Method for predicting geological abnormal area of fully mechanized coal mining face based on multi-source data

Abstract

The invention relates to the field of mine geophysical exploration, and discloses a fully mechanized coal mining face geological abnormal area prediction method based on multi-source data, and the method comprises the steps: carrying out the detection of a fully mechanized coal mining face through a radio pit penetrating device and a slot wave seismograph, and obtaining a working face CT image and a slot wave energy attenuation coefficient image; analyzing a distribution rule of a plurality of abnormal prompt areas in the working face CT image, and delimiting a plurality of candidate abnormal areas to obtain an average abnormal amplitude value of each candidate abnormal area; analyzing a plurality of slot wave attenuation high value abnormal regions in the slot wave energy attenuation coefficient imaging, and carrying out overlapping region calculation on the plurality of candidate abnormal regions to obtain a plurality of overlapping regions; carrying out drilling point arrangement and drilling according to the plurality of overlapped areas, and measuring a plurality of drilling trajectories and drilling data corresponding to each drilling point through a trajectory instrument; and according to the plurality of drilling trajectories and the drilling data corresponding to each drilling point, abnormal region prediction data is generated, so that accurate prediction of the abnormal region is realized.

Description

Technical Field

[0001] The present invention relates to the field of mine geophysical exploration, and particularly to a prediction method for geological abnormal areas in fully mechanized coal mining faces based on multi-source data. Background Art

[0002] In fully mechanized coal mining faces, the accurate detection of geological anomalies is the key to ensuring safe production. Currently, commonly used methods include: surface geophysical exploration techniques (such as electrical methods and seismic exploration), but the resolution is relatively low, and it is difficult to accurately reflect small-scale anomalies underground. Drilling and sampling can directly obtain geological information, but when the layout of drilling points is unreasonable, the cost is high and the efficiency is low. Trough wave detection can detect the internal structure of coal seams and is applicable to the coal mine environment, but the detection results are uncertain due to factors such as coal seam thickness and water content. Radio wave penetration technology in mines uses radio waves to penetrate coal seams to detect underground structure information and can be used for the identification of abnormal areas, but the accuracy is affected by signal attenuation. Therefore, there is an urgent need for an effective means to solve the problems of low accuracy of single detection means and poor accuracy of abnormal area prediction.

[0003] During the process of detecting abnormal areas in the existing technology, the following technical problems often exist: First, the existing detection means are single, and errors are likely to occur in the identification of abnormal areas; moreover, the layout of traditional drilling points depends on experience or a single geological model, resulting in problems such as unreasonable layout of drilling points and low drilling accuracy, leading to inaccurate and unreliable identification of abnormal areas; Second, there are differences in the spatial resolution and data dimensions of different detection means, making it difficult to directly superimpose and analyze; moreover, geological anomalies often show discrete and fragmented characteristics in detection data, resulting in blurred boundaries of abnormal areas; and traditional static division of abnormal areas cannot reflect dynamic risks. Summary of the Invention

[0004] This part of the content of the present invention is used to briefly introduce the concepts, which will be described in detail in the subsequent detailed implementation part. This part of the content of the present invention does not aim to identify the key features or essential features of the claimed technical solution, nor is it intended to limit the scope of the claimed technical solution.

[0005] The present invention proposes a prediction method for geological abnormal areas in fully mechanized coal mining faces based on multi-source data to solve one or more of the technical problems mentioned in the above background art part.

[0006] The present invention provides a prediction method for geological abnormal areas in fully mechanized coal mining faces based on multi-source data, including: respectively using radio wave penetration equipment in mines and trough wave seismographs to detect the fully mechanized coal mining face, and obtaining the CT map of the working face and the imaging of the trough wave energy attenuation coefficient; Analyze the distribution law of multiple abnormal prompt areas in the CT image of the working face, delimit multiple candidate abnormal areas, and obtain the average abnormal amplitude value of each candidate abnormal area; Analyze multiple high-value abnormal areas of slot wave attenuation in the imaging of slot wave energy attenuation coefficient, and calculate the overlapping areas with multiple candidate abnormal areas to obtain multiple overlapping areas; Arrange and drill boreholes according to multiple overlapping areas. During the drilling process, measure multiple drilling trajectories and drilling data corresponding to each drilling point through a trajectory instrument; Generate prediction data of abnormal areas in the fully-mechanized mining face according to multiple drilling trajectories and drilling data corresponding to each drilling point.

[0007] Optionally, analyzing the distribution law of multiple abnormal prompt areas in the CT image of the working face, delimiting multiple candidate abnormal areas, and obtaining the average abnormal amplitude value of each candidate abnormal area includes: Calculate the distance between each abnormal prompt area and the nearest neighboring abnormal prompt area and record it as the minimum adjacent distance; according to the minimum adjacent distance of each abnormal prompt area, determine the aggregation area of abnormal prompt areas, generate the radiation range coefficient corresponding to the aggregation area of abnormal prompt areas, and radiate outward from the aggregation area of abnormal prompt areas according to the corresponding radiation range coefficient to obtain candidate abnormal areas; Statistically analyze the abnormal amplitude values of all abnormal prompt areas within each candidate abnormal area, and calculate the average abnormal amplitude value of each candidate abnormal area.

[0008] Optionally, analyzing multiple high-value abnormal areas of slot wave attenuation in the imaging of slot wave energy attenuation coefficient includes: If there is an area in multiple high-value abnormal areas of slot wave attenuation whose area range is less than the preset range and the distance between the center point of the abnormal area and the roadway opening is less than the preset distance, then delete the corresponding high-value abnormal area of slot wave attenuation, and conduct a joint analysis with single-shot seismic records and actual geological information to obtain the actual position information, influence range, regional shape, and abnormal type corresponding to each remaining high-value abnormal area of slot wave attenuation.

[0009] Optionally, deleting the corresponding high-value abnormal area of slot wave attenuation and conducting a joint analysis with single-shot seismic records and actual geological information includes: Conduct a spectral analysis on the waveform amplitude and frequency characteristics of single-shot seismic records, and extract the abnormal characteristics of the seismic wave reflection interface; spatially match the abnormal characteristics of the reflection interface with the known fault and lithology boundary data in the geological information to generate a matching degree score; if the matching degree score is lower than the preset threshold, then determine that the high-value abnormal area of slot wave attenuation is interference signal and delete it; conduct boundary smoothing processing and regional connectivity analysis on the remaining high-value abnormal areas of slot wave attenuation to generate the final set of effective abnormal areas.

[0010] Optionally, perform spatial matching between the abnormal characteristics of the reflection interface and the known faults and lithological boundary data in the geological information to generate a matching degree score, including: Establish a mapping relationship table between seismic reflection characteristics and geological anomaly types; dynamically adjust the matching degree weights according to the mapping relationship table, and verify the credibility of the matching results through a convolutional neural network.

[0011] Optionally, calculate the distance between each abnormal prompt area and the nearest neighboring abnormal prompt area, including: Judge its spatial dispersion degree by measuring the maximum distance between the boundaries of the abnormal prompt area; when the dispersion degree exceeds the preset threshold, mark this abnormal prompt area as an independent abnormal area; perform boundary fusion on adjacent abnormal areas with a dispersion degree lower than the preset threshold to generate a continuous candidate abnormal area.

[0012] Optionally, the boundary fusion includes: establishing a vector boundary buffer for the abnormal area, and when the buffers of adjacent abnormal areas overlap, using the minimum bounding rectangle algorithm to generate a fused continuous candidate abnormal area, and retaining the extreme points of the abnormal amplitude of the original abnormal area as feature markers.

[0013] Optionally, a method for predicting geological abnormal areas in fully mechanized coal mining faces based on multi-source data according to the present invention further includes: After the continuous candidate abnormal area is generated, obtain the top and bottom plate shape data of the abnormal area through a three-dimensional laser scanner, perform spatial registration on the scanned point cloud and the results of the trough wave imaging, correct the three-dimensional spatial shape of the candidate abnormal area, and mark the potential danger level.

[0014] The present invention has the following beneficial effects: 1. The prediction accuracy is improved. Specifically, by using a radio pit penetration device and a trough wave seismograph for large-scale and non-destructive detection to obtain the distribution information of geological anomalies, the detection accuracy of abnormal areas is improved. The radio pit penetration device provides CT imaging data, and the trough wave seismograph provides energy attenuation coefficient imaging, which respectively reflect different characteristics of underground anomalies (such as density changes and the hardness of rock layers). By fusing the data of the two detection methods, the accuracy of abnormal area identification can be improved, and the influence of errors of a single detection method can be avoided. And based on the superposition analysis of multi-source data (CT abnormal area and trough wave abnormal area), calculate the overlapping area, accurately plan the drilling points, and improve the effectiveness of drilling. Record the actual drilling trajectory of each drilling point through a trajectory instrument, and combine the geological parameters obtained during the drilling process to construct a more accurate three-dimensional abnormal area model, dynamically optimize the abnormal area prediction, and improve the prediction accuracy.

[0015] 2. The recognition accuracy and reliability of abnormal areas are improved. Specifically, through multi-source data fusion (such as CT images, slot wave imaging, geological information, etc.) and spatial registration technology, the recognition accuracy and reliability of abnormal areas can be effectively improved. Reducing the influence of interference signals: By setting distance rules to delete small-scale slot wave attenuation high-value abnormal areas that are close to the roadway opening, the influence of interference signals on the prediction of geological abnormal areas is reduced. Combining 3D laser scanning data to correct the 3D spatial morphology of abnormal areas can more accurately reflect the actual geological conditions within the working face. By marking potential danger levels, effective risk assessment and early warning can be provided for fully mechanized mining faces, helping to formulate safer operation plans. By automatically calculating parameters such as the minimum adjacent distance, abnormal amplitude, and radiation range of abnormal areas through algorithms, as well as automated boundary fusion and matching degree scoring, the efficiency of abnormal area analysis and prediction is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In combination with the accompanying drawings and with reference to the following specific embodiments, the above and other features, advantages, and aspects of the embodiments of the present invention will become more apparent. Throughout the accompanying drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic and the elements and elements are not necessarily drawn to scale.

[0017] Figure 1 is a flowchart of a method for predicting geological abnormal areas in a fully mechanized mining face based on multi-source data of the present invention; Figure 2 is a schematic diagram of the formation principle of slot waves of a method for predicting geological abnormal areas in a fully mechanized mining face based on multi-source data of the present invention; Figure 3 is a schematic diagram of the coordinate system for detecting slot waves in a working face of a method for predicting geological abnormal areas in a fully mechanized mining face based on multi-source data of the present invention; Figure 4 is a schematic diagram of candidate abnormal areas of a method for predicting geological abnormal areas in a fully mechanized mining face based on multi-source data of the present invention; Figure 5 is a schematic diagram of slot wave attenuation high-value abnormal areas of a method for predicting geological abnormal areas in a fully mechanized mining face based on multi-source data of the present invention; Figure 6 is a schematic diagram of the drilling trajectory of a method for predicting geological abnormal areas in a fully mechanized mining face based on multi-source data of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] The present invention will be described in more detail below with reference to the accompanying drawings. Although some embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. On the contrary, these embodiments are provided to more thoroughly and completely understand the present invention. It should be understood that the drawings and embodiments of the present invention are only for exemplary purposes and are not used to limit the protection scope of the present invention.

[0019] In addition, it should be noted that for ease of description, only the parts related to the relevant invention are shown in the drawings. Without conflict, the embodiments in the present invention and the features in the embodiments can be combined with each other.

[0020] It should be noted that the concepts such as "first" and "second" mentioned in the present invention are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence relationship of the functions performed by these devices, modules or units.

[0021] It should be noted that the modifications of "one" and "multiple" mentioned in the present invention are illustrative rather than restrictive. Those skilled in the art should understand that unless otherwise clearly specified in the context, it should be understood as "one or more".

[0022] The names of the messages or information exchanged between multiple devices of the present invention are only for illustrative purposes and are not used to limit the scope of these messages or information.

[0023] The present invention will be described in detail below with reference to the accompanying drawings and in combination with embodiments.

[0024] As Figure 1 shown, a flowchart of a method for predicting geological abnormal areas in fully-mechanized coal mining faces based on multi-source data of the present invention is shown, which specifically includes the following steps: Step 101, respectively use a radio penetration device and a shear wave seismograph to detect the fully-mechanized coal mining face, and obtain a working face CT map and a shear wave energy attenuation coefficient imaging.

[0025] In some embodiments, the execution entity of a prediction method for geological abnormal areas in fully mechanized coal mining faces based on multi-source data of the present invention can be a back-end server. A fully mechanized coal mining face refers to a coal mining face equipped with fully mechanized equipment in underground coal mines. When mining, for safety reasons, it is generally mined from front to back. First, the roadway is dug through to the far end of the planned mining area. The working face during formal mining is called the coal mining face. The working face refers to the front-line area directly for coal mining operations in underground coal mines and is the core place for coal production. The fully mechanized equipment includes equipment such as coal shearers, hydraulic supports, and scraper conveyors. The radio penetration equipment is an instrument for underground detection based on the principle of radio waves. It sends electromagnetic waves (radio waves) underground and then receives the electromagnetic wave signals returned from underground media, thereby imaging and analyzing the underground structure. On this basis, the radio penetration equipment sends electromagnetic waves of a specific frequency underground. When the electromagnetic waves propagate in underground media, they will be reflected, refracted, or absorbed by different substances (such as rock layers, cavities, cracks, etc.). The electromagnetic properties (such as dielectric constant, conductivity) of different underground substances determine the propagation speed and attenuation characteristics of electromagnetic waves. The equipment receives the returned electromagnetic wave signals and analyzes the properties of different underground areas through parameters such as the intensity and propagation time of the signals, generates corresponding CT images and sends them to the back-end server. These images can reflect the abnormal areas of the underground structure, such as broken rock layers, cracks, cavities, etc. The CT image is essentially a two-dimensional visualization expression of the physical property parameters of underground media (such as electromagnetic wave absorption coefficient, seismic wave attenuation coefficient, etc.). The trough wave seismograph can be the KDZ31114 mine distributed shock wave exploration instrument. Among them, the KDZ31114 mine distributed shock wave exploration instrument is composed of a data recorder, a blasting start recorder, and sensors, etc. The data recorder and the blasting start recorder are synchronized using a high-precision clock. After synchronous timekeeping underground, all devices will work independently with the same time rhythm. The data recorder continuously collects and stores seismic data without interruption. The blasting start recorder records the detonator initiation time. After all excitation work is completed, data recovery is carried out. All devices form a bus network, and the seismic traces are "cut out" from the data records with the blasting start time of the blasting recorder as 0, and a single-shot record is formed.

[0026] In some embodiments, a trough wave seismograph is a device specifically used for exciting and receiving trough wave signals. On this basis, the trough wave seismograph usually uses explosions as the seismic source. In the middle of the roadway coal wall (about 1.4 m from the roadway floor), a shot point is arranged every 20 m along the roadway direction as the seismic source point, and a geophone (i.e., sensor) is arranged every 10 m on the working face. The geophone is connected to the bolt in the middle of the coal seam, and a adapter is used to fix the geophone on the bolt. The installation direction of the geophone is parallel to the dip of the coal seam, perpendicular to the coal wall and pointing into the working face, and the connection directions of all geophones are the same. At the seismic source point, an explosion is set off to excite the trough wave. The trough wave propagates along the coal seam or the low-velocity layer and is affected by factors such as coal seam thickness, fractures, and water content. If there are anomalies inside the coal seam, such as cracks, faults, goafs, or water-bearing areas, the trough wave will undergo energy attenuation, velocity change, or scattering, forming abnormal signals. The geophone receives the trough wave signal, converts the trough wave signal into an electrical signal and sends it to the data recorder. The data recorder records and stores the trough wave signal, preprocesses the trough wave signal to obtain trough wave data and sends it to the background server. Among them, the trough wave is a seismic wave that propagates between underground rock interfaces and usually appears in layered media such as coal seams. In the coal-bearing strata, the coal seam is a relatively low-velocity seismic trough. The interface between the coal seam and its surrounding rock generally presents a good reflecting surface. Compared with its roof and floor surrounding rocks, the coal seam always appears with a low velocity, low density, and thus a low wave impedance. When a seismic wave is excited in the coal seam, the longitudinal wave and transverse wave excited both spread around the seismic source in the form of spherical body waves and enter the roof and floor interfaces at different angles, as Figure 2 shown, which shows a schematic diagram of the trough wave formation principle. When the incident angle is less than the critical angle, most of the energy will penetrate into the surrounding rock, and only a small part of the energy will be reflected back into the coal seam. The energy returning to the coal seam will be quickly attenuated (leakage mode) due to multiple reflections and transmissions in the coal seam; when the incident angle is greater than or equal to the critical angle, the seismic wave energy incident on the roof and floor interfaces will be completely reflected back into the coal seam and reflected multiple times in the coal seam, and finally be confined in the coal seam (normal mode). The seismic wave diffuses and propagates outward in this low-velocity trough of the coal seam. Among them, the upward and downward waves interfere and superimpose with each other in the coal seam, and most of the harmonic components cancel each other out and weaken, and gradually disappear; only various harmonics that meet certain conditions undergo constructive interference in the trough to form a standing wave perpendicular to the coal seam surface and continuously propagate forward in the coal seam, which forms the trough wave, also known as the coal seam wave.

[0027] In practice, the background server converts the format of the slot wave data, imports it into the slot wave processing software, and establishes a processing project. And preprocess the data. Preprocessing refers to abnormal trace inspection, rejection of bad traces, and empty traces. A large number of abnormal traces will affect the subsequent processing effect, especially the abnormal traces with strong energy, which have extremely adverse effects on multi-trace processing such as broadband filtering and imaging: it will affect the attenuation law of the amplitude, making it difficult to accurately obtain the compensation coefficient. Then establish an observation system. When establishing the observation system, specify the spatial coordinates of the source excitation point and the receiving point, and establish the corresponding relationship between the shot and the receiver. As an example, when processing the slot wave seismic penetration detection data of the 21211 working face, the position 1610m backward from the cut-through of the return air drift is taken as the origin of the plane coordinates, the direction along the return air drift is the positive direction of the X-axis, and the direction into the face along the orthogonal return air drift is the positive direction of the Y-axis, and coordinates are established for each data acquisition point (such as Figure 3 shown). Among them, there is a certain time difference from the detonation of the detonator to the explosion of the explosive. In addition, there are also time differences in the detonation delays of each detonator. This difference affects the first arrival time of the slot wave, and thus affects the imaging quality during slot wave velocity imaging. Therefore, before processing, it is necessary to correct the first arrival time of each trace according to the seismic wave propagation theory. Then perform geometric spreading energy compensation on the data, and then perform dispersion analysis, velocity analysis, narrowband filtering, and slot wave extraction, and analyze the single-shot record to obtain the slot wave energy attenuation coefficient imaging.

[0028] Step 102, analyze the distribution law of multiple abnormal prompt areas in the working face CT diagram, delimit multiple candidate abnormal areas, and obtain the average abnormal amplitude value of each candidate abnormal area.

[0029] In some embodiments, after the background server receives the working face CT diagram, color contrast analysis is used to identify high / low velocity areas (i.e., abnormal areas). Combining the seismic wave attenuation characteristics, screen possible abnormal areas (such as faults, collapse columns, water-rich areas, etc.). Use edge detection algorithms (such as Sobel operator or Canny algorithm) to outline the boundaries of the abnormal areas. Statistically analyze the spatial distribution characteristics (position coordinates, shape, area) of all abnormal prompt areas. Analyze the trend direction of the abnormal areas, such as distributed along the fault zone or concentrated near certain stratigraphic interfaces. Calculate the density of the abnormal areas and perform cluster analysis to identify regular patterns. Then, according to the distribution law of the abnormal areas, delimit candidate abnormal areas, such as Figure 4 shown, Figure 4 where the candidate abnormal areas are outlined in magenta, and the blue blocks are the abnormal prompt areas. Use connected component analysis to remove too small or isolated abnormal points. Combine historical data (such as existing collapse column or fault distributions) to screen high-risk abnormal areas. Extract the abnormal amplitude values within each candidate abnormal area. Calculate the mean value of each abnormal area and record it in the database. If the abnormal amplitude values show uneven distribution, the weighted average value can be calculated to analyze the change trend of the abnormal amplitude.

[0030] Step 103: Analyze multiple high-value abnormal areas of slot wave attenuation in the imaging of slot wave energy attenuation coefficient, and calculate the overlapping areas with multiple candidate abnormal areas to obtain multiple overlapping areas.

[0031] In some embodiments, the background server comprehensively demarcates the range of the abnormal area according to methods such as transmission slot wave CT imaging, and obtains multiple high-value abnormal areas of slot wave attenuation (such as Figure 5 the orange area shown). Among them, the larger the attenuation coefficient, the more factors that weaken the seismic wave energy exist near this area in the coal seam; the smaller the attenuation coefficient, that is, the coal seam occurrence is relatively stable, and the energy attenuation of the seismic wave during propagation is relatively small; when the slot wave encounters a tectonic anomaly, the energy has a large attenuation. Store multiple high-value abnormal areas of slot wave attenuation and multiple candidate abnormal areas in vector format. Among them, the vector format is a format that uses geometric elements such as points, lines, and surfaces (polygons) to represent spatial data. Store data based on coordinate points. Use Shapely for vector space calculation, calculate the intersection of multiple candidate abnormal areas and multiple high-value abnormal areas of slot wave attenuation to obtain multiple discrete overlapping areas, and then use the connected component labeling algorithm to divide these areas. Output the boundary coordinates and area of each overlapping area.

[0032] Step 104: Arrange and drill exploration points according to multiple overlapping areas. During the drilling process, measure multiple drilling trajectories and drilling data corresponding to each exploration point through a trajectory instrument.

[0033] In some embodiments, according to the preset spacing, select the center points or edges of multiple overlapping areas as exploration points to ensure that key areas are all covered. Query the borehole dip angle, borehole depth, and borehole strike of each overlapping area from the pre-stored geological structure database and drilling history data. And during the drilling process, measure multiple drilling trajectories and drilling data corresponding to each exploration point through a trajectory instrument. As Figure 6 shown, Figure 6 the blue lines in are the trajectories. Among them, the trajectory instrument is generally installed inside the drill pipe, and continuously measures the attitude and position of the borehole during the drilling process. The measurement method is gyroscopic inclinometry.

[0034] Step 105: Generate prediction data of the abnormal area of the fully mechanized coal mining face according to multiple drilling trajectories and drilling data corresponding to each exploration point.

[0035] In some embodiments, each drilling point corresponds to multiple drilling trajectories, and each drilling trajectory generates corresponding drilling data, including: Trajectory data: the dip angle, azimuth angle, and depth of the borehole, which are used to determine the three-dimensional spatial path of the borehole. Geological data: such as trough wave attenuation anomaly values, rock formation hardness, water content, coal seam thickness, fracture development conditions, etc. Spatial coordinate data: the positions of each drilling point and its trajectory in a three-dimensional coordinate system (X, Y, Z). The drilling point data measured by the trajectory instrument is used to calculate the path of each borehole in three-dimensional space and form a borehole network. Combining the geological data, the formation interfaces traversed by the boreholes are determined to form a three-dimensional geological model of the working face. Combining the imaging data of the trough wave energy attenuation coefficient, the high-value anomaly areas of trough wave attenuation are extracted. The three-dimensional coordinates of the trough wave anomaly areas are matched with the borehole trajectory data to analyze which boreholes have traversed the trough wave anomaly areas and the corresponding depths and rock formation characteristics. Through interpolation methods (such as Kriging interpolation or inverse distance weighting (IDW) interpolation), the borehole data is spatially interpolated to infer the trough wave anomaly conditions in the un-drilled areas. Combining the abnormal parameters obtained during the drilling process (such as coal seam thickness changes, fracture density, etc.), a three-dimensional abnormal area prediction model is constructed in the fully mechanized mining face. Calculate the spatial distribution characteristics of the abnormal areas: calculate the center, boundary, volume, and shape of the abnormal areas. Analyze the relationship between the abnormal areas and the coal seam in the fully mechanized mining face to evaluate whether it affects production safety. Calculate the average abnormal amplitude value of each abnormal area to evaluate the severity of the abnormal area. Through machine learning (such as random forests, neural networks, etc.), the corresponding relationship between the trough wave anomaly characteristics and the actual geological anomalies is trained to improve the prediction accuracy. Finally, the prediction data of the abnormal areas in the fully mechanized mining face is generated, which may include: the three-dimensional coordinate range (X, Y, Z) of the abnormal areas. The geological characteristics of the abnormal areas (such as lithology, fracture development, water content, etc.). Trough wave anomaly intensity data (attenuation coefficient, abnormal amplitude). Possible geological risk predictions (such as water inrush risk, gas enrichment area, fault influence area, etc.).

[0036] In these embodiments, the prediction accuracy is improved. Specifically, by using radio penetration equipment and trough wave seismographs for large-scale and non-destructive detection, the distribution information of geological anomalies is obtained, and the detection accuracy of the abnormal areas is improved. The radio penetration equipment provides CT imaging data, and the trough wave seismograph provides energy attenuation coefficient imaging, which respectively reflect different characteristics of underground anomalies (such as density changes, hardness of rock formations). By fusing the data of the two detection methods, the accuracy of abnormal area identification can be improved, and the influence of errors in a single detection method can be avoided. And based on the overlay analysis of multi-source data (CT abnormal areas and trough wave abnormal areas), the overlapping areas are calculated, and the drilling points are accurately planned to improve the effectiveness of drilling. The actual drilling trajectories of each drilling point are recorded by the trajectory instrument, and combined with the geological parameters obtained during the drilling process, a more accurate three-dimensional abnormal area model is constructed to dynamically optimize the prediction of abnormal areas and improve the prediction accuracy.

[0037] In some embodiments, in order to further solve the technical problems described in the background art, that is, "the spatial resolution and data dimensions of different detection means are different, making it difficult to directly superimpose and analyze; and geological anomalies often show discrete and fragmented characteristics in the detection data, resulting in blurred boundaries of the anomaly areas; and the traditional static anomaly area division cannot reflect dynamic risks", in some embodiments of the present invention, the distribution law of multiple anomaly prompt areas in the working face CT image is analyzed, multiple candidate anomaly areas are delimited, and the average anomaly amplitude value of each candidate anomaly area is obtained, including: Step 1: Calculate the distance between each anomaly prompt area and the nearest neighboring anomaly prompt area, denoted as the minimum adjacent distance; according to the minimum adjacent distance of each anomaly prompt area, determine the aggregation area of the anomaly prompt areas, generate a radiation range coefficient corresponding to the aggregation area of the anomaly prompt areas, and radiate outward from the aggregation area of the anomaly prompt areas according to the corresponding radiation range coefficient to obtain candidate anomaly areas.

[0038] In some embodiments, the minimum adjacent distance refers to the Euclidean distance between a certain anomaly prompt area and its nearest other anomaly prompt area. By calculating the minimum adjacent distance of each anomaly prompt area, the spatial distribution characteristics of the anomaly prompt areas can be analyzed. Set an aggregation determination threshold. If the minimum adjacent distances of multiple anomaly prompt areas are less than this threshold, these anomaly prompt areas are classified into the same aggregation area. This can eliminate isolated anomaly points and only retain anomaly groups with actual geological significance. The radiation range coefficient can be calculated based on factors such as the distribution density and anomaly amplitude of the anomaly prompt areas. For example, in areas with a higher anomaly density, the radiation range may be larger, while the radiation range of isolated anomaly points is smaller. By radiating and expanding outward, a more complete candidate anomaly area can be obtained, avoiding missing some boundary anomaly points.

[0039] Step 2: Count the anomaly amplitude values of all the anomaly prompt areas in each candidate anomaly area, and calculate the average anomaly amplitude value of each candidate anomaly area.

[0040] In some embodiments, the anomaly amplitude value usually comes from data such as seismic waves, trough waves, and CT imaging, and can be understood as a numerical expression of the degree of geological anomaly. For example, in a CT image, the change in gray value can reflect the formation density anomaly, and the attenuation degree of trough waves can be used to identify underground fault zones. By adding up the anomaly amplitude values of all the anomaly prompt areas and dividing by the number of anomaly prompt areas in the candidate anomaly area, the average anomaly amplitude value of each candidate anomaly area is obtained.

[0041] Among them, analyzing multiple high-value anomaly areas of trough wave energy attenuation coefficient imaging includes: If there is an area smaller than the preset range among multiple high-value anomaly areas of slot wave attenuation and the distance between the center point of the anomaly area and the tunnel entrance is less than the preset distance, the corresponding high-value anomaly area of ​​slot wave attenuation will be deleted, and a joint analysis will be performed with the single-shot seismic record and actual geological information to obtain the actual location information, affected range, regional shape, and anomaly type corresponding to each remaining high-value anomaly area of ​​slot wave attenuation.

[0042] In some embodiments, if the area of ​​a high-value anomaly zone of slot wave attenuation is smaller than a preset range, it may be a measurement error, noise or a minor anomaly and will not be considered. For example, if the threshold is set to 500m², and a certain anomaly zone is only 100m², the anomaly zone is considered too small and should be removed. If the center point of a high-value anomaly zone of slot wave attenuation is close to the tunnel entrance (less than the preset value), it may be due to the tunnel structure affecting the slot wave signal, resulting in a false anomaly. For example, if the preset distance threshold is 50m, and the distance between the center point of a certain anomaly zone and the tunnel entrance is 30m, the anomaly zone may be caused by structural reflection or signal interference around the tunnel and should be deleted. Single-shot seismic records contain the propagation, reflection and scattering characteristics of seismic waves, which can be used to verify the authenticity of slot wave anomalies. Comparing the slot wave anomaly zone with the high-energy reflection points in the single-shot seismic record, if the two match well, it means that the slot wave anomaly may be a true reflection of the underground structure. If the two match poorly, it may be an interference signal. Combined with known geological data, the authenticity of the slot wave anomaly zone is further verified. For example, if a channel wave anomaly area coincides with the known fault position in historical drilling data, it means that the anomaly area may be real. If an anomaly area coincides with a known artificial support structure, it may be human interference and should be deleted. After screening and joint analysis, the effective channel wave attenuation high-value anomaly area needs to further extract its key spatial information. The key spatial information includes the actual location information. The specific geographical location of the anomaly area is extracted through the channel wave imaging coordinate system (usually the mining area coordinate system or the geographic coordinate system). Impact range, for example, a channel wave anomaly area with an impact range of more than 2000m² may need to be focused on, while one less than 500m² may be a secondary anomaly. Regional shape, through the boundary extraction algorithm, the boundary morphology of the anomaly area is calculated. Shape characteristics can help determine the nature of the anomaly area. For example, an elongated anomaly area may correspond to an underground fault zone. A blocky anomaly area may correspond to a rock fracture zone or a collapse zone. Anomaly type, combined with seismic records and geological information, the specific type of the channel wave attenuation high-value anomaly area is determined, such as a fault fracture zone: it usually has a higher channel wave energy attenuation, and the shape extends along the fault direction. Collapse of goaf area: may show regular low trough wave energy attenuation, accompanied by other geological anomalies. Structural stress concentration area: Trough wave energy attenuation is obvious, and corresponds to the stress concentration location.

[0043] Among them, the high-value abnormal areas of the corresponding trough wave attenuation are deleted and jointly analyzed with single-shot seismic records and actual geological information, including: Perform spectral analysis on the waveform amplitude and frequency characteristics of the single-shot seismic record, and extract the abnormal characteristics of the seismic wave reflection interface; spatially match the abnormal characteristics of the reflection interface with the known fault and lithology boundary data in the geological information to generate a matching degree score; if the matching degree score is lower than the preset threshold, it is determined that the high-value abnormal area of the trough wave attenuation is interference signal and is deleted; perform boundary smoothing processing and regional connectivity analysis on the remaining high-value abnormal areas of the trough wave attenuation to generate the final set of effective abnormal areas.

[0044] In some embodiments, the single-shot seismic record refers to exciting seismic waves at a certain fixed source position and recording the seismic wave signals measured by the receivers around the source. Collect the single-shot seismic record, which includes waveform amplitude and frequency. Then convert the time-domain data of the seismic wave signal into frequency-domain data and extract the geological feature information therein. By analyzing the spectral characteristics, the possible reflection interfaces of the abnormal area are extracted. For example, sudden amplitude change may be a fault or a fracture, extremely high low-frequency energy may be a mined-out area or a collapse zone, and disappearance of high-frequency components may be a lithology change or a soft and fractured zone. Obtain the known geological information, including fault data: underground geological exploration, historical drilling data, and mining records of the mining area. Lithology boundary data: interfaces of different rock layers, such as sandstone-mudstone boundary, roof and floor of coal seams. 3D geological model: Use GIS (Geographic Information System) or mining area modeling software to generate a 3D geological model of the mining area.

[0045] In some embodiments, calculate the spatial coordinates (such as X, Y, Z) of the high-value abnormal area of the trough wave attenuation. Perform three-dimensional spatial matching with the known geological faults and lithology boundaries. Use the DTW (Dynamic Time Warping) algorithm to measure the similarity between the seismic signal pattern and the geological fault characteristics. Use the Hausdorff distance to calculate the proximity between the boundary of the seismic abnormal area and the boundary of the known fault. Combine statistical methods (such as similarity index, cross-correlation analysis) to calculate the matching degree score. If the matching degree score of the high-value abnormal area of the trough wave attenuation is lower than the preset threshold, it indicates that this abnormal area may be interference signal or noise and should be excluded. For example, if the threshold is set to 0.6 and the matching degree score of a certain abnormal area is only 0.4, then this abnormal area may be interference and should be deleted. If the matching degree score is 0.85, it indicates that the abnormal area may be related to the real geological anomaly and should be retained. After removing the interference signal, perform boundary smoothing processing through Gaussian filtering. And detect the spatial connectivity of the abnormal area through connected component analysis to remove isolated small-area abnormal points. Finally, obtain the real high-value abnormal area of the trough wave attenuation, that is, the set of effective abnormal areas. Among them, the set of effective abnormal areas includes position coordinates, influence range, geological anomaly type, and spatial form.

[0046] Among them, spatially matching the abnormal features of the reflection interface with the known faults and lithological boundary data in the geological information to generate a matching degree score, including: Establish a mapping relationship table between seismic reflection features and geological anomaly types; dynamically adjust the matching degree weights according to the mapping relationship table, and verify the credibility of the matching results through a convolutional neural network.

[0047] In some embodiments, a mapping relationship table between seismic reflection features and geological anomaly types is established. Among them, the mapping relationship table includes seismic reflection features, corresponding geological anomaly types, and corresponding characteristic numerical ranges. Then, dynamically adjust the matching degree weights according to the mapping relationship table, and verify the credibility of the matching results through a convolutional neural network.

[0048] Among them, calculating the distance between each abnormal prompt area and the nearest neighboring abnormal prompt area includes: Judging its spatial dispersion degree by measuring the maximum distance between the boundaries of the abnormal prompt area; when the dispersion degree exceeds a preset threshold, mark this abnormal prompt area as an independent abnormal area; perform boundary fusion on adjacent abnormal areas with a dispersion degree lower than the preset threshold to generate a continuous candidate abnormal area.

[0049] In some embodiments, the abnormal prompt area refers to a suspected abnormal area identified during the exploration process, which may be an underground fault, a mined-out area, or other geological anomalies. When performing data analysis, it is necessary to calculate the distance between each abnormal prompt area and its nearest neighboring abnormal prompt area. The maximum distance refers to the longest distance between a certain boundary point of one abnormal prompt area and a certain boundary point of another abnormal prompt area. Through this maximum distance, the separation degree between two abnormal prompt areas can be measured. If the maximum distance between two abnormal prompt areas is large, it means that they are relatively independent in space and have a high dispersion degree. If the maximum distance is small, it means that these abnormal prompt areas are close to each other and have a low spatial dispersion degree. According to the measured maximum distance, the spatial dispersion degree of each abnormal prompt area can be judged. For abnormal prompt areas with a high dispersion degree, if the maximum distance between them and neighboring abnormal prompt areas exceeds a preset threshold (usually set by experience or experimental data), it can be considered that these abnormal prompt areas are independent abnormal areas, which are relatively independent in space and should not be merged. When the maximum distance exceeds the threshold, the abnormal prompt area is marked as an independent abnormal area, that is, they represent different geological anomaly bodies or different regions and should be processed separately. If the maximum distance between two or more abnormal prompt areas is less than the preset threshold, it means that they are close in space and can be regarded as a continuous abnormal area. At this time, the boundaries of these adjacent abnormal prompt areas will be fused to form a continuous candidate abnormal area.

[0050] Among them, boundary fusion includes: establishing a vector boundary buffer for the abnormal area. When the buffers of adjacent abnormal areas overlap, the minimum bounding rectangle algorithm is used to generate a continuous candidate abnormal area after fusion, and the extreme points of the abnormal amplitude of the original abnormal area are retained as feature markers.

[0051] In some embodiments, the vector boundary refers to the outer contour of the abnormal area, and the buffer refers to the area established within a certain distance around this boundary. The buffer is an area formed by expanding the boundary of each abnormal area by a certain distance, and the width of the buffer is usually a preset distance value. This buffer can help identify the relationship between adjacent abnormal areas and facilitate subsequent fusion analysis. When the buffers of multiple abnormal areas are adjacent or overlapping, it indicates that these abnormal areas may be continuous in space or their boundaries are very close. This overlapping relationship is a key factor in determining whether two abnormal areas need to be fused. The overlap judgment is made by detecting whether the buffers intersect or overlap, which can determine whether two abnormal areas belong to the same geological body in space and whether they should be merged into a larger abnormal area. If it is determined that the buffers of two or more abnormal areas overlap, the minimum bounding rectangle algorithm is used for fusion. The purpose of this algorithm is to find a minimum rectangle that can contain all the overlapping areas. When performing boundary fusion, it is also necessary to retain the extreme points of the abnormal amplitude of the original abnormal area as feature markers. Among them, the extreme points of the abnormal amplitude: refer to the points with the strongest or weakest amplitude values within each abnormal area. These points can well reflect the core characteristics of geological anomalies. Feature markers: By retaining these extreme points, it can help subsequent analysts find the most representative abnormal signals in the continuous candidate abnormal area, and then conduct further processing and decision-making.

[0052] Among them, a method for predicting geological abnormal areas in fully mechanized coal mining faces based on multi-source data according to the present invention further includes: After the continuous candidate abnormal area is generated, the top and bottom plate shape data of the abnormal area is obtained by a 3D laser scanner, and the scanned point cloud is spatially registered with the results of trough wave imaging to correct the three-dimensional spatial shape of the candidate abnormal area and mark the potential danger level.

[0053] In some embodiments, a three-dimensional laser scanner (LiDAR) is a technology that obtains precise three-dimensional data of a target object through laser measurement. The three-dimensional laser scanner is used to scan the morphological data of the roof and floor of the abnormal area. Specifically, the laser scanner can calculate the distance to the object surface by reflecting the time of the laser beam, thereby obtaining high-precision three-dimensional point cloud data. After the continuous candidate abnormal area is generated, the morphological data of the roof and floor of the abnormal area is obtained through the three-dimensional laser scanner. The morphological data of the roof and floor refers to the shape of the fully-mechanized mining face in space, including the geometric forms of the upper and lower layers. By scanning the roof and floor, the spatial structure within the working face can be accurately captured, providing detailed geometric information for subsequent spatial analysis. After the continuous candidate abnormal area is generated. Through spatial registration, the point cloud data can be accurately aligned with the crosshole tomography image, so that the two match in space, and consistent geological information can be extracted from the two data sources. After spatial registration, the point cloud data and the crosshole tomography results can be analyzed in the same coordinate system. The purpose of this step is to use the morphological data of the roof and floor obtained by three-dimensional laser scanning to correct the spatial morphology of the candidate abnormal area. By analyzing the characteristics of the spatial position, morphology, crosshole tomography attenuation, etc. of the abnormal area, and combining the actual geological and working face conditions, the risk level of each abnormal area is evaluated. For example, some abnormal areas may have a higher risk due to being too shallow or adjacent to the mining area and need special attention.

[0054] In these embodiments, the recognition accuracy and reliability of the abnormal area are improved. Specifically, through multi-source data fusion (such as CT images, crosshole tomography, geological information, etc.) and spatial registration technology, the recognition accuracy and reliability of the abnormal area can be effectively improved. Reducing the influence of interference signals: By setting distance rules to delete small-range abnormal areas with high crosshole tomography attenuation values close to the roadway opening, the influence of interference signals on the prediction of geological abnormal areas is reduced. Combining the three-dimensional laser scanning data to correct the three-dimensional spatial morphology of the abnormal area can more accurately reflect the actual geological conditions within the working face. By marking the potential risk levels, effective risk assessment and early warning can be provided for the fully-mechanized mining face, helping to formulate a safer operation plan. By automatically calculating parameters such as the minimum adjacent distance, abnormal amplitude, and radiation range of the abnormal area through algorithms, as well as automatic boundary fusion and matching degree scoring, the efficiency of abnormal area analysis and prediction is improved.

[0055] The above description is only some preferred embodiments of the present invention and an explanation of the applied technical principles. Those skilled in the art should understand that the scope of the invention involved in the present invention is not limited to the technical solutions formed by the specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above inventive concept. For example, the technical solutions formed by mutually replacing the above features with the technical features (but not limited to) disclosed in the present invention that have similar functions.

Claims

1. A prediction method for geological abnormal areas in fully-mechanized mining faces based on multi-source data, characterized in that, Including: Using a radio penetration device and a trough wave seismograph to detect a fully mechanized coal mining face respectively, obtaining a CT map of the working face and an imaging of the trough wave energy attenuation coefficient; Analyzing the distribution law of multiple abnormal prompt areas in the CT map of the working face, demarcating multiple candidate abnormal areas, and obtaining the average abnormal amplitude value of each candidate abnormal area; Analyzing multiple high-value abnormal areas of trough wave attenuation in the imaging of the trough wave energy attenuation coefficient, and calculating the overlapping areas with multiple candidate abnormal areas to obtain multiple overlapping areas; Arranging and drilling boreholes according to multiple overlapping areas. During the drilling process, measuring multiple drilling trajectories and drilling data corresponding to each borehole through a trajectory instrument; Generating prediction data of abnormal areas of the fully mechanized coal mining face according to multiple drilling trajectories and drilling data corresponding to each borehole.

2. The prediction method for geological abnormal areas in fully-mechanized coal mining faces based on multi-source data according to claim 1, wherein, The analysis of the distribution law of multiple abnormal prompt areas in the CT map of the working face, demarcating multiple candidate abnormal areas, and obtaining the average abnormal amplitude value of each candidate abnormal area includes: Calculating the distance between each abnormal prompt area and the nearest adjacent abnormal prompt area and recording it as the minimum adjacent distance; according to the minimum adjacent distance of each abnormal prompt area, determining the aggregation area of abnormal prompt areas, generating a radiation range coefficient corresponding to the aggregation area of abnormal prompt areas, and radiating outward from the aggregation area of abnormal prompt areas according to the corresponding radiation range coefficient to obtain candidate abnormal areas; Counting the abnormal amplitude values of all abnormal prompt areas in each candidate abnormal area and calculating the average abnormal amplitude value of each candidate abnormal area.

3. The prediction method for geological abnormal areas in fully mechanized coal mining faces based on multi-source data according to claim 2, characterized in that The analysis of multiple high-value abnormal areas of trough wave attenuation in the imaging of the trough wave energy attenuation coefficient includes: If there is an area in multiple high-value abnormal areas of trough wave attenuation whose area range is less than the preset range and the distance between the center point of the abnormal area and the roadway opening is less than the preset distance, then deleting the corresponding high-value abnormal area of trough wave attenuation, and conducting a joint analysis with single-shot seismic records and actual geological information to obtain the actual position information, influence range, regional shape, and abnormal type corresponding to each remaining high-value abnormal area of trough wave attenuation.

4. The method for predicting geological abnormal areas in fully mechanized coal mining faces based on multi-source data according to claim 3, characterized in that, The deletion of the corresponding high-value abnormal area of trough wave attenuation and the joint analysis with single-shot seismic records and actual geological information include: Conducting a spectral analysis of the waveform amplitude and frequency characteristics of single-shot seismic records, and extracting abnormal characteristics of seismic wave reflection interfaces; spatially matching the abnormal characteristics of the reflection interfaces with known fault and lithology demarcation data in geological information to generate a matching degree score; if the matching degree score is lower than the preset threshold, then determining that the high-value abnormal area of trough wave attenuation is interference signal and deleting it; conducting boundary smoothing processing and regional connectivity analysis on the remaining high-value abnormal areas of trough wave attenuation to generate a final set of effective abnormal areas.

5. The prediction method for geological abnormal areas in fully-mechanized mining faces based on multi-source data according to claim 4, characterized in that The spatial matching of the abnormal characteristics of the reflection interfaces with known fault and lithology demarcation data in geological information to generate a matching degree score includes: Establishing a mapping relationship table between seismic reflection characteristics and geological abnormal types; dynamically adjusting the matching degree weights according to the mapping relationship table, and verifying the credibility of the matching results through a convolutional neural network.

6. The prediction method for geological abnormal areas in fully-mechanized coal mining faces based on multi-source data according to claim 5, characterized in that, The calculation of the distance between each abnormal prompt area and the nearest adjacent abnormal prompt area includes: Judge the spatial dispersion degree by measuring the maximum distance between the boundaries of the abnormal prompt area; when the dispersion degree exceeds the preset threshold, mark the abnormal prompt area as an independent abnormal area; perform boundary fusion on adjacent abnormal areas with a dispersion degree lower than the preset threshold to generate a continuous candidate abnormal area.

7. The prediction method for geological abnormal areas in fully-mechanized mining faces based on multi-source data according to claim 6, characterized in that The boundary fusion includes: establishing a vector boundary buffer for the abnormal area. When the buffers of adjacent abnormal areas overlap, use the minimum bounding rectangle algorithm to generate a continuous candidate abnormal area after fusion, and retain the extreme points of the abnormal amplitude of the original abnormal area as feature markers.

8. The prediction method for geological abnormal areas in fully mechanized coal mining faces based on multi-source data according to claim 7, characterized in that, It also includes: After the continuous candidate abnormal area is generated, obtain the top and bottom plate morphology data of the abnormal area through a 3D laser scanner, perform spatial registration on the scanned point cloud and the trough wave imaging result, correct the three-dimensional spatial morphology of the candidate abnormal area, and mark the potential danger level.

Images