Long-time-sequence mining area ground surface settlement prediction method, system, medium and equipment

Through the improved Informer model and InSAR technology, combined with iTranformer Block and attention distillation mechanism, the problem of long-term dependence of deep learning methods in the prediction of surface settlement in long-term mining areas is solved, and high accuracy prediction of mining area settlement information is achieved. Technical support.

CN118940014AActive Publication Date: 2024-11-12GEOPHYSICAL SURVEY TEAM OF SHANDONG COALFIELD GEOLOGY BUREAU

Patent Information

Application Number
CN202411418799.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-10-12
Publication Date
2024-11-12
Estimated Expiration
2044-10-12

AI Technical Summary

Technical Problem

Existing deep learning methods are difficult to effectively deal with long-term dependence phenomena in surface settlement prediction of long-term mining areas, resulting in difficulty in model training or loss of information, lack of integrated monitoring-prediction models, and cannot accurately obtain settlement prediction values ​​of more than 12 timestamps. .

Method used

Using improved Informer model and InSAR technology, high-quality long-term data is obtained through denoising and time series difference processing, threshold partitioning is used to extract spatial neighborhood information, and combined with iTranformer Block and attention distillation mechanism to reduce time complexity and memory usage to achieve long-term prediction of mining area settlement information.

Benefits of technology

Long-term prediction of settlement information in mining areas is realized, auxiliary decision-making and technical support is provided, the time complexity and memory usage of the model are reduced, and prediction accuracy is improved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118940014A_ABST
    Figure CN118940014A_ABST
Patent Text Reader

Abstract

The invention provides a long-time-sequence mining area ground surface settlement prediction method and system, a medium and equipment, and relates to the technical field of InSAR and geological disaster prevention and control, the initial ground surface settlement rate of a mining area is obtained, and after the initial ground surface settlement rate is preprocessed, long-time-sequence InSAR settlement data at equal time intervals are obtained; calculating the accumulated settlement amount of the earth surface of the mining area by using the InSAR settlement data; partitioning the mining area by using threshold division according to the accumulated settlement, and obtaining spatial neighborhood information of point targets in different mining area partitions according to mining area partitioning results; inputting spatial neighborhood information into an improved Informer model to learn long-time-sequence nonlinear features of local positions in different partitions, and outputting a time sequence settlement prediction value of each pixel to realize long-time-sequence prediction of mining area settlement information; according to the method and the device, the time complexity and the memory occupancy rate of the original Transform under long sequence input are reduced.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of InSAR and geological disaster prevention and control, and in particular to a method, system, medium and equipment for predicting surface subsidence in mining areas over a long time series. Background Art

[0002] The statements in this section merely provide background information related to the present disclosure and do not necessarily constitute prior art.

[0003] Coal is the world's main basic energy and important raw material, and has great strategic significance for the development of an economy. However, long-term coal mining has also brought serious negative ecological and environmental effects to coal resource areas, such as geological disasters such as surface collapse, landslides, and soil erosion induced by coal mining, which have caused damage to roads, bridges, farmland, houses and other structures, seriously threatening the safe production and living environment of local residents. Therefore, the long-term prediction of surface subsidence in mining areas is of great significance for the prevention and control of geological disasters, the restoration of mining subsidence areas, and the safety of life and property of surrounding residents.

[0004] InSAR uses the phase information of two or more radar images covering the same area to quickly extract elevation and subsidence information in a wide range of ground satellite line of sight. It has the advantages of low cost, high temporal and spatial resolution, and no reliance on control points, which is unmatched by traditional geodetic techniques such as leveling, three-dimensional laser scanning, and global positioning systems.

[0005] In existing research, InSAR is often used as a means of data acquisition for surface subsidence monitoring, and there is a lack of an integrated monitoring-prediction model. In fact, in the prevention and control of geological disasters in mines, monitoring and prediction are inseparable. Only by integrating mine subsidence monitoring and prediction can the surface be monitored to timely understand the deformation situation and future deformation characteristics, providing a reasonable basis for mine management decisions. With the rapid development of artificial intelligence, deep learning methods have been widely used in fields such as computer vision. The deep learning network is a multi-layer perceptron containing multiple hidden layers. Through sample learning, a deep nonlinear network structure is obtained to achieve the approximation of complex functions. For complex data, deep learning can extract the essential characteristics of the data by combining low-level features and abstract high-level features, providing a new solution for the prediction of surface deformation in mining areas.

[0006] Currently, most of the deep learning methods widely used for time series prediction are based on recurrent structure networks, such as RNN (Recurrent Neural Network, RNN, recurrent neural network) and LSTM (Short-Term Memory, LSTM, long short-term memory network). Although this type of model can better solve the problem of sequence prediction than other deterministic and traditional statistical models, it is difficult or impossible to train the model because it extracts time series information sequentially and continuously transmits it backwards. At the same time, in the process of sequential calculation, once information is lost, although the gated recurrent unit is able to alleviate the long-term dependency to a certain extent, it is difficult to completely handle the long-term dependency phenomenon. Summary of the invention

[0007] In order to solve the above problems, the present invention proposes a long-time series mining area surface subsidence prediction method, system, medium and equipment. Based on the improved Informer and InSAR, the long-time series prediction of the mining area subsidence change trend is carried out, and the time series prediction values ​​of the settlement points in the study area with more than 12 timestamps are obtained to realize the long-time series prediction of the mining area subsidence information.

[0008] According to some embodiments, the present disclosure adopts the following technical solutions: A long-term mining area surface subsidence prediction method, comprising: Obtain the initial surface subsidence rate of the mining area and preprocess the initial surface subsidence rate, including denoising and time series difference, to fill in missing values ​​and improve data quality, so as to obtain long-time series InSAR subsidence data with equal time intervals and significant time series change characteristics; The InSAR subsidence data is used to calculate the accumulated subsidence of the mining area; the mining area is partitioned according to the accumulated subsidence using threshold division, and the partition basis includes the size of the subsidence and the change of the subsidence rate; then, the spatial neighborhood information of the point target is extracted in each mining area partition; The spatial neighborhood information is input into the improved Informer model to learn the long-term nonlinear characteristics of the settlement points in different partitions, and the time-series settlement prediction value of each settlement point is output to realize the long-term prediction of the settlement information of the mining area.

[0009] According to some embodiments, the present disclosure adopts the following technical solutions: A long-time series mining area surface subsidence prediction system, comprising: Data acquisition module, which obtains the initial surface subsidence rate of the mining area, including denoising and time series difference, to fill in missing values ​​and improve data quality, so as to obtain long-time series InSAR subsidence data with equal time intervals and significant time series change characteristics; The InSAR subsidence data is used to calculate the accumulated subsidence of the mining area; the mining area is partitioned according to the accumulated subsidence using threshold division, and the partition basis includes the size of the subsidence and the change of the subsidence rate; then, the spatial neighborhood information of the point target is extracted in each mining area partition; The prediction module is used to input the spatial neighborhood information into the improved Informer model to learn the long-term nonlinear characteristics of settlement points in different partitions, output the time-series settlement prediction value of the settlement point, and realize the long-term prediction of the settlement information of the mining area.

[0010] According to some embodiments, the present disclosure adopts the following technical solutions: A non-transitory computer-readable storage medium is used to store computer instructions. When the computer instructions are executed by a processor, a long-time series mining area surface subsidence prediction method is implemented.

[0011] According to some embodiments, the present disclosure adopts the following technical solutions: An electronic device comprises: a processor, a memory and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory so that the electronic device implements the long-time series mining area surface subsidence prediction method.

[0012] Compared with the prior art, the present invention has the following beneficial effects: This paper proposes a long-time series mining area surface subsidence prediction method, specifically a long-time series mining area surface subsidence prediction method based on improved Informer and InSAR. The settlement prediction values ​​of more than 12 timestamps are extracted through the existing InSAR mining area surface subsidence data in the study area, and the iTranformer Block and attention distillation mechanism proposed by the improved Informer model are used to reduce the time complexity and memory occupancy rate of the original Transformer under long sequence input. Finally, the long-time series prediction of mining area subsidence can be realized, providing auxiliary decision-making and technical support for mining area subsidence disasters. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] The accompanying drawings constituting a part of the present disclosure are used to provide a further understanding of the present disclosure. The illustrative embodiments of the present disclosure and their descriptions are used to explain the present disclosure and do not constitute an improper limitation on the present disclosure.

[0014] Figure 1 It is a schematic diagram of the overall process of the method for predicting surface subsidence in a mining area over a long time series according to an embodiment of the present disclosure; Figure 2 This is a schematic diagram of the mining area zoning results of an embodiment of the present disclosure; Figure 3 It is a schematic diagram of the structure of an improved Informer model according to an embodiment of the present disclosure; Figure 4 Schematic diagram of the iTranformer block structure of an embodiment of the present disclosure. DETAILED DESCRIPTION

[0015] The present disclosure is further described below in conjunction with the accompanying drawings and embodiments.

[0016] It should be noted that the following detailed descriptions are all illustrative and are intended to provide further explanation of the present disclosure. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present disclosure belongs.

[0017] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present disclosure. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0018] Example 1 In one embodiment of the present disclosure, a method for predicting surface subsidence in a mining area over a long time series is provided, comprising the following steps: Step 1: Obtain the initial surface subsidence rate of the mining area, and pre-process the initial surface subsidence rate, including denoising and time series difference, to fill in missing values ​​and improve data quality, so as to obtain long-term InSAR subsidence data with equal time intervals and significant time series change characteristics; Step 2: Calculate the accumulated surface settlement of the mining area using the InSAR settlement data; partition the mining area according to the normalized accumulated settlement and the change trend of the settlement rate, and obtain the spatial neighborhood information of point targets in different mining area partitions according to the mining area partition results; construct a training data set and a test data set; Step 3: Input the spatial neighborhood information into the improved Informer model to learn the long-term nonlinear characteristics of the settlement points in different partitions, output the time series settlement prediction value of the settlement points, and realize the long-term prediction of the settlement information in the mining area.

[0019] As an embodiment, a long-time series mining area surface subsidence prediction method based on Infomer and InSAR can realize the long-time series prediction of mining area subsidence, which specifically includes: Step 1: Obtain the initial surface subsidence rate of the mining area through InSAR time series processing technology, filter and interpolate the initial surface subsidence rate, and obtain InSAR subsidence rate data with equal time intervals; Step 2: Calculate the accumulated surface settlement of the mining area using the InSAR settlement rate data; After normalizing the accumulated settlement results, the mining area is partitioned according to the normalized accumulated settlement results and the change trend of the settlement rate, and the spatial neighborhood information of the point targets in different partitions is processed according to the partition results; Construct training and test data sets; train different models in each partition to learn the long-term nonlinear characteristics of local locations in different partitions, and evaluate the network performance in the test data; Step 3: Use the trained improved Informer model to make advance predictions and obtain the settlement prediction values ​​of each settlement point for more than 12 timestamps to achieve long-term prediction of the settlement information in the mining area.

[0020] In the above implementation, the process of acquiring long time series InSAR sedimentation data with equal time intervals includes: The initial surface subsidence rate of the mining area was obtained using InSAR time series processing technology; The twice local median absolute deviation of the surface subsidence rate is used as the threshold, and the new surface subsidence rate is obtained by filtering the abnormal data of the original surface subsidence rate by setting a sliding window method; The filtered surface sedimentation rate is interpolated by cubic spline to solve the problem of uneven time intervals of monitoring data caused by irregular revisits of SAR satellites. The time series data with equal time intervals after interpolation are defined as X = {X1, X2, . . ., X m+p}.

[0021] In step 2, the construction of the training data set and the test data set includes: According to the time series data X with equal time intervals, the cumulative settlement data is calculated and normalized using the standard deviation normalization method. The study area is spatially partitioned based on the normalized results, such as Figure 2 As shown in the figure, the study area is divided into four sub-areas. The subsidence characteristics of each sub-area are similar and the spatial locations are close. Training in each sub-area separately helps the model to obtain more reliable results. Extract the spatial neighborhood information of the point target contained in the time series data X, and save it into the training data set Y according to the partition result. Assume that there are n sample points with a time series length of m+p, and take the data of the m time series before the sample point as the training set, denoted as ,in represents partition, and the data on the remaining p time series are used as the test set, denoted as ,but and They can be expressed as:

[0022]

[0023] In step 3, the trained improved Informer model is used for advance prediction, such as Figure 3 As shown, the overall structure of the improved Informer model includes two parts: an encoder and a decoder. Specifically, the encoder and the decoder include an iTranformer block and a self-attention distillation layer, wherein the iTranformer block mainly includes an embedding layer, a multivariate attention mechanism, a normalization layer, a feedforward network, and a fully connected layer. Different from the traditional Transformer module, the embedding layer of the iTranformer block inputs all data on a sample point time series into the network as a group, and the time series can be regarded as an independent process; the multivariate attention mechanism uses linear projection through the self-attention module to obtain Q (query vector), K (key vector) and V (value vector), comprehensively extracts the time series features of the sample point, and reveals the correlation between variables; the normalization layer can reduce the difference between different sample points; the feedforward network is used to extract nonlinear features; the projection layer performs a linear projection operation on the processed results to obtain the final predicted value; the self-attention distillation layer uses KL divergence to extract data with large information entropy, thereby reducing the dimension and the number of network parameters.

[0024] The input of the encoder is a long sequence of spatial neighborhood information of each point target in the mining subsidence area within the time period of known subsidence information. First, the iTranformer block is used to comprehensively extract the time series features and fully explain the correlation between variables. Then, the self-attention distillation layer is used to reduce the dimension. Finally, the output of the encoder is obtained.

[0025] The input of the decoder consists of two parts: the first part is a combination of a short sequence at the end of the time series and a 0 value with the same prediction step length, where the 0 value input to the decoder is used as a placeholder for the mining subsidence prediction value; the second part is the output of the encoder. After the iTranformer block is used to extract features from the first part, it is combined with the second part, and the multi-head attention module is used for deep feature extraction. Finally, the data dimension is adjusted through the fully connected layer to obtain the surface subsidence prediction result of the mining area. The output prediction result is back-gradient propagated after calculating the mean square error (MSE) loss to continuously optimize the network parameters.

[0026] In the above implementation, the settlement prediction includes: Performing a partition operation on the acquired data according to the partition method to obtain verification data of different partitions; After processing the validation data of different partitions according to the data filtering and interpolation methods, the processed data are input into the trained network to obtain the long-term settlement prediction value of each settlement point with more than 12 time stamps.

[0027] Example 2 In one embodiment of the present disclosure, a long-term mining area surface subsidence prediction system is provided, comprising: The data acquisition module obtains the initial surface settlement rate of the mining area, and obtains the long-time series InSAR settlement data with equal time intervals after preprocessing the initial surface settlement rate; Calculate the accumulated surface settlement of the mining area using the InSAR settlement data; partition the mining area using threshold value according to the accumulated settlement, and obtain spatial neighborhood information of point targets in different mining area partitions based on the mining area partition results; The prediction module is used to input the spatial neighborhood information into the improved Informer model to learn the long-term nonlinear characteristics of local positions in different partitions, output the time-series settlement prediction value of each settlement point with more than 12 timestamps, and realize the long-term prediction of the settlement information of the mining area.

[0028] Example 3 In one embodiment of the present disclosure, a non-transitory computer-readable storage medium is provided, wherein the non-transitory computer-readable storage medium is used to store computer instructions. When the computer instructions are executed by a processor, the long-time series mining area surface subsidence prediction method is implemented.

[0029] Example 4 In one embodiment of the present disclosure, an electronic device is provided, including: a processor, a memory and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory, so that the electronic device executes the long-time series mining area surface subsidence prediction method.

[0030] The present disclosure 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 present disclosure. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the 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 generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0031] 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, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0032] Although the above describes the specific implementation methods of the present disclosure in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present disclosure. Technical personnel in the relevant field should understand that on the basis of the technical solution of the present disclosure, various modifications or variations that can be made by those skilled in the art without creative work are still within the scope of protection of the present disclosure.

Claims

1. A method for predicting surface subsidence in a mining area over a long time series, characterized in that: include: Obtain the initial surface subsidence rate of the mining area, and after preprocessing the initial surface subsidence rate, obtain long-time series InSAR subsidence data with equal time intervals; Calculating the accumulated surface subsidence of the mining area using the InSAR subsidence data; According to the accumulated settlement, the mining area is divided into zones using threshold values, and according to the mining area division results, the spatial neighborhood information of point targets in different mining area divisions is obtained; The spatial neighborhood information is input into the improved Informer model to learn the long-term nonlinear characteristics of the settlement points in different partitions, and the time-series settlement prediction values ​​of the settlement points are output to realize the long-term prediction of the settlement information in the mining area.

2. A method for predicting surface subsidence in a mining area over a long time series as claimed in claim 1, characterized in that: The process of obtaining long-time series InSAR sedimentation data with equal time intervals includes: The InSAR time series processing technology is used to obtain the initial surface subsidence rate of the mining area over a long time series; The twice local median absolute deviation of the surface subsidence rate is used as the threshold, and the new surface subsidence rate is obtained by filtering the abnormal data of the original surface subsidence rate by setting a sliding window method; The filtered surface subsidence rate is subjected to cubic spline interpolation to obtain interpolated long-time series InSAR subsidence data with equal time intervals.

3. A method for predicting surface subsidence in a mining area over a long time series as claimed in claim 1, characterized in that: It includes using spatial neighborhood information to construct training data sets and test data sets. Specifically, the cumulative settlement data is calculated based on time series data with equal time intervals, and the cumulative settlement data is normalized using the standard deviation normalization method. The study area is spatially partitioned based on the normalized results to obtain sub-areas with similar settlement characteristics and adjacent spatial locations. Extract the spatial neighborhood information of the point target contained in the time series data and save it into the training data set according to the partition results. Assuming that there are n sample points with a time series length of m+p, the data on the m time series before the sample point is used as the training set, recorded as ,in represents partition; the data on the remaining p time series are used as the test set.

4. A method for predicting surface subsidence in a mining area over a long time series as claimed in claim 1, characterized in that: The overall structure of the improved Informer model includes an encoder and a decoder.

5. A method for predicting long-term mining area surface subsidence as claimed in claim 4, characterized in that: The encoder and decoder include a position information embedding layer, an iTranformer Block and a convolutional distillation layer.

6. A method for predicting long-term mining area surface subsidence as claimed in claim 5, characterized in that: The position embedding layer is used to time encode the input sequence, and the input training set and test set are encoded; the data is input into the encoder and decoder. The input of the encoder is a long sequence composed of the spatial neighborhood information of each point target in the mining subsidence area within the time period of known settlement information. A multi-channel feature map is obtained through iTranformer Block, and the input of the decoder is composed of a short sequence at the end of the time series and a 0 value equal to the prediction step length. A neural network module composed of a multi-head iTranformer Block and a convolutional distillation layer is then used to perform deep feature extraction. Finally, the data dimension is adjusted through the fully connected layer to obtain the surface settlement prediction result of the mining area.

7. A method for predicting long-term mining area surface subsidence according to claim 6, characterized in that: in, The 0 value input to the decoder is used as a placeholder for the mining subsidence prediction value.

8. A long-term mining area surface subsidence prediction system, characterized in that: include: The data acquisition module obtains the initial surface settlement rate of the mining area, and obtains the long-time series InSAR settlement data with equal time intervals after preprocessing the initial surface settlement rate; Calculating the accumulated surface subsidence of the mining area using the InSAR subsidence data; According to the accumulated settlement, the mining area is divided into zones using threshold values, and according to the mining area division results, the spatial neighborhood information of point targets in different mining area divisions is obtained; The prediction module is used to input the spatial neighborhood information into the improved Informer model to learn the long-term nonlinear characteristics of local positions in different partitions, output the time-series settlement prediction value of each settlement point, and realize the long-term prediction of the settlement information of the mining area.

9. A non-transitory computer-readable storage medium, characterized in that: The non-transitory computer-readable storage medium is used to store computer instructions. When the computer instructions are executed by the processor, a long-time series mining area surface subsidence prediction method as described in any one of claims 1-7 is implemented.

10. An electronic device, characterized in that: include: A processor, a memory and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory so that the electronic device executes a long-time series mining area surface subsidence prediction method as described in any one of claims 1-7.

Citation Information

Patent Citations

  • Method for predicting ground surface settlement based on Informer model

    CN117910338A

  • Mining area mining subsidence prediction method, readable storage medium and computer equipment

    CN117910820A

  • Method for predicting surrounding ground surface settlement caused by foundation pit precipitation based on neural network

    CN118350106A

  • Landslide deformation prediction method and device, medium and product

    CN118746808A

  • Prediction device, prediction method, and prediction program

    JP7489502B1

Cited By

  • Urban land subsidence prediction method based on improved Transform model

    CN120724159A