Mine spontaneous combustion risk intelligent identification method and system based on gas concentration field

Abstract

The present application provides a mine spontaneous combustion risk intelligent identification method and system based on gas concentration field, and relates to the technical field of mine safety monitoring and mining. The present application obtains the spatial position of the goaf gas monitoring point, the time sequence of the gas concentration, the working face advancing position, the ventilation operation and the oxidation response data, constructs the first gas concentration field and the second gas concentration field; determines the first gas dilution area and the first boundary according to the gas concentration reduction and the spatial gradient, generates the first path and determines the first end area; combines the goaf age of each spatial calculation unit to form the second area; determines the first, second or third spontaneous combustion risk level according to the concentration field anomaly persistence after the ventilation adjustment and the gas dilution and oxidation response time lag relationship along the first path, and updates the spatial position, moving direction and moving distance of the second area at fixed time intervals to form the dynamic identification result of the mine spontaneous combustion risk changing with the working face advancing.

Description

Technical Field

[0001] This invention relates to the field of mine safety monitoring and mining technology, specifically to a method and system for intelligent identification of mine spontaneous combustion risk based on gas concentration field. Background Technology

[0002] During underground coal mining, residual coal in the goaf is prone to slow oxidation and heat accumulation under continuous oxygen supply. When this heat cannot be dissipated in time, it may further increase the risk of spontaneous combustion. Due to the enclosed space and complex collapse structure within the goaf, air typically migrates into the goaf through the working face ends, roadway fissures, and pores in the collapsed rock mass, resulting in significant spatial differences in the oxygen supply conditions of the residual coal. Therefore, accurately determining the spatial location where continuous oxidation and heat accumulation may occur, based on the gas migration state within the goaf, is a crucial issue in identifying spontaneous combustion risks in mines.

[0003] Existing mine spontaneous combustion monitoring typically uses parameters such as carbon monoxide concentration, oxygen concentration, temperature, and hydrocarbon indicator gases. Thresholds are set or trends are analyzed to determine if coal seam oxidation anomalies have occurred. Some methods further combine gas concentration, ventilation parameters, and historical monitoring data to establish risk assessment models. While these methods can reflect gas or temperature anomalies at local monitoring locations, gas concentrations in the goaf are easily affected by factors such as face advancement, changes in ventilation volume, gas extraction, and localized air leakage. Changes in concentration at a single monitoring point cannot directly characterize the spatial source of the anomaly. Especially when ventilation conditions are adjusted, a decrease in gas concentration may be caused by normal airflow dilution or may reflect new air leakage extending deeper into the goaf. If risk assessment is based solely on gas concentration values ​​or multi-parameter weighted results, it is easy to confuse normal ventilation responses with abnormal changes caused by continuous air leakage and oxygen replenishment.

[0004] Furthermore, existing methods do not adequately utilize the spatial continuity and temporal sequence of monitoring data from different locations. They typically struggle to deduce the migration direction of air leakage and oxygen supply based on the spatial distribution of gas concentration within the goaf, and also find it difficult to further combine this with the differences in goaf age resulting from working face advancement to determine the region where the end of air leakage and oxygen supply coincides with the oxidation time conditions of residual coal. Therefore, it is necessary to convert gas concentration from discrete point values ​​to a spatial concentration field, and further utilize the spatial variation of the concentration field to identify the air leakage and oxygen supply path and its terminal location. Based on this, and combining the goaf age, the persistence of concentration field anomalies, and the time lag relationship of oxidation response, the risk of spontaneous combustion in the mine can be dynamically identified. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for intelligent identification of mine spontaneous combustion risk based on gas concentration field, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: A method and system for intelligent identification of spontaneous combustion risk in mines based on gas concentration fields, the specific steps of which include: S1. Acquire spatial location data of gas monitoring points, time series data of gas concentration, time series data of working face advancement position, ventilation operation data, and oxidation response data of the target mine goaf, and divide the target mine goaf into several spatial measurement units; construct a first gas concentration field based on the time series data of gas concentration of each gas monitoring point within the current time window, and construct a second gas concentration field based on the historical stable state gas concentration data corresponding to the current ventilation operation status; S2. Based on the first gas concentration field and the second gas concentration field, obtain the gas concentration change results and gas concentration spatial gradient of each spatial measurement unit, determine the first gas dilution region according to the gas concentration change results and gas concentration spatial gradient, and obtain the first boundary according to the first gas dilution region; generate the first path according to the spatial connection relationship of each spatial measurement unit in the first gas dilution region and the direction of the gas concentration spatial gradient, and determine the first end region according to the first path; S3. Based on the time sequence data of the working face advancement position, determine the time when each spatial calculation unit enters the goaf and obtain the goaf age of each spatial calculation unit; based on the first path, the first end area and the goaf age of each spatial calculation unit, filter the first end area and its adjacent spatial calculation units to generate the second area. S4. Based on ventilation operation data and the first gas concentration field corresponding to the continuous time window, obtain the abnormal persistence judgment result of the first gas concentration field; based on the gas concentration change and oxidation response data on the first path, obtain the time delay response judgment result along the first path; according to the second region, the abnormal persistence judgment result and the time delay response judgment result, determine the first spontaneous combustion risk level, the second spontaneous combustion risk level or the third spontaneous combustion risk level corresponding to the goaf of the target mine.

[0007] A mine spontaneous combustion risk intelligent identification system based on gas concentration field, used to implement a mine spontaneous combustion risk intelligent identification method based on gas concentration field, including: The data acquisition module is used to acquire spatial location data of gas monitoring points, time series data of gas concentration, time series data of working face advancement position, ventilation operation data and oxidation response data of the target mine goaf, divide the target mine goaf into several spatial measurement units, and construct the first gas concentration field and the second gas concentration field. The first identification module is used to obtain the gas concentration change results and gas concentration spatial gradient of each spatial measurement unit based on the first gas concentration field and the second gas concentration field, determine the first gas dilution region and the first boundary, generate the first path according to the spatial connection relationship of each spatial measurement unit and the direction of the gas concentration spatial gradient, and determine the first end region according to the first path. The second identification module is used to determine the mined-out age of each spatial measurement unit based on the time sequence data of the working face advancement position, and to generate a second region based on the first path, the first end region and the mined-out age of each spatial measurement unit. The risk assessment module is used to obtain the abnormal persistence assessment result based on ventilation operation data and the first gas concentration field corresponding to the continuous time window, obtain the time delay response assessment result based on the gas concentration change and oxidation response data on the first path, and determine the first spontaneous combustion risk level, the second spontaneous combustion risk level, or the third spontaneous combustion risk level according to the second region, the abnormal persistence assessment result, and the time delay response assessment result.

[0008] Compared with the prior art, the beneficial effects of the present invention are: This invention maps the current gas concentration at discrete gas monitoring points to historical stable gas concentrations under the same ventilation operating conditions, respectively, into a unified spatial measurement unit to construct a first gas concentration field and a second gas concentration field. It then calculates the gas concentration reduction and spatial gradient for each spatial measurement unit, determining a spatially continuous first gas dilution region and its first boundary. This improves the utilization of the spatial distribution characteristics of gas changes in goaf areas. Compared to judging solely based on the concentration level of a single monitoring point, this invention can identify the spatial morphology of gas dilution extending deeper into the goaf by combining the direction of concentration change and gradient continuity between adjacent spatial units. This transforms discrete monitoring data into spatial results that reflect the dilution intrusion range, providing continuous spatial evidence for subsequent judgment of air leakage and oxygen replenishment paths, and reducing the impact of local sensor fluctuations and instantaneous concentration changes on risk identification results.

[0009] This invention improves the targeting of identifying the spatial migration direction and end location of air leakage and oxygen replenishment by generating candidate connection paths according to spatial connectivity and the spatial gradient direction of gas concentration within the first gas dilution region, and determining the first path by combining the continuous connection length and the consistency of the gradient direction. The first terminal region is then determined by the intersection of the first path and the first boundary. Based on this, the time when each spatial calculation unit enters the goaf is determined using the working face advancement position, and the goaf age is calculated. Then, second spatial calculation units within the goaf age range corresponding to continuous changes in historical oxidation response are selected from the first terminal region and its adjacent areas to form a second region. This allows risk assessment to simultaneously consider the location of air leakage and oxygen replenishment and the exposure time of residual coal, thereby distinguishing between "existence of air leakage" and "having oxidation and heat storage time conditions," reducing the likelihood of directly identifying general air leakage ends as high-risk spontaneous combustion areas.

[0010] This invention tracks the continuous first gas concentration field after ventilation operation parameters are adjusted, compares the first duration required for the first boundary to stabilize with the second duration corresponding to similar normal historical adjustments, and combines the remaining spatial offset after stabilization to determine whether there is an abnormal continuous offset in the concentration field. Simultaneously, it compares the timing of the gas dilution response with the oxidation response such as carbon monoxide or temperature along the first path to obtain the first response time lag and its change relationship along the path, improving the ability to distinguish between normal ventilation disturbances and continuous air leakage and oxygen replenishment anomalies. Based on this, it further combines the second region, the abnormal persistence judgment results, and the time lag response judgment results to determine the first, second, or third spontaneous combustion risk level. Furthermore, as the working face advances and the gas concentration field is updated, it synchronously calculates the movement direction and distance of the second region, enabling the risk identification results to correspond to the continuous changes in the spatial state of the goaf. Attached Figure Description

[0011] Figure 1 This is a schematic diagram of the overall method flow of the present invention; Figure 2 This is a schematic diagram of the spatial distribution of the gas monitoring point in the goaf, the first gas dilution zone, the first boundary, the first path, the first end zone, and the second zone of the present invention. Figure 3 This is the process for constructing the first and second gas concentration fields according to the present invention; Figure 4 This is a flowchart of the first boundary, first path, and first end region of the first gas dilution zone according to the present invention. Figure 5 This is a flowchart for generating the second region of the present invention; Figure 6 This is a flowchart of the abnormal persistence judgment process of the present invention; Figure 7 This is a flowchart of the time-delay response judgment process of the present invention; Figure 8This is a flowchart of the spontaneous combustion risk level and dynamic identification results of the present invention. Detailed Implementation

[0012] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0013] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0014] Please see Figures 1 to 8 The present invention provides a technical solution: A method for intelligent identification of spontaneous combustion risk in mines based on gas concentration fields, characterized by the following specific steps: S1. Acquire spatial location data of gas monitoring points, time series data of gas concentration, time series data of working face advancement position, ventilation operation data, and oxidation response data of the target mine goaf, and divide the target mine goaf into several spatial measurement units; construct a first gas concentration field based on the time series data of gas concentration of each gas monitoring point within the current time window, and construct a second gas concentration field based on the historical stable state gas concentration data corresponding to the current ventilation operation status; S2. Based on the first gas concentration field and the second gas concentration field, obtain the gas concentration change results and gas concentration spatial gradient of each spatial measurement unit, determine the first gas dilution region according to the gas concentration change results and gas concentration spatial gradient, and obtain the first boundary according to the first gas dilution region; generate the first path according to the spatial connection relationship of each spatial measurement unit in the first gas dilution region and the direction of the gas concentration spatial gradient, and determine the first end region according to the first path; S3. Based on the time sequence data of the working face advancement position, determine the time when each spatial calculation unit enters the goaf and obtain the goaf age of each spatial calculation unit; based on the first path, the first end area and the goaf age of each spatial calculation unit, filter the first end area and its adjacent spatial calculation units to generate the second area. S4. Based on ventilation operation data and the first gas concentration field corresponding to the continuous time window, obtain the abnormal persistence judgment result of the first gas concentration field; based on the gas concentration change and oxidation response data on the first path, obtain the time delay response judgment result along the first path; according to the second region, the abnormal persistence judgment result and the time delay response judgment result, determine the first spontaneous combustion risk level, the second spontaneous combustion risk level or the third spontaneous combustion risk level corresponding to the goaf of the target mine.

[0015] The following describes the specific implementation of the identification method of this application using an exemplary fully mechanized longwall face in a high-gas coal mine. Examples one through eight all use the same longwall face, the same mine spatial coordinate system, the same spatial measurement unit numbering system, and the same ventilation adjustment event as continuous samples. The example date is uniformly set to the current day: the longwall face ventilation adjustment occurs at 10:00 on the current day, and the ventilation operating parameters stabilize at 10:12; Examples one through four use the 10:12-10:17 time window as the initial identification time window, and Examples five through eight continue to track this event backward to 10:50. The deep direction of the goaf is defined as the direction in which the X-coordinate of the unified coordinate system increases. In Table 1, P01-P08 are located near spatial measurement units U01-U08 and have a fixed mapping relationship. Tables 2 and 3 continue to use the same spatial grid numbering, only adding candidate branch units U09-U16 obtained by spatial interpolation. Table 4 only continues to search for U18-U24 around the end of the first path U07 according to the same grid. Tables 5-8 continuously update the time window data based on the above spatial results. Therefore, Tables 1-8 are not eight independent sets of samples, but a continuous data chain in the same mine scenario: "monitoring point → concentration field → dilution area → path → end → goaf age → persistence → time lag → risk level → dynamic migration". The dimensions, monitoring values, and thresholds below are example values ​​used to illustrate the data transfer relationship between steps and do not constitute a limitation on the protection scope.

[0016] Figure 1 Using the working face, intake airway, return airway, and goaf behind the working face of the target mine as the main spatial components, the goaf is divided into several spatial measurement units. Gas monitoring points and oxidation response monitoring points are set at different spatial locations. The spatial distribution relationship between the first and second gas concentration fields is represented by a spatial grid and concentration distribution lines. The area continuously extending from one side of the working face into the depth of the goaf represents the first gas dilution area determined based on gas concentration changes and the spatial gradient of gas concentration. Its boundary along the direction of the depth of the goaf corresponds to the first boundary. The continuous connection structure formed along the spatial connection relationship of each spatial measurement unit and the direction of the spatial gradient of gas concentration corresponds to the first path. The intersection of the first path and the first boundary, and its adjacent spatial measurement units on the deep side of the goaf, correspond to the first end region. The goaf age of each spatial measurement unit is determined by combining the time-series data of the working face advance position, and a second region is formed based on the first path, the first end region, and the goaf age. Figure 1The four flowcharts on the right correspond to steps S1 to S4 in sequence: The first flowchart indicates the acquisition of spatial location data of gas monitoring points, time-series data of gas concentration, time-series data of working face advancement position, ventilation operation data, and oxidation response data, and the construction of a first gas concentration field and a second gas concentration field; the second flowchart indicates the acquisition of gas concentration change results and gas concentration spatial gradient based on the first and second gas concentration fields, the determination of the first gas dilution area and the first boundary, and the generation of the first path and the determination of the first end area; the third flowchart indicates the acquisition of spatial location data of each gas monitoring point based on the time-series data of working face advancement position. The calculation unit determines the age of the goaf and generates a second region by combining the first path and the first end region; the fourth flowchart indicates that the abnormal persistence judgment result is obtained based on the ventilation operation data and the first gas concentration field corresponding to the continuous time window, and the time delay response judgment result is obtained based on the gas concentration change and oxidation response data on the first path. Based on this, the first spontaneous combustion risk level, the second spontaneous combustion risk level, or the third spontaneous combustion risk level is determined. As the working face advances and the gas concentration time series data is updated, the spatial location, movement direction, and movement distance of the second region are updated to form the dynamic identification result of spontaneous combustion risk in the goaf of the mine.

[0017] Example 1 In this embodiment, steps S1 and S11-S14 are used to convert discrete gas concentration point values ​​into a first gas concentration field and a second gas concentration field that can be continuously compared under a unified spatial coordinate system, providing a basis for subsequent identification of the gas dilution intrusion front.

[0018] In this application scenario, the range of 0-120m behind the working face is selected as the current goaf measurement range, and the spatial measurement units are divided into 5m, 5m, and 2m units along the strike, dip, and vertical directions, respectively. The original sampling period of the gas monitoring point is 10s, and CO and temperature data are correlated using the same timestamp; the fixed-length sliding time window is 5min, and the step size of adjacent time windows is 1min. The initial identification time window in this embodiment is 10:12-10:17 of the current day. To allow direct reference in subsequent embodiments, P01-P08 are mapped sequentially to U01-U08, and this numbering relationship will not be changed subsequently.

[0019] S11. Transform the three-dimensional spatial coordinates of each gas monitoring point to the mine spatial coordinate system. Use the first spatial corner point of the measurement area as the reference. Starting from the index, for the first... For each gas monitoring point, the index of its spatial measurement unit in the direction of orientation is: The directional and vertical indices are determined in the same way. Indicates the first The coordinates of the direction of each gas monitoring point The length of the spatial measurement unit indicating the direction of travel. In this embodiment... By using three directional indices, a unique association is established between the monitoring point and the specific spatial measurement unit.

[0020] S12. Divide the gas concentration time series data into time windows. Let the current time window contain the segmented data. There are a total of [number] monitoring points. If there are 1 valid sample value, the first gas concentration is recorded as . The calculation is as follows: Screening is performed on historical stable time windows that are the same as the current ventilation operation status, assuming that the obtained If there are 10 historical stable sampling values, then the second gas concentration is recorded as 10. The calculation is as follows: The selection criteria for the historical stable time window are: consistent main ventilation fan operation combination, working face air distribution within the allowable fluctuation range of the current air distribution, consistent gas extraction branch status, and no abnormal coal spontaneous combustion recorded before or after the historical time window. The resulting second gas concentration is not a fixed constant, but rather a historical stable benchmark corresponding to the current ventilation operation status.

[0021] S13. Using the inverse distance weighted spatial interpolation method, the first gas concentration of the current time window is mapped to each spatial measurement unit. For the first... Each spatial measurement unit's center point is selected, and its predetermined neighborhood is used for selection. At each gas monitoring point, the first concentration value is recorded as... The calculation is as follows: in, Indicates the first From the first gas monitoring point to the first The Euclidean distance between the center points of each spatial measurement unit This represents the distance weighting index, which is taken in this embodiment. After obtaining the first concentration values ​​of all spatial measurement units in sequence, the first gas concentration field is constructed based on the spatial index.

[0022] S14. Using the same spatial interpolation neighborhood, weighting index, and spatial grid as S13, interpolate the second gas concentration to obtain the second concentration value for each spatial measurement unit. A second gas concentration field was constructed. Using the same grid and interpolation rules can avoid subsequent differences arising from variations in spatial discretization. In this embodiment, CO concentration and temperature were simultaneously acquired as oxidation response data. Table 1 shows the monitoring samples for the initial identification time window of 10:12-10:17, where P01-P08 correspond one-to-one with U01-U08.

[0023] Table 1: Monitoring Samples for Current Time Window and Historical Stable Time Window As shown in Table 1, the first gas concentration at positions P01-P07 is significantly lower than the second gas concentration under the same ventilation conditions, and the first gas concentration gradually increases from P01 to P07; the decrease at P08 is significantly smaller. Since P01-P08 correspond to U01-U08 respectively, the first and second gas concentrations in Table 1 will be directly used as the basis data for the first and second concentration values ​​of U01-U08 in Table 2, thereby ensuring the continuity of the concentration reduction and gradient calculation in the next embodiment with this embodiment.

[0024] Example 2 This embodiment corresponds to steps S21-S24, which are used to identify the first gas dilution region and the first boundary. Based on the two sets of gas concentration fields obtained in Embodiment 1, the risk of spontaneous combustion is not directly judged by the gas concentration level of a single monitoring point, but rather the spatial dilution structure of the current gas concentration field relative to the historical stable state is first identified.

[0025] S21, Regarding the first Each spatial measurement unit calculates the reduction in gas concentration. : when When the concentration is below the historical stable level, it indicates that the current concentration is lower than the historical stable level. To avoid misinterpreting normal fluctuations in the sensor as dilution, this embodiment statistically analyzes the concentration difference from historical stable samples that have never experienced ventilation adjustments or spontaneous combustion anomalies, obtaining the mean value. and standard deviation First threshold Determined according to the following formula: Example historical data calculation , Therefore, the first threshold is taken as .

[0026] S22. For each spatial measurement unit, calculate the spatial gradient of gas concentration using the first concentration value of its six neighboring spatial measurement units. Taking a regular spatial grid as an example, the... The gradient vector of each spatial measurement unit is represented as: The gradient components in the three directions are calculated according to the following formulas: in, and They represent the first The first concentration value of each spatial measurement unit along the positive and negative directions of its adjacent units. , , and The meanings are the same. The gradient direction points to the direction of the fastest increase in gas concentration, so in this application scenario, it can be used to describe the spatial transition direction from a low-concentration dilution zone to a higher-concentration zone deep within the goaf.

[0027] Adjacent spatial measurement unit and The gradient direction angle between them The calculation is as follows: This embodiment is based on the statistical analysis of the angle between adjacent gradient directions using historically confirmed normal air leakage dilution samples, with its 95th percentile value being... Therefore, This serves as a threshold for determining the continuity of the gradient direction.

[0028] S23, will The spatial measurement unit was determined as the first spatial measurement unit; subsequently, a three-dimensional adjacency search method was used to select adjacent faces or edges whose gradient directions were not more than [value missing]. The first spatial measurement units are connected. If multiple first spatial measurement units form a continuous connecting body extending from the working face side into the depth of the goaf, then this connecting body is defined as the first gas dilution zone.

[0029] S24. In the first gas dilution zone, the side directly connected to the working face or the shallow boundary of the goaf is designated as the first side. Spatial measurement units within the first gas dilution zone are searched layer by layer along the deeper part of the goaf. A set of outer spatial measurement units that are farthest from the first side and still belong to the first gas dilution zone is selected, and their common outer boundary constitutes the first boundary. The first boundary thus represents the deepest spatial location that the current gas dilution intrusion can stably reach.

[0030] This embodiment calculates on the same spatial grid as in Embodiment 1. For U01-U08, the first and second gas concentrations of the corresponding monitoring points in Table 1 are directly inherited; for branch units U09-U16 where no monitoring point is located exactly in the center, the first and second concentration values ​​are obtained according to the same inverse distance weighted interpolation rule in S13-S14 of Embodiment 1. The calculation results of some spatial measurement units are shown in Table 2.

[0031] Table 2: Samples identified in the first gas dilution zone According to Table 2, the gas concentration reduction of U01-U07 and U09-U16 all reach the first threshold of 0.20%, and the gradient angle along their respective connecting directions does not exceed 25°, thus forming a first gas dilution zone extending from the first side into the depth of the goaf in space. The reduction of U08 is only 0.17%, and it does not belong to the first spatial calculation unit. The first gas dilution zone forms multiple outer edge units in the depth, of which U07 and U16 are both located on the first boundary. In the subsequent embodiment three, no new spatial samples are regenerated, but candidate paths are sorted among these first spatial calculation units in Table 2.

[0032] Example 3 In this embodiment, steps S25-S28 are used to further convert the first gas dilution area into a first path that can describe the migration direction of the leaking oxygen supply space, and the first end area is determined by the intersection of the first path and the first boundary.

[0033] S25. Select the unit that connects to the air leakage inlet space on the working face side from the first spatial measurement units on the first side as the first starting unit. In this embodiment, U01, which is closest to the edge of the goaf on the return air side and has the largest reduction in gas concentration, is selected as the first starting unit. If there are multiple units that meet the conditions, they are used as the starting points of different candidate connection paths.

[0034] S26. Start expanding element by element from the first initial element. For the current element... In this embodiment, a unit is retained among its spatially adjacent units that simultaneously meets the following two conditions: first, the first concentration value along the deep direction of the goaf is higher than that of the current unit; second, the angle between the gradient direction of the adjacent unit and the gradient direction of the current unit does not exceed a second threshold. In this embodiment, the second threshold is taken from the value obtained in Embodiment 2. Each expansion records the cell center coordinates and gradient direction, thereby forming multiple candidate connection paths.

[0035] S27. Calculate the continuous connection length and gradient direction consistency of the candidate connection paths respectively. For the first... The number of candidate connection paths and their consecutive connection lengths for: in, Indicates the first The number of spatial measurement units contained in each candidate connection path. Indicates the first The three-dimensional coordinates of the center of each spatial measurement unit.

[0036] Gradient direction consistency Determined by the cosine mean of the angle between adjacent gradient directions: To facilitate sorting, the path length is normalized and then combined with the degree of consistency with the gradient direction to obtain the first sorting value. : in, This represents the maximum contiguous connection length among all current candidate connection paths. The first condition is set as follows: the candidate connection path can reach the first boundary, and the first ranking value is the largest among the candidate connection paths that can reach the first boundary. This avoids simply selecting a short path that is locally very straight but does not extend to the dilution front.

[0037] S28. Obtain the spatial measurement unit where the first path intersects with the first boundary, and using this intersecting unit as the center, select spatial measurement units that are adjacent in appearance and edge, with a center distance not exceeding the diagonal length of one spatial measurement unit on the deep side of the goaf area of ​​the first boundary. These units, together with the intersecting unit, constitute the first terminal region. This region does not indicate that spontaneous combustion of coal has occurred, but rather represents the terminal spatial range that the current air leakage dilution effect can reach along the first path to the depths.

[0038] Within the same first gas dilution region obtained in Table 2, seven candidate connection paths are formed starting from U01, and the calculation results are shown in Table 3.

[0039] Table 3: Ranking Samples of Candidate Connection Paths As shown in Table 3, R1, R3, R5, and R6 can all reach the first boundary determined in Example 2. Among them, R1 has the highest first ranking value of 0.984, so R1 is determined as the first path. R1 completely follows U01-U07 in Table 2 and intersects with the first boundary at U07. Subsequently, only according to S28, the search for adjacent units continues on the same spatial grid on the deep side of U07, and U07 and U18, U19, and U20 adjacent to U07 are determined as the first end region. U18-U20 are not new independent sample sets, but adjacent spatial measurement units obtained by extending from the end of the first path U07 in Table 2 to the deep side. The next embodiment directly uses U07 and U18-U20 as the screening starting point.

[0040] Example 4 This embodiment corresponds to steps S31-S34, which are used to add the exposure time condition of residual coal in the goaf area based on the first end area, so as to avoid directly judging all air leakage ends as spontaneous combustion hazard areas.

[0041] S31. The working face advancement position system records the working face position every 10 minutes. For any spatial measurement unit j, the time when the centerline of the working face first crosses the projected position of the unit's center along the working face advancement direction is denoted as the first time tj. This embodiment uses the initial identification time window of embodiments one to three, and fixes the current identification time to 10:17 of the current day, denoted as t. now Then the spatial measurement unit sampling age A j Calculated by day: If the working face crosses the cell center within one time recording interval, linear interpolation is used to obtain the result from two adjacent advance positions. .

[0042] S32. Retrieve confirmed historical mine spontaneous combustion monitoring samples under the same coal seam or similar coal quality conditions, and record the goaf age at the corresponding spatial location when the CO concentration or temperature begins to rise continuously in each sample, forming a historical goaf age set. To reduce the impact of individual abnormal historical samples on the extent, the lower limit of the first mining-out age is set. and the upper limit of the first mining age Using 10% and 90% quantiles respectively: This embodiment is based on 32 sets of historical samples. , .

[0043] S33. Based on the first end area, extend the space to a 15m radius around the end of the first path to form a set to be screened. Calculate the mining age of each spatial measurement unit within the set, and determine the units with a mining age between 15.0 and 30.0 days as the second spatial measurement units.

[0044] S34. Perform a spatial connectivity search on the second spatial measurement units. Connect adjacent or edge-adjacent second spatial measurement units into the same connected region; if multiple independent connected regions exist, retain the connected region that directly intersects with the first terminal region or is closest to the end of the first path, and generate the second region. The second region simultaneously satisfies the two conditions of "near the end of the air leakage and oxygen replenishment" and "within the range of mining age where continuous historical oxidation changes are likely to occur," thus corresponding to a potential oxidation heat storage region.

[0045] This embodiment uses the first terminal region U07 and U18-U20 obtained in Embodiment 3 as the core, and continues to search for U21-U24 in the same grid within a 15m range from the end of the first path. All mined-out ages are calculated with 10:17 of the current day as the unified current time, and the screening results are shown in Table 4.

[0046] Table 4: Samples selected based on mined-out age near the first terminal region In Table 4, U07 and U18-U21 are all located within a preset 15m space range at the end of the first terminal area or the end of the first path, and their mined-out age is between 15.0 and 30.0 days, and they are spatially continuous, thus forming a second area. U22 has a mined-out age of 14.1 days, which has not yet reached the lower limit of the first mined-out age; U23 has a mined-out age of 32.8 days, exceeding the upper limit of the first mined-out age; and although U24 has a mined-out age of 21.6 days, it is 16.8m from the end of the first path, exceeding the preset 15m space range. Therefore, the second area ultimately consists of U07, U18, U19, U20, and U21, and serves as the same risk space object continuously tracked in Examples 5 to 8.

[0047] Example 5 This embodiment corresponds to steps S41-S45, which are used to distinguish between "normal concentration field response after ventilation parameter adjustment" and "the ventilation state has stabilized but the gas dilution front is still abnormally and continuously shifting".

[0048] S41. Continuing with the same event described above, the ventilation system records that the air volume at the working face was adjusted from 1650 m³ / min to 1780 m³ / min at 10:00 AM on the same day, and 10:00 AM is designated as the second time point. Subsequently, the coefficient of variation of the air volume is calculated in real time for 3 consecutive minutes. When the coefficient of variation is not greater than 2% and the fan frequency and damper status remain unchanged, the ventilation operating parameters are considered to have reached stability. In this embodiment, the above conditions are met at 10:12 AM, and 10:12 AM is designated as the third time point. This third time point is also the starting point of the initial identification time window in Embodiment 1.

[0049] S42. Starting from the third time point, continue constructing the first gas concentration field according to the 5-minute sliding time window of Example 1, and determine the first boundary in each time window according to the method of Example 2. To characterize the overall spatial position of the first boundary, calculate the components that make up the first boundary. The average value of the center coordinates of the first outer edge spatial measurement unit is used as the first... The center position of the first boundary of the time window : S43, Spatial displacement between the first boundaries of two adjacent time windows The calculation is as follows: This embodiment uses historical stable samples without any anomalies to determine the boundary stability judgment value. When two consecutive time windows All are not greater than When the end time of the second time window is determined as the fourth time window. First duration for: S44. Select ventilation adjustment events from the historical database that have an increase in air distribution volume of 5%-10%, have the same extraction system status, and have not experienced spontaneous combustion abnormalities. Use the same method as in S41-S43 to obtain the historical duration of each event. To mitigate the impact of individual events, the second duration... Using the median of historical duration: This embodiment obtained 12 similar historical adjustment events. .

[0050] S45. Obtain the center position of the first boundary corresponding to the fourth time. And obtain the historical stable boundary center position of the second gas concentration field under the same first threshold condition. Calculate the remaining space offset : The third threshold is determined based on the mean and standard deviation of the remaining offset after the historical normal adjustment events have stabilized; in this embodiment, it is set to 3.0m. When and If the condition is met, a first persistence judgment result is generated; otherwise, a second persistence judgment result is generated.

[0051] The first gas concentration field and the first boundary are continuously updated starting from the third time 10:12, and the calculation results are shown in Table 5; where row 10:17 corresponds to the end time of the initial identification time window used in Examples 1 to 4.

[0052] Table 5: Samples of continuous response at the first boundary after ventilation adjustment From Table 5, we can see that t3=10:12 and t4=10:37, the first duration T1=25min, which is greater than the second duration T2=12min; at 10:37, the remaining spatial offset of the center of the first boundary relative to the historical stable boundary of the second gas concentration field is about 4.9m, which is greater than the third threshold of 3.0m, thus generating the first duration judgment result. Here, the first boundary is not a redefined object, but the updated position of the first boundary determined in Example 2 within the continuous time window: at 10:17, its center direction coordinate is 1936.8m, which corresponds to the intersection position of the first path near U07 in Example 3, and then continues to shift towards the depth of the goaf until 10:37.

[0053] Example 6: Obtaining the time-delay response judgment result along the first path This embodiment corresponds to steps S46-S49, which are used to verify whether the gas dilution migration reflected by the first path has a temporal relationship with the subsequent oxidation response along the same spatial direction, thereby further reducing misjudgments caused by single-point gas sensor malfunctions or transient ventilation disturbances.

[0054] S46. Following the direction from the first side to the first end region along the first path R1 determined in Example 3, the gas monitoring points corresponding to R1 spatially are sorted. Since Example 1 has established a one-to-one mapping relationship between P01-P08 and U01-U08, the effective gas monitoring points corresponding to R1=U01-U02-U03-U04-U05-U06-U07 are P01-P07, a total of 7 monitoring points. The first gas concentration of P01-P07 is retained for the continuous time window after 10:12, forming the first concentration change sequence.

[0055] S47, regarding the first At each monitoring point, the second gas concentration Based on this, establish the first response sequence. : An increase in the first response sequence indicates a greater dilution of the current gas concentration relative to its historical steady state. In this embodiment, CO is used as the primary oxidation response quantity, and the historical stable CO concentration at the corresponding monitoring point is used as the reference. Establish a second response sequence based on the baseline. : Temperature data is also retained for verification, which improves the confirmation of a sustained oxidation response when both CO and temperature continue to increase, but does not change the time delay acquisition method defined in the claims.

[0056] S48. Determine the fluctuation thresholds for the first and second response sequences based on historical stable data. When the first response sequence exceeds its threshold for three consecutive sampling windows, the time of the first time it exceeds the threshold is determined as the first response time. When the second response sequence exceeds its threshold for three consecutive sampling windows, the time of the first time the threshold is exceeded is determined as the second response time. First response delay for: This indicates that gas dilution enhancement occurs first, followed by enhanced oxidation response.

[0057] S49. Count the number of monitoring points with a first response time delay greater than zero on the first path. Compared with the total number of valid monitoring points To obtain the first ratio : In this embodiment, the fourth threshold is set to 0.75. When Furthermore, when arranged according to the first path from shallow to deep, the first response time delay satisfies a non-decreasing relationship, that is... If the result is positive, a first time-delay response judgment result is generated; otherwise, a second time-delay response judgment result is generated.

[0058] This embodiment still uses P01-P07 and the same 10:00 ventilation adjustment event. The timing results of the first response and the second response are shown in Table 6.

[0059] Table 6: Response Time and First Response Delay Judgment Results of Monitoring Points along the First Path As shown in Table 6, the first response delays of the seven valid monitoring points (P01-P07) are 7 min, 8 min, 9 min, 10 min, 11 min, 12 min, and 13 min, respectively, all greater than 0. The first ratio η = 7 / 7 = 1.00, which is higher than the fourth threshold of 0.75. Furthermore, starting from P02, the first response delay of each monitoring point is not less than the first response delay of the previous monitoring point. Therefore, the first path from the first side to the first end region satisfies a non-decreasing change relationship. By 10:35, the second response time corresponding to P07 has appeared, and the first delay response judgment result is generated accordingly.

[0060] Example 7 This embodiment combines the results of whether a second region is generated, the determination of the continuity of anomalies, and the determination of the time-delay response into a clear risk level, without adopting the single-point threshold rule of "the higher the gas concentration, the higher the risk".

[0061] To facilitate system execution, whether a second region is generated is denoted as... When generated When not generated The result of the first persistence judgment is denoted as The result of the second persistence judgment is denoted as The result of the first time-delay response judgment is recorded as... The result of the second time-delay response judgment is denoted as Spontaneous combustion risk level It can be expressed logically as follows: in, Corresponding to the first level of spontaneous combustion risk, Corresponding to the second level of spontaneous combustion risk, This corresponds to the third level of spontaneous combustion risk.

[0062] Example 4 has generated a second region consisting of U07 and U18-U21. Example 5 completed the first persistence judgment result at 10:37, therefore, the persistence judgment can be recorded as the first starting from the next complete 5-minute identification time window at 10:40; Example 6 completed the first time-delay response judgment result for P01-P07 along the first path at 10:35. Therefore, this example no longer uses the unrelated abstract samples W01-W06, but directly uses the continuous time window of 10:20-10:50 to determine the risk level window by window according to the aforementioned results.

[0063] Table 7 shows the risk level combination determination for seven consecutive time windows under the same event.

[0064] Table 7: Sample of Combinations for Determining Spontaneous Combustion Risk Levels As shown in Table 7, although the second region existed between 10:20 and 10:30, the first persistence judgment result and the first time-delay response judgment result had not yet been formed simultaneously, thus it was classified as the first spontaneous combustion risk level. At 10:35, Example 6 had confirmed the first time-delay response judgment result, while the first persistence judgment result of Example 5 had not yet been completed, so it was upgraded to the second spontaneous combustion risk level. After 10:40, both types of evidence from Examples 5 and 6 were valid, so it was upgraded to the third spontaneous combustion risk level. The risk level is thus connected to the same second region in Table 4, the continuous boundary response in Table 5, and the same path time-delay result in Table 6 on a time-window basis.

[0065] Example 8 This embodiment is used to continue tracking the spatial migration of the same second region U07, U18-U21 and its subsequent extended units, based on the continuous risk assessment in Embodiment 7. The time windows are still 10:20, 10:25, 10:30, 10:35, 10:40, 10:45 and 10:50, which is completely consistent with Table 7, and no new independent sample time axis is set.

[0066] This embodiment updates the working face advancement position time-series data and gas concentration time-series data at fixed 5-minute time intervals, and re-executes embodiments one through seven. For the second region obtained from two adjacent time windows... and First, obtain the degree of spatial overlap between the two. : in, This indicates the number of spatial measurement units contained in the region. This is used to determine whether the second region of adjacent time windows has spatial continuity.

[0067] Then, the center position of the second region is calculated. Let the... The second region of the time window contains Each spatial measurement unit, its central location for: Region shift vector of adjacent time windows and distance traveled They are respectively: Will Unit vector in the direction of face advance The projection on the surface is used as the direction criterion: when When the direction of movement is the same as that of the deeper goaf area, the second region is determined to have migrated towards the deeper goaf area; otherwise, it is determined to have moved back towards the working face. The direction of movement, the distance of movement, and the corresponding risk level are correlated to form a dynamic identification result.

[0068] The risk levels in Table 7 are correlated with the center location, spatial overlap, and movement distance of the second region within the same time window. The continuous results are shown in Table 8.

[0069] Table 8: Dynamic migration samples in the second region As shown in Table 8, the second region maintained a high degree of spatial overlap throughout the continuous time window from 10:20 to 10:50, with its center continuously migrating from a strike coordinate of 1941.2m to 1945.0m deeper into the goaf. The risk level was completely consistent with Table 7, escalating from the first spontaneous combustion risk level to the second and then to the third spontaneous combustion risk level. At 10:50, the system output the dynamic identification result: "Third spontaneous combustion risk level, the center of the second region migrating deeper into the goaf, and a movement of 0.5m in the last 5 minutes." Thus, the monitoring data from P01-P08 in Table 1, after processing through the spatial concentration field in Table 2, path inversion in Table 3, goaf age screening in Table 4, anomaly persistence judgment in Table 5, time lag judgment in Table 6, and risk classification in Table 7, were finally continuously transmitted to the dynamic identification result in Table 8.

[0070] Example 9 The intelligent identification system for spontaneous combustion risk in mines based on gas concentration fields includes a data acquisition module, a first identification module, a second identification module, and a risk judgment module. Each module transmits data through a unified spatial measurement unit number and a time window number, thereby avoiding errors in reference relationships between modules due to inconsistencies in coordinate or time references.

[0071] The data acquisition module obtains spatial location and time-series data of gas monitoring points from the gas monitoring system, time-series data of advance position from the working face production record system, air volume, fan frequency, and damper status from the ventilation monitoring system, and CO and temperature data from the mine gas and temperature monitoring system. The data acquisition module performs the processing corresponding to Example 1 and outputs the first gas concentration field, the second gas concentration field, and data from spatially numbered measurement units.

[0072] The first identification module receives the first gas concentration field and the second gas concentration field, performs the corresponding processing in Examples 2 and 3, and outputs the first gas dilution region, the first boundary, the first path, and the first end region in sequence. The key output of this module is not a single gas concentration alarm, but the air leakage and oxygen replenishment path and its end position obtained by inverting the spatial morphology of the gas concentration field.

[0073] The second identification module receives the working face advance position timing data, the first path and the first end area, performs the corresponding processing in Example 4, generates the mining age for each spatial measurement unit, and outputs the second area.

[0074] The risk assessment module receives the first gas concentration field, ventilation operation data, second area, first path and oxidation response data in a continuous time window, executes Examples 5 and 6 respectively to obtain the continuous assessment results and the time delay response assessment results, then obtains the first, second or third spontaneous combustion risk level according to Example 7, and forms the dynamic identification result according to Example 8.

[0075] Table 9 shows an example of data transfer relationships between system modules.

[0076] Table 9 System Module Input / Output Samples The system's workflow is a progressive identification process: "gas concentration point value → spatial concentration field → first gas dilution zone → first path → first terminal area → second area → spontaneous combustion risk level". This data stream allows the gas concentration field to primarily serve the purpose of inverting air leakage and oxygen supply conditions, rather than directly equating gas concentration itself with coal spontaneous combustion risk; then, the spatial inversion results are constrained layer by layer by the goaf age, the persistence of concentration field anomalies, and the time-delay response along the path.

[0077] It should be noted that all calculation formulas in this application employ regression analysis, including but not limited to machine learning algorithms, to deeply analyze the collected parameters and identify their natural trends and interrelationships. Specialized software, such as Python's Scikit-learn library or the R language, is used to automatically generate mathematical models that match the data. Then, cross-validation and other methods are used to objectively evaluate the model performance, and continuous feedback and optimization are combined to ensure that the created formulas truly reflect the inherent laws of the data, thereby guaranteeing their effectiveness and accuracy. In all calculation formulas in this application, the parameters in each formula undergo dimensionless processing within a consistent range to ensure that different physical quantities are compared on the same scale; dimensionless processing techniques include, but are not limited to, min-max-normalization and Z-score standardization. The algorithm of this invention is implemented as a Python script. Before executing the core logic, the program first executes a data loading module (e.g., using the widely used pandas library in Python) configured to read the aforementioned spreadsheet file and load its contents into the program's working memory (e.g., a DataFrame data structure). Subsequent algorithm steps will directly query and retrieve the required configuration parameters from this in-memory data structure.

[0078] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for intelligent identification of spontaneous combustion risk in mines based on gas concentration field, characterized in that, The specific steps include: S1. Acquire spatial location data of gas monitoring points, time series data of gas concentration, time series data of working face advancement position, ventilation operation data, and oxidation response data of the target mine goaf, and divide the target mine goaf into several spatial measurement units; construct a first gas concentration field based on the time series data of gas concentration of each gas monitoring point within the current time window, and construct a second gas concentration field based on the historical stable state gas concentration data corresponding to the current ventilation operation status; S2. Based on the first gas concentration field and the second gas concentration field, obtain the gas concentration change results and gas concentration spatial gradient of each spatial measurement unit, determine the first gas dilution region according to the gas concentration change results and gas concentration spatial gradient, and obtain the first boundary according to the first gas dilution region; generate the first path according to the spatial connection relationship of each spatial measurement unit in the first gas dilution region and the direction of the gas concentration spatial gradient, and determine the first end region according to the first path; S3. Based on the time sequence data of the working face advancement position, determine the time when each spatial calculation unit enters the goaf and obtain the goaf age of each spatial calculation unit; based on the first path, the first end area and the goaf age of each spatial calculation unit, filter the first end area and its adjacent spatial calculation units to generate the second area. S4. Based on ventilation operation data and the first gas concentration field corresponding to the continuous time window, obtain the abnormal persistence judgment result of the first gas concentration field; based on the gas concentration change and oxidation response data on the first path, obtain the time delay response judgment result along the first path; according to the second region, the abnormal persistence judgment result and the time delay response judgment result, determine the first spontaneous combustion risk level, the second spontaneous combustion risk level or the third spontaneous combustion risk level corresponding to the goaf of the target mine.

2. The intelligent identification method for spontaneous combustion risk in mines based on gas concentration field according to claim 1, characterized in that, Step S1 includes: S11. Obtain the three-dimensional spatial coordinates of each gas monitoring point, and map each gas monitoring point to the corresponding spatial calculation unit according to the unified mine spatial coordinate system. S12. Divide the time series data of gas concentration at each gas monitoring point into segments according to a fixed-length sliding time window, obtain the first gas concentration of each gas monitoring point in the current time window, and obtain the second gas concentration of each gas monitoring point in the historical stable time window under the current ventilation operation state. S13. Based on the three-dimensional spatial coordinates of each gas monitoring point and the first gas concentration, the first concentration value corresponding to each spatial measurement unit is obtained by spatial interpolation method, and the first gas concentration field is constructed according to each first concentration value. S14. Based on the three-dimensional spatial coordinates and the second gas concentration of each gas monitoring point, the second concentration value corresponding to each spatial measurement unit is obtained by using the same spatial interpolation method as the first gas concentration field, and the second gas concentration field is constructed according to each second concentration value. The oxidation response data includes at least one of carbon monoxide concentration data and temperature data.

3. The intelligent identification method for spontaneous combustion risk in mines based on gas concentration field according to claim 2, characterized in that, Determining the first gas dilution zone and the first boundary in step S2 includes: S21. For any i-th spatial measurement unit, obtain the difference between the second concentration value corresponding to the second gas concentration field and the first concentration value corresponding to the first gas concentration field, and use the difference as the gas concentration reduction amount of the i-th spatial measurement unit. S22. Based on the first concentration value of the i-th spatial measurement unit and its spatially adjacent units, the spatial gradient calculation method is used to obtain the spatial gradient of the gas concentration of the i-th spatial measurement unit and the corresponding gradient direction. S23. The spatial measurement unit in which the gas concentration reduction reaches the first threshold is determined as the first spatial measurement unit. Multiple first spatial measurement units that are spatially adjacent and have continuous gradient directions are connected to form the first gas dilution zone. S24. Obtain the side of the first gas dilution zone that connects to the working face or goaf boundary as the first side, and obtain the side boundary of the first gas dilution zone that is farthest from the first side along the deep direction of the goaf, and determine the boundary as the first boundary.

4. The intelligent identification method for spontaneous combustion risk in mines based on gas concentration field according to claim 3, characterized in that, The step S2, which generates the first path and determines the first terminal region, includes: S25. Select a spatial measurement unit connected to the first spatial measurement unit from the first side of the first gas dilution zone as the first starting unit; S26. Starting from the first starting unit, in adjacent spatial units, select spatial measurement units in sequence where the first concentration value increases along the direction of the depth of the goaf and the angle between the spatial gradient direction of the gas concentration and the gradient direction of the previous spatial measurement unit does not exceed the second threshold, to form multiple candidate connection paths. S27. Obtain the continuous connection length and gradient direction consistency between adjacent spatial measurement units for each candidate connection path. Sort the candidate connection paths according to the continuous connection length and gradient direction consistency. Determine the candidate connection path that satisfies the first condition as the first path. S28. Obtain the spatial measurement unit where the first path intersects with the first boundary, and jointly determine the spatial measurement unit that is spatially adjacent to the spatial measurement unit and located on the deep side of the goaf area of ​​the first boundary as the first end region. Wherein, the first path represents the air leakage and oxygen supply path corresponding to the first gas dilution area, and the first terminal area represents the terminal position of the air leakage and oxygen supply extending along the first path into the depth of the goaf.

5. The intelligent identification method for spontaneous combustion risk in mines based on gas concentration field according to claim 4, characterized in that, Step S3 includes: S31. Based on the time sequence data of the working face advance position, obtain the first time when the working face passes through the corresponding position of each spatial calculation unit in sequence, and determine the time interval between the current time and the first time corresponding to each spatial calculation unit as the mining age of each spatial calculation unit. S32. Obtain the age of the goaf at the spatial location where the oxidation response data in the historical mine spontaneous combustion monitoring data begins to change continuously, and obtain the lower limit of the first goaf age and the upper limit of the first goaf age based on the goaf age. S33. From the spatial calculation units within the preset spatial range of the first end area and the first path end, select the spatial calculation units whose goaf age is between the lower limit of the first goaf age and the upper limit of the first goaf age, and determine them as the second spatial calculation units. S34. Connect the adjacent second spatial measurement units to generate a second region; The second region represents the potential oxidation heat storage region where the leakage oxygen supply end of the first path and the oxidation time condition of the residual coal are simultaneously satisfied.

6. The intelligent identification method for spontaneous combustion risk in mines based on gas concentration field according to claim 5, characterized in that, The step S4, which involves obtaining the result of the abnormal persistence judgment of the first gas concentration field, includes: S41. Based on the ventilation operation data, obtain the second time when the ventilation operation parameters are adjusted, and obtain the third time when the ventilation operation parameters reach a stable state after the adjustment; S42. Starting from the third time, construct the corresponding first gas concentration fields according to the continuous sliding time windows, and determine the first boundary corresponding to each first gas concentration field according to the method described in claim 3. S43. Obtain the spatial displacement between the first boundary corresponding to adjacent time windows, and obtain the fourth time when the first boundary reaches a stable state based on multiple consecutive spatial displacements, and determine the time interval between the third time and the fourth time as the first duration. S44. Based on historical monitoring data that are the same type of adjustment as the current ventilation operation parameters and have not shown any spontaneous combustion abnormalities, the second duration is obtained using the same method as steps S41-S43. S45. When the first duration is longer than the second duration, and the remaining spatial offset of the first boundary corresponding to the fourth time relative to the corresponding position of the second gas concentration field reaches the third threshold, a first duration judgment result is generated; otherwise, a second duration judgment result is generated. The first continuous judgment result indicates that the gas concentration field still has an abnormal and continuous shift after the ventilation operation status has stabilized, while the second continuous judgment result indicates that the change in the gas concentration field is consistent with the normal response corresponding to the adjustment of ventilation operation parameters.

7. The intelligent identification method for spontaneous combustion risk in mines based on gas concentration field according to claim 6, characterized in that, The step S4, obtaining the time-delay response judgment result along the first path, includes: S46. Sort the gas monitoring points on the first path according to the direction from the first side to the first end area, and obtain the first concentration change sequence corresponding to each gas monitoring point. S47. Obtain the decrease in the first concentration change sequence of each gas monitoring point relative to the second concentration value corresponding to the second gas concentration field, and determine the change sequence of the decrease over time as the first response sequence; obtain the oxidation response data of the same gas monitoring point or spatially adjacent to its corresponding spatial measurement unit, and determine the change sequence of the increase in the oxidation response data relative to the historical stable value over time as the second response sequence; S48. For each gas monitoring point, obtain the first response time when the first response sequence begins to increase continuously, and obtain the second response time when the second response sequence begins to increase continuously. Determine the time difference between the second response time and the first response time as the first response time delay. S49. According to the sorting results of gas monitoring points on the first path, compare the first response time delays of each gas monitoring point; when the first response time delay corresponding to the gas monitoring point that reaches the fourth threshold number ratio is greater than zero, and the first response time delay along the first path towards the first end area satisfies the non-decreasing change relationship, generate the first time delay response judgment result; otherwise, generate the second time delay response judgment result. The first time-delay response judgment result indicates that the gas dilution change and the subsequent oxidation response form a time-delay correlation with spatial order along the first path, while the second time-delay response judgment result indicates that the time-delay correlation has not been formed.

8. The intelligent identification method for spontaneous combustion risk in mines based on gas concentration field according to claim 7, characterized in that, Determining the spontaneous combustion risk level in step S4 includes: When a second region is not generated, or when a second region is generated but the abnormal persistence judgment result is the second persistence judgment result and the time delay response judgment result is the second time delay response judgment result, it is determined to be the first spontaneous combustion risk level; When a second region has been generated, and the abnormal persistence judgment result is either the first persistence judgment result or the time delay response judgment result is either the first time delay response judgment result, it is determined to be the second spontaneous combustion risk level. When a second region has been generated, and the abnormal persistence judgment result is the first persistence judgment result, and the time delay response judgment result is the first time delay response judgment result, it is determined to be the third spontaneous combustion risk level. Among them, the spontaneous combustion risk level corresponding to the third spontaneous combustion risk level is higher than that of the second spontaneous combustion risk level, and the spontaneous combustion risk level corresponding to the second spontaneous combustion risk level is higher than that of the first spontaneous combustion risk level.

9. The intelligent identification method for spontaneous combustion risk in mines based on gas concentration field according to claim 8, characterized in that, Also includes: Update the working face advance position time series data and gas concentration time series data at fixed time intervals, and reconstruct the first gas concentration field for each updated time window; Update the first gas dilution zone, the first boundary, the first path, the first terminal zone, the goaf age of each spatial measurement unit, and the second zone respectively; Obtain the spatial overlap of the second region corresponding to adjacent time windows and the center position of the region, and obtain the moving direction and moving distance of the second region relative to the working surface based on the continuous change of the center position of the region. The movement direction and distance of the second area are correlated with the corresponding first, second, or third spontaneous combustion risk level to generate a dynamic identification result of spontaneous combustion risk in the goaf of the target mine.

10. A mine spontaneous combustion risk intelligent identification system based on a gas concentration field, applied to the mine spontaneous combustion risk intelligent identification method based on a gas concentration field as described in any one of claims 1 to 9, characterized in that, include: The data acquisition module is used to acquire spatial location data of gas monitoring points, time series data of gas concentration, time series data of working face advancement position, ventilation operation data and oxidation response data of the target mine goaf, divide the target mine goaf into several spatial measurement units, and construct the first gas concentration field and the second gas concentration field. The first identification module is used to obtain the gas concentration change results and gas concentration spatial gradient of each spatial measurement unit based on the first gas concentration field and the second gas concentration field, determine the first gas dilution region and the first boundary, generate the first path according to the spatial connection relationship of each spatial measurement unit and the direction of the gas concentration spatial gradient, and determine the first end region according to the first path. The second identification module is used to determine the mined-out age of each spatial measurement unit based on the time sequence data of the working face advancement position, and to generate a second region based on the first path, the first end region and the mined-out age of each spatial measurement unit. The risk assessment module is used to obtain the abnormal persistence assessment result based on ventilation operation data and the first gas concentration field corresponding to the continuous time window, obtain the time delay response assessment result based on the gas concentration change and oxidation response data on the first path, and determine the first spontaneous combustion risk level, the second spontaneous combustion risk level, or the third spontaneous combustion risk level according to the second region, the abnormal persistence assessment result, and the time delay response assessment result.

Images