Mine intelligent alarm method and system based on multi-dimensional perception information

Abstract

The application provides a mine intelligent alarm method and system based on multi-dimensional perception information, relates to the technical field of intelligent alarm, extracts abnormal driving criteria corresponding to mine abnormal events from all mine safety state information, associates and judges the abnormal driving criteria with the protection response limit value of the partition where the monitoring terminal is located, obtains the alarm situation index of the region where the mine abnormal event is located, when the abnormal driving criteria trigger the alarm condition, generates dynamic alarm information within the period from the occurrence of the abnormality to the cancellation through the alarm judgment rule, classifies and disposes the dynamic alarm information, obtains the alarm marker value corresponding to each abnormality level, determines the risk warning level in the mine according to the alarm situation index and all alarm marker values, and performs linkage alarm according to the risk warning level. The application can construct a full-link collaborative alarm mechanism for global linkage disposal of mine abnormal events, so as to improve the accuracy of mine underground safety risk prevention and control.

Description

Technical Field

[0001] This application relates to the field of intelligent alarm technology, and more specifically, to a mine intelligent alarm method and system based on multi-dimensional sensing information. Background Technology

[0002] Intelligent alarms rely on various deployed field sensing devices to collect multi-dimensional environmental, equipment, and operational condition monitoring data. This data, combined with digital analysis methods, replaces the traditional single-threshold trigger-based alarm safety management approach. This method can perform time-series analysis, feature filtering, and correlation judgment on multi-source sensing information such as gas, dust, roadway pressure, and ventilation parameters. Based on preset hierarchical judgment rules, it completes the identification of abnormal events and can generate time-series alarm content according to the dynamic process of risk development. Different alarm levels are divided according to the severity of the abnormality. Intelligent alarms can match corresponding emergency response strategies with the protection and control requirements of the monitoring area, accurately push hierarchical alarm information to the corresponding control personnel and underground linkage equipment, and complete regional collaborative alarm response through communication transmission links. This realizes intelligent alarm management and control of the entire process of mine safety anomalies, from identification, hierarchical classification, information push to on-site linkage execution.

[0003] However, in the existing underground mine safety monitoring and alarm management process, the lack of a full-link collaborative judgment mechanism for multi-zone emergency response results in fragmented execution logic among multiple underground devices. This leads to a high probability of misjudgment and missed detection of abnormal events, risk warning results that fail to align with the risk distribution characteristics of the underground space, and poor adaptability of alarm response strategies to zone protection levels, making it difficult to achieve proactive risk warning and comprehensive collaborative emergency response. Therefore, how to construct a full-link collaborative alarm mechanism for comprehensive response to abnormal events in mines to improve the accuracy of underground mine safety risk prevention and control is a problem facing the industry. Summary of the Invention

[0004] This application provides a mine intelligent alarm method and system based on multi-dimensional perception information, which can construct a full-link collaborative alarm mechanism for the full-domain linkage and handling of abnormal events in mines, so as to improve the accuracy of mine underground safety risk prevention and control.

[0005] In a first aspect, this application provides a mine intelligent alarm method based on multi-dimensional sensing information, the alarm method comprising the following steps:

[0006] Collect safety status information from monitoring terminals deployed in various underground sections of the mine;

[0007] Extract the anomaly driving criteria corresponding to the abnormal events in the mine from all mine safety status information, and associate the anomaly driving criteria with the protection response limit of the zone where the monitoring terminal is located to obtain the alarm status index of the area where the abnormal event in the mine is located.

[0008] When the anomaly driving criterion triggers the alarm condition, dynamic alarm information is generated from the occurrence of the anomaly to the period of its resolution through the alarm judgment rule. The dynamic alarm information is then classified and processed to obtain the alarm flag value corresponding to each anomaly level.

[0009] The risk warning level within the mine is determined based on the alarm status index and all alarm marker values, and a linkage alarm is triggered according to the risk warning level.

[0010] In this embodiment, the specific criteria for extracting the anomaly-driven judgment corresponding to the abnormal events in the mine from all mine safety status information include:

[0011] Based on all mine safety status information, mutual information causal screening of multi-source monitoring variables is performed to obtain abnormal situation characteristics;

[0012] The abnormal driving factors of the abnormal events are determined by the abnormal situation characteristics and the time sequence information of the abnormal events.

[0013] The anomaly driving criteria corresponding to the abnormal events in the mine are determined based on the aforementioned anomaly driving factors and the composite disaster change attributes of the mine.

[0014] In this embodiment, the alarm status index of the area where the mine anomaly event is located is obtained by associating the anomaly driving criterion with the protection response limit of the partition where the monitoring terminal is located. Specifically, this includes:

[0015] Based on the anomaly driving criterion and the protection response limit of the partition where the monitoring terminal is located, determine the partition response association feature base;

[0016] The correlation situation mapping is performed on the partition response correlation feature base and the real-time sensing data of the multi-source monitoring terminals within the partition to obtain the alarm situation hot zone gradient of the abnormal event.

[0017] The alarm status hot zone gradient and the dynamic margin threshold of the protection response limit are compensated for by risk attenuation to generate an alarm status index for the area where the mine abnormal event is located.

[0018] In this embodiment, determining the partition response association feature base based on the anomaly driving criterion and the protection response limit of the partition where the monitoring terminal is located specifically includes:

[0019] The early warning coupling tensor is determined based on the critical combination of parameters in the anomaly driving criterion and the graded threshold range of the protection response limit;

[0020] By performing basis projection on the causal origin features of the warning coupling tensor and the anomaly driving criterion, a partitioned response association feature basis is generated.

[0021] In this embodiment, when the anomaly-driven criterion triggers the alarm condition, generating dynamic alarm information for the period from the occurrence of the anomaly to its resolution through alarm determination rules specifically includes:

[0022] Based on the alarm conditions triggered by the anomaly driving criteria and the combination of criteria parameters, the initial state vector and temporal evolution quantity of the abnormal event are determined, and the abnormal situation evolution interval is obtained.

[0023] By using the alarm level mapping data in the abnormal situation evolution interval and alarm judgment rules, the various stages of the abnormal occurrence are hierarchically associated to generate a dynamic alarm level sequence from the occurrence of the abnormality to the time period of its resolution.

[0024] Extract the level transition nodes and duration features from the dynamic alarm level sequence, match them with the risk push constraints of the protection response limit, and generate dynamic alarm information.

[0025] In this embodiment, the dynamic alarm information is processed in a hierarchical manner to obtain the alarm flag value corresponding to each anomaly level, specifically including:

[0026] Based on the hierarchical handling rule base of the dynamic alarm information and the zone protection response limit, determine the hierarchical handling coupling coefficient when the alarm level sequence matches the handling strategy;

[0027] Based on the hierarchical processing coupling coefficient, the hierarchical jump nodes in the dynamic alarm information are strategically encoded to generate a preliminary alarm marker sequence;

[0028] By performing boundary correction using the preliminary alarm tag sequence and the initial state vector of the anomaly driving criterion, the alarm tag value corresponding to each anomaly level is obtained.

[0029] In this embodiment, the strategy encoding of the level transition nodes in the dynamic alarm information based on the hierarchical handling coupling coefficient to generate a preliminary alarm marker sequence specifically includes:

[0030] The node handling encoding vector is determined based on the hierarchical handling coupling coefficient and the abnormal time period corresponding to the level jump node in the dynamic alarm information.

[0031] The joint coding features of the handling strategy and response parameters are extracted from the temporal evolution of the node handling coding vector and the anomaly driving criterion to generate a preliminary alarm tag sequence.

[0032] In this embodiment, determining the risk warning level within the mine based on the alarm status index and all alarm marker values ​​specifically includes:

[0033] The regional risk intensity characteristics are determined based on the alarm status index and the regional distribution vector of all alarm marker values.

[0034] By mapping the regional risk intensity characteristics and the mine risk benchmark threshold, a corrected risk warning level is generated.

[0035] The risk warning level within the mine is obtained by making a transition judgment on the predicted trend of the modified risk warning level and alarm status index.

[0036] In this embodiment, triggering a linked alarm based on the risk warning level specifically includes:

[0037] The alarm strategy coupling vector is determined based on the aforementioned risk warning level and alarm linkage response conditions;

[0038] The alarm strategy coupling vector is used to fuse the collaborative triggering parameters of abnormal events in the mine, generate a linkage alarm command, and execute it.

[0039] Secondly, this application provides a mine intelligent alarm system based on multi-dimensional sensing information, used to execute a mine intelligent alarm method based on multi-dimensional sensing information, the alarm system comprising:

[0040] The data acquisition module is used to collect safety status information from monitoring terminals deployed in various sections of the underground mine.

[0041] The judgment module is used to extract the anomaly driving criteria corresponding to the abnormal events in the mine from all mine safety status information, and to associate the anomaly driving criteria with the protection response limit of the zone where the monitoring terminal is located to obtain the alarm status index of the area where the abnormal event in the mine is located.

[0042] The processing module is used to generate dynamic alarm information from the occurrence to the resolution period through alarm judgment rules when the abnormality driving criterion triggers the alarm condition, and to perform hierarchical processing on the dynamic alarm information to obtain the alarm flag value corresponding to each abnormality level.

[0043] The execution module is used to determine the risk warning level in the mine based on the alarm status index and all alarm marker values, and to trigger a linkage alarm based on the risk warning level.

[0044] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects:

[0045] Collecting safety status information collected from monitoring terminals arranged in each subarea of an underground mine; extracting an anomaly driving criterion corresponding to a mine anomaly event from all mine safety status information, performing association judgment on the anomaly driving criterion and a protection response limit of the subarea where the monitoring terminal is located, and obtaining an alarm situation index of the area where the mine anomaly event is located; when the anomaly driving criterion triggers an alarm condition, generating dynamic alarm information in a period from anomaly occurrence to anomaly elimination through an alarm judgment rule, performing graded disposal on the dynamic alarm information, and obtaining an alarm marking value corresponding to each anomaly level; determining a risk early warning level in the mine according to the alarm situation index and all alarm marking values, and performing linked alarm according to the risk early warning level.

[0046] It can be seen from this that in the present application, linked alarm can be performed according to the risk early warning level; wherein, obtaining the alarm situation index can yield a comprehensive risk quantitative characterization result of the whole area and spatially continuous distribution of the subarea where the mine anomaly event is located, thereby breaking the limitation that traditional single-point monitoring data can only characterize the risk of local points, restoring the real risk diffusion range and intensity gradient in the underground mine based on multi-dimensional protection limit constraints and spatial risk attenuation correction logic, realizing horizontal quantitative benchmarking of risk levels among different underground subareas, providing a standardized global risk reference for unified research and judgment of global alarm levels and overall allocation of cross-regional emergency resources, ensuring the objectivity and balance of whole-mine risk assessment from the spatial dimension, and enabling the coverage range and response priority of linked disposal to accurately match the actual distribution characteristics of risks; obtaining the alarm marking value can yield standardized risk priority quantitative identifiers corresponding to different anomaly event levels, thereby converting unstructured alarm evolution information under dynamic time-series evolution into standardized numerical characteristics that can be weighted and aggregated, realizing vertical quantitative description of risk severity in the whole cycle of occurrence, development and regression of anomaly events, accurately anchoring the adaptive relationship of differentiated disposal strategies corresponding to different anomaly levels, providing local time-series risk support for correction and optimization of the global risk early warning level, enabling global linked disposal to hierarchically match push objects, starting sequence and equipment linkage range according to the severity of anomaly evolution, and improving the fine disposal logic of full-link coordinated alarm from the time-series grading dimension. In conclusion, the technical solution adopted by the present application can construct a full-link coordinated alarm mechanism for global linked disposal of mine anomaly events, so as to improve the accuracy of safety risk prevention and control in underground mines. Description of Drawings

[0047] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following briefly introduces the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only the embodiments of the present application, and those skilled in the art can also obtain other drawings according to these drawings without creative effort.

[0048] Figure 1 This is an exemplary flowchart of a mine intelligent alarm method based on multi-dimensional sensing information provided in this application;

[0049] Figure 2 This is a flowchart illustrating the process of determining alarm flag values ​​provided in this application;

[0050] Figure 3 This is a schematic diagram of the mine linkage alarm principle provided in this application;

[0051] Figure 4 This is a modular structure diagram of a mine intelligent alarm system based on multi-dimensional sensing information provided in this application. Detailed Implementation

[0052] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0053] This application provides a mine intelligent alarm method and system based on multi-dimensional sensing information. Its core is to collect safety status information from monitoring terminals deployed in various underground sections of the mine; extract anomaly driving criteria corresponding to mine anomalies from all mine safety status information; correlate these anomaly driving criteria with the protection response limits of the section where the monitoring terminals are located to obtain an alarm status index for the area where the mine anomaly occurs; when the anomaly driving criteria trigger an alarm condition, generate dynamic alarm information for the period from the occurrence of the anomaly to its resolution through alarm judgment rules; classify and process the dynamic alarm information to obtain an alarm marker value corresponding to each anomaly level; determine the risk warning level within the mine based on the alarm status index and all alarm marker values; and trigger a linked alarm based on the risk warning level.

[0054] Example 1: To better understand the above technical solution, the following will provide a detailed description of the technical solution in conjunction with the accompanying drawings and specific implementation methods. (Refer to...) Figure 1 As shown in the figure, this is an exemplary flowchart of a mine intelligent alarm method based on multi-dimensional sensing information according to this embodiment of the present application. The alarm method includes the following steps:

[0055] In step S1, safety status information is collected from monitoring terminals deployed in various underground sections of the mine.

[0056] In practice, the collection of safety status information from monitoring terminals deployed in various underground sections of the mine can be achieved in the following way: The mine is divided into independent zones according to its functional layout, such as mining faces, transport roadways, electromechanical chambers, and return air roadways. Monitoring terminals adapted to the monitoring needs are deployed at corresponding points in each zone. The terminals are connected to the nearest underground data substation via an RS485 bus. The ground monitoring host sends collection instructions to each substation according to a preset sampling cycle. Gas parameters such as methane and carbon monoxide are sampled once per second, while parameters such as dust concentration, roof pressure, and roadway wind speed are sampled once per minute. Each set of raw data is synchronously bound to a unique terminal number, its zone code, and a collection timestamp. After being aggregated by the underground substations, the data is uploaded to the ground monitoring platform via an industrial Ethernet ring network. Throughout the upload process, cyclic redundancy checks are used to verify data integrity, automatically removing invalid data exceeding the sensor's physical range and lost data. Valid data that passes the checks is written into a real-time database to form mine safety status information.

[0057] It should be noted that, in this application, the monitoring terminal refers to the field sensing equipment deployed in various underground zones of the mine; the safety status information refers to the real-time monitoring data and status parameter set of the environmental and equipment operation safety status in various underground zones of the mine; wherein, the real-time monitoring data is the quantitative measured data periodically collected by various underground sensing terminals, bound with zone codes and timestamps, and is divided into three categories according to the monitoring object: environmental gas monitoring data, surrounding rock and roadway monitoring data, and electromechanical equipment operating condition monitoring data, specifically including: environmental gas monitoring data, methane concentration, carbon monoxide concentration, oxygen concentration, carbon dioxide concentration, hydrogen sulfide real-time concentration, underground air temperature and humidity; surrounding rock and roadway monitoring data, roof subsidence, roadway cross-sectional convergence deformation, support stress, mine pressure load, respirable dust / total dust concentration, roadway ventilation wind speed, and wind pressure; electromechanical equipment operating condition monitoring data, operating current, voltage, speed, start-stop status of main ventilation fan, drainage pump, and belt conveyor, substation cable temperature, switch fault alarm signal, and equipment overload load parameters. Each set of monitoring data is accompanied by a unique device number of the acquisition terminal, the code of the underground zone to which it belongs, and a data acquisition timestamp accurate to the second. After verification by the underground substation to remove out-of-range, lost, or invalid data, it is stored in the real-time database.

[0058] In step S2, the abnormal driving criteria corresponding to the abnormal events in the mine are extracted from all mine safety status information. The abnormal driving criteria are then correlated with the protection response limit of the zone where the monitoring terminal is located to obtain the alarm status index of the area where the abnormal events in the mine are located.

[0059] In this embodiment, the anomaly-driven criteria corresponding to abnormal events in mines can be extracted from all mine safety status information using the following steps:

[0060] Based on all mine safety status information, mutual information causal screening of multi-source monitoring variables is performed to obtain abnormal situation characteristics;

[0061] The abnormal driving factors of the abnormal events are determined by the abnormal situation characteristics and the time sequence information of the abnormal events.

[0062] The anomaly driving criteria corresponding to the abnormal events in the mine are determined based on the aforementioned anomaly driving factors and the composite disaster change attributes of the mine.

[0063] In practice, firstly, based on the multi-source monitoring variables covered by all mine safety status information, a complete continuous time series of each variable is constructed. The complete continuous time series samples are taken from the measured monitoring data of the mine's historical normal working conditions and archived abnormal working conditions. Based on the standard calculation formula of mutual information in information theory, the mutual information value between variables is calculated group by group. The mutual information value is non-negative and the larger the value, the stronger the correlation between variables. This is used to quantify the nonlinear correlation between variables. Then, the causal orientation relationship between variables is determined by combining the transfer entropy method. Redundant variables that have no causal relationship with the abnormal evolution are eliminated, and the set of variables that are highly sensitive to abnormal states and have a clear causal driving relationship is retained to obtain the abnormal situation characteristics. Then, the time series of abnormal situation characteristics are aligned with the time series information of abnormal events to locate the characteristic fluctuation time window before and after each historical abnormal event. The change amplitude, advance response duration, and fluctuation stability of each abnormal situation characteristic throughout the entire abnormal evolution cycle are statistically analyzed. Regular anomalies that last for no less than 30 minutes before the occurrence of an abnormal event are screened out. The regular anomalies include three standardized judgment conditions: continuous unidirectional increase or unidirectional decrease of monitoring parameters; parameter fluctuation amplitude exceeding the safety baseline by 20% within 5 minutes; periodic oscillation frequency ≥ 5 times within 10 minutes, which has a stable leading indication effect on the occurrence of anomalies. The stable leading indication effect is determined by two quantifiable indicators. The function is identified when both indicators are met: the feature advance response duration ≥ 15 minutes, which means that the occurrence time of the feature anomaly is significantly earlier than the node of the disaster anomaly outbreak; and the calculated value of the mutual information correlation between the feature and the abnormal event ≥ 0.7, which means that there is a strong nonlinear causal driving relationship between the two feature variables, and these are identified as the anomaly driving factors of the corresponding abnormal events.Finally, based on the complex hazard characteristics of mines, the critical change ranges and trend constraints of each anomaly driving factor are clarified under single-hazard and multi-hazard coupled scenarios. The critical change ranges are the risk threshold ranges defined in the mine safety regulations: under a single-hazard scenario, methane 0.8%–1.0%, carbon monoxide 24 ppm–30 ppm, roof support stress 12 MPa–15 MPa, and total dust concentration 4 mg / m³–6 mg / m³; under a multi-hazard coupled scenario, the critical ranges of various parameters are lowered by 20% overall, triggering risk assessment in advance for each type of mine anomaly. The system matches threshold combinations of corresponding abnormal driving factors. These threshold combinations are categorized into three standardized matching forms: single-factor over-limit combination: any driving factor reaching the lower limit of the critical change range triggers an early warning; dual-factor coupling combination: two related disaster driving factors simultaneously reach 80% of their respective critical range thresholds; and multi-factor linkage combination: three or more disaster driving factors simultaneously exceed the safety baseline by 50%. The system also includes a change rate limit, defined as the parameter increase per unit time threshold: gas parameters increase by ≥0.1% per minute, mine pressure stress increase by ≥1 MPa per hour, and dust concentration increase by ≥2 mg / m³ per 10 minutes. Furthermore, the system specifies abnormal duration requirements, categorized as follows: early warning level abnormal duration ≥5 min, general alarm level abnormal duration ≥3 min, and major risk alarm level abnormal duration ≥1 min. This forms quantifiable and executable standardized abnormal judgment rules, which serve as the abnormal driving criteria for mine abnormal events.

[0064] It should be noted that, in this application, "mining abnormal events" refers to various abnormal working conditions and disasters that occur underground in mines and endanger personnel and production safety; "multi-source monitoring variables" refers to quantifiable monitoring parameters from different types of monitoring terminals underground, corresponding to different safety dimensions; "mutual information causal screening" refers to information analysis methods that identify the causal relationship between monitoring variables and abnormal evolution and eliminate redundant variables; "abnormal situation characteristics" refers to the set of monitoring variables that are causally related to abnormal evolution; "abnormal event time series information" refers to information data that records the time nodes of occurrence, development, and dissipation of various abnormal events; "abnormal driving factors" refers to characteristic variables when mining abnormal events occur; "composite disaster change attributes" are inherent attributes that reflect the coupled evolution law between different disasters in the mine; and "abnormal driving criteria" refers to the quantitative rules and threshold standards for determining whether a mining abnormal event has occurred.

[0065] In this embodiment, the alarm status index of the area where the mine abnormal event is located is obtained by associating the anomaly driving criterion with the protection response limit of the zone where the monitoring terminal is located. This can be achieved by the following steps:

[0066] Based on the anomaly driving criterion and the protection response limit of the partition where the monitoring terminal is located, determine the partition response association feature base;

[0067] The correlation situation mapping is performed on the partition response correlation feature base and the real-time sensing data of the multi-source monitoring terminals within the partition to obtain the alarm situation hot zone gradient of the abnormal event.

[0068] The alarm status hot zone gradient and the dynamic margin threshold of the protection response limit are compensated for by risk attenuation to generate an alarm status index for the area where the mine abnormal event is located.

[0069] In practice, firstly, based on the protection level, personnel density, and disaster type of each underground zone in the mine, the pre-set protection response limits for each zone are retrieved. These limits are then categorized and organized according to the zone dimension. The previously obtained anomaly-driven criteria are decomposed into independent judgment parameters, and each parameter is matched against the corresponding zone's protection response limit. The deviation ratio, response priority weight, and over-limit triggering order of each criterion parameter are calculated. The matching results are then integrated into a structured feature set according to the zone dimension, forming the zone response association feature base for the corresponding zone. Next, real-time sensing data from all multi-source monitoring terminals within the zone are collected. After aligning the data timestamps, the data is matched to the corresponding parameter dimension of the zone response association feature base. The over-limit ratio of the real-time monitoring value relative to the protection response limit is calculated for each point. The underground spatial coordinates of each monitoring terminal are retrieved, and the existing commonly used inverse distance weighted interpolation method is used to map the over-limit ratio of discrete points into a continuous situational distribution field across the entire zone. Multiple gradient intervals are divided according to the degree of over-limit, resulting in the alarm situation hot zone gradient of the abnormal event. Finally, based on the protection level of the zone and the fluctuation characteristics of the monitoring parameters, a dynamic margin threshold corresponding to the protection response limit is set as a buffer zone for instantaneous data fluctuations. Combining the underground roadway topology, ventilation conditions and obstacle distribution, the spatial risk attenuation coefficient of different locations is calculated. The gradient values ​​of each gradient of the alarm status hot zone are attenuated and corrected one by one. After eliminating the fluctuation interference within the dynamic margin threshold range, the alarm status index of the area where the mine abnormal event is located is obtained by weighted summation according to the coverage area of ​​each gradient.

[0070] It should be noted that, in this application, the protection response limit refers to the safety critical threshold set for each zone in the mine; the correlation determination refers to the analytical logic of matching the anomaly driving criterion with the zone protection limit; the zone response correlation feature base refers to the set of benchmark features that integrates the correspondence between the anomaly criterion and the zone limit; real-time sensing data refers to the real-time safety parameter data collected on-site by the monitoring terminal; the correlation situation mapping refers to the processing method of converting discrete monitoring data into a continuous regional situation; the alarm situation hot zone gradient refers to the gradient level of the spatial distribution difference of abnormal risks; the dynamic margin threshold refers to the buffer threshold set next to the protection response limit; the risk attenuation compensation refers to the processing method of attenuating and correcting the spatial situation value; and the alarm situation index refers to the comprehensive index of the strength of the regional alarm situation.

[0071] In addition, in this embodiment, the determination of the partition response association feature basis based on the anomaly driving criterion and the protection response limit of the partition where the monitoring terminal is located can be achieved by the following steps:

[0072] The early warning coupling tensor is determined based on the critical combination of parameters in the anomaly driving criterion and the graded threshold range of the protection response limit;

[0073] By performing basis projection on the causal origin features of the warning coupling tensor and the anomaly driving criterion, a partitioned response association feature basis is generated.

[0074] In practical implementation, firstly, the critical combinations of parameters included in the anomaly driving criteria are decomposed, clarifying the critical numerical boundaries of each set of parameters and the collaborative triggering logic between parameters. Based on the graded threshold ranges corresponding to the protection response limits of the monitoring terminal's zone, different levels of threshold ranges are defined according to the warning level from low to high. A three-dimensional tensor structure is constructed with the anomaly parameter category, protection threshold level, and parameter coupling weight as three dimensions. The critical values ​​of each set of parameters are matched to the corresponding graded threshold ranges, filled into the corresponding tensor units, and the coupling triggering relationship between units is marked to form the warning coupling tensor. Then, the causal tracing features corresponding to the anomaly driving criteria are extracted, clarifying the causal hierarchy and driving priority order of each anomaly parameter. The causal tracing features are transformed into projection vectors and mapped into the multi-dimensional space of the warning coupling tensor. According to the causal driving order, the units of the tensor are weighted and their dimensions are filtered. The core feature dimensions directly related to the zone protection response are retained, and redundant coupling terms without actual driving effect are removed to obtain the zone response association feature basis of the corresponding zone.

[0075] It should be noted that, in this application, the parameter critical combination refers to the set of triggering conditions composed of multiple parameter critical values ​​in the anomaly driving criterion; the graded threshold interval refers to the multiple threshold ranges of the protection response limit divided according to the warning level; the warning coupling tensor refers to the multi-dimensional data structure that integrates the coupling relationship between the anomaly parameter critical conditions and the graded protection threshold; the causal tracing feature is the feature information that characterizes the causal hierarchy and driving order among the anomaly driving parameters; and the basis projection refers to the processing method of mapping causal features to the coupling tensor and extracting dimensions.

[0076] In step S3, when the anomaly driving criterion triggers the alarm condition, dynamic alarm information for the period from the occurrence of the anomaly to its resolution is generated through alarm judgment rules. The dynamic alarm information is then classified and processed to obtain the alarm flag value corresponding to each anomaly level.

[0077] In this embodiment, when the anomaly-driven criterion triggers the alarm condition, the generation of dynamic alarm information for the period from the occurrence of the anomaly to its resolution through alarm determination rules can be achieved through the following steps:

[0078] Based on the alarm conditions triggered by the anomaly driving criteria and the combination of criteria parameters, the initial state vector and temporal evolution quantity of the abnormal event are determined, and the abnormal situation evolution interval is obtained.

[0079] By using the alarm level mapping data in the abnormal situation evolution interval and alarm judgment rules, the various stages of the abnormal occurrence are hierarchically associated to generate a dynamic alarm level sequence from the occurrence of the abnormality to the time period of its resolution.

[0080] Extract the level transition nodes and duration features from the dynamic alarm level sequence, match them with the risk push constraints of the protection response limit, and generate dynamic alarm information.

[0081] In specific implementation, firstly, when the anomaly-driven criterion meets the alarm conditions, the real-time monitoring values ​​of each parameter in the current criterion parameter combination are extracted and arranged in order according to parameter category to form the initial state vector of the anomaly event. Each element of the vector corresponds to the current value and degree of exceedance of a single parameter. Subsequently, subsequent monitoring data of the criterion parameters are continuously collected at a fixed time step, and the change amplitude and rate of change of each parameter per unit time are calculated to obtain the time-series evolution quantity. Combined with the evolution statistics of similar anomalies in the mine's history, the estimated time range from the anomaly triggering time to the parameter returning to the safety threshold is defined, forming the anomaly situation evolution interval. Then, the alarm level mapping data preset in the alarm judgment rules is retrieved to clarify the alarm level classification standards corresponding to different parameter exceedance degrees and change rates. The anomaly situation evolution interval is then divided into continuous independent analysis periods at a fixed time step. The comprehensive exceedance level of the criterion parameters in each period is calculated, and the alarm level is matched with the alarm level mapping data to obtain the alarm level of the corresponding period. The alarm levels of each period are arranged in chronological order to form a dynamic alarm level sequence from the occurrence of the anomaly to its deactivation. Finally, the complete dynamic alarm level sequence is traversed to identify the time points when the alarm level changes between adjacent time periods and mark them as level jump nodes. At the same time, the continuous duration of each alarm level in the sequence is counted to extract the duration feature. The risk push constraints corresponding to the protection response limit are retrieved to clarify the alarm push frequency, push objects and content specifications under different levels and different durations. The duration feature is adapted and integrated according to the push constraints to generate dynamic alarm information.

[0082] It should be noted that in this application, alarm conditions refer to the critical rules for determining whether an abnormal event meets the requirements for triggering an alarm; alarm judgment rules refer to a set of pre-defined rules that include alarm level classification, mapping logic, and output specifications; criterion parameter combination refers to a set of multiple monitoring parameters that constitute the anomaly-driven criterion; initial state vector refers to a structured parameter of the initial risk state at the time of anomaly triggering; temporal evolution quantity is a characteristic parameter used to characterize the trend of abnormal parameters changing over time; abnormal situation evolution interval refers to the time range from the occurrence of an anomaly to its estimated resolution; alarm level mapping data refers to the benchmark data that records the correspondence between parameter states and alarm levels; hierarchical level association refers to the association mechanism that matches the anomaly evolution stage with the corresponding alarm level; dynamic alarm level sequence refers to a temporal sequence that presents the dynamic changes of the anomaly level throughout its entire cycle; level jump node refers to the key time point that identifies a sudden change in the degree of anomaly risk; duration characteristic refers to the characteristic information of the continuous duration of each alarm level; risk push constraint refers to the constraint conditions that regulate the method, object, and frequency of alarm information push; dynamic alarm information refers to the temporalized alarm content that covers the entire anomaly cycle.

[0083] It should also be noted that, in this application, alarm conditions can be set in the following manner: based on relevant industry regulations on mine safety and the protection level requirements of each underground zone, combined with historical mine disaster monitoring data and archived accident cases, trigger boundaries are set for each monitoring parameter corresponding to each type of anomaly driving criterion. First, a single parameter over-limit trigger condition is set. When the real-time value of a single monitoring parameter exceeds the protection response limit of the corresponding zone, the basic alarm trigger requirement is met. Second, a multi-parameter coupling trigger condition is set. When two or more related parameters simultaneously reach the set proportion of the protection response limit, an alarm can be triggered even if a single parameter does not reach the over-limit threshold. Finally, a parameter change rate trigger condition is set. When the change amplitude of the monitoring parameter within the set time exceeds the rate threshold, a warning alarm condition is triggered. All conditions are adjusted according to the zone protection level, and the trigger threshold is appropriately lowered for high-risk zones. The alarm judgment rules can be set in the following way: referring to the general classification standard for mine safety risk classification and control, and combining the protection response level and on-site handling capability of each underground zone, the alarm level is divided into four progressive risk levels. For each alarm level, the corresponding parameter over-limit ratio range, the requirement for the number of coupled related parameters, and the threshold for the duration of abnormal state are clearly defined as the core basis for level judgment. The timing judgment logic for level upgrading and downgrading is set. When the number of sampling cycles in which the abnormal state continuously meets the conditions of a higher level reaches the set value, the alarm level is automatically upgraded. When the parameter falls back to the lower level range and remains at the set duration, the alarm level is downgraded step by step. At the same time, the information push objects, handling response time limits, and linkage equipment ranges corresponding to different levels are set to form alarm judgment rules.

[0084] Preferably, in this embodiment, the dynamic alarm information is processed in a hierarchical manner to obtain an alarm flag value corresponding to each anomaly level, with reference to... Figure 2 As shown in the figure, this is a schematic flowchart of the process for determining the alarm flag value in some embodiments of this application. In this embodiment, the alarm flag value can be determined by the following steps:

[0085] In step S31, based on the hierarchical handling rule base of the dynamic alarm information and the partition protection response limit, the hierarchical handling coupling coefficient when the alarm level sequence matches the handling strategy is determined;

[0086] In step S32, the level transition nodes in the dynamic alarm information are strategically encoded based on the hierarchical handling coupling coefficient to generate a preliminary alarm tag sequence;

[0087] In step S33, boundary correction is performed using the preliminary alarm tag sequence and the initial state vector of the anomaly driving criterion to obtain the alarm tag value corresponding to each anomaly level.

[0088] In practice, the process begins by retrieving the hierarchical handling rule library corresponding to the zone protection response limits. This rule library pre-defines standardized handling strategy entries for different alarm levels and different zone protection scenarios. The dynamic alarm level sequence contained in the dynamic alarm information is broken down into consecutive time periods of the same level, and each segment is matched against the corresponding handling strategy in the rule library. A quantitative score is calculated from three dimensions: level matching degree, zone protection level adaptability, and anomaly duration fit. The matching degree value between each level sequence and the corresponding handling strategy is obtained by weighting and summing according to preset weights; this is the hierarchical handling coupling coefficient, which ranges from 0 to 1. Next, all level transition nodes are extracted from the dynamic alarm information. The direction of level increase / decrease and the temporal position of each node are determined. Based on the hierarchical handling coupling coefficient obtained earlier, the preset coding rules of the corresponding handling strategy are matched, and a corresponding standardized coding value is assigned to each level transition node. The coding value corresponds one-to-one with the alarm level and handling priority. The coding values ​​of all nodes are arranged sequentially according to time, forming a preliminary alarm marker sequence from anomaly occurrence to resolution. Finally, the initial state vector corresponding to the anomaly driving criterion is retrieved, and the two core boundary parameters contained in the vector, namely the initial over-limit ratio and the number of associated parameters, are extracted. The initial over-limit ratio and the number of associated parameters are used as the amplitude correction weight and the dimension correction weight, respectively, to perform weighted correction on the encoding values ​​of the start and end segments of the preliminary alarm labeling sequence, thereby eliminating the labeling error caused by the initial state deviation. After the correction is completed, the encoding values ​​in the sequence are classified and aggregated according to the anomaly level, and the comprehensive label value corresponding to each anomaly level is calculated to obtain the alarm label value corresponding to each anomaly level.

[0089] It should be noted that, in this application, graded handling refers to the processing flow of matching corresponding handling strategies according to the abnormal alarm level; zoned protection response limit refers to the set of safety critical thresholds set for each zone downhole; graded handling rule base refers to the set of rules that pre-set the correspondence between alarm levels and handling strategies; handling strategy refers to the standardized handling scheme corresponding to different alarm levels; graded handling coupling coefficient is a numerical indicator that quantifies the degree of matching between alarm level sequence and handling strategy; strategy coding refers to the processing method of converting handling strategy into standardized numerical coding; preliminary alarm mark sequence refers to the set of alarm node codes arranged in chronological order; initial state vector is a structured parameter that represents the initial risk state at the time of abnormal triggering; abnormal level refers to the abnormal event level category divided according to the degree of risk; alarm mark value refers to the quantitative mark value corresponding to each abnormal level.

[0090] In addition, in this embodiment, the generation of a preliminary alarm marker sequence by strategy encoding the level transition nodes in the dynamic alarm information based on the hierarchical handling coupling coefficient can be achieved by the following steps:

[0091] The node handling encoding vector is determined based on the hierarchical handling coupling coefficient and the abnormal time period corresponding to the level jump node in the dynamic alarm information.

[0092] The joint coding features of the handling strategy and response parameters are extracted from the temporal evolution of the node handling coding vector and the anomaly driving criterion to generate a preliminary alarm tag sequence.

[0093] In specific implementation, firstly, all level transition nodes are extracted one by one from the dynamic alarm information. The abnormal time period corresponding to each node is located, and the alarm level, level transition direction, and duration of the time period in the entire abnormal cycle are recorded. The standardized handling code table built into the hierarchical handling rule base is retrieved, and a unique corresponding numerical code is assigned to the matched handling strategy. The hierarchical handling coupling coefficient is used as the weight dimension and combined with the handling strategy code value, level transition direction identifier, and abnormal time period time sequence weight in a fixed dimension order to form the node handling code vector of the corresponding node. Then, the node handling code vector corresponding to each level transition node is timestamped with the time sequence evolution of the abnormal driving criterion within the same time period. The parameter change rate and the over-limit increase are extracted from the time sequence evolution. Using the known data processing method of feature dimension splicing, the handling code dimension and the response parameter dimension are merged in a fixed order to extract the joint coding feature that simultaneously covers the handling strategy attribute and the abnormal evolution attribute. According to the chronological order of the abnormal event development, the joint coding features of all nodes are arranged sequentially to generate a preliminary alarm mark sequence.

[0094] It should be noted that, in this application, the abnormal time period refers to the continuous time interval covered by the level jump node; the node disposal coding vector refers to the structured vector that integrates the multi-dimensional disposal attributes of the jump node; the response parameter refers to the feature parameter that reflects the degree of evolution of the abnormal parameter; and the joint coding feature refers to the composite coding unit that integrates the disposal attribute and the evolution attribute.

[0095] In step S4, the risk warning level within the mine is determined based on the alarm status index and all alarm marker values, and a linkage alarm is triggered according to the risk warning level.

[0096] In this embodiment, determining the risk warning level within the mine based on the alarm status index and all alarm marker values ​​can be achieved through the following steps:

[0097] The regional risk intensity characteristics are determined based on the alarm status index and the regional distribution vector of all alarm marker values.

[0098] By mapping the regional risk intensity characteristics and the mine risk benchmark threshold, a corrected risk warning level is generated.

[0099] The risk warning level within the mine is obtained by making a transition judgment on the predicted trend of the modified risk warning level and alarm status index.

[0100] In practice, firstly, all alarm marker values ​​are collected according to the spatial division of each zone in the mine. The weighted sum of all alarm marker values ​​in each zone is calculated and arranged in the fixed number order of the zones to form a regional distribution vector consistent with the number dimension of the zones. The alarm situation index is used as the global situation correction coefficient and multiplied element-wise with the regional distribution vector. At the same time, the inherent weights of personnel density and equipment importance pre-set for each zone are superimposed to calculate the quantified risk intensity value of each zone. The quantified risk intensity value of each zone is used as the regional risk intensity feature. Then, the pre-set mine risk benchmark threshold is retrieved. This mine risk benchmark threshold is defined according to the mine safety industry regulations and historical risk classification statistics. It is divided into multiple continuous threshold intervals according to risk from low to high. The risk intensity values ​​of each zone included in the regional risk intensity feature are compared with the benchmark threshold interval one by one to obtain the initial warning level of the corresponding interval. Then, the initial level is adapted and corrected by combining the zone protection level weight. After summarizing the comprehensive risk level of the entire mine, the corrected risk warning level is generated. Finally, based on the historical time series data of the alarm situation index, the known prediction method of sliding window linear extrapolation is adopted. A historical data window of a set length is selected, and the direction and rate of change of the index within a set future time period are calculated to obtain the predicted trend of the alarm situation index. The rate of change of the predicted trend is compared with the preset level jump threshold. If the trend shows that the risk continues to rise and the jump condition is met, the risk warning level is adjusted and corrected in advance. Otherwise, the original level is maintained, and the risk warning level in the mine is finally obtained.

[0101] It should be noted that, in this application, the regional distribution vector refers to the structured vector carrying the spatial distribution of alarm marker values ​​in each zone; the regional risk intensity characteristic refers to the feature set of the degree of risk intensity in each zone underground; the mine risk benchmark threshold refers to the standardized critical value for dividing risk level intervals; the level mapping is the processing logic for converting quantified risk values ​​into graded early warning levels; the corrected risk early warning level refers to the preliminary early warning level after being corrected by zone attribute adaptation; the alarm situation index prediction trend refers to the trend information reflecting the future evolution of the alarm situation index; the transition judgment refers to the judgment logic for determining whether the early warning level needs to be adjusted in advance; and the risk early warning level refers to the overall risk alarm classification result of the mine.

[0102] In this embodiment, the linkage alarm based on the risk warning level can be implemented using the following steps:

[0103] The alarm strategy coupling vector is determined based on the aforementioned risk warning level and alarm linkage response conditions;

[0104] The alarm strategy coupling vector is used to fuse the collaborative triggering parameters of abnormal events in the mine, generate a linkage alarm command, and execute it.

[0105] In practice, the process begins by retrieving the alarm linkage response conditions pre-configured according to the mine safety emergency response procedures. These conditions are categorized by risk warning level and pre-set with the scope of linkage equipment, response priority, activation sequence constraints, and handling authority. The currently determined risk warning level is matched one by one with the categorized items in the response conditions. The core parameters of the successfully matched linkage strategy are extracted and quantified from three dimensions: level fit, area coverage, and response timeliness. The quantified values ​​are then combined with the strategy identifier and linkage range code in a fixed-dimensional order to form an alarm strategy coupling vector. Each dimension of the alarm strategy coupling vector has a clear mapping rule. Then, the collaborative triggering parameters corresponding to the current mine abnormal event are extracted, including the partition code of the abnormality, the address of the on-site linkage equipment, the duration of the trigger and the conditions for release. The weight values ​​of each dimension in the alarm strategy coupling vector are used as adaptation coefficients to screen and correct the collaborative triggering parameters item by item, eliminate triggering items that do not meet the current warning level, generate linkage alarm commands that include equipment action type, execution sequence and feedback requirements, and send them to the corresponding on-site equipment and control terminals through the underground industrial Ethernet ring network, and simultaneously verify the execution status of the command feedback.

[0106] It should be noted that, in this application, alarm linkage response conditions refer to a set of linkage response rules that are pre-defined at different levels; alarm strategy coupling vector refers to a structured vector that integrates early warning levels and linkage strategy adaptation attributes; collaborative triggering parameters refer to a set of parameters for abnormal on-site linkage triggering conditions; and linkage alarm instructions refer to executable instructions sent to on-site equipment and control terminals.

[0107] In this embodiment, reference Figure 3 As shown in the diagram, this is a schematic diagram of the mine linkage alarm principle. The diagram illustrates the signal flow path and logical judgment mechanism of the mine's underground linkage alarm process. The left side of the diagram is the risk warning level input unit, which carries the overall mine risk warning level signal obtained from the previous analysis. It is transmitted to the core logic judgment unit via the input module. The judgment unit uses the pre-set alarm linkage response conditions as the matching benchmark, and verifies and matches the input risk warning level with the hierarchical linkage rules one by one. It completes the construction of the alarm strategy coupling vector and the adaptation and screening of the collaborative triggering parameters, and generates the linkage control logic of the corresponding level. The matched linkage control signal is transmitted to the field linkage execution terminal via the output module, driving the alarm broadcast, emergency communication and control equipment of the corresponding underground zone to execute linkage actions according to the preset time sequence. The logic judgment node at the bottom of the diagram forms a closed-loop feedback link, receiving the status feedback from the execution terminal and feeding back the execution information to the input end. This is used to support the dynamic adjustment of the warning level and the optimization and correction of the linkage strategy, forming a complete and controllable closed-loop linkage alarm control mechanism.

[0108] It can be seen from this that in the present application, linked alarm can be performed according to the risk warning level; wherein, by determining the alarm situation index, a comprehensive risk quantitative characterization result of the entire area and continuous space of the subarea where the mine abnormal event is located can be obtained, thereby breaking the limitation that traditional single-point monitoring data can only characterize the risk of local points. Relying on multi-dimensional protection limit constraints and spatial risk attenuation correction logic, the real risk diffusion range and intensity gradient in the underground mine can be restored, and the horizontal quantitative benchmarking of risk levels between different underground subareas can be realized, which provides standardized global risk reference for unified research and judgment of global alarm levels and overall allocation of cross-regional emergency resources, ensures the objectivity and balance of the whole mine risk assessment from the spatial dimension, and enables the coverage and response priority of linked disposal to accurately match the actual distribution characteristics of risks; by determining the alarm marking values, standardized quantitative identifiers of risk priorities corresponding to different abnormal event levels can be obtained, thereby converting unstructured alarm evolution information under dynamic time series evolution into standardized numerical features that can be weighted and aggregated, realizing vertical quantitative description of the risk severity in the whole cycle of occurrence, development and regression of abnormal events, and accurately anchoring the adaptation relationship of differentiated disposal strategies corresponding to different abnormal levels. It provides local time series risk evidence for the correction and optimization of the global risk warning level, so that the global linked disposal can match and push objects, start time sequences and equipment linkage ranges according to the severity of abnormal evolution, and improve the refined disposal logic of full-link cooperative alarm from the time series classification dimension.

[0109] In summary, the technical solution adopted in the present application can construct a full-link cooperative alarm mechanism for global linked disposal of mine abnormal events, so as to improve the accuracy of safety risk prevention and control in underground mines.

[0110] In the second embodiment, the present application provides a mine intelligent alarm system based on multi-dimensional perceptual information, with reference to Figure 4 shown in the figure, this figure is a module structure diagram of the mine intelligent alarm system based on multi-dimensional perceptual information according to the embodiment of the present application, the alarm system comprises:

[0111] an acquisition module 100, configured to acquire safety status information at monitoring terminals arranged in each subarea of an underground mine;

[0112] a judgment module 200, configured to extract an abnormal driving criterion corresponding to a mine abnormal event from all mine safety status information, perform correlation judgment on the abnormal driving criterion and a protection response limit of the subarea where the monitoring terminal is located, and obtain an alarm situation index of the area where the mine abnormal event is located;

[0113] a processing module 300, configured to, when the abnormal driving criterion triggers an alarm condition, generate dynamic alarm information from the occurrence to the release period of the abnormality through an alarm judgment rule, perform graded disposal on the dynamic alarm information, and obtain an alarm marking value corresponding to each abnormality level;

[0114] The execution module 400 is used to determine the risk warning level in the mine based on the alarm status index and all alarm marker values, and to trigger a linkage alarm based on the risk warning level.

[0115] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0116] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, including read-only memory (ROM), random access memory (RAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), one-time programmable read-only memory (OTPROM), electrically-Erasable Programmable Read-Only Memory (EEPROM), compactdisc read-only memory (CD-ROM) or other optical disc storage, disk storage, magnetic tape storage, or any other computer-readable medium capable of carrying or storing data.

[0117] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

Claims

1. A mine intelligent alarm method based on multi-dimensional sensing information, characterized in that, The alarm method includes the following steps: Collect safety status information from monitoring terminals deployed in various underground sections of the mine; Extract the anomaly driving criteria corresponding to the abnormal events in the mine from all mine safety status information, and associate the anomaly driving criteria with the protection response limit of the zone where the monitoring terminal is located to obtain the alarm status index of the area where the abnormal events in the mine are located. When the anomaly driving criterion triggers the alarm condition, dynamic alarm information is generated from the occurrence of the anomaly to the period of its resolution through the alarm judgment rule. The dynamic alarm information is then classified and processed to obtain the alarm flag value corresponding to each anomaly level. The risk warning level within the mine is determined based on the alarm status index and all alarm marker values, and a linkage alarm is triggered according to the risk warning level.

2. The intelligent mine alarm method based on multi-dimensional sensing information as described in claim 1, characterized in that, The specific criteria for extracting anomaly-driven judgments corresponding to abnormal events in mines from all mine safety status information include: Based on all mine safety status information, mutual information causal screening of multi-source monitoring variables is performed to obtain abnormal situation characteristics; The abnormal driving factors of the abnormal events are determined by the abnormal situation characteristics and the time sequence information of the abnormal events. The anomaly driving criteria corresponding to the abnormal events in the mine are determined based on the aforementioned anomaly driving factors and the composite disaster change attributes of the mine.

3. The intelligent mine alarm method based on multi-dimensional sensing information as described in claim 1, characterized in that, The alarm status index of the area where the mine anomaly event is located is obtained by associating the anomaly driving criterion with the protection response limit of the zone where the monitoring terminal is located. Specifically, this includes: Based on the anomaly driving criterion and the protection response limit of the partition where the monitoring terminal is located, determine the partition response association feature base; The correlation situation mapping is performed on the partition response correlation feature base and the real-time sensing data of the multi-source monitoring terminals within the partition to obtain the alarm situation hot zone gradient of the abnormal event. The alarm status hot zone gradient and the dynamic margin threshold of the protection response limit are compensated for by risk attenuation to generate an alarm status index for the area where the mine abnormal event is located.

4. The intelligent mine alarm method based on multi-dimensional sensing information as described in claim 3, characterized in that, Based on the aforementioned anomaly-driven criterion and the protection response limit of the partition where the monitoring terminal is located, the partition response association feature base is determined to specifically include: The early warning coupling tensor is determined based on the critical combination of parameters in the anomaly driving criterion and the graded threshold range of the protection response limit; By performing basis projection on the causal origin features of the warning coupling tensor and the anomaly driving criterion, a partitioned response association feature basis is generated.

5. The intelligent mine alarm method based on multi-dimensional sensing information as described in claim 1, characterized in that, When the anomaly-driven criterion triggers the alarm condition, the dynamic alarm information generated according to the alarm determination rules for the period from the occurrence of the anomaly to its resolution specifically includes: Based on the alarm conditions triggered by the anomaly driving criteria and the combination of criteria parameters, the initial state vector and temporal evolution quantity of the abnormal event are determined, and the abnormal situation evolution interval is obtained. By using the alarm level mapping data in the abnormal situation evolution interval and alarm judgment rules, the various stages of the abnormal occurrence are hierarchically associated to generate a dynamic alarm level sequence from the occurrence of the abnormality to the time period of its resolution. Extract the level transition nodes and duration features from the dynamic alarm level sequence, match them with the risk push constraints of the protection response limit, and generate dynamic alarm information.

6. The intelligent mine alarm method based on multi-dimensional sensing information as described in claim 1, characterized in that, The dynamic alarm information is processed in a hierarchical manner to obtain the alarm flag value corresponding to each anomaly level, specifically including: Based on the hierarchical handling rule base of the dynamic alarm information and the zone protection response limit, determine the hierarchical handling coupling coefficient when the alarm level sequence matches the handling strategy; Based on the hierarchical processing coupling coefficient, the hierarchical jump nodes in the dynamic alarm information are strategically encoded to generate a preliminary alarm marker sequence; By performing boundary correction using the preliminary alarm tag sequence and the initial state vector of the anomaly driving criterion, the alarm tag value corresponding to each anomaly level is obtained.

7. A mine intelligent alarm method based on multi-dimensional sensing information as described in claim 6, characterized in that, Based on the hierarchical handling coupling coefficient, strategy encoding is performed on the level transition nodes in the dynamic alarm information to generate a preliminary alarm marker sequence, specifically including: The node handling encoding vector is determined based on the hierarchical handling coupling coefficient and the abnormal time period corresponding to the level jump node in the dynamic alarm information. The joint coding features of the handling strategy and response parameters are extracted from the temporal evolution of the node handling coding vector and the anomaly driving criterion to generate a preliminary alarm tag sequence.

8. The intelligent mine alarm method based on multi-dimensional sensing information as described in claim 1, characterized in that, Determining the risk warning level within the mine based on the aforementioned alarm status index and all alarm marker values ​​specifically includes: The regional risk intensity characteristics are determined based on the alarm situation index and the regional distribution vector of all alarm marker values. By mapping the regional risk intensity characteristics and the mine risk benchmark threshold, a corrected risk warning level is generated. The risk warning level within the mine is obtained by making a transition judgment on the predicted trend of the modified risk warning level and alarm status index.

9. A mine intelligent alarm method based on multi-dimensional sensing information as described in claim 1, characterized in that, The triggering of a linked alarm based on the aforementioned risk warning level specifically includes: The alarm strategy coupling vector is determined based on the aforementioned risk warning level and alarm linkage response conditions; The alarm strategy coupling vector is used to fuse the collaborative triggering parameters of abnormal events in the mine, generate a linkage alarm command, and execute it.

10. A mine intelligent alarm system based on multi-dimensional sensing information, used to execute the mine intelligent alarm method based on multi-dimensional sensing information as described in any one of claims 1 to 9, characterized in that, The alarm system includes: The data acquisition module is used to collect safety status information from monitoring terminals deployed in various sections of the underground mine. The judgment module is used to extract the anomaly driving criteria corresponding to the abnormal events in the mine from all mine safety status information, and to associate the anomaly driving criteria with the protection response limit of the zone where the monitoring terminal is located to obtain the alarm status index of the area where the abnormal event in the mine is located. The processing module is used to generate dynamic alarm information from the occurrence to the resolution period through alarm judgment rules when the abnormality driving criterion triggers the alarm condition, and to perform hierarchical processing on the dynamic alarm information to obtain the alarm flag value corresponding to each abnormality level. The execution module is used to determine the risk warning level in the mine based on the alarm status index and all alarm marker values, and to trigger a linkage alarm based on the risk warning level.

Images