Methods, devices, equipment and storage media for risk prevention and control in mining goaf areas

Abstract

This application discloses a method, device, equipment, and storage medium for risk prevention and control in mining goaf areas, relating to the field of safety monitoring technology, and aiming to solve the problems of low accuracy in goaf risk early warning and poor targeting in prevention and control. The method includes: collecting multi-source monitoring data through a sensor network deployed in different spaces within the goaf area, including microseismic data, ground pressure data, gas data, and displacement data; performing spatiotemporal alignment and standardization processing on the multi-source monitoring data to generate a fused dataset with a unified spatiotemporal benchmark; extracting risk feature vectors from the fused dataset based on a deep learning model, and analyzing the risk feature vectors using time-series prediction and clustering algorithms to assess the risk level and risk evolution trend of the goaf area; and generating and outputting risk prevention and control instructions matching specific areas of the goaf area based on the risk level and risk evolution trend.

Description

Technical Field

[0001] This application relates to the field of safety monitoring technology, and in particular to a method, device, equipment and storage medium for risk prevention and control in mining goaf areas. Background Technology

[0002] During mining operations, the formation of goaf areas is often accompanied by safety hazards such as rock instability, surface subsidence, and accumulation of harmful gases, which seriously threaten the safety of underground workers and equipment.

[0003] Currently, common methods for monitoring goaf areas include seismic wave detection, ground radar interferometry, and borehole stress monitoring. Specifically, existing monitoring systems mostly rely on single-type sensors or isolated data analysis methods. Although they can achieve deformation and stress state perception within a certain range, they still generally suffer from problems such as large monitoring blind spots, lagging data updates, and low degree of multi-source information fusion. Especially in complex mining environments with high stress and strong interference, it is difficult to achieve real-time and comprehensive perception of the risk status of goaf areas at all times and in all places, and it is impossible to accurately predict the risk development trend and carry out targeted prevention and control.

[0004] Therefore, how to construct a monitoring scheme that can achieve accurate prediction and targeted prevention and control has become a key technical problem that urgently needs to be solved in the field of mine safety. Summary of the Invention

[0005] The purpose of this application is to provide a method, device, equipment and storage medium for risk prevention and control in mining goaf areas, aiming to solve the problems of low accuracy of risk early warning and poor targeted prevention and control in goaf areas.

[0006] To achieve the above objectives, this application adopts the following technical solution: This application provides a method for risk prevention and control in mining goaf areas. The method includes: collecting multi-source monitoring data, including microseismic data, ground pressure data, gas data, and displacement data, through a sensor network deployed in different spaces within the goaf area; performing spatiotemporal alignment and standardization on the multi-source monitoring data to generate a fused dataset with a unified spatiotemporal benchmark; extracting risk feature vectors from the fused dataset based on a deep learning model, and analyzing the risk feature vectors using time series prediction and clustering algorithms to assess the risk level and risk evolution trend of the goaf area; and generating and outputting risk prevention and control instructions matching specific areas of the goaf area based on the risk level and risk evolution trend.

[0007] The mine goaf risk prevention and control method provided in this application deploys a multi-spatial sensor network to collaboratively collect multi-source monitoring data, achieving multi-dimensional perception of the rock mass stability, gas environment, and deformation state of the goaf. Then, it performs spatiotemporal alignment and standardization processing on the multi-source data to construct a fusion dataset with a unified spatiotemporal benchmark, effectively solving the problem of integrating and comprehensively utilizing heterogeneous data. Next, it uses a deep learning model to extract risk feature vectors and combines them with time-series prediction and clustering algorithms for analysis, accurately identifying risk patterns, assessing risk levels, and predicting their evolution trends from massive amounts of data, thus improving the early identification and dynamic assessment capabilities of potential risks in goaf areas. Finally, by generating prevention and control instructions precisely matched to specific areas based on risk levels and evolution trends, it achieves a shift from generalized early warning to precise control, improving the accuracy of goaf risk early warning and enabling targeted prevention and control deployment.

[0008] In some embodiments, the above-mentioned collection of multi-source monitoring data by a sensor network deployed in different spaces of the goaf includes: dividing the goaf into multiple risk monitoring areas according to the geomechanical model of the goaf; deploying corresponding sensors in each risk monitoring area, including multiple types of sensors such as microseismic sensors, stress sensors, gas sensors and displacement sensors; controlling each sensor to collect data synchronously according to a unified time reference, and aggregating the data to obtain multi-source monitoring data.

[0009] Based on this, this application deploys multiple types of sensors in different zones according to the geomechanical model and collects data synchronously, realizing spatial adaptation between the monitoring network and the rock mechanical behavior of the goaf, and effectively improving the comprehensiveness and synergy of data collection.

[0010] In some embodiments, the above-mentioned spatiotemporal alignment and standardization processing of multi-source monitoring data to generate a fusion dataset with a unified spatiotemporal reference includes: correcting the time delay and spatial deviation in the multi-source monitoring data caused by differences in sensor deployment locations; removing noise and outliers from the multi-source monitoring data; and mapping the corrected and modified multi-source monitoring data to a unified dimension according to a preset dimension to form a fusion dataset.

[0011] Based on this, this application eliminates data bias caused by sensor heterogeneity and deployment differences by performing spatiotemporal correction, noise removal and dimension unification on multi-source data, and constructs a high-quality, spatiotemporally consistent fusion dataset, thereby improving the accuracy of subsequent risk feature vector extraction and analysis.

[0012] In some embodiments, the above-mentioned extraction of risk feature vectors from the fused dataset based on a deep learning model, and analysis of the risk feature vectors using time series prediction and clustering algorithms to assess the risk level and risk evolution trend of the goaf, includes: extracting spatial features from the fused dataset using a convolutional neural network and extracting temporal features from the fused dataset using a recurrent neural network to construct a risk feature vector, which includes rock mass fracture features, stress concentration features, and gas seepage anomaly features; inputting the risk feature vector into a time series prediction model to obtain the risk evolution trend; and using a clustering algorithm to match the risk evolution trend with the features of historical risk cases to output the risk level.

[0013] Based on this, this application extracts spatial and temporal features by fusing convolutional neural networks and recurrent neural networks respectively, and combines temporal prediction with historical case clustering and matching to achieve in-depth mining of multiple risk feature vectors such as rock mass fracture, stress concentration and gas seepage, thereby enhancing the comprehensiveness of risk assessment.

[0014] In some embodiments, the risk prevention and control method for mining goaf areas provided in this application may further include: using an expert system based on rock mechanics rules to cross-validate the risk level, correcting the deviation of the single algorithm evaluation, and obtaining the validated risk level; wherein the expert system has built-in rock mass fracture threshold rules, stress mutation judgment rules, and gas abnormal diffusion rules.

[0015] Based on this, this application introduces an expert system based on rock mechanics rules to cross-validate the algorithm evaluation results, which effectively corrects the misjudgments that may occur in a single data-driven model and improves the reliability of risk level output.

[0016] In some embodiments, the above-mentioned generation and output of risk prevention and control instructions matching specific areas of the goaf based on risk level and risk evolution trend includes: determining the dominant risk type matching specific areas of the goaf based on risk evolution trend, wherein the dominant risk type includes rock mass instability risk, gas anomaly risk or structural deformation risk; and generating risk prevention and control instructions adapted to the dominant risk type for different risk levels and dominant risk types.

[0017] Based on this, this application identifies the dominant risk type by combining risk evolution trends and generates differentiated prevention and control instructions accordingly, realizing the transformation from generalized response to precise policy implementation, and improving the matching degree and execution effectiveness of prevention and control measures with actual risk scenarios.

[0018] In some embodiments, the above-mentioned generation of risk prevention and control instructions adapted to the dominant risk type for different risk levels and dominant risk types includes: outputting a first risk prevention and control instruction when the dominant risk type is rock mass instability risk, the first risk prevention and control instruction including the reinforcement range and support strength corresponding to the risk level; outputting a second risk prevention and control instruction when the dominant risk type is gas anomaly risk, the second risk prevention and control instruction including the ventilation volume adjustment range and monitoring frequency corresponding to the risk level; and outputting a third risk prevention and control instruction when the current risk type is structural deformation risk, the third risk prevention and control instruction including the evacuation range and monitoring level corresponding to the risk level.

[0019] Based on this, this application adaptively adjusts the specific parameters of the prevention and control instructions (such as reinforcement scope, ventilation volume, evacuation scope, etc.) according to the risk level, so that the prevention and control measures not only target the risk type, but also match the severity of the risk, thereby improving the utilization efficiency of prevention and control resources and the level of intelligence in emergency response.

[0020] This application provides a risk prevention and control device for mining goaf areas. The device includes: a data acquisition unit for acquiring multi-source monitoring data through a sensor network deployed in different spaces of the goaf area, including microseismic data, ground pressure data, gas data, and displacement data; a generation unit for performing spatiotemporal alignment and standardization processing on the multi-source monitoring data to generate a fused dataset with a unified spatiotemporal benchmark; a processing unit for extracting risk feature vectors from the fused dataset based on a deep learning model, and analyzing the risk feature vectors using time series prediction and clustering algorithms to assess the risk level and risk evolution trend of the goaf area; and an output unit for generating and outputting risk prevention and control instructions matching specific areas of the goaf area based on the risk level and risk evolution trend.

[0021] In some embodiments, the above-mentioned acquisition unit is specifically used for: dividing multiple risk monitoring areas according to the geomechanical model of the goaf; deploying corresponding sensors in each risk monitoring area, including multiple types of sensors such as microseismic sensors, stress sensors, gas sensors and displacement sensors; controlling each sensor to synchronously acquire data according to a unified time reference, and aggregating the data to obtain multi-source monitoring data.

[0022] In some embodiments, the above-mentioned generation unit is specifically used to: correct the time delay and spatial deviation caused by the difference in sensor deployment location in the multi-source monitoring data; remove noise and outliers in the multi-source monitoring data; and map the corrected and modified multi-source monitoring data to a unified dimension according to a preset dimension to form a fused dataset.

[0023] In some embodiments, the above processing unit is specifically used to: extract spatial features from the fused dataset using a convolutional neural network and extract temporal features from the fused dataset using a recurrent neural network to form a risk feature vector, the risk feature vector including rock mass fracture features, stress concentration features and gas seepage anomaly features; input the risk feature vector into a time series prediction model to obtain the risk evolution trend; and use a clustering algorithm to match the risk evolution trend with the features of historical risk cases to output the risk level.

[0024] In some embodiments, the processing unit is further configured to use an expert system based on rock mechanics rules to cross-validate the risk level, correct the deviation of the single algorithm evaluation, and obtain the validated risk level; wherein the expert system has built-in rock mass fracture threshold rules, stress change judgment rules, and gas abnormal diffusion rules.

[0025] In some embodiments, the above-mentioned output unit is specifically used to: determine the dominant risk type matching the specific area of ​​the goaf based on the risk evolution trend, wherein the dominant risk type includes rock mass instability risk, gas anomaly risk or structural deformation risk; and generate risk prevention and control instructions adapted to the dominant risk type for different risk levels and dominant risk types.

[0026] In some embodiments, the above-mentioned output unit is specifically used to: output a first risk prevention and control instruction when the dominant risk type is rock mass instability risk, the first risk prevention and control instruction including the reinforcement range and support strength corresponding to the risk level; output a second risk prevention and control instruction when the dominant risk type is gas anomaly risk, the second risk prevention and control instruction including the ventilation volume adjustment range and monitoring frequency corresponding to the risk level; and output a third risk prevention and control instruction when the current risk type is structural deformation risk, the third risk prevention and control instruction including the evacuation range and monitoring level corresponding to the risk level.

[0027] This application provides an electronic device, including: a processor; and a memory for storing processor-executable instructions; wherein the processor is configured to execute instructions to implement the above-described method for risk prevention and control of mining goaf areas.

[0028] This application provides a computer-readable storage medium storing instructions that, when executed on a terminal, cause the terminal to perform the mine goaf risk prevention and control method described above.

[0029] This application provides a computer program product containing instructions that, when executed by a computer, cause the computer to perform the mine goaf risk prevention and control method described above.

[0030] This application provides a chip, which includes a processor and a communication interface. The communication interface and the processor are coupled together. The processor is used to run computer programs or instructions to implement the risk prevention and control method for goaf areas in mines described above.

[0031] Specifically, the chip provided in this application embodiment also includes a memory for storing computer programs or instructions. Attached Figure Description

[0032] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0033] Figure 1 A flowchart illustrating a method for risk prevention and control in mining goaf areas, provided in this application embodiment; Figure 2 A structural diagram of a mine goaf risk prevention and control device provided in this application embodiment; Figure 3 This is a structural diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0034] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0035] In the description of this application, it should be understood that the terms "upper," "lower," "left," "right," "front," "rear," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or relative positional relationship shown in the accompanying drawings. They are used only for the convenience of describing this application and for simplification, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application. Unless otherwise specified, the above-mentioned orientational descriptions can be flexibly set in practical applications, provided that the relative positional relationships shown in the accompanying drawings are satisfied.

[0036] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0037] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," "linking," and "communication" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection. They can refer to a direct connection or an indirect connection through an intermediate medium, or a connection within two components. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0038] In some embodiments, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, article, or apparatus that includes that element.

[0039] In some embodiments, the words "exemplary" or "for example" are used to indicate that something is an example, illustration, or illustration. Any embodiment or design described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the words "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0040] In the description of this specification, specific features, structures, materials, or characteristics may be combined in any suitable manner in one or more embodiments or examples.

[0041] Currently, common methods for monitoring goaf areas include seismic wave detection, ground radar interferometry, and borehole stress monitoring. Although these technologies can achieve deformation and stress state perception within a certain range, they still generally suffer from problems such as large monitoring blind spots, delayed data updates, and low degree of multi-source information fusion, making it difficult to achieve real-time and comprehensive perception of the risk status of goaf areas at all times and in all spaces.

[0042] Furthermore, existing monitoring systems largely rely on single-type sensors or isolated data analysis methods, lacking the ability to collaboratively process and intelligently analyze multi-dimensional and heterogeneous monitoring data such as microseismic activity, ground pressure, gas concentration, and rock mass displacement. Especially in complex mining environments with high stress and strong interference, traditional methods struggle to identify early signs of risk evolution in goaf areas in a timely and accurate manner, let alone accurately predict and provide graded early warnings of risk development trends, resulting in a lack of foresight and specificity in prevention and control decisions.

[0043] Therefore, how to build a risk prevention and control system for goaf areas that can achieve deep integration of multi-source monitoring data, has real-time analysis and intelligent assessment capabilities, and supports hierarchical response and dynamic control has become a key technical problem that urgently needs to be solved in the field of mine safety.

[0044] Against this backdrop, to address the issues of low accuracy in early warning and poor targeting of goaf risk in related technologies, this application provides a method, device, equipment, and storage medium for goaf risk prevention and control. By integrating data from multiple sensors, including microseismic, ground pressure, gas, and displacement sensors, and performing spatiotemporal alignment and standardization, a unified dataset is formed. A deep learning model is used to extract risk feature vectors, and combined with time-series prediction and clustering algorithms, accurate assessment of the risk level and evolution trend of goafs is achieved. Finally, based on the assessment results, prevention and control instructions matching the specific regional risk are generated.

[0045] The following is a reference. Figure 1 The method for risk prevention and control of mining goaf areas provided in the embodiments of this application is described.

[0046] Figure 1 The flowchart of the risk prevention and control method for mining goaf areas provided in the embodiments of this application is shown. The subject executing the method can be an electronic device or various devices / modules in the electronic device, such as integrated circuits or chips. The embodiments of this application do not specifically limit this.

[0047] For example, such as Figure 1 As shown, the mine goaf risk prevention and control method provided in this application embodiment may include the following S101 to S104: S101. Collect multi-source monitoring data through a sensor network deployed in different spaces of the goaf.

[0048] In the embodiments of this application, the multi-source monitoring data includes microseismic data, ground pressure data, gas data, and displacement data.

[0049] For example, microseismic data refers to the elastic wave signals released during the rock mass fracturing process collected by a microseismic monitoring system, including parameters such as event occurrence time, source coordinates (X, Y, Z), released energy E, and event frequency.

[0050] In practical implementation, the South African Integrated Seismic System (ISS) microseismic monitoring system can be used, deploying 16 microseismic sensors with a sampling frequency of 2000Hz, a dynamic range ≥120dB, and the ability to monitor energy ranges from 10² to 10⁻⁶. 8 Microseismic events in J. By identifying the arrival times of longitudinal waves (P-waves) and transverse waves (S-waves), the Geiger algorithm is used to locate the seismic source with an accuracy of within 10 meters.

[0051] For example, ground pressure data refers to the stress state data of the surrounding rock in the goaf collected by a vibrating wire stress sensor and a hydraulic pressure cell, including the absolute stress value σ and the stress change rate d. σ / d t wait.

[0052] In practice, 20 stress monitoring points can be set up in key parts of the top, bottom and sides of the goaf. The sensors have a range of 0~40MPa and an accuracy of 0.1%FS. They are connected to form a network via RS-485 bus and collect data every 5 minutes to monitor stress concentration in real time.

[0053] For example, gas data refers to the volume concentration and its variation gradient dc / dt of gases (such as CH4 (methane), CO (carbon monoxide), O2 (oxygen), CO2 (carbon dioxide)) in the goaf area collected by multiple gas sensors.

[0054] In practice, a CH4 sensor based on the principle of infrared spectroscopy (range 0-100%, accuracy ±2%FS) and a CO sensor based on the principle of electrochemical spectroscopy (range 0-500ppm, accuracy ±1ppm) can be used. Eight monitoring points are set up in the upper corner of the goaf and the ventilation dead zone. Data is collected every 30 seconds and transmitted to the monitoring center via industrial Ethernet.

[0055] For example, displacement data refers to surface and rock mass displacement data collected by a global navigation satellite system (GNSS) receiver, total station, and laser rangefinder, including three-dimensional displacement (ΔX, ΔY, ΔZ) and displacement rate V.

[0056] In practice, 12 GNSS monitoring stations can be set up on the surface above the goaf, with a sampling frequency of 1Hz, a horizontal accuracy of ±2mm+0.5ppm, and an elevation accuracy of ±4mm+0.5ppm; 5 total stations can be set up in key underground roadways, with a measurement accuracy of ±1mm, to monitor the deformation of the surrounding rock in real time.

[0057] In some embodiments, when collecting multi-source monitoring data, multiple risk monitoring areas can be divided according to the geomechanical model of the goaf.

[0058] For example, a geomechanical model of the goaf can be established using fast lagrangian analysis of continua in 3 dimensions (FLAC3D) numerical simulation software. Based on the stress distribution cloud map, the extent of the plastic zone, and historical collapse, the goaf can be divided into three monitoring areas with different risk levels: "high stress concentration zone", "fracture development zone" and "relatively stable zone".

[0059] For example, the coal pillar area around the goaf is designated as a high stress concentration area, the collapse area in the middle of the goaf is designated as a fracture development area, and the solid coal area far from the goaf is designated as a relatively stable area.

[0060] In some embodiments, corresponding sensors are deployed in each risk monitoring area, and each sensor is controlled to collect data synchronously according to a unified time base, and the data is aggregated to obtain multi-source monitoring data.

[0061] The sensors include various types such as micro-vibration sensors, stress sensors, gas sensors, and displacement sensors.

[0062] In this embodiment, a microseismic sensor is used to collect microseismic data; a stress sensor is used to collect ground pressure data; a gas sensor is used to collect gas data; and a displacement sensor is used to collect displacement data.

[0063] In one example, microseismic sensors and stress sensors are deployed in areas of high stress concentration to monitor rock mass fracture and stress abrupt changes, with the sensor spacing controlled within 50 meters.

[0064] In another example, gas sensors and displacement sensors are deployed in the fracture development zone to monitor gas escape and rock strata movement, with the sensor spacing controlled within 30 meters.

[0065] It should be noted that all the above sensors achieve time synchronization through a Network Time Protocol (NTP) server, with time synchronization accuracy down to the millisecond level, ensuring that the data collected by different sensors have a unified time reference.

[0066] Thus, this application deploys multiple types of sensors in different zones based on the geomechanical model and collects data synchronously, achieving spatial adaptation between the monitoring network and the rock mechanical behavior of the goaf, effectively improving the comprehensiveness and synergy of data collection.

[0067] S102. Perform spatiotemporal alignment and standardization on multi-source monitoring data to generate a fusion dataset with a unified spatiotemporal benchmark.

[0068] In some embodiments, time delays and spatial biases caused by differences in sensor deployment locations in multi-source monitoring data can be corrected first.

[0069] For example, for microseismic data, the source location error can be corrected using the travel time difference of P-waves and S-waves based on the relative position of the sensor and the seismic source. The correction formula is ΔT = T s –T p The distance to the seismic source is calculated using the travel time difference ΔT.

[0070] Among them, T s When the S wave arrives, T p When the P-wave arrives, ΔT is the travel time difference.

[0071] For example, for displacement data, the local coordinate system of each monitoring point can be transformed into the mining geodetic coordinate system through coordinate transformation. The transformation parameters include 7 parameters (such as 3 translation parameters, 3 rotation parameters and 1 scale parameter) to ensure that all displacement data have a unified spatial reference benchmark.

[0072] In other embodiments, noise and outliers in multi-source monitoring data can be removed.

[0073] For example, a wavelet threshold denoising (WTD) algorithm can be used to filter out random noise in microseismic signals. The sym8 wavelet basis is used for 5-level decomposition, and the detail coefficients are filtered by the Stein unbiased risk estimation threshold.

[0074] For example, Grubbs' criterion can be used to identify and remove outliers in gas concentration data. The significance level is set to 0.05, and gas concentration data is judged as outliers when it is greater than the critical value of 0.05G.

[0075] For example, linear interpolation can also be used to fill in missing temperature and humidity data caused by sensor failure, ensuring the continuity and integrity of the data.

[0076] Furthermore, the corrected and revised multi-source monitoring data can be mapped to a unified dimension according to a preset dimension to form a fused dataset.

[0077] For example, the Min-Max standardization method can be used to normalize monitoring data of different dimensions to the [0,1] interval, and the normalization formula is X. norm =(XX min ) / (X max -X minThis process ultimately results in a multidimensional time-series dataset with a unified dimension, containing fields such as timestamps, spatial coordinates, and monitoring values.

[0078] Among them, the stress data is in MPa, the displacement data is in mm, the gas concentration data is in ppm, and the microseismic energy is in J.

[0079] Thus, by performing spatiotemporal correction, noise removal, and dimensional unification on multi-source data, this application eliminates data bias caused by sensor heterogeneity and deployment differences, constructs a high-quality, spatiotemporally consistent fusion dataset, and significantly improves the accuracy of subsequent risk feature vector extraction and analysis.

[0080] S103. Based on a deep learning model, risk feature vectors are extracted from the fused dataset, and time series prediction and clustering algorithms are combined to analyze the risk feature vectors and assess the risk level and risk evolution trend of the goaf.

[0081] In this embodiment of the application, the deep learning model may include convolutional neural networks and recurrent neural networks.

[0082] In some embodiments, convolutional neural networks can be used to extract spatial features from a fused dataset.

[0083] For example, spatial characteristics refer to the pattern characteristics that reflect the spatial distribution of rock mass fracturing, stress concentration and gas seepage in the goaf.

[0084] Specifically, a three-dimensional convolutional neural network (3D-CNN) can be used to perform convolution operations on the spatial distribution of microseismic events and the stress field of ground pressure monitoring points. The network structure includes four convolutional layers (e.g., convolutional kernel size 3×3×3, stride 1) and two pooling layers (e.g., pooling window 2×2×2), and finally extracts a 128-dimensional spatial feature vector, which includes features such as the geometric center of microseismic event clusters and the spatial range of stress anomaly zones.

[0085] In other embodiments, recurrent neural networks can be used to extract temporal features from the fused dataset.

[0086] For example, time series characteristics refer to the trends and periodic patterns that reflect the changes in the aforementioned monitoring data over time.

[0087] Specifically, a long short-term memory (LSTM) network can be used to process time-series data on gas concentrations. The network structure includes two LSTM layers (e.g., 64 hidden units) and one fully connected layer. Long-term dependencies in concentration changes are captured through forget gates, input gates, and output gates. Examples include the gradient characteristics of continuously rising CH4 concentrations and the periodic changes in CO concentrations.

[0088] Furthermore, spatial features and temporal features are fused together to form a risk feature vector.

[0089] In this embodiment of the application, the risk feature vector includes rock mass fracture features, stress concentration features, and gas seepage anomaly features.

[0090] For example, rock mass fracture characteristics refer to the energy release rate of microseismic events and the fracture propagation rate per unit time, calculated by the formula E. rate =ΣE i / Δt.

[0091] Among them, E i Let Δt be the energy of a single event, and Δt be the time window.

[0092] For example, stress concentration characteristics refer to the ratio of the current stress value to the rock compressive strength and its rate of change, calculated using the formula K. σ =σ / σ c .

[0093] Where σ is the measured stress, σ c It represents the compressive strength of the rock.

[0094] For example, the abnormal gas seepage characteristic refers to the ratio of the difference between the actual gas seepage velocity and the theoretical seepage rate, calculated by the formula A. g =(V 实际值 -V 理论值 ) / V 理论值。

[0095] Specifically, the 128-dimensional spatial features extracted by CNN and the 64-dimensional temporal features extracted by LSTM can be concatenated to form a 192-dimensional comprehensive risk feature vector.

[0096] In some embodiments, after obtaining the risk feature vector, the risk feature vector can be input into a time series prediction model to obtain the risk evolution trend.

[0097] In the embodiments of this application, the time series prediction model can be an autoregressive integrated moving average model (ARIMA) or a temporal convolutional network (TCN).

[0098] For example, a TCN model can be used, which has a structure of 8 residual blocks, each containing 2 dilated convolutional layers (e.g., dilation coefficients of 1, 2, 4, 8, 16, 32, 64, 128 respectively), a kernel size of 3, and 64 hidden units. The risk feature vector sequence of the past 7 days (containing 1008 data points) is input into the model to predict the changing trend of risk indicators in the goaf area in the next 24 hours.

[0099] Furthermore, clustering algorithms can be used to match risk evolution trends with the characteristics of historical risk cases to output risk levels.

[0100] In the embodiments of this application, the clustering algorithm can be the density-based spatial clustering of applications with noise (DBSCAN) algorithm or the K-Means algorithm.

[0101] In practical implementation, an improved DBSCAN algorithm can be used, with parameters set as follows: neighborhood radius eps = 0.5, minimum number of samples min. samples =5.

[0102] In this embodiment of the application, the historical risk case features may include risk feature vectors from the 24 hours preceding a historical roof fall, rib collapse, or gas outburst accident. For example, 50 historical cases were collected, each containing a 192-dimensional feature vector.

[0103] For example, the Euclidean distance between the current risk feature vector and the feature cluster center of historical cases can be calculated, and the current risk can be divided into three levels according to the distance (e.g., a distance less than 0.3 is "low risk", a distance between 0.3 and 0.7 is "medium risk", and a distance greater than 0.7 is "high risk").

[0104] Thus, this application achieves in-depth mining of multiple risk feature vectors such as rock mass fracture, stress concentration, and gas seepage by integrating convolutional neural networks and recurrent neural networks to extract spatial and temporal features respectively, and combining temporal prediction with historical case clustering and matching, thereby enhancing the comprehensiveness of risk assessment.

[0105] S104. Based on the risk level and risk evolution trend, generate and output risk prevention and control instructions that match the specific area of ​​the goaf.

[0106] In some embodiments, the dominant risk type matching a specific area of ​​the goaf can be determined based on the risk evolution trend.

[0107] In the embodiments of this application, the dominant risk type may include rock mass instability risk, gas anomaly risk, or structural deformation risk.

[0108] For example, rock mass instability risk refers to the risk of sudden collapse or plastic deformation of the roof or surrounding rock in a goaf under high stress. Specifically, in the stress concentration factor K... σ >0.8 and microseismic energy release rate E rate >10 5 When the speed is J / h, it is judged as a risk of rock mass instability.

[0109] For example, gas anomaly risk refers to the risk of combustion or explosion due to the excessive and continuous accumulation of CH4 or CO concentrations in the goaf. Specifically, a gas anomaly risk is identified when the CH4 concentration is >1.0% or the CO concentration is >24ppm and the concentration gradient dc / dt is >5ppm / min.

[0110] For example, structural deformation risk refers to the risk of ground subsidence or collapse above the goaf, endangering surface buildings and facilities. Specifically, structural deformation risk is identified when the displacement rate V > 5 mm / d and the cumulative displacement > 50 mm.

[0111] Specifically, the weight of each indicator can be determined by analyzing the contribution of various indicators in the risk feature vector and using the entropy weight method. The calculation formula is w. j =(1-e j ) / Σ(1-e j The risk type corresponding to the indicator with the highest weight is selected as the dominant risk type.

[0112] Among them, e j Let be the information entropy of the j-th indicator.

[0113] For example, if the weights of stress concentration characteristics and rock mass fracture characteristics reach 0.6, while the weight of gas seepage anomaly characteristics is only 0.2, then the current dominant risk type is determined to be rock mass instability risk.

[0114] Furthermore, risk control instructions adapted to the dominant risk type can be generated for different risk levels and dominant risk types.

[0115] In some embodiments, the urgency and scope of risk control instructions can be determined based on the level of risk.

[0116] In one example, for high-risk levels (e.g., distance > 0.7), immediate area lockdown and evacuation are required, with a response time of no more than 10 minutes.

[0117] In another example, for medium-risk levels (e.g., distance 0.3 to 0.7), reinforcement work is required to be completed within 8 hours.

[0118] In another example, for low-risk levels (e.g., distance <0.3), it is required to increase the monitoring frequency within 24 hours.

[0119] Thus, by combining risk evolution trends to identify dominant risk types and generating differentiated prevention and control instructions accordingly, this application achieves a shift from generalized response to precise policy implementation, improving the matching degree and execution effectiveness of prevention and control measures with actual risk scenarios.

[0120] In the mine goaf risk prevention and control method provided in this application embodiment, multi-source monitoring data is collected collaboratively by deploying a multi-space sensor network, realizing multi-dimensional perception of rock mass stability, gas environment, and deformation state in the goaf. Then, the multi-source data undergoes spatiotemporal alignment and standardization to construct a fusion dataset with a unified spatiotemporal benchmark, effectively solving the problem of integrating and comprehensively utilizing heterogeneous data. Next, a deep learning model is used to extract risk feature vectors, which are then analyzed in conjunction with time-series prediction and clustering algorithms to accurately identify risk patterns, assess risk levels, and predict their evolution trends from massive amounts of data, improving the early identification and dynamic assessment capabilities of potential risks in goaf areas. Finally, by generating prevention and control instructions precisely matched to specific areas based on risk levels and evolution trends, a shift from generalized early warning to precise control is achieved, improving the accuracy of goaf risk early warning and enabling targeted prevention and control deployment.

[0121] Optionally, after determining the risk level in S103 above, the risk level can be cross-validated using an expert system based on rock mechanics rules to correct the bias of the single algorithm evaluation and obtain the validated risk level.

[0122] In this embodiment, the expert system has built-in rules for rock mass fracture threshold, stress mutation determination, and gas anomaly diffusion.

[0123] For example, the rock mass fracture threshold rule refers to the rule that when the cumulative energy released by microseismic events exceeds 10... 5 When the frequency of an event exceeds 50 times within 24 hours, it is determined to be a critical state of rock mass fracture.

[0124] For example, the stress mutation judgment rule means that when the rate of change of the stress value at the monitoring point exceeds 15% of its compressive strength within 1 hour, it is judged as a stress mutation.

[0125] For example, the gas abnormal diffusion rule means that when the CO concentration gradient exceeds 5 ppm / min and the O2 concentration is below 18%, it is determined to be gas abnormal diffusion.

[0126] Specifically, the risk level output by the deep learning model can be weighted and fused with the result of rule-based reasoning from the expert system. The weight allocation can be determined using the analytic hierarchy process (AHP), and a judgment matrix can be constructed to calculate the weight of each indicator.

[0127] Thus, by introducing an expert system based on rock mechanics rules to cross-validate the algorithm evaluation results, this application effectively corrects the misjudgments that may occur in a single data-driven model and improves the reliability of risk level output.

[0128] Optionally, if the results of cross-validation of risk levels using an expert system based on rock mechanics rules differ by more than one level, a manual review process can be initiated, whereby three senior engineers independently evaluate the risk levels and a majority vote is used to determine the final risk level.

[0129] Optionally, the above-mentioned risk control instructions, adapted to the dominant risk type, can be generated for different risk levels and dominant risk types. Specifically, these instructions may include: In one alternative implementation, a first risk control instruction can be output when the dominant risk type is rock mass instability risk.

[0130] The first risk prevention and control instruction includes the reinforcement scope and support strength corresponding to the risk level.

[0131] For example, for a high-risk level, the instruction could be: "Immediately carry out grouting reinforcement within 50m of the western roof of the goaf, with a grouting pressure of not less than 10MPa, a grouting volume of not less than 50m³, and use 425# ordinary Portland cement with a water-cement ratio of 0.8:1."

[0132] For example, for a medium-risk level, the instruction could be: "Within 24 hours, install No. 36 U-shaped steel supports within 30m of the surrounding rock on the east side of the goaf, with a spacing of 0.8m, a torque of 300N·m, and a preload of not less than 100kN."

[0133] In another alternative implementation, a second risk control instruction can be output when the dominant risk type is gas anomaly risk.

[0134] The second risk control instruction includes the adjustment range of ventilation volume and monitoring frequency corresponding to the risk level.

[0135] For example, for a high-risk level, the instruction could be: "Immediately increase the air volume of ventilation shaft No. 3 from 1200 m³ / min to 1800 m³ / min (an increase of 50%), turn on the standby ventilation fan, and monitor the gas concentration in the goaf every 10 minutes, using an infrared spectrometer for precise measurement."

[0136] For example, for a medium-risk level, the instruction could be: "Turn on the auxiliary ventilation fan in roadway No. 5, adjust the air volume to 800 m³ / min, monitor the gas concentration once per hour, and use a portable gas detector for routine monitoring."

[0137] In another alternative implementation, a third risk control instruction can be output when the current risk type is structural deformation risk.

[0138] The third risk prevention and control instruction includes the evacuation scope and monitoring level corresponding to the risk level.

[0139] For example, for high-risk levels, the instruction could be: "Immediately evacuate all personnel within a 500m radius of the surface above the goaf, establish a warning line, and initiate a drone swarm patrol every hour, using lidar for three-dimensional scanning with an accuracy of ±2cm."

[0140] For example, for a medium-risk level, the instruction could be: "Restrict personnel from entering within a 200m radius of the ground surface above the goaf, set up warning signs, arrange for a dedicated person to patrol every 4 hours, and use a total station to monitor displacement with a measurement accuracy of ±1mm."

[0141] Thus, by adaptively adjusting the specific parameters of the prevention and control instructions (such as reinforcement scope, ventilation volume, evacuation scope, etc.) according to the risk level, this application enables the prevention and control measures to not only target the type of risk but also match the severity of the risk, thereby improving the utilization efficiency of prevention and control resources and the level of intelligence in emergency response.

[0142] The above primarily describes the solutions provided in the embodiments of this application from a methodological perspective. To achieve the above functions, the mine goaf risk prevention and control device or electronic device includes corresponding hardware structures and / or software modules for performing each function. Those skilled in the art should readily recognize that, based on the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0143] This application embodiment can, based on the above method, exemplarily divide a mine goaf risk prevention and control device or electronic device into functional modules. For example, the mine goaf risk prevention and control device or electronic device may include various functional modules corresponding to each functional division, or two or more functions may be integrated into one processing module. The integrated module can be implemented in hardware or as a software functional module. It should be noted that the module division in this application embodiment is illustrative and only represents one logical functional division; in actual implementation, there may be other division methods.

[0144] Figure 2 This is a structural diagram of a mine goaf risk prevention and control device provided in an embodiment of this application. The mine goaf risk prevention and control device 200 includes: a data acquisition unit 201, a generation unit 202, a processing unit 203, and an output unit 204.

[0145] The system comprises: a data acquisition unit 201, used to acquire multi-source monitoring data through a sensor network deployed in different spaces of the goaf, including microseismic data, ground pressure data, gas data, and displacement data; a generation unit 202, used to perform spatiotemporal alignment and standardization processing on the multi-source monitoring data to generate a fusion dataset with a unified spatiotemporal benchmark; a processing unit 203, used to extract risk feature vectors from the fusion dataset based on a deep learning model, and to analyze the risk feature vectors by combining time series prediction and clustering algorithms to assess the risk level and risk evolution trend of the goaf; and an output unit 204, used to generate and output risk prevention and control instructions that match the specific area of ​​the goaf based on the risk level and risk evolution trend.

[0146] In some embodiments, the acquisition unit 201 is specifically used to: divide multiple risk monitoring areas according to the geomechanical model of the goaf; deploy corresponding sensors in each risk monitoring area, including multiple types of sensors such as microseismic sensors, stress sensors, gas sensors and displacement sensors; control each sensor to synchronously acquire data according to a unified time reference, and aggregate the data to obtain multi-source monitoring data.

[0147] In some embodiments, the generation unit 202 is specifically used to: correct the time delay and spatial deviation caused by the difference in sensor deployment locations in the multi-source monitoring data; remove noise and outliers in the multi-source monitoring data; and map the corrected and modified multi-source monitoring data to a unified dimension according to a preset dimension to form a fused dataset.

[0148] In some embodiments, the processing unit 203 is specifically used to: extract spatial features from the fused dataset using a convolutional neural network and extract temporal features from the fused dataset using a recurrent neural network to form a risk feature vector, the risk feature vector including rock mass fracture features, stress concentration features and gas seepage anomaly features; input the risk feature vector into a time series prediction model to obtain the risk evolution trend; and use a clustering algorithm to match the risk evolution trend with the features of historical risk cases to output the risk level.

[0149] In some embodiments, the processing unit 203 is further configured to use an expert system based on rock mechanics rules to cross-validate the risk level, correct the deviation of the single algorithm evaluation, and obtain the validated risk level; wherein the expert system has built-in rock mass fracture threshold rules, stress change judgment rules and gas abnormal diffusion rules.

[0150] In some embodiments, the output unit 204 is specifically used to: determine the dominant risk type matching a specific area of ​​the goaf based on the risk evolution trend, wherein the dominant risk type includes rock mass instability risk, gas anomaly risk or structural deformation risk; and generate risk prevention and control instructions adapted to the dominant risk type for different risk levels and dominant risk types.

[0151] In some embodiments, the output unit 204 is specifically used to: output a first risk prevention and control instruction when the dominant risk type is rock mass instability risk, the first risk prevention and control instruction including the reinforcement range and support strength corresponding to the risk level; output a second risk prevention and control instruction when the dominant risk type is gas anomaly risk, the second risk prevention and control instruction including the ventilation volume adjustment range and monitoring frequency corresponding to the risk level; and output a third risk prevention and control instruction when the current risk type is structural deformation risk, the third risk prevention and control instruction including the evacuation range and monitoring level corresponding to the risk level.

[0152] In the mine goaf risk prevention and control device provided in this application embodiment, multi-source monitoring data is collected collaboratively by deploying a multi-space sensor network, realizing multi-dimensional perception of the rock mass stability, gas environment, and deformation state of the goaf. Then, the multi-source data undergoes spatiotemporal alignment and standardization to construct a fusion dataset with a unified spatiotemporal benchmark, effectively solving the problem of integrating and comprehensively utilizing heterogeneous data. Next, a deep learning model is used to extract risk feature vectors, which are then analyzed in conjunction with time-series prediction and clustering algorithms to accurately identify risk patterns, assess risk levels, and predict their evolution trends from massive amounts of data, improving the early identification and dynamic assessment capabilities of potential risks in goaf areas. Finally, by generating prevention and control instructions precisely matched to specific areas based on risk levels and evolution trends, a shift from generalized early warning to precise control is achieved, improving the accuracy of goaf risk early warning and enabling targeted prevention and control deployment.

[0153] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0154] Figure 3 This is a structural diagram of an electronic device provided in an embodiment of this application. Figure 3 As shown, the electronic device 300 includes, but is not limited to, a processor 301 and a memory 302.

[0155] The aforementioned memory 302 is used to store the executable instructions of the aforementioned processor 301. It is understood that the aforementioned processor 301 is configured to execute instructions to implement the mine goaf risk prevention and control method in the above embodiments.

[0156] It should be noted that those skilled in the art will understand that Figure 3 The electronic device structure shown does not constitute a limitation on the electronic device; the electronic device may include, but is not limited to, other electronic devices. Figure 3 This may indicate more or fewer components, or combinations of certain components, or different component arrangements.

[0157] Processor 301 is the control center of the electronic device. It connects various parts of the electronic device via various interfaces and lines. By running or executing software programs and / or modules stored in memory 302, and by calling data stored in memory 302, it performs various functions and processes data, thereby providing overall monitoring of the electronic device. Processor 301 may include one or more processing units. Optionally, processor 301 may integrate an application processor and a modem processor. The application processor mainly handles the operating system, user interface, and applications, while the modem processor mainly handles wireless communication. It is understood that the modem processor may not be integrated into processor 301.

[0158] The memory 302 can be used to store software programs and various data. The memory 302 may mainly include a program storage area and a data storage area, wherein the program storage area may store the operating system, application programs required by at least one functional module (such as a determination unit, a processing unit, etc.), etc. In addition, the memory 302 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0159] In an exemplary embodiment, a computer-readable storage medium including instructions is also provided, such as a memory 302 including instructions, which can be executed by a processor 301 of an electronic device 300 to implement the mine goaf risk prevention and control method in the above embodiments.

[0160] In actual implementation, Figure 2 The steps performed by the acquisition unit 201, generation unit 202, processing unit 203, and output unit 204 can all be performed by... Figure 3 The processor 301 calls the computer program stored in the memory 302 to implement the process. The specific execution process can be found in the method section of the previous embodiment, and will not be repeated here.

[0161] Optionally, the computer-readable storage medium may be a non-transitory computer-readable storage medium, such as a read-only memory (ROM), random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device.

[0162] In an exemplary embodiment, this application also provides a computer program product including one or more instructions, which can be executed by the processor 301 of an electronic device to complete the mine goaf risk prevention and control method in the above embodiments.

[0163] It should be noted that when one or more instructions in the computer-readable storage medium or computer program product are executed by the processor of an electronic device, they implement the various processes of the above method embodiments and achieve the same technical effect as the above method. To avoid repetition, they will not be described again here.

[0164] Through the above description of the embodiments, those skilled in the art can clearly understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0165] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another apparatus, or some features may be ignored or not executed. Furthermore, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0166] The units described as separate components may or may not be physically separate. A component shown as a unit can be one or more physical units; that is, it can be located in one place or distributed in multiple different locations. Some or all of the classified units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0167] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0168] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a readable storage medium. Based on this understanding, the technical solution of the embodiments of this application, essentially, or the part that contributes to the prior art, or a complete or partial classification of the technical solution, can be embodied in the form of a software product. This software product is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0169] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for risk prevention and control in mining goaf areas, characterized in that, The method includes: Multi-source monitoring data is collected by a sensor network deployed in different spaces of the goaf. The multi-source monitoring data includes microseismic data, ground pressure data, gas data, and displacement data. The multi-source monitoring data is spatiotemporally aligned and standardized to generate a fusion dataset with a unified spatiotemporal reference. Risk feature vectors are extracted from the fused dataset based on a deep learning model, and the risk feature vectors are analyzed by combining time series prediction and clustering algorithms to assess the risk level and risk evolution trend of the goaf area. Based on the risk level and the risk evolution trend, risk prevention and control instructions matching the specific area of ​​the goaf are generated and output.

2. The method for risk prevention and control in mining goaf areas according to claim 1, characterized in that, The collection of multi-source monitoring data through a sensor network deployed in different spaces within the goaf includes: Multiple risk monitoring zones were divided based on the geomechanical model of the goaf area; A corresponding sensor is deployed in each of the aforementioned risk monitoring areas, and the sensor includes multiple types of sensors such as micro-vibration sensors, stress sensors, gas sensors, and displacement sensors; The system controls each sensor to collect data synchronously according to a unified time reference, and then aggregates the data to obtain the multi-source monitoring data.

3. The method for risk prevention and control in mining goaf areas according to claim 1, characterized in that, The process of performing spatiotemporal alignment and standardization on the multi-source monitoring data to generate a fusion dataset with a unified spatiotemporal benchmark includes: Correcting the time delay and spatial deviation in the multi-source monitoring data caused by differences in sensor deployment locations; Remove noise and outliers from the multi-source monitoring data; The corrected and revised multi-source monitoring data are mapped to a unified dimension according to a preset dimension to form the fused dataset.

4. The method for risk prevention and control in mining goaf areas according to claim 1, characterized in that, The process of extracting risk feature vectors from the fused dataset using a deep learning model, and analyzing these risk feature vectors using time-series prediction and clustering algorithms to assess the risk level and risk evolution trend of the goaf area includes: Spatial features in the fused dataset are extracted using a convolutional neural network, and temporal features are extracted using a recurrent neural network to construct the risk feature vector, which includes rock mass fracture features, stress concentration features, and gas seepage anomaly features. The risk feature vector is input into the time series prediction model to obtain the risk evolution trend; Clustering algorithms are used to match the risk evolution trend with the characteristics of historical risk cases, and the risk level is output.

5. The method for risk prevention and control in mining goaf areas according to claim 4, characterized in that, The method further includes: The risk level is cross-validated using an expert system based on rock mechanics rules to correct the bias of the single algorithm assessment and obtain the validated risk level. The expert system includes built-in rules for rock mass fracture threshold, stress mutation determination, and gas anomaly diffusion.

6. The method for risk prevention and control in mining goaf areas according to claim 1 or 5, characterized in that, The step of generating and outputting risk prevention and control instructions matching the specific area of ​​the goaf based on the risk level and the risk evolution trend includes: Based on the risk evolution trend, the dominant risk type matching the specific area of ​​the goaf is determined. The dominant risk type includes rock mass instability risk, gas anomaly risk, or structural deformation risk. For different risk levels and dominant risk types, generate risk prevention and control instructions that are adapted to the dominant risk type.

7. The method for risk prevention and control in mining goaf areas according to claim 6, characterized in that, The generation of risk control instructions adapted to the dominant risk type for different risk levels and dominant risk types includes: When the dominant risk type is the rock mass instability risk, a first risk prevention and control instruction is output, which includes the reinforcement range and support strength corresponding to the risk level. When the dominant risk type is the gas anomaly risk, a second risk control instruction is output, which includes the ventilation volume adjustment range and monitoring frequency corresponding to the risk level. If the current risk type is the structural deformation risk, a third risk prevention and control instruction is output, which includes the evacuation range and monitoring level corresponding to the risk level.

8. A risk prevention and control device for goaf areas in mines, characterized in that, The device includes: The acquisition unit is used to acquire multi-source monitoring data through a sensor network deployed in different spaces of the goaf. The multi-source monitoring data includes microseismic data, ground pressure data, gas data, and displacement data. The generation unit is used to perform spatiotemporal alignment and standardization processing on the multi-source monitoring data to generate a fusion dataset with a unified spatiotemporal reference. The processing unit is used to extract risk feature vectors from the fused dataset based on a deep learning model, and to analyze the risk feature vectors in combination with time series prediction and clustering algorithms to evaluate the risk level and risk evolution trend of the goaf area. The output unit is used to generate and output risk prevention and control instructions that match the specific area of ​​the goaf, based on the risk level and the risk evolution trend.

9. An electronic device, characterized in that, include: processor; Memory used to store the processor's executable instructions; The processor is configured to execute the instructions to implement the mine goaf risk prevention and control method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing instructions, characterized in that, When the computer executes the instruction, the computer performs the risk prevention and control method for goaf areas in mines as described in any one of claims 1 to 7.

Images