Coal mining subsidence land mining while treatment method and system based on multi-source monitoring and dynamic prediction

Abstract

The invention provides a coal mining subsidence land mining while treatment method and system based on multi-source monitoring and dynamic prediction, and belongs to the technical field of coal mining governance, the coal mining subsidence land mining while treatment system based on multi-source monitoring and dynamic prediction comprises a sensing layer, a data layer, an intelligent layer and an application layer; wherein the sensing layer comprises a'satellite-air-ground-well 'integrated monitoring network; wherein the data layer comprises a fusion and processing network; wherein the intelligent layer comprises an analysis and prediction network; wherein the application layer comprises an execution and feedback network. Through multi-source data fusion and a dynamic prediction model, treatment opportunity selection is changed from experience judgment to data-driven scientific decision, and the problem of intervention opportunity is solved; according to the dynamic backfill volume calculation model, filling according to needs and accurate regulation and control are achieved, it is predicted that earthwork waste caused by excessive design or secondary treatment can be reduced, and the engineering cost is remarkably reduced.

Description

Technical Field

[0001] This invention belongs to the field of coal mining management technology, and in particular relates to a method and system for simultaneous mining and management of coal mining subsidence areas based on multi-source monitoring and dynamic prediction. Background Technology

[0002] Currently, in the comprehensive management of coal mining subsidence areas, it is necessary to manage coal mining subsidence areas that have not yet settled after mining. As coal mining continues, the number of unsettled coal mining subsidence areas continues to increase. Therefore, it is very necessary to study the technology of simultaneous mining and management of unsettled subsidence areas.

[0003] The existing methods and systems for simultaneously mining and treating coal mining subsidence areas still have the following problems: 1. Insufficient integration of monitoring technologies hinders the formation of a collaborative perception and unified understanding of the entire subsidence process. Currently, multi-source monitoring methods provide abundant data sources, but these technologies often operate independently, with inconsistent data formats, spatiotemporal benchmarks, and analytical standards. The lack of an effective multi-source heterogeneous data fusion engine makes it difficult to correlate and calibrate massive amounts of data within a unified spatiotemporal framework, preventing the formation of a continuous and three-dimensional perception of the unstable subsidence area across multiple scales: "global-local-point."

[0004] 2. The timing of remediation decisions lacks scientific basis and carries high risks. Currently, the selection of remediation timing relies heavily on experience-based judgment. If intervention is too early, the surface may not be sufficiently deformed, and large-scale earthwork projects may be misjudged by satellite remote sensing as farmland destruction, leading to enforcement risks. If intervention is too late, topsoil resources will be lost, fertility will decline, and the soil may even become submerged and ineffective, resulting in a sharp increase in remediation costs.

[0005] 3. Settlement prediction and treatment design are disconnected, and the backfill volume calculation has a large deviation. The existing method statically designs the backfill volume based on the current settlement volume, which fails to accurately predict the final state after the settlement stabilizes. As a result, the treated area sinks again during the continuous settlement process, resulting in the dilemma of "treatment followed by sinking", making project acceptance difficult and the treatment effect difficult to sustain. Summary of the Invention

[0006] To address the aforementioned technical problems, this invention provides a method and system for simultaneous mining and remediation of coal mining subsidence areas based on multi-source monitoring and dynamic prediction. It constructs an integrated closed-loop intelligent governance system encompassing monitoring, prediction, decision-making, and control. The core of this system lies in utilizing multi-source data fusion technology to achieve dynamic and accurate prediction of the subsidence process, thereby guiding adaptive backfilling control. Data after backfilling is fed back to the prediction and decision-making modules for further optimization. The system comprises a perception layer, a data layer, an intelligence layer, and an application layer. The sensing layer includes an integrated monitoring network that combines space, ground, and well monitoring. The data layer includes a fusion and processing network; The intelligent layer includes analysis and prediction networks; The application layer includes the execution and feedback network.

[0007] Preferably, the integrated "space-air-ground-well" monitoring network includes: The space-based monitoring module enables wide-area screening. The airborne monitoring module enables precise snapshot functionality; The surface monitoring module enables continuous real-time monitoring. The downhole monitoring module enables source sensing.

[0008] Preferably, the fusion and processing network includes: a data access and spatiotemporal matching module, a multi-source data fusion module, and a four-dimensional subsidence field model database.

[0009] Preferably, the analysis and prediction network includes: a dynamic settlement prediction model library, an optimal treatment timing decision algorithm module, and a dynamic backfill volume calculation method module.

[0010] Preferably, the execution and feedback network includes: a governance timing early warning module, a dynamic backfill design drawing generation module, a smart construction guidance module, and a post-governance monitoring module.

[0011] A method for simultaneous mining and remediation of coal mining subsidence areas based on multi-source monitoring and dynamic prediction specifically includes the following steps: Step 1: Obtain multi-source heterogeneous data through the perception layer; Step 2: The data layer processes the multi-source heterogeneous data and the verification data fed back from the application layer to obtain the fused time-series data; Step 3: The intelligent layer processes the fused time-series data and the verification data fed back from the application layer, and transmits the calculated prediction results and decision instructions to the application layer.

[0012] Step 4: Obtain verification data through the application layer to determine the optimal timing for remediation, generate dynamic backfill design drawings to guide intelligent construction, continuously monitor after remediation and feed the data back to the data layer and intelligent layer for continuous optimization.

[0013] Preferably, in step one: Step 1: Space-based monitoring. Using SAR satellite data such as Sentinel-1, and through PS-InSAR technology (Permanent Scatter Radar Interferometry), a large-scale, long-term surface deformation rate field is obtained to identify potential subsidence basin boundaries and key monitoring areas. The second step is aerial monitoring, which involves regularly using drones equipped with differential GNSS modules and LiDAR to conduct aerial surveys and quickly acquire a digital elevation model (DEM) of the treatment area with centimeter-level accuracy. This model is used to quantify the total settlement and earthwork volume in each stage. Step 3: Surface monitoring. Deploy IoT GNSS surface displacement monitoring stations to form a real-time monitoring network, continuously collect changes in the three-dimensional coordinates of the surface, and transmit the data back through 4G / 5G networks to provide continuous deformation time series of key points. Step 4: Downhole monitoring. Inclination sensors and microseismic monitoring arrays are deployed in key strata above the goaf to monitor the fracture and migration of overburden structures, providing early warning and mechanism explanation for surface subsidence.

[0014] Preferably, in step two: Step 1: Data access and spatiotemporal registration; establish a unified data interface and pipeline to automatically and securely aggregate raw data from different platforms and formats to the central data processing platform; and unify all spatial data under the same geographic coordinate system.

[0015] The second step is to fuse multi-source data to establish a unified spatiotemporal benchmark. This involves fusing InSAR deformation rate, UAV DEM differential results, GNSS time series data, and downhole exploration data using Kalman filtering to generate a high-precision, high-spatiotemporal-resolution "four-dimensional (space + time) subsidence field" model.

[0016] Preferably, in step three: Step 1: Dynamic settlement prediction. Based on the fused time series data, an improved time function model is introduced and coupled with machine learning algorithms to dynamically and continuously predict the future settlement of each computing unit. This not only predicts the final settlement, but more importantly, it predicts the time curve required for the settlement to reach stability and the settlement at any time point. The second step is to run the optimal treatment timing decision algorithm. The system sets a threshold for predicting the amount and speed of settlement. When the monitoring data indicates that a certain area is about to enter the decline period, the system automatically triggers a treatment warning and recommends the best intervention window, thus fundamentally avoiding the risks of treating too early or too late. Step 3: Calculate the dynamic backfill volume. The calculation method is as follows: The backfill volume is dynamically determined by the following formula: Dynamic backfill volume = [Predicted final settlement - Current actual settlement] × Area × Soil compaction coefficient.

[0017] Preferably, in step four: Step 1: Early warning of the timing of governance; Step 2: Generate dynamic backfill design drawings; Step 3: Guiding intelligent construction; Step 4: Post-treatment monitoring. After backfilling, the monitoring network continues to track and monitor the backfill area, comparing the actual settlement data with the predicted values. When deviations occur, the data is automatically fed back to the prediction model for correction and optimization, forming a closed loop and continuously improving the accuracy of subsequent predictions. Beneficial effects

[0018] Compared with the prior art, the present invention has the following beneficial effects: 1. Scientific nature: By integrating multi-source data and using dynamic prediction models, the timing of governance is transformed from "experience-based judgment" to "data-driven scientific decision-making," thus solving the problem of intervention timing.

[0019] 2. Precision: The dynamic backfill volume calculation model realizes "filling on demand and precise control", which is expected to reduce the waste of earthwork caused by over-design or secondary treatment and significantly reduce project costs.

[0020] 3. Foresight: By accurately predicting the subsidence process, the remediation design can "base itself on the present and look to the future", ensuring the long-term stability of the reclaimed land and significantly improving the success rate of the remediation project and its long-term effects.

[0021] 4. Systematic: It has constructed a complete, automated, and intelligent technology system covering the entire chain from monitoring and prediction to regulation, providing a replicable and scalable complete solution for the "mining while treating" model, effectively helping to achieve the goal of "reducing the increase and reducing the stock" in the treatment of unstable coal mining subsidence areas. Attached Figure Description

[0022] Figure 1 This is a flowchart of a method and system for simultaneous mining and treatment of coal mining subsidence areas based on multi-source monitoring and dynamic prediction. Detailed Implementation

[0023] The present invention will be further described below with reference to the accompanying drawings: As attached Figure 1 As shown: A coal mining subsidence area remediation system based on multi-source monitoring and dynamic prediction includes: The perception layer, data layer, intelligence layer, and application layer; The sensing layer includes an integrated monitoring network that combines space, ground, and well monitoring. The data layer includes a fusion and processing network; The intelligent layer includes analysis and prediction networks; The application layer includes the execution and feedback network.

[0024] In this implementation plan, specifically, the integrated "space-ground-well" monitoring network includes: The space-based monitoring module enables wide-area screening. The airborne monitoring module enables precise snapshot functionality; The surface monitoring module enables continuous real-time monitoring. The downhole monitoring module enables source sensing.

[0025] In this implementation scheme, specifically, the fusion and processing network includes: a data access and spatiotemporal matching module, a multi-source data fusion module, and a four-dimensional subsidence field model database.

[0026] In this implementation plan, specifically, the analysis and prediction network includes: a dynamic settlement prediction model library, an optimal treatment timing decision algorithm module, and a dynamic backfill volume calculation method module.

[0027] In this implementation plan, specifically, the execution and feedback network includes: a remediation timing early warning module, a dynamic backfill design drawing generation module, a smart construction guidance module, and a remediation post-remediation monitoring module.

[0028] A method for simultaneous mining and remediation of coal mining subsidence areas based on multi-source monitoring and dynamic prediction specifically includes the following steps: Step 1: Obtain multi-source heterogeneous data through the perception layer; Step 2: The data layer processes the multi-source heterogeneous data and the verification data fed back from the application layer to obtain the fused time-series data; Step 3: The intelligent layer processes the fused time-series data and the verification data fed back from the application layer, and transmits the calculated prediction results and decision instructions to the application layer.

[0029] Step 4: Obtain verification data through the application layer to determine the optimal timing for remediation, generate dynamic backfill design drawings to guide intelligent construction, continuously monitor after remediation and feed the data back to the data layer and intelligent layer for continuous optimization.

[0030] In this implementation plan, specifically in step one: Step 1: Space-based monitoring. Using SAR satellite data such as Sentinel-1, and through PS-InSAR technology (Permanent Scatter Radar Interferometry), a large-scale, long-term surface deformation rate field is obtained to identify potential subsidence basin boundaries and key monitoring areas. The second step is aerial monitoring, which involves regularly using drones equipped with differential GNSS modules and LiDAR to conduct aerial surveys and quickly acquire a digital elevation model (DEM) of the treatment area with centimeter-level accuracy. This model is used to quantify the total settlement and earthwork volume in each stage. Step 3: Surface monitoring. Deploy IoT GNSS surface displacement monitoring stations to form a real-time monitoring network, continuously collect changes in the three-dimensional coordinates of the surface, and transmit the data back through 4G / 5G networks to provide continuous deformation time series of key points. Step 4: Downhole monitoring. Inclination sensors and microseismic monitoring arrays are deployed in key strata above the goaf to monitor the fracture and migration of overburden structures, providing early warning and mechanism explanation for surface subsidence.

[0031] In this implementation plan, specifically in step two: Step 1: Data access and spatiotemporal registration; establish a unified data interface and pipeline to automatically and securely aggregate raw data from different platforms and formats to the central data processing platform; and unify all spatial data under the same geographic coordinate system.

[0032] The second step is to fuse multi-source data to establish a unified spatiotemporal benchmark. This involves fusing InSAR deformation rate, UAV DEM differential results, GNSS time series data, and downhole exploration data using Kalman filtering to generate a high-precision, high-spatiotemporal-resolution "four-dimensional (space + time) subsidence field" model.

[0033] In this implementation plan, specifically in step three: Step 1: Dynamic settlement prediction. Based on the fused time series data, an improved time function model is introduced and coupled with machine learning algorithms to dynamically and continuously predict the future settlement of each computing unit. This not only predicts the final settlement, but more importantly, it predicts the time curve required for the settlement to reach stability and the settlement at any time point. The second step is to run the optimal treatment timing decision algorithm. The system sets a threshold for predicting the amount and speed of settlement. When the monitoring data indicates that a certain area is about to enter the decline period, the system automatically triggers a treatment warning and recommends the best intervention window, thus fundamentally avoiding the risks of treating too early or too late. Step 3: Calculate the dynamic backfill volume. The calculation method is as follows: The backfill volume is dynamically determined by the following formula: Dynamic backfill volume = [Predicted final settlement - Current actual settlement] × Area × Soil compaction coefficient; This calculated value is dynamically adjusted as monitoring data is updated and the prediction model is corrected to ensure long-term stability after one treatment.

[0034] In this implementation plan, specifically in step four: Step 1: Early warning of the timing of governance; Step 2: Generate dynamic backfill design drawings; Step 3: Guiding intelligent construction; Step 4: Post-treatment monitoring. After backfilling, the monitoring network continues to track and monitor the backfill area, comparing the actual settlement data with the predicted values. When deviations occur, the data is automatically fed back to the prediction model for correction and optimization, forming a closed loop and continuously improving the accuracy of subsequent predictions.

[0035] Specifically, this implementation plan proposes a new paradigm of "dynamic prediction-threshold determination-adaptive control" for simultaneous collection and treatment, breaking the traditional linear process of "monitoring-static design-construction" and creating an intelligent closed-loop feedback system with a prediction model at its core, thereby realizing dynamic optimization and forward-looking management of governance behavior.

[0036] In this implementation plan, specifically, the spatiotemporal fusion and collaborative analysis technology of multi-source heterogeneous monitoring data, and the Kalman filter algorithm, solve the problem of fusion between data from different sources, with different precisions and different sampling rates, such as InSAR, UAV LiDAR, and GNSS, generating a "holographic map of the subsidence field" that far exceeds the capabilities of any single technology, providing a unique and reliable data foundation for accurate prediction.

[0037] Specifically, this implementation plan constructs an intelligent subsidence prediction model that couples mining mechanisms with data-driven approaches. It combines the traditional time function model based on rock movement laws with the powerful machine learning LSTM model, which utilizes the interpretability of physical mechanisms and leverages the ability of big data learning to capture complex nonlinear relationships, thereby significantly improving the accuracy of medium- and long-term predictions in unstable subsidence areas.

[0038] Specifically, this implementation plan develops a real-time calculation algorithm for backfill volume based on dynamic prediction results. By using the predicted future settlement as the core input variable, it achieves "on-demand, dynamic, and accurate" calculation of backfill volume, fundamentally avoiding secondary settlement and material waste. This is a key step in the transformation of the remediation project from "extensive" to "precise".

[0039] In this implementation plan, there is low dependence on specific equipment brands. Different levels of sensors and drone platforms can be flexibly selected according to the project budget. It can be applied to large-scale key governance projects as well as small and medium-sized mining areas, thus expanding the applicability of the technology.

[0040] In this implementation plan, the high groundwater level mining areas have extremely high requirements for the timing and precision of the treatment, where the value of this technology is most evident. It is also generally applicable to the treatment of subsidence areas caused by underground mining, thus increasing its regional applicability.

[0041] Any technical solution that achieves the above-mentioned technical effects by utilizing the technical solutions described in this invention, or by designing similar technical solutions by those skilled in the art under the inspiration of the technical solutions described in this invention, falls within the protection scope of this invention.

Claims

1. A system for simultaneous mining and remediation of coal mining subsidence areas based on multi-source monitoring and dynamic prediction, characterized in that, include: The perception layer, data layer, intelligence layer, and application layer; The sensing layer includes an integrated monitoring network encompassing "space-ground-well" systems. The data layer includes a fusion and processing network; The intelligent layer includes analysis and prediction networks; The application layer includes the execution and feedback network.

2. The coal mining subsidence area treatment system based on multi-source monitoring and dynamic prediction as described in claim 1, characterized in that, The integrated "space-ground-well" monitoring network includes: The space-based monitoring module enables wide-area screening. The airborne monitoring module enables precise snapshot functionality; The surface monitoring module enables continuous real-time monitoring. The downhole monitoring module enables source sensing.

3. The coal mining subsidence area treatment system based on multi-source monitoring and dynamic prediction as described in claim 1, characterized in that, The fusion and processing network includes: a data access and spatiotemporal matching module, a multi-source data fusion module, and a four-dimensional subsidence field model database.

4. The coal mining subsidence area treatment system based on multi-source monitoring and dynamic prediction as described in claim 1, characterized in that, The analysis and prediction network includes: a dynamic settlement prediction model library, an optimal treatment timing decision algorithm module, and a dynamic backfill volume calculation method module.

5. The coal mining subsidence area treatment system based on multi-source monitoring and dynamic prediction as described in claim 1, characterized in that, The execution and feedback network includes: a remediation timing early warning module, a dynamic backfill design drawing generation module, a smart construction guidance module, and a remediation post-remediation monitoring module.

6. A method for simultaneous mining and remediation of coal mining subsidence areas based on multi-source monitoring and dynamic prediction, characterized in that, Specifically, the following steps are included: Step 1: Obtain multi-source heterogeneous data through the perception layer; Step 2: The data layer processes the multi-source heterogeneous data and the verification data fed back from the application layer to obtain the fused time-series data; Step 3: Through the intelligent layer, the fused time series data and the verification data fed back from the application layer are processed, and the calculated prediction results and decision instructions are transmitted to the application layer. Step 4: Obtain verification data through the application layer to determine the optimal timing for remediation, generate dynamic backfill design drawings to guide intelligent construction, continuously monitor after remediation and feed the data back to the data layer and intelligent layer for continuous optimization.

7. The method for simultaneous mining and remediation of coal mining subsidence areas based on multi-source monitoring and dynamic prediction as described in claim 6, characterized in that, In step one: Step 1: Space-based monitoring. Using SAR satellite data and PS-InSAR technology (Permanent Scatter Radar Interferometry), a large-scale, long-term surface deformation rate field is obtained to identify potential subsidence basin boundaries and key monitoring areas. The second step is aerial monitoring, which involves regularly using drones equipped with differential GNSS modules and LiDAR to conduct aerial surveys and quickly acquire a digital elevation model (DEM) of the treatment area with centimeter-level accuracy. This model is used to quantify the total settlement and earthwork volume in each stage. Step 3: Surface monitoring. Deploy IoT GNSS surface displacement monitoring stations to form a real-time monitoring network, continuously collect changes in the three-dimensional coordinates of the surface, and transmit the data back through 4G / 5G networks to provide continuous deformation time series of key points. Step 4: Downhole monitoring. Inclination sensors and microseismic monitoring arrays are deployed in key strata above the goaf to monitor the fracture and migration of overburden structures, providing early warning and mechanism explanation for surface subsidence.

8. The method for simultaneous mining and remediation of coal mining subsidence areas based on multi-source monitoring and dynamic prediction as described in claim 6, characterized in that, In step two: Step 1: Data Access and Spatiotemporal Registration; Establish a unified data interface and pipeline to automatically and securely aggregate raw data from different platforms and formats to the central data processing platform; And unify all spatial data under the same geographic coordinate system; The second step is to fuse multi-source data to establish a unified spatiotemporal benchmark. InSAR deformation rate, UAV DEM differential results, GNSS time series data and downhole exploration data are fused using Kalman filtering to generate a high-precision, high spatiotemporal resolution "four-dimensional (space + time) subsidence field" model.

9. The method for simultaneous mining and remediation of coal mining subsidence areas based on multi-source monitoring and dynamic prediction as described in claim 6, characterized in that, In step three: Step 1: Dynamic settlement prediction. Based on the fused time series data, an improved time function model is introduced and coupled with machine learning algorithms to dynamically and continuously predict the future settlement of each computing unit. This not only predicts the final settlement, but more importantly, it predicts the time curve required for the settlement to reach stability and the settlement at any time point. The second step is to run the optimal treatment timing decision algorithm. The system sets a threshold for predicting the amount and speed of settlement. When the monitoring data indicates that a certain area is about to enter the decline period, the system automatically triggers a treatment warning and recommends the best intervention window, thus fundamentally avoiding the risks of treating too early or too late. Step 3: Calculate the dynamic backfill volume. The calculation method is as follows: The backfill volume is dynamically determined by the following formula: Dynamic backfill volume = [Predicted final settlement - Current actual settlement] × Area × Soil compaction coefficient.

10. The method for simultaneous mining and remediation of coal mining subsidence areas based on multi-source monitoring and dynamic prediction as described in claim 6, characterized in that, In step four: Step 1: Early warning of the timing of governance; Step 2: Generate dynamic backfill design drawings; Step 3: Guiding intelligent construction; Step 4: Post-treatment monitoring. After backfilling, the monitoring network continues to track and monitor the backfill area, comparing the actual settlement data with the predicted values. When deviations occur, the data is automatically fed back to the prediction model for correction and optimization, forming a closed loop and continuously improving the accuracy of subsequent predictions.

Images