Road settlement data prediction method and device based on GRU model and medium

Abstract

The invention provides a road settlement data prediction method and device based on a GRU model, equipment and a medium, and the method comprises the steps: collecting road settlement data, carrying out the preprocessing, and arranging the road settlement data of each measurement point according to a time sequence, so as to form time sequence data; extracting a feature construction rule from the time series data, and processing the time series data into structured data according to the rule; constructing a GRU prediction model, configuring a multi-layer perceptron structure, using Dense as a hidden layer, using ReLu and a linear function as activation functions, using a mean square error as a loss function and an Adam optimizer for training, and using structured data as a data set to train the GRU prediction model to obtain a trained GRU prediction model; and performing road settlement data prediction by adopting the trained GRU prediction model, and generating a prediction result of a road settlement trend so as to solve the problem that the road facilities are damaged due to the fact that the deformation trend of the road facilities cannot be predicted in time.

Description

Technical Field

[0001] The present invention relates to the technical field of bridge monitoring, and in particular to a method, device, equipment and medium for predicting road settlement data based on a GRU model. Background Art

[0002] Urban road infrastructure, the lifeblood and foundation of a city's development, is one of its "lifelines." If a city is likened to a living organism, then the crisscrossing network of roads is its blood vessels. In recent years, with the acceleration of urbanization and the rapid development of underground engineering construction, the application of open-cut technology has become increasingly restricted, significantly increasing the demand for trenchless technology.

[0003] Since underground engineering construction inevitably disturbs the surrounding strata, causing deformation or settlement of the strata, which in turn causes deformation and settlement of existing road facilities within its impact range. If the deformation or settlement exceeds the control indicators, it will cause damage to the facilities, thereby affecting the safe operation of the road facilities.

[0004] Currently, the industry's primary regulatory approach for trenchless construction crossing existing road infrastructure involves third-party monitoring during underground engineering projects. This third-party monitoring of existing road infrastructure allows for timely measurement of deformation indicators, and construction guidance based on this data. However, this method of protecting road infrastructure has its drawbacks. When monitoring data exceeds standards or exhibits anomalies, damage to the infrastructure has already occurred, necessitating measures such as suspension of work and grouting to prevent further damage. This process also results in delayed warnings and high-cost measures.

[0005] Therefore, it is urgent to propose a road settlement data prediction method based on the GRU model to solve the technical problem of damage to road facilities caused by the inability to timely predict the deformation trend of road facilities. Summary of the Invention

[0006] In order to overcome the problems existing in the related art, the present disclosure provides a road settlement data prediction method, device, equipment and medium based on the GRU model to solve the technical problem in the related art that the deformation trend of road facilities cannot be predicted in time, resulting in damage to road facilities.

[0007] One or more embodiments of this specification provide a method for predicting road subsidence data based on a GRU model, comprising the following steps:

[0008] Collect road settlement data, and after pre-processing, arrange the road settlement data of each measuring point in chronological order to form time series data;

[0009] Extracting features from the time series data to construct rules, and processing the time series data into structured data according to the rules;

[0010] Construct a GRU prediction model, configure a multi-layer perceptron structure, use Dense as the hidden layer and ReLu and linear functions as activation functions, use mean square error as the loss function and Adam optimizer for training, use the structured data as a data set to train the GRU prediction model, and obtain a trained GRU prediction model;

[0011] The trained GRU prediction model is used to predict road subsidence data and generate prediction results of road subsidence trends.

[0012] Preferably, the rules include cumulative difference, daily difference and risk assessment.

[0013] Preferably, the method of using the structured data as a data set to train a GRU prediction model specifically includes the following steps:

[0014] Convert the structured data into sequences, delete null values, and construct a data set from sequences with a length greater than 1;

[0015] Split the dataset into time series of fixed length, which are used as training set and validation set respectively;

[0016] The training set is input into the GRU prediction model for training, and the validation set is input into the GRU prediction model for performance evaluation.

[0017] Preferably, the method further comprises the following steps:

[0018] The prediction results are visualized and risk warning information is displayed.

[0019] One or more embodiments of this specification provide a road subsidence data prediction device based on a GRU model, including a data acquisition module, a rule construction module, a model training module, and a prediction module;

[0020] The data acquisition module is used to collect road settlement data and arrange the road settlement data of each measuring point in chronological order after pre-processing to form time series data;

[0021] The rule construction module is used to extract features from the time series data to construct rules, and process the time series data into structured data according to the rules;

[0022] The model training module is used to build a GRU prediction model, configure a multi-layer perceptron structure, use Dense as the hidden layer and ReLu and linear functions as activation functions, use mean square error as the loss function and Adam optimizer for training, and use the structured data as a data set to train the GRU prediction model to obtain a trained GRU prediction model;

[0023] The prediction module is used to use the trained GRU prediction model to predict road subsidence data and generate prediction results of road subsidence trends.

[0024] Preferably, the rules include cumulative difference, daily difference and risk assessment.

[0025] Preferably, the model training module includes a data set construction unit, a data set segmentation unit and a model training unit;

[0026] The data set construction unit is used to convert the structured data into sequences, delete null values, and construct a data set from sequences with a length greater than 1;

[0027] The data set segmentation unit is used to segment the data set into time series of fixed lengths, which are used as a training set and a validation set respectively;

[0028] The model training unit is used to input the training set into the GRU prediction model for training, and input the verification set into the GRU prediction model for performance evaluation.

[0029] Preferably, it also includes a display module for visually displaying the prediction results and risk warning information.

[0030] One or more embodiments of the present specification provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the road subsidence data prediction method based on the GRU model as described above is implemented.

[0031] One or more embodiments of the present specification provide a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the steps of the above-mentioned method for predicting road subsidence data based on the GRU model are implemented.

[0032] The present disclosure provides a road subsidence data prediction method, device, equipment and medium based on the GRU model, which has the advantages of collecting road subsidence data, arranging the road subsidence data of each measuring point in chronological order after preprocessing to form time series data, ensuring the temporal sequence and integrity of the data, and laying a solid foundation for subsequent feature extraction and model training; extracting features from the time series data to construct rules, and processing the time series data into structured data according to the rules, which can effectively capture the key features in the road subsidence data and improve the input quality of the model; constructing a GRU prediction model, configuring a multi-layer perceptron structure, using Dense as the hidden layer and ReLu and linear functions as activation functions, using mean square error as the loss function and Adam optimizer for training, and using the structured data as a data set to train the GRU prediction model, thereby obtaining a trained GRU prediction model, which can effectively handle the long-term dependencies of the time series data and improve the prediction accuracy of the model. Furthermore, the use of mean squared error as a loss function and the Adam optimizer for training further improved the model's convergence speed and prediction performance. The trained GRU prediction model was used to predict road subsidence data and generate prediction results of road subsidence trends, providing a scientific basis for road maintenance and management. This method not only has high prediction accuracy but also adapts to complex road subsidence variations, demonstrating its strong practicality and potential for promotion. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate one or more embodiments of this specification 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 recorded in this specification. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0034] Figure 1 A schematic flow chart of a road subsidence data prediction method based on a GRU model provided in one or more embodiments of this specification;

[0035] Figure 2 A schematic diagram for visually displaying prediction results provided in one or more embodiments of this specification;

[0036] Figure 3 A schematic diagram of the structure of a road subsidence data prediction device based on a GRU model provided in one or more embodiments of this specification;

[0037] Figure 4 A schematic diagram of the structure of a computer device provided in one or more embodiments of this specification. DETAILED DESCRIPTION

[0038] In order to enable those skilled in the art to better understand the technical solutions in one or more embodiments of this specification, the technical solutions in one or more embodiments of this specification will be clearly and completely described below in conjunction with the drawings in one or more embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this specification, not all of the embodiments. Based on one or more embodiments of this specification, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of this invention document.

[0039] The present invention will be described in detail below with reference to specific implementation methods and the accompanying drawings.

[0040] Method Example

[0041] According to an embodiment of the present invention, a road subsidence data prediction method based on a GRU (Gated Recurrent Unit) model is provided. Figure 1 FIG. 1 is a flow chart of a method for predicting road subsidence data based on a GRU model according to an embodiment of the present invention. The method for predicting road subsidence data based on a GRU model according to an embodiment of the present invention includes the following steps:

[0042] S110. Collect road settlement data from the data obtained from the previous monitoring measurements. After preprocessing, arrange the road settlement data of each measuring point in chronological order to form time series data. The preprocessing includes cleaning measuring points with a large number of missing values. For measuring points with a small number of missing values, fill them in using the mean or interpolation method.

[0043] S120: Extract meaningful features from the time series data to construct rules, and process the time series data into structured data according to the rules.

[0044] S130. Build a GRU prediction model, configure a multi-layer perceptron structure, use Dense as the hidden layer and ReLu and linear functions as activation functions, use mean square error as the loss function and Adam optimizer for training, use the structured data as a data set to train the GRU prediction model, and obtain a trained GRU prediction model.

[0045] S140: Use the trained GRU prediction model to predict road subsidence data and generate a prediction result of the road subsidence trend.

[0046] The method provided in this embodiment collects road subsidence data, arranges the road subsidence data of each measuring point in chronological order after preprocessing, and forms time series data, thereby ensuring the time sequence and integrity of the data and laying a solid foundation for subsequent feature extraction and model training; extracts features from the time series data to construct rules, and processes the time series data into structured data according to the rules, which can effectively capture the key features in the road subsidence data and improve the input quality of the model; constructs a GRU prediction model, configures a multi-layer perceptron structure, uses Dense as the hidden layer and ReLu and linear functions as activation functions, uses mean square error as the loss function and Adam optimizer for training, and uses the structured data as a data set to train the GRU prediction model, thereby obtaining a trained GRU prediction model, which can effectively handle the long-term dependency of time series data and improve the prediction accuracy of the model. At the same time, using mean square error as the loss function and Adam optimizer for training further optimizes the convergence speed and prediction performance of the model; uses the trained GRU prediction model to predict road subsidence data and generate prediction results of road subsidence trends, which can generate prediction results of road subsidence trends and provide a scientific basis for road maintenance and management. This method not only has high prediction accuracy, but also can adapt to complex changes in road settlement, and has strong practicality and promotion value.

[0047] In one embodiment, the rules include cumulative differences, daily differences, and risk assessments.

[0048] Specifically, the Cumulative Difference function calculates the cumulative change in settlement at each measurement point since its initial moment. This feature can effectively reveal long-term trends. By using the Cumulative Difference function, the model can capture the overall changes in settlement, rather than just short-term fluctuations.

[0049] Daily Difference calculates the difference in settlement between each time point and the previous time point, reflecting the short-term fluctuations in settlement. Through daily difference, the model can capture sudden changes in settlement and help identify sudden settlement events.

[0050] Risk assessment rules set certain rules to assess risk. For example, when the cumulative difference exceeds a certain threshold, it is marked as a potential risk point; when the daily difference exceeds a set value, it is marked as an emergency risk. These rules can help identify areas requiring special attention. By setting thresholds, outliers or high-risk points in the data can be marked for subsequent analysis and processing.

[0051] Assume we have the following raw data:

[0052] Settlement value of measuring point A at time point

[0053]

[0054] Cumulative difference:

[0055] Time point 1:10

[0056] Time point 2: 12-10=2

[0057] Time point 3: 15-10=5

[0058] Time point 4: 14-10=4

[0059] Time point 5: 18-10=8

[0060] Daily difference:

[0061] Time point 1: None (because there is no previous time point)

[0062] Time point 2: 12-10=2

[0063] Time point 3: 15-12=3

[0064] Time point 4: 14-15 = -1

[0065] Time point 5: 18-14=4

[0066] Risk Assessment Rules:

[0067] Assume the cumulative difference threshold is 5 and the daily difference threshold is 3:

[0068] The cumulative difference between time points 3 and 5 exceeds 5 and is marked as a potential risk point.

[0069] The daily difference between time points 3 and 5 exceeds 3, which is marked as an emergency risk.

[0070] The method provided in this embodiment converts the original data into structured data including cumulative differences, daily differences and risk assessments, thereby providing important features for subsequent model training and supporting the functions of settlement prediction and risk warning.

[0071] In one embodiment, the structured data is used as a data set to train a GRU prediction model, specifically including the following steps:

[0072] The structured data is converted into sequences, and null values ​​are deleted, and a data set is constructed from sequences with a length greater than 1.

[0073] The dataset is divided into time series of fixed length, for example, sequences of 60 time units, and these sequences are extracted from the dataset as training sets and validation sets respectively.

[0074] The training set is input into the GRU prediction model for training, and the validation set is input into the GRU prediction model for performance evaluation.

[0075] The method provided in this embodiment constructs a valid dataset by converting structured data into sequences and removing null values, improving data quality and adaptability and providing reliable input for the model. Splitting the data into fixed-length time series to form training and validation sets leverages the GRU model's strengths in processing sequential data, enabling the model to deeply learn the temporal dependencies of settlement data. Using the training set for training and the validation set for evaluation allows for efficient optimization of model parameters, ensuring good generalization and prediction accuracy, and providing solid support for road settlement prediction.

[0076] like Figure 2 As shown, this is a schematic diagram of visually displaying the prediction results provided by this embodiment. In one embodiment, the following steps are also included:

[0077] The prediction results and risk warning information are visualized. Specifically, a prediction trend chart is drawn using Matplotlib. The chart contains the following content:

[0078] Blue line: represents historical settlement data, that is, the actual observed settlement value.

[0079] Yellow lines: Show the input data points, which are the data used for prediction.

[0080] Green line: represents the model's prediction results, that is, the possible future subsidence trend.

[0081] Red dotted line: represents the warning line. When the prediction result exceeds this line, it indicates that there may be risks.

[0082] Device embodiment

[0083] According to an embodiment of the present invention, a road subsidence data prediction device based on a GRU model is provided. Figure 3 As shown, this is a structural diagram of the road subsidence data prediction device based on the GRU model provided in this embodiment. The road subsidence data prediction device based on the GRU model according to an embodiment of the present invention includes a data acquisition module 31, a rule construction module 32, a model training module 33 and a prediction module 34.

[0084] The data acquisition module 31 is used to collect road settlement data, and arrange the road settlement data of each measuring point in chronological order after pre-processing to form time series data.

[0085] The rule construction module 32 is used to extract features from the time series data to construct rules, and process the time series data into structured data according to the rules, wherein the rules include cumulative difference, daily difference and risk assessment.

[0086] The model training module 33 is used to build a GRU prediction model, configure a multi-layer perceptron structure, use Dense as the hidden layer and ReLu and linear functions as activation functions, use mean square error as the loss function and Adam optimizer for training, and use the structured data as a data set to train the GRU prediction model to obtain a trained GRU prediction model.

[0087] The prediction module 34 is used to use the trained GRU prediction model to predict road subsidence data and generate prediction results of road subsidence trends.

[0088] It also includes a display module for visually displaying the prediction results and risk warning information.

[0089] In the device provided by this embodiment, the data acquisition module 31 collects road subsidence data, and after pre-processing, arranges the road subsidence data of each measuring point in chronological order to form time series data, thereby ensuring the timeliness and integrity of the data and laying a solid foundation for subsequent feature extraction and model training; the rule construction module 32 extracts features from the time series data to construct rules, and processes the time series data into structured data according to the rules, which can effectively capture the key features in the road subsidence data and improve the input quality of the model; the model training module 33 constructs a GRU prediction model, configures a multi-layer perceptron structure, uses Dense as the hidden layer and ReLu and linear functions as activation functions, uses mean square error as the loss function and Adam optimizer for training, and uses the structured data as a data set to train the GRU prediction model to obtain a trained GRU prediction model, which can effectively handle the long-term dependencies of time series data and improve the prediction accuracy of the model. Furthermore, the use of mean squared error as a loss function and the Adam optimizer for training further improved the model's convergence speed and prediction performance. Prediction module 34 uses the trained GRU prediction model to predict road subsidence data and generate prediction results of road subsidence trends. This method can generate prediction results of road subsidence trends and provide a scientific basis for road maintenance and management. This method not only has high prediction accuracy but also can adapt to complex road subsidence variations, making it highly practical and promising for promotion.

[0090] In one embodiment, the model training module 33 includes a data set construction unit, a data set segmentation unit and a model training unit.

[0091] The data set construction unit is used to convert the structured data into sequences, delete null values, and construct a data set from sequences with a length greater than 1.

[0092] The data set segmentation unit is used to segment the data set into time series of fixed lengths, which are used as a training set and a validation set respectively.

[0093] The model training unit is used to input the training set into the GRU prediction model for training, and input the verification set into the GRU prediction model for performance evaluation.

[0094] The device provided in this embodiment constructs a valid dataset by converting structured data into sequences and removing null values, improving data quality and adaptability and providing reliable input for the model. Splitting the data into fixed-length time series to form training and validation sets leverages the GRU model's strengths in processing sequential data, enabling the model to deeply learn the temporal dependencies of settlement data. Using the training set for training and the validation set for evaluation allows for efficient optimization of model parameters, ensuring good generalization and prediction accuracy, and providing solid support for road settlement prediction.

[0095] The embodiment of the present invention is an apparatus embodiment corresponding to the above-mentioned method embodiment. The specific operations of the processing steps of each module can be understood by referring to the description of the method embodiment, and will not be repeated here.

[0096] like Figure 4 As shown, the present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the road subsidence data prediction method in the above embodiment is implemented, or when the computer program is executed by a processor, the road subsidence data prediction method based on the GRU model in the above embodiment is implemented.

[0097] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. 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 memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0098] Each embodiment in this specification is described in a progressive manner. The same or similar parts between the embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments. In particular, for the device or system embodiments, since they are basically similar to the method embodiments, the description is relatively simple. For the relevant parts, refer to the partial description of the method embodiments. The device and system embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the scheme of this embodiment. A person of ordinary skill in the art can understand and implement it without making any creative efforts.

[0099] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention, and the contents not described in detail in the specification of the present invention belong to the common knowledge of those skilled in the art.

Claims

1. A road subsidence data prediction method based on the GRU model, characterized in that: The following steps are involved: Collect road settlement data, and after pre-processing, arrange the road settlement data of each measuring point in chronological order to form time series data; Extracting features from the time series data to construct rules, and processing the time series data into structured data according to the rules; Construct a GRU prediction model, configure a multi-layer perceptron structure, use Dense as the hidden layer and ReLu and linear functions as activation functions, use mean square error as the loss function and Adam optimizer for training, use the structured data as a data set to train the GRU prediction model, and obtain a trained GRU prediction model; The trained GRU prediction model is used to predict road subsidence data and generate prediction results of road subsidence trends.

2. The road subsidence data prediction method based on the GRU model according to claim 1, characterized in that: The rules include cumulative difference, daily difference and risk assessment.

3. The road subsidence data prediction method based on the GRU model according to claim 1, characterized in that: The method of using the structured data as a data set to train the GRU prediction model specifically includes the following steps: Convert the structured data into sequences, delete null values, and construct a data set from sequences with a length greater than 1; Split the dataset into time series of fixed length, which are used as training set and validation set respectively; The training set is input into the GRU prediction model for training, and the validation set is input into the GRU prediction model for performance evaluation.

4. The road subsidence data prediction method based on the GRU model according to claim 1, characterized in that: The following steps are also included: The prediction results are visualized and risk warning information is displayed.

5. A road settlement data prediction device based on the GRU model, characterized in that: Includes data acquisition module, rule building module, model training module and prediction module; The data acquisition module is used to collect road settlement data and arrange the road settlement data of each measuring point in chronological order after pre-processing to form time series data; The rule construction module is used to extract features from the time series data to construct rules, and process the time series data into structured data according to the rules; The model training module is used to build a GRU prediction model, configure a multi-layer perceptron structure, use Dense as the hidden layer and ReLu and linear functions as activation functions, use mean square error as the loss function and Adam optimizer for training, and use the structured data as a data set to train the GRU prediction model to obtain a trained GRU prediction model; The prediction module is used to use the trained GRU prediction model to predict road subsidence data and generate prediction results of road subsidence trends.

6. The road subsidence data prediction device based on the GRU model according to claim 5, characterized in that: The rules include cumulative difference, daily difference and risk assessment.

7. The road subsidence data prediction device based on the GRU model according to claim 5, characterized in that: The model training module includes a data set construction unit, a data set segmentation unit and a model training unit; The data set construction unit is used to convert the structured data into sequences, delete null values, and construct a data set from sequences with a length greater than 1; The data set segmentation unit is used to segment the data set into time series of fixed lengths, which are used as a training set and a validation set respectively; The model training unit is used to input the training set into the GRU prediction model for training, and input the verification set into the GRU prediction model for performance evaluation.

8. The road subsidence data prediction device based on the GRU model according to claim 5, characterized in that: It also includes a display module for visually displaying the prediction results and risk warning information.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the road subsidence data prediction method based on the GRU model as described in any one of claims 1 to 4 is implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the road subsidence data prediction method based on the GRU model as claimed in any one of claims 1 to 4 are implemented.

Images