Mine hidden disaster-causing factor census method based on multi-source heterogeneous data fusion

Abstract

The present application relates to the technical field of data mining software, and discloses a mine hidden disaster-causing factor census method based on multi-source heterogeneous data fusion, comprising the following steps: obtaining mine account, mining map, geophysical anomaly, drilling verification, gas parameter, hydrogeology, field investigation and mining planning data, converting the data into hidden disaster-causing factor data objects containing mine well identification, coal seam position, spatial coordinates, elevation range, disaster type, evidence source and verification state, performing spatial position fusion on the hidden disaster-causing factor data objects and mining planning area, updating the abnormal verification state according to position matching and position matching, obtaining gas risk identification results according to gas content, gas pressure and buried depth, and generating hidden disaster-causing factor census results and field verification tasks. Through object processing, position fusion and state updating, the present application forms traceable correlation among data, risk objects and verification tasks.

Description

Technical Field

[0001] This invention relates to the field of data mining software technology, and in particular to a method for investigating hidden disaster-causing factors in mines based on the fusion of multi-source heterogeneous data. Background Technology

[0002] Data mining software technology typically revolves around the collection, cleaning, storage, association, retrieval, statistics, modeling, and visualization of multi-source structured, semi-structured, and unstructured data. Data can be obtained from sensors, measuring instruments, business terminals, databases, map files, and manual data entry records. It is then retrieved and calculated by local or server programs according to fields, coordinates, time, category, thresholds, and mapping relationships. Traditional methods for investigating hidden mine disaster factors refer to the technical process of verifying data, conducting on-site investigations, and compiling results for objects such as goaf areas, poorly sealed boreholes, geological structures, water sources and channels, gas, geological hazards, construction projects, and roof structures. Generally, investigators collect geological reports, hydrological observations, geophysical drilling data, gas parameters, mining engineering maps, monitoring records, and on-site verification records. They then input coal seam elevations, mine boundaries, goaf areas, borehole locations, aquifer distribution, fault information, and mining plans into databases or graphical systems. Based on spatial overlay, attribute comparison, record verification, and threshold judgment, they generate investigation result maps, factor lists, and treatment guidelines.

[0003] Traditional methods for investigating hidden disaster-causing factors in mines rely on manual compilation of mine ledgers, mining maps, geophysical results, drilling records, gas parameters, hydrogeological data, and field investigation records. Data from different sources are inconsistent in terms of coordinates, stratigraphy, elevation, time, and disaster type. The correlation between anomaly identification, verification records, and mining plans requires manual comparison. The verification standards for data at different stages are scattered, the boundaries between verified objects and objects to be verified are confused, the survey results are updated late, and the correspondence between field verification tasks and risk objects is unclear. Summary of the Invention

[0004] The purpose of this invention is to provide a method for investigating hidden disaster-causing factors in mines based on multi-source heterogeneous data fusion. This method uses data mining software to uniformly objectify data such as mine ledgers, mining maps, geophysical anomalies, drilling verification, gas parameters, hydrogeology, field investigations, and mining plans. It establishes correlations between coordinates, stratigraphy, elevation, time, disaster type, evidence source, and verification status, so that the results of the investigation of hidden disaster-causing factors and the field verification tasks form a traceable data correspondence.

[0005] To achieve the purpose of this invention, the present invention provides a method for investigating hidden disaster-causing factors in mines based on multi-source heterogeneous data fusion, comprising the following steps: acquiring multi-source heterogeneous mine investigation data, including mine ledger data, mining map data, geophysical anomaly data, drilling verification data, gas parameter data, hydrogeological data, field survey data, and mining planning data; converting the multi-source heterogeneous mine investigation data into hidden disaster-causing factor data objects, including mine identification, coal seam location, spatial coordinates, elevation range, time attribute, disaster type, evidence source, and verification status; and based on the spatial coordinates, coal seam location, and elevation range, investigating the hidden disaster-causing factors... The disaster-causing factor data objects are spatially and stratigraphically fused with the mining planning area to obtain disaster factor correlation results. When the anomaly area corresponding to the geophysical anomaly data meets the location matching and stratigraphic matching of the drilling verification data, the anomaly area is marked as verified. When the anomaly area does not meet the location matching or stratigraphic matching, the anomaly area is marked as pending review. Gradient prediction is performed based on the gas content, gas pressure and burial depth information in the gas parameter data, and the prediction results are compared with fixed thresholds and regional mean multiple thresholds to obtain gas risk identification results. Based on the disaster factor correlation results, the anomaly area verification status and the gas risk identification results, the hidden disaster-causing factor survey results and on-site verification tasks are generated.

[0006] Furthermore, the process of converting the multi-source heterogeneous mine survey data into hidden disaster-causing factor data objects includes: extracting the goaf number, poorly sealed borehole number, sealing location, formation time, and mine identification from the mine ledger data; extracting the mining boundary, coal seam floor contour lines, roadway location, and mining planning area from the mining map data; extracting the anomaly zone boundary, verification point coordinates, verification strata, and verification record time from the geophysical anomaly data and drilling verification data; aligning the extracted data according to the mine identification, coal seam strata, spatial coordinates, elevation range, time attribute, and disaster type; merging the same disaster type into the same hidden disaster-causing factor data object when the spatial coordinates and coal seam strata meet the consistency condition; and generating hidden disaster-causing factor data objects with different evidence sources when the consistency condition is not met.

[0007] Furthermore, the field alignment includes coordinate alignment, stratigraphic alignment, and time alignment. The coordinate alignment uses the mining area boundary coordinates in the mining map data as a benchmark, converting the text positions in the mine ledger data and the positioning points in the field survey data into the same plane coordinate system. The stratigraphic alignment uses the coal seam strata and elevation range as common constraints, assigning geophysical anomaly data, drilling verification data, gas parameter data, and hydrogeological data to the corresponding coal seam. The time alignment determines the evidence priority according to the recording time from most recent to oldest. When the same hidden disaster-causing factor data object has the same field with different recording times, the field value with the recording time closer to the generation time of the mining planning data is retained.

[0008] Furthermore, based on spatial coordinates, coal seam strata, and elevation range, the process of spatially merging the data objects of the hidden disaster-causing factors with the mining planning area includes: classifying the data objects of hidden disaster-causing factors corresponding to goaf areas, water accumulation areas, poorly sealed boreholes, faults, aquifers, water-conducting channels, and gas-rich areas into a disaster factor object set; classifying the mining area, tunneling area, and boundary coal pillar area in the mining planning data into a mining planning object set; calculating the planar overlap relationship, boundary distance relationship, vertical elevation intersection relationship, and coal seam strata consistency between the disaster factor object set and the mining planning object set; and generating a disaster factor association result of the direct association type when the planar overlap relationship, vertical elevation intersection relationship, and coal seam strata consistency all meet the association conditions.

[0009] Furthermore, when calculating the boundary distance relationship between the disaster factor object set and the mining planning object set, the outer boundary of the mining planning object set is used as the distance calculation benchmark. The shortest distance from the boundary of the poorly sealed borehole, fault, water accumulation area, and gas-rich area to the outer boundary of the mining planning object set is calculated respectively. When the shortest distance is within the preset verification distance range and the coal seam consistency meets the association condition, the corresponding disaster factor association result is marked as the neighbor verification type. When the shortest distance exceeds the preset verification distance range, the evidence source and verification status of the corresponding hidden disaster-causing factor data object are retained, and no neighbor verification type disaster factor association result is generated.

[0010] Furthermore, after the time caliber is aligned, evidence consistency verification is performed on the data objects of hidden disaster-causing factors. Evidence consistency verification uses the following parallel judgment conditions: spatial coordinates falling into the same grid cell, coal seam position being the same, and elevation ranges intersecting. When at least two types of evidence sources for the same hidden disaster-causing factor data object meet the parallel judgment conditions, the verification status is updated to the review and confirmation status. When only one type of evidence source for the same hidden disaster-causing factor data object meets the parallel judgment conditions, the verification status is retained as the pending review status, and the fields that do not meet the parallel judgment conditions are written into the field set to be verified.

[0011] Further, the process of marking the abnormal area as the verified state includes: extracting the outer boundary and abnormal strata of the abnormal area from the geophysical anomaly data; extracting the verification point coordinates, verification strata, and verification record time from the drilling verification data; when the verification point coordinates fall within the outer boundary of the abnormal area and the verification strata are consistent with the abnormal strata, it is determined that both the location matching and the strata matching are valid, and the verification state of the abnormal area is updated to the verified state; when the verification point coordinates do not fall within the outer boundary of the abnormal area, or the verification strata are inconsistent with the abnormal strata, the verification state of the abnormal area is updated to the pending review state.

[0012] Furthermore, the process of gradient prediction based on the gas content, gas pressure, and burial depth information in the gas parameter data includes: screening gas measurement records under the same mine identifier and the same coal seam level from the gas parameter data; the gas measurement records include the coordinates of the measuring point, the burial depth of the measuring point, the gas content, the gas pressure, and the recording time; when there are shallow and deep measuring points in the gas measurement records, the gas content gradient is determined according to the conversion result of the difference in gas content and the difference in burial depth per 100 meters, and the gas pressure gradient is determined according to the conversion result of the difference in gas pressure and the difference in burial depth per 100 meters; when there are no shallow and deep measuring points in the gas measurement records that satisfy the burial depth increasing relationship, the generation of gradient prediction results for the corresponding coal seam level is stopped.

[0013] Furthermore, the process of comparing the prediction results with a fixed threshold and a multiple threshold of the regional average includes: generating a predicted gas content based on the gas content gradient and the maximum burial depth of the mining planning area, and generating a predicted gas pressure based on the gas pressure gradient and the maximum burial depth of the mining planning area; first comparing the predicted gas content with a fixed threshold of gas content, and then comparing the predicted gas pressure with a fixed threshold of gas pressure; when the fixed threshold comparison is not valid, then comparing the predicted gas content with a multiple threshold of the regional average gas content, and then comparing the predicted gas pressure with a multiple threshold of the regional average gas pressure, and generating the gas risk identification result based on the two comparison results.

[0014] Furthermore, the process of generating the hidden disaster-causing factor survey results and on-site verification tasks based on the disaster factor association results, the anomaly area verification status, and the gas risk identification results includes: using the direct association type and adjacent verification type in the disaster factor association results as task generation objects, using the pending review status in the anomaly area verification status as task triggering conditions, and using the high gas content area and high gas pressure area in the gas risk identification results as risk coverage conditions; when the task generation object meets the task triggering conditions or risk coverage conditions, an on-site verification task containing the task object, task location, disaster type, verification basis, and task priority is generated; when the task generation object does not meet the task triggering conditions and does not meet the risk coverage conditions, the corresponding hidden disaster-causing factor data object is written into the hidden disaster-causing factor survey results.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, mine ledger data, mining map data, geophysical anomaly data, drilling verification data, gas parameter data, hydrogeological data, field survey data, and mining planning data are converted into hidden disaster-causing factor data objects, so that coordinates, strata, elevations, times, disaster types, evidence sources, and verification status are processed under the same caliber. Disaster factor correlation results are obtained through spatial strata fusion, the verification status of anomaly areas is updated through location matching and strata matching, and gas risk identification results are obtained through gas gradient prediction and threshold comparison. This makes the survey results and field verification tasks form a traceable data correspondence, reducing correlation omissions caused by manual caliber conversion. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is the main flowchart of the mine hidden disaster-causing factor survey of the present invention; Figure 2 This is a flowchart of the multi-source data objectification conversion process of the present invention; Figure 3 This is a logic diagram for field alignment and evidence consistency verification in this invention; Figure 4 This is a spatial layer fusion and association type determination diagram for the present invention; Figure 5 This is a data flow diagram showing the update of geophysical anomalies and drilling verification status in this invention. Figure 6This invention generates a linkage diagram for the gas risk identification and on-site verification tasks. Detailed Implementation

[0018] The technical solutions described in this invention are explained below with reference to embodiments and newly added drawings. The terms "multi-source heterogeneous mine survey data," "hidden disaster-causing factor data objects," "mining planning areas," "disaster factor correlation results," "anomaly zone verification status," "gas risk identification results," "hidden disaster-causing factor survey results," and "on-site verification tasks" mentioned in the protection scope text and invention content have the same meaning in this section. Multi-source heterogeneous mine survey data refers to mine ledger data, mining map data, geophysical anomaly data, drilling verification data, gas parameter data, hydrogeological data, on-site investigation data, and mining planning data used for the survey of hidden disaster-causing factors in mines; hidden disaster-causing factor data objects refer to data objects that, after being organized with unified fields, can characterize disaster objects such as goaf areas, water accumulation areas, poorly sealed boreholes, faults, aquifers, water-conducting channels, and gas-rich areas; mining planning areas refer to the mining areas, tunneling areas, and boundary coal pillar areas determined by the mining planning data; disaster factor correlation results refer to... The association type between the hazard object and the mining planning area is determined by the spatial coordinates, coal seam position, and elevation range; the verification status of the anomaly area refers to the verified status or the status pending review formed after the anomaly area corresponding to the geophysical anomaly data is checked by the drilling verification data; the gas risk identification result refers to the identification result formed by the gradient prediction and threshold comparison of gas content, gas pressure, and burial depth information; the on-site verification task refers to the task object, task location, hazard type, verification basis, and task priority generated based on the hazard factor association result, the anomaly area verification status, and the gas risk identification result.

[0019] Please see Figures 1 to 6 This embodiment provides a method for investigating hidden disaster-causing factors in mines based on multi-source heterogeneous data fusion, applicable to the unified processing of regional mine hidden disaster-causing factor investigation data. This method uses data already generated or obtainable during the mine investigation process as input, without altering the fundamental sources of mine geology, hydrology, geophysical exploration, drilling, and gas measurement. Through objectification, spatial stratigraphic fusion, verification status updates, gas risk identification, and task generation, it transforms scattered data into a traceable data chain. This method includes steps S1 to S6.

[0020] S1. Acquire multi-source heterogeneous mine survey data. This data includes mine ledger data, mining map data, geophysical anomaly data, drilling verification data, gas parameter data, hydrogeological data, field survey data, and mining planning data. Mine ledger data provides goaf numbers, poorly sealed borehole numbers, sealing locations, formation times, and mine identification. Mining map data provides mining boundaries, coal seam floor contour lines, roadway locations, and planned mining areas. Geophysical anomaly data provides anomaly zone boundaries and anomaly strata. Drilling verification data provides verification point coordinates, verification strata, and verification recording time. Gas parameter data provides measuring point coordinates, measuring point depth, gas content, gas pressure, and recording time. Hydrogeological data provides information on aquifers, water-conducting channels, and corresponding coal seams or elevations. Field survey data provides location points, survey objects, and survey recording time. Mining planning data provides the corresponding future mining areas, tunneling areas, and boundary coal pillar areas.

[0021] S2 converts multi-source heterogeneous mine survey data into hidden hazard-causing factor data objects. These data objects include mine identification, coal seam location, spatial coordinates, elevation range, time attribute, hazard type, evidence source, and verification status. The input for this step is the multi-source heterogeneous mine survey data obtained in S1, and the output is a data object with a unified field structure. Mine identification distinguishes different mine sources; coal seam location distinguishes high-carbon coal seams, lower-carbon coal seams, or other coal seam objects; spatial coordinates express planar location; elevation range expresses vertical range; time attribute expresses the chronological order of data formation or recording; hazard type distinguishes objects such as goaf, water accumulation area, poorly sealed boreholes, faults, aquifers, water-conducting channels, and gas-rich areas; evidence source records data originating from ledgers, maps, geophysical exploration, drilling, field investigations, gas measurement, or hydrogeological data; and verification status indicates whether the data object is supported by corresponding verification data.

[0022] S201 extracts the goaf number, poorly sealed borehole number, sealing location, formation time, and mine identifier from the mine ledger data. The goaf number serves as the object identifier for the goaf, the poorly sealed borehole number serves as the object identifier for the borehole, the sealing location helps determine the spatial extent of the goaf or the closed area of ​​the roadway, the formation time determines the temporal attribute of the ledger evidence, and the mine identifier determines the mine to which the data object belongs. After extraction, the mine ledger data is no longer compared in its original table state, but rather uses the object fields for subsequent field alignment.

[0023] S202 extracts mining boundaries, coal seam floor contour lines, roadway locations, and planned mining areas from mining map data. Mining boundaries define the planar extent of historical or planned mining operations; coal seam floor contour lines supplement coal seam elevation relationships; roadway locations identify spatial connections between existing engineering works and hidden hazard-causing factors; and the planned mining area serves as the object for subsequent spatial stratigraphic fusion. After extraction, the mining map data forms boundary data that can be used in conjunction with ledger data, geophysical anomaly data, and drilling verification data.

[0024] S203 extracts the anomaly zone boundary, verification point coordinates, verification strata, and verification record time from geophysical anomaly data and drilling verification data. The anomaly zone boundary is used to limit the suspected spatial range reflected by the geophysical anomaly; the verification point coordinates are used to determine the drilling verification location; the verification strata are used to determine whether the drilling verification object corresponds to the anomaly stratum; and the verification record time is used to determine the temporal attribute of the drilling verification evidence. The spatial relationship between the anomaly zone boundary and the verification point coordinates, and the stratigraphic relationship between the anomaly stratum and the verification stratum, are used as the basis for status updates in S4.

[0025] S204. The extracted data is aligned by field according to mine identification, coal seam location, spatial coordinates, elevation range, time attribute, and disaster type. Field alignment includes coordinate alignment, stratigraphic alignment, and time alignment. Coordinate alignment uses the mine boundary coordinates in the mining map data as a benchmark, converting the text positions in the mine ledger data and the positioning points in the field survey data to the same plane coordinate caliber. Stratigraphic alignment uses the coal seam location and elevation range as common constraints, assigning geophysical anomaly data, drilling verification data, gas parameter data, and hydrogeological data to the corresponding coal seam. Time alignment prioritizes evidence according to the recording time from most recent to oldest. When the same hidden disaster-causing factor data object has the same field with different recording times, the field value with the recording time closer to the generation time of the mining planning data is retained, ensuring that subsequent association judgments are consistent with the time caliber of the mining planning data.

[0026] S205: When the same type of disaster meets the consistency criteria in both spatial coordinates and coal seam location, the corresponding data are merged into a single hidden disaster-causing factor data object. When the consistency criteria are not met, separate hidden disaster-causing factor data objects with different evidence sources are generated. The spatial coordinate consistency criteria are based on overlapping object boundaries, location points falling within object boundaries, or positions within the same grid cell. The coal seam location consistency criteria are based on identical coal seam names or elevation ranges belonging to the same coal seam. The merged hidden disaster-causing factor data object retains each evidence source and corresponding time attribute; unmerged data objects retain their independent object identity to avoid data from different layers or spatial locations being mixed up.

[0027] S206. After aligning the time caliber, perform evidence consistency verification on the data objects of hidden disaster-causing factors. Evidence consistency verification uses the following parallel judgment conditions: spatial coordinates falling into the same grid cell, identical coal seam positions, and overlapping elevation ranges. When at least two types of evidence sources for the same hidden disaster-causing factor data object meet the parallel judgment conditions, the verification status is updated to the review and confirmation status; when only one type of evidence source for the same hidden disaster-causing factor data object meets the parallel judgment conditions, the verification status is retained as pending review, and fields that do not meet the parallel judgment conditions are written into the field set to be verified. The field set to be verified is used to record fields in coordinates, seams, elevations, or time that have not yet received consistent evidence support, serving as one of the bases for subsequent on-site verification tasks.

[0028] S3, based on spatial coordinates, coal seam position, and elevation range, performs spatial stratigraphic fusion of the hidden disaster-causing factor data objects and the mining planning area to obtain the disaster factor correlation results. This step matches the hidden disaster-causing factor data objects obtained in S2 with the mining planning data in S1. The inputs are the hidden disaster-causing factor data objects and the mining planning area, and the output is the disaster factor correlation results. Spatial stratigraphic fusion considers both planar position and vertical stratigraphic position, and does not use the result of a single planar superposition as the correlation conclusion.

[0029] S301 categorizes the hidden disaster-causing factors corresponding to goaf areas, waterlogged areas, poorly sealed boreholes, faults, aquifers, water-conducting channels, and gas-rich areas into a disaster factor object set. Each object in the disaster factor object set includes a mine identifier, coal seam location, spatial coordinates, elevation range, time attribute, disaster type, evidence source, and verification status. After entering the disaster factor object set, objects are organized according to disaster type and coal seam location to match them item by item with the mining planning area.

[0030] S302 categorizes the mining planning data, including the longwall mining area, tunneling area, and boundary coal pillar area, into a mining planning object set. Each object in the mining planning object set includes a planar boundary, corresponding coal seam level, elevation range, and planning time attribute. The longwall mining area represents the planned longwall mining range, the tunneling area represents the planned tunneling range, and the boundary coal pillar area represents the boundary range that requires spatial verification with adjacent hazardous objects.

[0031] S303 calculates the planar overlap relationship, boundary distance relationship, vertical elevation intersection relationship, and coal seam consistency between the disaster factor object set and the mining planning object set. The planar overlap relationship is used to determine whether the disaster factor objects and the mining planning objects overlap in planar scope; the boundary distance relationship is used to determine whether the shortest distance from the boundary of the disaster factor object to the outer boundary of the mining planning object falls within the preset verification distance range; the vertical elevation intersection relationship is used to determine whether the elevation range of the disaster factor objects and the elevation range of the mining planning objects intersect; and the coal seam consistency relationship is used to determine whether the disaster factor objects and the mining planning objects correspond to the same coal seam. These relationships form intermediate judgment results, and the disaster factor association results are generated by combining the intermediate judgment results.

[0032] S304: When the planar overlap, vertical elevation intersection, and coal seam horizon consistency all meet the association conditions, a direct association type of disaster factor association result is generated. Direct association type indicates that the disaster factor object and the mining planning area have a corresponding relationship in the planar, vertical, and coal seam horizon dimensions. The direct association type of disaster factor association result records the candidate task object, the corresponding mining planning area, the disaster type, and the source of evidence, serving as one of the inputs for generating the field verification task in S6.

[0033] S305, when calculating the boundary distance relationship between the disaster factor object set and the mining planning object set, the outer boundary of the mining planning object set is used as the distance calculation benchmark. The shortest distance from the boundary of the sealed poor borehole, fault, water accumulation area, and gas-rich area to the outer boundary of the mining planning object set is calculated respectively. When the shortest distance is within the preset verification distance range and the coal seam consistency meets the association condition, the corresponding disaster factor association result is marked as the neighboring verification type; when the shortest distance exceeds the preset verification distance range, the evidence source and verification status of the corresponding hidden disaster-causing factor data object are retained, and no neighboring verification type disaster factor association result is generated. The neighboring verification type and the direct association type are entered into S6 as mutually exclusive association types.

[0034] In step S4, when the anomaly area corresponding to the geophysical anomaly data satisfies both location and stratigraphic matching with the drilling verification data, the anomaly area is marked as verified; when the anomaly area does not satisfy either location or stratigraphic matching, it is marked as pending review. The inputs to this step are geophysical anomaly data, drilling verification data, and the hidden disaster-causing factor data object formed in step S2. The output is the anomaly area verification status. The anomaly area verification status is used to distinguish between anomaly areas already supported by drilling verification data and anomaly areas that still require review.

[0035] S401 extracts the outer boundary of the anomaly area and the anomaly stratum from the geophysical anomaly data, and extracts the verification point coordinates, verification stratum, and verification record time from the drilling verification data. The outer boundary of the anomaly area is used to define the planar range of the geophysical anomaly, the anomaly stratum is used to define the location of the coal seam or aquifer corresponding to the geophysical anomaly, the verification point coordinates are used to determine whether the drilling verification location falls within the outer boundary of the anomaly area, the verification stratum is used to determine whether the drilling verification object corresponds to the anomaly stratum, and the verification record time is used to record the time attribute of the verification evidence.

[0036] S402, when the coordinates of the verification point fall within the outer boundary of the anomaly zone and the verification layer matches the anomaly layer, both location matching and layer matching are determined to be valid, and the verification status of the anomaly zone is updated to "verified". The "verified" status is written to the verification status field of the corresponding hidden disaster-causing factor data object and is saved together with the evidence source and verification record time of the drilling verification data. This update result can be traced back to the corresponding anomaly zone and the corresponding verification point in subsequent hidden disaster-causing factor survey results.

[0037] S403: When the coordinates of the verification point do not fall within the outer boundary of the anomaly area, or the verification layer is inconsistent with the anomaly layer, the verification status of the anomaly area is updated to "pending review". The "pending review" status is written to the verification status field of the corresponding hidden disaster-causing factor data object, and the reason for not meeting the matching conditions is recorded in the "to be verified" field. The "pending review" status is used as a task trigger condition in S6, enabling the on-site verification task to correspond with the differences between the outer boundary of the anomaly area, the anomaly layer, the coordinates of the verification point, and the verification layer.

[0038] S5, based on gas content, gas pressure, and burial depth information from gas parameter data, performs gradient prediction and compares the prediction results with a fixed threshold and a threshold multiple of the regional mean to obtain the gas risk identification result. The inputs to this step are gas parameter data and the maximum burial depth of the mining planning area, and the output is the gas risk identification result. The gas risk identification result is not directly generated from data from a single measuring point, but rather a gradient prediction result is first formed based on gas measurement records under the same mine identification and the same coal seam level, and then the identification result is formed through threshold comparison.

[0039] S501, select gas measurement records from the gas parameter data that are from the same mine and coal seam. Gas measurement records include measurement point coordinates, measurement point depth, gas content, gas pressure, and recording time. During selection, gas measurement records with inconsistent mine identification or coal seam locations are not included in the same gradient prediction process; measurement point depth and recording time are used to determine shallow and deep measurement points and the time sequence of evidence; gas content and gas pressure are used to form gas content gradients and gas pressure gradients, respectively.

[0040] S502, when both shallow and deep measuring points exist in the gas measurement records, the gas content gradient is determined based on the conversion of the gas content difference to the burial depth difference over 100 meters, and the gas pressure gradient is determined based on the conversion of the gas pressure difference to the burial depth difference over 100 meters. Shallow and deep measuring points are determined according to an increasing relationship between their burial depths. The burial depth difference is used to convert the depth difference between the two measuring points to a 100-meter scale. The gas content gradient is formed by mapping the gas content difference to the 100-meter scale, and the gas pressure gradient is formed by mapping the gas pressure difference to the 100-meter scale. The gas content gradient and gas pressure gradient serve as intermediate results in subsequent predictions.

[0041] S503: When no shallow or deep measuring points satisfying the depth-increasing relationship exist in the gas measurement records, the generation of gradient prediction results for the corresponding coal seam level is stopped. When gradient prediction results are stopped, the corresponding gas measurement records are still retained in the gas parameter data, but are not included in the gradient prediction chain of the gas risk identification results. This process avoids using data lacking a depth-increasing basis to form unverifiable prediction results.

[0042] S504: Predicted gas content is generated based on the gas content gradient and the maximum burial depth of the planned mining area; predicted gas pressure is generated based on the gas pressure gradient and the maximum burial depth of the planned mining area. The predicted gas content is based on the gas content at deep measuring points, combined with the increase in burial depth relative to the maximum burial depth of the planned mining area and the gas content gradient. The predicted gas pressure is based on the gas pressure at deep measuring points, combined with the increase in burial depth relative to the maximum burial depth of the planned mining area and the gas pressure gradient. The predicted gas content and predicted gas pressure serve as inputs for threshold comparison.

[0043] S505: First, compare the predicted gas content with a fixed threshold for gas content, and then compare the predicted gas pressure with a fixed threshold for gas pressure. When the fixed threshold comparison is valid, a gas risk identification result is generated for the corresponding coal seam level and mining planning area. When the fixed threshold comparison is invalid, then compare the predicted gas content with a multiple threshold of the regional average gas content, and then compare the predicted gas pressure with a multiple threshold of the regional average gas pressure. Based on the results of the two comparisons, a gas risk identification result is generated. The fixed threshold comparison and the regional average multiple threshold comparison are performed in sequence. If either comparison results in a high gas content area or a high gas pressure area, the corresponding identification content is written into the gas risk identification result.

[0044] S6 generates the results of the hidden disaster-causing factor survey and on-site verification tasks based on the disaster factor correlation results, anomaly area verification status, and gas risk identification results. The inputs to this step are the disaster factor correlation results obtained in S3, the anomaly area verification status obtained in S4, and the gas risk identification results obtained in S5. The outputs are the hidden disaster-causing factor survey results and the on-site verification tasks. The hidden disaster-causing factor survey results record the survey content including the formed data objects, correlations, evidence sources, and verification status. The on-site verification tasks are used to identify risk objects or objects awaiting further on-site verification.

[0045] S601 uses the direct association type and adjacent verification type in the disaster factor association results as the task generation object, the pending review status in the anomaly area verification status as the task triggering condition, and the high gas content area and high gas pressure area in the gas risk identification results as the risk coverage condition. The task generation object is used to limit the scope of objects that can generate on-site verification tasks, the task triggering condition is used to indicate that the object still lacks a verification loop, and the risk coverage condition is used to indicate that there are gas risk identification results in the area corresponding to the object.

[0046] S602, when a task-generated object meets the task triggering condition or risk coverage condition, a field verification task is generated, including the task object, task location, disaster type, verification basis, and task priority. The task object comes from the hidden disaster-causing factor data object corresponding to a directly related type or adjacent verification type; the task location comes from spatial coordinates, the outer boundary of the anomaly zone, the boundary of the mining planning area, or the coordinates of the verification point; the disaster type comes from the disaster type field of the hidden disaster-causing factor data object; the verification basis comes from the evidence source, the set of fields to be verified, the verification status of the anomaly zone, and the gas risk identification results; the task priority is determined based on the hit relationship between directly related types, adjacent verification types, the status to be reviewed, high gas content areas, and high gas pressure areas. If the same task-generated object simultaneously meets both the task triggering condition and the risk coverage condition, the task priority covers task-generated objects that only meet a single condition.

[0047] S603: When the task-generated object does not meet the task triggering conditions and risk coverage conditions, the corresponding hidden disaster-causing factor data object will be written into the hidden disaster-causing factor survey results. The written content includes mine identification, coal seam location, spatial coordinates, elevation range, time attribute, disaster type, evidence source, verification status, and disaster factor correlation results. The hidden disaster-causing factor survey results do not replace the on-site verification task; when subsequent new drilling verification data, on-site investigation data, or gas parameter data enter S1, the relevant hidden disaster-causing factor data objects will be reprocessed through S2 to S6, and updated verification status, gas risk identification results, hidden disaster-causing factor survey results, or on-site verification tasks will be generated based on the new evidence.

[0048] Through steps S1 to S6 above, the data from the survey of hidden mine disaster-causing factors enters a unified data object from raw, multi-source data. This data then undergoes spatial stratification fusion, anomaly zone verification status updates, gas risk identification, and on-site verification tasks, forming a closed loop between input sources, processing rules, intermediate states, output results, and verification basis. The intermediate data generated in each step has a clear source and subsequent use, and can be traced through the evidence source, verification status, and set of fields to be verified within the hidden disaster-causing factor data object.

[0049] The above-described embodiments, including their processes, field states, processing order, judgment conditions, parameter sources, and data correspondences, are used to explain the possible implementations of the present invention and should not be construed as limiting the invention to the specific embodiments listed. Without departing from the scope of the technical solutions described in this invention, any equivalent substitutions, modifications, combinations, order adjustments, equivalent transformations of field names, equivalent inheritance of the executing entity, or equivalent changes in the carrier form that can be conceived by those skilled in the art should fall within the scope of protection of this patent; however, such modifications should not extend to unclaimed topics, nor should they alter the substantive correspondence of the technical objects through name changes.

Claims

1. A method for investigating hidden disaster-causing factors in mines based on multi-source heterogeneous data fusion, characterized in that, Acquire multi-source heterogeneous mine survey data, which includes mine ledger data, mining map data, geophysical anomaly data, drilling verification data, gas parameter data, hydrogeological data, field survey data, and mining planning data; The multi-source heterogeneous mine survey data is converted into hidden disaster-causing factor data objects, which include mine identification, coal seam location, spatial coordinates, elevation range, time attribute, disaster type, evidence source and verification status; Based on spatial coordinates, coal seam level and elevation range, the data objects of the hidden disaster-causing factors are spatially fused with the mining planning area to obtain the disaster factor correlation results. When the anomaly area corresponding to the geophysical anomaly data meets the location matching and stratigraphic matching of the drilling verification data, the anomaly area is marked as verified; when the anomaly area does not meet the location matching or stratigraphic matching, the anomaly area is marked as pending review. Gradient prediction is performed based on gas content, gas pressure and burial depth information in gas parameter data, and the prediction results are compared with fixed threshold and regional mean multiple threshold to obtain gas risk identification results. Based on the disaster factor correlation results, the verification status of abnormal areas, and the gas risk identification results, the results of the survey of hidden disaster-causing factors and the on-site verification tasks are generated.

2. The method for investigating hidden disaster-causing factors in mines based on multi-source heterogeneous data fusion according to claim 1, characterized in that, The process of converting the multi-source heterogeneous mine survey data into hidden disaster-causing factor data objects includes: Extract the goaf number, poorly sealed borehole number, sealing location, formation time, and mine identifier from the mine ledger data; Extract mining boundaries, coal seam floor contour lines, roadway locations, and planned mining areas from mining map data; Extract the boundary of the anomaly zone, coordinates of the verification point, verification layer, and verification record time from geophysical anomaly data and drilling verification data; The extracted data are aligned by field according to mine identification, coal seam location, spatial coordinates, elevation range, time attribute, and disaster type. When the same disaster type meets the consistency condition in spatial coordinates and coal seam location, it is merged into the same hidden disaster-causing factor data object. When the consistency condition is not met, hidden disaster-causing factor data objects with different evidence sources are generated separately.

3. The method for investigating hidden disaster-causing factors in mines based on multi-source heterogeneous data fusion according to claim 2, characterized in that, The field alignment includes coordinate alignment, layer alignment, and time alignment; The coordinate alignment is based on the boundary coordinates of the mining area in the mining map data, and converts the text positions in the mine ledger data and the positioning points in the field survey data into the same plane coordinate system. The alignment of the stratigraphic caliber is constrained by both the coal seam stratum and elevation range, and geophysical anomaly data, drilling verification data, gas parameter data and hydrogeological data are assigned to the corresponding coal seam. The time caliber alignment determines the evidence priority according to the recording time from the most recent to the oldest. When the same hidden disaster-causing factor data object has the same field with different recording times, the field value with the recording time closer to the mining planning data generation time is retained.

4. The method for investigating hidden disaster-causing factors in mines based on multi-source heterogeneous data fusion according to claim 1, characterized in that, The process of spatially merging the data objects of the hidden disaster-causing factors with the mining planning area based on spatial coordinates, coal seam level, and elevation range includes: The data objects of hidden disaster-causing factors corresponding to goaf areas, water accumulation areas, poorly sealed boreholes, faults, water-bearing bodies, water-conducting channels, and gas-rich areas are classified into the disaster factor object set; The mining planning data includes the mining area, tunneling area, and boundary coal pillar area, which are then incorporated into the mining planning object set. Calculate the planar overlap relationship, boundary distance relationship, vertical elevation intersection relationship, and coal seam horizon consistency between the disaster factor object set and the mining planning object set, respectively; When the planar overlap relationship, vertical elevation intersection relationship, and coal seam horizon consistency all meet the association conditions, a disaster factor association result of the direct association type is generated.

5. The method for investigating hidden disaster-causing factors in mines based on multi-source heterogeneous data fusion according to claim 4, characterized in that, When calculating the boundary distance relationship between the disaster factor object set and the mining planning object set, the outer boundary of the mining planning object set is used as the distance calculation benchmark. The shortest distance from the boundary of the sealed poor borehole, fault, water accumulation area and gas-rich area to the outer boundary of the mining planning object set is calculated respectively. When the shortest distance is within the preset verification distance range and the coal seam stratum consistency meets the correlation conditions, the corresponding disaster factor correlation result will be marked as the proximity verification type. When the shortest distance exceeds the preset verification distance range, the evidence source and verification status of the corresponding hidden disaster-causing factor data object are retained, and no disaster factor association results of the adjacent verification type are generated.

6. The method for investigating hidden disaster-causing factors in mines based on multi-source heterogeneous data fusion according to claim 3, characterized in that, After the time caliber is aligned, the data objects of hidden disaster-causing factors are subjected to evidence consistency verification. The evidence consistency verification uses the spatial coordinates falling into the same grid cell, the coal seam level being the same, and the elevation range having an intersection as parallel judgment conditions. When at least two types of evidence sources for the same hidden disaster-causing factor data object meet the parallel judgment conditions, the verification status will be updated to the review and confirmation status. When only one type of evidence source meets the parallel judgment condition for the same hidden disaster-causing factor data object, the verification status is retained as pending review, and the fields that do not meet the parallel judgment condition are written into the field set to be verified.

7. The method for investigating hidden disaster-causing factors in mines based on multi-source heterogeneous data fusion according to claim 1, characterized in that, The process of marking the abnormal region as the verified state includes: Extract the outer boundary and anomalous strata of the anomalous area from the geophysical anomaly data, and extract the coordinates of the verification point, the verification strata, and the verification record time from the drilling verification data; When the coordinates of the verification point fall within the outer boundary of the abnormal area and the verification layer is consistent with the abnormal layer, it is determined that both the position matching and the layer matching are valid, and the verification status of the abnormal area is updated to the verified status. When the coordinates of the verification point do not fall within the outer boundary of the abnormal area, or when the verification layer is inconsistent with the abnormal layer, the verification status of the abnormal area is updated to the status to be reviewed.

8. The method for investigating hidden disaster-causing factors in mines based on multi-source heterogeneous data fusion according to claim 1, characterized in that, The process of gradient prediction based on the gas content, gas pressure, and burial depth information in the gas parameter data includes: From the gas parameter data, gas measurement records under the same mine identification and the same coal seam are selected. The gas measurement records include the coordinates of the measuring point, the burial depth of the measuring point, the gas content, the gas pressure, and the recording time. When there are shallow and deep measuring points in the gas measurement record, the gas content gradient is determined according to the conversion result of the gas content difference and the burial depth difference per 100 meters, and the gas pressure gradient is determined according to the conversion result of the gas pressure difference and the burial depth difference per 100 meters. When there are no shallow or deep measuring points in the gas measurement record that satisfy the increasing burial depth relationship, the generation of gradient prediction results for the corresponding coal seam level is stopped.

9. The method for investigating hidden disaster-causing factors in mines based on multi-source heterogeneous data fusion according to claim 8, characterized in that, The process of comparing the prediction results with a fixed threshold and a threshold that is a multiple of the regional mean includes: The gas content is predicted based on the gas content gradient and the maximum burial depth of the mining planning area, and the gas pressure is predicted based on the gas pressure gradient and the maximum burial depth of the mining planning area. First, compare the predicted gas content with a fixed threshold for gas content, and then compare the predicted gas pressure with a fixed threshold for gas pressure. When the fixed threshold comparison is not valid, the predicted gas content is compared with the multiple threshold of the regional average gas content, and the predicted gas pressure is compared with the multiple threshold of the regional average gas pressure. The gas risk identification result is generated based on the results of the two comparisons.

10. The method for investigating hidden disaster-causing factors in mines based on multi-source heterogeneous data fusion according to claim 1, characterized in that, The process of generating the survey results of hidden disaster-causing factors and the on-site verification task based on the disaster factor correlation results, the anomaly area verification status, and the gas risk identification results includes: The direct association type and adjacent verification type in the disaster factor association results are used as task generation objects, the pending review status in the abnormal area verification status is used as task triggering conditions, and the high gas content area and high gas pressure area in the gas risk identification results are used as risk coverage conditions. When the task generation object meets the task triggering conditions or risk coverage conditions, an on-site verification task is generated, which includes the task object, task location, disaster type, verification basis, and task priority. When the task-generated object does not meet the task triggering conditions and the risk coverage conditions, the corresponding hidden disaster-causing factor data object will be written into the hidden disaster-causing factor survey results.

Images