Exponential function model-based mining subsidence residual deformation prediction method

Abstract

The invention discloses a mining subsidence residual deformation prediction method based on an exponential function model, a prediction model has real-time updating and correcting capabilities, internal parameters of the prediction model can be continuously adjusted according to latest acquired monitoring data, a simulated deformation curve is ensured to be closely attached to the ground deformation condition observed on the spot, and the mining subsidence residual deformation prediction method based on the exponential function model is realized. The accuracy and sensitivity of prediction are remarkably improved, and the problem of prediction errors caused by the fact that parameters cannot be adjusted in real time in a prediction method in the prior art is effectively solved. The specific time when the residual settlement deformation reaches the set threshold value for the first time can be accurately estimated, and the space-time dual prediction function is achieved.

Description

Technical Field

[0001] The invention relates to the technical field of ground subsidence residual deformation prediction, and is particularly suitable for a mining subsidence residual deformation prediction method based on an exponential function model. Background Art

[0002] The surface subsidence deformation caused by underground mining can be divided into two stages: the initial severe deformation period and the subsequent residual settlement deformation period. The subsequent residual settlement deformation of the surface restricts the management and sustainable development of land resources and is crucial to land reclamation, ecological restoration and land redevelopment. Therefore, it is necessary to accurately predict the residual settlement deformation.

[0003] At present, the common methods used to predict the residual deformation of mining subsidence include probability integral method, limit equilibrium method, finite element method, etc. However, these methods require three-dimensional modeling of the overlying rock and soil and the mining working surface. In the simplified mechanical model, the geological structure, rock and soil characteristics, mining technology and other factors are assumed, and the residual deformation of mining subsidence is calculated based on parameters such as soil compression coefficient, elastic modulus, Poisson's ratio and groundwater level change.

[0004] However, in actual engineering, the nonlinearity and anisotropy of geological conditions are very complex. At the same time, due to objective factors such as measurement level and site, some data in the above methods cannot be accurately obtained or even cannot be measured, which directly leads to the unreliability of the prediction results.

[0005] Moreover, the factors that cause surface deformation are dynamic and variable. When the surrounding conditions change, the original models, calculation parameters, assumptions, etc. need to be adjusted, resulting in the inability to adjust the existing methods for predicting residual deformation of mining subsidence in a timely manner and make rapid predictions for new situations. Summary of the invention

[0006] The present invention aims to provide a method for predicting residual deformation of mining subsidence based on an exponential function model, so as to solve the problems of complex model, multiple required parameters, low prediction accuracy and poor real-time performance in the current prediction of residual deformation of surface mining subsidence.

[0007] To achieve the above object, the present invention adopts the following technical solutions: The mining subsidence residual deformation prediction method based on the exponential function model of the present invention is based on formula (1), based on the real-time observation data of mining subsidence residual deformation, nonlinear iterative optimization is used to perform data fitting to determine the parameters of formula (1); t in formula (1) is made to tend to infinity, and the final residual settlement deformation of the surface is obtained; Formula (1), where t is the cumulative observation time, δ is the cumulative settlement deformation; a, b, c, d are parameters.

[0008] Furthermore, the real-time observation data should include more than 12 valid observation data for three consecutive months, and be updated in real time with the observation time.

[0009] Furthermore, the units of the accumulated observation time include hours, days, weeks, ten days, and months according to needs.

[0010] Furthermore, the cumulative observation time and cumulative settlement deformation curve is obtained according to formula (1), and the cumulative settlement deformation corresponding to any cumulative observation time can be determined; combined with the exact date of the first real-time observation data involved in the fitting of formula (1), the settlement deformation and daily settlement on any date can be determined.

[0011] Furthermore, according to formula (1), the acceleration change curve of the residual deformation of mining subsidence can also be determined to intuitively reflect the danger of subsidence.

[0012] The advantage of the present invention is that the prediction model has the ability to update and correct in real time, and can continuously adjust its internal parameters according to the latest collected monitoring data, ensuring that the simulated deformation curve closely fits the ground deformation conditions observed in the field, significantly improving the accuracy and sensitivity of the prediction, and effectively overcoming the prediction error problem caused by the inability to adjust the parameters in the prediction method in the prior art in real time. The present invention can also accurately estimate the specific time when the residual settlement deformation first reaches the set threshold conditions, realizing the dual prediction function of time and space. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 The present invention is a specific implementation flow chart of the method.

[0014] Figure 2 The figure is a schematic diagram of an application program interface based on the method of the present invention.

[0015] Figure 3 This is a schematic diagram of the monitoring point data file format in Example 3 of the present invention.

[0016] Figure 4 Schematic diagram of calculating the final residual settlement deformation in Example 3 of the present invention.

[0017] Figure 5 The figure shows an output file of an application program based on the method of the present invention.

[0018] Figure 6 The figure is a schematic diagram of an output file output by an application program based on the method of the present invention.

[0019] Figure 7 This is a graph of cumulative observation time and cumulative settlement deformation outputted by Example 3 of the present invention.

[0020] Figure 8This is a schematic diagram of the outputdata file output by Example 3 of the present invention.

[0021] Fig. 9 This is the daily precipitation curve outputted by Example 3 of the present invention.

[0022] Fig.10 This is the sedimentation acceleration-time relationship curve outputted by Example 3 of the present invention.

[0023] Fig.11 This is a schematic diagram of calculating the set settlement threshold value according to Example 3 of the present invention.

[0024] Fig.12 This is a schematic diagram of an output file that calculates a daily settling rate threshold according to Example 3 of the present invention.

[0025] Fig.13 This is a schematic diagram of calculating the daily settling rate threshold that meets the set value according to Example 3 of the present invention.

[0026] Fig.14 This is a schematic diagram of an output file that calculates a daily settling rate threshold according to Example 3 of the present invention. DETAILED DESCRIPTION

[0027] The technical solutions in the embodiments of the present invention are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0028] Example 1 The mining subsidence residual deformation prediction method based on the exponential function model described in the present invention constructs a nonlinear exponential function that is highly consistent with the actual observation data according to the law of surface residual settlement deformation monitoring data, as shown in formula (1).

[0029] Formula (1) Among them, t is the cumulative observation time, δ is the cumulative settlement deformation; a, b, c, d are parameters.

[0030] The present invention combines the Levenberg-Marquardt (LM) nonlinear iterative optimization algorithm, performs analytical fitting based on the real-time observation data of mining subsidence residual deformation at the monitoring point, determines the parameter values ​​of a, b, c, and d in formula (1), and establishes a mining subsidence residual deformation prediction model. Generally, due to the low monitoring frequency in the residual settlement deformation stage, in order to ensure the representativeness and effectiveness of the data, the rationality of the fitting calculation and the accuracy of the model prediction, the collected monitoring data should include at least three consecutive months of monitoring data and more than 12 valid observation data.

[0031] The collected real-time observation data include the cumulative monitoring time t and the cumulative surface settlement deformation δ corresponding to the cumulative monitoring time t. The units of the cumulative observation time include hours, days, weeks, ten days, and months according to the needs. The unit of the cumulative surface settlement deformation δ is generally mm (millimeter). At the same time, the collected real-time observation data can participate in the analytical fitting of the nonlinear iterative optimization algorithm in real time to further correct the parameter values ​​of a, b, c, and d in formula (1), so that high-precision prediction of the settlement can be completed in a short time, and the influence of the current environment and settlement conditions on the subsequent persistent settlement can be reflected in real time, which solves the problem that the existing technology needs to construct a three-dimensional model with multiple parameters in predicting the residual settlement, and cannot be adjusted in time when facing new situations and new data, resulting in prediction errors and delays.

[0032] For the determined formula (1), when t tends to infinity, the convergence value of formula (1) is the final residual settlement deformation of the surface. In essence, formula (1) reflects the cumulative observation time and cumulative settlement deformation curve. According to this curve, the cumulative settlement deformation corresponding to any cumulative observation time can be determined. Combined with the exact date of the first real-time observation data analytically fitted by the residual formula (1), the settlement deformation and daily settlement on any date can be determined.

[0033] At the same time, based on the daily settlement amount, the acceleration change curve of the residual deformation of mining subsidence can be further calculated to intuitively reflect the rate of settlement and determine the danger brought by settlement.

[0034] Example 2 like Figure 1 As shown, the specific implementation steps of the mining subsidence residual deformation prediction method based on the exponential function model of the present invention are as follows: Step 1: Monitoring data processing. Taking a certain monitoring point as a unit, separate the cumulative monitoring time and the residual settlement deformation of the surface shape with commas, ensure that the format is clear and there are no spaces, and then enter it into a txt format file. The file name should follow the "monitoring point number. start date.txt". The data should contain at least three consecutive months of monitoring data and 12 valid observation points to ensure the representativeness and validity of the data.

[0035] Step 2: Data item usage. In the process of model fitting calculation, accurate and high-quality monitoring data are selected to ensure the prediction accuracy of the model. According to the data characteristics, monitoring data within a specific time range are selected to participate in model training. At the same time, recent data are reserved as a validation set. The validation set data does not directly participate in the calculation of model parameters, but is used to evaluate the prediction performance of the model. By applying the model to the reserved data, the settlement in the corresponding period is predicted and compared with the actual observation data, the prediction accuracy and reliability of the model can be verified.

[0036] Step 3: Construct a nonlinear exponential function that is highly consistent with the actual observed data , depicting the relationship curve between surface residual settlement deformation and time. The monitoring data determined in step 2 are iteratively optimized using Levenberg-Marquardt (LM) nonlinear iterative optimization fitting, so that the monitoring data and the exponential function gradually approach the best fitting state, and finally the parameter values ​​of the various parameters required by the exponential function model are determined, laying a solid foundation for subsequent prediction analysis.

[0037] Step 4: Substitute the parameter values ​​of each parameter in the exponential function model determined in step 3 into the exponential function model to obtain the curve of cumulative observation time and cumulative settlement deformation of the monitoring point. According to the curve, when time t tends to infinity, the convergence value of the exponential function is the final cumulative settlement. The residual settlement deformation can be obtained by subtracting the current cumulative settlement from the final cumulative settlement.

[0038] According to this curve, a rate curve that directly shows the change of daily deformation of the monitoring point with time t can also be drawn. In order to accurately grasp the time when the settlement of the monitoring point reaches a certain threshold, the starting monitoring date (that is, the first data involved in data fitting) is used as the starting node. According to the exact date corresponding to the starting node, the exact date when a certain settlement threshold is reached can be known. Or the exact date when the daily settlement reaches a certain threshold, or the exact date when the daily settlement rate reaches a certain threshold. According to the prediction, it can be judged when the settlement of the monitoring point meets the requirements of engineering technical specifications, which is convenient for calculating the exact date of suitable construction conditions. It is of great significance for evaluating the stability of surface settlement, planning engineering projects, and ensuring construction safety and quality. It is an indispensable tool in engineering decision-making and helps to achieve environmental protection and efficient use of resources.

[0039] Example 3 Taking the measured data of the ground subsidence monitoring point D01 of a coal mine as an example, the implementation steps of the present invention are demonstrated to carry out accurate prediction of the surface residual settlement deformation.

[0040] In order to improve user experience and ease of operation, in this embodiment, an application is developed according to the mining subsidence residual deformation prediction method based on the exponential function model of the present invention, and its interface is as follows: Figure 2 shown.

[0041] The specific operations are as follows: Step 1: For the D01 monitoring point data, Figure 3 The required monitoring data is organized in txt format by default. The basic format is as follows: Cumulative monitoring time, cumulative settlement displacement Cumulative monitoring time, cumulative settlement displacement Cumulative monitoring time, cumulative settlement displacement .... .... .... The D01 monitoring point has now entered the residual settlement and deformation stage. A total of 13 sets of monitoring data were collected from March 1, 2020 to June 25, 2021. The monitoring data were sorted and the data file was saved as a file named "D01.20200301.txt".

[0042] Step 2: Browse the data file to find the "D01.20200301.txt" file and click to open it. Set the monitoring time corresponding to the first collected monitoring data as the starting date, that is, enter "20200301" into the window; select "Use all data", that is, use all the data saved in the data file as the basis for nonlinear iterative optimization analytical fitting. First select "Do not set threshold" in the threshold setting module, and select "Yes" in the module to save the calculation results; determine the fitting formula in the model selection module ; Finally, click to start the calculation. Figure 4 After the calculation is completed, the calculation results will be displayed in the "Final residual settlement deformation" window corresponding to "No threshold set" in the model interface. Figure 4 It can be seen that the final cumulative residual settlement deformation of the D01 monitoring point is 109.7cm. At the same time, after the calculation is completed, a folder "output+generation time" will be generated in the directory where the D01.20200301.txt file is located. It contains three graphs of displacement (cumulative observation time and cumulative settlement deformation curve), velocity (daily settlement), acceleration (daily settlement acceleration) and two files, output.txt and outputdata.txt, such as Figure 5 Open the output.txt file to see the detailed parameters of the fitting, as shown in Figure 6 As shown, including , , , The values ​​of the four parameters, the residuals of the fitted function model, the average residuals, the correlation coefficients and the asymptotes. , , , Substituting the values ​​of the four parameters into the original fitting formula, we get the prediction model, and then we can draw the curve of the cumulative observation time and the cumulative settlement deformation during the whole process from the initial monitoring period to the settlement stabilization of the monitoring point, as shown in Figure 7 The correlation coefficient R = 0.993 is the degree of fit consistency between the prediction model and the measured data.

[0043] In the curve of cumulative observation time and cumulative settlement deformation of the monitoring point from the initial monitoring period to the settlement stabilization, except for the data involved in the nonlinear iterative optimization analytical fitting in the early stage, the later data of the curve are predicted values. The asymptote Line = 109.7 obtained by the asymptotic trend of the curve is the final cumulative settlement value of the monitoring point.

[0044] The outputdata.txt file contains the cumulative settlement displacement, daily settlement rate and settlement acceleration of the historical data observation node of the D01 monitoring point, as shown in Figure 8 As shown. Figure 8 The last measured data of the monitoring point is 106.6cm, which corresponds to June 24, 2021. Therefore, there is a residual settlement deformation of 109.7-106.6=3.1cm from June 24, 2021 until the final settlement is stable. The daily settlement curve in the fitting result file is shown in Fig. 9 The relationship curve between the sedimentation acceleration and time is shown in Fig.10 As shown, other settlement and deformation characteristics of the monitoring point can be intuitively displayed.

[0045] Step 3: After the final cumulative settlement deformation of the D01 monitoring point is obtained through the above operations, select "Set settlement 109.7 as threshold" in the threshold setting module and click to start the calculation. After the calculation is completed, the calculation result will be displayed in the "Date that meets the set threshold condition" window corresponding to "Set settlement threshold" in the model interface, such as Fig.11 shown.

[0046] The basic principle is: Substitute the set cumulative settlement 109.7 into The time t corresponding to the completion of the cumulative settlement can be obtained. Based on the initial date 20200301, it can be calculated that the date that meets the set threshold condition is April 24, 2024. Click View Fitting Calculation Results in the lower right corner of the model interface. When setting the settlement threshold model fitting results (output.txt file) as shown Fig.12 The corresponding results can also be viewed in the last row shown.

[0047] Step 4: In the field of engineering practice, the residual settlement and deformation of the ground surface is a gradual and long process. Under normal circumstances, once the daily settlement rate drops below the preset safety threshold, even if the settlement has not reached a completely stable state at this time, it will not affect the construction of the project. This is because after appropriate foundation treatment measures, even if there is continued trace settlement, the safety and stability of the engineering structure can still be guaranteed. This method balances the progress of the project with safety requirements and avoids unnecessary delays that may be caused by waiting for absolute stability. Therefore, the present invention also provides a daily settlement rate prediction function. In the threshold setting module, select "Set daily settlement rate 0.01 as threshold" and click to start calculation. After the calculation is completed, the calculation results will be displayed in the "Date that meets the set threshold conditions" window corresponding to the "Set daily settlement rate threshold" in the model interface. Fig.13 shown.

[0048] The basic principle is to substitute the set daily sedimentation rate of 0.01 into the rate equation (The first derivative of the settlement displacement-time function) can be used to obtain the time t corresponding to the settlement rate for the day. Based on the initial date 20200301, it can be calculated that the date that meets the set threshold condition is November 8, 2021. Click View fitting calculation results in the lower right corner of the model interface. When setting the daily settlement rate threshold model fitting results (output.txt file) as shown below Fig.14 The corresponding results can also be viewed in the last row shown.

[0049] In summary, the present invention proposes a residual subsidence deformation prediction method based on an exponential function model, which has a real-time data response mechanism and unique spatiotemporal prediction capability, can simply, quickly and efficiently calculate the surface residual subsidence deformation, greatly improves the surface deformation management efficiency and decision-making accuracy in underground mining activities, and provides strong technical support for engineering builders.

[0050] The present invention is based on a large amount of rich and reliable measured data. After in-depth analysis and refined processing of mining subsidence field monitoring data, a mathematical model of surface residual settlement deformation characteristics is successfully extracted. The model has a simple architecture, efficient operation, and excellent performance, providing a powerful tool for the prediction of residual deformation of underground mining subsidence.

[0051] The present invention updates the daily monitoring data in real time, accurately adjusts the fitting parameters, and instantly presents the latest deformation trend curve, ensuring accurate grasp of the deformation dynamics and achieving efficient prediction and immediate warning.

[0052] The present invention can not only accurately calculate the residual deformation of the ground surface, but also innovatively predict the specific time when the project sets the safety threshold. This forward-looking prediction is of great significance to key links such as project design, construction arrangement and maintenance planning, enabling project managers to formulate strategies in advance to ensure the safety and efficiency of the project, providing strong support for the smooth progress of the project.

[0053] The present invention is fast, efficient, and easy to operate. It can process accurate settlement data acquired in real time and use measured data to complete automatic correction, cleverly avoiding the need for accurate parameters required for complex theoretical calculations of traditional models. It can quickly generate accurate prediction results when running.

Claims

1. A mining subsidence residual deformation prediction method based on an exponential function model, characterized in that: According to formula (1), based on the real-time observation data of residual deformation of mining subsidence, nonlinear iterative optimization is used to fit the data and determine the parameters of formula (1); Let t in formula (1) tend to infinity, and obtain the final residual settlement deformation of the surface; Formula (1), where t is the cumulative observation time, δ is the cumulative settlement deformation; a, b, c, d are parameters.

2. The mining subsidence residual deformation prediction method based on the exponential function model according to claim 1 is characterized by: The real-time observation data should include more than 12 valid observation data for three consecutive months, and be updated in real time with the observation time.

3. The mining subsidence residual deformation prediction method based on the exponential function model according to claim 1 is characterized by: The units of the cumulative observation time include hours, days, weeks, ten days, and months according to needs.

4. The mining subsidence residual deformation prediction method based on the exponential function model according to claim 1 is characterized by: According to formula (1), the cumulative observation time and cumulative settlement deformation curve can be obtained, and the cumulative settlement deformation corresponding to any cumulative observation time can be determined; By combining the exact date of the first real-time observation data involved in fitting formula (1), the settlement deformation and daily settlement on any date can be determined.

5. The mining subsidence residual deformation prediction method based on the exponential function model according to claim 1 is characterized by: According to formula (1), the acceleration change curve of the residual deformation of mining subsidence can also be determined to intuitively reflect the danger of subsidence.

Images