Laboratory hazardous gas leakage monitoring method and system

Abstract

The invention relates to the technical field of laboratory safety monitoring, and discloses a laboratory hazardous gas leakage monitoring method and system, and the method comprises the steps: configuring an electronic identity tag for a hazardous gas related article, deploying a read-write device in an operation region, and automatically generating a temporary operation voucher with timeliness when the article enters the operation region; the environment concentration is monitored in real time through a gas sensor, and when abnormity is detected, whether a matching voucher exists or not is checked firstly so as to distinguish normal operation and abnormal leakage; according to the method, through a dynamic coupling mechanism of the operation certificate and the area sensing, controlled release and accidental leakage are distinguished, the balance problem of traditional threshold value alarm between sensitivity and false alarm rate is improved, and the reliability of the system is improved. And meanwhile, the accuracy of leakage detection and the robustness of the system are improved by combining near-field gradient verification and an air flow abrupt change self-adaptive strategy.

Description

TECHNICAL FIELD

[0001] The present application relates to a kind of laboratory dangerous gas leakage monitoring method and system, belong to laboratory safety monitoring technical field. BACKGROUND

[0002] Current laboratory gas monitoring mainly relies on fixed threshold alarm system, by deploying gas sensor network real-time monitoring environmental concentration, when detection value exceeds preset threshold, it triggers alarm, this indiscriminate alarm mode has inherent defects: on the one hand, to capture trace leakage needs to set lower threshold, leading to normal experiment operation such as reagent access frequently triggers false alarm;On the other hand, increase threshold although reduce false alarm will miss the slow accumulation of dangerous leakage.

[0003] Existing improvement attempts include using multi-sensor data fusion or machine learning algorithm to optimize threshold, but none of them solve the main problem-the system lacks the ability to distinguish the context judgment of controlled release and accidental leakage, specifically, the existing technology mainly has the following main limitations: 1, cannot establish logical connection between gas concentration signal and experimental operation behavior;2, lack of self-adaptive ability to sudden airflow disturbance and other environmental interference;3, the problem of declining monitoring reliability caused by sensor performance degradation has not been systematically solved. Therefore, how to build an intelligent monitoring system with operation context awareness ability, while ensuring high detection sensitivity, to realize the essential distinction of leakage event, has become the technical problem to be solved by the present application. SUMMARY

[0004] The present application provides a kind of laboratory dangerous gas leakage monitoring method and system, its main purpose is to solve the core technical problems that the existing monitoring system cannot distinguish controlled release and accidental leakage due to lack of operation context judgment ability, resulting in the mutual restraint of false alarm rate and missed detection rate.

[0005] To achieve the above purpose, the present application provides a kind of laboratory dangerous gas leakage monitoring method, the method comprises the following steps: Electronic identity tag is configured for the articles used by the dangerous gas in laboratory, and label reading and writing device is deployed in the article operation area of the articles used by the dangerous gas; When the articles used by the dangerous gas enter the article operation area, the label reading and writing device automatically reads the identity information of the electronic identity tag, and generates a temporary operation voucher with a given effective time length based on the identity information and the article operation area information; Real-time monitoring of gas concentration in the environment is carried out by gas sensor, and the area information where the gas sensor is located is obtained;When the gas concentration detected by any gas sensor exceeds the first concentration threshold, it is judged whether there is an effective temporary operation voucher that completely matches the area information where the gas sensor is located and the detected gas type; If there is a valid temporary operation credential that matches the area information where the gas sensor is located and the detected gas type, the system determines that the concentration increase is caused by normal operation and does not trigger a first-level warning, and records the data; If there is no valid temporary operation credential that matches the area information where the gas sensor is located and the detected gas type, the system determines that the concentration increase is caused by an abnormal leak and triggers a first-level warning. Regardless of whether there is a valid temporary operation credential, once the concentration of the gas detected by any gas sensor is higher than a second concentration threshold, which is higher than the first concentration threshold and is defined as a high-risk concentration that causes immediate harm to the human body, the system immediately triggers the highest-level alarm.

[0006] Preferably, after triggering the first-level warning, the system further includes the following steps: sending a wake-up instruction to at least one adjacent gas sensor in a dormant state around the gas sensor that triggered the first-level warning; the woken-up adjacent gas sensor performs concentration measurement and returns a concentration reading; comparing the concentration reading of the gas sensor that triggered the first-level warning with the concentration reading of the woken-up adjacent gas sensor, and when the ratio of the two is greater than a preset concentration gradient threshold, determining the source location of the abnormal leak or the reliability of the first-level warning.

[0007] Preferably, after triggering the first-level warning, before sending the wake-up instruction to the adjacent gas sensor, the system further includes the following steps: acquiring a physical state signal indicating that the macroscopic air flow state in the laboratory has changed abruptly in real time; when the physical state signal indicates that the macroscopic air flow is in an abrupt state, suspending the execution of the logic of triggering the first-level warning within a specified time period if there is no valid temporary operation credential that matches the area information where the gas sensor is located and the detected gas type.

[0008] Preferably, the electronic identity tag is a passive ultra-high frequency radio frequency identification electronic tag.

[0009] Preferably, when comparing the concentration reading of the gas sensor that triggered the first-level warning with the concentration reading of the woken-up adjacent gas sensor, if both concentration readings are lower than a preset verification threshold or the spatial distribution of the two concentration readings does not meet the specified spatial gradient characteristics, the system temporarily suspends the alarm and enters a continuous observation mode.

[0010] Preferably, the method further comprises the following steps: within a predetermined maintenance period, the system actively releases a standardized dose of the detection substance near the target gas sensor; obtaining the response data of the target gas sensor to the detection substance, the response data including the response time from receiving the detection substance signal to the peak reading and the peak height; calculating the current performance status of the target gas sensor by comparing with the pre-stored reference response data, the performance status being the time delay deviation rate and the sensitivity decay rate; and adjusting the first concentration threshold value used by the target gas sensor for determining whether there is a valid temporary operation credential or adjusting the weight factor of the target gas sensor in near field verification according to the performance status and a preset compensation algorithm.

[0011] Preferably, the method further comprises the following steps: recording the communication time delay of the adjacent gas sensor from receiving the wake-up instruction to returning the concentration measurement value; calculating the wireless communication interference intensity of the area where the adjacent gas sensor is located according to the deviation, trend and amplitude of the communication time delay from the preset reference time delay within a predetermined observation period, and generating a physical interference area label according to the interference intensity; and when transmitting a first-level warning or a highest-level alarm information, preferentially selecting a node in a non-interference area for information routing or instructing the gas sensor in the physical interference area to increase the number of information retransmissions.

[0012] Preferably, the generation of the temporary operation credential and the context decision of the alarm are both executed by a lightweight edge computing gateway, the lightweight edge computing gateway is integrated with a rule engine, the rule engine judges and processes the temporary operation credential and the gas sensor data according to a preset set of logical rules, and outputs an alarm decision instruction to control the execution of the alarm logic.

[0013] A laboratory hazardous gas leakage monitoring system, the system comprising: A tag reading and writing device for automatically reading the identity information of an electronic identity tag arranged on a hazardous gas used article when the hazardous gas used article enters an article operation area; A gas sensor for monitoring the gas concentration in the environment in real time and obtaining the area information where the gas sensor is located; an edge computing gateway in communication connection with the tag reading and writing device and the gas sensor, the edge computing gateway comprising: A credential generation module for generating a temporary operation credential with a predetermined valid duration based on the identity information and the article operation area information; A concentration judgment module for judging whether there is a valid temporary operation credential that completely matches the area information where the gas sensor is located and the detected gas type when the gas concentration detected by any gas sensor exceeds a first concentration threshold value; The alarm decision module is used for determining that the concentration increase is caused by normal operation if there is a valid temporary operation credential matching the area information where the gas sensor is located and the detected gas type, recording data without triggering a first-level early warning; determining that the concentration increase is caused by abnormal leakage if there is no valid temporary operation credential matching the area information where the gas sensor is located and the detected gas type; and triggering the highest-level alarm as soon as the gas concentration detected by any gas sensor is higher than a second concentration threshold, which is higher than the first concentration threshold and is defined as a high-concentration that causes instant harm to human bodies.

[0014] Compared with the prior art, the present application has the following beneficial effects: 1. When the gas sensor detects abnormal concentration, the system first verifies whether there is a matching valid operation credential in the area. If there is a credential, it indicates that the gas is from the current authorized operation, and the system automatically classifies it as a safe event. If there is no matching credential, it is determined to be an abnormal leakage. This mechanism based on operation context and physical location double verification distinguishes between controlled release and accidental leakage in principle, avoids the inherent contradiction between sensitivity and false alarm rate of traditional threshold alarm, and makes it possible to accurately capture early micro-leakage. After triggering the first-level early warning, the system wakes up the adjacent dormant sensor for concentration comparison. By analyzing the concentration gradient relationship between the trigger point and the adjacent point, such as the concentration of the trigger point being significantly higher than that of the adjacent point, the system can infer the location of the leakage source and verify the validity of the early warning. This mechanism does not rely on complex diffusion models, but is based on the physical law of natural decay of gas concentration in space, and realizes preliminary positioning of the leakage source and secondary filtering of false alarms with simple hardware cooperation.

[0015] 2. When the door and window are opened or the strong exhaust is started, the magnetic reed switch marks the airflow mutation event in real time, and the system automatically enters the conservative monitoring mode based on this: suspending the judgment of gas without credentials based on spatial logic, increasing the gradient threshold of near-field verification, and strengthening data recording. This logic switching mechanism triggered based on physical events enables the system to avoid misjudgment in airflow chaos and seamlessly restore high sensitivity mode after the environment stabilizes, ensuring the robustness of decision-making under complex working conditions.

[0016] 3. The system periodically releases standardized detection substances such as trace ethanol to the target sensor. By analyzing the response time and peak value changes, the sensor time delay bias rate and sensitivity decay rate are inverted. This implicit performance indicator is used in real time to dynamically compensate the alarm threshold or adjust the sensor weight. This closed loop of active detection-state inversion-parameter correction enables the system to have self-maintenance capability, ensuring the long-term effectiveness of the core monitoring logic. BRIEF DESCRIPTION OF DRAWINGS

[0017] Fig. 1A flowchart of a laboratory dangerous gas leakage monitoring method and system of the present application; Fig. 2 A graph of a laboratory chloroform leakage concentration gradient ratio R changing with time of the present application; Fig. 3 An interactive flowchart of a laboratory dangerous gas leakage monitoring system of the present application.

[0018] The purposes, functional features and advantages of the present application will be further described with reference to the accompanying drawings in conjunction with embodiments. DETAILED DESCRIPTION

[0019] In order to make the purposes, technical solutions and advantages of the present application more clear, the technical solutions of the present application are described in detail. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should fall within the protection scope of the present application.

[0020] The laboratory dangerous gas leakage monitoring method and system disclosed in the embodiments of the present application mainly have three core parts for collaborative operation in the overall architecture: first, electronic identity tags fixed on dangerous gas related articles and tag reading and writing devices deployed in specific article operation areas, second, a gas sensor network distributedly deployed in the laboratory for real-time monitoring of environmental gas concentration, and third, a lightweight edge computing gateway responsible for data fusion and decision arbitration; the gateway is connected with the aforementioned tag reading and writing devices and gas sensors through wireless or wired communication, and serves as an intelligent hub of the system, performing generation of temporary operation credentials, context matching judgment of gas concentration and operation credentials, and decision logic of final hierarchical alarm.

[0021] In existing laboratory safety monitoring practices, a fundamental challenge is the inability to establish a direct logical association between gas concentration signals in the environment and specific experimental operations. This limitation makes it difficult for the monitoring system to distinguish between controlled gas releases generated by normal experiments and accidental dangerous leaks, making it difficult to achieve a good balance between false alarm rate and missed detection rate. To meet this challenge, the core procedure of the solution of the present invention begins with equipping each item used for hazardous gases, such as high-pressure gas cylinders or high-volatility reagent bottles, with a passive ultra-high frequency wireless radio frequency identification electronic tag, which stores the globally unique identity information of the item. Accordingly, tag reading and writing devices are deployed in all areas where related material operations may be carried out, such as inside fume hoods, dedicated laboratory tables or reagent weighing areas. When an item used for hazardous gases carrying an electronic identity tag is moved into the effective electromagnetic induction field of any item operation area, the area The tag reading and writing device is activated, automatically completing the contactless reading of the identity information of the electronic identity tag, and immediately transmitting the identity information together with its own regional location information to the edge computing gateway; after receiving the data packet, the credential generation module in the edge computing gateway calculates and generates a temporary operation credential with a specific validity period based on a preset calculation model, taking into account the chemical type of the items used for the hazardous gas, the environmental ventilation properties of the current item operation area, and the duration of the established standard experimental procedures. The data structure of the credential contains at least the operation area information and the corresponding gas type as well as a start and end timestamp. The establishment of this procedure ensures that any compliant experimental operation is digitally authorized and recorded by the system at the beginning of the behavior, providing key contextual basis for the subsequent interpretation of gas concentration signals, thereby laying a solid foundation for distinguishing between controlled releases and accidental leaks in principle.

[0022] During the generation of temporary operation vouchers, their established validity period The determination is based on a linear weighted model The result of the calculation is Defined as the benchmark duration to ensure minimum safe operation time, usually set at 60 seconds; To characterize the quantitative coefficient of the inherent risk of the hazardous gas type, its value is normalized to the range of 1 to 10 by referring to the toxicity classification and explosion limit data of each gas in the hazardous chemicals catalog and using the risk assessment matrix. For example, the toxicity of the highly toxic gas hydrogen chloride is 9, non-toxic and non-flammable gas nitrogen is 1; In order to reflect the quantitative coefficient of ventilation efficiency in the article operation area, the value is normalized to a range of 1 to 5 after measuring the average surface wind speed (unit: m / s) in different areas of the laboratory using an anemometer and testing the number of air changes per hour (unit: times / hour) with tracer gas (such as sulfur hexafluoride). 5, ordinary open experimental table is 2; Defined as the sensor performance degradation, through the delay deviation rate and sensitivity attenuation rate The calculated risk correction factor is calculated as follows: ,in and is a pre-set weight coefficient, such as ; are the weight factors of each item, which are determined by iteratively solving with a genetic algorithm in a large-scale simulation environment containing thousands of simulation scenarios (covering different gas types, ventilation conditions of the operating area and the aging degree of the sensor), with minimizing the false alarm rate and the missed alarm rate as the objective function and taking the convenience of operation as the constraint condition to obtain the optimal solution set. For example, in a specific laboratory environment, This procedure ensures that the generation of the validity period of the voucher has a clear quantitative basis and engineering adjustment process.

[0023] Given that authorized operation credentials alone are not sufficient to deal with all security scenarios, the system must establish a rigorous logic that can dynamically judge gas concentration signals to effectively filter out environmental interference while ensuring high sensitivity; to this end, the gas sensors in the system, which can be electrochemical gas sensors or semiconductor gas sensors, are configured to continuously monitor the gas concentration in their microenvironment in real time and record the concentration readings. The concentration judgment module of the gateway processes these data streams in real time, and its built-in rule engine executes a dual-threshold comparison logic: when the concentration value reported by any sensor is When a preset lower first concentration threshold is exceeded for the first time, the system does not trigger an alarm immediately, but instead first searches the credential database to see if there is a temporary operation credential that fully matches the area information of the sensor and the currently detected gas type and whose timestamp is still within the valid window; if such a matching valid credential is retrieved, the alarm decision module will determine that the concentration increase event is caused by normal operation, and will not trigger a level one warning, but will only record and archive it as operation process data; on the contrary, if there is no matching valid credential, the event will be preliminarily determined to be an abnormal leak, and a level one warning will be triggered immediately; however, regardless of whether there is a valid temporary operation credential, once the gas concentration detected by any gas sensor exceeds Above a second concentration threshold defined as a high-risk concentration that can cause immediate harm to human body, the system will bypass all contextual judgment logic and directly trigger the highest level of alarm to ensure fast response speed in extreme dangerous situations; through this set of judgment mechanism based on the dynamic coupling of operation credentials and double concentration thresholds, the system can accurately identify early-stage leakage in most scenarios, while exempting benign concentration fluctuations caused by normal experimental operations, improving the accuracy of the alarm and the robustness of the system's decision-making.

[0024] Furthermore, after the first-level warning is triggered, in order to further confirm the authenticity of the leakage event and preliminarily locate the source of leakage, the system is configured to perform a set of procedures based on adjacent verification, which aims to deal with false alarms caused by single sensor failure or local airflow vortex; while sending alarm signals to the area triggering the first-level warning, the edge computing gateway will send a forced wake-up instruction to at least one adjacent gas sensor in a dormant or low-power state around the sensor; the awakened adjacent sensor will immediately perform a high-precision concentration measurement after receiving the instruction, and return the reading to the gateway; the system will then compare the concentration reading of the initial alarm sensor with the concentration reading of the awakened adjacent sensor When the ratio of the two is greater than a concentration gradient threshold preset based on the gas diffusion model, it can be confirmed that the source of abnormal leakage is near the initial alarm sensor, and the validity of the first-level warning is finally confirmed; as a supplementary constraint of this procedure, if in the comparison, the concentration readings of the two sensors are both below a preset validity check threshold or the spatial distribution of the concentration readings of the two sensors does not meet the established spatial gradient characteristics, for example, the concentration of the adjacent point is higher, the system will temporarily suspend the alarm and enter a continuous observation mode to monitor the area more closely; this near-field gradient verification mechanism takes advantage of the physical law of natural diffusion and decay of gas in space, adding an effective false alarm filtering and leakage positioning means to the system with lower communication and computing overhead.

[0025] The concentration gradient threshold used in the near-field gradient verification is set through a strict offline calibration procedure, which covers the target operating area where the gas sensor is deployed, accurately releases a specific kind of target gas at a known and stable rate (e.g. 0.01 liters / minute) through a micro-flow controller, and synchronously records the gas concentration readings at the initial alarm sensor and at least one adjacent dormant sensor; this process is repeated under multiple typical airflow organization forms to collect a large number of Ratio data; through statistical analysis of these experimental data, including drawing ROC curve to evaluate the true positive rate and false positive rate under different thresholds, finally determine the optimal ratio that can distinguish between local leakage and long-distance diffusion, this ratio is set as the concentration gradient threshold for the system to use for the first level of early warning follow-up verification, for example, in the chloroform leakage scenario, the threshold is 3.5; when the macroscopic airflow physical state signal indicates that mutation occurs, the system dynamically adjusts the gradient threshold to 1.5 times the original value, this adjustment is maintained until the airflow returns to a stable state, which is determined by the physical state sensor signal being unchanged for 30 seconds continuously, this dynamic adjustment strategy ensures the decision robustness of the system under environmental disturbance.

[0026] Considering that sudden opening or closing of doors and windows or starting and stopping of strong exhaust systems in the laboratory environment will cause severe disturbance of macroscopic airflow, which is enough to make the gas concentration distribution chaotic and disordered in a short time, thereby interfering with the aforementioned judgment logic based on stable diffusion model; for this reason, the system also integrates a set of airflow mutation adaptive strategy, by installing physical state sensors that can instantly reflect their opening and closing state at key positions in the laboratory, such as door frames, window frames and ventilation cabinet regulating valves, for example, magnetic reed switches, to obtain physical state signals indicating that the macroscopic airflow state in the laboratory has mutated in real time; when the system receives such signals indicating that the macroscopic airflow is in a state of mutation, its rule engine will automatically suspend the execution of the core logic that triggers the first level of early warning if there is no matching credentials within a certain period of time, for example, thirty to ninety seconds, because during this period, the gas in a region is likely to come from a normal operation far away that is wrapped by turbulence; under this conservative monitoring mode, the system will correspondingly raise the gradient threshold for near-field verification, and intensify the recording of each point concentration data, after the physical state signal returns to a stable state, the system automatically and seamlessly switches back to the original high-sensitivity monitoring mode; this dynamic degradation and recovery mechanism of decision logic triggered by physical events ensures the decision reliability of the system under complex conditions, and avoids false positives caused by environmental changes.

[0027] To maintain the monitoring effectiveness and data reliability of the entire monitoring system for a long period of time, the system also has a closed-loop sensor self-diagnosis and compensation correction mechanism to counter the inevitable performance degradation of sensors due to long-term operation; within a certain system maintenance period, for example, a predetermined maintenance window once a quarter, the system can automatically or under the instruction of the administrator, actively release a standardized dose of harmless probe material near a target gas sensor through a miniature piezoelectric pump or an electrically controlled valve; the system then accurately records and obtains the complete dynamic response data of the target sensor to this standard probe material, key indicators include the response time from receiving the probe material signal to the peak reading , and the final peak height ; By comparing these real-time measured data with the reference response data recorded at the sensor factory or last calibration, the system can calculate the current performance state parameters of the sensor, specifically represented as time delay deviation rate and sensitivity attenuation rate ; these parameters are then input into a preset compensation algorithm, which dynamically compensates the first concentration threshold used by the sensor for judgment or adjusts the weight factor it occupies in the near-field verification multi-point data fusion algorithm according to the attenuation rate; this closed-loop workflow of active detection-state inversion-parameter correction enables the system to have self-maintenance capability, actively identify and quantify the performance drift of the core sensor element, and eliminate its adverse effects on monitoring accuracy in the form of algorithm compensation, thereby ensuring the long-term stability and effectiveness of the core monitoring logic of the entire system.

[0028] The implementation of the sensor self-diagnosis and compensation correction mechanism relies on the system to accurately release a standardized dose of detection material, such as 50ppm ethanol gas, near the target gas sensor through a micro piezoelectric pump every quarter, for example, with a release amount of 0.5ml and a duration of 5 seconds; this dose does not affect the laboratory environment and has no long-term damage to the sensor; the system accurately records the response time of the sensor from receiving the detection material signal to the peak value of the reading and the final peak height ; these data are then compared with the reference response time and the reference peak height recorded by the sensor at the initial installation, to calculate the time delay deviation rate and the sensitivity attenuation rate ; according to the calculated performance state, the system compensates and adjusts the first concentration threshold of the sensor, with the new threshold , where is the compensation coefficient, whose value is determined by experiment to ensure that when the sensor sensitivity attenuates , the threshold is correspondingly reduced to maintain detection capability, for example, when , if the threshold needs to be reduced , then , in addition, the weight factor of the sensor in the near-field verification is also dynamically adjusted according to its performance attenuation rate, specifically , where is the weight attenuation coefficient, for example , this closed-loop mechanism realizes the continuous calibration of sensor performance through automated procedures, ensuring the long-term stability and precision of the monitoring system.

[0029] Finally, to ensure that the system can still reliably deliver critical alarm information in a complex electromagnetic environment, the system is also configured with passive sensing capability for wireless communication channel quality; while performing routine communication tasks such as waking up neighboring sensors and returning data, the system continuously records the complete communication delay of each neighboring gas sensor node from receiving the wake-up instruction to successfully returning the concentration measurement value; by analyzing the deviation of this communication delay from a preset baseline delay within a given observation period, its trend over time, and the jitter amplitude in the short term, the system can calculate the wireless communication interference intensity in the area where the neighboring sensor is located according to a set of preset decision rules, and generate a dynamic physical interference zone map accordingly; when the system needs to transmit high-priority first-level warning or highest-level alarm information, its routing logic will actively query the marker map, preferentially selecting nodes in non-interference areas for information relay and routing or directly instructing gas sensor nodes located in marked physical interference zones to actively increase the number of information retransmissions when sending alarm information until an upper-layer confirmation receipt is received; this mechanism converts the byproduct data of communication delay into effective insights into the network environment state and applies it to the reliability protection strategy for critical information transmission, further enhancing the overall survival and execution capability of the system in harsh environments.

[0030] Embodiment 1: This embodiment aims to deploy and operate the aforementioned technical solutions in a specific application scenario. Specifically, in a high-throughput drug research laboratory environment, multiple fume hoods are deployed side by side along the wall, and different researchers may be conducting different chemical experiments in adjacent fume hoods at the same time. This scenario poses a serious challenge to the gas monitoring system, namely how to effectively capture weak and unexpected dangerous gas leaks without interfering with normal and authorized experiments; at a certain moment, researcher A completes the scheduled operation in fume hood A and leaves a bottle of chloroform reagent with an electronic identity tag in the cabinet, and the scheduled temporary operation credential has expired on time after the operation is completed, while the valve of the reagent bottle has a small flaw and starts to leak chloroform gas at a slow rate; at the same time, researcher B is using a bottle of ether reagent with an electronic identity tag in the adjacent fume hood B for a compliant operation, and the system has generated an effective ether temporary operation credential for him in the B fume hood area.

[0031] With the passage of time, the chloroform gas concentration in the A fume hood slowly accumulates and eventually exceeds the first concentration threshold preset by the system, at which point the gas sensor deployed in the A fume hood reports this over-threshold event to the edge computing gateway; after the concentration judgment module of the gateway receives this signal, it immediately starts the context decision procedure, which first searches the valid credential database to query whether there is a valid temporary operation credential that matches the location of the A fume hood and the type of chloroform gas, and the query result is no; based on this judgment of no-credential over-threshold, the alarm decision module qualifies this event as an abnormal leakage and triggers a level-1 early warning, and the execution of this procedure first converts a physical concentration signal without distinction into a logical event with an unauthorized attribute.

[0032] At the moment of triggering the level-1 early warning, the system cooperatively starts the near-field gradient verification procedure, and the edge computing gateway sends a wake-up instruction to the adjacent gas sensors around the A fume hood, including those located between the A and B fume hoods. After the adjacent sensors are woken up, their readings are affected by both the trace amount of chloroform leaked from the A fume hood and the normal operation of the B fume hood releasing ether; however, when the system calculates the concentration gradient, it is based on the type of gas, and it compares the chloroform concentration reading of the A fume hood sensor with the chloroform concentration reading of the adjacent sensor Since the leakage source is inside the A fume hood, the ratio of the two is much larger than the preset concentration gradient threshold, thereby confirming that there is a real, local leakage source in the A fume hood; at the same time, the ether concentration increase event of the B fume hood is judged by the system as normal operation from beginning to end without any interference signal, because there is a completely matched valid temporary operation credential; this dynamic coupling of the temporary operation credential mechanism and the near-field gradient verification mechanism enables the system not only to distinguish between authorized and unauthorized gas release, but also to make a secondary confirmation of the authenticity and source of the unauthorized signal through the concentration distribution characteristics of the physical space, thereby resolving the inherent contradiction between high sensitivity and high false alarm rate in traditional monitoring technology.

[0033] Finally, the system does not issue a full-area or indiscriminate alarm, but a red warning light is lit on the operation panel of the A fume hood, and an accurate alarm information is sent to the terminal of the laboratory safety administrator through the network, which clearly indicates the specific location of the leakage, the gas type and the event qualification, so that the safety personnel can intervene quickly and accurately, and the normal experimental workflow of the B fume hood is not affected at all, and the overall operation efficiency of the laboratory is maintained; The results show that the technical scheme of the present application introduces the logical dimension of operation authorization, and converts the traditional gas concentration monitoring problem from a pure physical signal measurement problem to an abnormal behavior audit problem based on a four-tuple of location-substance-time-authorization, so that the dangerous signals and background noise that are difficult to distinguish or mixed together are effectively decoupled and identified under the new judgment framework.

[0034] Example 2: In order to objectively verify the actual effectiveness of the technical scheme of the present application in distinguishing real micro-leakage from authorized operation interference under complex working conditions, the verification test described in this embodiment is designed and performed; The purpose of the test is to quantitatively evaluate the accuracy and selectivity of the alarm decision of the system in the parallel authorized and unauthorized gas release event scenarios with spatial proximity, which directly corresponds to the most challenging false alarm and missed alarm risk points in actual application.

[0035] The test is carried out in a closed environment cabin with a volume of 15 cubic meters, which is internally provided with two simulated article operation areas with a distance of 1.5 meters, marked as area A and area B; Each area is provided with a tag reading and writing device and a high-precision electrochemical gas sensor for a specific gas species, respectively denoted as sensor and sensor , and a same type of adjacent gas sensor is arranged at the geometric center point of the two areas for gradient verification; The whole system is controlled by a lightweight edge computing gateway; In order to ensure the engineering rationality of the test conditions, the setting of the key parameters follows a strict decision logic chain, taking the determination of the first concentration threshold as an example, which needs to balance the sensitivity of early detection and the stability of resisting environmental noise, and its value rule is defined as a weighted sum of the sensor inherent baseline noise and the target gas short-term exposure limit, which is 10 ppm for chloroform gas in this test, and the sensor baseline noise is 0.5 ppm, and a sensitivity factor of 0.15 is set, so the first concentration threshold is determined as Likewise, the concentration gradient threshold for near-field gradient validation, which is set to balance the accuracy of the leak source localization and the reliability of the validation process, is calibrated offline, and is confirmed to be greater than 5.0 under the current sensor layout, so a concentration gradient threshold with a redundant fault tolerance space of 3.5 is set.

[0036] The test process is as follows: at , the background in the environmental cabin is clean; at , an ether reagent bottle configured with an electronic identity tag is placed in area B, the system immediately reads the tag and generates a temporary operation voucher for ether in area B, which is valid for 5 minutes, at the same time, the mass flow controller releases ether to area B at a stable rate, so that the concentration of ether reaches and maintains about 50 ppm at the sensor ; at , a separate mass flow controller is started in area A to simulate a voucher-free and slow chloroform leak, and the leak rate is precisely controlled to make the chloroform concentration reach the threshold of 2.0 ppm after about 120 seconds; during the whole process, the system continuously records the readings of all sensors and the internal decision state of the edge computing gateway, and at , the system captures a set of key state data with typical representative significance, as shown in Table 1.

[0037] Table 1: System key state data table at seconds.

[0038] As shown in Table 1, at seconds, the chloroform reading of sensor first exceeds the first concentration threshold of 2.0 ppm, and the system immediately determines this event as a first-level warning and triggers near-field gradient validation because no matching temporary operation voucher is found; at this moment, the ether event in area B with a concentration as high as 51.3 ppm is correctly identified by the system as normal operation because there is a valid voucher, and no alarm interference is generated; in other words, the system calculates the concentration gradient ratio by comparing the chloroform readings of and , and the value of 5.25 is definitely greater than the preset concentration gradient threshold of 3.5, thereby confirming the locality and authenticity of the leak source at the physical level, and finally outputting the accurate alarm decision for area A; the record data of this process show that the inherent cooperative mechanism of the present scheme, i.e. the temporary operation voucher mechanism, first removes the interference of authorized operation from the logical level, so that the system can focus on the voucher-free signal, and then the near-field gradient validation mechanism provides spatial dimension evidence for the voucher-free signal from the physical level.

[0039] Embodiment 3: This embodiment is combined with Figs. 1 to 3 A laboratory hazardous gas leakage monitoring method and system are implemented. As shown in Fig. 1 In the field perception and interaction layer, the hazardous gas articles are configured with electronic identity tags, and the tag reading and writing devices are deployed in the operation area; when the hazardous gas articles enter the operation area, the tag reading and writing devices automatically read the identity information of the electronic identity tags and transmit them to the edge computing gateway; the gas sensor monitors the environmental gas concentration in real time and obtains the location information thereof, and transmits the concentration reading to the edge computing gateway; in the edge computing and decision layer, the concentration data acquisition module is responsible for obtaining the sensor reading and the location; the certificate generation module generates a time-limited operation certificate according to the article identity information and the operation area information; the existence of a matching valid certificate module verifies the context of the operation authorization; if yes, it is determined as normal operation, the data is recorded, and no warning is given; if no, a first-level warning is triggered, and it is determined as abnormal leakage; at the same time, the system judges that the concentration > the first threshold value; whether there is a valid certificate or not, once the concentration > the second threshold value (high concentration), the system will immediately trigger the highest level alarm, realizing high concentration, immediate response; in addition, the system also includes a closed-loop self-diagnosis and compensation correction mechanism, which actively detects and corrects the performance decay of the sensor, and dynamically compensates the first concentration threshold value / adjusts the weight; in the system response and adaptive layer, after triggering the first-level warning, the system enters the near-field gradient verification link, wakes up the adjacent sensor to compare the concentration and locate the source; then, the gradient > threshold value is judged to verify the authenticity and locality of the leakage; if yes, the leakage source is confirmed, the alarm is locked, and the accurate position is sent; if no, the alarm is temporarily suspended and enters the continuous observation mode.

[0040] As shown in Fig. 2 , the concentration gradient ratio R changes with time (seconds), where the horizontal axis represents time (from 160 seconds to 200 seconds), and the vertical axis represents the concentration gradient ratio. The solid line in the figure represents the real-time measurement value of the concentration gradient ratio R, and the dashed line represents the preset gradient threshold value (3.5). During the test, the chloroform leakage started at T=60 seconds and reached the first concentration threshold value of 2.0 ppm at about 120 seconds (i.e. at T=180 seconds). At T=185 seconds, the chloroform reading of sensor SA reached 2.1 ppm, the system triggered the near-field gradient verification, and the concentration gradient ratio R was calculated as 5.25. As can be seen from the figure, the concentration gradient ratio R gradually increased and exceeded the gradient threshold value 3.5 from about 170 seconds, and reached 5.25 at T=185 seconds, which is greater than the preset threshold value, thereby confirming the locality and authenticity of the leakage source.

[0041] As shown in Fig. 3, first, the operator moves the item with the RFID tag into the operating area, the RFID tag sends the ID (gas type / chloroform) to the reader, the reader identifies the area B + chloroform tag, and then transmits the information to the edge gateway, the edge gateway calculates the validity period (according to the gas cylinder type / ventilation level), generates a voucher (valid for 300s), and the voucher information can be synchronized to the database, at the same time, the gas sensor continuously and in parallel monitors the gas concentration in the environment, when detecting the concentration data (area B: chloroform 2.1ppm), it transmits it to the edge gateway, which checks the voucher (area B + chloroform); if the voucher is valid, the operation log is recorded to the database, if the voucher is invalid, the edge gateway issues a first-level warning (code E102) and requests a proximity verification, when an abnormal handling branch occurs, such as tag reading failure or voucher generation timeout, etc.

[0042] In a semiconductor process laboratory where the monitoring system of the present application has been deployed and has been running continuously for more than twelve months, one of the gas sensors located in the etching area has experienced irreversible chemical aging of its sensing element due to long-term exposure to trace amounts of main product of corrosion environment, resulting in a slower response speed and a decrease in sensitivity. This gradual performance degradation poses a hidden risk to the reliability of the monitoring system, and if not corrected, it may result in a de facto increase in the detection threshold for slow leaks, or in the output of distorted spatiotemporal data due to response delay when performing near-field gradient verification, thereby causing false negatives or false positives.

[0043] To address such systematic risks caused by hardware aging, the closed-loop self-diagnostic mechanism built into the present application is activated at a predetermined maintenance cycle for the sensor; the system instructs a calibration unit located near the sensor to accurately release a standardized dose of the probe substance, and begins recording the response curve of the sensor; the system extracts two key performance indicators from the response curve, namely the response time from receiving the signal to the peak value of the reading , and the peak height ; the edge computing gateway then retrieves the baseline response time and the baseline peak height from its stored initial calibration database for the sensor, and performs a quantitative evaluation of the performance state according to the following procedures: first, calculate the time delay deviation rate ; second, calculate the sensitivity decay rate ; these two dimensionless parameters objectively depict the degree of deviation of the current performance of the sensor from its ideal state.

[0044] Based on the calculated performance state parameters, the compensation algorithm in the edge computing gateway is triggered to adaptively adjust the monitoring logic of the sensor; this compensation algorithm includes a set of pre-set rules, first, if the sensitivity decay rate If the sensor's performance exceeds the preset threshold, i.e. 20%, the system will accordingly lower the weight factor of this sensor in future participation in any multi-point collaborative analysis, such as near-field gradient verification, to reduce its interference with the overall decision-making; secondly, the first concentration threshold of the sensor is dynamically adjusted for compensation, and the basis for adjustment is a correlation compensation function with the sensitivity decay rate , i.e. the new threshold is set to the first concentration threshold , wherein is a pre-calibrated compensation coefficient, which functions to moderately lower the alarm threshold of the aging sensor to offset the loss of detection capability caused by the decline in sensitivity; this complete closed loop of active detection, state inversion, and parameter correction enables the system to autonomously identify and adapt to the performance evolution of its physical layer hardware, thereby maintaining the effectiveness of the monitoring logic over a long period.

[0045] Further, the aforementioned preset calculation model for generating a temporary operation credential can also integrate the health status information of the sensor for risk adjustment in its specific implementation; specifically, the predetermined validity period of the temporary operation credential may be determined by a linear weighting model, i.e. ; wherein, and are quantitative coefficients representing the gas risk level and the ventilation condition of the operation area, respectively, is a risk correction factor calculated according to the health status of the area sensor, i.e. according to and , and , , are weight factors for each item; in this way, the worse the health status of the sensors in an area, the larger the risk correction factor , and the shorter the validity period of the generated temporary operation credential, which prompts the operation to be carried out under stricter time monitoring, thereby compensating for the hardware aging effect in another dimension of risk management.

[0046] Example 5: To ensure the accuracy of the baseline of all subsequent monitoring and diagnostic functions after the first deployment of the monitoring system of the application in a new laboratory environment or after replacing any of the gas sensors therein, the system is set to first perform a set of pre-deployment calibration procedures; in this procedure, the system administrator places the system in a specific debugging mode through the maintenance interface and specifies the newly created sensors that need to be initialized one by one; for each specified target sensor, the edge computing gateway instructs its adjacent calibration unit to release a standardized dose of the detection substance, and records the response time and peak height of the new sensor to this standard excitation in full. These two initial measurement values are then solidified and stored in the gateway database, and are associated with the unique identifier of the sensor, respectively as the baseline response time for subsequent performance degradation calculations and the baseline peak height The execution of this procedure establishes a digital archive of the initial health status of each individual sensing element, thereby providing an accurate and non-arbitrary original baseline for the subsequent closed-loop self-diagnosis mechanism.

[0047] Further, to ensure that the calculation model used to generate the effective duration of the temporary operation credentials has objective basis, the various quantitative coefficients relied upon by the model are systematically constructed and filled through a series of offline calibration experiments before system deployment; specifically, the coefficient used to represent the risk level of the gas is a quantitative value calculated through a risk assessment matrix after releasing each target gas in a controlled standard environment and measuring its diffusion rate, dispersion trend and other physicochemical parameters; while the coefficient used to represent the nature of the article operation area is a risk index assigned after standardized testing of the air exchange efficiency and airflow organization form of different types of fume hoods or isolated operation boxes; finally, the various weight factors , and in the model are determined through optimization algorithm iteration in a large-scale simulation containing thousands of simulated scenarios covering different gases, different areas and different sensor aging degrees, in order to seek the optimal balance point between safety redundancy and operation convenience. This entire offline calibration and data filling procedure ensures that each parameter of the calculation model has its physical or statistical source as support, thereby guaranteeing the scientificity and consistency of the system's decision-making in actual application.

[0048] Example 6: To ensure the accuracy of the baseline for all subsequent monitoring and diagnostic functions, after the monitoring system of the present application is first deployed in a new laboratory environment or any of its gas sensors are replaced, the system is set to first perform a set of pre-deployment calibration procedures; in this procedure, the system administrator, through the maintenance interface, puts the system into a specific debugging mode and specifies the newly created sensors that need to be initialized one by one; for each designated target sensor, the edge computing gateway instructs its adjacent calibration unit to release a standardized dose of the detection substance, and records the response time and peak height of the new sensor to this standard excitation in full, these two initial measurement values are then solidified and stored in the gateway's database, and are associated with the unique identifier of the sensor, respectively as the baseline response time for subsequent performance degradation calculations and the baseline peak height The execution of this procedure establishes a digital archive of the initial health status of each individual sensing element, providing a non-arbitrary original baseline for the subsequent closed-loop self-diagnosis mechanism to accurately compare.

[0049] Further, to ensure that the calculation model used to generate the effective duration of the temporary operation credentials has objective basis, the various quantitative coefficients it relies on are systematically constructed and filled before system deployment through a series of offline calibration experiments; Specifically, the coefficient used to represent the risk level of the gas is a quantitative value calculated through a risk assessment matrix after releasing each target gas in a controlled standard environment and measuring its diffusion rate, dispersion trend and other physicochemical parameters; while the coefficient used to represent the nature of the item operating area is a risk index assigned after standardized testing of the air exchange efficiency and airflow organization form of different types of fume hoods or isolated operation boxes; finally, the weight factors , and in the model are determined through optimization algorithm iteration in a large-scale simulation containing thousands of simulated scenarios covering different gases, different areas and different sensor aging degrees, to seek the optimal balance point between safety redundancy and operation convenience; this entire offline calibration and data filling procedure ensures that each parameter of the calculation model has its physical or statistical source as support, thus guaranteeing the scientificity and consistency of the system's decision-making in actual application, all of which belong to the extended implementation manners known to those skilled in the art.

[0050] It is obvious to those skilled in the art that the present application is not limited to the details of the above exemplary embodiments, and can be implemented in other specific forms without departing from the spirit or essential characteristics of the present application.

[0051] It should be pointed out finally that the above embodiments are only used to illustrate the technical solutions of the present application but not to limit the present application. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or equivalently replaced without departing from the spirit and scope of the present application.

Claims

1. A laboratory hazardous gas leakage monitoring method, characterized in that: The method comprises the following steps: Equip items used with hazardous gases in the laboratory with electronic identity tags, and deploy tag reading and writing devices in the operation areas of items used with hazardous gases; When an item used for hazardous gas enters the item operation area, the tag reading and writing device automatically reads the identity information of the electronic identity tag and generates a temporary operation certificate with a predetermined validity period based on the identity information and the item operation area information; The gas concentration in the environment is monitored in real time by gas sensors, and information about the area where the gas sensors are located is obtained; when the gas concentration detected by any gas sensor exceeds a first concentration threshold, a determination is made as to whether there is a valid temporary operation certificate that fully matches the area information of the gas sensor and the type of gas detected; If there is a valid temporary operation certificate that completely matches the gas sensor location and the type of gas detected, the concentration increase will be determined to be caused by normal operation, and the first-level warning will not be triggered, and the data will be recorded; If there is no valid temporary operation certificate that fully matches the gas sensor location and the type of gas detected, the concentration increase will be considered an abnormal leak and trigger a level 1 warning. Regardless of whether there is a valid temporary operation certificate, once the gas concentration detected by any gas sensor is higher than the second concentration threshold, the second concentration threshold is higher than the first concentration threshold and is defined as a high-risk concentration that causes immediate harm to the human body, the highest level alarm will be triggered immediately.

2. A laboratory hazardous gas leakage monitoring method according to claim 1, characterized in that: After triggering the first-level warning, the following steps are also included: sending a wake-up command to at least one adjacent gas sensor in a dormant state around the gas sensor that triggered the first-level warning; the awakened adjacent gas sensor performs concentration measurement and returns a concentration reading; compares the concentration reading of the gas sensor that triggered the first-level warning with the concentration reading of the awakened adjacent gas sensor, and when the ratio of the two is greater than a preset concentration gradient threshold, determines the source location of the abnormal leakage or the reliability of the first-level warning.

3. A laboratory hazardous gas leakage monitoring method according to claim 2, characterized in that: After triggering the first-level warning, before sending a wake-up command to the adjacent gas sensor, the following steps are also included: real-time acquisition of a physical status signal indicating that the macroscopic airflow state of the laboratory has undergone a sudden change; when the physical status signal indicates that the macroscopic airflow is in a sudden change state, within a predetermined time period, the execution of the logic of triggering the first-level warning if there is no valid temporary operation certificate that completely matches the area information of the gas sensor and the detected gas type is suspended.

4. A laboratory hazardous gas leakage monitoring method according to claim 1, characterized in that: The electronic identity tag is a passive ultra-high frequency radio frequency identification electronic tag.

5. A laboratory hazardous gas leakage monitoring method according to claim 2, characterized in that: When comparing the concentration reading of the gas sensor that triggered the first-level warning with the concentration reading of the awakened adjacent gas sensor, if the concentration readings of both are lower than the preset verification threshold or the spatial distribution of the concentration readings of the two does not meet the established spatial gradient characteristics, the system temporarily suspends the alarm and enters the continuous observation mode.

6. A laboratory hazardous gas leakage monitoring method according to claim 1, characterized in that: The following steps are also included: During a predetermined maintenance cycle, the instruction system proactively releases a standardized dose of detection material near the target gas sensor; obtains the response data of the target gas sensor to the detection material, the response data including the response time from receipt of the detection material signal to the reading reaching a peak value, and the peak height; calculates the current performance status of the target gas sensor by comparing it with the pre-stored benchmark response data, the performance status being the time delay deviation rate and the sensitivity attenuation rate; based on the performance status, uses a preset compensation algorithm to make compensatory adjustments to the first concentration threshold used by the target gas sensor to determine whether there is a valid temporary operation credential, or adjusts its weight factor in near-field verification.

7. A laboratory hazardous gas leakage monitoring method according to claim 2, characterized in that: The method also includes: recording the communication delay of the adjacent gas sensor from receiving the wake-up command to returning the concentration measurement value; calculating the wireless communication interference intensity of the area where the adjacent gas sensor is located according to the deviation, change trend and change amplitude of the communication delay from the preset baseline delay during a predetermined observation period, and generating a physical interference area mark based on the interference intensity; when transmitting the first-level warning or the highest-level alarm information, according to the physical interference area mark, giving priority to selecting nodes in the non-interference area for information routing or instructing the gas sensor located in the physical interference area to increase the number of information retransmissions.

8. A laboratory hazardous gas leakage monitoring method according to claim 1, characterized in that: The generation of temporary operation credentials and the contextual judgment of alarms are both performed by the lightweight edge computing gateway. The lightweight edge computing gateway is integrated with a rule engine. The rule engine judges and processes temporary operation credentials and gas sensor data based on a preset set of logical rules, and outputs alarm decision instructions to control the execution of the alarm logic.

9. A laboratory hazardous gas leakage monitoring system, characterized in that: The system comprises: A tag reading and writing device is used to automatically read the identity information of the electronic identity tag configured on the hazardous gas item when the hazardous gas item enters the item operation area; The gas sensor is used to monitor the gas concentration in the environment in real time and obtain information about the area where the gas sensor is located. The edge computing gateway communicates with the tag reading and writing device and the gas sensor. The edge computing gateway includes: A credential generation module is used to generate a temporary operation credential with a predetermined validity period based on the identity information and the item operation area information; a concentration determination module, configured to determine whether there is a valid temporary operation certificate that fully matches the area information of the gas sensor and the type of gas detected when the gas concentration detected by any gas sensor exceeds a first concentration threshold; The alarm decision module is used to determine that the concentration increase is caused by normal operation if there is a valid temporary operation certificate that completely matches the information of the area where the gas sensor is located and the type of gas detected, and not trigger the first-level warning, and record the data; if there is no valid temporary operation certificate that completely matches the information of the area where the gas sensor is located and the type of gas detected, then the concentration increase is determined to be an abnormal leakage and trigger the first-level warning; and regardless of whether there is a valid temporary operation certificate, once the gas concentration detected by any gas sensor is higher than the second concentration threshold, the second concentration threshold is higher than the first concentration threshold and is defined as a high-risk concentration that causes immediate harm to the human body, the highest level alarm is immediately triggered.

Images