A method and system for detecting a hidden fire zone in an underground coal mine

Abstract

The application discloses a kind of underground coal mine hidden fire area detection method and system, related to coal mine fire detection technical field, to the continuous variation amplitude of underground gas concentration in space discrete position in underground coal mine, airflow deflection angle change and local pressure difference fluctuation scale are collected and handled, continuous variation amplitude, airflow deflection angle change and local pressure difference fluctuation scale are converted into multistage drift scale, based on multistage drift scale, gas abnormal jump sensitive interval is demarcated, and diffusion traction direction mark is labeled in each space discrete position.The application demarcates abnormal jump sensitive interval by multistage drift scale, realizes the continuous identification and space aggregation of gas anomaly, avoids abnormal point dispersion, improves the consistency and range definition stability of fire area positioning;And by subsection adjustment, reverse compensation and direction convergence processing, realize abnormal distribution continuous and unified reconstruction, determine fire source area and expansion direction, improve the reliability and management pertinence of fire evolution determination.

Description

Technical Field

[0001] This application relates to the field of coal mine fire detection technology, specifically to a method and system for detecting concealed fire zones in underground coal mines. Background Technology

[0002] Underground coal mines refer to a type of mine that mines by entering the coal seam through underground development methods such as shafts and inclined shafts. Their operating space is located deep underground, including goaf areas, roadways, coal pillars, and various concealed structural areas. They are characterized by strong spatial enclosure, complex ventilation conditions, and frequent geological disturbances. In such environments, concealed fire zones often develop latently through low-temperature oxidation or slow heat accumulation, lacking obvious external characteristics in the early stages, making them easily overlooked and gradually evolving into large-scale disaster sources. Therefore, introducing intelligent sensing systems to detect concealed fire zones in underground coal mines can continuously perceive and collaboratively interpret temperature changes, gas anomalies, and changes in strata properties. This allows for accurate identification of abnormal areas and definition of fire zone boundaries. Simultaneously, it enables forward-looking analysis of fire evolution trends, significantly improving the early identification capability of hidden dangers and the level of disaster warning, providing key technical support for safe mine operation and precise fire control.

[0003] During the detection of concealed fire zones in underground coal mines, when complex disturbances occur in the underground ventilation system, causing unstable changes in the direction and velocity of airflow in the local space, the abnormal gas signals, influenced by the airflow pull, will exhibit discontinuous distribution and abrupt diffusion characteristics during propagation. In this situation, if the intelligent sensing system, when sensing and interpreting changes in gas concentration, still relies on the assumption of continuous gas concentration changes, it is prone to misjudging originally correlated abnormal gas distributions as multiple scattered independent anomaly points, thus failing to effectively establish the spatial correlation between these anomaly points.

[0004] As the above state continues to evolve, the spatial continuity of abnormal signals is broken, and the system's perception of the overall distribution pattern of hidden fire zones gradually deviates from the actual situation. Ultimately, this leads to the abnormal area corresponding to the same fire source being incorrectly divided into multiple independent areas, resulting in a fragmented expression of the fire zone range.

[0005] This not only makes it difficult to accurately locate the center of the fire, but also misleads the development trend of the fire zone in the subsequent fire situation assessment process, thereby affecting the accuracy of the identification results of hidden fire zones and the reliability of subsequent governance decisions.

[0006] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0007] The purpose of this application is to provide a method and system for detecting concealed fire zones in underground coal mines, in order to solve the problems mentioned in the background art.

[0008] To achieve the above objectives, this application provides the following technical solution: a method for detecting concealed fire zones in underground coal mines, comprising the following steps:

[0009] The continuous variation amplitude, airflow offset angle variation, and local pressure difference fluctuation scale of underground gas concentration at spatially discrete locations in coal mines are collected and processed. The continuous variation amplitude, airflow offset angle variation, and local pressure difference fluctuation scale are converted into multi-level drift scales. Based on the multi-level drift scales, the sensitive interval of gas abnormal jump is defined, and diffusion traction direction is marked at each spatially discrete location.

[0010] Segmented release regulation is performed on the gas concentration distribution within the sensitive range of gas anomaly jumps. A reverse diffusion compensation amount is introduced at the spatial discrete position where the diffusion traction direction marker is offset and exceeds the preset scale. The abnormal gas distribution is continuously repaired to obtain a set of controlled anomaly associations.

[0011] Based on the controlled anomaly association set, the spatial discrete locations are subjected to misalignment stretching and rearrangement processing. The originally isolated anomaly points are extended by gradient according to the diffusion traction direction. At the same time, continuous spacing constraints are applied between adjacent spatial discrete locations to construct a continuous diffusion distribution pattern.

[0012] Based on the continuous diffusion distribution pattern, directional convergence traction adjustment is implemented along each spatial discrete position. An angle limit range is set according to the offset trend of the diffusion traction direction indicator. Reverse traction is superimposed at spatial discrete positions that deviate from the main diffusion direction to limit abnormal diffusion offset.

[0013] Based on the abnormal distribution state after directional convergence, the overall gas abnormal region is subjected to contour splicing and deduction processing to reconstruct the dispersed abnormal region into a single continuous abnormal segment, thereby determining the concentrated area of ​​the fire source and its expansion direction.

[0014] Preferably, the sensitive range for gas anomaly jumps is defined based on a multi-level drift scale, and the specific steps are as follows:

[0015] Based on the direction of roadway extension and the spatial distribution of goaf in the underground working area, the monitoring area is divided into spatially discrete locations. At each spatially discrete location, gas concentration is recorded, airflow direction angle is obtained, and pressure difference is recorded and processed. Based on this, the continuous change amplitude of gas concentration, change of airflow offset angle, and local pressure difference fluctuation scale are obtained.

[0016] The scale of continuous change in gas concentration, change in airflow offset angle, and local pressure difference fluctuation is divided into intervals. At each spatial discrete position, the gas concentration level, airflow offset angle level, and local pressure difference level are compared item by item. The maximum level is selected as the comprehensive level and defined as a multi-level drift scale.

[0017] By comparing the differences between adjacent spatial discrete positions based on multi-level drift scales, positions that meet the difference conditions are marked as abnormal jump points and gradually expanded to form a gas abnormal jump sensitive area. At the same time, the direction of the change in airflow offset angle of spatial discrete positions in the area is extracted and the diffusion traction direction is marked, thereby completing the delineation of the gas abnormal jump sensitive area based on multi-level drift scales.

[0018] Preferably, continuous repair processing is performed on the abnormal gas distribution to obtain a controlled abnormality association set. The specific steps are as follows:

[0019] The gas anomaly jump sensitive interval is traversed in the spatial discrete position numbering order. The continuous spatial discrete positions form the processing unit. The gas concentration of the reference spatial discrete position is used to perform difference comparison on the subsequent spatial discrete positions and perform segmented release adjustment so that the gas concentration converges to the reference value and maintains unidirectional change, thereby obtaining a continuous monotonic gas concentration distribution.

[0020] Based on the gas concentration distribution after segmented release adjustment, the angle difference of the diffusion traction direction markers at each spatial discrete position is determined. At the spatial discrete position where the offset occurs, the gas concentration is extracted from the adjacent spatial discrete position in the opposite direction and superimposed to form the reverse diffusion compensation amount. The offset position is then continuously repaired.

[0021] After introducing the reverse diffusion compensation, the spatial discrete positions within the gas anomaly jump sensitive range are sequentially scanned. The correlation chain is constructed starting from the spatial discrete position and is connected and truncated according to the gas concentration difference and the difference in the diffusion traction direction indicator angle to form a controlled anomaly correlation set.

[0022] Preferably, in the step of determining the angle difference of the diffusion traction direction markers at each spatial discrete location and introducing the reverse diffusion compensation amount: for the spatial discrete location where the offset occurs, spatial positioning is performed along the reverse angle direction of the diffusion traction direction marker, gas concentration is extracted from adjacent spatial discrete locations and superimposed on the current spatial discrete location, and after the compensation is completed, gas concentration extraction and superposition processing is performed along the same reverse path until the angle difference of the diffusion traction direction markers meets the preset conditions and the compensation path remains consistent in direction.

[0023] Preferably, a continuous spacing constraint is applied between adjacent spatial discrete locations to construct a continuous diffusion distribution pattern. The specific steps are as follows:

[0024] Based on the controlled anomaly association set, the starting extension point of the anomaly association chain is selected according to the spatial discrete position numbering order. The gas concentration value and diffusion traction direction mark are extracted as extension references. The spatial discrete position is scanned in one direction along the direction corresponding to the diffusion traction direction mark. The spatial discrete position that meets the direction difference condition is selected as the extension object. The extension reference is updated in one direction according to the gas concentration value difference to form a continuous gradient extension path.

[0025] Based on the completed gradient extension path, the spatial discrete positions are sequentially reordered by staggered stretching. By comparing the actual spatial distance between adjacent spatial discrete positions with the minimum continuous spacing, the spatial discrete positions are translated or inserted into intermediate spatial discrete positions so that the spacing constraint relationship between adjacent spatial discrete positions is satisfied.

[0026] Based on the results of the misaligned stretching and rearrangement process, a unidirectional scan is performed on the spatial discrete positions. A unified correction process is then applied to the spatial distance between adjacent spatial discrete positions, and the gas concentration values ​​are adjusted simultaneously to ensure that the gas concentration remains monotonically changing along the spatial arrangement direction, thereby constructing a continuous diffusion distribution pattern.

[0027] Preferably, the process of unidirectionally scanning the spatial discrete positions along the direction corresponding to the diffusion traction direction and selecting the spatial discrete positions whose directional differences meet the conditions as the extension objects further includes: when there are multiple spatial discrete positions that meet the conditions, only the spatial discrete position with the smallest number is selected as the unique extension object, and the value is kept to change in one direction without reverse correction during the gas concentration value adjustment process, until the difference between the extension object and the extension reference meets the preset relationship, thereby ensuring the uniqueness and continuity of the gradient extension path.

[0028] Preferably, based on the continuous diffusion distribution pattern, directional convergence traction adjustment is implemented along each spatial discrete position. The specific steps are as follows:

[0029] The continuous diffusion distribution pattern is traversed in a unidirectional manner according to the spatial discrete position number from small to large. The diffusion traction direction of the spatial discrete position with the smallest number is determined as the main diffusion direction. Based on the main diffusion direction, the angle difference of the diffusion traction direction of the subsequent spatial discrete positions is calculated to complete the initial direction classification.

[0030] For the spatially discrete positions after the initial direction classification is completed, the diffusion traction direction markers that deviate from the main diffusion direction are adjusted to the boundary of the angle limiting interval along the minimum rotation path to obtain the diffusion traction direction markers after unified constraints, while keeping the gas concentration value unchanged.

[0031] Based on the unified constraint of the diffusion traction direction identifier, the reverse traction amount superposition process is performed on each spatial discrete position, and unidirectional convergence is carried out along the spatial discrete position numbering order, so that the diffusion traction direction identifier forms a single convergence direction and maintains the continuous change of gas concentration value, so as to complete the limitation of abnormal diffusion offset.

[0032] Preferably, during the process of adjusting the diffusion traction direction marker that deviates from the main diffusion direction along the minimum rotation path to the boundary of the angle limiting range, only the diffusion traction direction marker is subjected to angle correction processing, and after the adjustment is completed, the corrected diffusion traction direction marker is used as the unique direction expression of the spatial discrete position, so that the diffusion traction direction markers of all spatial discrete positions are restricted to a uniform angle range.

[0033] Preferably, the scattered anomalous areas are reconstructed into a single continuous anomalous segment, thereby determining the concentrated area of ​​fire sources and the direction of their spread. The specific steps are as follows:

[0034] The spatial discrete positions in the abnormal distribution state after directional convergence are traversed in numerical order, the gas concentration values ​​are normalized and mapped, and the spatial discrete positions where the normalized gas concentration values ​​meet the preset conditions are determined as anomaly points. The anomaly points are used as the starting splicing points to construct the initial splicing path along the direction of increasing number.

[0035] Based on the initial splicing path obtained, the spatial discrete positions at the end and the spatial discrete positions at the beginning of adjacent paths are selected as connection endpoints. Path connection processing is performed according to the spatial distance between the two endpoints, and the spatial discrete positions after connection are rearranged in numerical order to obtain a set of candidate paths for continuous abnormal segments.

[0036] For the candidate path set of continuous abnormal sections, the spatial discrete position with the largest gas concentration value in the path is selected as the center point of the fire source concentration area. The expansion direction is determined along the numbering direction based on the center point, thereby realizing the reconstruction of the dispersed abnormal area into a single continuous abnormal section and determining the fire source concentration area and expansion direction.

[0037] A system for detecting hidden fire zones in underground coal mines includes a multi-level drift scale construction module, an abnormal distribution continuous repair module, a diffusion pattern rearrangement construction module, a directional convergence traction adjustment module, and a fire zone contour deduction and judgment module.

[0038] The multi-level drift scale construction module collects and processes the continuous variation amplitude, airflow offset angle change, and local pressure difference fluctuation scale of underground coal mine gas concentration at spatially discrete locations. It converts the continuous variation amplitude, airflow offset angle change, and local pressure difference fluctuation scale into a multi-level drift scale. Based on the multi-level drift scale, it delineates the sensitive range of abnormal gas jumps and marks the diffusion traction direction at each spatially discrete location.

[0039] The abnormal distribution continuous repair module performs segmented release adjustment on the gas concentration distribution within the sensitive range of abnormal gas changes. It introduces reverse diffusion compensation at spatial discrete positions where the diffusion traction direction marker is offset and exceeds the preset scale, performs continuous repair processing on the abnormal gas distribution, and obtains a set of controlled abnormal associations.

[0040] The diffusion pattern rearrangement construction module performs misaligned stretching rearrangement on each spatial discrete location based on the controlled anomaly association set. The originally isolated anomaly points are gradient extended according to the diffusion traction direction. At the same time, continuous spacing constraints are applied between adjacent spatial discrete locations to construct a continuous diffusion distribution pattern.

[0041] The directional convergence traction adjustment module performs directional convergence traction adjustment along each spatial discrete position based on the continuous diffusion distribution pattern. It sets the angle limit range according to the offset trend of the diffusion traction direction indicator and superimposes the reverse traction amount at the spatial discrete position that deviates from the main diffusion direction to limit abnormal diffusion offset.

[0042] The fire zone contour deduction and judgment module performs contour splicing and deduction processing on the overall gas anomaly area based on the abnormal distribution state after directional convergence, reconstructs the dispersed abnormal area into a single continuous abnormal segment, and determines the concentrated area of ​​fire source and its expansion direction accordingly.

[0043] The technical effects and advantages provided by this application in the above technical solution are as follows:

[0044] This application constructs a multi-level drift scale and delineates sensitive intervals for gas anomaly jumps accordingly, enabling the correlation and expression of downhole gas concentration changes, gas flow deviation states, and pressure difference disturbance characteristics on a unified scale. Even when gas flow disturbances cause discontinuous distribution of abnormal signals, continuous identification and spatial aggregation of abnormal changes can still be achieved, thereby avoiding the situation where gas anomalies are split into multiple independent points. This ensures that the spatial distribution of concealed fire zones can be presented in a holistic form, which is beneficial to improving the consistency of anomaly zone location and the stability of fire zone range definition.

[0045] This application utilizes a synergistic approach of segmented release regulation, reverse diffusion compensation, and directional convergence traction regulation to transform the spatial distribution of abnormal gases from a discrete state to a continuous diffusion pattern. Based on this, it achieves a unified reconstruction of the abnormal region through contour splicing and deduction, enabling the determination of the concentrated area of ​​the fire source and its expansion direction under a continuous distribution state. This reduces the diffusion offset caused by airflow interference, resulting in a more consistent evolution path for the fire development trend, which is conducive to the targeted formulation and implementation of subsequent governance strategies. Attached Figure Description

[0046] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this application. For those skilled in the art, other drawings can be obtained based on these drawings.

[0047] Figure 1 This is a flowchart illustrating the overall method of this application.

[0048] Figure 2 This is a flowchart illustrating the process of defining the gas abnormal change sensitive range based on a multi-level drift scale in this application.

[0049] Figure 3 This is a flowchart for obtaining the controlled exception association set for this application.

[0050] Figure 4 This is a schematic diagram of the system modules of this application. Detailed Implementation

[0051] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the description of this disclosure will be more complete and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0052] like Figures 1 to 3 As shown, this application provides a method for detecting concealed fire zones in underground coal mines, including the following steps:

[0053] Step 1: Collect and process the continuous variation amplitude, airflow offset angle variation, and local pressure difference fluctuation scale of underground coal mine gas concentration at spatially discrete locations. Convert the continuous variation amplitude, airflow offset angle variation, and local pressure difference fluctuation scale into multi-level drift scales. Delineate the sensitive interval of abnormal gas change based on the multi-level drift scales, and mark the diffusion traction direction at each spatially discrete location.

[0054] The sensitive range for gas anomalies is determined based on a multi-level drift scale. The specific steps are as follows:

[0055] Within the underground working area, based on the direction of the roadway extension and the spatial distribution of the goaf, the monitoring area is divided into spatially discrete locations at fixed intervals. The fixed interval is any value within the range of 5 to 10 meters, so that a continuous coverage relationship is formed between adjacent spatially discrete locations.

[0056] Gas concentration is recorded at each spatially discrete location at fixed time intervals. The fixed time interval is any value within the range of 30 to 120 seconds. The difference between the gas concentration values ​​of two adjacent time nodes is calculated, and the differences within five consecutive time intervals are accumulated to obtain the continuous change amplitude of gas concentration.

[0057] For the airflow direction, the airflow direction angle is recorded at each time point, and the angle between the airflow direction at the current time and the previous time is taken as the airflow offset angle. The angle changes over three consecutive time intervals are accumulated and summed to obtain the change in airflow offset angle.

[0058] Within the same time scale, the pressure difference at spatially discrete locations is recorded. The difference between the maximum and minimum pressure difference over four consecutive time intervals is taken as the local pressure difference fluctuation range. The pressure difference range is divided into three levels: 0 to 20 Pa, 20 to 50 Pa, and above 50 Pa, each corresponding to a different local pressure difference fluctuation scale.

[0059] Through the above processing, the amplitude of continuous change in gas concentration, the change in airflow offset angle, and the scale of local pressure difference fluctuation are made to correspond within a unified time window.

[0060] After obtaining the continuous variation amplitude, airflow offset angle variation, and local pressure difference fluctuation scale at each spatially discrete location, the continuous variation amplitude of gas concentration is divided into intervals: 0 to 0.3 is classified as Level 1, 0.3 to 0.7 as Level 2, and above 0.7 as Level 3. The airflow offset angle variation is also divided into angular intervals: 0 to 15 degrees is Level 1, 15 to 45 degrees as Level 2, and above 45 degrees as Level 3. The local pressure difference fluctuation scale is directly mapped to these three levels: Level 1, Level 2, and Level 3.

[0061] At each spatially discrete location, the gas concentration level, airflow offset angle level, and local pressure difference level are compared item by item, and the highest level among the three is selected as the comprehensive level of that location. This comprehensive level is defined as a multi-level drift scale.

[0062] By uniformly adopting the highest level, the multi-level drift scale can reflect the factors most affected by disturbances during gas transport, thereby ensuring that the drift scale at each spatial discrete position has unique determinism.

[0063] After assigning values ​​to the multi-level drift scale, each spatial discrete position is compared one by one with its adjacent spatial discrete positions in all directions. The difference in multi-level drift scale between adjacent positions is calculated. When the difference reaches two or more levels, the position is marked as an anomalous jump point. Starting from the anomalous jump point, the range is gradually expanded to its adjacent spatial discrete positions. When the difference in multi-level drift scale between the current point and the adjacent position is consistently greater than or equal to one level, the adjacent position is included in the same region. The expansion stops when an adjacent position with a difference of 0 is found. This determines the boundary of the gas anomalous jump sensitive region.

[0064] During the regional expansion process, each spatially discrete location within the included range is repeatedly assessed to ensure the complete closure of the boundary range, thereby giving the gas anomaly change sensitive area a clear spatial outline.

[0065] After the sensitive range of gas anomaly jump is determined, the direction of the change in airflow offset angle at each spatial discrete position in the region is extracted. The angle change with the same direction of airflow offset angle change in three consecutive time intervals is taken as the dominant direction. When there are multiple directional changes, the number of occurrences of each direction is counted, and the direction with the most occurrences is determined as the dominant direction.

[0066] When two directions appear the same number of times, the direction with the larger total angle change is selected as the final direction. The dominant direction is mapped to a pointing angle in space, and this pointing angle is used as the diffusion traction direction identifier and assigned to the corresponding discrete spatial position. Between adjacent discrete spatial positions, the direction identifiers are aligned. When the direction difference between two adjacent positions is less than 15 degrees, their directions are adjusted to a unified direction. When the direction difference is greater than 15 degrees, their respective direction identifiers are retained. This ensures the continuous expression of the diffusion traction direction identifier in space and provides a stable directional basis for the subsequent continuous recovery of the distribution of anomalous gases.

[0067] Step 2: Perform segmented release adjustment on the gas concentration distribution within the sensitive range of gas anomaly jumps. Introduce reverse diffusion compensation at spatial discrete positions where the diffusion traction direction marker shifts and exceeds the preset scale to continuously repair the abnormal gas distribution and obtain a set of controlled anomaly associations.

[0068] The abnormal gas distribution is continuously repaired to obtain a controlled set of anomaly associations. The specific steps are as follows:

[0069] Within the sensitive range of gas anomaly jumps, a one-way traversal is performed according to the spatial discrete position numbering order from smallest to largest. The first spatial discrete position is taken as the starting position, and three consecutive spatial discrete positions form a fixed-length processing unit. The processing unit is always selected in the order of current position, subsequent position, and next-to-last position, without skipping any positions.

[0070] When there are fewer than three remaining spatial discrete locations, the last two spatial discrete locations are used as the final processing units. In each processing unit, the gas concentration at the first spatial discrete location is used as a reference value. The gas concentrations at the second and third spatial discrete locations are calculated by subtracting the reference value from the gas concentration at each location. When the difference is positive, it indicates that the gas concentration at that location is higher than the reference value. In this case, the gas concentration at that location is gradually decreased by a fixed step size of 0.05. When the difference is negative, it indicates that the gas concentration at that location is lower than the reference value. In this case, the gas concentration at that location is gradually increased by the same step size until the absolute value of the difference between that location and the reference value does not exceed 0.2.

[0071] During the adjustment process, the value at each position is updated synchronously with each step change, and the adjustment direction is always converged towards the reference value; reverse adjustment is not allowed. After completing the current processing unit, the entire processing unit is moved backward by one spatially discrete position and the above process is repeated until the entire gas anomaly jump sensitive range has been traversed, thus making the gas concentration exhibit a continuous monotonic change in space.

[0072] Based on the results after the segmented release adjustment, the diffusion traction direction marker is offset according to fixed rules. Taking each spatial discrete position as the current processing point, the angle difference between the diffusion traction direction marker at this position and the diffusion traction direction marker at the previous spatial discrete position is calculated. The angle difference is uniformly determined by the minimum included angle method. When the angle difference is greater than 20 degrees, the position is marked as the offset position. For each offset position, spatial positioning is performed along the opposite angle direction of the diffusion traction direction marker at that position. The opposite direction is defined as the direction obtained by normalizing the current direction angle plus 180 degrees within the range of 0 to 360 degrees.

[0073] Two adjacent spatial discrete positions are strictly selected along the reverse direction. If there is only one spatial discrete position in the reverse direction, only that position is selected. If there is no corresponding position, the compensation process for that position is terminated. 20% of the gas concentration value of the first selected reverse spatial discrete position is extracted as the first compensation amount, and 15% of the gas concentration value of the second reverse spatial discrete position is extracted as the second compensation amount. The first compensation amount and the second compensation amount are added together and superimposed on the gas concentration value of the current offset position.

[0074] After compensation is completed, the directional angle difference between the current position and the previous position is recalculated. If it is still greater than 20 degrees, the next spatial discrete position is selected along the same reverse path, and its gas concentration is extracted at a ratio of 10% as an additional compensation amount for superposition. This process is repeated until the directional angle difference is less than or equal to 20 degrees, and the compensation path must not change direction, thus ensuring that the compensation process is uniquely determined.

[0075] After the segmented release regulation and reverse diffusion compensation are introduced, the spatial discrete positions within the entire gas abnormal jump sensitive range are sequentially scanned. Starting from the position with the smallest number, this position is set as the starting point of the current association chain, and its gas concentration and diffusion traction direction indicator are used as the initial reference. Then, its subsequent spatial discrete positions are checked. When the gas concentration difference between the subsequent position and the current starting point does not exceed 0.2 and the angle difference of the diffusion traction direction indicator does not exceed 20 degrees, the subsequent position is included in the current association chain and updated as the new current reference position.

[0076] If the gas concentration difference exceeds 0.2 or the directional angle difference exceeds 20 degrees, the current association chain construction is terminated, and the subsequent position is used as the starting point for a new association chain construction. Within each association chain, the gas concentration is required to maintain a monotonically changing direction along the spatial arrangement direction. If the gas concentration at a certain position changes in the opposite direction to the previous position, the association chain is truncated at that position.

[0077] Through the above point-by-point judgment and series processing, the entire gas anomaly jump sensitive range is divided into multiple continuous anomaly association chains. All association chains are then aggregated and organized in spatial order to form a controlled anomaly association set, so that the spatial relationship between each anomaly point has continuity and directional consistency after processing.

[0078] Step 3: Based on the controlled anomaly association set, perform misalignment stretching and rearrangement processing on each spatial discrete location. The originally isolated anomaly points are extended by gradient according to the diffusion traction direction. At the same time, continuous spacing constraints are applied between adjacent spatial discrete locations to construct a continuous diffusion distribution pattern.

[0079] To construct a continuous diffusion distribution pattern, continuous spacing constraints are applied between adjacent spatial discrete locations. The specific steps are as follows:

[0080] Once the controlled anomaly association set is determined, each anomaly association chain is processed in ascending order of its spatial discrete location number. The spatial discrete location with the smallest number is fixed as the starting extension point, and no reverse selection is performed. At this starting extension point, the gas concentration value and the diffusion traction direction identifier are extracted as the unique extension reference direction.

[0081] Subsequently, a unidirectional scan is performed along the spatial direction corresponding to the diffusion traction direction marker. The scanning order is fixed to search point by point in the direction of increasing spatial discrete position number. Among all candidate spatial discrete positions, only the spatial discrete position with the smallest number and an angle difference of no more than 15 degrees between the diffusion traction direction marker and the current extension reference direction is selected as the unique extension object. Multiple candidate points are not allowed to be selected at the same time.

[0082] Once the target for extension is determined, its gas concentration value is compared with the current extension reference point. When the gas concentration value of the target is lower than the reference value, the value is increased step by step in increments of 0.03. When the gas concentration value of the target is higher than the reference value, the value is decreased step by step in increments of 0.03. During the adjustment process, the value changes in one direction only and no reverse correction is allowed, until the absolute value of the difference between the two values ​​equals 0.15. When the adjustment is complete, the target is updated as the new extension reference point, and the extension process continues to be executed according to the same rules until all spatially discrete positions in the current anomaly association chain have been processed. This allows the originally isolated anomaly points to form a continuous gradient extension path along a uniquely determined diffusion traction direction.

[0083] After completing the gradient extension process, a misalignment stretching and rearrangement process is performed on each abnormal association chain. The processing order is fixed to proceed point by point along the direction of spatial discrete position number from small to large. At each spatial discrete position, the actual spatial distance between the position and the previous spatial discrete position is compared with the fixed minimum continuous spacing of 4 meters.

[0084] When the actual distance is less than 4 meters, the current spatial discrete position is translated along the direction corresponding to the diffusion traction direction marker. The translation distance is exactly equal to 4 meters minus the current actual distance, ensuring that the adjusted spatial spacing is strictly equal to 4 meters. When the actual distance is greater than 4 meters but not exceeding 8 meters, the current spatial discrete position is translated 1 meter towards the previous spatial discrete position, gradually bringing the spacing closer to 4 meters. When the actual distance is greater than 8 meters, only one intermediate spatial discrete position is inserted between the two positions for calculation. This intermediate spatial discrete position is only used for spacing equalization in this step. At the inserted position, the gas concentration value is taken as the arithmetic mean of the gas concentration values ​​of the two spatial discrete positions before and after, and the diffusion traction direction marker is taken as the direction marker of the previous spatial discrete position. After the insertion is completed, it immediately participates in the current spacing judgment process, but this intermediate spatial discrete position does not participate in any subsequent steps after this step is completed.

[0085] When the distance between three consecutive spatially discrete positions is less than 4 meters, the middle position is fixed and adjusted first, and then the next position is adjusted. This avoids spatial overlap or intersection and ensures that the misalignment stretching process maintains a unique spatial order.

[0086] Once the misalignment, stretching, and rearrangement of all spatial discrete positions are completed, a unidirectional scan is performed on the entire abnormal association chain. The scanning direction is fixed in the direction of increasing numbering. During the scan, the spatial distance between adjacent spatial discrete positions is checked pair by pair. When the distance between any pair of adjacent positions is not equal to 4 meters, a unified correction process is performed.

[0087] If the distance is less than 4 meters, the next spatial discrete position is translated along the direction of the diffusion traction to a position with a distance of 4 meters; if the distance is greater than 4 meters, the next spatial discrete position is translated along the direction of the previous position to a position with a distance of 4 meters. It is not allowed to insert a new intermediate spatial discrete position again.

[0088] While completing the spatial distance correction, the gas concentration value at the corresponding position is synchronously adjusted according to a fixed step size of 0.02, so that the gas concentration between adjacent spatial discrete positions always maintains a monotonically increasing or monotonically decreasing relationship. If a position is detected to have a change direction opposite to the previous position, the gas concentration at that position is adjusted to a value with a difference of 0.02 from the previous position and keeps changing in the same direction.

[0089] After completing all the scanning processes, all spatial discrete positions are rearranged in numerical order to obtain the spatial distribution state where the distance between each spatial discrete position is strictly equal to 4 meters and the gas concentration changes continuously, thus uniquely determining the continuous diffusion distribution pattern.

[0090] Step 4: Based on the continuous diffusion distribution pattern, implement directional convergence traction adjustment along each spatial discrete position, set the angle limit range according to the offset trend of the diffusion traction direction mark, and superimpose the reverse traction amount at the spatial discrete position that deviates from the main diffusion direction to limit abnormal diffusion offset.

[0091] Based on the continuous diffusion distribution pattern, directional convergence and traction adjustment are implemented along each spatial discrete location. The specific steps are as follows:

[0092] After the continuous diffusion distribution pattern is formed, a one-way traversal is performed according to the spatial discrete position numbers in ascending order. The diffusion traction direction identifier of the spatial discrete position with the smallest number is directly determined as the unique main diffusion direction, and this main diffusion direction is kept unchanged in subsequent processing. During the traversal, the diffusion traction direction identifier of each subsequent spatial discrete position is extracted, and the angle difference between it and the unique main diffusion direction is calculated. The angle difference is uniformly represented by the minimum included angle between the two directions.

[0093] When the angle difference is less than or equal to 10 degrees, the diffusion traction direction marker of the spatial discrete position is directly corrected to the main diffusion direction; when the angle difference is greater than 10 degrees and less than or equal to 20 degrees, the diffusion traction direction marker of the spatial discrete position is not corrected, and the angle difference is recorded as the offset trend value; when the angle difference is greater than 20 degrees, the spatial discrete position is marked as a position deviating from the main diffusion direction and enters the subsequent processing steps, so that the spatial discrete positions in the entire continuous diffusion distribution pattern complete the initial direction classification according to the unified rules.

[0094] An angle limiting interval constraint is applied to the spatial discrete positions marked as deviating from the main diffusion direction. The angle limiting interval is fixed within ±20 degrees of the main diffusion direction. For each spatial discrete position deviating from the main diffusion direction, its current diffusion traction direction indicator is adjusted along the minimum rotation path to a position exactly 20 degrees away from the main diffusion direction, so that its direction indicator strictly falls on the boundary of the angle limiting interval and is not allowed to fall outside the interval or other positions inside the interval.

[0095] During the adjustment process, no changes are made to the gas concentration values; only the angle of the diffusion traction direction indicator is corrected. After the adjustment is completed, the corrected diffusion traction direction indicator is used as the unique directional representation for that spatial discrete location, thus restricting the diffusion traction direction indicators of all spatial discrete locations to a uniform angular range.

[0096] After the angle-limited range constraint processing is completed, reverse traction superposition processing is performed on all spatial discrete positions. The reverse traction is limited to the gas concentration value, and the restriction on abnormal diffusion offset is achieved only by adjusting the gas concentration value. Specifically:

[0097] For each spatial discrete location, the angle difference between its diffusion traction direction indicator and the main diffusion direction is recalculated. When the angle difference is greater than 10 degrees and less than or equal to 20 degrees, a first-level reverse traction amount is introduced at that spatial discrete location. The reverse traction amount is 10% of the current gas concentration value at that location, and the value is transferred from the current spatial discrete location to its previous spatial discrete location in the main diffusion direction according to a fixed rule.

[0098] When the angular difference equals 20 degrees, a second-level reverse traction is introduced at that spatial discrete position. The reverse traction is taken as 20% of the current gas concentration at that position, and the same transfer rules are applied. During the transfer, if the previous spatial discrete position in the main diffusion direction does not exist, the reverse traction is not superimposed at that spatial discrete position. After completing one transfer, the reverse traction is not superimposed on the same spatial discrete position, thus ensuring the uniqueness and non-repeatability of the adjustment process.

[0099] After completing the reverse traction superposition processing, a one-way convergence and reorganization process is performed on the entire continuous diffusion distribution pattern. The processing order is fixed as follows: the spatial discrete position numbers are processed in ascending order. At each spatial discrete position, the diffusion traction direction indicator is compared with the diffusion traction direction indicator of the previous spatial discrete position. When the angle difference between the two is greater than 10 degrees, the diffusion traction direction indicator of the current spatial discrete position is forcibly corrected to be consistent with the previous spatial discrete position. At the same time, the gas concentration value of the spatial discrete position is adjusted step by step in the direction of the gas concentration value of the previous spatial discrete position with a fixed step size of 0.02. When the angle difference is less than or equal to 10 degrees, it remains unchanged.

[0100] The entire convergence and reorganization process is executed in only one round without any looping. After all spatial discrete positions are processed, the diffusion traction direction markers in the continuous diffusion distribution form a single convergence direction in space. At the same time, the gas concentration values ​​remain in a continuously changing state in space, thereby limiting abnormal diffusion offsets.

[0101] Step 5: Based on the abnormal distribution state after directional convergence, perform contour splicing and deduction processing on the overall gas abnormal region to reconstruct the dispersed abnormal region into a single continuous abnormal segment, thereby determining the concentrated area of ​​the fire source and the direction of expansion.

[0102] The scattered anomalous areas are reconstructed into a single continuous anomalous segment, and the concentrated area of ​​fire sources and the direction of expansion are determined accordingly. The specific steps are as follows:

[0103] In the abnormal distribution state after directional convergence, a one-way traversal is performed according to the spatial discrete position number in ascending order. The gas concentration value of each spatial discrete position is normalized. All gas concentration values ​​are linearly mapped to the interval of 0 to 1 according to the maximum and minimum values ​​in the current abnormal distribution range. Spatial discrete positions with normalized gas concentration values ​​greater than or equal to 0.6 are uniquely identified as abnormal points. All spatial discrete positions with values ​​less than 0.6 are not included in the subsequent splicing process.

[0104] Each anomaly is processed point by point in ascending order of its number. The current anomaly is used as the starting splicing point. The subsequent spatial discrete position is searched only in the ascending direction of the number. No reverse search is performed. When the subsequent spatial discrete position meets the three conditions of a spatial spacing of 4 meters, a normalized gas concentration difference of no more than 0.2, and a diffusion traction direction identification angle difference of no more than 10 degrees, the subsequent spatial discrete position is included in the current splicing path and used as a new splicing reference point to continue processing.

[0105] If any condition is not met, the current stitching path construction is immediately terminated, and the subsequent spatial discrete position is used as the new starting stitching point to re-execute the same rules, ensuring that each abnormal point participates in path construction only once and is not processed repeatedly, thereby forming several non-overlapping initial stitching paths.

[0106] After all the initial splicing paths are constructed, a unique connection process is performed between any two adjacent splicing paths. The connection order is fixed to process each pair in ascending order of the starting spatial discrete position numbers in the path. For each pair of splicing paths, only the end spatial discrete position of the previous splicing path and the starting spatial discrete position of the next splicing path are selected as unique connection endpoints, and the spatial distance between the two endpoints is calculated.

[0107] When the spatial distance is equal to 4 meters, the two splicing paths are directly merged into one continuous path. When the spatial distance is greater than 4 meters and less than or equal to 8 meters, an intermediate connecting spatial discrete position is inserted between the two endpoints. The gas concentration value of the intermediate connecting spatial discrete position is uniquely taken as the arithmetic mean of the gas concentration values ​​of the two endpoints, and its diffusion traction direction indicator is uniquely taken as the direction indicator of the spatial discrete position at the end of the previous splicing path. Furthermore, the intermediate connecting spatial discrete position only participates in the path splicing and sorting process in this step, and after this step, it is used as a formal spatial discrete position in subsequent steps.

[0108] When the spatial distance is greater than 8 meters, no connection operation is performed, and the two spliced ​​paths remain independent and do not participate in subsequent connection processing. After all paths are connected, all spliced ​​paths are reordered according to their spatial discrete position numbers, ensuring that the spatial discrete position numbers within each path are strictly increasing, thus obtaining a unique set of candidate paths for continuous abnormal segments.

[0109] After obtaining the set of candidate paths for continuous abnormal sections, contour splicing and deduction processing is performed on each candidate path. The processing order is fixed as follows: the candidate paths are processed one by one in ascending order of their initial spatial discrete position numbers. For each candidate path, the spatial discrete position with the largest gas concentration value is selected as the center point of the fire source concentration area. When there are multiple maximum values ​​with the same gas concentration, only the spatial discrete position with the smallest number is selected as the center point.

[0110] Once the center point is determined, the expansion direction is determined along the increasing and decreasing numbering directions, respectively, based on the center point. In each direction, only three consecutive spatial discrete positions are selected sequentially for determination. If the gas concentration values ​​at these three spatial discrete positions satisfy a strict decreasing relationship and the diffusion traction direction markings are completely consistent, the direction is uniquely determined as the ignition source expansion direction. Otherwise, the determination of the direction is immediately terminated and no further extension is allowed.

[0111] After determining the two directions, all directions that meet the conditions are output as the fire source expansion direction, and the corresponding candidate paths are uniquely determined as a single continuous abnormal segment, thereby completing the outline splicing and deduction of the overall gas abnormal area and realizing the unique determination of the fire source concentration area and expansion direction.

[0112] In the implementation of this invention, the values ​​of spatial spacing, time interval, gas concentration change amplitude range, airflow offset angle range, pressure difference fluctuation scale range, direction determination angle threshold, concentration difference threshold, and proportional adjustment parameters are all preferred parameter ranges selected under specific working conditions to achieve continuous identification of abnormal gas distribution and characterization of diffusion trends. Their values ​​are set according to the ventilation disturbance characteristics of underground coal mines, the time scale of gas diffusion response, and the requirements for continuous spatial distribution, so as to ensure that representative change characteristics can be stably obtained under complex airflow conditions.

[0113] The above values ​​are not the only limitation on the technical solution of the present invention. Without departing from the overall technical concept of the present invention, those skilled in the art can make proportional amplification, reduction or interval shifting adjustments to the relevant parameters according to different mine ventilation intensities, roadway scales, gas types and monitoring accuracy requirements. For example, adjusting the spatial spacing at the same scale, synchronously expanding or compressing the time window, making relative proportional corrections to the concentration threshold, and making consistent offsets to the angle limit range, etc. As long as the adjusted parameters still meet the technical objectives of identifying abnormal gas change trends, constraining diffusion directions and restoring spatial continuity, they should all be considered to fall within the protection scope of the present invention.

[0114] Meanwhile, the parameters are not set in isolation, but rather form a collaborative constraint based on a unified scale relationship. When adjusting the parameters, the relative proportion remains unchanged, so that equivalent technical effects can be obtained in different application scenarios, thus giving the invention good adaptability and scalability.

[0115] like Figure 4 As shown, this application provides a hidden fire zone detection system for underground coal mines, including a multi-level drift scale construction module, an abnormal distribution continuous repair module, a diffusion pattern rearrangement construction module, a directional convergence traction adjustment module, and a fire zone contour deduction and judgment module.

[0116] The multi-level drift scale construction module collects and processes the continuous variation amplitude, airflow offset angle change, and local pressure difference fluctuation scale of underground coal mine gas concentration at spatially discrete locations. It converts the continuous variation amplitude, airflow offset angle change, and local pressure difference fluctuation scale into a multi-level drift scale. Based on the multi-level drift scale, it delineates the sensitive range of abnormal gas jumps and marks the diffusion traction direction at each spatially discrete location.

[0117] The abnormal distribution continuous repair module performs segmented release adjustment on the gas concentration distribution within the sensitive range of abnormal gas changes. It introduces reverse diffusion compensation at spatial discrete positions where the diffusion traction direction marker is offset and exceeds the preset scale, performs continuous repair processing on the abnormal gas distribution, and obtains a set of controlled abnormal associations.

[0118] The diffusion pattern rearrangement construction module performs misaligned stretching rearrangement on each spatial discrete location based on the controlled anomaly association set. The originally isolated anomaly points are gradient extended according to the diffusion traction direction. At the same time, continuous spacing constraints are applied between adjacent spatial discrete locations to construct a continuous diffusion distribution pattern.

[0119] The directional convergence traction adjustment module performs directional convergence traction adjustment along each spatial discrete position based on the continuous diffusion distribution pattern. It sets the angle limit range according to the offset trend of the diffusion traction direction indicator and superimposes the reverse traction amount at the spatial discrete position that deviates from the main diffusion direction to limit abnormal diffusion offset.

[0120] The fire zone contour deduction and judgment module performs contour splicing and deduction processing on the overall gas anomaly area based on the abnormal distribution state after directional convergence, reconstructs the dispersed abnormal area into a single continuous abnormal segment, and determines the concentrated area of ​​fire source and its expansion direction accordingly.

[0121] The embodiment of this application provides a method for detecting concealed fire zones in underground coal mines, which is implemented through the aforementioned system for detecting concealed fire zones in underground coal mines. For details of the specific methods and processes of the system for detecting concealed fire zones in underground coal mines, please refer to the embodiment of the method for detecting concealed fire zones in underground coal mines, which will not be repeated here.

[0122] The foregoing has only described certain exemplary embodiments of this application by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of this application. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of this application.

Claims

1. A method for detecting concealed fire zones in underground coal mines, characterized in that, Includes the following steps: The continuous variation amplitude, airflow offset angle variation, and local pressure difference fluctuation scale of underground gas concentration at spatially discrete locations in underground coal mines are collected and processed. The continuous variation amplitude, airflow offset angle variation, and local pressure difference fluctuation scale are converted into multi-level drift scales. Based on the multi-level drift scales, the sensitive range of abnormal gas jumps is defined. Segmented release regulation is performed on the gas concentration distribution within the sensitive range of abnormal gas changes to continuously repair the abnormal gas distribution and obtain a set of controlled abnormal correlations. Based on the controlled anomaly association set, the spatial discrete locations are subjected to misalignment stretching and rearrangement processing, while continuous spacing constraints are applied between adjacent spatial discrete locations to construct a continuous diffusion distribution pattern; Based on the continuous diffusion distribution pattern, directional convergence traction adjustment is implemented along each spatial discrete position, and reverse traction is superimposed at spatial discrete positions that deviate from the main diffusion direction to limit abnormal diffusion offset. Based on the abnormal distribution state after directional convergence, the overall gas abnormal region is subjected to contour splicing and deduction processing to reconstruct the dispersed abnormal region into a single continuous abnormal segment, thereby determining the concentrated area of ​​the fire source and its expansion direction.

2. The method for detecting concealed fire zones in underground coal mines according to claim 1, characterized in that, The sensitive range for gas anomalies is determined based on a multi-level drift scale. The specific steps are as follows: Based on the direction of roadway extension and the spatial distribution of goaf in the underground working area, the monitoring area is divided into spatially discrete locations. At each spatially discrete location, gas concentration is recorded, airflow direction angle is obtained, and pressure difference is recorded and processed. Based on this, the continuous change amplitude of gas concentration, change of airflow offset angle, and local pressure difference fluctuation scale are obtained. The scale of continuous change in gas concentration, change in airflow offset angle, and local pressure difference fluctuation is divided into intervals. At each spatial discrete position, the gas concentration level, airflow offset angle level, and local pressure difference level are compared item by item. The maximum level is selected as the comprehensive level and defined as a multi-level drift scale. Based on the multi-level drift scale, the difference between adjacent spatial discrete positions is compared. Positions where the difference meets the condition are marked as abnormal jump points and gradually expanded to form a gas abnormal jump sensitive range.

3. The method for detecting concealed fire zones in underground coal mines according to claim 2, characterized in that, The abnormal gas distribution is continuously repaired to obtain a controlled set of anomaly associations. The specific steps are as follows: The gas anomaly jump sensitive interval is traversed in the spatial discrete position numbering order. The continuous spatial discrete positions form the processing unit. The gas concentration of the reference spatial discrete position is used to perform difference comparison on the subsequent spatial discrete positions and perform segmented release adjustment so that the gas concentration converges to the reference value and maintains unidirectional change, thus obtaining a continuous monotonic gas concentration distribution. Based on the gas concentration distribution after segmented release adjustment, the angle difference of the diffusion traction direction markers at each spatial discrete position is determined. At the spatial discrete position where the offset occurs, the gas concentration is extracted from the adjacent spatial discrete position in the opposite direction and superimposed to form the reverse diffusion compensation amount. After introducing the reverse diffusion compensation, the spatial discrete positions within the gas anomaly jump sensitive range are sequentially scanned. The correlation chain is constructed starting from the spatial discrete position and is connected and truncated according to the gas concentration difference and the difference in the diffusion traction direction indicator angle to form a controlled anomaly correlation set.

4. The method for detecting concealed fire zones in underground coal mines according to claim 3, characterized in that, In the step of determining the angle difference of the diffusion traction direction markers at each spatial discrete location and introducing the reverse diffusion compensation amount: for the spatial discrete location where the offset occurs, spatial positioning is performed along the reverse angle direction of the diffusion traction direction markers. Gas concentration is extracted from adjacent spatial discrete locations and superimposed on the current spatial discrete location. After the compensation is completed, gas concentration extraction and superposition processing is performed along the same reverse path until the angle difference of the diffusion traction direction markers meets the preset conditions and the compensation path remains consistent in direction.

5. The method for detecting concealed fire zones in underground coal mines according to claim 3, characterized in that, To construct a continuous diffusion distribution pattern, continuous spacing constraints are applied between adjacent spatial discrete locations. The specific steps are as follows: Based on the controlled anomaly association set, the starting extension point of the anomaly association chain is selected according to the spatial discrete position numbering order. The gas concentration value and diffusion traction direction mark are extracted as extension references. The spatial discrete position is scanned in one direction along the direction corresponding to the diffusion traction direction mark. The spatial discrete position that meets the direction difference condition is selected as the extension object. The extension reference is updated in one direction according to the gas concentration value difference to form a continuous gradient extension path. The spatial discrete positions are sequentially rearranged by staggering and stretching according to their numbering order. By comparing the actual spatial distance between adjacent spatial discrete positions with the minimum continuous spacing, the spatial discrete positions are translated or inserted into intermediate spatial discrete positions. Based on the results of the misaligned stretching and rearrangement process, a unidirectional scan is performed on the spatial discrete positions. A unified correction process is then applied to the spatial distance between adjacent spatial discrete positions, and the gas concentration values ​​are adjusted simultaneously to ensure that the gas concentration remains monotonically changing along the spatial arrangement direction, thereby constructing a continuous diffusion distribution pattern.

6. The method for detecting concealed fire zones in underground coal mines according to claim 5, characterized in that, The process of unidirectionally scanning spatial discrete positions along the direction of diffusion traction and selecting spatial discrete positions whose directional differences meet the conditions as extension objects also includes: when there are multiple spatial discrete positions that meet the conditions, only the spatial discrete position with the smallest number is selected as the only extension object, and the value of the gas concentration is kept to change in one direction without reverse correction during the adjustment process, until the difference between the extension object and the extension reference meets the preset relationship.

7. A method for detecting concealed fire zones in underground coal mines according to claim 5, characterized in that, Based on the continuous diffusion distribution pattern, directional convergence and traction adjustment are implemented along each spatial discrete location. The specific steps are as follows: The continuous diffusion distribution pattern is traversed in a unidirectional manner according to the spatial discrete position number from small to large. The diffusion traction direction of the spatial discrete position with the smallest number is determined as the main diffusion direction. Based on the main diffusion direction, the angle difference of the diffusion traction direction of the subsequent spatial discrete positions is calculated to complete the initial direction classification. For the spatial discrete positions after the initial direction classification is completed, the diffusion traction direction markers that deviate from the main diffusion direction are adjusted to the boundary of the angle limiting interval along the minimum rotation path to obtain the diffusion traction direction markers after unified constraints. Based on the unified constraint of the diffusion traction direction identifier, the reverse traction amount superposition process is performed on each spatial discrete position, and unidirectional convergence is carried out along the spatial discrete position numbering order, so that the diffusion traction direction identifier forms a single convergence direction, thus completing the restriction on abnormal diffusion offset.

8. A method for detecting concealed fire zones in underground coal mines according to claim 7, characterized in that, During the process of adjusting the diffusion traction direction marker that deviates from the main diffusion direction to the boundary of the angle limiting range along the minimum rotation path, only the diffusion traction direction marker is subjected to angle correction processing. After the adjustment is completed, the corrected diffusion traction direction marker is used as the unique direction expression of the spatial discrete position, so that the diffusion traction direction markers of all spatial discrete positions are restricted to a uniform angle range.

9. A method for detecting concealed fire zones in underground coal mines according to claim 7, characterized in that, The scattered anomalous areas are reconstructed into a single continuous anomalous segment, and the concentrated area of ​​fire sources and the direction of expansion are determined accordingly. The specific steps are as follows: The spatial discrete positions in the abnormal distribution state after directional convergence are traversed in numerical order, the gas concentration values ​​are normalized and mapped, and the spatial discrete positions where the normalized gas concentration values ​​meet the preset conditions are determined as anomaly points. The anomaly points are used as the starting splicing points to construct the initial splicing path along the direction of increasing number. Based on the initial splicing path obtained, the spatial discrete positions at the end and the spatial discrete positions at the beginning of adjacent paths are selected as connection endpoints. Path connection processing is performed according to the spatial distance between the two endpoints, and the spatial discrete positions after connection are rearranged in numerical order to obtain a set of candidate paths for continuous abnormal segments. For the candidate path set of continuous abnormal sections, the spatially discrete location with the largest gas concentration value in the path is selected as the center point of the concentrated fire source area, and the expansion direction is determined along the numbering direction based on the center point.

10. A concealed fire zone detection system for underground coal mines, used to implement the concealed fire zone detection method for underground coal mines as described in any one of claims 1-9, characterized in that, It includes a multi-level drift scale construction module, an abnormal distribution continuous repair module, a diffusion pattern rearrangement construction module, a direction convergence traction adjustment module, and a fire zone contour deduction and judgment module; The multi-level drift scale construction module collects and processes the continuous variation amplitude, airflow offset angle change, and local pressure difference fluctuation scale of underground coal mine gas concentration at spatially discrete locations. It converts the continuous variation amplitude, airflow offset angle change, and local pressure difference fluctuation scale into a multi-level drift scale. Based on the multi-level drift scale, it delineates the sensitive range of abnormal gas jumps and marks the diffusion traction direction at each spatially discrete location. The abnormal distribution continuous repair module performs segmented release adjustment on the gas concentration distribution within the sensitive range of abnormal gas changes. It introduces reverse diffusion compensation at spatial discrete positions where the diffusion traction direction marker is offset and exceeds the preset scale, performs continuous repair processing on the abnormal gas distribution, and obtains a set of controlled abnormal associations. The diffusion pattern rearrangement construction module performs misaligned stretching rearrangement on each spatial discrete location based on the controlled anomaly association set. The originally isolated anomaly points are gradient extended according to the diffusion traction direction. At the same time, continuous spacing constraints are applied between adjacent spatial discrete locations to construct a continuous diffusion distribution pattern. The directional convergence traction adjustment module performs directional convergence traction adjustment along each spatial discrete position based on the continuous diffusion distribution pattern. It sets the angle limit range according to the offset trend of the diffusion traction direction indicator and superimposes the reverse traction amount at the spatial discrete position that deviates from the main diffusion direction to limit abnormal diffusion offset. The fire zone contour deduction and judgment module performs contour splicing and deduction processing on the overall gas anomaly area based on the abnormal distribution state after directional convergence, reconstructs the dispersed abnormal area into a single continuous abnormal segment, and determines the concentrated area of ​​fire source and its expansion direction accordingly.

Images