Intelligent combustible gas alarm system and alarm method

Abstract

This invention belongs to the field of gas safety monitoring technology and is used to solve the problem that existing technologies cannot integrate multi-dimensional information and adaptively adjust monitoring strategies. Specifically, it is a combustible gas intelligent alarm system and alarm method applied to wireless gas detectors, including: acquiring the current concentration data, real-time airflow parameters and scene information of the detector's location, and the status and spatial information of at least one neighboring detector; correcting the current concentration data to obtain equivalent concentration data; correcting the basic risk level to obtain the corrected risk level; determining whether the concentration data meets the preset transient triggering conditions; and dynamically adjusting the detector's data reporting cycle and reporting content according to the current risk level. This invention solves the technical contradiction of fixed thresholds being unable to simultaneously achieve high sensitivity and low false alarm rate in dynamic operating environments by introducing scene information to adaptively adjust the monitoring threshold.

Description

Technical Field

[0001] This invention belongs to the field of gas safety monitoring technology, specifically a combustible gas intelligent alarm system and alarm method. Background Technology

[0002] In industrial production and residential life, combustible gas leak monitoring is a key link in ensuring the safety of life and property. With the popularization of Internet of Things (IoT) technology, wireless gas detectors are widely deployed in various places. By periodically collecting concentration data and reporting it to the monitoring center, real-time perception of leak risks is achieved. Existing technologies usually use a fixed threshold comparison method to trigger alarms. That is, when the concentration value collected by the detector exceeds the preset alarm threshold, it is judged as a leak event and an alarm is issued. This judgment logic based on a single threshold can work effectively in a steady-state environment. When a leak occurs slowly, the concentration gradually increases and eventually triggers an alarm.

[0003] However, the actual operating environment in industrial settings is far from steady-state. First, the operational status changes frequently—the probability of leakage and the allowable response time vary significantly in different scenarios such as hot work, loading and unloading operations, personnel inspections, and unattended operation. Using a uniform fixed threshold cannot balance the contradiction between high sensitivity and low false alarm rate. Second, the sensors themselves have performance drift, and their sensitivity may decrease after long-term operation, causing alarms that should be triggered to be delayed or even missed. More importantly, gas leaks often have sudden and diffuse characteristics. Isolated judgment by a single-point detector cannot capture the spatial propagation trend of the leak. When a rapid jet leak occurs in a certain area, the concentration may soar from zero to the lower explosive limit within seconds, while the detector is still in long-cycle reporting mode, resulting in the delayed transmission of critical alarm data.

[0004] Furthermore, industrial sites commonly experience airflow disturbances caused by fan operation, ventilation system startup and shutdown, or ambient winds. These airflows significantly impact the transport of gases in space. When the detector is located downwind of the leak source, the airflow can blow away the leaking gas, resulting in a lower detected concentration than the actual leak concentration, potentially leading to false alarms. Conversely, when the detector is upwind of the leak source, the airflow may transport higher concentrations of gas from other areas to the detector's location, causing a higher detected concentration than the actual background concentration, potentially triggering false alarms. This airflow convection effect is coupled with the leak itself, making it difficult for single-point concentration measurements to accurately reflect the true leak risk, further exacerbating the risk of misjudgment by the monitoring system in complex environments.

[0005] The combined effects of the above problems manifest in real-world engineering scenarios as follows: during high-risk periods such as hot work operations, insufficient sensitivity of threshold settings leads to the failure to detect minute leaks in a timely manner; during unattended periods, sensor drift causes frequent false alarms, resulting in alarm fatigue; and when multiple areas experience anomalies simultaneously, channel congestion further exacerbates alarm delays. The root cause of these problems lies in the fact that existing technologies employ a linear processing logic of "single point - single threshold - fixed period," lacking the ability to collaboratively perceive the work scenario, sensor status, spatial correlation, and sudden characteristics. This makes it difficult for the system to accurately characterize risk levels and dynamically adapt communication resources under non-steady-state conditions. Therefore, how to construct a combustible gas leak early warning method that can integrate multi-dimensional information and adaptively adjust monitoring strategies has become a pressing technical problem to be solved in this field. Summary of the Invention

[0006] The purpose of this invention is to provide an intelligent alarm system and method for combustible gases, which solves the problem that existing technologies cannot integrate multi-dimensional information and adaptively adjust monitoring strategies.

[0007] The technical problem to be solved by this invention is: how to provide a combustible gas intelligent alarm system and alarm method that can integrate multi-dimensional information and adaptively adjust the monitoring strategy.

[0008] The objective of this invention can be achieved through the following technical solutions:

[0009] On one hand, the present invention provides an intelligent alarm method for combustible gas, applied to a wireless gas detector, comprising:

[0010] Acquire the current concentration data, real-time airflow parameters and scene information of the detector's location, and the status and spatial information of at least one neighboring detector; based on the scene information, determine the rate of change threshold and concentration threshold that match the current scene; and correct the current concentration data based on the real-time airflow parameters and the spatial information of the neighboring detectors to obtain equivalent concentration data.

[0011] A basic risk level is generated based on the comparison results between equivalent concentration data and concentration thresholds, as well as the comparison results between the rate of change of equivalent concentration data and the rate of change thresholds.

[0012] Based on the state information of neighbor detectors, the basic risk level is corrected according to the preset spatial coordination rules to obtain the corrected risk level.

[0013] Determine whether the equivalent concentration data meets the preset transient triggering condition. The transient triggering condition is that the concentration increase in two consecutive sampling intervals exceeds the preset transient change threshold.

[0014] If the transient triggering condition is met, the detector's risk level will be forcibly set to the highest level, and a hold timer will be started; otherwise, the corrected risk level will be used as the current risk level.

[0015] Based on the current risk level, the data reporting cycle and content of the detectors are dynamically adjusted, and data packets containing the current risk level and transient trigger flags are sent to the network coordinator.

[0016] On the other hand, the present invention provides a combustible gas intelligent alarm system, comprising:

[0017] Basic information acquisition module: acquires the current concentration data, real-time airflow parameters and scene information of the detector's location area, and the status and spatial information of at least one neighboring detector; based on the scene information, determines the rate of change threshold and concentration threshold that match the current scene; and corrects the current concentration data based on the real-time airflow parameters and the spatial information of neighboring detectors to obtain equivalent concentration data.

[0018] Basic Risk Analysis Module: Generates a basic risk level based on the comparison results between equivalent concentration data and concentration thresholds, as well as the comparison results between the rate of change of equivalent concentration data and the rate of change thresholds.

[0019] Risk correction processing module: Based on the status information of neighbor detectors, the basic risk level is corrected according to the preset spatial coordination rules to obtain the corrected risk level;

[0020] Transient trigger decision module: Determines whether the equivalent concentration data meets the preset transient trigger conditions. The transient trigger conditions are that the concentration increase in two consecutive sampling intervals exceeds the preset transient change threshold.

[0021] If the transient triggering condition is met, the detector's risk level will be forcibly set to the highest level, and a hold timer will be started; otherwise, the corrected risk level will be used as the current risk level.

[0022] Data reporting module: Based on the current risk level, dynamically adjust the data reporting cycle and content of the detector, and send a data packet containing the current risk level and transient trigger flag to the network coordinator.

[0023] The present invention has the following beneficial effects:

[0024] 1. By introducing scenario information to adaptively adjust the monitoring threshold, the technical contradiction of fixed thresholds being unable to balance high sensitivity and low false alarm rate in dynamic operating environments is resolved. In high-risk scenarios such as hot work and loading / unloading operations, the system automatically adopts stricter change rate and concentration thresholds, enabling timely detection of even minor leaks. In low-risk scenarios such as unattended operation, the thresholds are appropriately relaxed to effectively suppress frequent false alarms caused by environmental fluctuations or sensor drift. This dynamic matching mechanism between scenario and threshold allows the detector to maintain optimal monitoring sensitivity under different operating conditions, avoiding the problems of missed or false alarms caused by improper threshold settings in traditional methods.

[0025] 2. By introducing an airflow disturbance factor to physically correct the concentration data, the problem of gas transport interference caused by fan operation, ventilation changes, or environmental airflow disturbances in industrial sites is solved. When the wind direction is consistent with the concentration gradient direction, it indicates that the airflow is transporting higher concentration gas from the outside to this detector, causing the observed concentration to be too high. The system uses downward correction to restore the true background concentration. When the wind direction is opposite to the concentration gradient direction, it indicates that the airflow is transporting gas from this detector to the outside, causing the observed concentration to be too low. The system uses upward correction to compensate for the diluted part. This differential correction mechanism based on the gradient-wind direction relationship enables the corrected equivalent concentration data to truly reflect the concentration level under conditions without airflow interference, effectively avoiding misjudgment or omission caused by airflow convection.

[0026] 3. By constructing a spatial collaborative correction mechanism, the risk status of neighboring nodes is integrated into the local risk level determination process, solving the technical defect that single-point isolated judgment cannot identify the leakage spread trend; when multiple surrounding nodes show a consistent increase in risk or there is a transient triggering event, the local node will be linked to raise its risk level even if the concentration has not exceeded the standard, forming a regional linkage early warning propagation; this risk perception method based on group consensus enables the system to enter the alert state in advance before the leaked gas actually arrives, effectively making up for the spatial perception blind spot caused by the spacing of sensor deployment points, and gaining valuable lead time for emergency response;

[0027] 4. By setting up a transient trigger detection path independent of periodic reporting, the technical challenge of capturing sudden high-speed leaks in real time during conventional sampling cycles is solved; based on the criterion that the concentration increase exceeds a fixed threshold for two consecutive sampling intervals, the system can accurately identify continuous drastic changes and effectively distinguish them from single random noise; once a transient event is triggered, the detector immediately forces the highest risk level and starts a hold timer to ensure a stable and reliable alarm state; at the same time, the transient information is transmitted to surrounding nodes through a broadcast mechanism to form a rapid response emergency link;

[0028] 5. By establishing a dynamic mapping relationship between risk levels and reporting strategies, optimized adaptation of communication resources to monitoring needs is achieved. At low risk, only compressed summary data is uploaded, maximizing bandwidth savings. At medium to high risk, data granularity and reporting frequency are gradually increased to provide real-time decision-making support for the monitoring center. At the highest risk, channel resources are preempted with minimal delay to ensure critical information is prioritized. This tiered reporting strategy, in industrial wireless scenarios with limited network resources, ensures both long-term operational stability for daily monitoring and reliable transmission of critical data at critical moments, significantly improving the overall resource utilization efficiency of the system. Attached Figure Description

[0029] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0030] Figure 1 This is a flowchart of the overall method of Embodiment 1 of the present invention;

[0031] Figure 2 This is a flowchart illustrating the basic risk level correction process in Embodiment 1 of the present invention;

[0032] Figure 3 This is a system block diagram of Embodiment 2 of the present invention. Detailed Implementation

[0033] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0034] In the field of industrial safety, accurate detection and timely alarm of combustible gas leaks are a key line of defense against major accidents such as explosions and fires. Current mainstream gas monitoring systems generally employ a fixed threshold comparison logic: the detector collects gas concentration data from the environment at a constant frequency, and triggers an alarm signal when the concentration exceeds a preset alarm threshold. This single-point, single-threshold, fixed-cycle monitoring mode works effectively under ideal steady-state conditions. However, when a leak occurs slowly, the concentration gradually accumulates and eventually exceeds the threshold, triggering an alarm.

[0035] However, the real industrial environment is far from an ideal steady state, but is full of dynamic changes and uncertainties. Frequent switching of operating states—hot work, loading and unloading operations, personnel inspections, unmanned operation, and other different scenarios have fundamentally different tolerances for leakage risks. Using a uniform fixed threshold cannot solve the fundamental contradiction between high sensitivity and low false alarm rate. Sensors themselves have performance drift, and their sensitivity will slowly decrease after long-term operation, causing alarms that should be triggered to be delayed or even missed. More importantly, gas leaks are sudden and spatially diffuse, and the isolated judgment of a single detector cannot capture the spatial propagation trend of the leak. When a high-speed jet leak occurs in a certain area, the concentration may soar from zero to the lower explosive limit in a few seconds, while the detector is still in long-cycle reporting mode, causing critical alarm data to be delayed in transmission and missing the best emergency response window.

[0036] The combined effect of the above factors manifests in practical engineering scenarios as follows: during high-risk periods such as hot work operations, insufficiently sensitive threshold settings prevent the timely detection of minor leaks; during unattended periods, frequent false alarms caused by sensor drift lead to alarm fatigue among personnel; when anomalies occur simultaneously in multiple areas, channel congestion further exacerbates the delay of critical alarm information; more seriously, when obvious signs of leakage have appeared in surrounding areas, individual nodes may remain in a low-risk state due to their own concentration not reaching the threshold, failing to provide early warning and allowing the leakage to continue to spread undetected. For example... In the actual operation of a certain petrochemical park, there was an incident where a valve leaked slightly during loading and unloading operations. The concentration of the gas cloud increased slightly on multiple detectors in the vicinity, but none of them reached the single alarm threshold, and the system judged it to be risk-free. It was not until the leak intensified that an alarm was triggered at a certain node. At this time, the gas cloud had spread to the adjacent hot work area, which almost caused a major accident. Another example is when a pipeline suddenly ruptured, the concentration soared from 0% LEL to 15% LEL in 3 seconds. However, because the detectors used a 60-second reporting cycle, the alarm information was delayed by 57 seconds before reaching the monitoring center, missing the golden time for remote valve shut-off.

[0037] If the above problems are not addressed, the gas monitoring system will remain in a sub-healthy operating state that appears normal but is actually dangerous: the disconnect between scenarios and risks results in insufficient sensitivity during high-risk periods and frequent false alarms during low-risk periods; the fragmentation of spatial information makes it impossible to identify the leakage and spread trend in advance, and the blind spots of single-point judgment allow hidden dangers to accumulate continuously; the mismatch between transient changes and periodic reporting makes it impossible to capture sudden leaks in real time, and alarm delays become the norm; as a result, the system's monitoring capabilities will fail at the critical moments when they are most needed, failing to provide reliable decision-making basis for emergency response, and ultimately causing preventable accidents to get out of control due to delayed perception.

[0038] Example 1: As Figure 1-2 As shown, a smart alarm method for combustible gas, applied to a wireless gas detector, includes:

[0039] Step S1: Obtain the current concentration data, scene information of the area where the detector is located, and status information of at least one neighboring detector; the concentration data refers to the combustible gas concentration value obtained by the detector after temperature and humidity compensation and filtering, and the unit is %LEL;

[0040] Scene information includes at least one of the following: hot work sign, loading and unloading sign, personnel inspection sign, and no-man's-land sign;

[0041] In the specific implementation of this embodiment, the data acquisition and preliminary processing steps are performed first. The detector acquires the original sensor signal at a fixed sampling frequency of 1Hz. This signal is compensated and corrected by data from the built-in temperature and humidity sensors. In addition, the detector has a built-in sensor self-diagnostic function to continuously monitor the validity of the original signal. If the concentration data does not change within 10 consecutive sampling periods (the change is less than 0.1% LEL), or a negative value appears ( If the value is <0 and the duration exceeds 5 seconds, the sensor is considered to have malfunctioned. The current risk level is immediately forcibly set to the highest level (L=3), and a fault alarm data packet is sent to the network coordinator, while the local fault indicator flashes. Once the sensor returns to normal and data from three consecutive sampling cycles is valid, the fault state is automatically cleared and normal monitoring resumes, eliminating the influence of environmental factors on the gas-sensitive element. The calibration concentration value at the current time t is obtained and denoted as [value missing]. The unit is %LEL. To suppress the inherent high-frequency noise of the sensor and occasional transient spike interference, the detector performs median filtering on three consecutive sampled values: Let , , Given the original calibration concentrations at the previous two time points, the previous time point, and the current time point, the final concentration data used at the current time point is the median value among the three, i.e.: The filtered concentration data C(t) serves as the basic input for calculating the equivalent concentration data; and is stored in a local circular buffer, retaining at least the data from the most recent 10 sampling points for subsequent rate of change calculation and transient detection.

[0042] Meanwhile, the detector connects in real time to the factory information system (such as Manufacturing Execution System, personnel positioning system, and equipment management system) via industrial Ethernet or a dedicated wireless interface to obtain scene information of its current location. This scene information is presented in the form of Boolean flags, specifically including: hot work operation flags. A value of 1 indicates that hot work (such as welding or cutting) is currently in progress; otherwise, it is 0. Loading / unloading operation flag. A value of 1 indicates that a combustible gas loading / unloading operation is in progress (such as tanker filling or tank emptying); personnel inspection sign. A value of 1 indicates that inspection personnel have entered the area; no-person time period indicator. A value of 1 indicates that the system is currently in a preset unattended period (e.g., from 10:00 PM to 6:00 AM the next day); otherwise, it is 0. These flags are updated in real time by the factory's production scheduling system or personnel positioning system; and are pushed to the detectors through periodic broadcasts or event triggers to ensure the real-time nature and accuracy of the information.

[0043] Real-time airflow parameters are acquired, and the concentration data C(t) at the current moment is corrected based on the real-time airflow parameters and neighbor state information to obtain equivalent concentration data. .

[0044] First, the detector obtains the real-time wind speed v(t) and wind direction unit vector by using its built-in wind speed and direction sensors or by receiving data broadcast from regional weather stations via a wireless network. The wind direction unit vector points in the direction of airflow, that is, from upwind to downwind.

[0045] Secondly, extract the concentration data of all valid neighbors from the locally maintained neighbor status table. (If the neighbor's reporting cycle is not synchronized with the local one, the most recently reported concentration value will be used.) The spatial location information of the neighbor relative to this detector is also included. Location information can be obtained through coordinate calibration during pre-deployment or estimated in real-time using wireless ranging technology. The unit direction vector from this detector to the i-th neighbor detector is calculated. and the spatial distance between the two .

[0046] Based on the above data, calculate the concentration gradient vector at the current moment. This vector represents the direction and rate of change of gas concentration in space. The gradient approximation of discrete nodes is performed using the inverse distance weighted method, calculated as follows: ;

[0047] Where N is the total number of effective neighbors, and C(t) is the concentration data of this detector at the current moment. For the concentration data of the i-th neighbor, The distance between the two, It is a unit direction vector. The weighted average is the inverse distance weight, defined as follows: This weight ensures that the closer the neighbor is to the detector, the greater its contribution to the gradient calculation, and the sum of all weights is 1.

[0048] Next, based on the wind speed v(t) and the unit vector of wind direction... and concentration gradient vector The airflow disturbance factor B(t) is calculated, which physically represents the change in concentration at the detector's location caused by airflow convection per unit time, expressed as %LEL / s. The calculation formula is: ;

[0049] In the formula, The dot product is the projection of the concentration gradient vector onto the wind direction, representing the rate of change of concentration along the wind direction. The sign of this dot product has a clear physical meaning: if the dot product is greater than zero, it means that the concentration gradient direction is consistent with the wind direction, that is, the pollution source is located upwind, and the airflow transports the pollutants to this detector; if the dot product is less than zero, it means that the concentration gradient direction is opposite to the wind direction, that is, the pollution source is located downwind, and the airflow blows the pollutants away from this detector.

[0050] The concentration data C(t) at the current moment is corrected based on the airflow disturbance factor to obtain the equivalent concentration data. This equivalent concentration characterizes the background concentration "when there is no airflow convection," and is used for subsequent risk level determination. The correction formula is: ;in, The sampling interval is 1 second in this embodiment, and k is a preset compensation coefficient used to adjust the compensation intensity. The value of the compensation coefficient k is dynamically adjusted according to the directional relationship between the wind direction and the concentration gradient.

[0051] when When the concentration is greater than 0, the pollution source is located upwind, and the convection causes the concentration observed by this detector to be too high, requiring downward correction. In this case, a smaller compensation coefficient is used, denoted as . The value range is [0.3, 0.7]. In this embodiment, we take... =0.5;

[0052] when When the concentration is less than 0, the pollution source is located downwind. Convection causes the concentration observed by this detector to be lower than expected, requiring upward correction. In this case, a larger compensation coefficient is used, denoted as . The value range is [0.7, 1.0]. In this embodiment, we take... =0.8;

[0053] when When k=0 or the wind speed is lower than the preset wind speed threshold (e.g., v(t) < 0.1m / s), the influence of airflow can be ignored, so k=0 is set and no correction is made to the concentration data.

[0054] As input for generating the basic risk level in step S2, this correction effectively eliminates the interference of airflow convection on concentration measurements, making the risk perception closer to the actual leakage situation.

[0055] Step S2: Based on the equivalent concentration data obtained in step S1 The rate of change threshold matched with the scenario determined in step S1 and concentration threshold A basic risk level is generated through comparison; subsequent comparisons of concentration data involved in the judgment logic are all based on... The specific process is as follows:

[0056] The detectors maintain continuous communication with surrounding devices via a wireless network. Each detector includes its current risk level L (defined below) and a transient trigger flag in its data packet with each data report. Meanwhile, the detector continuously listens to the channel, receives broadcasts from neighboring nodes, and maintains a neighbor state table. This neighbor state table, indexed by the neighbor node ID, records the risk level most recently reported by each neighbor. Transient trigger flag And the corresponding receiving timestamp; in addition, if the neighboring node also includes its current rate of change threshold in the broadcast. or concentration change rate If no update is received from a neighbor within 60 seconds, the neighbor's record is removed from the neighbor status table to ensure the timeliness of information; in this way, the detector can grasp the overall risk situation of the surrounding area in real time.

[0057] The status information of the neighbor detectors specifically includes: the current risk level of the neighbor detectors (denoted as...). ) and transient trigger flag (denoted as The risk level ranges from 0 to 3, corresponding to low risk, medium risk, high risk, and the highest level, respectively; the transient trigger flag is either 0 or 1, with 1 indicating that the neighboring detector has triggered a transient event; additionally, the status information of the neighboring detector may optionally include its concentration change rate (…). ) and rate of change threshold ( This information is obtained through periodic broadcast data packets from neighbor detectors.

[0058] Based on the scenario information obtained above, the detector begins to generate monitoring thresholds that match the current operational risk. A threshold mapping table is pre-stored in the detector's internal firmware. This table was obtained through prior industrial field calibration and is constructed based on the differences in tolerance for leakage risk under different operational scenarios. For example, during hot work operations, there is an open flame source, and even a small leak can lead to serious consequences, thus requiring the most sensitive threshold; while during unattended periods, the threshold can be appropriately relaxed to avoid unnecessary interference. The specific content of the mapping table is as follows: When the hot work operation flag is valid, the rate of change threshold... Take 0.3% LEL / s, concentration threshold Take 5% LEL; when the loading / unloading operation mark is valid, Take 0.5% LEL / s, Take 8% LEL; when the personnel inspection sign is valid, Take 0.8% LEL / s, Take 9% LEL; when the no-occupancy period flag is valid, Take 1.5% LEL / s. Take 10% LEL; if there are no special scenario markers, i.e., under normal operating conditions, Take 1.0% LEL / s. The threshold is set to 10% LEL. If multiple scene flags are received simultaneously, the system prioritizes safety and selects the combination with the lowest threshold, i.e., the most stringent monitoring standard. For example, if both hot work and loading / unloading operations are in progress, the threshold corresponding to hot work will be used. To avoid frequent threshold changes due to brief fluctuations in scene flags, the detector employs a delayed confirmation mechanism for scene information: any scene flag must remain unchanged for three consecutive sampling periods (i.e., 3 seconds) to be considered valid and the threshold updated; if the flag disappears during this period, no switching occurs. This mechanism effectively filters out transient interference and ensures threshold stability.

[0059] The calibration process for this threshold mapping table is as follows: First, standard gas release devices are deployed in typical industrial scenarios (including hot work areas, loading and unloading areas, inspection channels, unmanned warehouses, etc.) to release known concentrations of combustible gases (such as methane and propane) at different rates (0.1% to 5% LEL / s). Simultaneously, the time required for the sensor to reach the alarm threshold under different scenarios, as well as the required safe reaction time for on-site operators, are recorded. Based on a large amount of measured data, statistical analysis methods are used to determine the optimal threshold combination for each scenario: the optimization objective is to trigger an alarm within 10 seconds of a leak occurring with a false alarm rate of less than 0.1%. The specific values ​​of the rate of change threshold and concentration threshold are determined through ROC curve analysis. After calibration, the threshold combinations are written into the detector firmware in key-value pairs, where the key is the scenario flag combination (e.g., ...). =1, =0, =0, =0), the value is the corresponding ( , Numerical pairs.

[0060] The process of generating a basic risk level includes:

[0061] In step S1, the detector has determined the rate of change threshold for matching the current scene based on the real-time acquired scene information by querying a preset threshold mapping table. and concentration threshold The original concentration data was filtered to obtain the equivalent concentration data at the current moment. Step S2 will use these threshold and concentration data to generate a basic risk level, which reflects the local single node's initial assessment of the current leakage risk.

[0062] First, the detector calculates the concentration change rate r(t) at the current moment, which physically represents the change in concentration within a unit sampling interval, characterizing the instantaneous trend of gas concentration change. The calculation formula is: ;in This represents the equivalent concentration data at the current moment. This represents the equivalent concentration data from the previous time step. A positive value for r(t) indicates an increase in concentration, while a negative value indicates a decrease. For the decreasing process, since the risk is receding, this step only considers it as a risk-free trigger, but the value is still retained for subsequent trend analysis.

[0063] It is important to note that when the detector is first activated or awakened from a long period of dormancy, due to the lack of effective concentration data from the previous moment, the concentration change rate r(t) is set to 0 by default, and the risk level determination for the current sampling point is skipped until sufficient data is accumulated in the next sampling cycle before calculation. Furthermore, if two or more consecutive sampling points are missing due to communication failure or sensor malfunction, the change rate is not calculated for the first sampling point after data recovery; only the current equivalent concentration data is considered. With concentration threshold The risk level is determined by comparing the results, and the rate of change determination is activated only after the continuous sampling points return to normal.

[0064] Next, the detector according to With concentration threshold The comparison results and r(t) versus the rate of change threshold Based on the comparison results, a basic risk level is generated according to a progressive rule. The rule categorizes risks into three levels: low risk (…). =0), medium risk ( =1) and high risk ( =2). The specific determination logic is as follows: if the current concentration is... ≥ Or the current rate of change r(t) ≥ If the basic risk level is determined to be high risk, it is denoted as [missing information]. =2; if the above conditions are not met, but ≥0.5× Or r(t)≥0.5× If so, it is classified as a medium-risk level. =1; if neither of these conditions is met, it is classified as a low-risk level. =0. For example, suppose the current scenario is a hot work operation. =5%LEL, =0.3%LEL / s, if the detector measures =6%LEL, then it is directly judged as high risk; if =3%LEL but r(t)=0.4%LEL / s, it is also judged as high risk; if =3%LEL and r(t)=0.2%LEL / s, then because >0.5× =2.5%LEL, judged as medium risk; it is worth noting that half of the threshold (0.5 times) in the above judgment rule is a configurable parameter. In this embodiment, 0.5 is used as the default value. Its physical meaning is that 50% of the threshold is used as the warning trigger point, which retains sufficient reaction time and avoids too frequent warnings. In practical applications, this coefficient can be adjusted according to the safety requirements of different industrial sites; for the concentration decrease process (r(t)<0), the processing principle of this embodiment is: if the current equivalent concentration data Still above the concentration threshold Regardless of the rate of decrease, the risk level will remain high until the concentration falls below the threshold. This is because even if the concentration is decreasing, as long as it remains within the dangerous concentration range, an alert should be maintained.

[0065] The generated basic risk level The data will be temporarily stored in the detector's memory as input for the subsequent S3 spatial collaborative correction step. This level is based solely on local single-point data and does not yet consider the status of surrounding neighbors. Therefore, it needs to be corrected by incorporating neighbor information to more accurately reflect the true leakage situation. Through this hierarchical and progressive judgment method, the detector can adopt differentiated response strategies under different risk levels, laying the foundation for subsequent dynamic reporting and resource scheduling.

[0066] Step S3: As Figure 2 As shown, based on the state information of neighboring detectors, the basic risk level is corrected according to the preset spatial coordination rules to obtain the corrected risk level.

[0067] In step S2, the detector has generated a basic risk level based on local single-point data. However, gas leaks exhibit spatial diffusion characteristics, and isolated assessments of a single node may not accurately reflect the true extent and evolution trend of the leak. Therefore, step S3 utilizes the state information of neighboring detectors and, according to preset spatial coordination rules, [the system]... The revised risk level was obtained by making adjustments to better reflect the overall risk in the region. This revision process incorporates both group risk consistency assessment and risk acceleration and diffusion determination to further improve the accuracy of risk perception.

[0068] First, the detector reads the latest status information of all valid neighbors from the locally maintained neighbor status table. Each neighbor record contains at least the following fields: neighbor node ID, and the risk level reported by that neighbor. Transient trigger flag Concentration change rate (If included in the neighbor broadcast) and the neighbor's own rate of change threshold. (This threshold can be obtained directly from the configuration parameters of the neighbor broadcast, or calculated based on the neighbor's scenario information using the same threshold mapping table.) If a neighbor has not updated for more than 60 seconds, it is considered invalid and will not participate in this correction.

[0069] However, isolated judgments by a single node may overlook the spatial propagation effect of leaks. Therefore, the detector needs to utilize the state information of neighboring nodes to perform spatially coordinated corrections to the basic risk level; the detector executes the following judgment logic sequentially... Preliminary revisions made:

[0070] The detector reads the current risk level of all valid neighbors from the neighbor status table. and transient trigger flag And execute the following correction logic:

[0071] Highest-level neighbor forced upgrade: If any neighbor has a risk level of the highest level ( =3), then regardless of the local basic risk level, the basic risk level will be directly upgraded to the high risk level. =2), because the highest level means that a critical leak has occurred in the neighboring area and the surrounding area must be on high alert;

[0072] High-risk neighbor aggregation enhancement: If the above conditions are not met, further check whether there are at least two neighbors with a high-risk level. =2), or there exists a neighboring property with a high-risk level and its concentration change rate exceeds the neighboring property's own change rate threshold ( If any one of these conditions is met, it indicates that there are relatively consistent risk signals in the surrounding area, which may be spreading. In this case, the basic risk level is raised to the medium risk level, i.e., =1. Otherwise, the basic risk level remains unchanged, i.e. = .

[0073] Alert timer triggering and threshold sensitivity adjustment: If any neighbor's transient trigger flag is triggered... If the detection is effective, it indicates a drastic change in the surrounding environment, requiring increased local monitoring sensitivity to detect potential diffuse gases in advance. The detector immediately activates a watchdog timer with a preset duration of 30 seconds; if the watchdog timer is already running, it is reset, restarting the countdown from 30 seconds. This means that as long as new transient events continue to occur in the surrounding environment, the alert status will remain in place until no new transient events are triggered within 30 seconds after the last transient event. This design ensures continuous sensitivity maintenance in scenarios of continuous sudden leaks. During the effective period of the watchdog timer, the rate of change threshold used in subsequent steps... It will be temporarily adjusted to 0.5 times the original value, that is... This means that even a small rate of concentration change could trigger an upgrade in the risk level, thus enabling a proactive response to potential diffusion risks. For example, the original... =1.0%LEL / s, reduced to 0.5%LEL / s during the alert period. It is worth noting that the adjustment of the change rate threshold by the alert timer only affects the risk level determination of the current detector during the alert period, and the adjusted threshold must not be lower than the preset minimum protection threshold. In this embodiment, the minimum change rate threshold is set to 0.1%LEL / s. Even if the original threshold is adjusted by 0.5 times and falls below this value, it will still be executed at 0.1%LEL / s. This lower limit protection ensures that even under extremely sensitive conditions, the risk level will not be frequently triggered by minor sensor noise. The alert timer and the hold timer in step S4 are independent of each other: the alert timer is used to temporarily adjust the change rate threshold to enhance sensitivity, and the threshold automatically recovers after its timeout; the hold timer is used to maintain the duration of the highest risk level, and its duration is affected by the risk acceleration diffusion flag. Both can run simultaneously without interference.

[0074] The temporary risk level obtained after the above basic revisions This will serve as the basis for further revisions.

[0075] The process of further revising the basic risk level also includes:

[0076] To more precisely capture the group characteristics of regional risk, the detector further calculates a group risk consistency index. This index is obtained by statistically analyzing all valid neighbors in the neighbor status table, identifying those with a risk level no lower than medium risk (i.e.,...). The number of nodes with a value greater than or equal to 1 is denoted as . At the same time, count the total number of neighbors, and record it as... The group risk consistency index P is defined as: This indicator reflects the prevalence of risk signals in the surrounding area. If P exceeds a preset diffusion threshold... (In this embodiment, we take) =0.5, meaning more than half of the neighbors are at risk), and the current temporary risk level is... Lower than the high-risk level (i.e.) <2), then will Promoted to the next level. The promotion rules are: If... =0 then increase to 1, if =1 will be upgraded to 2, but the upgraded level will not exceed the high-risk level (Level 2). This modification reflects the principle of "group consensus": when most neighbors perceive the risk, even if the local indicator has not yet reached the threshold, the risk should be considered to exist; to ensure the reliability of the statistical results, the total number of neighbors should be determined first before conducting a group risk consistency assessment. Whether the minimum statistical sample size has been reached is determined; in this embodiment, the minimum sample size is set to 3. If the value is less than 3, it is considered that there is currently no statistical basis for conducting a population consistency assessment. Therefore, this step of correction is skipped, and the basic correction from step S3 is directly adopted. This serves as a revised risk level. This design avoids statistical bias caused by an insufficient number of neighbors.

[0077] For example, suppose a detector has 8 neighbors, 5 of which have a risk level of medium or higher, then P = 5 / 8 = 0.625 > 0.5. If currently... =1, then upgrade to 2; if =0, then increase to 1; if If it is already 2, then it will not be increased further.

[0078] Based on the consensus assessment of group risk, the detector further analyzes the rate of change characteristics of risky neighbors to determine whether there is an accelerating spread trend. All neighbors with a risk level not lower than the medium risk level are selected from the neighbor status table. For neighbors of (≥1), obtain their concentration change rates. Calculate the arithmetic mean of these rates of change. : If equivalent concentration data is missing at certain times within a reporting period due to sampling failure or communication anomalies, the average value will be calculated using only the actual collected valid data points as the denominator. An effective data ratio field will be appended to the summary data packet for the monitoring center to assess the completeness and reliability of the data within that period. If the effective data ratio is lower than a preset completeness threshold (e.g., 50%), the data for that period will be considered invalid, not reported, and a data loss event will be recorded for subsequent diagnostic purposes.

[0079] like =0, then Defined as 0. Preset mean rate of change threshold. In this embodiment, industrial experience is taken as the basis for selection. =1.0%LEL / s. If > If the current risk level is determined to be accelerated, the detector is identified as being in a state of accelerated spread. This determination does not directly correct the current risk level, but rather affects the hold timer setting in subsequent step S4. Specifically, when an accelerated spread state is determined, the detector records an "accelerated spread flag." =1, this flag will be used in step S4 to extend the preset base hold duration of the hold timer. If accelerated diffusion is not determined, then =0.

[0080] After the above three steps, the result is (This may have been further increased by the herd consensus indicator) This is the final revised risk level, denoted as... This value will serve as one of the baseline inputs for the transient trigger determination in step S4. If no transient is triggered in step S4, the current risk level L will be directly set to... If a transient condition is triggered, L will be forcibly set to the highest level, 3, and the duration of the timer will be adjusted accordingly. The logo has been adjusted.

[0081] Through the aforementioned spatial collaborative correction, the detector not only considers the data of a single node, but also incorporates the risk status of neighbors, group consistency, and diffusion trends, enabling the risk level to more accurately reflect the true leakage situation and providing a reliable decision-making basis for subsequent dynamic reporting and emergency response.

[0082] Step S4: Determine whether the equivalent concentration data meets the preset transient triggering condition. The transient triggering condition is that the concentration increase in two consecutive sampling intervals exceeds the preset transient change threshold.

[0083] If the transient triggering condition is met, the detector's risk level will be forcibly set to the highest level, and a hold timer will be started; otherwise, the corrected risk level will be used as the current risk level.

[0084] In step S3, the detector has obtained the corrected risk level through spatial cooperative correction. And recorded the acceleration indicators based on the risk acceleration status. Step S4 will perform transient trigger detection, which is independent of the aforementioned risk level generation logic and is specifically designed to capture sudden, high-rate concentration surge events, thereby determining the final current risk level L and maintaining the running state of the timer.

[0085] At each sampling time t, the detector not only calculates the current equivalent concentration data It also retrieves the equivalent concentration data from the first two time points from the local cache. and Based on these three consecutive sampling points, the first upper amplitude is calculated. With the second upper amplitude : = - , = - The physical meaning of these two amplitudes is the absolute increase in concentration within a unit sampling interval (1 second), expressed in %LEL. The system presets a fixed transient change threshold. This threshold is set based on a combination of the hazard rise rate of a typical combustible gas near its lower explosive limit and industrial safety response time requirements. In this embodiment, The value is set at 2% LEL, based on the following: if the concentration rise exceeds 2% LEL within two consecutive 1-second intervals, the leakage rate is at least 2% LEL / s and continues to accelerate or maintain a high speed, sufficient to approach the lower explosive limit (e.g., 10% LEL) within seconds, necessitating immediate triggering of the highest level alarm. This threshold has been verified through industrial field testing and can effectively distinguish between occasional sensor noise (usually isolated spikes) and genuine leakage (manifested as a continuous rise). It should be noted that this embodiment uses a transient change threshold. =2%LEL is set based on a 1Hz sampling frequency. The physical meaning of this threshold is: the concentration increase per second exceeds 2%LEL, and this increase continues for two consecutive seconds. If the sampling frequency is adjusted in practical applications, then... The sampling frequency should be adjusted proportionally accordingly. For example, if the sampling frequency is reduced to 0.5Hz (sampling once every 2 seconds), then... It should be adjusted to 4% LEL to maintain the same sensitivity to leakage rate. This embodiment uses a fixed sampling frequency of 1Hz by default. Fixed at 2% LEL.

[0086] The logic for determining transient triggering conditions is as follows:

[0087] like > and > If the transient triggering condition is met, the transient triggering flag is set. For effective ( =1);

[0088] otherwise, Keep as invalid ( =0).

[0089] For example, suppose that at a certain moment it is measured =0.5%LEL, =3.0%LEL =6.0%LEL, then =2.5%LEL =3.0%LEL, both exceed 2%LEL, therefore a transient is triggered. If =0.5%LEL, =2.2%LEL =3.8%LEL, then =1.7%LEL does not exceed the threshold. =1.6%LEL is not exceeded, so it is not triggered. This design effectively filters out single random noise interference.

[0090] once Once enabled, the detector immediately performs the following actions:

[0091] The current risk level is forcibly set to the highest level, L=3. This level corresponds to the emergency reporting strategy in subsequent step S5.

[0092] A hold timer is activated. This timer ensures the stability and reliability of the alarm status, preventing frequent changes in risk level due to short-term concentration fluctuations. The duration of the hold timer is not fixed but dynamically adjusted based on the risk acceleration state determined in step S3. Specifically, the system presets a basic hold duration. In this embodiment, 30 seconds is used; at the same time, an extension duration is preset. In this embodiment, 30 seconds is used. If it has been determined in step S3 that the current state is one of accelerated risk diffusion (i.e.) If =1), then hold the actual duration of the timer. The sum of base duration and extended duration: ;otherwise, This design is based on the following considerations: when multiple neighboring areas are at risk and the average rate of change is high, the leak may be spreading rapidly, requiring a longer hold time to ensure a sufficient window for emergency response; if it is only an isolated transient event, a 30-second hold time is sufficient to cover the initial response phase of a typical emergency.

[0093] Once the timer is started, during the timing period, even if the equivalent concentration data at subsequent sampling times decreases or the rate of change decreases, the current risk level L will always remain at the highest level, 3, until the timer expires. Simultaneously, the detector immediately broadcasts a transient trigger message via the wireless network. This message contains the local ID and a transient flag. Upon receiving this message, neighboring nodes will activate their respective warning timers according to the spatial coordination rules described in step S3, thereby creating a spatial warning propagation.

[0094] If the transient trigger flag is invalid ( If L = 0), then the current risk level of the detector is directly adopted from the corrected risk level in step S3, i.e., L = At this point, the timer remains off. If it was previously started and has timed out, it remains off. If the timer is running but has not yet timed out, and no new transient is triggered in subsequent sampling, the timer continues to count down. L remains at the highest level until timeout. After timeout, if no new transient is triggered, L will gradually downgrade according to the risk decay logic.

[0095] It is important to emphasize that transient trigger detection and risk level correction in step S3 are two parallel processing paths: the former identifies drastic changes based on a fixed high threshold, while the latter assesses progressive risks based on scenario-adaptive thresholds and spatial collaborative evaluation. The two are coupled through a transient trigger flag and a hold timer—once a transient event occurs, its priority is higher than any conventional judgment result, thus ensuring the system has the highest response speed to sudden leaks. The final determined current risk level L and transient trigger flag... It will be encapsulated into a data packet for use in the reporting decision and network coordinator scheduling in step S5. Through this design, the detector can sensitively capture both slowly accumulating micro-leaks and instantly pinpoint high-speed jet leaks, achieving comprehensive coverage of various leakage modes.

[0096] Step S5: Based on the current risk level, dynamically adjust the data reporting cycle and reporting content of the detector, and send a data packet containing the current risk level and transient trigger flag to the network coordinator.

[0097] In step S4, the detector has determined the final current risk level L based on the satisfaction of the transient triggering conditions and the corrected risk level, and recorded the transient triggering flag. Step S5 will dynamically adjust the data reporting strategy based on the current risk level and send data packets containing risk information to the network coordinator to achieve precise matching of communication resources and risk levels.

[0098] The detector has a built-in reporting parameter configuration table that maps four risk levels to different reporting cycles and data contents. This mapping relationship is designed based on the following engineering considerations: For low-risk situations, minimize communication load and upload only summary information for long-term trend analysis; for medium-risk situations, upload complete sequences to support detailed assessment by the monitoring center; for high-risk situations, increase the reporting frequency to track leak evolution in real time; and for the highest level, upload single-point data as quickly as possible to ensure that critical events are captured immediately. The specific mapping rules are as follows:

[0099] When the current risk level L=0 (low risk), the detector adopts the first reporting cycle. Sending data, in this embodiment =600 seconds. The reported content is compressed summary data, which includes at least the average concentration, peak value, and duration within one reporting period. Specifically, the detector stores 600 concentration data points collected over the past 600 seconds in its local cache (due to a sampling frequency of 1Hz), and calculates the arithmetic mean of these data points. Maximum value It also counts the duration for which the concentration exceeds a preset attention threshold (e.g., 1% LEL). The final summary data packet contains only these three values ​​and a timestamp, resulting in a very small data volume, which can effectively reduce channel occupancy during long-term operation.

[0100] When the current risk level L=1 (medium risk), the detector adopts the second reporting cycle. Sending data, in this embodiment =60 seconds. The reported content is the complete concentration sequence within the most recent reporting period, that is, 60 concentration data points collected within the past 60 seconds. This sequence is arranged in chronological order, allowing the monitoring center to analyze concentration change trends and calculate derivative indicators such as the rate of change. Less than This reflects the increased demand for real-time data as risks rise.

[0101] When the current risk level L=2 (high risk), the detector adopts the third reporting cycle. Sending data, in this embodiment =10 seconds. The reported content is the concentration sequence within the most recent reporting period, that is, 10 concentration data points collected in the past 10 seconds. Further shorten the timeframe to ensure that the monitoring center can track the leakage evolution process with second-level resolution, providing real-time data support for emergency decision-making.

[0102] When the current risk level L=3 (the highest level), the detector adopts the fourth reporting cycle. Sending data, in this embodiment =1 second. The reported content is real-time single-point concentration data, that is, the concentration at each sampling time. All reports were submitted independently, without any compression or aggregation. With a minimum reporting period consistent with the sampling period, zero-latency data transmission is achieved. Simultaneously, data packets at this level are marked as having the highest priority at the network layer, ensuring resource preemption during channel congestion.

[0103] The above four reporting periods satisfy a strict size relationship: In this embodiment, we take... =600s =60s =10s =1s, but in practical applications, it can be adjusted according to the safety requirements and communication conditions of specific industrial sites.

[0104] After determining the reporting cycle and data content, the detector generates data packets periodically according to that cycle. Each data packet contains at least the following fields: detector ID, timestamp, current risk level L, and transient trigger flag. This includes the reporting content (summary data or concentration sequence) determined according to the risk level. Specifically, this includes the current risk level L and the transient trigger flag. It is a key input for the network coordinator to perform global resource scheduling and for neighboring nodes to perform spatial collaborative correction, and its accuracy must be ensured.

[0105] After the data packet is encapsulated, the detector sends it to the network coordinator via the wireless network. During transmission, the detector follows standard channel access protocols (such as CSMA / CA). To cooperate with the network coordinator's global scheduling strategy, the detector also listens for broadcast commands on the channel. If a resource scheduling command is received from the network coordinator, such as a "silent command" broadcast when the channel load exceeds 70%, detectors at level 0 must temporarily suspend data reporting but continue listening to the channel until a resumption command is received or the load decreases. Detectors at level 1 and above are not affected by the silent command, ensuring the continuous transmission of medium- to high-risk data. This mechanism effectively avoids the impact of channel congestion on critical alarm information.

[0106] Through the above steps, the detector achieves dynamic adaptation between risk level and communication resources: maximizing bandwidth conservation during low-risk periods, gradually increasing data granularity and frequency during medium-to-high-risk periods, and seizing channel resources with minimal delay during the highest-risk periods. This tiered reporting strategy not only optimizes the overall network load but also ensures that critical information is prioritized and transmitted to the monitoring center in real time during emergencies, buying valuable time for emergency response.

[0107] For complex industrial site operations characterized by dynamic and ever-changing scenarios, and the spatial diffusion and suddenness of gas leaks, existing monitoring logic based on fixed thresholds and single-point judgments has significant limitations under non-steady-state conditions. Firstly, the varying tolerances for leak risks across different operational scenarios mean that a uniform threshold cannot simultaneously achieve high sensitivity and low false alarm rates. Secondly, fragmented spatial information prevents the early identification of leak diffusion trends, and single-node judgments create blind spots. Furthermore, the mismatch between periodic reporting mechanisms and transient changes makes it difficult to capture high-speed jet leaks in real time. This method addresses these limitations by constructing a scenario-adaptive threshold generation mechanism to dynamically match monitoring sensitivity with real-time operational risks; incorporating the risk status of neighboring nodes into local judgments through spatial collaborative correction, forming regionally coordinated early warning propagation; and achieving immediate capture of sudden leaks through transient trigger detection independent of periodic reporting. The nonlinear fusion of these multi-dimensional information allows the detector to accurately characterize leak risks under different operating conditions and dynamically adjust reporting strategies based on risk levels, optimizing the adaptation of communication resources to monitoring needs. This significantly improves the timeliness and reliability of flammable gas leak detection in complex industrial environments.

[0108] Example 2: Figure 3 As shown, a combustible gas intelligent alarm system includes:

[0109] Basic information acquisition module: acquires the current concentration data, real-time airflow parameters and scene information of the detector's location area, and the status and spatial information of at least one neighboring detector; based on the scene information, determines the rate of change threshold and concentration threshold that match the current scene; and corrects the current concentration data based on the real-time airflow parameters and the spatial information of neighboring detectors to obtain equivalent concentration data.

[0110] Basic Risk Analysis Module: Generates a basic risk level based on the comparison results between equivalent concentration data and concentration thresholds, as well as the comparison results between the rate of change of equivalent concentration data and the rate of change thresholds.

[0111] Risk correction processing module: Based on the status information of neighbor detectors, the basic risk level is corrected according to the preset spatial coordination rules to obtain the corrected risk level;

[0112] Transient trigger decision module: Determines whether the equivalent concentration data meets the preset transient trigger conditions. The transient trigger conditions are that the concentration increase in two consecutive sampling intervals exceeds the preset transient change threshold.

[0113] If the transient triggering condition is met, the detector's risk level will be forcibly set to the highest level, and a hold timer will be started; otherwise, the corrected risk level will be used as the current risk level.

[0114] Data reporting module: Based on the current risk level, dynamically adjust the data reporting cycle and content of the detector, and send a data packet containing the current risk level and transient trigger flag to the network coordinator.

[0115] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0116] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to any specific implementation. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention.

Claims

1. A method for intelligent alarm of combustible gas, characterized in that, Applications in wireless gas detectors include: The system acquires the current concentration data, real-time airflow parameters and scene information of the area where the detector is located, and status and spatial information of at least one neighboring detector. Based on the scene information, it determines a rate of change threshold and a concentration threshold that match the current scene. Based on the real-time airflow parameters and the spatial information of the neighboring detectors, it corrects the current concentration data to obtain equivalent concentration data. The current concentration data is the combustible gas concentration value collected by the detector at the current moment. The status information includes the risk level and transient trigger flag of the neighboring detectors. A basic risk level is generated based on the comparison results between the equivalent concentration data and the concentration threshold, and the comparison results between the change rate of the equivalent concentration data and the change rate threshold. Based on the status information of the neighbor detectors, the basic risk level is corrected according to the preset spatial coordination rules to obtain the corrected risk level. Determine whether the equivalent concentration data meets the preset transient triggering condition, wherein the transient triggering condition is that the concentration increase in two consecutive sampling intervals exceeds the preset transient change threshold. If the transient triggering condition is met, the risk level of the detector is forcibly set to the highest level, and a hold timer is started; otherwise, the corrected risk level is used as the current risk level. Based on the current risk level, the data reporting cycle and reporting content of the detector are dynamically adjusted, and a data packet containing the current risk level and transient trigger flag is sent to the network coordinator.

2. The intelligent alarm method for combustible gas according to claim 1, characterized in that, The scene information includes at least one of the following: hot work sign, loading and unloading sign, personnel inspection sign, and no-man's-land sign; The hot work sign is used to indicate whether hot work is being carried out in the area where the detector is located; The loading and unloading operation sign is used to indicate whether a combustible gas loading and unloading operation is being carried out in the area. The personnel patrol sign is used to indicate whether patrol personnel have entered the area; The unattended time period marker is used to indicate whether the area is in a preset unattended time period.

3. The intelligent alarm method for combustible gas according to claim 2, characterized in that, The process of determining the rate of change threshold and concentration threshold that match the current scenario includes: Based on the scenario information, a preset threshold mapping table is queried. The threshold mapping table stores the correspondence between different scenario combinations and change rate thresholds and concentration thresholds. If the scene information contains multiple scene markers, then the combination with the most stringent thresholds in the threshold mapping table is selected as the change rate threshold and concentration threshold for the current scene matching. The threshold mapping table is pre-set through industrial field calibration and stored in the detector's memory.

4. The intelligent alarm method for combustible gas according to claim 1, characterized in that, The process of correcting the current concentration data to obtain equivalent concentration data includes: The real-time airflow parameters include wind speed and wind direction; the spatial information of the neighboring detectors includes the spatial distance and relative orientation between the detector and its neighboring detectors. Based on the concentration data of the detector at the current moment and the concentration data of each neighboring detector, combined with the spatial distance and relative orientation, the concentration gradient vector at the current moment is calculated. The concentration gradient vector is used to characterize the direction and rate of change of gas concentration in space. Based on the wind speed, wind direction, and concentration gradient vector, an airflow disturbance factor is calculated. The airflow disturbance factor is used to characterize the contribution of airflow convection to the concentration data at the current moment. The concentration data at the current moment is corrected using the airflow disturbance factor to obtain equivalent concentration data.

5. The intelligent alarm method for combustible gas according to claim 4, characterized in that, The concentration data at the current moment is corrected using the aforementioned airflow disturbance factor, specifically including: Based on the directional relationship between the wind direction and the concentration gradient vector, determine the correction relationship between the detector's concentration data and the background concentration when there is no convection influence: If the wind direction is consistent with the direction of the concentration gradient vector, it is determined that the concentration data at the current moment is higher than the background concentration when there is no convection, and the concentration data at the current moment needs to be corrected downward. If the wind direction is opposite to the direction of the concentration gradient vector, it is determined that the concentration data at the current moment is lower than the background concentration when there is no convection, and the concentration data at the current moment needs to be corrected upward. The amount of upward or downward correction is determined based on the wind speed and the preset compensation coefficient. The correction amount is added to the concentration data at the current moment to obtain the equivalent concentration data; When the wind speed is lower than the preset wind speed threshold or the relationship between the wind direction and the concentration gradient vector cannot be determined, the compensation coefficient is zero and the concentration data at the current moment is not corrected.

6. The intelligent alarm method for combustible gas according to claim 5, characterized in that, The process of generating a basic risk level includes: The difference between the equivalent concentration data at the current moment and the equivalent concentration data at the previous moment is marked as the concentration change rate at the current moment; If the equivalent concentration data is greater than or equal to the concentration threshold or the concentration change rate is greater than or equal to the change rate threshold, then the basic risk level at the current moment is determined to be a high-risk level. Otherwise, if the equivalent concentration data is greater than or equal to half of the concentration threshold or the concentration change rate is greater than or equal to half of the change rate threshold, the basic risk level at the current moment is determined to be medium risk. Otherwise, the basic risk level at the current moment is determined to be low risk.

7. The intelligent alarm method for combustible gas according to claim 1, characterized in that, The specific process for revising the basic risk level includes: Obtain the risk level and transient trigger flags of each neighboring detector; If any neighboring detector has the highest risk level, then the basic risk level will be raised to the high risk level. Otherwise, if there are at least two neighboring detectors with a high risk level, or if there is a neighboring detector with a high risk level and the concentration change rate obtained from the status information of that neighboring detector exceeds its corresponding change rate threshold, then the basic risk level will be raised to a medium risk level. Otherwise, the aforementioned basic risk level remains unchanged; If the transient trigger flag of any neighboring detector is valid, a warning timer is started, and during the valid period of the warning timer, the change rate threshold at the current moment is temporarily adjusted to a preset multiple of the change rate threshold, wherein the preset multiple is less than 1.

8. The intelligent alarm method for combustible gas according to claim 1, characterized in that, The process of determining transient triggering conditions includes: Calculate the concentration increase between two consecutive sampling intervals and label them as the first upper amplitude and the second upper amplitude, respectively; the first upper amplitude is the difference between the equivalent concentration data at the current time and the equivalent concentration data at the previous time, and the second upper amplitude is the difference between the equivalent concentration data at the previous time and the equivalent concentration data at the previous two times. If both the first upper amplitude and the second upper amplitude are greater than the preset transient change threshold, then the transient triggering condition is determined to be met, and the transient triggering flag is set to valid. Otherwise, set the transient trigger flag to invalid.

9. The intelligent alarm method for combustible gas according to claim 1, characterized in that, The specific process of dynamically adjusting the data reporting cycle and reporting content of the detector includes: When the current risk level of the detector is low risk, compressed summary data is sent in the first reporting cycle. The summary data includes at least the average concentration, peak value and duration within one reporting cycle. When the current risk level of the detector is medium risk, the complete concentration sequence is sent in a second reporting period, which is shorter than the first reporting period; When the current risk level of the detector is high risk, the concentration sequence of the most recent reporting period is sent in the third reporting period, which is shorter than the second reporting period. When the current risk level of the detector is the highest level, real-time single-point concentration data is sent in the fourth reporting cycle, which is shorter than the third reporting cycle.

10. A combustible gas intelligent alarm system, using a combustible gas intelligent alarm method as described in any one of claims 1-9, characterized in that, include: Basic information acquisition module: acquires the current concentration data, real-time airflow parameters and scene information of the area where the detector is located, and the status and spatial information of at least one neighboring detector; Based on the scene information, a change rate threshold and a concentration threshold matching the current scene are determined; based on the real-time airflow parameters and the spatial information of neighboring detectors, the concentration data at the current moment is corrected to obtain equivalent concentration data; Basic risk analysis module: Based on the comparison results of the equivalent concentration data and the concentration threshold, and the comparison results of the change rate of the equivalent concentration data and the change rate threshold, a basic risk level is generated; Risk correction processing module: Based on the status information of the neighbor detector, the basic risk level is corrected according to the preset spatial coordination rules to obtain the corrected risk level; Transient trigger decision module: determines whether the equivalent concentration data meets the preset transient trigger condition, wherein the transient trigger condition is that the concentration increase in two consecutive sampling intervals exceeds the preset transient change threshold; If the transient triggering condition is met, the risk level of the detector is forcibly set to the highest level and a hold timer is started; otherwise, the corrected risk level is used as the current risk level. Data reporting module: Based on the current risk level, dynamically adjust the data reporting cycle and reporting content of the detector, and send a data packet containing the current risk level and transient trigger flag to the network coordinator.

Images