obtain the corresponding bridge dynamic response and the effective load of the vehicle-track-bridge system, and extract the overall stiffness matrix of the vehicle-track-bridge system;

[0008] S3: Construct a training sample set, where each training sample in the training sample set contains vehicle speed, track irregularity, the corresponding bridge dynamic response, and the effective load of the vehicle-track-bridge system;

[0009] S4: Using vehicle speed and track irregularity as input and bridge dynamic response as output, a BP neural network prediction model is constructed;

[0010] S5: Construct a fitness function that takes the payload into consideration, use a genetic algorithm to optimize the weights and thresholds of the BP neural network prediction model based on the training sample set, initialize the BP neural network prediction model using the obtained optimal weights and thresholds, and train the initialized BP neural network prediction model based on the training sample set to obtain a bridge dynamic response prediction model;

[0011] S6: Use the bridge dynamic response prediction model to predict the dynamic response of the bridge.

[0012] According to the first aspect, in some possible implementations, in S5, the fitness function considering the payload is expressed as follows:

[0013]

[0014] In the formula, fitness represents the fitness value; represents the vth output prediction value obtained by the bridge dynamic response prediction model for the kth training sample, Indicates the actual value of the vth output corresponding to the kth training sample; represents the number of training samples, Indicates the number of outputs of the bridge dynamic response prediction model; represents the overall stiffness matrix of the vehicle-track-bridge system; represents the effective load prediction value corresponding to the v-th output obtained by the bridge dynamic response prediction model for the k-th training sample, represents the actual value of the effective load corresponding to the kth training sample; when the vth output is the bridge displacement response, , when the vth output is the bridge acceleration response, ;

[0015] When calculating the fitness value of a chromosome, first use formula [1] to calculate the fitness value. If the value obtained by formula [1] ,in, The result obtained by formula [1] is used as the fitness value of the chromosome; if the result obtained by formula [1] , then use formula [2] to calculate the fitness value, and take the larger value of the results calculated by formula [1] and formula [2] as the fitness value of the chromosome.

[0016] According to the first aspect, in some possible implementations, in S5, the process of optimizing the weights and thresholds of the BP neural network prediction model using a genetic algorithm includes:

[0017] Step 1: Initialize the population. Each chromosome in the population is encoded by the weight and threshold of the BP neural network prediction model.

[0018] Step 2: Use the fitness function to calculate the fitness value of each chromosome;

[0019] Step 3: Selection operation, based on the fitness value of the chromosome, select a certain number of chromosomes to enter the next generation;

[0020] Step 4: Crossover operation: according to the crossover probability, two chromosomes are randomly selected from the chromosomes after the selection operation to cross over and generate new chromosomes;

[0021] Step 5: Mutation operation: According to the mutation probability, randomly select one chromosome from the chromosomes after the crossover operation to perform one or more gene mutations to generate a new chromosome;

[0022] Step 6: Repeat steps 2 to 5 until the termination condition is reached, and output the optimal weights and thresholds of the BP neural network prediction model.

[0023] According to the first aspect, in some possible implementations, in S1, track unevenness is constructed using a random harmonic function, which is expressed as follows:

[0024]

[0025] Where, 、 、 They represent track irregularity, horizontal irregularity and height irregularity respectively, among which, represents the random variable point set space, and x represents the representative point set of random variables; represents the amplitude of the orbital irregularity wavelength, represents random spatial frequencies, Indicates random phase, I takes y, , z represent track direction, horizontal direction and elevation direction respectively; q=1,2,…,n pt; N, i represent the dimensions of the random vector of track irregularity spatial frequency and phase angle and the i-th dimension of the vector respectively; M, j represent the dimensions of the random variable of structural parameters and the j-th dimension of the vector respectively; Indicates the total number of track irregularity sampling points.

[0026] According to the first aspect, in some possible implementations, in S1 , the vehicle speed ranges from 100 to 350 km / h.

[0027] In a second aspect, a railway vehicle-track-bridge system dynamic response prediction system is provided, comprising:

[0028] A vehicle speed and track irregularity sample acquisition module is used to obtain a sample set containing vehicle speed and track irregularity;

[0029] The physical model calculation module is used to sequentially input the samples in the sample set into the vehicle-track-bridge system coupled random distribution physical model to obtain the corresponding bridge dynamic response and the effective load of the vehicle-track-bridge system, and to extract the overall stiffness matrix of the vehicle-track-bridge system;

[0030] A training sample set construction module is used to construct a training sample set, where each training sample in the training sample set contains vehicle speed, track irregularity, corresponding bridge dynamic response, and effective load of the vehicle-track-bridge system;

[0031] BP neural network prediction model building module, which is used to build a BP neural network prediction model with vehicle speed and track irregularity as input and bridge dynamic response as output;

[0032] A training module is used to construct a fitness function that takes the payload into consideration, optimize the weights and thresholds of the BP neural network prediction model using a genetic algorithm based on a training sample set, initialize the BP neural network prediction model using the obtained optimal weights and thresholds, and train the initialized BP neural network prediction model based on the training sample set to obtain a bridge dynamic response prediction model;

[0033] The prediction module is used to predict the dynamic response of the bridge using the bridge dynamic response prediction model.

[0034] According to the second aspect, in some possible implementations, in the training module, the fitness function considering the payload is expressed as follows:

[0035]

[0036] In the formula, fitness represents the fitness value; represents the vth output prediction value obtained by the bridge dynamic response prediction model for the kth training sample, Indicates the actual value of the vth output corresponding to the kth training sample; represents the number of training samples, Indicates the number of outputs of the bridge dynamic response prediction model; represents the overall stiffness matrix of the vehicle-track-bridge system; represents the effective load prediction value corresponding to the v-th output obtained by the bridge dynamic response prediction model for the k-th training sample, represents the actual value of the effective load corresponding to the kth training sample; when the vth output is the bridge displacement response, , when the vth output is the bridge acceleration response, ;

[0037] When calculating the fitness value of a chromosome, first use formula [1] to calculate the fitness value. If the value obtained by formula [1] ,in , then the result obtained by formula [1] is used as the fitness value of the chromosome; if the result obtained by formula [1] , then use formula [2] to calculate the fitness value, and take the larger value of the results calculated by formula [1] and formula [2] as the fitness value of the chromosome.

[0038] According to the second aspect, in some possible implementations, in the training module, the process of optimizing the weights and thresholds of the BP neural network prediction model using the genetic algorithm includes:

[0039] Step 1: Initialize the population. Each chromosome in the population is encoded by the weight and threshold of the BP neural network prediction model.

[0040] Step 2: Use the fitness function to calculate the fitness value of each chromosome;

[0041] Step 3: Selection operation, based on the fitness value of the chromosome, select a certain number of chromosomes to enter the next generation;

[0042] Step 4: Crossover operation: according to the crossover probability, two chromosomes are randomly selected from the chromosomes after the selection operation to cross over and generate new chromosomes;

[0043] Step 5: Mutation operation: According to the mutation probability, randomly select one chromosome from the chromosomes after the crossover operation to perform one or more gene mutations to generate a new chromosome;

[0044] Step 6: Repeat steps 2 to 5 until the termination condition is reached, and output the optimal weights and thresholds of the BP neural network prediction model.

[0045] In a third aspect, a railway vehicle-track-bridge system dynamic response prediction system is provided, comprising:

[0046] A data acquisition module is used to obtain the vehicle speed and track irregularity to be predicted and input them into the prediction module;

[0047] The prediction module is configured with a bridge dynamic response prediction model trained in the railway vehicle-track-bridge system dynamic response prediction method as described in any one of the first aspects, and is used to input the data obtained by the data acquisition module into the bridge dynamic response prediction model to predict the bridge dynamic response.

[0048] In a fourth aspect, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the method for predicting the dynamic response of a railway vehicle-track-bridge system as described in any one of the first aspects is implemented. Beneficial effects

[0049] The present invention proposes a method, system and medium for predicting the dynamic response of a railway vehicle-track-bridge system. By introducing the effective load in the vehicle-track-bridge system into the fitness function of the genetic algorithm, an organic combination of a neural network model and a vehicle-track-bridge physical model is achieved, and an implicit relationship between the two is found, which effectively improves the prediction accuracy and solves the technical problem that existing prediction methods cannot take into account the internal factors of the vehicle-track-bridge system, which essentially limits the improvement of prediction accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0051] FIG1 is a flow chart of a method for predicting dynamic response of a railway vehicle-track-bridge system according to an embodiment of the present invention;

[0052] 2 is a comparison diagram of bridge acceleration response predicted by a traditional GA-BP neural network prediction model and an improved GA-BP neural network prediction model according to an embodiment of the present invention;

[0053] FIG3 is a comparison diagram of bridge displacement responses predicted by a traditional GA-BP neural network prediction model provided by an embodiment of the present invention and an improved GA-BP neural network prediction model of the present invention. Modes for Carrying Out the Invention

[0054] To make the objectives, technical solutions, and advantages of the present invention more apparent, the technical solutions of the present invention will be described in detail below. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other implementations obtained by those of ordinary skill in the art without inventive effort are within the scope of protection of the present invention.

[0055] In order to solve the technical problem that existing prediction methods cannot consider the internal factors of the vehicle-track-bridge system, which essentially limits the improvement of prediction accuracy, the present invention proposes a dynamic response prediction scheme for the railway vehicle-track-bridge system based on an improved genetic algorithm to optimize the BP neural network. By introducing the fitness function of the vehicle-track-bridge system into the genetic algorithm, an organic combination of the neural network model and the vehicle-track-bridge physical model is achieved, and an implicit relationship between the two is found, thereby improving prediction accuracy. The technical scheme of the present invention is described in detail below with reference to specific embodiments.

[0056] As shown in FIG1 , an embodiment of the present invention provides a method for predicting the dynamic response of a railway vehicle-track-bridge system, including:

[0057] S1: Obtain a sample set containing vehicle speed and track irregularity.

[0058] By randomly acquiring vehicle speed samples, in this embodiment, the vehicle speed range is 100-350 km / h; and using the random harmonic function method to acquire track irregularity samples, the vehicle speed samples and track irregularity samples are paired and combined to obtain a sample set containing vehicle speed and track irregularity.

[0059] The track unevenness is constructed using a random harmonic function, which is expressed as follows:

[0060]

[0061] Where, 、 、 They represent track irregularity, horizontal irregularity and height irregularity respectively, among which, represents the random variable point set space, and x represents the representative point set of random variables; represents the amplitude of the orbital irregularity wavelength, represents random spatial frequencies, Indicates random phase, I takes y, , z represent track direction, horizontal direction and elevation direction respectively; q=1,2,…,n pt ; N, i represent the dimensions of the random vector of track irregularity spatial frequency and phase angle and the i-th dimension of the vector respectively; M, j represent the dimensions of the random variable of structural parameters and the j-th dimension of the vector respectively; Indicates the total number of track irregularity sampling points.

[0062] From the random harmonic function of the track irregularity power spectrum, we know that its amplitude is:

[0063]

[0064] Where, Represent the power spectral density functions of track irregularity, horizontal irregularity, and height irregularity, are roughness coefficients, Both represent the cutoff frequency,

[0065] ,

[0066] ;

[0067] Indicates the spatial frequency of track irregularities.

[0068] S2: Input the samples in the sample set into the coupled random distribution physical model of the vehicle-track-bridge system in sequence to obtain the corresponding bridge dynamic response and the effective load of the vehicle-track-bridge system, and extract the overall stiffness matrix of the vehicle-track-bridge system.

[0069] The stochastic distribution physical model of the coupled vehicle-track-bridge system is currently available and will not be further elaborated here. Each sample containing vehicle speed and track irregularity is input into the stochastic distribution physical model to calculate the corresponding bridge dynamic response (which serves as the true value of the bridge dynamic response during subsequent training) and the corresponding vehicle-track-bridge system payload. The bridge dynamic response includes at least one of a displacement response and an acceleration response.

[0070] S3: Construct a training sample set. Each training sample in the training sample set contains vehicle speed, track irregularity, corresponding bridge dynamic response, and the effective load of the vehicle-track-bridge system.

[0071] S4: Using vehicle speed and track irregularity as inputs and bridge dynamic response as output, a BP neural network prediction model is constructed. The BP neural network is the core component of the feedforward network. It features a simple structure, numerous adjustable parameters, multiple trainable algorithms, and excellent operability. It also offers excellent prediction capabilities. Global optimization using a genetic algorithm can improve prediction accuracy.

[0072] S5: Construct a fitness function that takes the payload into consideration, use a genetic algorithm to optimize the weights and thresholds of the BP neural network prediction model based on the training sample set, initialize the BP neural network prediction model using the obtained optimal weights and thresholds, and train the initialized BP neural network prediction model based on the training sample set to obtain a bridge dynamic response prediction model.

[0073] Specifically, the fitness function considering the payload is expressed as follows:

[0074]

[0075] In the formula, fitness represents the fitness value; represents the vth output prediction value obtained by the bridge dynamic response prediction model for the kth training sample, Indicates the actual value of the vth output corresponding to the kth training sample; represents the number of training samples, Indicates the number of outputs of the bridge dynamic response prediction model; represents the overall stiffness matrix of the vehicle-track-bridge system; represents the effective load prediction value corresponding to the v-th output obtained by the bridge dynamic response prediction model for the k-th training sample, represents the actual value of the effective load corresponding to the kth training sample; when the vth output is the bridge displacement response, , when the vth output is the bridge acceleration response, ;

[0076] When calculating the fitness value of a chromosome, first use formula [1] to calculate the fitness value. If the value obtained by formula [1] ,in, The result obtained by formula [1] is used as the fitness value of the chromosome; if the result obtained by formula [1] , then use formula [2] to calculate the fitness value, and take the larger value of the results calculated by formula [1] and formula [2] as the fitness value of the chromosome.

[0077] The fitness function of a traditional genetic algorithm (GA) only considers the error between the predicted value and the actual value of the prediction model. This results in the GA-BP neural network being unable to consider the essential characteristics of the prediction object, making the prediction model lacking robustness. In this embodiment, the fitness function of the vehicle-track-bridge system is introduced into the genetic algorithm. The improved fitness function can comprehensively consider the internal parameters of the vehicle-track-bridge system, providing the genetic algorithm with more physical model data, facilitating the acquisition of the optimal weights and thresholds for the BP neural network prediction model. This achieves an organic combination of the neural network model and the vehicle-track-bridge physical model, improving the robustness and prediction accuracy of the prediction model.

[0078] The process of using genetic algorithm to optimize the weights and thresholds of the BP neural network prediction model includes:

[0079] Step 1: Initialize the population. Each chromosome in the population is encoded by the weight and threshold of the BP neural network prediction model. The initial values ​​of the weight and threshold are set to random numbers between [-0.5, 0.5]. Initialize parameters such as population size, number of evolutions, crossover probability, and mutation probability.

[0080] Step 2: Calculate the fitness value of each chromosome using the above fitness function considering payload.

[0081] Step 3: Selection operation, based on the fitness value of the chromosome, a certain number of chromosomes are selected to enter the next generation using the roulette wheel selection method.

[0082] Step 4: Crossover operation. According to the crossover probability, two chromosomes are randomly selected from the chromosomes after the selection operation to perform crossover and generate new chromosomes. For two chromosomes A and B with a length of l, the crossover operation can be expressed as follows:

[0083]

[0084] Where, and They represent the pth genes of the two new chromosomes generated after crossover, and represents the pth gene of chromosome A and B respectively, r is any value between [0,1], .

[0085] Step 5: Mutation operation: According to the mutation probability, randomly select one chromosome from the chromosomes after the crossover operation to perform one or more gene mutations to generate a new chromosome.

[0086] Step 6: Repeat steps 2 to 5 until the termination condition is reached, and output the optimal weights and thresholds of the BP neural network prediction model. The termination condition can be reaching the maximum number of evolutions.

[0087] After obtaining the optimal weights and thresholds, they are substituted into the weights and thresholds of the BP neural network prediction model to initialize them. The initialized BP neural network prediction model is then trained based on the training sample set to obtain a bridge dynamic response prediction model. During training, parameters such as the learning rate, convergence error, and maximum convergence times are set. Minimizing the prediction error is the goal, with vehicle speed (i.e., track irregularity) as input and bridge response as output for training. In this embodiment, the improved genetic algorithm optimizes the BP neural network parameters as shown in the following table:

[0088]

[0089] S6: Use the bridge dynamic response prediction model to predict the bridge dynamic response, that is, obtain the vehicle speed and track irregularity to be predicted, and input them into the bridge dynamic response prediction model to predict the bridge dynamic response.

[0090] Figures 2 and 3 show comparisons of the bridge acceleration response and displacement response predicted by the traditional GA-BP neural network prediction model and the improved GA-BP neural network prediction model of the present invention, respectively. The horizontal axis represents the number of train steps, and the vertical axis represents the corresponding dynamic response of the bridge. The three curves represent the actual value, the predicted value by the traditional GA-BP neural network prediction model, and the predicted value by the improved GA-BP neural network prediction model of the present invention, respectively. As can be seen from the figures, the predicted value of the improved GA-BP neural network prediction model of the present invention is closer to the actual value and has higher prediction accuracy.

[0091] An embodiment of the present invention further provides a railway vehicle-track-bridge system dynamic response prediction system, comprising:

[0092] A vehicle speed and track irregularity sample acquisition module is used to obtain a sample set containing vehicle speed and track irregularity;

[0093] The physical model calculation module is used to sequentially input the samples in the sample set into the vehicle-track-bridge system coupled random distribution physical model to obtain the corresponding bridge dynamic response and the effective load of the vehicle-track-bridge system, and to extract the overall stiffness matrix of the vehicle-track-bridge system;

[0094] A training sample set construction module is used to construct a training sample set, where each training sample in the training sample set contains vehicle speed, track irregularity, corresponding bridge dynamic response, and effective load of the vehicle-track-bridge system;

[0095] BP neural network prediction model building module, which is used to build a BP neural network prediction model with vehicle speed and track irregularity as input and bridge dynamic response as output;

[0096] A training module is used to construct a fitness function that takes the payload into consideration, optimize the weights and thresholds of the BP neural network prediction model using a genetic algorithm based on a training sample set, initialize the BP neural network prediction model using the obtained optimal weights and thresholds, and train the initialized BP neural network prediction model based on the training sample set to obtain a bridge dynamic response prediction model;

[0097] The prediction module is used to predict the dynamic response of the bridge using the bridge dynamic response prediction model.

[0098] Specifically, in the training module, the fitness function considering the payload is expressed as follows:

[0099]

[0100] In the formula, fitness represents the fitness value; represents the vth output prediction value obtained by the bridge dynamic response prediction model for the kth training sample, Indicates the actual value of the vth output corresponding to the kth training sample; represents the number of training samples, Indicates the number of outputs of the bridge dynamic response prediction model; represents the overall stiffness matrix of the vehicle-track-bridge system; represents the effective load prediction value corresponding to the v-th output obtained by the bridge dynamic response prediction model for the k-th training sample, represents the actual value of the effective load corresponding to the kth training sample; when the vth output is the bridge displacement response, , when the vth output is the bridge acceleration response, When calculating the fitness value of the chromosome, first use formula [1] to calculate the fitness value. If the value obtained by formula [1] ,in , then the result obtained by formula [1] is used as the fitness value of the chromosome; if the result obtained by formula [1] , then use formula [2] to calculate the fitness value, and take the larger value of the results calculated by formula [1] and formula [2] as the fitness value of the chromosome.

[0101] In the training module, the process of using genetic algorithm to optimize the weights and thresholds of the BP neural network prediction model includes:

[0102] Step 1: Initialize the population. Each chromosome in the population is encoded by the weight and threshold of the BP neural network prediction model.

[0103] Step 2: Use the fitness function to calculate the fitness value of each chromosome;

[0104] Step 3: Selection operation, based on the fitness value of the chromosome, select a certain number of chromosomes to enter the next generation;

[0105] Step 4: Crossover operation: according to the crossover probability, two chromosomes are randomly selected from the chromosomes after the selection operation to cross over and generate new chromosomes;

[0106] Step 5: Mutation operation: According to the mutation probability, randomly select one chromosome from the chromosomes after the crossover operation to perform one or more gene mutations to generate a new chromosome;

[0107] Step 6: Repeat steps 2 to 5 until the termination condition is reached, and output the optimal weights and thresholds of the BP neural network prediction model.

[0108] It should be understood that the functional unit modules in each embodiment of the present invention may be integrated into a single processing unit, or each unit module may exist physically separately, or two or more unit modules may be integrated into a single unit module, and may be implemented in the form of hardware or software. For other details in this embodiment, please refer to the same or similar parts in the previous embodiment, and will not be repeated here.

[0109] An embodiment of the present invention further provides a railway vehicle-track-bridge system dynamic response prediction system, comprising:

[0110] A data acquisition module is used to obtain the vehicle speed and track irregularity to be predicted and input them into the prediction module;

[0111] The prediction module is configured with a bridge dynamic response prediction model obtained by training in the railway vehicle-track-bridge system dynamic response prediction method provided in the above embodiment, and is used to obtain the data obtained by the data acquisition module and input it into the bridge dynamic response prediction model to predict the bridge dynamic response.

[0112] An embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for predicting the dynamic response of a railway vehicle-track-bridge system provided in the above embodiment is implemented.

[0113] It can be understood that the same or similar parts of the above embodiments can be referenced to each other, and the contents not described in detail in some embodiments can refer to the same or similar contents in other embodiments.

[0114] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0115] This application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate a device for implementing the functions specified in one or more processes in the flowchart and / or one or more blocks in the block diagram.

[0116] These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce a product including an instruction device that implements the functions specified in one or more processes in the flowchart and / or one or more boxes in the block diagram.

[0117] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more processes in the flowchart and / or one or more boxes in the block diagram.

[0118] Although the embodiments of the present invention have been shown and described above, it will be understood that the above embodiments are illustrative and are not to be construed as limitations on the present invention. A person skilled in the art may change, modify, replace and modify the above embodiments within the scope of the present invention.

Claims

1. A method for predicting the dynamic response of a railway track bridge system, characterized in that, Including: S1: Obtain a sample set including vehicle speed and track irregularity; S2: Input the samples in the sample set into the vehicle-track-bridge system coupled random distribution physical model in sequence to obtain the corresponding bridge dynamic response and the effective load of the vehicle-track-bridge system, and extract the overall stiffness matrix of the vehicle-track-bridge system; S3: Construct a training sample set, where each training sample in the training sample set includes vehicle speed, track irregularity, and the corresponding bridge dynamic response and the effective load of the vehicle-track-bridge system; S4: Construct a BP neural network prediction model with vehicle speed and track irregularity as inputs and bridge dynamic response as output; S5: Construct a fitness function considering the effective load, optimize the weights and thresholds of the BP neural network prediction model using a genetic algorithm based on the training sample set, initialize the BP neural network prediction model with the obtained optimal weights and thresholds, and train the initialized BP neural network prediction model based on the training sample set to obtain a bridge dynamic response prediction model; S6: Use the bridge dynamic response prediction model to predict the bridge dynamic response.

2. The dynamic response prediction method for a railway track bridge system according to claim 1, wherein In S5, the fitness function considering the effective load is expressed as follows: ； Wherein, fitness represents the fitness value; Denote the v-th output prediction value obtained by the k-th training sample through the bridge dynamic response prediction model. Denote the v-th actual output value corresponding to the k-th training sample; Indicates the number of training samples, Indicates the number of outputs of the bridge dynamic response prediction model; Denote the overall stiffness matrix of the vehicle-track-bridge system; Denote the payload prediction value corresponding to the v-th output obtained by the k-th training sample through the bridge dynamic response prediction model. denote the actual value of the payload corresponding to the k-th training sample; when the v-th output is the bridge displacement response, , when the v-th output is the bridge acceleration response, ； When calculating the fitness value of chromosomes, first calculate the fitness value using Equation [1]. If the result obtained from Equation [1] , where, Then, the result obtained from Equation [1] is used as the fitness value of this chromosome; if the result obtained from Equation [1] , then calculate the fitness value using Equation [2], and take the larger value of the calculation results of Equation [1] and Equation [2] as the fitness value of this chromosome.

3. The method for predicting the dynamic response of a railway track bridge system according to claim 2, wherein In S5, the process of optimizing the weights and thresholds of the BP neural network prediction model using a genetic algorithm includes: Step 1: Initialize the population, where each chromosome in the population is encoded by the weights and thresholds of the BP neural network prediction model; Step 2: Calculate the fitness value of each chromosome using the fitness function; Step 3: Selection operation, based on the fitness values of the chromosomes, select a certain number of chromosomes to enter the next generation; Step 4: Crossover operation, randomly select two chromosomes from the chromosomes after the selection operation according to the crossover probability to perform crossover and generate new chromosomes; Step 5: Mutation operation, randomly select one chromosome from the chromosomes after the crossover operation according to the mutation probability to perform one or more gene mutations and generate new chromosomes; Step 6: Repeat Step 2 to Step 5 until the termination condition is reached, and output the optimal weights and thresholds of the BP neural network prediction model.

4. The dynamic response prediction method for a railway track bridge system according to claim 1, wherein In S1, the track irregularity is constructed using a random harmonic function, which is expressed as follows: ； In the formula, 、 、 respectively represent track alignment irregularity, level irregularity and cross-level irregularity, where Denote the space of the random variable point sets, and \(x\) represents the representative point set of the random variable; Represents the amplitude of the track irregularity wavelength, represents a random spatial frequency, Denotes a random phase, I takes y, , z represent track alignment, horizontal direction and vertical alignment respectively; q = 1, 2, …, n pt ; N and i represent the dimension of the random vector of the spatial frequency and phase angle of track irregularities and the i-th dimension of this vector respectively; M and j represent the dimension of the random variable of the structural parameter and the j-th dimension of this vector respectively; represents the total number of track irregularity sampling points.

5. The dynamic response prediction method of the railway track bridge system according to claim 1, characterized in that In S1, the vehicle speed ranges from 100 to 350 km / h.

6. A dynamic response prediction system for a railway track bridge system, characterized in that, Including: A vehicle speed and track irregularity sample acquisition module for obtaining a sample set including vehicle speed and track irregularity; A physical model calculation module for inputting the samples in the sample set into the vehicle-track-bridge system coupled random distribution physical model in sequence to obtain the corresponding bridge dynamic response and the effective load of the vehicle-track-bridge system, and extracting the overall stiffness matrix of the vehicle-track-bridge system; A training sample set construction module for constructing a training sample set, where each training sample in the training sample set includes vehicle speed, track irregularity, and the corresponding bridge dynamic response and the effective load of the vehicle-track-bridge system; A BP neural network prediction model construction module for constructing a BP neural network prediction model with vehicle speed and track irregularity as inputs and bridge dynamic response as output; A training module, configured to construct a fitness function considering payloads, optimize the weights and thresholds of a BP neural network prediction model using a genetic algorithm based on a training sample set, initialize the BP neural network prediction model with the obtained optimal weights and thresholds, and train the initialized BP neural network prediction model based on the training sample set to obtain a bridge dynamic response prediction model; A prediction module, configured to predict the bridge dynamic response using the bridge dynamic response prediction model.

7. The dynamic response prediction system of the railway track bridge system according to claim 6, wherein In the training module, the fitness function considering payloads is expressed as follows: ； In the formula, fitness represents the fitness value; Denote the v-th output prediction value obtained by the k-th training sample through the bridge dynamic response prediction model. Denote the v-th actual output value corresponding to the k-th training sample; Indicates the number of training samples, Indicates the number of outputs of the bridge dynamic response prediction model; Denote the overall stiffness matrix of the vehicle-track-bridge system; Denote the payload prediction value corresponding to the v-th output obtained by the k-th training sample through the bridge dynamic response prediction model. Indicates the actual value of the payload corresponding to the k-th training sample; when the v-th output is the bridge displacement response, , when the v-th output is the bridge acceleration response, ； When calculating the fitness value of a chromosome, first calculate the fitness value using Equation [1]. If the result obtained from Equation [1] , where , then the result obtained from Equation [1] is used as the fitness value of this chromosome; if the result obtained from Equation [1] , then the fitness value is calculated using Equation [2], and the larger value of the calculation results of Equation [1] and Equation [2] is used as the fitness value of this chromosome.

8. The dynamic response prediction system for a railway track bridge system according to claim 6, wherein In the training module, the process of optimizing the weights and thresholds of the BP neural network prediction model using a genetic algorithm includes: Step 1: Initialize the population, where each chromosome in the population is encoded by the weights and thresholds of the BP neural network prediction model; Step 2: Calculate the fitness value of each chromosome using the fitness function; Step 3: Selection operation, based on the fitness values of the chromosomes, select a certain number of chromosomes to enter the next generation; Step 4: Crossover operation, according to the crossover probability, randomly select two chromosomes from the chromosomes after the selection operation for crossover to generate new chromosomes; Step 5: Mutation operation, according to the mutation probability, randomly select one chromosome from the chromosomes after the crossover operation for one or more gene mutations to generate new chromosomes; Step 6: Repeat Step 2 to Step 5 until the termination condition is reached, and output the optimal weights and thresholds of the BP neural network prediction model.

9. A dynamic response prediction system for a railway track bridge system, characterized in that, including: A data acquisition module, configured to acquire the vehicle speed and track irregularity of the response to be predicted and input them into the prediction module; A prediction module, on which a bridge dynamic response prediction model trained by the railway vehicle-track-bridge system dynamic response prediction method described in any one of claims 1 to 5 is configured, and is used to input the data acquired by the data acquisition module into the bridge dynamic response prediction model to predict the bridge dynamic response.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the railway vehicle-track-bridge system dynamic response prediction method described in any one of claims 1 to 5.

Images