A method for comprehensive prevention and control of close coal seam mining based on multi-disaster collaborative early warning

Abstract

The application discloses a kind of near distance coal seam mining integrated prevention and control method based on multi-disaster collaborative early warning.Belongs to mine prevention and control technical field.The method includes: according to the projection of overlying goaf coal pillar, the stope is divided into first area and second area;Deploy monitoring points in each area, and stress sensor, microseismic sensor and gas sensor are arranged at monitoring points;Build reference feature vector;According to the feature vector of each monitoring point in the second area and the reference feature vector, a first risk monitoring point set is obtained;Gas emission abnormal monitoring points are screened out using gradient anomaly algorithm;Spatial clustering analysis is carried out, and disaster warning is carried out based on the morphology of the cluster obtained by clustering.The application divides the stope into first area and second area by overlying goaf coal pillar projection, and clearly defines the risk dominant factor of different areas, only the second area is subjected to stress, microseismic, gas progressive collaborative screening, to avoid misjudgment of insensitive gas indicators under coal pillar.

Description

Technical Field

[0001] This invention belongs to the field of mine prevention and control technology, specifically relating to a comprehensive prevention and control method for close-range coal seam mining based on multi-hazard collaborative early warning. Background Technology

[0002] Mining close to coal seams is a common form of underground coal mining, especially when the overlying coal seam has been mined out and the underlying coal seam continues to be mined. The coal pillar left in the overlying goaf will transmit high-concentration stress to the underlying coal seam, forming a significant area of ​​increased projected stress. The mechanical behavior of the coal and rock mass and the gas occurrence state in this area and its surroundings change drastically, making it highly susceptible to various disasters such as rockbursts, coal and gas outbursts, and roof water hazards. More complexly, these disaster factors often coexist and interact: stress concentration leads to coal body fracturing, which generates microseismic signals; simultaneously, the development of fractures promotes the desorption and release of adsorbed gas, and the seepage characteristics of the water-bearing medium also change. However, the dominant risk factors differ significantly in different locations—the area directly beneath the coal pillar is dominated by high-stress compression failure, where gas is usually squeezed out, posing a relatively low risk; while in the decompression zones far from the coal pillar, the coal body fractures develop, and a large amount of gas is desorbed, becoming the main source of risk. Currently, coal mine sites generally use single stress monitoring, microseismic monitoring, or gas monitoring systems, which operate independently and lack multi-physics collaborative analysis and precursor fusion. Summary of the Invention

[0003] One objective of this invention is to provide a comprehensive prevention and control method for near-distance coal seam mining based on multi-hazard collaborative early warning, which can solve the aforementioned technical problems in the prior art.

[0004] According to a first aspect of the present invention, a comprehensive prevention and control method for near-field coal seam mining based on multi-hazard collaborative early warning is provided, comprising the following steps:

[0005] Based on the projection of the coal pillar in the overlying goaf, the mining area is divided into the first area and the second area.

[0006] Monitoring points are deployed in each area, and stress sensors, micro-vibration sensors, and gas sensors are installed at these monitoring points.

[0007] Based on the stress change rate and microseismic energy release rate of static monitoring points, a benchmark feature vector is constructed under normal mining conditions.

[0008] Based on the feature vectors of each monitoring point in the second region and the baseline feature vector, a first set of risk monitoring points is obtained;

[0009] Based on the gas concentration sequence of each monitoring point in the first risk monitoring point set, the gradient anomaly algorithm is used to screen out monitoring points with abnormal gas emission.

[0010] Spatial cluster analysis was performed on the selected monitoring points with abnormal gas outbursts, and disaster early warning was made based on the morphology of the clusters obtained from the clustering.

[0011] Wherein, the first region is the region directly below the projection of the coal pillar of the overlying goaf and its outer edge region, and the distance between the outer edge region and the edge of the region directly below the projection is not less than a first threshold.

[0012] The outer edges of the second region and the first region are not less than the second threshold.

[0013] The static monitoring point is located at least 100m from the outer edge of the first area.

[0014] Optionally, the reference feature vector further includes the envelope entropy reference distribution of the modal components;

[0015] The modal components are obtained by performing variational mode decomposition on the stress signals collected from the static monitoring points;

[0016] The envelope entropy baseline distribution is the statistical distribution of the envelope entropy of each modal component.

[0017] Optionally, the benchmark feature vector may further include a multi-field coupling correlation factor between stress and microseismic sequence;

[0018] The process of obtaining the multi-field coupling correlation factor includes:

[0019] Within the first time window, time-delay cross-correlation calculations are performed on the stress change rate and microseismic energy release rate obtained from static monitoring points;

[0020] The maximum time delay cross-correlation coefficient within the preset time delay window is used as the multi-field coupling correlation factor.

[0021] Optionally, the first set of risk monitoring points is determined by calculating the Euclidean distance or Mahalanobis distance between the base feature vector and the baseline feature vector.

[0022] Optionally, the monitoring points for the gas outburst anomaly are obtained through the following methods:

[0023] Calculate the characteristic value of gas concentration change based on the gas concentration increment, the first-order rate of change of gas concentration, and the second-order rate of change of gas concentration within the second time window;

[0024] If the characteristic value of the gas concentration change corresponding to the monitoring point continuously exceeds the third threshold, the monitoring point will be identified as a monitoring point with abnormal gas emission.

[0025] Optionally, the spatial clustering analysis employs the DBSCAN algorithm, and the number of members within the minimum cluster is determined based on the deployment density of the sensors.

[0026] Optionally, principal component analysis can be used to obtain the geometric characteristics of the anomalous clusters, and the risk type can be determined based on the geometric characteristics of the anomalous clusters.

[0027] Optionally, the geometric feature includes elongation;

[0028] The calculation process for the elongation includes:

[0029] Obtain the covariance matrix formed by the monitoring points within the anomaly cluster;

[0030] Calculate the first and second eigenvalues ​​of the covariance matrix;

[0031] The risk type is determined based on the first feature value and the second feature value;

[0032] Elongation is the ratio of the first characteristic value to the second characteristic value.

[0033] Optionally, in response to the ratio of the first feature value to the second feature value being greater than 3, the risk type is determined to be a stress concentration zone induced by the edge of the coal pillar;

[0034] In response to the ratio being between 1.1 and 1.5, the risk type is determined to be block fracture risk due to pressure.

[0035] Optionally, in response to the ratio of the convex hull area of ​​the abnormal cluster to the number of points within the abnormal cluster being greater than a fourth threshold, the monitoring point data within the abnormal cluster is determined to be abnormal data.

[0036] The beneficial effects of this invention are as follows: By projecting the coal pillar in the overlying goaf, the invention divides the mining area into a first region and a second region, and clarifies the dominant risk factors in different regions. The first region is mainly characterized by stress concentration, while the second region is mainly characterized by gas outburst. Based on this, only the second region undergoes a stepwise coordinated screening of stress, micro-seismic activity, and gas, avoiding misjudgments of insensitive gas indicators below the coal pillar. Attached Figure Description

[0037] Figure 1 This is a flowchart of a comprehensive prevention and control method for near-distance coal seam mining based on multi-hazard collaborative early warning in an embodiment of the present invention. Detailed Implementation

[0038] Various exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be noted that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps set forth in these embodiments do not limit the scope of the invention.

[0039] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit the invention or its application or use.

[0040] Techniques, methods, and apparatus known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and apparatus should be considered part of the specification. In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values.

[0041] It should be noted that similar labels and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be discussed further in subsequent figures.

[0042] In the specification of this invention, the terms "first" and "second" may explicitly or implicitly include one or more of the same feature. In the description of this invention, unless otherwise stated, "multiple" means two or more. Furthermore, in the specification, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0043] like Figure 1 As shown in the figure, this embodiment introduces a comprehensive prevention and control method for near-distance coal seam mining based on multi-hazard collaborative early warning, including the following steps 1100-1600.

[0044] Step 1100: Based on the projection of the coal pillar in the overlying goaf, divide the mining area into the first region and the second region.

[0045] The first region is the area directly below the projection of the coal pillar of the overlying goaf and its outer edge, wherein the distance between the outer edge and the edge of the area directly below the projection of the overlying goaf is not less than a first threshold. The distance between the outer edge of the second region and the outer edge of the first region is not less than a second threshold.

[0046] The risks vary depending on the location. For example, below a coal pillar, gas is not the dominant risk factor. However, in locations far from the coal pillar, gas is the primary source of risk.

[0047] The first threshold can be set to 20m, and the second threshold can be set to 50m. The first region is the area directly below the projection of the coal pillar in the overlying goaf and its outer edge. The first region directly bears the maximum vertical stress transmitted by the coal pillar and is a high-incidence area for rock bursts and media compression damage. However, the gas is often squeezed out, and the amount of residual gas is low. Therefore, gas is not the main risk source in this region.

[0048] The second zone is located to the side or in front of the coal pillar. Due to the stress redistribution caused by mining, the medium is depressurized, fractures develop, and gas is easily desorbed and released, making it the main danger zone for coal and gas outbursts.

[0049] Step 1200: Deploy monitoring points in each area and install stress sensors, micro-vibration sensors and gas sensors at the monitoring points.

[0050] Monitoring points can be deployed in a grid-like manner within the first and second zones. The spacing can be set to 10-20 meters, and the sampling interval can be as short as minutes.

[0051] Stress sensors, microseismic sensors, and gas sensors are installed at each monitoring point. The stress sensors can be borehole stress gauges or hollow inclusion stress gauges, capable of recording full stress components or vertical stress. The microseismic sensors primarily acquire energy in the 0-100Hz range, while also covering the high-frequency band above 100Hz. The gas sensors are used for continuous monitoring of gas concentration.

[0052] Step 1300: Construct a baseline feature vector under normal mining conditions based on the stress change rate and microseismic energy release rate of the static monitoring points. The static monitoring points are located at least 100m from the outer edge of the first region.

[0053] The distance between the static monitoring point and the outer edge of the first area exceeds 100m. The static monitoring point is far from the coal pillar influence zone and is also equipped with stress sensors and microseismic sensors. The stress change rate at the static monitoring point is obtained using the stress sensor, and the microseismic energy release rate is obtained using the microseismic sensor.

[0054] The baseline eigenvector is used to describe the characteristics under normal mining conditions. The baseline eigenvector may include stress value, mean stress, mean rate of change of stress, standard deviation of stress, mean microseismic energy release rate, and standard deviation of microseismic energy release rate.

[0055] Stress values, mean stress, mean stress change rate, and standard deviation of stress can be calculated from data collected by stress sensors. The mean and standard deviation of microseismic energy release rates can be obtained from data collected by microseismic energy sensors. The mean microseismic energy release rate is calculated based on energy in the 0-100Hz frequency band; high-frequency energy is not included in the calculation because high frequencies typically correspond to mechanical noise or minor fractures and have a weak correlation with coal pillar fracturing.

[0056] To facilitate subsequent calculations, the baseline feature vector is normalized, scaling the value of each dimension to the [0, 1] interval.

[0057] Step 1400: Based on the feature vectors of each monitoring point in the second region and the baseline feature vector, obtain the first risk monitoring point set.

[0058] In the second region, monitoring points with significantly different stress microseismic characteristics compared to those under normal mining conditions were identified. These monitoring points typically indicate stress concentration or increased coal body fracturing due to coal pillar pressure, which may subsequently induce gas outbursts.

[0059] The first set of risk monitoring points is determined by calculating the Euclidean distance or Mahalanobis distance between the baseline eigenvectors.

[0060] For any monitoring point in the second region Its eigenvectors With reference eigenvectors deviation value Defined as:

[0061] ;

[0062] in, This indicates the calculation of Euclidean distance or Mahalanobis distance.

[0063] Within each time window (e.g., one hour or one shift), the deviation of all monitoring points in the second region is calculated to form a deviation value sequence. , This indicates the number of monitoring points in the second area.

[0064] Sort the sequence of deviation values ​​and take the highest value as the initial deviation threshold.

[0065] In the subsequent k consecutive time windows (e.g., 3-5 time windows), the average or median of the highest deviation values ​​in each time window is taken as the dynamic deviation threshold. .

[0066] The dynamic deviation threshold is recalculated for each updated time window. The dynamic deviation threshold is typically reached after 3-5 iterations. It is trending towards stability.

[0067] satisfy Monitoring points that show significant changes are then added to the first set of risk monitoring points. This method is sensitive to monitoring points with large changes. If the changes are small, they will not be marked, thus avoiding false alarms.

[0068] Step 1500: Based on the gas concentration sequence of each monitoring point in the first risk monitoring point set, use the gradient anomaly algorithm to screen out monitoring points with abnormal gas outbursts.

[0069] In actual underground mining operations, operations such as face advancement, coal cutting by the mining machine, support movement, and ventilation disturbances simultaneously affect stress, microseismic activity, and gas readings. These non-catastrophic changes are intertwined with the characteristic changes caused by coal pillar fracturing. If a multimodal joint vector is directly constructed, the model will be extremely unstable and it will be difficult to distinguish the source of risk (whether it is coal pillar rupture or mining machine operation). Therefore, this scheme first uses relatively stable stress and microseismic data with clear physical mechanisms to identify anomalies that may be caused by coal pillar pressure, i.e., the first set of risk monitoring points. Then, the gas concentration sequence is analyzed at these anomaly points, and a gradient anomaly algorithm is used to eliminate periodic interferences from the mining machine, highlighting gas outbursts truly induced by stress rupture.

[0070] The monitoring points for abnormal gas emission are obtained in the following way: based on the gas concentration increment, the first-order rate of change of gas concentration and the second-order rate of change of gas concentration within the second time window, the characteristic value of gas concentration change is calculated; if the characteristic value of gas concentration change corresponding to the monitoring point continuously exceeds the third threshold, the monitoring point is identified as a monitoring point for abnormal gas emission.

[0071] The second time window can be set to 5-15 minutes, and can be adjusted according to the coal cutting cycle or the face advance speed. The length of the second time window should be sufficient to capture the continuous upward trend of gas concentration, while avoiding the mixing of multiple cutting cycles. The characteristics of gas concentration change are calculated within each window.

[0072] monitoring points exist The gas concentration at that moment was In the second time window Within, the sequence is obtained by sampling at equal intervals. , This represents the maximum number of samples.

[0073] The gas concentration increment is the total increase in gas concentration within the second time window, expressed as:

[0074] ;

[0075] in, This indicates the increase in gas concentration.

[0076] The first-order rate of change of gas concentration is the average rate of gas release, expressed as:

[0077] ;

[0078] in, This represents the first-order rate of change of gas concentration. This represents the gas concentration at the k-th sampling point. This represents the gas concentration at the (k-1)th sampling point. The sampling interval is denoted as .

[0079] The second rate of change of gas concentration is the derivative of the first rate of change, i.e., the change in slope between two adjacent sampling intervals. The second rate of change of gas concentration indicates the accelerating trend of gas release; a positive value indicates that the fissure is spreading at an accelerated rate.

[0080] The characteristic value of gas concentration is obtained by weighting the gas concentration increment, the first-order rate of change of gas concentration, and the second-order rate of change of gas concentration. For example, the weight corresponding to the gas concentration increment can be set to 0.4, the weight corresponding to the first-order rate of change of gas concentration can be set to 0.3, and the weight corresponding to the second-order rate of change of gas concentration can be set to 0.3.

[0081] For monitoring points If, within a continuous K-time window (K≥3), the characteristic value of the gas concentration in each window exceeds the third threshold, then the monitoring point will be... The monitoring point was identified as having an abnormal gas outburst. This was to exclude transient, single-event gas fluctuations (usually lasting only 1-2 windows) caused by the passing of the coal mining machine, while retaining anomalies caused by stress-fracture-driven gas outbursts.

[0082] For example, the sampling interval is set to 1 minute, and the second time window... The timeframe is set to 5 minutes, meaning 5 sampling points. The characteristic values ​​of gas concentration at a certain monitoring point within four consecutive windows are 0.72, 0.85, 0.91, and 0.88, respectively, with a third threshold of 0.8. Therefore, if the characteristic value of gas concentration at this monitoring point exceeds the third threshold for three consecutive windows, this monitoring point is considered to have an abnormal gas emission.

[0083] The increase in gas concentration reflects the total release, avoiding sensitivity to minor fluctuations. The first-order rate of change of gas concentration can distinguish between continuous release (stable rate) and transient pulses (extremely high but brief rate). The second-order rate of change of gas concentration is the most predictive indicator. When coal pillar fracturing leads to unstable fracture propagation, the gas desorption rate often accelerates, while gas fluctuations caused by coal mining machine cutting are usually released at a uniform or decelerating rate. By combining these three indicators, non-hazardous interferences can be effectively filtered out, highlighting gas anomalies synchronized with coal pillar stress emissions.

[0084] Step 1600: Perform spatial cluster analysis on the selected monitoring points with abnormal gas emission, and conduct disaster early warning based on the morphology of the clusters obtained from the clustering.

[0085] The selected monitoring points with abnormal gas outbursts are projected onto the mining area plan for spatial clustering, such as DBSCAN. The clustering results reflect the spatiotemporal correlation of gas release, while isolated points or sporadic points (such as single sensor failures or localized minor disturbances) do not participate in early warning decisions.

[0086] Cluster morphologies include banded, linear, and diffuse. For banded clusters, the anomaly extends along a specific direction, with its length significantly greater than its width. The risk interpretation is that the gas release communicates with other layers (such as fractures in the overlying layer and adjacent layers), potentially leading to interlayer gas transmission or gas inrush from old goaf areas. Corresponding warning actions include immediately halting progress, checking the gas composition in the overlying goaf, and increasing ventilation.

[0087] For linear clusters, anomalies appear as narrow, straight bands. The risk interpretation is that gas is concentrated and released along a specific structural band (fault, densely jointed zone, coal pillar boundary), potentially indicating slippage or shear failure. Corresponding early warning actions include implementing pressure relief drilling or water injection in the banded area to reduce the risk of outbursts.

[0088] For diffuse clusters, the anomalies are randomly distributed and lack clear directionality. The risk interpretation is that most are measurement errors or minor localized fractures, and do not represent precursors to systemic disasters. Corresponding early warning actions include maintaining normal monitoring but increasing verification sampling.

[0089] This invention divides the mining area into a first zone and a second zone by projecting the coal pillar in the overlying goaf, and identifies the dominant risk factors in each zone: the first zone is characterized by stress concentration, while the second zone is characterized by gas outburst. Based on this, only the second zone undergoes a stepwise coordinated screening of stress-induced microseismic gas, avoiding misjudgments of insensitive gas indicators below the coal pillar.

[0090] In this embodiment, the reference feature vector further includes the envelope entropy reference distribution of the modal components; the modal components are obtained by variational mode decomposition of the stress signals collected from the static monitoring points; the envelope entropy reference distribution is the statistical distribution of the envelope entropy of each modal component.

[0091] In underground mining operations, mechanical activities such as coal cutting by mining machines, operation of conveyors, and movement of supports generate periodic vibrations, which are superimposed on the actual stress sensor signals. Traditional methods require complex filtering or independent component analysis to separate mechanical interference, which is not only computationally intensive but also prone to filtering out true stress precursors (stress fluctuations before coal pillar rupture are often non-stationary).

[0092] Envelope entropy is a measure of the complexity and randomness of a signal. Signals with strong periodicity and regularity, such as mechanical vibrations, have low envelope entropy. Signals with strong randomness and complexity have high envelope entropy, such as stress fluctuations before coal and rock fracturing or stress pulsations caused by micro-fractures.

[0093] Therefore, directly calculating the envelope entropy of the stress signal can distinguish between mechanical disturbances and disaster precursors without preprocessing. Incorporating this index into the baseline feature vector is equivalent to adding a dimension of complexity to the stress signal.

[0094] Variational mode decomposition (VMD) is performed on stress signals acquired from static monitoring points, adaptively decomposing them into multiple modal components. Compared with empirical mode decomposition (EMD), VMD has a more rigorous mathematical foundation, less modal aliasing, and is more robust to noise. The number of modes can be automatically determined by observing the center frequency or using the modal overlap criterion; typically, the number of modes is set to 4-8.

[0095] For each modal component, its analytic signal is obtained through Hilbert transform, yielding the envelope signal. The envelope signal is then normalized to a probability distribution, and the envelope entropy is calculated based on the probability distribution.

[0096] Under normal mining conditions, the envelope entropy of each mode of the stress signal follows a certain distribution. When the coal pillar pressure causes precursors to rupture, the envelope entropy deviates significantly from this distribution. Specifically, the sample entropy of the envelope spectrum or the permutation entropy of the envelope can be used. Periodic signals have low permutation entropy, while random signals have high permutation entropy.

[0097] For example, phase space reconstruction is performed on the envelope signal sequence to calculate the permutation entropy. For periodic mechanical vibrations, the permutation entropy is low due to the repetition of their modes. For random rupture precursors, the permutation entropy is high due to the disordered modes.

[0098] At static monitoring points, stress signals under normal mining conditions are collected over a long period of time. The envelope entropy of all modal components can be calculated every 5 minutes to obtain the statistical distribution of the envelope entropy of each modal component, including the mean, standard deviation, quantiles, etc. This is the baseline distribution of the envelope entropy of the modal components.

[0099] The expanded reference eigenvector contains the reference distribution of envelope entropy for each modal component. For any monitoring point in the second region, its stress signal is also acquired, and variational mode decomposition is performed to obtain multiple modal components. The envelope entropy is calculated, and the envelope entropy deviation term is also calculated. The envelope entropy deviation term is then added to the aforementioned deviation value function. In this context, it is used to calculate the baseline feature vector.

[0100] Mechanical vibrations possess a fixed fundamental frequency and harmonics, and the envelope of their stress signal modal components exhibits significant periodicity with low envelope entropy. Under normal mining conditions, even with mechanical vibrations, their envelope entropy remains stable within a low range, forming part of the baseline distribution. When coal pillar pressure causes coal body fracturing, random transient pulses appear in the stress signal, disrupting the periodicity and significantly increasing the envelope entropy, thus deviating from the baseline. Therefore, without prior filtering of mechanical vibrations, directly comparing changes in envelope entropy can distinguish between normal mechanical background and disaster precursors.

[0101] In this embodiment, the reference feature vector further includes a multi-field coupling correlation factor between stress and microseismic sequence; the process of obtaining the multi-field coupling correlation factor includes: within a first time window, performing time-delay cross-correlation calculation on the stress change rate and microseismic energy release rate obtained based on static monitoring points; and taking the largest time-delay cross-correlation coefficient within a preset time delay window as the multi-field coupling correlation factor.

[0102] Under the influence of the pressure from the overlying coal pillar, the underlying coal body first undergoes elastic compression and stress concentration (stress sensors respond first). Subsequently, when the stress exceeds the local strength, the coal body generates micro-fractures, releasing elastic waves (micro-seismic sensors respond later). This process involves a physical lag time, the magnitude of which depends on the elastic wave velocity of the coal and rock (typically several kilometers per second, with negligible propagation time), the fracture incubation time (the delay between stress rise and fracture occurrence, typically on the order of seconds or minutes), and the spacing between sensor deployments relative to the geological structure.

[0103] Therefore, stress change preceding microseismic release is the normal sequence of disaster evolution. If the correlation between the two is weak or the sequence is reversed (e.g., microseismic release precedes stress change), it may indicate non-catastrophic vibrations. Non-catastrophic vibrations can include coal mining machine cutting, far-field blasting, etc.

[0104] Two sequences were continuously collected at the static monitoring point: a stress change rate sequence and a stress change rate sequence. and microseismic energy release rate sequence Using streaming computation, the two sequences are first standardized using Z-scores:

[0105] ;

[0106] ;

[0107] in, This is the standardized stress change rate sequence. The mean value of the rate of change of stress. The standard deviation of the rate of change of stress. This is the standardized microseismic energy release sequence. This represents the average energy release rate of microseismic events. This represents the variance of the microseismic energy release rate. Standardization eliminates the influence of dimensions, ensuring that cross-calculations only reflect waveform similarity.

[0108] The sampling time interval is In the first window Internally, calculate the time delay:

[0109] ;

[0110] ;

[0111] in, Due to time lag, This indicates the number of samples taken within the first time window. It is an integer, and its value range is 1. , This is the preset maximum physical lag time.

[0112] Sampling time interval Typically set to 1 minute, number of samples The value is set to 5-10, meaning the first time window length is 5-10 minutes. This window length is sufficient to encompass a complete stress-fracture-microseismic cycle (coal pillar pressure propagation, coal body rupture, elastic wave release) without smoothing out short-term dynamic characteristics. Time delay search range. Typically set to Or even less, because physical lag is usually no more than a few minutes.

[0113] The time-delay cross-correlation coefficient is:

[0114] ;

[0115] in, Let be the cross-correlation coefficient with time delay, when This indicates that the stress sequence comes first, followed by the microseismic sequence.

[0116] Within the preset delay time window Within this range, find the largest time-delay cross-correlation coefficient and its corresponding delay time:

[0117] ;

[0118] ;

[0119] in, The maximum time-delay cross-correlation coefficient, This represents the delay time corresponding to the maximum time-delay cross-correlation coefficient. The multi-field coupling correlation factor is precisely the maximum time-delay cross-correlation coefficient. .

[0120] The closer the multi-field coupling correlation factor is to 1, the stronger the linear correlation between stress changes and microseismic release, and the greater the time delay. Stability indicates that the pressure transmission of the coal pillar and the rupture response are well coupled (there may also be some coupling during normal mining, but the coupling will be significantly enhanced before a disaster).

[0121] A multi-field coupling correlation factor close to 0 or negative indicates that the two fields are unrelated or inverted, which may be due to mechanical interference, sensor failure, or non-pressure-dominated vibration.

[0122] This embodiment calculates the multi-field coupling correlation factor between stress and microseismic sequences, using it as a feature in the baseline feature vector. This respects the physical timeline of stress-rupture and avoids misjudging mechanical disturbances as precursors. Furthermore, Z-score normalization and cross-correlation calculations are insensitive to amplitude fluctuations, focusing only on waveform similarity, thus exhibiting strong noise resistance. Additionally, increases or decreases in the correlation coefficient have clear physical meanings, providing strong interpretability.

[0123] The spatial clustering analysis uses the DBSCAN algorithm, and the number of members in the minimum cluster is determined based on the deployment density of the sensors.

[0124] Sensor deployment density is divided into three levels: sparse, normal, and high density.

[0125] For the sparse level, the sensor spacing exceeds 20m, and the minimum number of members in a cluster is 2-3. For the normal level, the sensor spacing is 10-20m, and the minimum number of members in a cluster is 3-4. For the high-density level, the sensor spacing is less than 10m, and the minimum number of members in a cluster is 5-9.

[0126] The DBSCAN algorithm marks points that do not meet the minimum cluster member condition as noise. These isolated or sporadic gas outburst anomalies are not included in the early warning system. They may be caused by non-hazardous factors such as sensor drift, brief passage of the coal mining machine, or release from localized small extraction boreholes. Only clustered anomalies reflect spatially correlated gas releases induced by the same mechanical mechanism (such as coal pillar fracturing zones).

[0127] In this embodiment, principal component analysis is used to obtain the geometric features of the anomalous clusters, and the risk type is determined based on the geometric features of the anomalous clusters.

[0128] The geometric feature includes elongation; the calculation process of the elongation includes: obtaining the covariance matrix formed by each monitoring point within the anomaly cluster; calculating the first eigenvalue and the second eigenvalue of the covariance matrix; and determining the risk type based on the first eigenvalue and the second eigenvalue.

[0129] When the ratio of the first feature value to the second feature value is greater than 3, the risk type is determined to be a stress concentration zone induced by the edge of a coal pillar; when the ratio is between 1.1 and 1.5, the risk type is determined to be a block fracture risk caused by pressure.

[0130] If the ratio of the convex hull area of ​​the abnormal cluster to the number of points within the abnormal cluster is greater than a fourth threshold, the monitoring point data within the abnormal cluster is determined to be abnormal data.

[0131] When the DBSCAN algorithm identifies a cluster of anomalous gas outbursts in the second region, this cluster contains There are 1 monitoring point, each with planar coordinates. Since the variation in the coal seam thickness direction is small, it can be ignored. These points are then used to construct a coordinate matrix. :

[0132] ;

[0133] First, the coordinates are decentered (by subtracting the mean), then the covariance matrix is ​​calculated:

[0134] ;

[0135] ;

[0136] in, Let covariance matrix be the variance matrix. This represents the geometric center of the cluster.

[0137] Find the first eigenvalue of the covariance matrix Second eigenvalue and its corresponding eigenvectors. The principal direction of the corresponding cluster, the direction with the largest variance. The secondary direction of the corresponding cluster is perpendicular to the primary direction.

[0138] The elongation is the ratio of the first eigenvalue to the second eigenvalue, expressed as:

[0139] ;

[0140] in, Elongation. When the elongation is close to 1, the point set appears circular or square (isotropic). When the elongation is much greater than 1, the point set appears as a thin strip or line (anisotropic).

[0141] If the elongation is greater than 3, the cluster's geometric morphology exhibits a distinct strip-like or linear shape. Physically, this means that gas outburst points extend linearly along the coal pillar edge, fault, or fracture zone. The coal pillar edge is the location of stress concentration and maximum shear stress, easily forming a strip-like fracture channel that connects to gas from the overlying goaf. The corresponding risk type is a stress concentration zone induced by the coal pillar edge, indicating high risk and the possibility of coal and gas outbursts or rockbursts.

[0142] If the elongation is between 1.1 and 1.5, the cluster's geometry is approximately circular or elliptical. Physically, this means that gas outburst points are densely distributed within a localized block, with no dominant direction. This typically corresponds to a pressure-bearing fracture zone directly beneath or to the side of a coal pillar, where the coal body undergoes overall compression fracture and gas is uniformly desorbed. The corresponding risk type is pressure-induced block fracture risk, indicating moderate danger, with the possibility of spalling or localized outbursts.

[0143] If the elongation is less than 1.1, it will appear as an almost perfect circle. This is mostly due to measurement error or isolated noise points, and can be classified as diffuse, requiring no special action.

[0144] If the elongation is between 1.5 and 3, the geometric shape is transitional, possibly a transition zone between the edge of a coal pillar and a blocky fracture. It needs to be judged in conjunction with other indicators (such as microseismic energy), and monitoring needs to be strengthened.

[0145] For anomaly clusters, calculate their convex hull area, which is the area of ​​the smallest convex polygon that contains all points.

[0146] For normal clusters, the monitoring points are relatively evenly distributed, the convex hull area is directly proportional to the number of points in the cluster, and the ratio of the convex hull area to the number of points in the cluster is stable within a certain range, such as 20-40.

[0147] If the ratio of the convex hull area to the number of points within a cluster is significantly large, it indicates that although the number of points is small, the convex hull area is large, meaning the points are very sparse. This usually means that these outliers are not caused by the same mechanical mechanism, but rather that randomly occurring isolated points have been incorrectly merged into a cluster. Alternatively, it could be due to partial sensor aging, causing a few points that are far apart to still be classified as a cluster.

[0148] This invention utilizes sufficiently distant static monitoring points within the coal pillar projection area to construct a benchmark feature vector that includes stress statistics, microseismic energy release rate, and optional multi-field coupling correlation factors and modal component envelope entropy distributions. All benchmark distributions employ adaptive thresholds (dynamically calculated based on recent data to determine the highest deviation or quantile), eliminating the need for manual pre-setting and automatically adapting to the mining backgrounds of different working faces and mining stages.

[0149] This invention employs a two-step approach instead of directly constructing a multimodal vector: first, stress and microseismic activity are used to identify anomalous locations that may be caused by coal pillar pressure; then, only for these locations, a gas concentration gradient anomaly algorithm (gas concentration increment, first-order rate of change, and second-order rate of change continuously exceeding thresholds) is used to screen for outburst anomalies. This method effectively eliminates instantaneous gas fluctuations caused by periodic cutting by the coal mining machine, highlighting the continuous gas release synchronized with coal pillar rupture.

[0150] This invention performs DBSCAN spatial clustering on abnormal gas outburst locations and adaptively sets the neighborhood radius and minimum cluster size based on sensor deployment density. Only valid clusters, rather than isolated or sporadic points, participate in early warning decisions, significantly reducing false alarms caused by single-point failures or localized minor disturbances.

[0151] This invention performs principal component analysis on each anomalous cluster, extracting the elongation rate and the ratio of convex hull area to the number of points, and determines the risk type based on this ratio. Simultaneously, clusters with a convex hull area to point ratio exceeding a fourth threshold are automatically identified as anomalous sensor data clusters and removed. This quantitative method achieves automated and interpretable output of risk types.

[0152] This invention introduces a multi-field coupling correlation factor to examine the physical causal chain where stress precedes micro-seismic events. If the coupling factor deviates significantly from the baseline (too high or too low), an anomaly is indicated; if the coupling relationship is reversed (micro-seismic events precede), it is determined to be mechanical interference or non-catastrophic vibration. Furthermore, the optional high-frequency modal component envelope entropy distribution further distinguishes between periodic mechanical vibrations and random precursors of rupture from the perspective of the complexity of the stress signal itself, without requiring additional filtering preprocessing.

[0153] While specific embodiments of the present invention have been described in detail by way of examples, those skilled in the art should understand that the above examples are for illustrative purposes only and are not intended to limit the scope of the invention. Those skilled in the art should understand that modifications can be made to the above embodiments without departing from the scope and spirit of the invention.

[0154] Those skilled in the art will recognize that the modules and algorithm steps described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software 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 implementations should not be considered beyond the scope of this invention.

[0155] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the above-described apparatus and equipment can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0156] In the 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 is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0157] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; that is, they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of the embodiments of the present invention, depending on actual needs.

[0158] In addition, the functional modules in the embodiments of the present invention can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0159] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. 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.

[0160] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the inventive concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this application.

[0161] It should be understood that the sequence numbers of the steps in the invention's content and embodiments do not absolutely imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention. The foregoing description of embodiments of this disclosure has been provided for illustrative and descriptive purposes. The foregoing description is not exhaustive and is not intended to limit this disclosure to the exact form disclosed. Various modifications and variations may exist based on the foregoing teachings, or various modifications and variations may be derived from the practice of this disclosure. These embodiments were chosen and described to illustrate the principles of this disclosure and its practical application, so that those skilled in the art can utilize this disclosure in various implementations and modifications suitable for the specific purpose of the concept.

Claims

1. A method for comprehensive prevention and control of close coal seam mining based on multi-disaster collaborative early warning, characterized in that, include: Based on the projection of the coal pillar in the overlying goaf, the mining area is divided into the first area and the second area. Monitoring points are deployed in each area, and stress sensors, micro-vibration sensors, and gas sensors are installed at these monitoring points. Based on the stress change rate and microseismic energy release rate of static monitoring points, a benchmark feature vector is constructed under normal mining conditions. Based on the feature vectors of each monitoring point in the second region and the baseline feature vector, a first set of risk monitoring points is obtained; Based on the gas concentration sequence of each monitoring point in the first risk monitoring point set, the gradient anomaly algorithm is used to screen out monitoring points with abnormal gas emission. Spatial cluster analysis was performed on the selected monitoring points with abnormal gas outbursts, and disaster early warning was made based on the morphology of the clusters obtained from the clustering. Wherein, the first region is the region directly below the projection of the coal pillar of the overlying goaf and its outer edge region, and the distance between the outer edge region and the edge of the region directly below the projection is not less than a first threshold. The outer edges of the second region and the first region are not less than the second threshold. The static monitoring point is located at least 100m from the outer edge of the first area.

2. The multi-disaster collaborative early warning based comprehensive prevention and control method for close-range coal seam mining according to claim 1, characterized in that, The benchmark feature vector also includes the benchmark distribution of the envelope entropy of the modal components; The modal components are obtained by performing variational mode decomposition on the stress signals collected from the static monitoring points; The envelope entropy baseline distribution is the statistical distribution of the envelope entropy of each modal component.

3. The multi-disaster collaborative early warning based close-range coal seam mining comprehensive prevention and control method according to claim 1, characterized in that, The benchmark feature vector also includes a multi-field coupling correlation factor between stress and microseismic sequence; The process of obtaining the multi-field coupling correlation factor includes: Within the first time window, time-delay cross-correlation calculations are performed on the stress change rate and microseismic energy release rate obtained from static monitoring points; The maximum time delay cross-correlation coefficient within the preset time delay window is used as the multi-field coupling correlation factor.

4. The comprehensive prevention and control method for near-distance coal seam mining based on multi-hazard collaborative early warning as described in claim 1, characterized in that, The first set of risk monitoring points is determined by calculating the Euclidean distance or Mahalanobis distance between the baseline eigenvectors.

5. The comprehensive prevention and control method for near-distance coal seam mining based on multi-hazard collaborative early warning as described in claim 1, characterized in that, The monitoring points for the gas outburst anomaly were obtained through the following methods: Calculate the characteristic value of gas concentration change based on the gas concentration increment, the first-order rate of change of gas concentration, and the second-order rate of change of gas concentration within the second time window; If the characteristic value of the gas concentration change corresponding to the monitoring point continuously exceeds the third threshold, the monitoring point will be identified as a monitoring point with abnormal gas emission.

6. The comprehensive prevention and control method for near-distance coal seam mining based on multi-hazard collaborative early warning as described in claim 1, characterized in that, The spatial clustering analysis uses the DBSCAN algorithm, and the number of members in the minimum cluster is determined based on the deployment density of the sensors.

7. The comprehensive prevention and control method for near-distance coal seam mining based on multi-hazard collaborative early warning as described in claim 1, characterized in that, Principal component analysis is used to obtain the geometric characteristics of the anomalous clusters, and the risk type is determined based on the geometric characteristics of the anomalous clusters.

8. The comprehensive prevention and control method for near-distance coal seam mining based on multi-hazard collaborative early warning as described in claim 7, characterized in that, The geometric features include elongation; The calculation process for the elongation includes: Obtain the covariance matrix formed by the monitoring points within the anomaly cluster; Calculate the first and second eigenvalues ​​of the covariance matrix; The risk type is determined based on the first feature value and the second feature value; Elongation is the ratio of the first characteristic value to the second characteristic value.

9. A comprehensive prevention and control method for near-distance coal seam mining based on multi-hazard collaborative early warning, as described in claim 8, characterized in that, In response to the ratio of the first feature value to the second feature value being greater than 3, the risk type is determined to be a stress concentration zone induced by the edge of the coal pillar; In response to the ratio being between 1.1 and 1.5, the risk type is determined to be block fracture risk due to pressure.

10. A comprehensive prevention and control method for near-distance coal seam mining based on multi-hazard collaborative early warning, as described in claim 8, characterized in that, If the ratio of the convex hull area of ​​the abnormal cluster to the number of points within the abnormal cluster is greater than a fourth threshold, the monitoring point data within the abnormal cluster is determined to be abnormal data.

Images