High-speed railway subgrade settlement identification method based on multi-source data and knowledge dual drive

Abstract

The invention relates to a high-speed railway subgrade settlement identification method based on multi-source data and knowledge dual drive. The method comprises the following steps: firstly, acquiring roadbed settlement multi-source data of a high-speed railway roadbed settlement observation section, and preprocessing; then, dividing the preprocessed roadbed settlement multi-source data into a training set and a test set, and establishing an SVR support vector machine optimal basis evaluator and a BP neural network optimal basis evaluator based on the training set; then, a multi-factor B-S-N model is constructed based on the optimal basis evaluator, and the multi-factor B-S-N model comprises a strong evaluator of an SVR support vector machine, a strong evaluator of a BP neural network and an output module; and finally, inputting the test set into the multi-factor B-S-N model to obtain a roadbed settlement prediction result. More accurate high-speed railway subgrade settlement prediction in a complex natural environment is realized, the deformation condition of the subgrade can be quickly mastered, and coping and remedy can be carried out in time.

Description

Technical Field

[0001] The present application relates to the field of roadbed engineering technology, and in particular to a high-speed railway roadbed settlement identification method based on dual drive of multi-source data and knowledge. Background Art

[0002] During the construction of high-speed railways, the settlement of the roadbed is a key factor in measuring road stability and an important indicator for evaluating project quality. Therefore, dynamic monitoring and prediction of roadbed settlement is particularly critical during construction. Traditional roadbed displacement and settlement monitoring methods only analyze real-time monitoring data, ignoring early warning of possible displacement. When an accident occurs, only post-event remediation can be carried out, missing the best time for reinforcement, causing economic losses and even delaying construction progress. In order to quickly grasp the deformation of the roadbed, it is crucial to predict future settlement trends based on existing monitoring data.

[0003] In the tide of informatization, applying artificial intelligence methods to the field of settlement prediction is undoubtedly an important research direction. Among them, neural networks and support vector machine regression, as new methods for dealing with such complex nonlinear problems, have been widely used in roadbed settlement prediction.

[0004] The reason why neural networks have been widely used for a long time is mainly because they have the following advantages: First, neural networks have powerful nonlinear processing capabilities and can effectively solve nonlinear clustering problems. Second, the network has the ability to optimize calculations and can find a set of suitable parameter combinations under specific constraints to minimize the error of the determined objective function. Although neural networks have many advantages and have been widely used in engineering practice, they have been found to have some disadvantages during long-term use. For example, neural networks are sometimes difficult to obtain the optimal solution because the nonlinear transfer function in the multi-layer neural network may have multiple local optimal solutions. In addition, the search for the optimal point is closely related to the selection of the initial point. When the initial point is closer to the local optimal point than the global optimal point, it is difficult to obtain the global optimal solution.

[0005] Support vector machine regression (SVR) has advantages in solving nonlinear small sample problems and has strong adaptability. Compared with traditional artificial neural network models, SVR is not easy to fall into local optimality and can make up for the shortcomings of complex neural network structure. However, traditional SVR models have the defects of long running time and low stability, and also have certain limitations.

[0006] Therefore, in the relevant technology, there is an urgent need for a method that can improve the accuracy of high-speed railway subgrade settlement prediction in complex natural environments. Summary of the invention

[0007] Based on this, it is necessary to provide a high-speed railway subgrade settlement identification method based on multi-source data and knowledge dual drive, which can improve the prediction accuracy of high-speed railway subgrade settlement in complex natural environments, in response to the above technical problems.

[0008] In a first aspect, the present application provides a high-speed railway subgrade settlement identification method based on multi-source data and knowledge dual drive. The method comprises:

[0009] Obtain multi-source data of subgrade settlement of high-speed railway subgrade settlement observation section and perform preprocessing;

[0010] The pre-processed roadbed settlement multi-source data is divided into a training set and a test set, and an SVR support vector machine optimal basis estimator and a BP neural network optimal basis estimator are established based on the training set;

[0011] Constructing a multi-factor BSN model based on the optimal basis evaluator, the multi-factor BSN model includes a strong evaluator of an SVR support vector machine, a strong evaluator of a BP neural network and an output module, the parameters of the strong evaluator of the SVR support vector machine are determined by the parameters of the optimal basis evaluator of the SVR support vector machine, the parameters of the strong evaluator of the BP neural network are determined by the parameters of the optimal basis evaluator of the BP neural network, and the output module weights the outputs of the strong evaluator of the SVR support vector machine and the strong evaluator of the BP neural network to obtain a final output;

[0012] The test set is input into the multi-factor BSN model to obtain the roadbed settlement prediction result.

[0013] Optionally, in one embodiment of the present application, the preprocessing includes:

[0014] Abnormal data were eliminated, missing data were filled, and standardized.

[0015] Optionally, in one embodiment of the present application, constructing a multi-factor BSN model based on the optimal basis evaluator includes:

[0016] Determine the mean absolute error of the SVR support vector machine and the BP neural network respectively based on the optimal basis evaluator;

[0017] If the relative difference ratio of the mean absolute error is less than a preset threshold, the weight of the strong evaluator of the multi-factor BSN model is determined based on the mean absolute error, and the optimal prediction result of the roadbed settlement of the multi-factor BSN model is determined based on the weight of the strong evaluator.

[0018] Optionally, in one embodiment of the present application, the weight of the strong evaluator of the multi-factor BSN model is expressed as:

[0019]

[0020]

[0021] Among them, MAE BP is the mean absolute error, MAE, of the BP neural network strong evaluator model SVR is the mean absolute error of the SVR support vector machine strong estimator model.

[0022] Optionally, in one embodiment of the present application, the method further includes:

[0023] The root mean square error and mean absolute error are used to judge the degree of deviation between the predicted value and the measured value.

[0024] In the second aspect, the present application also provides a high-speed railway subsidence identification device based on multi-source data and knowledge dual drive. The device comprises:

[0025] The data acquisition module is used to acquire multi-source data of the subgrade settlement of the high-speed railway subgrade settlement observation section and perform preprocessing;

[0026] The model initial parameter adjustment module is used to divide the pre-processed roadbed settlement multi-source data into a training set and a test set, and establish an SVR support vector machine optimal basis estimator and a BP neural network optimal basis estimator based on the training set;

[0027] A roadbed settlement identification model construction module, used for constructing a multi-factor BSN model based on the optimal basis evaluator, the multi-factor BSN model includes a strong evaluator of an SVR support vector machine, a strong evaluator of a BP neural network and an output module, the parameters of the strong evaluator of the SVR support vector machine are determined by the parameters of the optimal basis evaluator of the SVR support vector machine, the parameters of the strong evaluator of the BP neural network are determined by the parameters of the optimal basis evaluator of the BP neural network, and the output module weights the outputs of the strong evaluator of the SVR support vector machine and the strong evaluator of the BP neural network to obtain a final output;

[0028] The roadbed settlement identification module is used to input the test set into the multi-factor BSN model to obtain the roadbed settlement prediction result.

[0029] In a third aspect, the present application further provides a computer device, wherein the computer device comprises a memory and a processor, wherein the memory stores a computer program, and the processor executes the steps of the methods described in the above embodiments.

[0030] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the method described in each of the above embodiments are implemented.

[0031] The high-speed railway subgrade settlement identification method based on dual-drive of multi-source data and knowledge first obtains the subgrade settlement multi-source data of the high-speed railway subgrade settlement observation section and performs preprocessing; then, the preprocessed subgrade settlement multi-source data is divided into a training set and a test set, and an SVR support vector machine optimal basis evaluator and a BP neural network optimal basis evaluator are established based on the training set; then, a multi-factor BSN model is constructed based on the optimal basis evaluator, and the multi-factor BSN model includes a strong evaluator of the SVR support vector machine, a strong evaluator of the BP neural network and an output module, the parameters of the strong evaluator of the SVR support vector machine are determined by the parameters of the SVR support vector machine optimal basis evaluator, the parameters of the strong evaluator of the BP neural network are determined by the parameters of the BP neural network optimal basis evaluator, and the output module weights the outputs of the strong evaluator of the SVR support vector machine and the strong evaluator of the BP neural network to obtain the final output; finally, the test set is input into the multi-factor BSN model to obtain the subgrade settlement prediction result. In other words, by combining the BP neural network with the support vector machine regression SVR, a multi-factor BSN model is constructed, and the roadbed settlement is predicted based on multi-source data of the roadbed. This achieves a more accurate prediction of high-speed railway roadbed settlement in a complex natural environment, enables the deformation of the roadbed to be quickly grasped, and timely response and remedial measures can be taken. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 This is an application environment diagram of a high-speed railway subgrade settlement identification method based on dual drive of multi-source data and knowledge in one embodiment;

[0033] Figure 2 A schematic diagram of a process of identifying high-speed railway subgrade settlement based on dual drive of multi-source data and knowledge in one embodiment;

[0034] Figure 3 is a schematic diagram of the structure of a multi-source data observation system in one embodiment;

[0035] Figure 4 is a schematic diagram of the prediction results of the SVR strong evaluator model in one embodiment;

[0036] Figure 5 A schematic diagram of a prediction result of a BP neural network strong evaluator model in one embodiment;

[0037] Figure 6A schematic diagram of a process for constructing a multi-factor BSN model in one embodiment;

[0038] Figure 7 It is a structural block diagram of a high-speed railway subgrade settlement identification device based on multi-source data and knowledge dual drive in one embodiment;

[0039] Figure 8 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0040] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0041] The high-speed railway subgrade settlement identification method based on multi-source data and knowledge dual drive provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or it can be placed on the cloud or other network servers. Among them, the terminal 102 can be but is not limited to various personal computers, laptops, smart phones, tablets, Internet of Things devices and portable wearable devices. The Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart car-mounted devices, etc. Portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. The server 104 can be implemented with an independent server or a server cluster consisting of multiple servers.

[0042] In one embodiment, Figure 2 As shown in the figure, a high-speed railway subgrade settlement identification method based on multi-source data and knowledge dual drive is provided. Figure 1 The server in the example is used to illustrate the following steps:

[0043] S201: Acquire multi-source data of roadbed settlement of a high-speed railway roadbed settlement observation section and perform preprocessing.

[0044] In the embodiment of the present application, firstly, the multi-source data of the subgrade settlement of the high-speed railway subgrade settlement observation section is obtained. The multi-source data of the subgrade settlement includes the soil stress value at 0.5m below the top surface of the subgrade, the middle of the subgrade and the bottom of the subgrade, the subgrade settlement value, the water content inside the soil, the average temperature of the day and the maximum temperature difference of the day. Specifically, the multi-source data of the subgrade settlement is obtained by real-time measurement by constructing a multi-source data observation system for the subgrade settlement, such as Figure 3The figure shows the structure diagram of the roadbed settlement multi-source data observation system, which includes a stress sensor, a static level, a moisture meter, a thermometer, a data transmission system and a power supply system, wherein the stress sensor is used to measure the soil stress value 0.5m below the top surface of the roadbed, in the middle of the roadbed and at the bottom of the roadbed, the static level is used to obtain the settlement data of the foundation and the roadbed pavement, the moisture meter is used to obtain the moisture content of the soil inside the foundation, and the thermometer is used to obtain the ambient temperature near the roadbed. After obtaining the roadbed settlement multi-source data through the roadbed settlement multi-source data observation system, the data needs to be preprocessed.

[0045] Specifically, in one embodiment of the present application, the preprocessing includes:

[0046] Abnormal data were eliminated, missing data were filled, and standardized.

[0047] In one embodiment of the present application, for the multi-source data of roadbed settlement obtained, first, the obvious abnormal data is eliminated, and then the missing data is filled with missing values ​​using random forest regression. Finally, the data is standardized, and its standardized calculation expression is shown in the following formula.

[0048]

[0049] Among them, X* is the standardized data, X is the original data, μ is the mean of the original data, and σ is the standard deviation of the original data.

[0050] S203: Divide the preprocessed roadbed settlement multi-source data into a training set and a test set, and establish an SVR support vector machine optimal basis estimator and a BP neural network optimal basis estimator based on the training set.

[0051] In the embodiment of the present application, the pre-processed multi-source data of roadbed settlement are randomly divided, wherein 70% of the data are used as training samples to form a training set, and 30% of the data are used as test samples to form a test set. Afterwards, the model is initially adjusted, and the optimal basis estimators of the SVR support vector machine and the BP neural network are established based on the training set. The SVR basis estimator is trained using the training set data and a grid search is performed to obtain the optimal hyperparameters, and the optimal basis estimator of the SVR support vector machine is established. Specifically, a grid search is performed on the hyperparameters of SVR: kernel (kernel function), gamma, degree, coef0, and C. Among them, the kernel function kernel selects four types: linear (linear kernel function), poly (polynomial kernel function), sigmoid (S-type kernel function), and rbf (radial basis function kernel). The expressions of the four kernel functions are as follows:

[0052] K linear =x Ty = x·y

[0053] K poly = (γ(x·y) + r) d

[0054] K sigmoid = tanh(γ(x·y) + r)

[0055]

[0056] Among them, γ is the hyperparameter gamma, d is the hyperparameter degree, r is the hyperparameter coef0, gamma is selected from 1, 2, 3; the value range of degree is [-10, 1]; the value range of cofe0 is [0, 5], and the value range of the penalty coefficient C of the relaxation coefficient is [0.01, 10].

[0057] For the optimal basis estimator of the BP neural network, the number of nodes in the input layer and output layer of the neural network depends on the specific problem of the actual research. In this method, the data to be input is seven-dimensional variable data, that is, there are seven input variables in the input layer, so the number of input layer nodes is determined to be seven. The output is the subgrade settlement prediction. According to the input variables, a relatively accurate subgrade settlement value is output, and the number of nodes in the output layer is set to one. First, an empirical formula is used to determine the range of the number of hidden layer neurons. The selected empirical formula is as follows:

[0058]

[0059] Among them, P is the number of hidden layer neurons, m and n are the number of nodes in the input layer and output layer respectively; a is a constant, and 1 < a < 10. The range of the number of hidden layer nodes is determined to be: 4 < P < 14. Then, Python is used to construct a learning curve to determine the mean square error of the model corresponding to the number of neurons in each hidden layer, so as to determine the optimal number of hidden layer neurons. Sigmoid, Purelin, tanh, and relu are respectively selected to train the model. By determining the mean square error of the model corresponding to each transfer function, the optimal transfer function is further determined. And the hyperparameters w and b are optimized with the Adam optimizer to determine the optimal hyperparameters w and b.

[0060] S205: Construct a multi-factor BSN model based on the optimal basis evaluator, the multi-factor BSN model includes a strong evaluator of an SVR support vector machine, a strong evaluator of a BP neural network and an output module, the parameters of the strong evaluator of the SVR support vector machine are determined by the parameters of the optimal basis evaluator of the SVR support vector machine, the parameters of the strong evaluator of the BP neural network are determined by the parameters of the optimal basis evaluator of the BP neural network, and the output module weights the outputs of the strong evaluator of the SVR support vector machine and the strong evaluator of the BP neural network to obtain a final output.

[0061] S207: Input the test set into the multi-factor BSN model to obtain a roadbed settlement prediction result.

[0062] In an embodiment of the present application, a strong evaluator of SVR and BP neural network is constructed based on the optimal basis evaluator of SVR and BP neural network, and then a multi-factor BSN model is created based on this and combined with the output module. Specifically, in an embodiment of the present application, the multi-factor BSN model is constructed based on the optimal basis evaluator, including:

[0063] S301: Determine the mean absolute errors of the SVR support vector machine and the BP neural network respectively based on the optimal basis evaluator.

[0064] S303: If the relative difference ratio of the mean absolute error is less than a preset threshold, the weight of the strong evaluator of the multi-factor BSN model is determined based on the mean absolute error, and the optimal prediction result of the roadbed settlement of the multi-factor BSN model is determined based on the weight of the strong evaluator.

[0065] In one embodiment of the present application, first, Bootstrap sampling is performed on the data set to form 20 groups of different sub-samples. Ten groups of sub-samples are selected, and the optimal hyperparameters of the SVR base evaluator are used as the hyperparameters of the 10 groups of sub-samples. Ten groups of independent SVR base evaluators are constructed, and the mean of the roadbed settlement predicted by the 10 groups of base evaluators is used as the prediction result of the SVR strong evaluator model, such as Figure 4 As shown, the test set is brought into the model to obtain the SVR strong evaluator prediction value y SVR , and get the mean absolute error MAE of the model SVR Another 10 groups of sub-samples are selected, and the best hyperparameters of the BP neural network base evaluator are used as the hyperparameters of the 10 groups of sub-samples. Ten groups of independent BP neural network base evaluators are constructed, and the mean of the roadbed settlement predicted by the 10 groups of base evaluators is used as the prediction result of the BP neural network strong evaluation model, as shown in Figure 2. Figure 5 As shown, the test set is brought into the model to obtain the BP neural network strong evaluator prediction value y BP, and get the mean absolute error MAE of the model BP .

[0066] Afterwards, the model is further optimized, such as Figure 6 As shown, a preset threshold value θ is defined. If the relative difference ratio of the mean absolute errors of the two algorithms is less than the preset threshold value, the weights of the two strong evaluators in the multi-factor BSN model are determined based on the mean absolute errors of the two strong evaluators, and the optimal prediction result of the roadbed settlement of the multi-factor BSN model is determined based on the weights of the strong evaluators. The specific formula is shown below.

[0067]

[0068] In one embodiment of the present application, the weight of the strong evaluator of the multi-factor BSN model is expressed as:

[0069]

[0070] Among them, MAE BP is the mean absolute error, MAE, of the BP neural network strong evaluator model SVR is the mean absolute error of the SVR support vector machine strong estimator model.

[0071] The prediction result y of the multi-factor BSN model is BSN for:

[0072] y BSN =W SVR y SVR +W BP y BP

[0073] Among them, y SVR The predicted value y for the SVR strong estimator SVR ,y BP It is the prediction value of BP neural network strong evaluator.

[0074] If the relative difference ratio of the mean absolute errors of the two algorithms is not less than the preset threshold, the mean absolute errors of the two algorithms are further compared to determine the weight of the strong evaluator of the corresponding multi-factor BSN model. The specific calculation process is as follows.

[0075] If MAE SVR >MAE BP , W SVR =0,W BP =1; if MAE SVR <MAE BP , W SVR =1,W BP =0.

[0076] The learning curve of the preset threshold value θ is established through python and the optimal threshold value θ of the multi-factor BSN model is determined to obtain the best prediction result of roadbed settlement.

[0077] In one embodiment of the present application, the method further includes:

[0078] The root mean square error and mean absolute error are used to judge the degree of deviation between the predicted value and the measured value.

[0079] In one embodiment of the present application, the root mean square error (RMSE) and the mean absolute error (MAE) are used to judge the degree of deviation between the predicted value and the measured value. RMSE reflects the average degree of deviation, that is, the square root of the ratio of the sum of the squares of the deviation between the predicted value and the actual value to the number of samples n. It is mainly used to indicate the degree of deviation between the predicted value and the actual value of the overall model, and is easily affected by the maximum and minimum values. MAE is the average value of the absolute error, which is used to indicate the error between the actual value and the predicted value of the overall model. In actual use, both are used to observe whether the predicted value of the model as a whole is close to the actual value.

[0080] The root mean square error (RMSE) formula is:

[0081]

[0082] The smaller the RMSE value, the higher the accuracy of the prediction model. Because the maximum or minimum error between the predicted value and the measured value will have a serious impact on its final value, the RMSE can better reflect the overall error deviation of the prediction model.

[0083] The mean absolute error (MAE) formula is:

[0084]

[0085] The smaller the MAE value, the better the performance of the prediction model, which can better reflect the overall average error of the model.

[0086] In the above-mentioned high-speed railway embankment settlement identification method based on dual drive of multi-source data and knowledge, first, the embankment settlement multi-source data of the high-speed railway embankment settlement observation section is obtained and preprocessed; then, the preprocessed embankment settlement multi-source data is divided into a training set and a test set, and an SVR support vector machine optimal basis evaluator and a BP neural network optimal basis evaluator are established based on the training set; then, a multi-factor BSN model is constructed based on the optimal basis evaluator, and the multi-factor BSN model includes a strong evaluator of the SVR support vector machine, a strong evaluator of the BP neural network and an output module, the parameters of the strong evaluator of the SVR support vector machine are determined by the parameters of the SVR support vector machine optimal basis evaluator, the parameters of the strong evaluator of the BP neural network are determined by the parameters of the BP neural network optimal basis evaluator, and the output module weights the outputs of the strong evaluator of the SVR support vector machine and the strong evaluator of the BP neural network to obtain the final output; finally, the test set is input into the multi-factor BSN model to obtain the embankment settlement prediction result. In other words, by combining the BP neural network with the support vector machine regression SVR, a multi-factor BSN model is constructed, and the roadbed settlement is predicted based on multi-source data of the roadbed. This achieves a more accurate prediction of high-speed railway roadbed settlement in a complex natural environment, enables the deformation of the roadbed to be quickly grasped, and timely response and remedial measures can be taken.

[0087] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0088] Based on the same inventive concept, the embodiment of the present application also provides a high-speed railway subgrade settlement identification device based on multi-source data and knowledge dual drive for implementing the high-speed railway subgrade settlement identification method based on multi-source data and knowledge dual drive mentioned above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more embodiments of the high-speed railway subgrade settlement identification device based on multi-source data and knowledge dual drive provided below can refer to the above limitations on the high-speed railway subgrade settlement identification method based on multi-source data and knowledge dual drive, and will not be repeated here.

[0089] In one embodiment, Figure 7 As shown, a high-speed railway subsidence identification device 700 based on dual drive of multi-source data and knowledge is provided, comprising: a data acquisition module 701, a model initial parameter adjustment module 703, a subsidence identification model construction module 705 and a subsidence identification module 707, wherein:

[0090] The data acquisition module 701 is used to acquire multi-source data of the subgrade settlement of the high-speed railway subgrade settlement observation section and perform pre-processing.

[0091] The model initial parameter adjustment module 703 is used to divide the pre-processed roadbed settlement multi-source data into a training set and a test set, and establish an SVR support vector machine optimal basis estimator and a BP neural network optimal basis estimator based on the training set.

[0092] The roadbed settlement identification model construction module 705 is used to construct a multi-factor BSN model based on the optimal basis evaluator. The multi-factor BSN model includes a strong evaluator of an SVR support vector machine, a strong evaluator of a BP neural network and an output module. The parameters of the strong evaluator of the SVR support vector machine are determined by the parameters of the optimal basis evaluator of the SVR support vector machine, and the parameters of the strong evaluator of the BP neural network are determined by the parameters of the optimal basis evaluator of the BP neural network. The output module weights the outputs of the strong evaluator of the SVR support vector machine and the strong evaluator of the BP neural network to obtain the final output.

[0093] The roadbed settlement identification module 707 is used to input the test set into the multi-factor BSN model to obtain the roadbed settlement prediction result.

[0094] In one embodiment of the present application, the data acquisition module is further used to:

[0095] Abnormal data were eliminated, missing data were filled, and standardized.

[0096] In one embodiment of the present application, the roadbed settlement identification model building module is also used to:

[0097] Determine the mean absolute error of the SVR support vector machine and the BP neural network respectively based on the optimal basis evaluator;

[0098] If the relative difference ratio of the mean absolute error is less than a preset threshold, the weight of the strong evaluator of the multi-factor BSN model is determined based on the mean absolute error, and the optimal prediction result of the roadbed settlement of the multi-factor BSN model is determined based on the weight of the strong evaluator.

[0099] In one embodiment of the present application, the weight of the strong evaluator of the multi-factor BSN model is expressed as:

[0100]

[0101] Among them, MAE BP is the mean absolute error, MAE, of the BP neural network strong evaluator model SVR is the mean absolute error of the SVR support vector machine strong estimator model.

[0102] In one embodiment of the present application, the method further includes:

[0103] The root mean square error and mean absolute error are used to judge the degree of deviation between the predicted value and the measured value.

[0104] Each module in the high-speed railway subgrade settlement identification device based on multi-source data and knowledge dual drive can be implemented in whole or in part by software, hardware and their combination. Each module can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to each module above.

[0105] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 8 As shown. The computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a high-speed railway subgrade settlement identification method based on multi-source data and knowledge dual drive is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a button, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse.

[0106] Those skilled in the art will understand that Figure 8The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0107] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0108] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0109] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0110] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0111] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0112] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0113] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A high-speed railway subgrade settlement identification method based on multi-source data and knowledge dual drive, characterized in that: The method comprises: Obtain multi-source data of subgrade settlement of high-speed railway subgrade settlement observation section and perform preprocessing; The pre-processed roadbed settlement multi-source data is divided into a training set and a test set, and an SVR support vector machine optimal basis estimator and a BP neural network optimal basis estimator are established based on the training set; Constructing a multi-factor BSN model based on the optimal basis evaluator, the multi-factor BSN model includes a strong evaluator of an SVR support vector machine, a strong evaluator of a BP neural network and an output module, the parameters of the strong evaluator of the SVR support vector machine are determined by the parameters of the optimal basis evaluator of the SVR support vector machine, the parameters of the strong evaluator of the BP neural network are determined by the parameters of the optimal basis evaluator of the BP neural network, and the output module weights the outputs of the strong evaluator of the SVR support vector machine and the strong evaluator of the BP neural network to obtain a final output; The test set is input into the multi-factor BSN model to obtain the roadbed settlement prediction result.

2. The high-speed railway subgrade settlement identification method based on multi-source data and knowledge dual drive according to claim 1 is characterized in that: The pre-processing comprises: Abnormal data were eliminated, missing data were filled, and standardized.

3. The high-speed railway subgrade settlement identification method based on multi-source data and knowledge dual drive according to claim 1 is characterized in that: The constructing of a multi-factor BSN model based on the optimal basis evaluator comprises: Determine the mean absolute error of the SVR support vector machine and the BP neural network respectively based on the optimal basis evaluator; If the relative difference ratio of the mean absolute error is less than a preset threshold, the weight of the strong evaluator of the multi-factor BSN model is determined based on the mean absolute error, and the optimal prediction result of the roadbed settlement of the multi-factor BSN model is determined based on the weight of the strong evaluator.

4. The high-speed railway subgrade settlement identification method based on multi-source data and knowledge dual drive according to claim 3 is characterized in that: The weight of the strong evaluator of the multi-factor BSN model is expressed as: Among them, MAE BP is the mean absolute error, MAE, of the BP neural network strong evaluator model SVR is the mean absolute error of the SVR support vector machine strong estimator model.

5. The high-speed railway roadbed settlement identification method based on multi-source data and knowledge dual drive according to claim 1 is characterized in that: The method further comprises: The root mean square error and mean absolute error are used to judge the degree of deviation between the predicted value and the measured value.

6. A high-speed railway subsidence identification device based on multi-source data and knowledge dual drive, characterized in that: The device comprises: The data acquisition module is used to acquire multi-source data of the subgrade settlement of the high-speed railway subgrade settlement observation section and perform preprocessing; The model initial parameter adjustment module is used to divide the pre-processed roadbed settlement multi-source data into a training set and a test set, and establish an SVR support vector machine optimal basis estimator and a BP neural network optimal basis estimator based on the training set; A roadbed settlement identification model construction module, used for constructing a multi-factor BSN model based on the optimal basis evaluator, the multi-factor BSN model includes a strong evaluator of an SVR support vector machine, a strong evaluator of a BP neural network and an output module, the parameters of the strong evaluator of the SVR support vector machine are determined by the parameters of the optimal basis evaluator of the SVR support vector machine, the parameters of the strong evaluator of the BP neural network are determined by the parameters of the optimal basis evaluator of the BP neural network, and the output module weights the outputs of the strong evaluator of the SVR support vector machine and the strong evaluator of the BP neural network to obtain a final output; The roadbed settlement identification module is used to input the test set into the multi-factor BSN model to obtain the roadbed settlement prediction result.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 5 are implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

Images