Volatile organic compound leakage real-time supervision system and method

Abstract

The invention discloses a volatile organic compound leakage real-time supervision system and method, and the system comprises a multi-dimensional sensing module which collects volatile organic compound concentration data and environmental meteorological data in real time, and forms an original monitoring signal containing a concentration value and meteorological parameters; the edge calculation module receives the original monitoring signal, performs data filtering, abnormal value elimination and concentration trend analysis operation, and forms a preprocessing signal marked with a potential abnormal state; the central processing module receives the preprocessed signal and generates a leakage positioning signal and a real-time risk level signal; the early warning management module receives the real-time risk level signal and generates a graded early warning signal containing an alarm level; and the visual platform module receives the leakage positioning signal and the grading early warning signal, dynamically renders a volatile organic compound concentration distribution cloud picture based on a factory digital twinborn model, and generates a visual supervision interface signal. According to the invention, the problem that VOCs leakage is difficult to monitor in real time, accurately position and intelligently warn can be solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of environmental monitoring and industrial safety technology, in particular to a volatile organic compound leakage real-time monitoring system and method. BACKGROUND

[0002] Currently, in the petroleum chemical industry, pharmaceutical industry, coating industry and other industries involving the use of volatile organic compounds, effective supervision of possible VOCs leakage of production devices, storage tanks and pipelines is the core link of environmental protection and safety production. However, the existing mainstream supervision technology and method has many significant defects. First, in the monitoring aspect, a large number of enterprises still rely on safety personnel to use portable detectors for regular or irregular patrol inspection. This method not only cannot realize all-weather uninterrupted coverage, has a large monitoring blind area and time gap, and is difficult to capture instantaneous or intermittent leakage events, but also relies heavily on the experience and responsibility of personnel, is low in efficiency and high in cost. Secondly, in the leakage positioning aspect, even if the concentration exceeds the standard phenomenon is found through fixed alarm or manual patrol, it can only provide a general direction or area, and cannot quickly and accurately locate the specific leakage equipment, such as a valve, flange, pump body or slight damage of the connection, so that the maintenance personnel need to spend a lot of time on site to investigate, delay the best disposal opportunity, and may make a small hidden danger into a big accident. Thirdly, in the early warning and decision-making aspect, the traditional alarm mechanism is usually based on a single concentration threshold. This method is easily disturbed by concentration fluctuations caused by environmental factors such as sudden change of wind direction and change of air pressure, thereby generating a large number of false alarms and missed alarms, reducing the credibility of the alarm signal, and easily leading to the paralysis of the staff. At the same time, due to the lack of comprehensive assessment of leakage risk, it is impossible to perform graded alarm and differentiated emergency response according to the severity of the leakage, the diffusion trend and the potential impact, resulting in waste of management resources or insufficient emergency response. In addition, the existing monitoring system is usually an information island, and there is a lack of effective linkage and integration between the monitoring data and the production management system, equipment maintenance system and safety emergency platform of the enterprise, the data value is not deeply mined, and a comprehensive information view for managers to provide traceability analysis, trend prediction and decision support cannot be provided. In summary, it is an urgent technical requirement in the industry to develop a VOCs leakage monitoring system that can realize real-time and accurate monitoring, rapid traceability positioning, intelligent graded early warning and visual integration. SUMMARY

[0003] In view of the above prior art defects, the purpose of the present application is to provide a volatile organic compound leakage real-time monitoring system and method for solving the problems of difficult real-time monitoring, accurate positioning and intelligent early warning of VOCs leakage. The present application acquires VOCs concentration and meteorological data in real time by deploying a multi-dimensional sensing network; uses an edge computing module to perform data preprocessing and preliminary anomaly judgment on the monitoring terminal side, improving response efficiency and reducing cloud load; the central processing module fuses multi-source information, uses a tracing algorithm based on the Gaussian plume model to accurately calculate the coordinates of the leakage source, and combines a machine learning model to evaluate the risk level; finally, the warning management module realizes graded alarm, and the visualization platform based on digital twinning dynamically displays the leakage diffusion situation and risk position, thereby realizing integrated intelligent monitoring from perception, positioning, early warning to visualization management.

[0004] The present application provides a volatile organic compound leakage real-time monitoring system, comprising: A multi-dimensional sensing module acquires volatile organic compound concentration data and environmental meteorological data in real time, forming original monitoring signals containing concentration values and meteorological parameters; An edge computing module processes the original monitoring signals, performs data filtering, outlier rejection and concentration trend analysis operations, and forms preprocessed signals indicating potential abnormal states; A central processing module receives the preprocessed signals, fuses multi-source monitoring data and real-time meteorological information, calculates the coordinates of potential leakage sources through a leakage tracing algorithm based on the Gaussian plume model, and generates leakage positioning signals and real-time risk level signals; A warning management module receives real-time risk level signals, generates graded warning signals containing alarm levels according to preset low, medium and high risk thresholds and corresponding action rules; A visualization platform module receives leakage positioning signals and graded warning signals, dynamically renders volatile organic compound concentration distribution cloud maps based on a plant digital twinning model, displays leakage source positions and risk levels in a three-dimensional virtual plant, and generates visualization management interface signals.

[0005] In an embodiment of the present application, the multi-dimensional sensing module includes a point sensor array arranged near key equipment in the production device area, a line scanning monitoring station arranged at the perimeter of the plant, and a planar area environmental micro station covering the entire plant overhead, the point sensor array is used to capture local high-concentration leakage signals, the line scanning monitoring station is used to monitor the diffusion trend of volatile organic compounds at the edge of the plant boundary, and the planar area environmental micro station is used to obtain the overall concentration background distribution of the plant. The multi-dimensional sensing module integrates monitoring data from different spatial scales into a unified original monitoring signal through the fusion of heterogeneous networks.

[0006] In an embodiment of the present application, the concentration trend analysis operation performed by the edge computing module is specifically to calculate the change slope and acceleration of the concentration sequence using a sliding time window, and when it is identified that the concentration value continues to rise and the change acceleration exceeds the adaptive threshold, it is determined as a potential abnormal state. The edge computing module also has a local caching mechanism to continuously record the preprocessed signals when the communication is interrupted and perform data retransmission after the communication is restored, ensuring the continuity and integrity of the monitoring data.

[0007] In an embodiment of the present application, the device start-stop state, pipeline internal pressure and temperature parameters provided by the enterprise production management system are used as priori knowledge by the central processing module to embed into the leakage tracing algorithm, which is used to correct the errors generated when tracing calculation is only relied on environmental meteorological information, thereby improving the calculation accuracy and reliability of the potential leakage source coordinates in complex working conditions.

[0008] In an embodiment of the present application, the real-time risk level signal generated by the central processing module is generated by a multi-factor comprehensive evaluation model, which simultaneously considers the real-time concentration exceeding multiple, concentration rising rate, estimated duration of leakage event, sensitivity of the location where the leakage source is located, and diffusion influence range under current weather conditions, outputs a quantitative comprehensive risk index by weighted fusion of these factors, and divides the risk level according to the index.

[0009] In an embodiment of the present application, the graded warning signal generated by the early warning management module also automatically triggers the disposal process linked thereto, automatically generates a patrol work order for the medium risk level and pushes it to the mobile patrol terminal, and automatically triggers the sound and light alarm for the high risk level and synchronously sends emergency information containing emergency disposal instructions to the communication equipment of the safety responsible person, realizing the rapid closed-loop response from early warning to disposal.

[0010] In an embodiment of the present application, the volatile organic compound concentration distribution cloud map dynamically rendered by the visualization platform module is dynamically simulated in combination with real-time meteorological wind direction and wind speed data, and the cloud map directly shows the concentration diffusion path and dilution process from the leakage source to the outside with different color gradients, and can replay the whole leakage diffusion process in a specific time period based on historical monitoring data, which is used for accident review and tracing analysis.

[0011] In an embodiment of the present application, the visualization platform module is also integrated with a device health management submodule, which establishes an electronic file for each monitored device, records its historical leakage events, maintenance records and operation parameters related to leakage, and calculates a dynamic health score for the device based on these data, and the health score is displayed as background information in the visualization supervision interface together with the real-time risk level signal.

[0012] In an embodiment of the present application, the system further comprises an audit tracking module connected to the early warning management module and the visualization platform module, the audit tracking module continuously receives and stores the graded early warning signals, the leakage positioning signals and the original data snapshots on which the early warning is based, forming a complete audit log with time stamp, which records the whole chain of events from the initial appearance of the anomaly, risk judgment, early warning release to subsequent disposal, supporting query and traceability by time, location or equipment.

[0013] The present application also includes a volatile organic compound leakage real-time monitoring method, comprising: S1: Real-time acquisition of volatile organic compound concentration data and environmental meteorological data to form original monitoring signals containing concentration values and meteorological parameters; S2: The original monitoring signals are subjected to data filtering, outlier rejection and concentration trend analysis operations to form preprocessed signals indicating potential abnormal states; S3: Receiving the preprocessed signals, fusing multi-source monitoring data and real-time meteorological information, calculating the coordinates of potential leakage sources through a leakage tracing algorithm based on the Gaussian plume model to generate leakage positioning signals and real-time risk level signals; S4: Receiving the real-time risk level signals, generating graded early warning signals containing alarm levels according to pre-set low, medium and high risk thresholds and corresponding action rules; S5: Receiving the leakage positioning signals and the graded early warning signals, dynamically rendering the volatile organic compound concentration distribution cloud map based on the digital twin model of the plant, and displaying the leakage source location and risk level in the three-dimensional virtual plant to generate a visual monitoring interface signal.

[0014] The volatile organic compound leakage real-time monitoring system and method provided by the present application acquires VOCs concentration and meteorological data in real time through the deployment of a multi-dimensional sensing network; uses an edge computing module to perform data preprocessing and preliminary anomaly judgment on the monitoring terminal side, improving response efficiency and reducing cloud load; the central processing module fuses multi-source information, uses a tracing algorithm based on the Gaussian plume model to accurately calculate the coordinates of the leakage source, and combines a machine learning model to evaluate the risk level; finally, the early warning management module realizes graded alarm, and the visualization platform based on digital twinning dynamically displays the leakage diffusion situation and risk location, thereby realizing integrated intelligent monitoring from perception, positioning, early warning to visualization management. BRIEF DESCRIPTION OF DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0016] Figure 1 System architecture diagram for volatile organic compound leakage real-time supervision system; Figure 2 Method flow diagram for volatile organic compound leakage real-time supervision method. DETAILED DESCRIPTION

[0017] The present application is herein described, by way of example only, with the assistance of the accompanying drawings in which:

[0018] It is to be understood that the drawings are designed solely for the purpose of illustration and are not intended to limit the scope of the present application in any way. Further, it should be understood that the figures are merely meant to illustrate the basic concept of the present application and as such they are not to be considered in a limiting sense. Specifically, the figures do not show the components in their actual number, shape and size, but only show the components that are relevant to the present application. The actual number, shape and size of the components can be arbitrarily changed and the layout of the components can be more complicated.

[0019] In the following description, numerous specific details are discussed in order to provide a thorough understanding of the embodiments of the present application. However, it will be apparent to one of ordinary skill in the art that the embodiments of the present application can be practiced without these specific details. In other instances, well-known structures and devices are shown in block diagram form in order to avoid obscuring the embodiments of the present application. References to items in the singular are not intended to be limited to a single item unless specifically so stated, but rather, exceptions can exist that are both singular and plural, and the use of "one" or "the other" can be taken to mean one, both, or either.

[0020] Reference will now be made to the drawings, in which Figures 1-2The image shows the real-time monitoring system and method for volatile organic compound (VOC) leakage according to the present invention. The real-time monitoring system for VOC leakage of the present invention includes a multi-dimensional sensing module, which collects VOC concentration data and environmental meteorological data in real time to form a raw monitoring signal containing concentration values ​​and meteorological parameters; an edge computing module, which performs data filtering, outlier removal, and concentration trend analysis on the raw monitoring signal to form a preprocessed signal indicating potential abnormal states; a central processing module, which receives the preprocessed signal, integrates multi-source monitoring data and real-time meteorological information, calculates the coordinates of potential leakage sources using a leakage tracing algorithm based on a Gaussian plume model, and generates a leakage location signal and a real-time risk level signal; an early warning management module, which receives the real-time risk level signal and generates a graded early warning signal containing alarm levels based on preset low, medium, and high risk thresholds and corresponding action rules; and a visualization platform module, which receives the leakage location signal and the graded early warning signal, dynamically renders a VOC concentration distribution cloud map based on a digital twin model of the plant area, and displays the leakage source location and risk level in a three-dimensional virtual plant area, generating a visualized monitoring interface signal.

[0021] like Figure 1As shown, the core framework of the system of this invention is constructed, which is a complete technical solution integrating data perception, edge processing, central decision-making, early warning management, and visualization. The operation of the system begins with the multi-dimensional sensing module, which is the "nerve ending" of the system's perception of the physical world. It continuously captures two key data from the monitoring area: volatile organic compound (VOC) concentration data and environmental meteorological data. The acquisition of concentration data relies on highly sensitive gas sensors, which can respond to VOC molecules at ppb or even lower levels, converting chemical signals into precise electrical signal readings. At the same time, meteorological sensors synchronously collect parameters such as wind speed, wind direction, ambient temperature, atmospheric humidity, and air pressure. These raw readings, after preliminary signal conditioning and analog-to-digital conversion, are encapsulated into a structured data packet, namely the raw monitoring signal. This signal not only contains the concentration value and meteorological parameters of each monitoring point at a specific timestamp, but also usually includes device identification and geographical location information, laying the foundation for subsequent data tracing and spatial analysis. Next, the raw monitoring signal is transmitted to the edge computing module. This module is a computing unit deployed at the monitoring site or locally on the equipment. Its core value lies in decentralizing computing power, performing preliminary intelligent processing at the source of data generation. After receiving the raw data stream from the sensor module, it first performs data filtering, using digital filters (such as Kalman filters or low-pass filters) to suppress random errors caused by electromagnetic interference, instantaneous sensor fluctuations, or environmental background noise, thereby smoothing the data curve and extracting the true concentration trend. Next is outlier removal. This module uses statistical process control methods, such as dynamic threshold judgment based on moving averages and standard deviations, to identify and remove abnormal data points that significantly deviate from the normal range and may be caused by non-leakage factors (such as the brief passage of vehicle exhaust). Finally, and most importantly, the edge computing module performs concentration trend analysis. It no longer focuses solely on the absolute concentration value at a single point in time, but instead calculates the rate of change (first derivative, i.e., slope) and acceleration of change (second derivative) of the concentration by analyzing historical data sequences within a sliding time window. When the system detects that the concentration value is not only continuously rising, but the acceleration of this rise exceeds an adaptive threshold dynamically calculated based on historical baselines, it determines the current state as a "potential anomaly" and labels the data accordingly. At this point, the massive, redundant, and noisy raw monitoring signals are refined into preprocessed signals that are smaller, have higher information density, and contain preliminary diagnostic conclusions. This process greatly reduces the burden on the backend communication network and central processing unit, and enables early, localized identification of anomalies.

[0022] Furthermore, the preprocessed signals, processed by the edge computing module, are aggregated to the system's "brain"—the central processing module—via wired or wireless networks. This is the core of complex data fusion and intelligent decision-making. This module receives preprocessed signals from all monitoring points within the region, and its primary task is multi-source data fusion. This means it doesn't simply display the concentration data of each point on a map, but rather performs spatiotemporal alignment and correlation analysis of concentration data, precise real-time meteorological information (especially wind direction and speed fields), and equipment operating status (such as pump start / stop and valve opening) that may be obtained from other system interfaces. The fused data is then fed into a leak source tracing algorithm based on a Gaussian plume model. This algorithm mathematically simulates the physical process of pollutant diffusion in the atmosphere, treating each possible leak source as a virtual point source and calculating the concentration distribution of pollutants emitted from that point source downwind based on the current wind field conditions. Then, the algorithm iteratively optimizes (e.g., using least squares) the location and intensity of the virtual leak sources to minimize the error between the calculated theoretical concentration distribution and the concentration distribution actually observed by the monitoring network. Ultimately, the location of the virtual leak source that minimizes the error is determined by the algorithm as the most likely potential leak source coordinates, generating a leak location signal. Simultaneously, the central processing module runs a more complex risk assessment model, such as a pre-trained machine learning classifier (e.g., random forest or support vector machine). This model uses the concentration trend characteristics in the pre-processed signal, the computational intensity of the leak source, the process hazard classification of the leak source location (e.g., whether it is near a fire source or densely populated area), and the potential impact range of pollutants under current meteorological conditions as input features. The model performs nonlinear weighted and comprehensive analysis on these multi-dimensional features, outputting a comprehensive risk assessment conclusion—a real-time risk level signal. This level is no longer a simple "exceeding the standard" or "not exceeding the standard," but can be subdivided into different levels such as "low risk," "medium risk," and "high risk," accurately reflecting the potential severity and urgency of the leak event. The leak location signal and real-time risk level signal generated by the central processing module are sent separately and synchronously to the early warning management module and the visualization platform module. The early warning management module acts as the system's "command center," internally pre-setting thresholds and action rule bases corresponding to different risk levels. Upon receiving a real-time risk level signal, it immediately matches it with preset rules. For example, for a "low-risk" signal, it may only generate a reminder message that requires further verification; for a "medium-risk" signal, it will generate a clear alarm and may automatically create a maintenance work order; for a "high-risk" signal, it will trigger the highest level alarm, generate a tiered warning signal that includes recommendations for emergency evacuation and process shutdown, and forcibly notify the preset safety officer and emergency response team through multiple channels such as audible and visual alarms, SMS, and application push notifications, ensuring the timeliness and effectiveness of information transmission.

[0023] Specifically, the system's presentation layer is implemented by a visualization platform module. This module receives leak location signals from the central processing module and tiered early warning signals from the early warning management module. It operates based on a digital twin model of the plant area—a three-dimensional virtual model that perfectly corresponds to the physical plant area. In this model, the system not only accurately marks the coordinates indicated by the leak location signals in the form of icons, but also dynamically renders a cloud map of the volatile organic compound (VOC) concentration distribution across the entire plant area based on real-time data. This cloud map is not static; it simulates the diffusion path and concentration gradient of pollutants in real time according to wind direction and speed, visually displaying changes in risk areas using different colors (e.g., from green to red). Simultaneously, the color and level of the tiered early warning signals are highlighted at the corresponding leak source locations and on the corresponding equipment. All this information together constitutes a visualized regulatory interface signal integrating real-time monitoring, alarm management, risk location, and historical tracing functions, presented on a large screen or terminal computer, enabling safety management personnel to have a comprehensive overview and achieve all-round, three-dimensional perception and command of leak events.

[0024] In one embodiment of the invention, the multi-dimensional sensing module is further defined in greater detail, elaborating on its spatial deployment strategy and collaborative working mechanism, which is crucial for achieving accurate monitoring and preliminary spatial identification. This module is not simply a collection of single-type sensors, but rather a carefully designed heterogeneous network of monitoring devices in three forms: "points," "lines," and "surfaces," based on spatial scale and functional positioning. First, there is a point sensor array deployed near key equipment in the production area. These "points" are the densest and most advanced units in the monitoring network, directly installed around static and dynamic sealing points most prone to leakage, such as pumps, compressors, valves, flanges, and connectors. Their core task is to capture instantaneously high-concentration leakage signals in a localized area, requiring extremely high sensitivity to react quickly to even minute leaks, serving as direct evidence for discovering the "first scene" of a leak. Second, there are linear scanning monitoring stations deployed along the perimeter of the plant. These "lines" constitute a protective barrier between the plant and the external environment. These devices typically employ open-path spectroscopy (such as non-dispersive infrared or ultraviolet differential absorption spectroscopy) or higher-precision scanning lidar to continuously monitor the plant boundary horizontally or vertically. Their focus is not on a specific leaking device, but rather on macroscopically monitoring whether volatile organic compounds (VOCs) are diffusing from the plant area into the external environment, and the overall trend, flux, and concentration levels of this diffusion. This provides data support for assessing the impact on the external environment and complements leaks that point arrays might miss. Finally, there are area-based environmental micro-stations covering the entire plant area. These "area" devices are usually installed at higher locations, such as towers or plant rooftops, providing an overview of the entire production area below. They provide background concentration distribution values ​​for the entire plant area, helping to identify large-scale, diffuse leaks, or clarifying the overall pollution situation when multiple point sources exist simultaneously. The key innovation lies in the fact that the multi-dimensional sensing module does not allow these three network layers to work independently. Instead, it uses a data fusion gateway to synchronize, spatially register, and standardize sensor data from points, lines, and surfaces at different spatial scales and potentially using different technological principles. This data is then integrated into a unified raw monitoring signal with spatial hierarchy information. This heterogeneous network fusion design enables the system to possess multi-perspective perception capabilities from the source of data acquisition, ranging from microscopic to macroscopic and from local to global perspectives. This provides a far richer and more reliable data foundation for the backend central processing module to perform more accurate source tracing calculations compared to a single-dimensional monitoring network.

[0025] like Figure 1As shown, the edge computing module has deepened and expanded its core algorithm functions and reliability assurance mechanisms. First, it defines in detail the specific implementation of the key operation, "concentration trend analysis." This operation is not a simple comparison of magnitudes, but a dynamic, time-series-based intelligent analysis process. It uses a sliding time window to continuously capture concentration data sequences over a recent period. Within this window, the module calculates two key mathematical features in real time: one is the slope of the concentration sequence, i.e., the average change in concentration per unit time, reflecting the speed of leak development; the other is the acceleration of change, i.e., the rate of change of the slope itself, which reveals whether the leak is accelerating, decelerating, or maintaining a constant expansion rate, crucial for determining the urgency of the leak. The system compares these two calculated feature values ​​with an adaptive threshold. This adaptive threshold is not fixed but dynamically adjusted based on historical baselines established through long-term sensor monitoring (including diurnal fluctuations, seasonal changes, etc.), enabling the system to adapt to different operating conditions and environmental backgrounds, reducing misjudgments. Only when the system identifies that the concentration value not only continues to rise but also that its acceleration exceeds the current adaptive threshold will it ultimately be determined as a "potentially abnormal state." This dual judgment mechanism based on trends and acceleration can effectively filter out instantaneous concentration peaks caused by brief changes in meteorological diffusion conditions, or background interference caused by drifting from distant leak sources, significantly improving the accuracy of anomaly identification. Furthermore, it endows the edge computing module with an important reliability feature—a local caching mechanism. Considering the complex electromagnetic environment of industrial sites, communication links (especially wireless networks) are at risk of interruption. This mechanism ensures that even in the event of a complete communication interruption, the edge computing module does not cease operation but utilizes its built-in storage space to continuously and uninterruptedly record all preprocessed signals and critical intermediate computational data. Once the communication link is detected to be restored, the module immediately initiates a data retransmission program, transmitting the cached data from the interruption period to the central processing module in an orderly and complete manner according to time sequence. This design completely eliminates the loss of monitoring data due to communication failures, ensuring that every valid data point acquired by the system is recorded and uploaded, thereby guaranteeing the continuity and integrity of monitoring data. This has invaluable significance for accident tracing analysis, compliance reporting, and historical trend studies. It is precisely these sophisticated algorithms and reliable engineering designs that enable the edge computing module to truly achieve a qualitative leap from a "data collector" to an "intelligent preprocessing node".

[0026] Furthermore, the preprocessed signals, after edge intelligent processing, are converged to the central processing module—the system's nervous system—through an industrial IoT network. This module undertakes the most complex data fusion, model calculation, and decision-making tasks of the entire system. It receives preprocessed signals from all edge nodes within the monitoring area and deeply integrates them with real-time meteorological information streams. This integration involves not only temporal alignment but also spatial correlation. Based on this, the core leak source tracing algorithm based on the Gaussian plume model is activated. Mathematically, this algorithm constructs a physically driven diffusion model that treats the atmosphere as a flowing medium and assumes a potential leak source as a point source with a specific location and intensity. By inputting real-time wind field data, the model can simulate and calculate the theoretical concentration distribution formed by the virtual leak source downwind. Subsequently, the algorithm continuously adjusts the location and emission intensity of this virtual leak source on the electronic map through iterative optimization techniques and compares its simulation results with the concentration distribution reported by the actual monitoring network to find the optimal solution that minimizes the difference between the two. The spatial coordinates corresponding to this optimal solution are determined as the coordinates of the most likely potential leak source, and a precise leak location signal is generated. Simultaneously, the central processing module also runs a more complex risk assessment engine. This engine, likely a machine learning model trained on extensive historical data, comprehensively considers multiple factors such as the real-time concentration exceedance multiple, the rate of concentration increase, the estimated duration of the leak, the process hazard level of the leak source location, and the diffusion impact range under current meteorological conditions. Through a non-linear comprehensive evaluation function, it outputs a quantitative real-time risk level signal categorized into different levels (e.g., low, medium, high). This signal accurately characterizes the severity and urgency of the current leak situation. This significantly deepens and expands the breadth of data fusion and the accuracy of the source tracing algorithm within the central processing module. The core of this claim lies in the fact that the multi-source monitoring data fused by the central processing module is no longer limited to environmental concentration and meteorological information provided by multi-dimensional sensing modules, but further incorporates rich and critical equipment operating status parameters provided by the enterprise's production management system. These parameters include, but are not limited to, the start-up and shutdown status of critical equipment, real-time pressure readings inside pipelines or containers, and temperature parameters of process fluids. This data from the production control domain provides unprecedented context and prior knowledge for leak source tracing. After receiving these equipment operating status parameters, the central processing module does not display them in isolation. Instead, it uses them as powerful constraints and correction factors, deeply embedding them into the leak tracing algorithm based on the Gaussian plume model. Specifically, when iteratively calculating the coordinates of the most likely leak source, the algorithm prioritizes potential sources associated with abnormal equipment states. For example, the probability of a newly started pump or a pipe section with an abnormally rising pressure reading being a leak source is significantly increased by the algorithm.Conversely, for a device in a standby state with zero pressure, even if nearby sensors detect concentration, the algorithm will more likely consider it to have originated from a distant leak source or be subject to other interference, thus reducing its weight as the primary suspect. This deep data fusion mechanism transforms the source tracing algorithm from a "geometric problem solver" relying solely on atmospheric physical diffusion models into an "analysis expert" capable of incorporating actual industrial operating conditions. It can effectively correct for potential errors or absurd results that contradict process logic when relying solely on environmental meteorological information for back-dive analysis. For example, under complex meteorological conditions such as calm winds or fluctuating wind directions, relying solely on concentration gradients for source tracing might point to a vague area or even the wrong direction. However, if a device happens to report a high-pressure alarm, the algorithm can quickly focus its computational resources on the area surrounding that device, combining it with weak concentration gradient signals to provide a more physically and technologically reliable and accurate leak source coordinate calculation result. This significantly improves the accuracy and engineering reliability of the potential leak source coordinate calculation under complex operating conditions, making the system's decision output more convincing and practical.

[0027] like Figure 2 As shown, the present invention also includes a real-time monitoring method for volatile organic compound (VOC) leaks, comprising: S1: real-time acquisition of VOC concentration data and environmental meteorological data to form a raw monitoring signal containing concentration values ​​and meteorological parameters; S2: performing data filtering, outlier removal, and concentration trend analysis on the raw monitoring signal to form a preprocessed signal indicating potential abnormal states; S3: receiving the preprocessed signal, fusing multi-source monitoring data and real-time meteorological information, calculating the coordinates of potential leak sources using a leak tracing algorithm based on a Gaussian plume model, and generating a leak location signal and a real-time risk level signal; S4: receiving the real-time risk level signal, generating a graded early warning signal containing alarm levels based on preset low, medium, and high risk thresholds and corresponding action rules; S5: receiving the leak location signal and graded early warning signal, dynamically rendering a VOC concentration distribution cloud map based on a digital twin model of the plant area, and displaying the leak source location and risk level in a three-dimensional virtual plant area, generating a visual monitoring interface signal.

[0028] Specifically, the inherent generation mechanism of the real-time risk level signal generated by the central processing module elevates it from a judgment that might be based on simple rules to an intelligent quantitative output that reflects the multidimensional characteristics of a leak event, based on a multi-factor comprehensive evaluation model. The design philosophy of this model is that the risk level of a leak event cannot be determined solely by the instantaneous concentration. Therefore, the model simultaneously considers and quantifies five key influencing factors. The first factor is the real-time concentration exceedance multiple, i.e., the extent to which the current monitored concentration exceeds the safety threshold or background value, reflecting the absolute severity of the leak. The second factor is the concentration rise rate, which reveals whether the leak is a slow seepage or a rapid ejection; the latter obviously indicates a more urgent situation. The third factor is the estimated duration of the leak event, which can be inferred by combining the duration of the sustained concentration exceedance with the time of the first alarm. Even if the concentration is not high, a prolonged leak will continuously increase its cumulative emissions and potential risks. The fourth factor is the sensitivity of the leak source's location. This is a weighted factor based on prior knowledge. For example, if the leak source is located near an open flame source, a densely populated area, a sensitive environmental receptor, or an area storing flammable or explosive materials, its inherent risk level will naturally be much higher. The fifth factor is the diffusion impact range under current meteorological conditions. Based on real-time wind speed, atmospheric stability, and other parameters, the system simulates and predicts the downwind area that the pollutant plume may affect. The larger the impact range, the higher the risk. These five factors are not evaluated in isolation but are input into a pre-defined weighted fusion function, which may be trained using historical machine learning data. This function assigns appropriate weights to each factor and outputs a quantified comprehensive risk index through a non-linear calculation process. This index is a continuous or piecewise continuous value that comprehensively reflects the combined effect of all the above factors. Finally, based on this calculated comprehensive risk index, the system maps it to several pre-defined risk level ranges, such as "low risk," "medium risk," "high risk," and even "extremely high risk." The advantage of this evaluation mechanism lies in the fact that it ensures that risk level classification is no longer subjective or arbitrary, but rather a traceable and scientifically quantified result based on multi-dimensional objective data. For example, a momentary leak with a very high concentration exceeding the standard but located in a remote area with no sensitive points downwind might be assessed as "medium risk" in terms of its comprehensive risk index; while a leak with a concentration not extremely exceeding the standard but continuous and spreading towards residential areas outside the plant might be assessed as "high risk." This refined risk assessment capability is the core of achieving accurate early warning and optimized emergency resource allocation. The function of the early warning management module extends from the information generation end to the action execution end, endowing it with the ability to trigger automated and process-oriented emergency responses, thereby achieving seamless connection from perception to disposal. This claim specifies in detail the automatic linkage mechanism between graded early warning signals and specific disposal procedures.Once the early warning management module generates a tiered early warning signal, it's no longer just a notification on the screen, but immediately becomes a trigger driving a series of subsequent actions. For leaks classified as medium-risk, the system doesn't just issue an alarm; it automatically generates a detailed inspection work order in the company's computerized maintenance management system or work order management platform according to preset rules. This work order automatically includes leak location information, risk level, and recommended inspection items, and is pushed to the mobile smart terminals of relevant inspection personnel via an interface, guiding them to the site for verification and initial handling, ensuring that alarm information is promptly and accurately translated into concrete actions. For high-risk signals, the system's response is even faster and more decisive. In addition to generating the highest-level alarm, it immediately and automatically triggers the on-site audible and visual alarms, alerting personnel in the area through visual and audible signals. Simultaneously, it sends an emergency message to the communication devices of key personnel such as safety managers and emergency command personnel. This emergency message not only informs of a "leak," but also includes emergency response guidelines, such as recommended process isolation steps, the scope of personnel to be notified, and evacuation recommendations, providing decision-makers with immediate support for incident handling. This automated closed-loop response mechanism, from early warning to handling, greatly reduces the time window from the discovery of anomalies to the initiation of response measures. It avoids delays, omissions, or misinterpretations that may occur due to manual information transmission, transforming the traditional passive management model of "people looking for problems" into an active early warning and automated response model of "problems looking for people," significantly improving the efficiency and safety of enterprises in responding to sudden leakage incidents.

[0029] Furthermore, the dynamic rendering and historical analysis capabilities of the visualization platform module place higher demands on it, evolving it from a static display tool into a strategic decision support platform capable of simulation, prediction, and retrospective analysis. This claim first defines the dynamically rendered volatile organic compound (VOC) concentration distribution cloud map, whose generation process must be closely integrated with real-time meteorological wind direction and speed data for dynamic simulation. This means the cloud map is not simply a contour map formed by interpolation based on monitoring point readings, but a predictive, dynamic visualization result based on a physical diffusion model. The system uses the leak source coordinates and intensity calculated by the central processing module as initial conditions, and real-time high-precision wind direction, wind speed, and atmospheric turbulence data as driving fields to perform rapid diffusion simulation in a digital twin model. The final generated concentration distribution cloud map, using different color gradients, not only intuitively displays the spatial distribution of concentration at the current moment, but more importantly, it can simulate and show the path of pollutants spreading downwind from the leak source, and the process of gradual dilution along its path with increasing distance. This dynamic cloud map allows managers to clearly predict the movement direction, impact range, and trajectory of the core high-concentration area of ​​the pollution zone, providing forward-looking information for emergency decisions such as evacuation and post-disaster deployment. Furthermore, claim 7 endows the module with powerful post-event analysis capabilities—replaying the entire leakage and diffusion process within a specific time period based on historical monitoring data. Users can select any past time period in which a leak occurred, and the system will extract all monitoring and meteorological data stored in the database for that period, arranged in time sequence, and re-enact the complete spatiotemporal evolution of the leak's occurrence, development, diffusion, and eventual dissipation in a digital twin model in the form of animation. This playback function is irreplaceable for accident review and source tracing analysis. Safety engineers can use slow motion, pause, and multi-angle observation to meticulously analyze the leak's starting point, the main influencing factors of diffusion, monitoring blind spots, and the effectiveness of emergency measures, thereby summarizing lessons learned, optimizing sensor layout, improving emergency plans, and ultimately enhancing the overall enterprise's safety management level and accident prevention capabilities. This makes the system of this invention not only a real-time monitoring tool but also a knowledge base and analysis platform for continuous improvement of the safety management system.

[0030] The present invention relates to a real-time monitoring system and method for volatile organic compound (VOC) leaks. This system collects VOC concentration and meteorological data in real time through a multi-dimensional sensor network. An edge computing module performs data preprocessing and preliminary anomaly assessment at the monitoring terminal, improving response efficiency and reducing cloud load. A central processing module integrates multi-source information, employs a Gaussian plume-based source tracing algorithm to accurately calculate the leak source coordinates, and combines this with a machine learning model to assess the risk level. Finally, a tiered alarm system is implemented through an early warning management module, and a digital twin-based visualization platform dynamically displays the leak diffusion trend and risk location, thereby achieving integrated intelligent monitoring from perception, location, early warning to visualization management.

[0031] Therefore, the real-time monitoring system and method for volatile organic compound (VOC) leakage of the present invention solves the problem of difficulty in real-time monitoring, accurate location and intelligent early warning of VOC leakage.

[0032] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. A volatile organic compound leak real-time monitoring system, characterized by, The method comprises the following steps: A multi-dimensional sensing module acquires volatile organic compound concentration data and environmental meteorological data in real time to form original monitoring signals containing concentration values and meteorological parameters; An edge computing module receives the original monitoring signals, performs data filtering, outlier rejection, and concentration trend analysis operations to form preprocessed signals indicating potential abnormal states; A central processing module receives the preprocessed signals, fuses multi-source monitoring data and real-time meteorological information, calculates potential leakage source coordinates through a leakage source tracing algorithm based on a Gaussian plume model, generates a leakage positioning signal and a real-time risk level signal; An early warning management module receives the real-time risk level signal, generates a hierarchical early warning signal containing alarm levels according to preset low, medium, and high risk thresholds and corresponding action rules; A visualization platform module receives the leakage positioning signal and the hierarchical early warning signal, dynamically renders a volatile organic compound concentration distribution cloud map based on a digital twin model of the plant, displays the leakage source location and the risk level in a three-dimensional virtual plant, and generates a visualization supervision interface signal.

2. The volatile organic compound leak real-time monitoring system of claim 1, wherein, The multi-dimensional sensing module comprises a point sensor array arranged near key equipment in the production device area, a line scanning monitoring station arranged at the perimeter of the plant, and a planar regional environmental micro station covering the entire plant overhead. The point sensor array is used to capture local high-concentration leakage signals, the line scanning monitoring station is used to monitor the diffusion trend of volatile organic compounds at the plant boundary edge, and the planar regional environmental micro station is used to obtain the concentration background distribution of the entire plant. The multi-dimensional sensing module integrates monitoring data from different spatial scales into a unified original monitoring signal through the fusion of heterogeneous networks.

3. The volatile organic compound leak real-time monitoring system of claim 1, wherein, The concentration trend analysis operation performed by the edge computing module specifically calculates the change slope and acceleration of the concentration sequence using a sliding time window. When a continuous rise in concentration value and a change acceleration exceeding an adaptive threshold are identified, a potential abnormal state is determined. The edge computing module also has a local caching mechanism that continuously records preprocessed signals during communication interruptions and performs data retransmission after communication is restored, ensuring the continuity and integrity of monitoring data.

4. The volatile organic compound leak real-time monitoring system of claim 1, wherein, The multi-source monitoring data fused by the central processing module also includes equipment start-stop states, pipeline internal pressure and temperature parameters provided by the enterprise production management system. The central processing module embeds the equipment operating state parameters as prior knowledge into the leakage source tracing algorithm to correct errors produced by tracing calculations relying solely on environmental meteorological information, thereby improving the calculation accuracy and reliability of potential leakage source coordinates in complex working conditions.

5. The volatile organic compound leak real-time monitoring system of claim 1, wherein, The real-time risk level signal generated by the central processing module is produced by a multi-factor comprehensive evaluation model that considers real-time concentration exceedance multiples, concentration rise rates, estimated duration of leakage events, sensitivity of the location of the leakage source, and the diffusion influence range under current meteorological conditions. A quantitative comprehensive risk index is output by weighted fusion of these factors, and the index is used to divide risk levels.

6. The volatile organic compound leak real-time monitoring system of claim 1, wherein, The early warning management module generates a graded early warning signal that automatically triggers a corresponding treatment process. For a medium risk level, a patrol work order is automatically generated and pushed to a mobile patrol terminal. For a high risk level, an audible and light alarm is automatically triggered, and an emergency message containing emergency treatment guidelines is sent to the safety supervisor's communication device, realizing a quick closed-loop response from early warning to treatment.

7. The volatile organic compound leak real-time monitoring system of claim 1, wherein, The visualization platform module dynamically renders a volatile organic compound concentration distribution cloud map that is dynamically simulated in combination with real-time meteorological wind direction and speed data. The cloud map visually displays the concentration diffusion path and dilution process from the leakage source to the outside with different color gradients. It can also replay the entire leakage diffusion process within a specific time period based on historical monitoring data, which is used for accident review and traceability analysis.

8. The volatile organic compound leak real-time monitoring system of claim 1, wherein, The visualization platform module also integrates a device health management submodule that establishes an electronic file for each monitored device, records its historical leakage events, maintenance records, and operation parameters related to leakage, and calculates a dynamic health score for the device based on these data. The health score is displayed as background information in the visualization management interface along with the real-time risk level signal.

9. The volatile organic compound leak real-time monitoring system of claim 1, wherein, The system also includes an audit tracking module connected to the early warning management module and the visualization platform module. The audit tracking module continuously receives and stores the graded early warning signal, the leakage positioning signal, and the original data snapshot on which the early warning is based, forming a complete audit log with a time stamp. This log records all chain events from the initial appearance of anomalies, risk assessment, early warning release, to subsequent treatment, supporting queries and traceability by time, location, or device.

10. The intelligent suppression management method of the volatile organic compound leakage real-time monitoring system according to claims 1-9, comprising: S1: Real-time acquisition of volatile organic compound concentration data and environmental meteorological data to form original monitoring signals containing concentration values and meteorological parameters; S2: Receive the original monitoring signals, perform data filtering, outlier rejection, and concentration trend analysis operations to form preprocessed signals indicating potential abnormal states; S3: Receive the preprocessed signals, fuse multi-source monitoring data and real-time meteorological information, calculate potential leakage source coordinates through a leakage traceability algorithm based on the Gaussian plume model, generate leakage positioning signals and real-time risk level signals; S4: Receive the real-time risk level signals, generate graded early warning signals containing alarm levels according to pre-set low, medium, and high risk thresholds and corresponding action rules; S5: Receive the leakage positioning signals and graded early warning signals, dynamically render volatile organic compound concentration distribution cloud maps based on the plant digital twin model, display the leakage source location and risk level in a three-dimensional virtual plant, and generate visualization management interface signals.

Images