A method, device, equipment and medium for predicting mine surface subsidence deformation

Abstract

The present application discloses a method, device, equipment and medium for predicting surface settlement and deformation of a mine, and relates to the field of surface settlement and deformation prediction of a mine. The method comprises: obtaining monitoring data of a historical time period of a mining area; calculating relative displacement data of the historical time period based on the monitoring data of the historical time period; processing the relative displacement data of the historical time period using an optimal noise reduction algorithm to obtain noise reduction data of the historical time period; the optimal noise reduction algorithm is an L2 regularization method; inputting the noise reduction data of the historical time period into an optimal prediction model to obtain surface settlement and deformation prediction data for a future time period; the optimal prediction model is a BP neural network prediction model based on particle swarm optimization, and the present application improves the efficiency and accuracy of surface settlement and deformation prediction.

Description

Technical Field

[0001] The present application relates to the field of mine surface subsidence deformation prediction, and in particular to a mine surface subsidence deformation prediction method, device, equipment and medium. Background Art

[0002] The rational development and application of mineral resources are key to promoting social, economic, and scientific and technological progress. With the continuous advancement of science and technology, the types of exploitable mineral resources are increasing. However, with the deepening of resource mining, the area of ​​goaf within the mining area has gradually expanded, which has led to increasingly serious surface subsidence and deformation problems. These subsidence deformations are mainly caused by changes in the stress structure of the rock and soil in the goaf roof, which causes the static pressure of the overlying rock mass to be concentrated on the surrounding areas, thereby causing the rock mass to bend, crack, and even collapse, and ultimately form varying degrees of subsidence on the surface. In-depth research on the evolution of surface subsidence deformation in mines is of great significance for preventing collapse accidents and reducing resource waste. Therefore, how to efficiently and accurately monitor the development characteristics and evolution laws of surface subsidence deformation in mines has become an important issue that needs to be solved urgently. Summary of the Invention

[0003] The purpose of this application is to provide a method, device, equipment and medium for predicting mine surface subsidence deformation, which can improve the accuracy and efficiency of mine surface subsidence deformation prediction.

[0004] To achieve the above objectives, this application provides the following solutions:

[0005] In a first aspect, the present application provides a method for predicting surface subsidence and deformation of a mine, comprising:

[0006] Obtain monitoring data for historical time periods in mining areas;

[0007] Calculating relative displacement data of the historical time period based on the monitoring data of the historical time period;

[0008] The relative displacement data of the historical time period is processed using the best noise reduction algorithm to obtain the noise reduction data of the historical time period; the best noise reduction algorithm is the L2 regularization method;

[0009] The noise reduction data of the historical time period is input into the optimal prediction model to obtain the surface settlement deformation prediction data of the future time period; the optimal prediction model is a BP neural network prediction model based on particle swarm optimization.

[0010] In a second aspect, the present application provides a device for predicting surface subsidence and deformation of a mine, comprising:

[0011] The monitoring data acquisition module is used to obtain monitoring data of the mining area in the historical period;

[0012] A relative displacement data calculation module is used to calculate the relative displacement data of the historical time period based on the monitoring data of the historical time period;

[0013] A noise reduction processing module is used to process the relative displacement data of the historical time period using an optimal noise reduction algorithm to obtain noise-reduced data of the historical time period; the optimal noise reduction algorithm is an L2 regularization method;

[0014] The surface settlement deformation prediction module is used to: input the noise reduction data of the historical time period into the optimal prediction model to obtain the surface settlement deformation prediction data of the future time period; the optimal prediction model is a BP neural network prediction model based on particle swarm optimization.

[0015] In a third aspect, the present application provides a computer device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-mentioned mine surface subsidence deformation prediction method.

[0016] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned method for predicting surface subsidence and deformation of a mine.

[0017] According to the specific embodiments provided in this application, this application discloses the following technical effects:

[0018] The present application provides a method, device, equipment and medium for predicting surface subsidence and deformation of a mine, which obtains monitoring data of a historical time period of a mining area; calculates relative displacement data of a historical time period based on the monitoring data of the historical time period; uses an optimal noise reduction algorithm to process the relative displacement data of the historical time period to obtain noise reduction data of the historical time period; the optimal noise reduction algorithm is an L2 regularization method; the noise reduction data of the historical time period is input into an optimal prediction model to obtain surface subsidence and deformation prediction data for a future time period; the optimal prediction model is a BP neural network prediction model based on particle swarm optimization. The present application improves the efficiency and accuracy of surface subsidence and deformation prediction by using an L2 regularization method to process and reduce the noise of the relative displacement data, and using a BP neural network prediction model based on particle swarm optimization to predict surface subsidence and deformation of the noise reduction data. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0020] Figure 1 This is an application environment diagram of a mine surface subsidence deformation prediction method in one embodiment of the present application;

[0021] Figure 2 A schematic flow chart of a method for predicting surface subsidence and deformation of a mine provided in one embodiment of the present application;

[0022] Figure 3 A schematic flow chart of a method for self-iterative settlement prediction modeling based on Beidou monitoring data for different mining conditions in mining areas provided in one embodiment of the present application;

[0023] Figure 4 A schematic diagram of a specific process of a method for self-iterative settlement prediction model of Beidou monitoring data for different mining conditions in mining areas provided in one embodiment of the present application;

[0024] Figure 5 A schematic diagram of the functional modules of a mine surface subsidence and deformation prediction device provided in one embodiment of the present application;

[0025] Figure 6 A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0026] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0027] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0028] The mine surface settlement deformation prediction method provided in the embodiment of the present application can be applied to Figure 1In 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 set up separately, integrated on the server 104, or placed on the cloud or other servers. The terminal 102 can send the monitoring data of the historical time period of the mining area to the server 104. After the server 104 receives the monitoring data of the historical time period of the mining area, for the monitoring data of the historical time period of the mining area, the server 104 calculates the relative displacement data of the historical time period based on the monitoring data of the historical time period, processes the relative displacement data of the historical time period using the L2 regularization method, obtains the noise reduction data of the historical time period, and inputs the noise reduction data of the historical time period into the optimal prediction model to obtain the surface settlement deformation prediction data of the future time period. The server 104 can feed back the obtained surface settlement deformation prediction data for the future time period of the mining area to the terminal 102. In addition, in some embodiments, the mine surface subsidence and deformation prediction method can also be implemented independently by the server 104 or the terminal 102. For example, the terminal 102 can directly perform surface subsidence and deformation prediction based on the monitoring data of the historical time period of the mining area, or the server 104 can obtain the monitoring data of the historical time period of the mining area from the data storage system and perform surface subsidence and deformation prediction based on the monitoring data of the historical time period of the mining area.

[0029] The terminal 102 may be, but is not limited to, various desktop computers and laptop computers. The server 104 may be implemented as an independent server or a server cluster consisting of multiple servers, or may be a cloud server.

[0030] In an exemplary embodiment, Figure 2 As shown, a method for predicting surface settlement and deformation of a mine is provided. The method is executed by a computer device, specifically a computer device such as a terminal or a server, or a terminal and a server. In the embodiment of the present application, the method is applied to Figure 1 The server 104 in the example is used as an example to illustrate the process, including the following steps 201 to 204.

[0031] Step 201: Obtain monitoring data for a historical period of the mining area.

[0032] Step 202: Calculate the relative displacement data of the historical time period based on the monitoring data of the historical time period.

[0033] Step 203: using an optimal denoising algorithm to process the relative displacement data of the historical time period to obtain denoised data of the historical time period; the optimal denoising algorithm is an L2 regularization method.

[0034] Step 204: inputting the noise reduction data of the historical time period into the optimal prediction model to obtain the surface settlement deformation prediction data of the future time period; the optimal prediction model is a BP neural network prediction model based on particle swarm optimization (PSO).

[0035] By implementing the above steps 201 to 204, the relative displacement data is processed and denoised by using the L2 regularization method, and the surface settlement deformation prediction is performed on the denoised data using the BP neural network prediction model based on particle swarm optimization, thereby improving the efficiency and accuracy of the surface settlement deformation prediction.

[0036] In another exemplary embodiment of the present application, historical monitoring data is obtained using the Beidou Navigation Satellite System (GNSS). Beidou satellite technology demonstrates significant advantages in surface deformation monitoring in mines. First, it offers real-time dynamic monitoring capabilities, providing continuous monitoring data, thereby more accurately depicting the ongoing development of surface deformation and providing detailed data support for predicting future deformation trends. Second, monitoring accuracy is significantly improved. In particular, when the Beidou system is combined with GNSS positioning technology, monitoring accuracy can reach millimeter levels, providing an accurate and reliable data foundation for in-depth analysis of surface deformation in mines. Furthermore, this technology reduces the impact of environmental factors on monitoring results, adapts to all-weather monitoring needs, and minimizes interference from external environmental changes on monitoring accuracy. Furthermore, the monitoring process is more automated, achieving efficient automation across all processes, from data collection, processing, and storage to data analysis, visualization, and early warning alerts. Through these advantages, Beidou satellite technology not only improves the efficiency and accuracy of surface deformation monitoring in mines, but also provides solid data support for timely preventive measures, reducing potential safety risks. This application makes full use of high-precision Beidou monitoring data to simulate the subsidence of the mining area and establish a prediction model.

[0037] Different mining areas are divided into surface mining areas, shallow mining areas, metal mines and non-metal mines.

[0038] like Figure 3 and Figure 4As shown in the figure, GNSS monitoring data of the mining area are collected and processed simply to convert them into relative displacement data; the relative displacement data are processed by the BP neural network prediction model based on particle swarm optimization, the long short-term memory (LSTM) algorithm and the differential integrated moving average autoregressive model (ARIMA), respectively, and a model is selected according to the calculation results to determine the training set and prediction set; the relative displacement data are denoised by the L2 regularization method, the wavelet algorithm and the singular spectrum analysis method to obtain denoised data; the denoised data are processed by the BP neural network prediction algorithm based on particle swarm optimization, the long short-term memory algorithm and the differential integrated moving average autoregressive model according to the training set and prediction set, and a combination is selected as the final combined prediction model according to the calculation results. This application can improve the ability to predict settlement and deformation in mining areas; determine the number of training sets and prediction sets suitable for the model based on the BP neural network prediction algorithm based on particle swarm optimization; use the L2 regularization method, wavelet algorithm and singular spectrum analysis method to reduce the noise of the data respectively, and use the BP neural network prediction algorithm based on particle swarm optimization, long short-term memory algorithm and differential integrated moving average autoregressive model to predict the noise-reduced data respectively, so as to realize multi-method optimization of the data.

[0039] Beidou Navigation Satellite System monitoring data was obtained from the mining area. The data obtained included data from ten stations, each with ten years of data. The first row of data from each station was used as the baseline, and the remaining rows were subtracted from the first row to obtain the relative displacement data required for the experiment. Each row of data from each station represents one day of data, and the data from each station has three columns, one for each north-south, east-west, and vertical direction. Therefore, the relative displacement data is the relative displacement in the north-south, east-west, or vertical direction.

[0040] Based on the Beidou monitoring solution results, where the high-precision solution results are millimeter-level high-precision relative displacements in the north-south, east-west, and vertical directions, a prediction model is selected based on a particle swarm optimized BP neural network prediction algorithm, a long short-term memory algorithm, and a differential integrated moving average autoregressive model to determine the training set and prediction set. In another exemplary embodiment of the present application, the above-mentioned training set and prediction set determination process may include the following steps 301 to 304:

[0041] Step 301: inputting the relative displacement data into the BP neural network prediction model based on particle swarm optimization, the long short-term memory algorithm and the differential integrated moving average autoregressive model respectively to obtain the first prediction data, the second prediction data and the third prediction data;

[0042] Step 302: Calculate the root mean square error and mean absolute error of the label data of the relative displacement data with the first prediction data, the second prediction data, and the third prediction data, respectively, to obtain a first error calculation result, a second error calculation result, and a third error calculation result;

[0043] Step 303: determining the best prediction model according to the first error calculation result, the second error calculation result, and the third error calculation result, specifically including: taking the prediction model corresponding to the error calculation result with the smallest value as the best prediction model.

[0044] Step 304: Determine the best training set and the best prediction set based on the best prediction model; the time of the historical time period is the same as the time of the relative displacement data in the best training set; the time of the future time period is the same as the time of the surface settlement deformation prediction data in the best prediction set.

[0045] The process of determining the best training set and the best prediction set is as follows:

[0046] Take ten years of data from three directions at N stations in the BeiDou Navigation Satellite System monitoring data. One year of relative displacement data in any direction at any station is used as a time series sample input to obtain the first training set. Two years of relative displacement data in any direction at any station is used as a time series sample input to obtain the second training set. Similarly, ten years of relative displacement data in any direction at any station is used as a time series sample input to obtain the tenth training set. The relative displacement data in these training sets correspond to different times.

[0047] Each training set is input into the optimal prediction model to obtain a prediction set corresponding to each training set. The optimal training set and optimal prediction set are determined based on all training sets and the prediction set corresponding to each training set. That is, the relative displacement data in the ultimately determined optimal training set is four or five years long, and the surface settlement deformation prediction data in the optimal prediction set is one year long. Therefore, the input of each prediction model is four or five years of relative displacement data in any direction at any station, and the output is the surface settlement deformation prediction data for the next year in any direction at any station. That is, the historical time period in step 201 can be four or five years, and the future time period can be one year.

[0048] Based on the Beidou monitoring solution results, i.e., relative displacement data, the data are respectively substituted into the L2 regularization method, the wavelet algorithm, and the singular spectrum analysis method to obtain denoised data. Based on the denoised data, the data are respectively substituted into the BP neural network prediction algorithm based on particle swarm optimization, the long short-term memory algorithm, and the differential integrated moving average autoregressive model according to the number of training sets, to form nine combined prediction models. The results obtained according to the number of prediction sets are evaluated using indicators to obtain the final combined prediction model. The process of determining the final combined prediction model may include the following steps 401 to 404:

[0049] Step 401: Processing the relative displacement data using an L2 regularization method, a wavelet algorithm, and a singular spectrum analysis method to obtain first denoised data, second denoised data, and third denoised data;

[0050] Step 402: Inputting the target denoised data into a BP neural network prediction model based on particle swarm optimization, a long short-term memory algorithm, and a differential integrated moving average autoregressive model, respectively, to obtain fourth predicted data, fifth predicted data, and sixth predicted data corresponding to the target denoised data; the target denoised data is the first denoised data, the second denoised data, or the third denoised data;

[0051] Step 403: Calculate the root mean square error and mean absolute error of the label data of the target denoised data with the fourth prediction data, the fifth prediction data, and the sixth prediction data corresponding to the target denoised data, respectively, to obtain the fourth error calculation result, the fifth error calculation result, and the sixth error calculation result corresponding to the target denoised data;

[0052] Step 404: Determine a final combined prediction model based on the fourth error calculation results, the fifth error calculation results, and the sixth error calculation results corresponding to all the target denoised data; the final combined prediction model is obtained by combining the optimal denoising algorithm and the optimal prediction model; the denoising algorithm includes an L2 regularization method, a wavelet algorithm, and a singular spectrum analysis method; the prediction model includes a BP neural network prediction model based on particle swarm optimization, a long short-term memory algorithm, and a differential integrated moving average autoregressive model.

[0053] The prediction model input is relative displacement data equivalent to the optimal training set time, and the prediction model output is data equivalent to the optimal training set time. The root mean square error and mean absolute error are then calculated for the relative displacement data at the training set location.

[0054] The BP neural network prediction algorithm based on particle swarm optimization determines the number of training sets and prediction sets, and then iterates continuously until the maximum number of iterations, i.e., 1000, is reached.

[0055] Based on the long short-term memory algorithm, the input data of the historical time period is provided to the LSTM model to obtain the initial hidden state and cell state. During the prediction process, the input data of the current time period and the hidden state of the previous time period are input to the model each time. The output result of the current time period is calculated by the model, and the output result of the current time period is used as the input of the next time period. This process is repeated until the predicted time range is reached or the stopping condition is met.

[0056] Based on the differential integrated moving average autoregressive model, the Akaike information criterion, the Bayesian information criterion and the maximum value of the model parameters are initialized. Through loop traversal, when the results of the Akaike information criterion and the Bayesian information criterion are both small, the traversal is ended and the model parameters at this time are used as the optimal parameters of the entire model to determine the prediction set results.

[0057] For the above three prediction models, by comparing the root mean square error and mean absolute error, the BP neural network prediction model based on particle swarm optimization is selected to determine the best training set and prediction set, and the training set is four or five years, and the prediction set is one year.

[0058] In another exemplary embodiment of the present application, monitoring data from three sites was used for an experiment. The average root mean square error (RMSE) values ​​of the LSTM for the three sites were calculated to be 7.4611, the average RMSE values ​​of the ARIMA were 2.1677, and the average RMSE values ​​of the BP neural network prediction model based on particle swarm optimization were 2.0239. The average mean absolute error (MAE) values ​​of the LSTM were calculated to be 4.5938, the average MAE values ​​of the ARIMA were 1.7825, and the average MAE values ​​of the BP neural network prediction model based on particle swarm optimization were 1.5803. It can be seen that the BP neural network prediction model based on particle swarm optimization is more effective, so the BP neural network prediction model based on particle swarm optimization is used to determine the training set and prediction set.

[0059] It should be noted that since the trend in the north-south direction is not obvious, it is not suitable for prediction using the BP neural network prediction algorithm based on particle swarm optimization. Therefore, this application only targets the vertical and east-west directions, that is, the relative displacement data used in this application includes relative displacement in the vertical and east-west directions.

[0060] Substitute the relative displacement data into L2 regularization, wavelet algorithm and singular spectrum analysis respectively to obtain the denoised data results;

[0061] The L2 regularization method can reduce the complexity of the model by optimizing the weight coefficients to be very small, thereby reducing the role of a single feature in the model and preventing a certain feature from dominating the entire prediction direction.

[0062] For the wavelet algorithm, the root mean square error and smoothness are normalized, and then the coefficient of variation weighting method is used. The composite evaluation index T is obtained by the linear combination method. When T is minimized, the decomposition level obtained is the optimal decomposition level. The Daubechies wavelet is selected as the wavelet basis, which is often used to decompose and reconstruct signals. Based on the optimal decomposition level, dbN is traversed and the optimal wavelet basis is selected. The 'heursure' threshold criterion and soft threshold processing function are used.

[0063] For the singular spectrum analysis method, the window length needs to be determined. The total length of the time series used in this embodiment is 3650. The window length N should be between (0, 1825) and preferably an integer multiple of the period. This is because sedimentation monitoring will vary in different seasons. After consulting relevant literature and comparing the root mean square error and mean absolute error, the window length N is selected as 365.

[0064] Preferably, by comparing the root mean square error and the mean absolute error, the combination of L2 regularization and the BP neural network prediction algorithm based on particle swarm optimization, and the combination of singular spectrum analysis and differential integrated moving average autoregressive model have better effects. However, because the data after singular spectrum analysis processing is too smooth and the loss of small features of the data is relatively large, the combination of L2 regularization and the BP neural network prediction algorithm based on particle swarm optimization is finally selected as the final combined prediction model.

[0065] The present application also provides an application scenario, which applies the above-mentioned mine surface settlement deformation prediction method. Specifically: the mine surface settlement deformation prediction method provided in this embodiment can be applied in the mine surface settlement deformation prediction scenario. The mine surface settlement deformation prediction scenario includes a data collection link and a surface settlement deformation prediction link; the monitoring data of the mining area enters the surface settlement deformation prediction link from the data collection link, and the corresponding surface settlement deformation prediction data for the future time period is obtained through human-computer collaboration. The mine surface settlement deformation prediction method provided in this embodiment belongs to the surface settlement deformation prediction link. Specifically, in the process of the surface settlement deformation prediction link for the mining area, the relative displacement data of the historical time period can be calculated based on the monitoring data of the historical time period, and the relative displacement data of the historical time period can be processed using the L2 regularization method to obtain the denoised data of the historical time period, and the denoised data of the historical time period is input into the optimal prediction model to obtain the surface settlement deformation prediction data for the future time period.

[0066] Based on the same inventive concept, the embodiments of the present application also provide a mine surface settlement and deformation prediction device for implementing the above-mentioned mine surface settlement and deformation prediction method. The implementation solution provided by this device is similar to the implementation solution described in the above-mentioned method. Therefore, the specific limitations of one or more embodiments of the mine surface settlement and deformation prediction device provided below can be found in the above-mentioned limitations of the mine surface settlement and deformation prediction method, and will not be repeated here.

[0067] In an exemplary embodiment, Figure 5 As shown, a mine surface settlement deformation prediction device is provided, which includes the following modules.

[0068] The monitoring data acquisition module T1 is used to obtain monitoring data of the mining area in the historical time period.

[0069] The relative displacement data calculation module T2 is used to calculate the relative displacement data of the historical time period based on the monitoring data of the historical time period.

[0070] The noise reduction processing module T3 is used to: process the relative displacement data of the historical time period using the best noise reduction algorithm to obtain the noise reduction data of the historical time period; the best noise reduction algorithm is the L2 regularization method.

[0071] The surface settlement deformation prediction module T4 is used to input the noise reduction data of the historical time period into the optimal prediction model to obtain the surface settlement deformation prediction data of the future time period; the optimal prediction model is a BP neural network prediction model based on particle swarm optimization.

[0072] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Figure 6 As shown. The computer device includes a processor, a memory, an input / output interface (I / O) and a communication interface. The processor, memory and input / output interface are connected via a system bus, and the communication interface is connected to the system bus via the input / output interface. 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, a computer program and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store surface subsidence deformation prediction data. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a method for predicting surface subsidence deformation in a mine is implemented.

[0073] Those skilled in the art will understand that Figure 6 The 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 shown in the figure, or combine certain components, or have a different component arrangement.

[0074] In an exemplary embodiment, a computer device is further provided, including a memory and a processor. The memory stores a computer program, and the processor implements the steps in the above method embodiments when executing the computer program.

[0075] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0076] 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, and the collection, use and processing of relevant data must comply with relevant regulations.

[0077] 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 above-mentioned embodiments. In particular, any reference to memory, database, or other media used in the embodiments provided in this 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), magnetic 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. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0078] The databases involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchains. The processors involved in the various embodiments provided herein may include, but are not limited to, general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic units, data processing logic units based on quantum computing, and the like.

[0079] The technical features of the above embodiments can 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.

[0080] This document uses specific examples to illustrate the principles and implementation methods of this application. The description of the above examples is only intended to help understand the method and core concept of this application. At the same time, for those skilled in the art, based on the concept of this application, there may be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A method for predicting surface subsidence and deformation of a mine, characterized in that: The mine surface subsidence deformation prediction method includes: Obtain monitoring data for historical time periods in mining areas; Calculating relative displacement data for the historical time period based on the monitoring data for the historical time period; the relative displacement data is relative displacement in the north-south, east-west or vertical direction; The relative displacement data are processed using an L2 regularization method, a wavelet algorithm, and a singular spectrum analysis method to obtain first denoised data, second denoised data, and third denoised data; inputting the target denoised data into a BP neural network prediction model based on particle swarm optimization, a long short-term memory algorithm, and a differential integrated moving average autoregressive model, respectively, to obtain fourth prediction data, fifth prediction data, and sixth prediction data corresponding to the target denoised data; the target denoised data is the first denoised data, the second denoised data, or the third denoised data; Calculate the root mean square error and mean absolute error of the label data of the target denoised data with the fourth prediction data, the fifth prediction data, and the sixth prediction data corresponding to the target denoised data, respectively, to obtain a fourth error calculation result, a fifth error calculation result, and a sixth error calculation result corresponding to the target denoised data; Determining a final combined prediction model based on the fourth error calculation results, the fifth error calculation results, and the sixth error calculation results corresponding to all the target denoised data; the final combined prediction model is obtained by combining the optimal denoising algorithm and the optimal prediction model; the denoising algorithm includes an L2 regularization method, a wavelet algorithm, and a singular spectrum analysis method; the prediction model includes a BP neural network prediction model based on particle swarm optimization, a long short-term memory algorithm, and a differential integrated moving average autoregressive model; The relative displacement data of the historical time period is processed using the best noise reduction algorithm to obtain the noise reduction data of the historical time period; the best noise reduction algorithm is the L2 regularization method; the best noise reduction algorithm is the L2 regularization method; The noise reduction data of the historical time period is input into the optimal prediction model to obtain the surface settlement deformation prediction data of the future time period; the optimal prediction model is a BP neural network prediction model based on particle swarm optimization; The mine surface subsidence deformation prediction method further includes: The relative displacement data are respectively input into the BP neural network prediction model based on particle swarm optimization, the long short-term memory algorithm and the differential integrated moving average autoregressive model to obtain the first prediction data, the second prediction data and the third prediction data; Calculate the root mean square error and mean absolute error of the label data of the relative displacement data with the first prediction data, the second prediction data, and the third prediction data to obtain a first error calculation result, a second error calculation result, and a third error calculation result; Determining the best prediction model based on the first error calculation result, the second error calculation result, and the third error calculation result; specifically, determining the best prediction model by taking the prediction model corresponding to the error calculation result with the smallest value as the best prediction model; The best training set and the best prediction set are determined according to the best prediction model; the time of the historical time period is the same as the time of the relative displacement data in the best training set; the time of the future time period is the same as the time of the surface settlement deformation prediction data in the best prediction set.

2. The mine surface subsidence deformation prediction method according to claim 1, characterized in that: The monitoring data for the historical time period is obtained by monitoring the Beidou navigation satellite system.

3. The mine surface subsidence deformation prediction method according to claim 1, characterized in that: Mining areas are divided into surface mining areas, shallow mining areas, metal mines and non-metal mines.

4. A mine surface settlement deformation prediction device, characterized in that: The mine surface subsidence deformation prediction device comprises: The monitoring data acquisition module is used to obtain monitoring data of the mining area in the historical period; A relative displacement data calculation module is used to calculate relative displacement data of a historical time period based on the monitoring data of the historical time period; the relative displacement data is relative displacement in the north-south direction, the east-west direction, or the vertical direction; The relative displacement data are processed using an L2 regularization method, a wavelet algorithm, and a singular spectrum analysis method to obtain first denoised data, second denoised data, and third denoised data; inputting the target denoised data into a BP neural network prediction model based on particle swarm optimization, a long short-term memory algorithm, and a differential integrated moving average autoregressive model, respectively, to obtain fourth prediction data, fifth prediction data, and sixth prediction data corresponding to the target denoised data; the target denoised data is the first denoised data, the second denoised data, or the third denoised data; Calculate the root mean square error and mean absolute error of the label data of the target denoised data with the fourth prediction data, the fifth prediction data, and the sixth prediction data corresponding to the target denoised data, respectively, to obtain a fourth error calculation result, a fifth error calculation result, and a sixth error calculation result corresponding to the target denoised data; Determining a final combined prediction model based on the fourth error calculation results, the fifth error calculation results, and the sixth error calculation results corresponding to all the target denoised data; the final combined prediction model is obtained by combining the optimal denoising algorithm and the optimal prediction model; the denoising algorithm includes an L2 regularization method, a wavelet algorithm, and a singular spectrum analysis method; the prediction model includes a BP neural network prediction model based on particle swarm optimization, a long short-term memory algorithm, and a differential integrated moving average autoregressive model; A noise reduction processing module is used to process the relative displacement data of the historical time period using an optimal noise reduction algorithm to obtain noise-reduced data of the historical time period; the optimal noise reduction algorithm is an L2 regularization method; The surface settlement deformation prediction module is used to: input the noise reduction data of the historical time period into the optimal prediction model to obtain the surface settlement deformation prediction data of the future time period; the optimal prediction model is a BP neural network prediction model based on particle swarm optimization.

5. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the mine surface subsidence deformation prediction method according to any one of claims 1 to 3.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for predicting mine surface subsidence and deformation according to any one of claims 1 to 3 is implemented.

Images