Mining ground subsidence dynamic prediction method based on geological trigger rebalancing mechanism

Abstract

The present application belongs to the technical field of mining engineering, and provides a mining ground subsidence dynamic prediction method based on a geological trigger rebalancing mechanism, comprising: obtaining multi-dimensional characteristic parameters and real-time subsidence monitoring data of mining ground subsidence; according to the obtained multi-dimensional characteristic parameters and real-time subsidence monitoring data, inverting real-time sensitivity of the characteristic parameters; based on the obtained real-time sensitivity, adaptively allocating parameter contribution degree, embedding the geological trigger rebalancing mechanism; combining the geological trigger rebalancing mechanism and the real-time subsidence monitoring data, automatically identifying a mining ground subsidence stage, calculating a mining ground subsidence dynamic prediction value, and completing dynamic prediction of the mining ground subsidence.

Description

Technical Field

[0001] This invention belongs to the field of mining engineering technology, specifically relating to a dynamic prediction method for mining ground subsidence based on a geologically triggered rebalancing mechanism. Background Technology

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

[0003] Large-scale coal mining activities trigger stress redistribution in the overlying strata, leading to continuous surface subsidence and deformation, which in turn poses a serious threat to surface infrastructure and ecological stability. In actual mining projects, the surface subsidence process exhibits significant stage characteristics, successively going through three stages: elastic deformation, accelerated collapse, and residual adjustment. The dominant influencing factors, subsidence rates, and deformation mechanisms differ significantly in each stage, which places stringent requirements on the dynamic adaptability of prediction methods.

[0004] Existing technologies can predict settlement by integrating the effects of soil arching and pipe curtain deflection, but they do not consider the multi-stage dynamic characteristics of mining settlement. Distributed fiber optic sensing technology can be used to identify overburden settlement, which can monitor deformation inside the overburden, but it lacks the ability to dynamically adjust parameters throughout the entire stage of surface settlement. It cannot effectively solve the prediction accuracy bottleneck caused by spatiotemporal heterogeneity and is difficult to meet the actual needs of safe mining and ecological protection in mining areas.

[0005] For existing mining ground subsidence prediction technologies, traditional empirical models are suitable for trend prediction due to their ease of calculation. However, because they ignore the coupling mechanism of multiple factors, they are difficult to accurately capture short- and medium-term subsidence patterns and cannot cope with nonlinear deformation problems under complex geological conditions. Numerical simulation methods are based on physical and mechanical mechanisms and can simulate the deformation characteristics of heterogeneous strata, but they suffer from the drawbacks of difficult parameter acquisition and high computational cost, and are not adaptable to extreme conditions such as water inrush and sand collapse. Remote sensing and InSAR technologies rely on satellite imagery to achieve large-scale, millimeter-level monitoring and can trace back historical deformation data, but they are susceptible to interference from atmospheric delay, missing imagery, and other factors, limiting their performance in real-time monitoring and small sample scenarios. Summary of the Invention

[0006] To address the aforementioned issues, this invention proposes a dynamic prediction method for mining ground subsidence based on a geologically triggered rebalancing mechanism. By effectively capturing the different stages of subsidence—elastic deformation, accelerated collapse, and residual adjustment—this method solves the spatiotemporal heterogeneity problem in dynamic prediction of ground subsidence, achieving high-precision dynamic prediction of all stages of mining ground subsidence.

[0007] According to some embodiments, the present invention provides a method for dynamic prediction of mining surface subsidence based on a geologically triggered rebalancing mechanism, employing the following technical solution:

[0008] A dynamic prediction method for mining surface subsidence based on a geologically triggered rebalancing mechanism includes:

[0009] Acquire multi-dimensional characteristic parameters of mining ground subsidence and real-time subsidence monitoring data;

[0010] Based on the acquired multi-dimensional characteristic parameters and real-time settlement monitoring data, the real-time sensitivity of the characteristic parameters is inverted;

[0011] Based on the obtained real-time sensitivity adaptive allocation parameter contribution, a geological triggering rebalancing mechanism is embedded.

[0012] By combining geologically triggered rebalancing mechanisms and real-time settlement monitoring data, the system automatically identifies the settlement stage of the mining surface, calculates the dynamic prediction value of the mining surface settlement, and completes the dynamic prediction of the mining surface settlement.

[0013] As a further technical limitation, the multi-dimensional characteristic parameters of mining surface subsidence obtained include at least geological parameters and probability integral parameters; wherein, the geological parameters include overlying lithology, coal seam burial depth, mining thickness, coal seam dip angle, loose layer thickness, mining degree, and repeated mining indicators.

[0014] Furthermore, the probability integral parameters include the subsidence coefficient, horizontal movement coefficient, inflection point deflection coefficient, inflection point offset distance, and mining influence propagation angle.

[0015] Furthermore, the identified mining surface subsidence stages include the elastic deformation stage, the accelerated collapse stage, and the residual adjustment stage; the highly sensitive parameters for the elastic deformation stage are the coal seam depth, overlying lithology, and mining thickness; the highly sensitive parameters for the accelerated collapse stage are the mining thickness, subsidence coefficient, and inflection point offset distance; and the highly sensitive parameters for the residual adjustment stage are the subsidence coefficient, inflection point deflection coefficient, and repeated mining indicator.

[0016] Furthermore, after obtaining the multi-dimensional characteristic parameters of the mining ground settlement, the obtained characteristic parameters are preprocessed. The preprocessing includes using the Raida criterion to remove outliers in the monitoring data; using linear interpolation to supplement missing data; and normalizing all characteristic parameters to the [0, 1] interval.

[0017] As a further technical limitation, in the process of acquiring real-time settlement monitoring data, based on the acquired multi-dimensional characteristic parameters of mining ground settlement and the XGBoost framework, the monitoring data of the mining ground settlement monitoring points in the mining area are obtained, that is, the real-time settlement monitoring data is obtained.

[0018] As a further technical limitation, the dynamic correlation coefficient between the characteristic parameters and the settlement change is calculated based on the characteristic parameters and the settlement change. The obtained dynamic correlation coefficient is used as a correction term for real-time sensitivity; , For the first Parameters of each monitoring point Values, For the first The amount of settlement change at each monitoring point within the same time period. and These are the sample means.

[0019] Furthermore, the real-time sensitivity of the feature parameters for ;in, These are the observations from the previous month. t For time.

[0020] Furthermore, the parameter contribution is adaptively assigned based on the obtained real-time sensitivity, i.e., parameter contribution. for ;in, This is the learning rate.

[0021] As a further technical limitation, when the settlement change exceeds the settlement change rate threshold, a settlement transition occurs, and a weight rebalancing operation is performed. This involves ranking the sensitivity of parameters at the current moment and resetting the weights of the top N most sensitive parameters. The weights of the low-sensitivity parameters in the bottom M positions are reset to... ;in, and These are the mean and standard deviation, calculated in real time based on the weights of all parameters at the current moment, respectively. M and N are both preset positive integers.

[0022] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0023] This invention employs a geologically triggered rebalancing mechanism to dynamically adapt to the parameter characteristic changes in the three stages of subsidence, effectively solving the problem of spatiotemporal heterogeneity in the prediction of ground subsidence caused by mining. After processing, the input data changes from static to dynamic evolution data, extending the prediction range to the subsidence value at any point in time.

[0024] This invention integrates a multi-dimensional system of geological and mining parameters with probabilistic integral method parameters, comprehensively covering key factors affecting settlement. By combining intelligent algorithms and geological mechanisms, it improves the interpretability and engineering applicability of the model. Attached Figure Description

[0025] The accompanying drawings, which form part of this embodiment, are used to provide a further understanding of this embodiment. The illustrative embodiments and their descriptions are used to explain this embodiment and do not constitute an improper limitation of this embodiment.

[0026] Figure 1 This is a flowchart of the dynamic prediction method for mining surface subsidence based on a geologically triggered rebalancing mechanism in an embodiment of the present invention.

[0027] Figure 2(a) is a schematic diagram of the monitoring results of monitoring line I in the study area in an embodiment of the present invention;

[0028] Figure 2(b) is a schematic diagram of the monitoring results of monitoring line K in the study area in an embodiment of the present invention;

[0029] Figure 3(a) is a distribution diagram of the parameter sensitivity and importance ratio in the elastic deformation stage in the embodiment of the present invention;

[0030] Figure 3(b) is a distribution diagram of the parameter sensitivity and importance ratio in the accelerated collapse stage in the embodiment of the present invention;

[0031] Figure 3(c) is a distribution diagram of the percentage of parameter sensitivity and importance in the residual adjustment stage in the embodiment of the present invention;

[0032] Figure 4 This is a schematic diagram comparing data results in an embodiment of the present invention. Detailed Implementation

[0033] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0034] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0035] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0036] In this invention, terms such as "upper," "lower," "left," "right," "front," "back," "vertical," "horizontal," "side," and "bottom" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. These terms are used only to facilitate the description of the structural relationships of the various components or elements of this invention and do not specifically refer to any component or element in this invention. They should not be construed as limiting the invention.

[0037] In this invention, terms such as "fixed connection," "connected," and "linked" should be interpreted broadly, indicating a fixed connection, an integral connection, or a detachable connection; a direct connection or an indirect connection through an intermediate medium. Those skilled in the art can determine the specific meaning of these terms in this invention based on the specific circumstances, and they should not be construed as limitations on the invention.

[0038] Example

[0039] Embodiment 1 of this invention introduces a dynamic prediction method for mining surface subsidence based on a geologically triggered rebalancing mechanism.

[0040] like Figure 1 The method for dynamic prediction of mining surface subsidence based on a geologically triggered rebalancing mechanism, as shown, includes:

[0041] Acquire multi-dimensional characteristic parameters of mining ground subsidence and real-time subsidence monitoring data;

[0042] Based on the acquired multi-dimensional characteristic parameters and real-time settlement monitoring data, the real-time sensitivity of the characteristic parameters is inverted;

[0043] Based on the obtained real-time sensitivity adaptive allocation parameter contribution, a geological triggering rebalancing mechanism is embedded.

[0044] By combining geologically triggered rebalancing mechanisms and real-time settlement monitoring data, the system automatically identifies the settlement stage of the mining surface, calculates the dynamic prediction value of the mining surface settlement, and completes the dynamic prediction of the mining surface settlement.

[0045] As one or more implementation methods, this embodiment selects key feature parameters of geology and probability integral method as model inputs when constructing a multi-dimensional feature parameter system.

[0046] This embodiment collects data through field exploration, laboratory testing, and mine production records: Based on rock compressive strength grading, it determines the overburden lithology f, measured coal seam depth H, mining thickness M, coal seam dip angle α, loose layer thickness L, mining degree n, and repeated mining marker Re; it calibrates relevant parameters such as the subsidence coefficient on-site using the probability integral method; monitoring points and lines are set up at the surface of the working face, with a control spacing of 30m between monitoring points; a GPS monitoring system is used to collect surface vertical subsidence data monthly for 12 consecutive months, forming a training and validation dataset.

[0047] It should be noted that the geological parameters include overlying lithology, coal seam depth, mining thickness, coal seam dip angle, loose layer thickness, mining degree, and repeated mining indicators; the probability integral method parameters include subsidence coefficient, horizontal movement coefficient, inflection point deflection coefficient, inflection point offset distance, and mining influence propagation angle, specifically:

[0048] a) The subsidence coefficient q has a non-linear relationship with the elastic modulus of the overlying strata, and the calculation formula is as follows: ;in, E The elastic modulus of the overlying strata. γ is the average elastic modulus of the multi-mining area, and γ is the average unit weight of the overburden. H For coal seam burial depth, M Coal seam thickness;

[0049] b) The horizontal movement coefficient b is dynamically controlled by the ratio of loose layer thickness to mining depth, and the calculation formula is as follows: ;in, L For the thickness of the loose layer, H The depth of the coal seam;

[0050] c) Inflection point deflection coefficient tan β The calculation formula is influenced by the coupling effect of coal seam dip angle and mining depth: Where D is the lithological influence coefficient, H Where α is the coal seam depth and α is the coal seam dip angle;

[0051] d) Inflection point offset distance s It is directly related to the geometry of the longwall face, and the calculation formula is: ;in, l 0 represents the actual size of the mining section. H The depth of the coal seam;

[0052] e) Angle of spread of mining influence θ It is significantly correlated with the rock mass integrity coefficient, and the calculation formula is as follows: ;in, k The rock hardness coefficient, α The angle of dip of the coal seam.

[0053] As one or more implementation methods, this embodiment selects the XGBoost algorithm for parameter configuration and model training; that is, it uses the Hyperopt tool to carry out automated parameter tuning of XGBoost, sets the parameter range and optimization target, automatically tries different parameter combinations and evaluates the effect, and finally outputs the optimal configuration; it uses the Laida criterion to remove abnormal data of monitoring data points, supplements missing data by linear interpolation, normalizes all feature parameters to the [0,1] interval, divides them into training set and test set according to a 7:3 ratio, and iterates training until the convergence accuracy meets the requirements.

[0054] This embodiment embeds a dynamic weighting layer on top of XGBoost and introduces a geologically triggered rebalancing mechanism; that is, it first integrates 12 input feature parameters and the monthly settlement rate of monitoring points in the study area into the database. y obs (t ),pass y obs ( t Sensitivity of each inversion parameter S i ( t If the settlement suddenly increases in a certain month, and is consistent with the parameters... X i Strong correlation, then S i ( t Increase, weight w i ( t The parameters should be increased accordingly. Therefore, this paper constructs parameters. X i With △ y Correlation coefficient Corr ( X i , △ y ), as S i ( t Corr() is a correction term. X i , △ y Dynamic correlation is achieved through a dataset of multiple monitoring points across the entire mining area to monitor △ y The surge in parameters is attributed to various factors. For example, if monitoring point A experiences a surge in settlement relative to monitoring point B, and at this time, the parameters of the two monitoring points show a significant difference. X i When one parameter changes while others remain largely unchanged, this can be identified by comparing historical data. X i It is the main mutagenic factor.

[0055] Specifically:

[0056] (1) Constructing the dynamic correlation coefficient between parameters and settlement changes As a sensitivity The correction term, namely ;in, For the first Parameters of each monitoring point Values, For the first The amount of settlement change at each monitoring point within the same time period. and These are the sample means;

[0057] (2) Real-time settlement rate data from monitoring points y obs ( t The real-time sensitivity of each parameter is inverted by calculating the settlement change. ,Right now ;in, These are the observations from the previous month. t The time frame is one month.

[0058] (3) Based on sensitivity S i ( t Adaptive allocation of parameter contribution ,Right now ;in, The learning rate is set to 0.2, and the adjustment range is controlled accordingly.

[0059] (4) When the change in settlement |△ y When the settlement change exceeds the threshold, a settlement transition is initiated, and the weights of high-sensitivity and low-sensitivity parameters are reset to handle the abrupt change. In this embodiment, when the settlement change exceeds the settlement rate threshold, a settlement transition occurs, and a weight rebalancing operation is performed. This involves ranking the sensitivity of the parameters at the current moment and resetting the weights of the top N high-sensitivity parameters. The weights of the low-sensitivity parameters in the bottom M positions are reset to... ;in, and These are the mean and standard deviation, calculated in real time based on the weights of all parameters at the current moment, respectively. M and N are both preset positive integers.

[0060] As one or more implementation methods, this embodiment automatically identifies the elastic deformation stage, accelerated collapse stage, and residual adjustment stage based on real-time settlement rate and parameter sensitivity changes, and dynamically adjusts the parameter contribution for different stages; inputs the feature parameters of the area to be predicted into XGBoost, and outputs real-time ground settlement prediction values ​​through the trained model.

[0061] In this embodiment, during the process of identifying the settlement stage, when the time dynamic of the monitoring point |△y| reaches the threshold, that is, when different settlement stages begin to transition, it is identified that the area where monitoring line I is located is in the elastic deformation stage from the 1st to the 3rd month, the accelerated collapse stage from the 4th to the 10th month, and the residual adjustment stage from the 11th to the 12th month. Monitoring line I and monitoring line K are shown in Figure 2(a) and Figure 2(b), respectively.

[0062] As shown in Figures 3(a), 3(b), and 3(c), the sensitivity of parameters retrieved based on settlement data at each stage is analyzed. S i ( tDuring the elastic deformation stage, H, f, and M showed the highest sensitivity, at 0.554, 0.325, and 0.283, respectively. During the accelerated collapse stage, M, q, and s showed the highest sensitivity, at 0.639, 0.120, and 0.118, respectively. During the residual adjustment stage, q, tanβ, and Re showed the highest sensitivity, at 0.481, 0.380, and 0.252, respectively. Based on the sensitivity calculation results, the weights of each parameter were assigned to different settlement stages.

[0063] This embodiment inputs the feature parameters of the test set into the trained prediction model, and combines the parameter weights dynamically adjusted by the geological triggering rebalancing mechanism to output the predicted settlement values ​​for all monitoring points at different time points over 12 months. Comparing the prediction results with other static models shows that the dynamic prediction method for different settlement stages in this embodiment has a high degree of agreement with the field monitoring data. Figure 4 As shown in the diagram, the predicted values ​​for the K-line pattern show varying degrees of underestimation compared to the actual monitored values. This is because the K-line area has not yet reached complete settlement, resulting in overestimated values. This embodiment dynamically adjusts the contribution of each parameter based on different settlement stages to accurately predict settlement at different time points, effectively improving the identification accuracy of the dominant settlement parameters. After numerous repeated experiments, it was found that this method retains the data-driven advantages of machine learning while incorporating the physical logic of geological engineering, significantly improving the accuracy and interpretability of ground settlement prediction, and demonstrating strong scalability and applicability.

[0064] This embodiment employs a geologically triggered rebalancing mechanism to dynamically adapt to the parameter characteristic changes in the three stages of settlement, effectively solving the problem of spatiotemporal heterogeneity in predicting ground settlement caused by mining. After processing, the input data changes from static to dynamically evolving data, extending the prediction range to settlement values ​​at any point in time. By integrating a multi-dimensional system of geological and mining parameters with probability integral method parameters, it comprehensively covers the key factors affecting settlement. Combined with intelligent algorithms and geological mechanisms, it improves the interpretability and engineering applicability of the model.

[0065] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

[0066] The above description is merely a preferred embodiment of this practice and is not intended to limit the scope of this practice. Various modifications and variations can be made to this practice by those skilled in the art. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of this practice should be included within the protection scope of this practice.

Claims

1. A method for dynamic prediction of mining surface subsidence based on a geologically triggered rebalancing mechanism, characterized in that, include: Acquire multi-dimensional characteristic parameters of mining ground subsidence and real-time subsidence monitoring data; Based on the acquired multi-dimensional characteristic parameters and real-time settlement monitoring data, the real-time sensitivity of the characteristic parameters is inverted; Based on the obtained real-time sensitivity adaptive allocation parameter contribution, a geological triggering rebalancing mechanism is embedded. By combining the geological triggering rebalancing mechanism and real-time settlement monitoring data, the mining ground settlement stage is automatically identified, the dynamic prediction value of mining ground settlement is calculated, and the dynamic prediction of mining ground settlement is completed. The embedded geologically triggered rebalancing mechanism is as follows: obtaining multi-dimensional characteristic parameters and monthly settlement rates at monitoring points within the study area. y obs ( t The sensitivity of each parameter was determined by inverting the monthly settlement rate. S i ( t If the monthly settlement rate suddenly increases and is consistent with the parameters X i Strong correlation leads to increased parameter sensitivity and weighting. w i ( t Corresponding improvements; through parameter construction X i With △ y Correlation coefficient Corr ( X i , △ y ), as S i ( t Correction terms; Correlation coefficient Corr( X i , △ y Dynamic correlation is achieved through a dataset of multiple monitoring points across the entire mining area to control the parameter Δ. y The surge was attributed to [the following]. Calculate the dynamic correlation coefficient between the characteristic parameters and the settlement change. The obtained dynamic correlation coefficient is used as a correction term for real-time sensitivity; , For the first Parameters of each monitoring point Values, For the first The amount of settlement change at each monitoring point within the same time period. and These are the sample means; The real-time sensitivity of the feature parameters for ;in, These are the observations from the previous month. t For time.

2. The method for dynamic prediction of mining surface subsidence based on a geologically triggered rebalancing mechanism as described in claim 1, characterized in that, The acquired multi-dimensional characteristic parameters of mining surface subsidence include at least geological parameters and probability integral parameters; wherein, the geological parameters include overlying lithology, coal seam burial depth, mining thickness, coal seam dip angle, loose layer thickness, mining degree, and repeated mining indicators.

3. The method for dynamic prediction of mining surface subsidence based on a geologically triggered rebalancing mechanism as described in claim 2, characterized in that, The probability integral parameters include the subsidence coefficient, horizontal movement coefficient, inflection point deflection coefficient, inflection point offset distance, and mining influence propagation angle.

4. The dynamic prediction method for mining surface subsidence based on a geologically triggered rebalancing mechanism as described in claim 3, characterized in that, The identified mining surface subsidence stages include the elastic deformation stage, the accelerated collapse stage, and the residual adjustment stage; the highly sensitive parameters for the elastic deformation stage are the coal seam depth, overlying lithology, and mining thickness; the highly sensitive parameters for the accelerated collapse stage are the mining thickness, subsidence coefficient, and inflection point offset; and the highly sensitive parameters for the residual adjustment stage are the subsidence coefficient, inflection point deflection coefficient, and repeated mining indicator.

5. The method for dynamic prediction of mining surface subsidence based on a geologically triggered rebalancing mechanism as described in claim 2, characterized in that, After obtaining the multi-dimensional characteristic parameters of mining ground settlement, the obtained characteristic parameters are preprocessed. The preprocessing includes using the Raida criterion to remove outliers in the monitoring data; using linear interpolation to supplement missing data; and normalizing all characteristic parameters to the [0, 1] interval.

6. The method for dynamic prediction of mining surface subsidence based on a geologically triggered rebalancing mechanism as described in claim 1, characterized in that, In the process of acquiring real-time settlement monitoring data, based on the acquired multi-dimensional characteristic parameters of mining ground settlement and the XGBoost framework, the monitoring data of the mining ground settlement monitoring points in the mining area are obtained, that is, the real-time settlement monitoring data is obtained.

7. The method for dynamic prediction of mining surface subsidence based on a geologically triggered rebalancing mechanism as described in claim 1, characterized in that, Based on the obtained real-time sensitivity, the parameter contribution is adaptively assigned. for ;in, This is the learning rate.

8. The method for dynamic prediction of mining surface subsidence based on a geologically triggered rebalancing mechanism as described in claim 1, characterized in that, When the settlement change exceeds the settlement change rate threshold, a settlement transition occurs, and a weight rebalancing operation is performed. This involves ranking the sensitivity of parameters at the current moment and resetting the weights of the top N most sensitive parameters. The weights of the low-sensitivity parameters in the bottom M positions are reset to... ;in, and These are the mean and standard deviation, calculated in real time based on the weights of all parameters at the current moment, respectively. M and N are both preset positive integers.

Images