Toxic gas emission monitoring method and system for electronic component workshop

Abstract

The invention relates to the technical field of gas monitoring, in particular to a toxic gas emission monitoring method and system for an electronic component workshop, and the method comprises the steps: constructing a hierarchical monitoring network in the electronic component workshop; the fixed monitoring nodes and the key area monitoring nodes are used for collecting poisonous gas concentration data and environment airflow velocity data of the positions where the fixed monitoring nodes and the key area monitoring nodes are located, meanwhile, the mobile inspection nodes are used for collecting poisonous gas concentration data and real-time space coordinate data on a moving track, and the data are packaged into a monitoring data packet to be transmitted to the central processing platform. Then preprocessing and standardization processing are carried out, a concentration distribution cloud picture of a workshop is dynamically generated based on standardized data, and when the concentration data exceed a preset concentration absolute value threshold value, traceability analysis is carried out to determine the position of a leakage source; and triggering trend early warning based on the concentration change trend, triggering grading alarm based on the concentration absolute value and / or the traceability analysis result, and performing rapid traceability and early warning on the toxic gas in the electronic component workshop.

Description

Technical Field

[0001] This invention relates to the field of gas monitoring technology, and in particular to a method and system for monitoring toxic gas emissions in an electronic component workshop. Background Technology

[0002] In electronic component manufacturing workshops, especially in critical processes such as etching, doping, chemical vapor deposition (CVD), and photoresist removal, large quantities of highly toxic, flammable, and explosive gases such as phosphine (PH3), arsine (AsH3), silane (SiH4), and chlorine (Cl2) are used and emitted. Even at very low concentrations, these gases pose a serious threat to the health of operators and present significant safety hazards.

[0003] Currently, the industry commonly uses a technology that involves installing a limited number of gas concentration sensors at fixed locations in the workshop for timed monitoring of ambient air quality. This traditional method typically only reports "concentration exceeding the standard at a certain location," but it cannot quickly and accurately pinpoint the exact location of the leak, making emergency response and troubleshooting difficult and prolonging the risk window. It also fails to provide rapid source tracing and early warning of toxic gases in electronic component workshops. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for monitoring the emission of toxic gases in electronic component workshops, enabling rapid source tracing and early warning of toxic gases in these workshops.

[0005] To achieve the above objectives, in a first aspect, the present invention provides a method for monitoring the emission of toxic gases in an electronic component workshop, comprising the following steps: A hierarchical monitoring network is constructed within the electronic component workshop. The hierarchical monitoring network includes multiple fixed monitoring nodes deployed in the workshop space and key process equipment, multiple key area monitoring nodes deployed in the gas accumulation area, and at least one mobile inspection node. The fixed monitoring nodes and key area monitoring nodes are used to collect data on the concentration of toxic gases and the speed of ambient airflow at their locations. At the same time, the mobile inspection nodes are used to collect data on the concentration of toxic gases and real-time spatial coordinates along the movement trajectory. All data is then encapsulated into monitoring data packets and transmitted to the central processing platform. The received monitoring data packets are preprocessed and standardized, and a concentration distribution cloud map of the workshop is dynamically generated based on the standardized data. When the concentration data exceeds the preset absolute value threshold, a source tracing analysis is performed to determine the location of the leak source. Trend warnings are triggered based on concentration change trends, and tiered alarms are triggered based on absolute concentration values ​​and / or source tracing analysis results.

[0006] The method further includes: When a high-level alarm is triggered, a gas diffusion prediction path is generated based on the location of the leak source and real-time ambient airflow velocity data.

[0007] The method further includes, after establishing a hierarchical monitoring network within the electronic component workshop: After all nodes are powered on and complete their hardware self-tests, the central processing platform instructs all nodes to collect the ambient background concentration under a preset leak-free state and store the normal range benchmark parameters of each node. The mobile inspection node is instructed to move to the designated fixed monitoring node, and the gas concentration data collected synchronously by the two nodes are compared to perform cross-validation of the sensor data.

[0008] After the central processing platform receives the monitoring data packet, the method further includes: By parsing the node's unique identifier and node type code, and matching the metadata in the preset node information database, the data source is identified and classified; the data is correlated in time and space by timestamp and node location information. The monitoring data packet includes the node's unique identifier, node type code, timestamp, sensor data payload, and node status word.

[0009] The process includes preprocessing and standardizing the received monitoring data packets, and dynamically generating a concentration distribution cloud map of the workshop based on the standardized data, including: Statistical outlier filtering and sliding window smoothing are applied to toxic gas concentration data. Using pre-stored sensor calibration parameters, the background concentration data of the environment is normalized in real time to obtain standardized concentration values; Unify the position information of all nodes to the same three-dimensional coordinate system; The concentration distribution cloud map of the workshop is dynamically generated based on the standardized data.

[0010] Among them, the dynamic generation of concentration distribution cloud maps in the workshop based on standardized data includes: A spatial interpolation algorithm was used, with real-time collected environmental airflow velocity data as an auxiliary variable, to calculate the standardized toxic gas concentration data and generate a concentration distribution cloud map of the workshop.

[0011] When the concentration data exceeds a preset absolute concentration threshold, a source tracing analysis is performed to determine the location of the leak source, including: Calculate the concentration gradient field of the concentration distribution cloud map; A dynamic airflow field model is constructed based on the environmental airflow velocity data collected by the fixed monitoring nodes. Starting from the high-concentration region, reverse particle tracking simulation is performed in the dynamic airflow field model; Based on the convergence of particle trajectories, a source probability heatmap representing the probability of a leakage source is generated to determine the location of the leakage source.

[0012] Among them, trend warnings are triggered based on concentration change trends, and tiered alarms are triggered based on absolute concentration values ​​and / or source tracing analysis results, including: Based on the dynamic data buffer corresponding to each monitoring node, the trend slope of the concentration-time series within the dynamic data buffer is calculated; A trend warning is triggered when the trend slope exceeds the trend warning threshold and the current concentration is lower than the first threshold. A tiered alarm is triggered based on the absolute concentration value and / or the source tracing analysis results.

[0013] Among them, tiered alarms triggered based on absolute concentration values ​​and / or source tracing analysis results include: A Level 1 warning is triggered when a trend warning is triggered or the concentration exceeds the first threshold but the source analysis results are unclear. A level 2 warning is triggered when the concentration exceeds the second threshold or when the concentration continues to be out of control after a level 1 warning.

[0014] Secondly, the present invention provides a toxic gas emission monitoring system for electronic component workshops, applied to the toxic gas emission monitoring method for electronic component workshops provided in the first aspect. The toxic gas emission monitoring system for electronic component workshops includes a hierarchical monitoring network, a data acquisition module, a source tracing and analysis module, and a monitoring and early warning module. The hierarchical monitoring network is set up in the electronic component workshop. The hierarchical monitoring network includes multiple fixed monitoring nodes deployed in the workshop space and key process equipment, multiple key area monitoring nodes deployed in the gas accumulation area, and at least one mobile inspection node. The data acquisition module is used to collect toxic gas concentration data and environmental airflow speed data at the location using the fixed monitoring nodes and key area monitoring nodes. At the same time, it uses the mobile inspection nodes to collect toxic gas concentration data and real-time spatial coordinate data on the movement trajectory, and encapsulates all the data into a monitoring data packet and transmits it to the central processing platform. The source tracing analysis module is used to preprocess and standardize the received monitoring data packets, dynamically generate a concentration distribution cloud map of the workshop based on the standardized data, and perform source tracing analysis to determine the location of the leakage source when the concentration data exceeds the preset absolute value threshold. The monitoring and early warning module is used to trigger trend warnings based on concentration change trends, and to trigger graded alarms based on absolute concentration values ​​and / or source tracing analysis results.

[0015] This invention discloses a method and system for monitoring toxic gas emissions in an electronic component workshop. The toxic gas emission monitoring system for electronic component workshops includes a hierarchical monitoring network, a data acquisition module, a source tracing and analysis module, and a monitoring and early warning module. A hierarchical monitoring network is constructed within the electronic component workshop. This network includes multiple fixed monitoring nodes deployed in the workshop space and key process equipment, multiple key area monitoring nodes deployed in gas accumulation areas, and at least one mobile inspection node. The fixed monitoring nodes and key area monitoring nodes collect toxic gas concentration data and ambient airflow velocity at their respective locations. Simultaneously, the mobile inspection nodes collect toxic gas concentration data and real-time spatial coordinate data along the movement trajectory, and encapsulate all data into monitoring data packets for transmission to the central processing platform. The received monitoring data packets are preprocessed and standardized, and a concentration distribution cloud map of the workshop is dynamically generated based on the standardized data. When the concentration data exceeds a preset absolute concentration threshold, source tracing analysis is performed to determine the location of the leak source. Trend warnings are triggered based on concentration change trends, and graded alarms are triggered based on absolute concentration values ​​and / or source tracing analysis results, enabling rapid source tracing and early warning of toxic gases in the electronic component workshop. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below.

[0017] Figure 1 This is a schematic diagram illustrating the steps of a method for monitoring toxic gas emissions in an electronic component workshop according to the first embodiment of the present invention.

[0018] Figure 2 This is a flowchart illustrating a method for monitoring toxic gas emissions in an electronic component workshop, provided by the present invention.

[0019] Figure 3 This is a schematic diagram of a toxic gas emission monitoring system for an electronic component workshop according to the second embodiment of the present invention.

[0020] Figure 4 This is a schematic diagram of the electronic device of the present invention.

[0021] In the diagram: 101 - Data acquisition module, 102 - Source tracing and analysis module, 103 - Monitoring and early warning module. Detailed Implementation

[0022] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application.

[0023] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used in this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0024] It should be understood that although the terms first, second, third, etc., may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0025] The first embodiment of this application is as follows: Please see Figures 1-2 This invention provides a method for monitoring toxic gas emissions in an electronic component workshop, comprising the following steps: S101. Construct a hierarchical monitoring network within the electronic component workshop. The hierarchical monitoring network includes multiple fixed monitoring nodes deployed in the workshop space and key process equipment, multiple key area monitoring nodes deployed in the gas accumulation area, and at least one mobile inspection node.

[0026] Specifically, the fixed monitoring nodes are deployed in a regular grid pattern on the workshop ceiling and load-bearing columns; and densely within one meter of potential leak points (such as valves and pipe interfaces) and exhaust outlets of all key process equipment (such as etching machines, doping equipment, and chemical vapor deposition reaction chambers). Each fixed monitoring node integrates a toxic gas concentration sensing unit and a miniature ultrasonic anemometer and wind direction sensing unit. All nodes are equipped with a unique network identifier (such as an IP address or ID code) and connect to the central processing platform via the workshop's wireless or wired IoT network. This forms the basic framework of the monitoring network, responsible for providing large-scale, continuous basic concentration data and airflow field data for the workshop.

[0027] Deployment locations of key monitoring nodes: Targeted deployments are made in areas prone to gas accumulation, determined based on fluid dynamics simulations or historical experience. These include, but are not limited to: return air vents of workshop ventilation systems, maintenance access points in underground process pipe corridors, corners formed by densely packed equipment, and areas with low elevations. These nodes use the same core sensing units as fixed monitoring nodes, but typically employ high-sensitivity gas concentration sensing units with lower detection limits to capture minute concentration accumulations. As a reinforcement to the fixed network, they are specifically designed to monitor high-risk blind spots and serve as a crucial line of defense for early leak warning.

[0028] Mobile inspection nodes utilize autonomous mobile robot platforms or ceiling-mounted track-based inspection vehicles as their mobile carriers. Each mobile carrier is equipped with a miniaturized toxic gas concentration sensor, a high-precision positioning module (such as a UWB or laser SLAM system), and a visible light camera. As the network's mobile force, it undertakes three core tasks: periodic patrols along preset routes; verification at designated locations upon receiving abnormal signals from fixed nodes; and temporary enhanced monitoring of temporary work areas difficult to cover by the fixed network.

[0029] When the monitoring node is physically installed and powered on for the first time, or when the system performs periodic maintenance as planned, a standardized self-test and initialization process will be automatically triggered to ensure that all sensors are in optimal working condition.

[0030] First, during the power-on self-test phase, after the node microcontroller starts up, it first diagnoses its own memory and communication modules. Then, it sends self-test commands to the gas concentration sensing unit and the wind speed and direction sensing unit. The internal firmware of the sensing unit performs basic functional tests such as circuit path checks and heater status assessments. The self-test results (normal / abnormal) are packaged together with the node identifier and immediately uploaded to the central processing platform. The platform monitoring interface will intuitively display the self-test status of all nodes (e.g., green light indicates success, red light indicates an alarm), facilitating engineers to quickly locate faulty hardware.

[0031] After passing the hardware self-test, the gas concentration sensing unit (especially sensors based on electrochemical or metal-oxide-semiconductor principles) will automatically enter the preheating phase to reach a stable operating temperature. The node continuously monitors the baseline signal output by the sensor. When the fluctuation of this signal stabilizes within a threshold range over a preset time period (e.g., 5 minutes), the node determines that it has reached the ready state and sends a "preheating complete" signal to the central processing platform.

[0032] After the system confirms that all nodes are ready, the central processing platform issues a command to initiate the baseline calibration procedure. At this point, the system assumes that the workshop environment is in a leak-free, safe state.

[0033] The platform commands all fixed and key monitoring nodes to simultaneously sample gas concentrations for a continuous period (e.g., 3 minutes), and uses the average reading as the baseline environmental concentration for that node. The platform collects these baseline values ​​from all nodes to create a background concentration distribution map of the workshop environment. This map allows the platform to intelligently identify nodes whose concentrations are generally slightly higher than other areas due to objective factors such as proximity to sewage outlets. The system learns and records the normal range baseline parameters for these nodes, rather than simply using a uniform absolute zero value. This effectively avoids false alarms caused by environmental background interference.

[0034] When the mobile inspection node returns to the charging dock, the dock contains a reference air chamber filled with clean air. When the mobile node docks for charging, it automatically extends its sensor probe into the reference air chamber to perform zero-point calibration and sensitivity checks, ensuring its measurement accuracy.

[0035] During system downtime (such as non-production hours), the central processing platform can instruct a mobile inspection node to sequentially travel to several key fixed monitoring nodes. The mobile and fixed nodes simultaneously sample the air at close range. The central processing platform compares the gas concentration readings returned by both. If the difference in readings exceeds a preset reasonable tolerance range, the platform will generate a sensor data consistency verification failure alarm, prompting manual verification and maintenance of the fixed node and / or the mobile node. This mechanism greatly enhances the reliability and trustworthiness of the entire system.

[0036] Once all the above steps are completed without error, the central processing platform stores the environmental background concentration benchmark value and normal range benchmark parameters of each node into the database, which will serve as the personalized benchmark for subsequent alarm calculations for that node. Subsequently, all monitoring nodes officially enter real-time monitoring operation mode.

[0037] S102. The fixed monitoring nodes and key area monitoring nodes are used to collect toxic gas concentration data and environmental airflow velocity data at their respective locations. At the same time, the mobile inspection nodes are used to collect toxic gas concentration data and real-time spatial coordinate data on the movement trajectory. All data is then encapsulated into monitoring data packets and transmitted to the central processing platform.

[0038] Specifically, each monitoring node (fixed, key, or mobile) collects a standardized data unit at a preset frequency (e.g., once per second). This unit contains the following core fields: Node Unique Identifier: A globally unique ID code used to accurately identify the source of data. This ID is burned into the node when it leaves the factory or joins the network and is bound to the node profile in the central processing platform database.

[0039] Node type code: Clearly identifies whether a node belongs to a "fixed monitoring node," a "key area monitoring node," or a "mobile inspection node." This helps the platform adopt different processing strategies for different types of data.

[0040] Timestamp: The UTC time of the collected data, accurate to milliseconds, is uniformly synchronized by the network time protocol to ensure that all node data has a unified time base.

[0041] Sensor data payload: For fixed and key monitoring nodes, it includes "toxic gas concentration value" and "airflow velocity vector" (including velocity magnitude and direction). For mobile inspection nodes, it includes "toxic gas concentration value" and "real-time spatial coordinates" (X, Y, Z three-dimensional coordinates provided by the positioning module).

[0042] Node status word: contains the device's own status information, such as battery level (for mobile nodes), sensor health status codes, communication signal strength, etc.

[0043] The collected data units are further encapsulated into standard "monitoring data packets". Each data packet has the following structure: [Packet header start character] + [Packet length] + [Node unique identifier] + [Node type code] + [Time stamp] + [Sensor data payload] + [Node status word] + [Data checksum] + [Packet header end character].

[0044] After receiving massive amounts of monitoring data packets, the central processing platform distinguishes, identifies, and correlates the data through the following mechanisms: First, the unique node identifier and node type code of each data packet are parsed. The platform maintains a "node information database" that stores detailed metadata corresponding to each identifier, including: the node's physical installation location (three-dimensional coordinates for fixed nodes, dynamically updated for mobile nodes), the type of gas monitored, sensor model, measurement range, calibration history, etc. By matching the metadata, the platform automatically classifies the data streams from different nodes and distributes them to different processing buffers. For example, airflow data from all fixed nodes are aggregated to build a workshop airflow field model.

[0045] To achieve spatiotemporal alignment between data, the platform performs core intelligent association operations: Spatial correlation: For fixed nodes, their location information is directly retrieved from the node information database; for moving nodes, their real-time spatial coordinates are extracted and bound to the gas concentration value, together forming a "data point" that can be precisely located in space. The platform maps the location information of all data points onto a unified electronic map of the workshop. Through this map, the platform can intuitively understand which concentration reading comes from which specific location in the workshop.

[0046] Time correlation: The timestamps of all data packets ensure that they are slices of the workshop state at the same or similar times. This allows the platform to make meaningful cross-sectional comparisons, such as analyzing whether the concentration increase at point A at "time T" is related to the airflow change at point B.

[0047] Logical Association: The platform can not only perform simple spatiotemporal matching, but also dynamically establish associated logical groups based on operating conditions. For example, when a mobile node is assigned to verify an alarmed fixed node, the platform will temporarily create a "verification task group," logically strongly associating the data streams of the two nodes and prioritizing the comparison of their concentration readings to quickly verify the authenticity of the alarm. Furthermore, based on airflow direction and equipment layout, the platform intelligently determines that certain nodes are physically downstream of potential leak sources. When an upstream node alarms first, the platform will immediately focus on data changes in its downstream associated node group, achieving relay of early warning and prediction of diffusion paths.

[0048] To ensure smooth data transmission in complex industrial electromagnetic environments, a multi-layered data transmission protection mechanism was adopted.

[0049] Fixed and key monitoring nodes prioritize wired industrial Ethernet communication to ensure bandwidth and stability. Simultaneously, all nodes (including fixed nodes) are equipped with industrial-grade wireless communication modules (such as Wi-Fi 6 or 5G) as backup communication paths. The central processing platform continuously monitors the link status of each node. When the wired network is interrupted, the system automatically and seamlessly switches the node's communication to the wireless link, ensuring uninterrupted data flow.

[0050] The platform manages data packets using priority tags. For example, data from alarm nodes and data transmitted back from mobile nodes during review tasks are marked as high priority and placed at the head of the transmission queue to ensure that emergency information can be processed immediately. Regular monitoring data is transmitted with normal priority. This mechanism ensures that security alerts are not affected during network congestion.

[0051] Each monitoring node has a small area of ​​non-volatile storage designed as a local data cache. When a node detects a complete network connection interruption to the platform, it automatically initiates a resume transmission mode. Newly collected monitoring data packets are temporarily stored in the local cache. Once the network is restored, the node automatically uploads the cached historical data to the platform in batches, in chronological order.

[0052] During wireless transmission, forward error correction (FEC) technology is employed, adding redundant checksums to data packets. This allows the receiver to correct minor bit errors automatically without retransmission, reducing latency. Upon receiving each data packet, the platform calculates its checksum and compares it with the checksum within the packet. Packets that fail checksum verification are immediately discarded, and retransmission can be requested from the sending node according to a policy, ensuring the accuracy of the data stored.

[0053] S103. The received monitoring data packets are preprocessed and standardized, and a concentration distribution cloud map of the workshop is dynamically generated based on the standardized data. When the concentration data exceeds the preset absolute value threshold, source tracing analysis is performed to determine the location of the leak source.

[0054] Specifically, when the central processing platform receives monitoring data packets from various monitoring nodes, it first enters the data preprocessing and standardization process, which aims to "purify" the data and lay a reliable foundation for subsequent advanced analysis.

[0055] The platform first unpacks each data packet according to a preset data packet encapsulation protocol, extracting all fields such as the node's unique identifier, timestamp, and sensor data payload. It then checks the logical validity of the data. For example, gas concentration values ​​should not be negative, and airflow velocity should not exceed the physical limits of the workshop ventilation system. Any data that violates basic physical logic will be marked as "invalid data" and temporarily stored in the log, not proceeding to further analysis.

[0056] Statistical outlier filtering: The platform builds a short-term historical model for each data type (such as concentration and wind speed) for each node. For newly introduced data points, if their values ​​suddenly deviate from the historical mean by more than a certain number of standard deviations (e.g., 5 standard deviations), they are considered impulse noise and filtered out.

[0057] Cross-validation based on node status words: If the gas concentration value of a node is abnormally high, but the node status word shows that the sensor health status is "warning" or the signal strength is extremely weak, the platform will reduce the weight of the data or mark it as "suspected abnormal" and wait for subsequent data or moving nodes to verify and confirm.

[0058] Sliding window smoothing: To suppress random fluctuations, a moving average window over time is used to process continuous concentration and wind speed data. For example, the average of the most recent 5 seconds of data is taken as the effective output value for the current second. This effectively smooths out spikes caused by brief disturbances, making the trend clearer.

[0059] Even when monitoring the same gas, gas concentration sensing units from different manufacturers and of different models may produce different raw output signals. The platform maintains a sensor calibration parameter database, which stores the "sensitivity coefficient" and "zero-point offset" of each sensing unit. The platform uses the following formula to standardize the raw concentration readings in real time: Standardized concentration value = (original reading - zero offset) * sensitivity coefficient.

[0060] The physical location information of all nodes (fixed and moving) is transformed into a unified three-dimensional coordinate system with a corner of the workshop as the origin. This allows all data points to be processed within the same spatial model.

[0061] Despite the timestamps, the platform employs interpolation alignment techniques to generate "snapshots" of the same moment. For data that is not collected in exactly the same millisecond, the platform uses an interpolation algorithm to calculate estimates of their values ​​at the same standard analysis time (such as whole seconds) based on data from moments before and after them.

[0062] Using preprocessed and standardized data, the platform began to build a dynamic model reflecting the gas distribution throughout the workshop.

[0063] The platform divides the workshop space into a fine two-dimensional or three-dimensional grid. It employs a Kriging spatial interpolation algorithm. This algorithm considers not only the concentration value of each known data point (i.e., the location of a monitoring node) but also the spatial correlation structure between these data points (modeled through a "variance function"). When calculating the concentration at a grid point, the algorithm assigns higher weight to monitoring node data upwind, as upwind data contributes more to the predicted concentration at that point. This makes the generated concentration distribution more consistent with the physical laws of gas diffusion, resulting in significantly higher accuracy than interpolation methods based solely on distance.

[0064] Through the interpolation calculations described above, each grid cell is assigned an estimated concentration value. The platform uses a color gradient (such as from blue to red) to visualize these concentration values, thereby generating an intuitive, dynamic concentration distribution cloud map on the electronic map of the workshop. This cloud map updates in real time, much like performing a "CT scan" on the workshop environment, making the previously invisible distribution of toxic gases readily apparent.

[0065] When the cloud map shows that the concentration at any node exceeds the standard, that is, when the concentration data exceeds the preset absolute concentration threshold, the platform initiates intelligent source tracing analysis to accurately locate the source of the leak. Based on the dynamic concentration distribution cloud map, the platform generates a concentration gradient field by calculating the rate of change of concentration of each grid point in its surrounding grids. The gradient direction points to the direction of the fastest increase in concentration, and the area where the arrows converge is the potential leak source area.

[0066] The platform utilizes airflow velocity vector data uploaded from all fixed monitoring nodes and generates a dynamic airflow field model of the workshop using a similar interpolation method. Starting from a detected high-concentration area (which can be considered a "target point"), the platform hypothetically releases several "virtual particles." These particles are then commanded to simulate motion against the current airflow direction. The trajectories of these reverse-moving particles converge in certain specific regions. These convergence regions are the most likely locations of the leak source. This is because, regardless of how the gas diffuses in the real physical world, its initial source must be where these trajectories converge.

[0067] Instead of simply outputting a single coordinate point, the platform generates a "source tracing probability heatmap." This map, based on the density of particle trajectory convergence, shows the probability of different regions being leakage sources (e.g., the red region has the highest probability, followed by the yellow region). Upgrading pure concentration analysis to probabilistic localization based on inverse fluid dynamics simulation greatly improves the scientific rigor and accuracy of the source tracing results, especially when multiple potential leakage sources or complex airflows exist.

[0068] The platform overlays the calculated source tracing probability heatmap with the dynamic concentration distribution cloud map to comprehensively determine the one or several most suspicious locations. It then immediately issues instructions to the mobile inspection nodes, guiding them to the area with the highest probability for close-range reconnaissance. The real-time video and high-precision concentration readings transmitted back by the mobile nodes provide the final "ground-based" verification for the platform's source tracing conclusions.

[0069] S104. Trigger trend warnings based on concentration change trends, and trigger tiered alarms based on absolute concentration values ​​and / or source tracing analysis results.

[0070] Specifically, for each monitoring node (especially those near key process equipment), the central processing platform maintains a dynamic data buffer that stores preprocessed, standardized concentration data over a past period (e.g., 3 to 5 minutes), forming a concentration-time series.

[0071] The platform periodically (e.g., every second) performs linear regression analysis on the concentration-time series within the buffer. A trend line is fitted using the least squares method, and the slope of this trend line is calculated. This slope quantitatively represents the average rate of change of concentration within the current time window. A positive slope indicates that the concentration is rising. The larger the slope, the faster the concentration is rising, and the faster the risk is accumulating.

[0072] The platform sets a trend warning threshold based on the slope of the concentration change trend (this is a threshold based on the slope value, distinct from the absolute concentration threshold used to determine whether a concentration exceeds the limit). When the system detects that the trend slope of a node's concentration sequence continuously exceeds this trend warning threshold, and its current standardized concentration value remains below the safe first-level concentration threshold, a trend warning is immediately triggered. The first-level concentration threshold is one of the preset absolute concentration thresholds mentioned above.

[0073] To prevent false alarms, the platform performs context verification when a trend alert is triggered. It checks whether the airflow near the node in the dynamic airflow field model is stable. If the airflow is stable but the concentration continues to rise rapidly, the reliability of the alert is further confirmed. Conversely, if the concentration fluctuation is caused by a brief shutdown of the ventilation system, the system may choose to increase the monitoring level instead of immediately issuing a global alert.

[0074] Tiered alarm systems ensure that emergency response resources are allocated in the most rational and effective manner. Alarm levels are defined as follows: Level 1 Warning: Potential Risk Alert. This corresponds to the trend warning mentioned above, or situations where the concentration exceeds the first threshold (warning threshold) but the source of the leak has not yet been identified.

[0075] Level 2 Alert: A confirmed emergency. This corresponds to a concentration exceeding a higher second threshold (hazard threshold), or a concentration that continues to rise uncontrollably to a hazardous level after a Level 1 alert has been confirmed.

[0076] Level 1 Early Warning Response Procedure: On the system's electronic map, the node icon that triggered the warning began to flash yellow. The system automatically sent a notification message to the mobile terminal of the area manager, containing the message: "[Area XX] The concentration of XX gas is showing a rapid upward trend. Please pay close attention." Simultaneously, the platform launched an intelligent source tracing analysis program to attempt to locate the possible leak source and instructed the mobile inspection node to go to the warning area for reconnaissance and confirmation.

[0077] Level 2 Alarm Response Procedure: The system triggers the workshop's audible and visual alarm system, issuing a clear evacuation or warning signal. An emergency alert is sent to all relevant management and safety teams, containing the precise leak location (from source tracing analysis) and current concentration. The platform, through predefined interfaces, automatically performs some or all of the following operations: It activates the ventilation system, initiating emergency exhaust mode to maximize the removal of toxic gases from the workshop; it sends a signal to the material management system to automatically shut off the gas solenoid valves upstream and downstream of the suspected leaking equipment; it locks the access control system in the area to prevent accidental entry; and it initiates the generation of gas diffusion prediction paths to guide evacuation and rescue efforts.

[0078] The method further includes: When a high-level alarm is triggered, a gas diffusion prediction path is generated based on the location of the leak source and real-time ambient airflow velocity data, specifically: The generated gas diffusion prediction path is a key function that provides scientific decision support for personnel evacuation and external rescue after a Level 2 alarm is triggered.

[0079] The system employs a simplified Gaussian plume diffusion model as its core algorithm. This model can describe the diffusion behavior of gases released from continuous point sources in turbulent atmospheres, making it highly suitable for simulating gas leaks within a workshop. The model's accuracy depends on precise perception of the current environmental conditions. The platform dynamically acquires the following parameters: Leakage source parameters: These are derived from the location of the highest probability point determined in the source probability heatmap and are considered the "source point" in the diffusion model. The leakage rate can be estimated based on the concentration increase trend and the nearby airflow velocity.

[0080] Environmental dynamic parameters: These are derived directly from a dynamic airflow field model covering the entire workshop, including the current average wind speed and prevailing wind direction. This is the core driving force behind the diffusion simulation.

[0081] Workshop geometry parameters: The model has a pre-stored 3D layout of the workshop, including information on obstacles such as walls, equipment, and passageways.

[0082] The platform does not perform a single simulation, but rather conducts short-term, multi-scenario simulations taking into account possible changes in airflow.

[0083] Main scenario: Simulation based on the current measured airflow field.

[0084] Boundary scenarios: Simulations are conducted assuming a certain percentage increase / decrease in wind speed or a slight shift in wind direction to assess the uncertainty of the prediction results.

[0085] The simulation results of all scenarios are overlaid to generate a diffusion probability cloud map over time. For example: Area affected in the next minute: Displayed as a red high-probability zone, indicating the area where the gas is highly likely to reach.

[0086] Areas affected in the next 2-3 minutes: Shown as yellow medium probability zones, indicating areas where the gas may spread.

[0087] Areas affected in the next 3-5 minutes: Shown as green low probability areas, indicating areas where the gas may have a small impact but vigilance is still required.

[0088] The predictive diffusion path has been upgraded to a probabilistic diffusion area prediction based on multi-scenario simulation, and visualized over time. This better reflects the uncertainties of the real world and provides policymakers with a more flexible and forward-looking action guide.

[0089] The generated gas diffusion prediction path includes the following steps: Determine the diffusion model parameters: take the location of the leak source determined by the source tracing analysis as the point source, estimate the initial leakage rate based on the concentration rise rate near the leak point and the ambient airflow velocity; use the average wind speed and prevailing wind direction in the real-time constructed dynamic airflow field model as environmental dynamic parameters; and call the pre-stored 3D layout map of the workshop as the geometric boundary conditions.

[0090] Perform multi-scenario diffusion simulations: The Gaussian plume diffusion model is used as the core algorithm. Multi-scenario simulations are performed, including: a main scenario simulation based on the current measured airflow field; and several boundary scenario simulations based on assumed wind speed increases or decreases or wind direction deflections, to assess prediction uncertainties.

[0091] Probability cloud maps over time are generated: by integrating simulation results from multiple scenarios, probability cloud maps of the gas diffusion impact area are generated for different time periods in the future (such as 1 minute, 2-3 minutes, 3-5 minutes), where high probability areas represent areas where the gas is most likely to reach.

[0092] Output of Risk Situation Map and Evacuation Routes: A dynamic risk situation map is created by overlaying a diffusion probability cloud map over time onto the workshop's electronic map. Based on this map, the system automatically calculates and highlights the optimal safe evacuation route. This route proactively avoids current and predicted high-concentration areas and guides passengers to the nearest safe exit. The risk situation map and recommended evacuation routes are pushed to personnel terminals and the command center in real time.

[0093] The diffusion probability cloud map over time is directly overlaid onto the electronic map of the workshop to form a dynamic "risk situation map". Based on the diffusion prediction, the system automatically calculates and highlights the best safe evacuation route on the map. This route will proactively avoid current and predicted high-concentration areas and guide personnel to the nearest safe exit or refuge room. The "risk situation map" containing the predicted areas and recommended evacuation routes is pushed in real time to the mobile terminals of on-site staff and the large screen of the security command center.

[0094] Achieving precise source tracing: By building a dynamic monitoring network, the location of the leak source can be quickly identified, greatly shortening the emergency response time and facilitating timely equipment maintenance and risk control.

[0095] Enhanced early warning capabilities: It has early warning functions and can issue an alarm in advance by analyzing the concentration change trend before the gas concentration reaches the dangerous threshold, realizing the transformation from "passive alarm" to "active early warning".

[0096] Eliminating monitoring blind spots: By combining fixed and mobile monitoring nodes, especially by strengthening monitoring of key areas, comprehensive and seamless coverage of the workshop space has been achieved.

[0097] Improve system reliability: The introduction of multi-parameter fusion analysis (concentration, airflow, process status) effectively filters out environmental interference, significantly reduces the false alarm rate and false alarm rate of the system, and makes alarms more accurate and reliable.

[0098] Intelligent decision support: The system can simulate gas diffusion paths, providing intuitive and scientific decision-making basis for personnel evacuation and ventilation system adjustments.

[0099] The second embodiment of this application is as follows: Please see Figure 3 This invention provides a toxic gas emission monitoring system for electronic component workshops, applied to a toxic gas emission monitoring method for electronic component workshops as provided in the first embodiment. The toxic gas emission monitoring system for electronic component workshops includes a hierarchical monitoring network, a data acquisition module 101, a source tracing and analysis module 102, and a monitoring and early warning module 103. The hierarchical monitoring network is set up in the electronic component workshop. The hierarchical monitoring network includes multiple fixed monitoring nodes deployed in the workshop space and key process equipment, multiple key area monitoring nodes deployed in the gas accumulation area, and at least one mobile inspection node. The data acquisition module 101 is used to collect toxic gas concentration data and environmental airflow speed data at the location using the fixed monitoring node and the key area monitoring node. At the same time, it uses the mobile inspection node to collect toxic gas concentration data and real-time spatial coordinate data on the moving trajectory, and encapsulates all the data into a monitoring data packet and transmits it to the central processing platform. The source tracing analysis module 102 is used to preprocess and standardize the received monitoring data packets, dynamically generate a concentration distribution cloud map of the workshop based on the standardized data, and perform source tracing analysis to determine the location of the leakage source when the concentration data exceeds the preset absolute value threshold. The monitoring and early warning module 103 is used to trigger trend early warning based on concentration change trends, and to trigger graded alarms based on absolute concentration values ​​and / or source tracing analysis results.

[0100] Regarding the system in the above embodiments, the specific ways in which each module performs operations have been described in detail in the embodiments related to the method, and will not be elaborated here.

[0101] For the system embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this application according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0102] Accordingly, this application also provides an electronic device, comprising: one or more processors; a memory for storing one or more programs; and, when the one or more programs are executed by the one or more processors, causing the one or more processors to implement the above-described method for monitoring toxic gas emissions in an electronic component workshop. Figure 4 The diagram shown is a hardware structure diagram of any device with data processing capabilities used in an electronic component workshop to provide a toxic gas emission monitoring system according to an embodiment of the present invention. Except for... Figure 4 In addition to the processor, memory, and network interface shown, any data processing device in the embodiment may also include other hardware depending on the actual function of the data processing device, which will not be described in detail here.

[0103] Accordingly, this application also provides a computer-readable storage medium storing computer instructions, which, when executed by a processor, implement the aforementioned method for monitoring toxic gas emissions in an electronic component workshop. The computer-readable storage medium can be an internal storage unit of any data-processing device as described in any of the foregoing embodiments, such as a hard disk or memory. The computer-readable storage medium can also be an external storage device, such as a plug-in hard disk, smart media card (SMC), SD card, flash card, etc., equipped on the device. Furthermore, the computer-readable storage medium can include both internal storage units of any data-processing device and external storage devices. The computer-readable storage medium is used to store the computer program and other programs and data required by the data-processing device, and can also be used to temporarily store data that has been output or will be output.

[0104] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the disclosure herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein.

[0105] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope.

Claims

1. A method of monitoring toxic gas emissions from an electronic components plant, characterized by, The method comprises the following steps: A hierarchical monitoring network is constructed in the electronic component workshop, which comprises a plurality of fixed monitoring nodes deployed in the workshop space and key process equipment, a plurality of key area monitoring nodes deployed in gas accumulation areas, and at least one mobile inspection node, wherein the mobile inspection node is used for periodic patrol according to a preset route; after receiving an abnormal signal of the fixed monitoring node, the mobile inspection node goes to the specified location for review; temporary intensive monitoring is performed on temporary work areas that are difficult to cover by the fixed monitoring nodes; The fixed monitoring nodes and the key area monitoring nodes are used to collect toxic gas concentration data and environmental airflow velocity data at the location, and the mobile inspection node is used to collect toxic gas concentration data and real-time spatial coordinate data on the moving track, and all data are packaged into a monitoring data packet and transmitted to a central processing platform; The received monitoring data packet is preprocessed and standardized, and a concentration distribution cloud map of the workshop is dynamically generated based on the standardized data, and when the concentration data exceeds a preset concentration absolute value threshold, a traceability analysis is performed to determine the leak source location; A trend warning is triggered based on the concentration change trend, and a graded alarm is triggered based on the concentration absolute value and / or the traceability analysis result; When the concentration data exceeds the preset concentration absolute value threshold, a traceability analysis is performed to determine the leak source location, comprising: Based on the dynamic concentration distribution cloud map, a concentration gradient field is generated by calculating the concentration change rate of each grid point in its surrounding grid, wherein the gradient direction points to the direction of the fastest concentration rise, and the area where the arrows converge is the potential leak source area; The vector data of the environmental airflow velocity data uploaded by all fixed monitoring nodes is used to generate a current dynamic airflow field model of the workshop by an interpolation method; Starting from the high-concentration area, reverse particle tracking simulation is performed in the dynamic airflow field model; According to the convergence of the particle trajectory, a traceability probability thermal map representing the probability of the leak source is generated, the calculated traceability probability thermal map is superimposed on the dynamic concentration distribution cloud map, and one or more points are comprehensively determined; then, an instruction is sent to the mobile inspection node to control the mobile inspection node to go to the area with the highest probability for close reconnaissance, and the leak source location is determined by combining the real-time video and high-precision concentration readings returned by the mobile node.

2. The electronic component manufacturing plant toxic gas emission monitoring method according to claim 1, wherein The method further comprises: When a high-level alarm is triggered, a gas diffusion prediction path is generated based on the leak source location and real-time environmental airflow velocity data.

3. The electronic component manufacturing plant toxic gas emission monitoring method according to claim 1, wherein After constructing a hierarchical monitoring network in the electronic component workshop, the method further comprises: After all nodes are powered on and complete hardware self-checking, the central processing platform instructs all nodes to collect the environmental background concentration in a preset non-leakage state, and stores the normal range reference parameters of each node; The mobile inspection node is instructed to move to a specified fixed monitoring node, and the gas concentration data collected by the two nodes are compared for cross-validation of sensor data.

4. The method of claim 1, wherein the toxic gas emission monitoring method is used in an electronic device manufacturing plant. After the central processing platform receives the monitoring data packet, the method further comprises: The data source is identified and classified by analyzing the node unique identifier and the node type code and matching the metadata in the preset node information database; and the data is associated in time and space by the timestamp and the node position information, wherein the monitoring data packet comprises a node unique identifier, a node type code, a timestamp, a sensing data payload and a node state word.

5. The method of claim 1, wherein the toxic gas emission monitoring method is used in an electronic device manufacturing plant. The received monitoring data packet is preprocessed and standardized, and a concentration distribution cloud map of the workshop is dynamically generated based on the standardized data, comprising: statistically filtering and sliding window smoothing the toxic gas concentration data; using the pre-stored sensor calibration parameters to perform real-time normalization processing on the environmental background concentration data to obtain standardized concentration values; unifying the position information of all nodes to the same three-dimensional coordinate system; dynamically generating a concentration distribution cloud map of the workshop based on the standardized data.

6. The method of claim 5, wherein the toxic gas emission monitoring method is used in an electronic device manufacturing plant. The concentration distribution cloud map of the workshop is dynamically generated based on the standardized data, comprising: using a spatial interpolation algorithm and taking the real-time collected environmental airflow velocity data as an auxiliary variable to calculate the standardized toxic gas concentration data to generate a concentration distribution cloud map of the workshop.

7. The method of claim 1, wherein the toxic gas emission monitoring method is used in an electronic device manufacturing plant. Triggering a trend warning based on the concentration change trend and triggering a hierarchical alarm based on the concentration absolute value and / or the traceability analysis result, comprising: calculating the trend slope of the concentration-time sequence in the dynamic data buffer corresponding to each monitoring node; when the trend slope exceeds the trend warning threshold and the current concentration is lower than the first threshold, triggering a trend warning; triggering a hierarchical alarm based on the concentration absolute value and / or the traceability analysis result.

8. The method of claim 7, wherein the toxic gas emission monitoring method is used in an electronic device manufacturing plant. Triggering a hierarchical alarm based on the concentration absolute value and / or the traceability analysis result, comprising: when the trend warning is triggered or the concentration exceeds the first threshold but the traceability analysis result is not clear, triggering a first-level warning; when the concentration exceeds the second threshold or the concentration continues to be out of control after the first-level warning, triggering a second-level warning.

9. A toxic gas emission monitoring system for an electronic component workshop, applied to the toxic gas emission monitoring method for an electronic component workshop as claimed in claim 1, characterized in that, the toxic gas emission monitoring system for an electronic component workshop comprises a hierarchical monitoring network, a data acquisition module, a traceability analysis module and a monitoring warning module, the hierarchical monitoring network is arranged in the electronic component workshop, and comprises a plurality of fixed monitoring nodes arranged in the workshop space and key process equipment, a plurality of key area monitoring nodes arranged in the gas accumulation area, and at least one mobile inspection node; the data acquisition module is used to collect toxic gas concentration data and environmental airflow velocity data at the location by using the fixed monitoring nodes and the key area monitoring nodes, and simultaneously collect toxic gas concentration data and real-time spatial coordinate data on the moving track by using the mobile inspection node, and encapsulate all data into a monitoring data packet and transmit it to a central processing platform; The traceability analysis module is configured to preprocess and standardize the received monitoring data packet, dynamically generate a concentration distribution cloud map of the workshop based on the standardized data, and perform traceability analysis to determine the location of the leakage source when the concentration data exceeds a preset absolute concentration threshold value. The monitoring and early warning module is configured to trigger trend early warning based on the concentration change trend, and trigger graded alarm based on the absolute concentration value and / or the traceability analysis result.

Images