AMC sampling point distribution optimization method based on singular point fusion of multiple pollution sources

Abstract

The invention discloses an AMC sampling point distribution optimization method based on multi-pollution source singular point fusion, and relates to the field of molecular pollution sampling, and the method comprises the steps: extracting the respiratory capacity and the skin secretion amount generated by an operator, and taking the respiratory capacity and the skin secretion amount as molecular pollution loads; extracting molecular pollutant concentration data in the historical operation process in the monitoring area, and coupling the molecular pollutant concentration data with the number of corresponding operators to obtain a causal relationship between the molecular pollution load and the molecular pollutant concentration; the number and distribution positions of operators in the monitoring area in the target monitoring time period are obtained, the molecular pollutant concentration of the monitoring area is predicted by combining the number of the operators with the causal relationship, and singular points of multiple pollution sources are determined; and generating an initial distribution diagram of the monitoring sampling points, and optimizing the initial distribution diagram to determine a final monitoring sampling point according to the application cost under the condition of the initial distribution diagram. According to the method, the monitoring cost can be reduced, the sampling point layout efficiency can be improved, and it is ensured that the key area of the monitoring area is sufficiently covered.

Description

Technical Field

[0001] The present invention relates to the field of molecular pollution sampling, and in particular to an AMC sampling point distribution optimization method based on the fusion of singular points of multiple pollution sources. Background Art

[0002] AMC refers to airborne molecular contaminants, including organic and inorganic pollutants in clean room air, which exist in the form of gas, vapor and airborne dust, such as acids, alkalis, polymer additives, organometallic compounds, etc. AMC pollutants are usually the key source of pollution in the high-precision manufacturing industry. The products of the high-precision manufacturing industry are very sensitive to tiny pollutants in the environment. Any tiny molecular pollutants (such as acids, alkalis, polymer additives, metal compounds, etc.) may affect the quality and performance of the products. In order to obtain the AMC concentration in the air, sampling monitors are usually used to monitor the air.

[0003] The purpose of AMC sampling is to monitor and control airborne molecular pollutants in controlled environments such as clean rooms to ensure that environmental quality meets the strict requirements of specific industries. However, the existing AMC monitoring sampling points are often designed according to the layout of the monitoring area during the layout process, and the monitoring application costs are ignored during the monitoring process. This makes it impossible to achieve maximum monitoring efficiency within narrow budget constraints, increases operating costs, and reduces the rationality of resource allocation. Summary of the Invention

[0004] In order to solve the above problems, the present invention proposes an AMC sampling point distribution optimization method based on multi-pollution source singular point fusion, which achieves the purpose of reducing monitoring costs while improving the efficiency of sampling point layout.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions: The present invention provides an AMC sampling point distribution optimization method based on multi-pollution source singular point fusion, comprising: Simulate the operator's working behavior in the monitoring area, and extract the operator's respiratory volume and skin secretion volume generated within the target time period based on the work behavior simulation results as the molecular pollution load; Extract the molecular pollutant concentration data of the historical operation process in the monitoring area and the corresponding number of operators for coupling, and obtain the causal relationship between molecular pollution load and molecular pollutant concentration based on the coupling results; Obtain the number and distribution of operators in the monitoring area during the target monitoring period, and combine the number of operators with causal relationships to predict the concentration of molecular pollutants in the monitoring area and determine the singular points of multiple pollution sources; An initial distribution map of monitoring sampling points is generated based on the singular points of multiple pollution sources and the distribution positions of operators. The initial distribution map is optimized to determine the final monitoring sampling points according to the application cost under the conditions of the initial distribution map.

[0006] Preferably, the working behavior of the operator in the monitoring area is simulated, and the respiratory volume and skin secretion volume generated by the operator within the target time are extracted according to the work behavior simulation results as the molecular pollution load, including: Generate operator work tasks based on the equipment work content in the monitoring area, decompose the work tasks into behavior action nodes, and set action operation parameters for the behavior action nodes to obtain execution condition nodes; Create a blank behavior tree, add an execution condition node to it to generate a behavior rule engine, associate the behavior rule engine with a physical operator, and run the blank behavior tree to obtain the operator's work behavior. The respiratory rate and sweating degree of the operator under working behavior conditions are obtained, and a behavior-physiology coupling model is constructed to predict the respiratory volume and skin secretion volume generated by the operator within the target time, and obtain the molecular pollution load.

[0007] Preferably, create a blank behavior tree, add an execution condition node to the blank behavior tree to generate a behavior rule engine, and associate the behavior rule engine with an actual operator. Running the blank behavior tree to obtain the operator's work behavior includes: Launch the Behavior Tree Creation Wizard based on the Autonomous Behavior Rule Editor to generate a blank Behavior Tree framework containing several root nodes and child node containers. Use the Ruler to define the behavior logic for the root nodes and child nodes. Obtain a blank behavior tree based on the behavior logic definition results, add execution condition nodes to the blank behavior tree nodes to extract operation instruction information, and generate a behavior rule engine based on the extracted results; Associating the behavior rule engine with the physical operator simulation model and sending operation instruction information to the physical operator simulation model to determine the behavior path of the physical operator in the virtual monitoring area; In the simulation environment, a blank behavior tree structure is run according to the behavior path of the physical operator, and the operator's work behavior in the virtual monitoring area is obtained based on the running results.

[0008] Preferably, the respiratory rate and sweating degree of the operator under working behavior conditions are obtained, a behavior-physiology coupling model is constructed, and the respiratory volume and skin secretion volume generated by the operator within the target time are predicted to obtain the molecular pollution load, including: The respiratory rate and sweating degree of the operator are extracted based on the working behavior in the virtual monitoring area, and the respiratory rate and sweating degree are combined to generate a bivariate observation sequence; The cluster centers are iteratively optimized using Euclidean distance as the similarity metric, and cluster analysis is performed on the bivariate observation sequence. The physiological behavior pattern categories of the bivariate observation sequence are then divided according to the cluster analysis results. Based on the physiological behavior pattern categories, the Bayesian prior distribution technology was used to establish a behavioral coupling model, and the Markov chain Monte Carlo technology was used to sample the posterior distribution of the behavioral coupling model. The temporal dynamic parameters are estimated based on the sampling results, and the operator's breathing rate and sweating degree within the target duration are predicted. The breathing rate and sweating degree are converted into the total volume of respiratory gas and skin sweat secretion through integral changes, generating a physiological output sequence that changes with time and obtaining the molecular pollution load.

[0009] Preferably, extracting molecular pollutant concentration data from historical operations within the monitoring area and the corresponding number of operators for coupling, and obtaining a causal relationship between molecular pollution load and molecular pollutant concentration based on the coupling result includes: Collect historical pollutant concentration data and the number of operators in the monitoring area, divide the historical molecular pollutant concentration data and the number of operators into several time periods, determine the fluctuation characteristics of the historical molecular pollutant concentration data and the change pattern of operators in each time period, and generate a time series data set; Based on time series data sets and molecular pollution loads, neural networks are used for time series modeling to capture the long-term dependency between historical molecular pollutant concentration changes and molecular pollution loads, generating a causal network. A sensitivity analysis was performed on the causal network to evaluate the contribution of the number of operators in the causal network. The causal network was optimized based on the contribution to capture the potential causal relationship between the time series dataset and the molecular pollution load.

[0010] Preferably, generating an initial distribution map of monitoring sampling points based on multiple pollution source singular points and operator distribution positions, and optimizing the initial distribution map to determine the final monitoring sampling points based on the application cost under the initial distribution map conditions includes: Determine the upper limit of molecular pollutant concentration in the monitoring area based on the singular points of multiple pollution sources, and determine the maximum number of monitoring sampling points based on the upper limit and the activity density of the distribution of operators; According to the maximum number of monitoring sampling points and the layout of the monitoring area, the axis alignment and oriented bounding box technology are used to construct a bounding box, and the initial monitoring sampling point distribution map is generated based on the bounding box; Obtain the purchase price of the monitoring sampler, and calculate the application cost generated by the initial monitoring sampling point distribution map based on the purchase price. Establish a minimum objective function based on the application cost, optimize the initial distribution map and determine the final monitoring sampling points.

[0011] Preferably, according to the maximum number of monitoring sampling points and the layout state of the monitoring area, a bounding box is constructed using axis alignment and oriented bounding box technology, and an initial monitoring sampling point distribution map is generated based on the bounding box, including: Obtain the maximum number of monitoring sampling points and the layout status of the monitoring area, and use the axis-aligned bounding box and oriented bounding box bounding box construction technology to determine the optimal coordinate axis direction in combination with the layout status of the monitoring area; Construct a bounding box that completely covers the monitoring area based on the optimal coordinate axis direction, and project the monitoring sampling points to the nearest plane of the bounding box according to the layout status of the monitoring area, and output the location of the monitoring sampling points; The distance between the monitoring sampling point and the entity in the monitoring area is determined according to the location of the monitoring sampling point, and the monitoring sampling points with a distance higher than the distance threshold are eliminated, and the eliminated monitoring sampling points are re-projected; Until the difference between the monitoring sampling point location and the physical distance within the monitoring area is less than or equal to the distance location, the initial monitoring sampling point distribution map is obtained.

[0012] The beneficial effects of the present invention are: 1. By establishing a causal relationship between pollution load and pollutant concentration, the present invention accurately reveals the dynamic changes between the two, providing a more scientific basis for future monitoring and prediction. In addition, after combining the number and distribution of operators in the monitoring area, the pollutant concentration is more accurately predicted within the target monitoring period, and the singular points of multiple pollution sources are accurately identified, laying the foundation for optimizing the sampling point layout. At the same time, the optimization of the initial monitoring sampling point distribution map can not only reduce monitoring costs, but also improve the efficiency of the sampling point layout, ensuring that key areas of the monitoring area are adequately covered, while avoiding the waste of redundant sampling points.

[0013] 2. The present invention determines the upper limit of molecular pollutant concentration within the monitoring area based on multiple pollution source singular points and determines the maximum number of monitoring sampling points in combination with the operator activity intensity. This not only accurately reflects the relationship between pollutant concentration and operator distribution, but also dynamically adjusts the number of sampling points according to the pollution load in different areas, ensuring that important areas receive adequate monitoring coverage, thereby avoiding excessive monitoring of ineffective areas and ensuring that key pollution sources are not missed, thereby optimizing monitoring efficiency. The actual input cost of each sampling point is evaluated, and a minimum objective function is established based on the actual input cost. The initial sampling point layout is further optimized to ensure that the final sampling point layout can meet the needs of efficient monitoring while achieving maximum monitoring efficiency within budget constraints. This can effectively reduce operating costs, improve the rationality of resource allocation, and make environmental monitoring more sustainable and efficient. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The accompanying drawings, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0015] Figure 1 4 is a flow chart of an AMC sampling point distribution optimization method based on multi-pollution source singular point fusion according to an embodiment of the present invention. DETAILED DESCRIPTION

[0016] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0017] It should be noted that the following detailed descriptions are exemplary and intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used in the present invention have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0018] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0019] In the absence of conflict, the embodiments of the present invention and the features thereof may be combined with each other.

[0020] See also Figure 1 The present invention provides an AMC sampling point distribution optimization method based on multi-pollution source singular point fusion, including: Step S1, simulating the working behavior of the operator in the monitoring area, and extracting the respiratory volume and skin secretion volume generated by the operator within the target time according to the working behavior simulation results as the molecular pollution load.

[0021] In one embodiment, simulating the work behavior of an operator in a monitoring area, and extracting the respiratory volume and skin secretion volume generated by the operator within a target duration as a molecular pollution load based on the work behavior simulation results includes: generating the operator's work tasks based on the equipment work content in the monitoring area, decomposing the work tasks into behavior action nodes, and setting action operation parameters for the behavior action nodes to obtain execution condition nodes; creating a blank behavior tree, adding the execution condition node to the blank behavior tree to generate a behavior rule engine, associating the behavior rule engine with the physical operator, and running the blank behavior tree to obtain the operator's work behavior; obtaining the operator's respiratory rate and sweating degree under the work behavior conditions, constructing a behavior-physiology coupling model, and predicting the respiratory volume and skin secretion volume generated by the operator within the target duration to obtain the molecular pollution load.

[0022] In one embodiment, a blank behavior tree is created, an execution condition node is added to the blank behavior tree to generate a behavior rule engine, the behavior rule engine is associated with a physical operator, and running the blank behavior tree to obtain the operator's work behavior includes: launching a behavior tree creation wizard based on an autonomous behavior rule editor to generate a blank behavior tree framework including a plurality of root nodes and child node containers, and using a rule generator to define behavior logic for the root nodes and child nodes; obtaining a blank behavior tree based on the behavior logic definition result, and adding an execution condition node to the blank behavior tree node to extract operation instruction information, and generating a behavior rule engine based on the extraction result; associating the behavior rule engine with a physical operator simulation model, and sending the operation instruction information to the physical operator simulation model to determine the physical operator's behavior path within a virtual monitoring area; running the blank behavior tree structure according to the physical operator's behavior path in the simulation environment, and obtaining the operator's work behavior within the virtual monitoring area based on the running result.

[0023] In one embodiment, the respiratory rate and sweating degree of the operator under working behavior conditions are obtained, a behavior-physiological coupling model is constructed, and the respiratory volume and skin secretion volume generated by the operator within the target duration are predicted to obtain the molecular pollution load, including: extracting the respiratory rate and sweating degree of the operator based on the working behavior in the virtual monitoring area, and merging the respiratory rate and sweating degree to generate a bivariate observation sequence; iteratively optimizing the clustering center using Euclidean distance as the similarity measure, performing cluster analysis on the bivariate observation sequence, and dividing the physiological behavior pattern categories of the bivariate observation sequence according to the cluster analysis results; establishing a behavior coupling model based on the physiological behavior pattern category using Bayesian prior distribution technology, and using Markov chain Monte Carlo technology to sample the posterior distribution of the behavior coupling model respectively; estimating time dynamic parameters based on the sampling results, and predicting the respiratory rate and sweating degree of the operator within the target duration, converting the respiratory rate and sweating degree into the total volume of respiratory gas and skin sweat secretion through integral change, generating a physiological output sequence that changes with time, and obtaining the molecular pollution load.

[0024] It should be explained that in the process of obtaining molecular contamination loads, a fine-grained task behavior model is constructed by simulating the work behavior of operators in the monitoring area. Specific operation tasks are generated based on the work content of the actual equipment in the area, and the operation tasks are subdivided into standardized behavior action nodes, such as handling chemicals, operating valves, and testing instruments. Action operation parameters are set for each node, such as action duration, tool use, contact area, task frequency, etc., to form execution condition nodes. At the same time, the behavior rule editor is used to start the behavior tree creation wizard, and a blank behavior tree framework containing a root node and multi-layer child node containers is constructed. The behavior logic is written for each node through the rule maker to obtain an executable behavior model. After binding the behavior model to the physical operator simulation model, the operation instructions from the behavior rule engine are received and executed, thereby reproducing the operator's behavior path in the simulation environment. The path information after the simulation is used to obtain the operator's behavior state under different spatial positions and operating conditions, and further extract key physiological parameters, including respiratory rate (such as 15 breaths per minute) and sweating level (such as sweat secretion reaching 150 ml per hour under high-intensity work), to generate a two-variable observation sequence.

[0025] To accurately simulate behavioral-physiological coupling, Euclidean distance was used as the clustering basis, and K-means++ was employed for initial center point optimization. Cluster analysis was then performed to categorize operators into different physiological-behavioral patterns, such as high metabolic activity, moderate interaction, and low exposure. A Bayesian prior distribution was established based on the cluster labels, and the Markov Chain Monte Carlo method was used to iteratively sample the posterior distribution to estimate temporal dynamic parameters. Integrating an integral transformation model, the temporal dynamics of respiratory rate (e.g., a gradual increase from 12 to 18 breaths per minute) and sweating intensity (e.g., a significant increase starting at the 30th minute of work) were converted into total respiratory gas volume (e.g., 2.4 cubic meters) and sweat secretion volume (e.g., 1.2 liters). This sequence of physiological outputs covering the entire target duration was constructed, representing the operator's molecular pollution burden.

[0026] At the same time, it is assumed that a hazardous chemical workshop is 15m long and 10m wide, with a total of 6 operating equipment, each equipment corresponds to an operator, and 6 tasks are generated according to the work plan (such as loading, stirring, sampling, etc.), which are decomposed into 18 behavioral action nodes. The average duration of each node is 8 minutes. The rule editor defines that the sampling node requires manual opening of the valve and is more than 0.5m close to the equipment, which is a high-contact and high-metabolism task. After the behavior tree is created, it is bound to 6 entity operator simulation models, and the behavior path simulation is run. The simulation data extraction shows that operator A is in a high-intensity state from the 10th to the 30th minute, with a stable respiratory rate of 18 times per minute and a sweat rate of about 160ml / h. Its bivariate sequence is synthesized and divided into three metabolic modes: high, medium, and low through K=3 cluster analysis. Operator A is classified into the high-metabolism group. At the same time, the behavioral coupling model sampling estimates that the total respiratory gas volume of operator A in 60 minutes is 2.6m 3 , the total sweat secretion is 1.5L, and its pollution load is output.

[0027] Step S2: extract the molecular pollutant concentration data during the historical operation process in the monitoring area and couple it with the corresponding number of operators, and obtain the causal relationship between the molecular pollution load and the molecular pollutant concentration based on the coupling result.

[0028] In one embodiment, the molecular pollutant concentration data of the historical operation process in the monitoring area and the corresponding number of operators are extracted for coupling, and the causal relationship between the molecular pollution load and the molecular pollutant concentration is obtained based on the coupling result, including: collecting the historical pollutant concentration data and the number of operators in the monitoring area, and dividing the historical molecular pollutant concentration data and the number of operators into several time periods, judging the fluctuation characteristics of the historical molecular pollutant concentration data in each time period, and the change pattern of the operators, and generating a time series data set; based on the time series data set and the molecular pollution load, using a neural network to perform time series modeling processing, capturing the long-term dependence of the historical molecular pollutant concentration changes and the molecular pollution load, and generating a causal network; performing sensitivity analysis on the causal network, evaluating the contribution of the number of operators in the causal network, optimizing the causal network according to the contribution, and capturing the potential causal relationship between the time series data set and the molecular pollution load.

[0029] It should be explained that in the process of capturing the potential causal relationship between the time series dataset and the molecular pollution load, historical molecular pollutant concentration data and the number of operators in the monitoring area are collected, specifically including the changes in pollutant concentrations at various time points in the area (such as sulfur dioxide and nitrogen oxide concentrations per hour) and the number of operators in each time period (such as the number of workers per hour). The historical data are divided into time periods, such as by hour or by shift, to obtain the fluctuation characteristics of pollutant concentrations and the changing pattern of the number of operators in each time period. By analyzing the pollutant concentrations and the number of operators in the time period, a time series dataset is generated. This time series dataset contains the number of operators and the corresponding pollutant concentration data in each time period, reflecting the changing trend of pollutant concentrations and the impact of operators on pollutant concentrations under different operating intensities.

[0030] Based on the changing trends of pollutant concentrations and the impact of operators on pollutant concentrations, neural networks are used for time series modeling to capture the long-term dependence between changes in pollutant concentrations and the number of operators. The model learns the time series characteristics of historical data and captures the dynamic change patterns between pollutant concentrations and the number of operators. During the model training process, the input is the number of operators and pollutant concentrations in the historical time period, and the output is the predicted future pollutant concentration or molecular pollution load value. The modeling results can identify potential causal relationships in the historical data and generate a causal network. The causal network can intuitively display the interaction between the number of operators and pollutant concentrations and reveal the mechanism of operators' role in the process of pollutant concentration changes. After the causal network is generated, a sensitivity analysis is performed to determine the contribution of the number of operators to the change in pollutant concentration by evaluating the impact of different variables (such as the number of operators) on each causal relationship in the causal network. Specifically, by modifying the values ​​of the number of operators or other related variables and observing the changes in pollutant concentrations in the causal network, the optimized causal network can more accurately capture the potential causal relationship between time series data and molecular pollution loads.

[0031] Assume that the sulfur dioxide (SO2) concentration and the number of operators in a chemical plant workshop are monitored as follows. Through historical data collection, the hourly change records of the sulfur dioxide concentration (unit: ppb) and the hourly change records of the number of operators (unit: person) in the workshop over a period of time (for example, 3 months) are as follows: From 8:00 to 9:00 on January 1, the sulfur dioxide concentration was 120 ppb, and the number of operators was 5; From 9:00 to 10:00 on January 1, the sulfur dioxide concentration was 130 ppb, and the number of operators was 8; From 10:00 to 11:00 on January 1, the sulfur dioxide concentration was 125 ppb, and the number of operators was 6; After the data is divided into time periods, a time series dataset is generated. Each time period includes the corresponding sulfur dioxide concentration and the number of operators. For example, changes in the number of operators within a certain time period may cause fluctuations in pollutant concentrations. By analyzing the time series data, a dataset is generated for neural network learning. The neural network model (such as LSTM) uses the learned temporal features to capture the long-term dependency between the number of operators and pollutant concentrations. The model predicts that when the number of operators increases from 5 to 8, the sulfur dioxide concentration increases by an average of 10 ppb. Based on the training results, a causal network is generated, in which the relationship between the number of operators and sulfur dioxide concentration is quantified to assess the impact of the number of operators on pollutant concentrations. By modifying the number of operators (for example, increasing the number of operators from 5 to 10), the magnitude of the change in sulfur dioxide concentration increases significantly, indicating that the number of operators plays a significant role in sulfur dioxide concentration fluctuations. Based on the analysis results, the causal network is optimized by removing unimportant variables (such as temperature). Ultimately, a precise causal network is generated, resulting in more accurate predictions of sulfur dioxide concentrations.

[0032] Furthermore, through time series modeling and sensitivity analysis of the neural network model, the causal impact of the number of operators on pollutant concentrations can be analyzed. The causal network can help decision makers more accurately predict the changes in pollutant concentrations under different operating intensities and personnel distributions, thereby providing strong data support for optimizing the layout of monitoring sampling points.

[0033] Step S3, obtaining the number and distribution positions of operators in the monitoring area during the target monitoring period, and combining the number of operators with the causal relationship to predict the concentration of molecular pollutants in the monitoring area, and determining the singular points of multiple pollution sources.

[0034] It should be explained that the number and distribution of operators in the monitoring area during the target monitoring period can be obtained. For example, the distribution information of operators in a monitoring area from 8:00 to 10:00 in the morning can be obtained through the factory's scheduling system. Specifically, the distribution of personnel in each position and work area is assumed to be 10 operators in the monitoring area during a certain period of time, distributed in different work areas, of which 3 are in the area close to the pollution source, 5 are in the middle area, and 2 are in the area far away from the pollution source. Through the distribution information, the relationship between the operators and the pollution source can be analyzed, and further input data can be provided for the prediction model of pollutant concentration. The number of operators is combined with the causal relationship to predict the molecular pollution in the monitoring area. The previously established causal relationship model is used in the prediction process. This model has captured the causal relationship between the number and distribution of operators and pollutant concentrations by analyzing the relationship between historical data and operating behavior and pollutant concentrations. Specifically, the causal relationship model predicts the pollutant concentration of a region based on the distribution of operators in different areas (such as whether the operators are concentrated near the pollution source). For example, if more operators are concentrated in the high pollution source area, the pollutant concentration in this area may be relatively high, while the pollutant concentration in other areas is relatively low. Based on this causal relationship model, the pollutant concentration in different areas during the target period is calculated in real time, thereby determining the distribution of pollutant concentrations in the monitoring area.

[0035] After determining the pollutant concentration, multiple pollution source singular points are identified through analysis of pollutant concentrations in different areas. Multiple pollution source singular points refer to areas where pollutant concentrations fluctuate abnormally or concentrate under certain special environmental conditions. These areas may become key points for pollution control. For example, some areas may exceed the pollutant concentration standard due to equipment failure, excessive operator density, or poor local ventilation, becoming potential pollution source singular points. Through comprehensive analysis of pollutant concentrations and combined with the distribution of operators, these singular points are identified to provide a reference for subsequent sampling point optimization.

[0036] Assume that in a production workshop of a chemical plant, the target monitoring period is one day's working hours (8:00-18:00). During this period, there are 12 operators in the monitoring area. Through the scheduling task, it is known that the operators are distributed in different areas, 3 in the high pollution source area, 5 in the medium pollution source area, and 4 in the low pollution source area. The causal model shows that for every additional operator near the pollution source, the pollutant concentration will increase by 15ppb (unit: pollutant concentration). Based on this causal relationship, the pollutant concentration in each area is first calculated. Assuming that from 8:00 to 10:0 During the period of 0, there were three operators in the high pollution source area, so the pollutant concentration in this area was 45 ppb (3 people × 15 ppb). In the medium pollution source area, the total pollutant concentration of the five operators was 75 ppb (5 people × 15 ppb), and in the low pollution source area, the pollutant concentration of the four operators was 60 ppb (4 people × 15 ppb). After calculating the pollutant concentration distribution in each area, the pollutant concentration in the medium pollution source area reached 75 ppb, exceeding the standard limit (such as 50 ppb), so it was identified as a pollution source singularity point.

[0037] Through real-time prediction and analysis, not only can potential high-pollution risk areas be discovered in advance, but it can also provide decision makers with a basis for optimizing monitoring sampling points. For example, the area with the highest pollutant concentration in the monitoring area is the heavy pollution source area, so the monitoring sampling points can be focused on this area and the sampling frequency can be increased to timely monitor and control the pollutant concentration in this area to ensure that it does not exceed the standard.

[0038] Step S4: generating an initial distribution map of monitoring sampling points based on the singular points of multiple pollution sources and the distribution positions of operators, and optimizing the initial distribution map to determine the final monitoring sampling points according to the application cost under the initial distribution map conditions.

[0039] In one embodiment, an initial distribution map of monitoring sampling points is generated based on multiple pollution source singular points and operator distribution positions, and the initial distribution map is optimized to determine the final monitoring sampling points based on the application cost under the initial distribution map conditions, including: determining an upper limit value of the molecular pollutant concentration in the monitoring area based on the multiple pollution source singular points, and determining the maximum number of monitoring sampling points based on the upper limit value and the activity density of the operator distribution positions; constructing a bounding box using axis alignment and oriented bounding box technology based on the maximum number of monitoring sampling points and the layout status of the monitoring area, and generating an initial monitoring sampling point distribution map based on the bounding box; obtaining the purchase price of the monitoring sampler, and calculating the application cost generated by the initial monitoring sampling point distribution map based on the purchase price, establishing a minimum objective function based on the application cost, and optimizing the initial distribution map to determine the final monitoring sampling points.

[0040] In one embodiment, according to the maximum number of monitoring sampling points and the layout status of the monitoring area, an axis-aligned and oriented bounding box technology is used to construct a bounding box, and an initial monitoring sampling point distribution map is generated based on the bounding box, including: obtaining the maximum number of monitoring sampling points and the layout status of the monitoring area, and using the axis-aligned bounding box and oriented bounding box bounding box construction technology to determine the optimal coordinate axis direction in combination with the layout status of the monitoring area; constructing a bounding box that completely covers the monitoring area based on the optimal coordinate axis direction, and projecting the monitoring sampling points to the nearest plane of the bounding box according to the layout status of the monitoring area, and outputting the monitoring sampling point position; judging the distance between the monitoring sampling point and the entity in the monitoring area according to the monitoring sampling point position, and eliminating the monitoring sampling points whose distance is greater than a distance threshold, and re-projecting the eliminated monitoring sampling points; until the difference between the distance between the monitoring sampling point position and the entity in the monitoring area is less than or equal to the distance position, an initial monitoring sampling point distribution map is obtained.

[0041] In one embodiment, an initial distribution map of monitoring sampling points is generated based on multiple pollution source singular points and the distribution positions of operators, and the final monitoring sampling points are determined based on the application cost optimization of the initial distribution map. The purpose is to reduce unnecessary monitoring points while ensuring that key pollution sources and operating behaviors in the monitoring area are fully covered, thereby improving the efficiency and cost-effectiveness of monitoring. The upper limit of pollutant concentration in the monitoring area is determined based on the multiple pollution source singular points, and the maximum number of monitoring sampling points is determined in combination with the distribution density and activity intensity of the operators. Through causal analysis and identification of pollution source singular points, the upper limit of pollutant concentration can be determined (such as the pollutant concentration near the pollution source reaches a maximum value, such as 80 ppb). Based on this concentration upper limit and the distribution of operators in different areas (for example, a high concentration of personnel in a certain area leads to a high pollutant concentration in that area), the required number of monitoring sampling points can be calculated.

[0042] Based on the maximum number of monitoring sampling points and the layout status of the monitoring area, the axis-aligned and oriented bounding box techniques (AABB and OBB techniques) are used to construct bounding boxes, and the initial monitoring sampling point distribution map is generated for the monitoring area in a rational manner. The axis-aligned bounding box and the oriented bounding box can efficiently perform spatial layout. When the geometric shape of the monitoring area is complex, the oriented bounding box can more accurately enclose the irregular area, thereby reducing redundant sampling points and ensuring sufficient coverage of key areas. Assuming that a certain monitoring area has an irregular polygonal shape, the use of the AABB method may lead to unnecessary monitoring points, while the OBB technology can obtain a more reasonable distribution of monitoring points and reduce redundant monitoring.

[0043] Assume that a chemical plant has a rectangular workshop that is 100 meters long and 50 meters wide. The target monitoring period is 24 hours. There are three main pollution sources distributed in the workshop. During the monitoring period, there are 30 operators distributed in different areas. Assume that pollution source 1 and pollution source 2 are relatively close, and pollution source 3 is located at the other end of the workshop. According to historical data, the upper limit of pollutant concentration in the chemical reactor area is 80ppb, the concentration in the evaporator area is 75ppb, and the exhaust gas emission pipe area is 50ppb.

[0044] According to the distribution density of operators, assuming that there are 12 operators in the chemical reactor area, 8 in the evaporator area, and 10 in the exhaust gas emission pipe area, it is calculated based on this distribution that more sampling points are needed to cover high-pollution source areas, such as chemical reactors and evaporators. Assuming that the maximum number of monitoring sampling points is 30, the sampling points are allocated to different pollution source areas, and the layout is optimized using axis alignment and oriented bounding box technology to construct a bounding box. This can rationalize the irregular layout in the workshop, optimize the sampling points to high-pollution areas, and effectively reduce redundant points.

[0045] In one embodiment, the purchase price of the monitoring sampler is obtained, and the application cost generated by the initial monitoring sampling point distribution map is calculated based on the purchase price, a minimum objective function is established based on the application cost, and the initial distribution map is optimized to determine the final monitoring sampling points, including: obtaining the purchase price of the monitoring sampler, and calculating the purchase cost of the monitoring sampler under the condition of the maximum number of monitoring sampling points, and at the same time predicting the deployment, operation and maintenance, and additional electricity costs of the initial monitoring sampling point distribution map during the monitoring process; constructing a total application cost function based on the additional cost and the purchase cost, and with the goal of minimizing the total application cost, combining monitoring coverage and deployment efficiency to construct a multi-objective function for optimizing the distribution of monitoring sampling points; encoding the monitoring sampling point distribution scheme according to the multi-objective function, and defining the monitoring sampling point distribution scheme as nodes, and using a non-dominated sorting reverse learning differential evolution algorithm to determine the final monitoring sampling points.

[0046] In one embodiment, a monitoring sampling point distribution scheme is encoded according to a multi-objective function, and the monitoring sampling point distribution scheme is defined as a node. The final monitoring sampling point is determined using a non-dominated sorting reverse learning differential evolution algorithm, including: encoding the monitoring sampling point distribution scheme as a node representation based on the multi-objective function, using the node to represent the configuration of the monitoring sampling point, and using a differential evolution algorithm to perform differential mutation processing on the node to generate a mutation vector; performing a crossover operation on the mutation vector according to the crossover probability to generate experimental individuals, and calculating the fitness value of each experimental individual, and calculating the performance of the corresponding node under the multi-objective function based on the fitness value; selecting the optimal solution from the experimental individuals based on the performance calculation result and the greedy selection strategy; outputting a Pareto optimal solution set according to the optimal solution, and selecting a monitoring sampling point distribution scheme with the lowest total cost, the highest monitoring coverage and deployment efficiency from the Pareto optimal solution set as the final monitoring sampling point.

[0047] In one embodiment, selecting the optimal solution from experimental individuals based on performance calculation results and a greedy selection strategy includes: ranking the experimental individuals on multiple objectives based on the performance results of the nodes under the multi-objective function, comparing the comprehensive performance of the experimental individuals under the multi-objective function to confirm the existence of the dominance relationship; hierarchically sorting the experimental individuals according to the existence of the dominance relationship, dividing the Pareto levels based on the hierarchical sorting results, and using a greedy selection strategy to gradually select the experimental individuals with the best performance from the Pareto levels to output the optimal solution.

[0048] It needs to be explained that the purchase price of the monitoring sampler is obtained. Assuming that the purchase price of each sampler is 2,000 yuan, the purchase cost of the sampler is calculated based on the area of ​​the monitoring area and the required number of sampling points. Assuming that the maximum number of sampling points in the monitoring area is 30, according to the purchase price of each sampler, the total purchase cost of the sampler is preliminarily calculated as: purchase cost = 30*2000 = 60,000 yuan.

[0049] In addition to the purchase cost, it is also necessary to predict the deployment, operation and maintenance, and additional electricity costs of the monitoring sampling points during the monitoring process. The costs include the daily maintenance costs of the sampling instruments, the electricity consumption costs (such as 10 yuan of electricity per instrument per day), and the operation and maintenance costs such as deployment, debugging and data analysis. Assuming that the annual operation and maintenance cost of each sampling point in the operation phase is 1,500 yuan, and the additional electricity cost is 10 yuan per month for each sampler, the annual operation and maintenance and additional electricity costs of the entire monitoring system are: Additional cost = 30*(1,500+10*12)=30*1,740=52,200 yuan.

[0050] Based on the purchase cost and additional cost, the total application cost function is constructed. The total application cost function can be expressed as: total application cost = purchase cost + additional cost = 60,000 + 52,200 = 112,200 yuan / year.

[0051] Based on the total application cost required for the distribution of monitoring sampling points, these costs need to be optimized. In this optimization process, the goal is not only to minimize the total application cost, but also to consider the monitoring coverage and deployment efficiency. A multi-objective function is constructed to consider these goals at the same time. This function combines the monitoring coverage (the range of pollution sources covered by each sampling point) and deployment efficiency (the density and cost of the monitoring points) to optimize the distribution of sampling points. The monitoring sampling point distribution scheme is encoded and represented as a node, and the differential evolution algorithm is applied for optimization. The differential evolution algorithm is a global optimization algorithm based on population, which is suitable for handling multi-objective optimization problems. In the algorithm, each individual represents a distribution scheme of a monitoring sampling point, and all individuals form a population. The fitness value of each individual is related to its performance under the multi-objective function.

[0052] Specifically, the differential evolution algorithm first performs mutation processing on each node to generate a mutation vector. The mutation vector is generated by selecting three random individuals from the current population, calculating the individual's difference vector and adding it to the current individual. Assuming that the current individual represents the layout of a certain sampling point, three individuals are selected to calculate their differences and generate a mutation vector, thereby forming a new experimental individual. The mutation vector is cross-operated according to the crossover probability to generate a new experimental individual. The experimental individual represents a new monitoring sampling point distribution scheme.

[0053] Calculate the fitness value of each experimental individual and evaluate its performance under the multi-objective function. The calculation of the fitness value will comprehensively consider the total application cost, monitoring coverage and deployment efficiency. If a solution has a low total application cost, but its ability to cover high pollution sources is weak, then its fitness value may be low. On the contrary, if a solution can effectively cover all pollution sources and has high deployment efficiency, then its fitness value is high. Based on the fitness value, the optimal solution is selected from the experimental individuals through the greedy selection strategy. The greedy selection strategy first sorts the experimental individuals, compares their advantages and disadvantages under multiple objectives, and selects them from the experimental individuals according to their priority. The best performing individual is selected from the individuals. The quality of the experimental individual is judged by its comprehensive performance under multiple objectives. The performance of each individual under all objective functions is compared, and individuals with strong dominance relationships are selected. These individuals are put into the next round of optimization and the Pareto optimal solution set is output. The Pareto optimal solution set means that in multi-objective optimization, the solution of other objectives cannot be improved by adjusting the value of any one objective. Finally, the monitoring sampling point distribution scheme with the lowest total cost, the highest monitoring coverage and deployment efficiency is selected as the final monitoring scheme. In this way, the sampling point layout scheme with the best comprehensive performance can be obtained within the framework of multi-objective optimization.

[0054] Those skilled in the art will appreciate that the units and algorithm steps of the various examples described in conjunction with this embodiment can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0055] Although the above describes the specific embodiments of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without any creative work are still within the scope of protection of the present invention.

Claims

1. AMC sampling point distribution optimization method based on multi-pollution source singular point fusion, characterized by: include: Simulate the operator's working behavior in the monitoring area, and extract the operator's respiratory volume and skin secretion volume generated within the target time period based on the work behavior simulation results as the molecular pollution load; Extract the molecular pollutant concentration data of the historical operation process in the monitoring area and the corresponding number of operators for coupling, and obtain the causal relationship between molecular pollution load and molecular pollutant concentration based on the coupling results; Obtain the number and distribution of operators in the monitoring area during the target monitoring period, and combine the number of operators with causal relationships to predict the concentration of molecular pollutants in the monitoring area and determine the singular points of multiple pollution sources; An initial distribution map of monitoring sampling points is generated based on the singular points of multiple pollution sources and the distribution positions of operators. The initial distribution map is optimized to determine the final monitoring sampling points according to the application cost under the conditions of the initial distribution map.

2. The AMC sampling point distribution optimization method based on multi-pollution source singular point fusion according to claim 1 is characterized in that: The simulated operator's working behavior in the monitoring area and the extraction of the operator's respiratory volume and skin secretion volume generated within the target time period based on the work behavior simulation results are used as molecular pollution loads, including: Generate operator work tasks based on the equipment work content in the monitoring area, decompose the work tasks into behavior action nodes, and set action operation parameters for the behavior action nodes to obtain execution condition nodes; Create a blank behavior tree, add an execution condition node to it to generate a behavior rule engine, associate the behavior rule engine with a physical operator, and run the blank behavior tree to obtain the operator's work behavior. The respiratory rate and sweating degree of the operator under working behavior conditions are obtained, and a behavior-physiology coupling model is constructed to predict the respiratory volume and skin secretion volume generated by the operator within the target time, and obtain the molecular pollution load.

3. The AMC sampling point distribution optimization method based on multi-pollution source singular point fusion according to claim 2 is characterized in that: Creating a blank behavior tree, adding execution condition nodes to the blank behavior tree to generate a behavior rule engine, and associating the behavior rule engine with a physical operator. Running the blank behavior tree to obtain the operator's work behavior includes: Launch the Behavior Tree Creation Wizard based on the Autonomous Behavior Rule Editor to generate a blank Behavior Tree framework containing several root nodes and child node containers. Use the Ruler to define the behavior logic for the root nodes and child nodes. Obtain a blank behavior tree based on the behavior logic definition results, add execution condition nodes to the blank behavior tree nodes to extract operation instruction information, and generate a behavior rule engine based on the extracted results; Associating the behavior rule engine with the physical operator simulation model and sending operation instruction information to the physical operator simulation model to determine the behavior path of the physical operator in the virtual monitoring area; In the simulation environment, a blank behavior tree structure is run according to the behavior path of the physical operator, and the operator's work behavior in the virtual monitoring area is obtained based on the running results.

4. The AMC sampling point distribution optimization method based on multi-pollution source singular point fusion according to claim 3 is characterized in that: The method of obtaining the operator's respiratory rate and sweating degree under working behavior conditions, building a behavior-physiology coupling model, and predicting the operator's respiratory volume and skin secretion volume within the target time period to obtain the molecular pollution load includes: The respiratory rate and sweating degree of the operator are extracted based on the working behavior in the virtual monitoring area, and the respiratory rate and sweating degree are combined to generate a bivariate observation sequence; The cluster centers are iteratively optimized using Euclidean distance as the similarity metric, and cluster analysis is performed on the bivariate observation sequence. The physiological behavior pattern categories of the bivariate observation sequence are then divided according to the cluster analysis results. Based on the physiological behavior pattern categories, the Bayesian prior distribution technology was used to establish a behavioral coupling model, and the Markov chain Monte Carlo technology was used to sample the posterior distribution of the behavioral coupling model. The temporal dynamic parameters are estimated based on the sampling results, and the operator's breathing rate and sweating degree within the target duration are predicted. The breathing rate and sweating degree are converted into the total volume of respiratory gas and skin sweat secretion through integral changes, generating a physiological output sequence that changes with time and obtaining the molecular pollution load.

5. The AMC sampling point distribution optimization method based on multi-pollution source singular point fusion according to claim 1 is characterized in that: The extracting of molecular pollutant concentration data during historical operation in the monitoring area and the corresponding number of operators are coupled, and obtaining the causal relationship between the molecular pollution load and the molecular pollutant concentration based on the coupling result includes: Collect historical pollutant concentration data and the number of operators in the monitoring area, divide the historical molecular pollutant concentration data and the number of operators into several time periods, determine the fluctuation characteristics of the historical molecular pollutant concentration data and the change pattern of operators in each time period, and generate a time series data set; Based on time series data sets and molecular pollution loads, neural networks are used for time series modeling to capture the long-term dependency between historical molecular pollutant concentration changes and molecular pollution loads, generating a causal network. A sensitivity analysis was performed on the causal network to evaluate the contribution of the number of operators in the causal network. The causal network was optimized based on the contribution to capture the potential causal relationship between the time series dataset and the molecular pollution load.

6. The AMC sampling point distribution optimization method based on multi-pollution source singular point fusion according to claim 1 is characterized in that: The generating of an initial distribution map of monitoring sampling points based on the singular points of multiple pollution sources and the distribution positions of operators, and optimizing the initial distribution map to determine the final monitoring sampling points according to the application cost under the initial distribution map conditions includes: Determine the upper limit of molecular pollutant concentration in the monitoring area based on the singular points of multiple pollution sources, and determine the maximum number of monitoring sampling points based on the upper limit and the activity density of the distribution of operators; According to the maximum number of monitoring sampling points and the layout of the monitoring area, the axis alignment and oriented bounding box technology are used to construct a bounding box, and the initial monitoring sampling point distribution map is generated based on the bounding box; Obtain the purchase price of the monitoring sampler, and calculate the application cost generated by the initial monitoring sampling point distribution map based on the purchase price. Establish a minimum objective function based on the application cost, optimize the initial distribution map and determine the final monitoring sampling points.

7. The AMC sampling point distribution optimization method based on multi-pollution source singular point fusion according to claim 6 is characterized in that: The method of constructing a bounding box based on the maximum number of monitoring sampling points and the layout state of the monitoring area using axis alignment and oriented bounding box technology, and generating an initial monitoring sampling point distribution map based on the bounding box includes: Obtain the maximum number of monitoring sampling points and the layout status of the monitoring area, and use the axis-aligned bounding box and oriented bounding box bounding box construction technology to determine the optimal coordinate axis direction in combination with the layout status of the monitoring area; Construct a bounding box that completely covers the monitoring area based on the optimal coordinate axis direction, and project the monitoring sampling points to the nearest plane of the bounding box according to the layout status of the monitoring area, and output the location of the monitoring sampling points; The distance between the monitoring sampling point and the entity in the monitoring area is determined according to the location of the monitoring sampling point, and the monitoring sampling points with a distance higher than the distance threshold are eliminated, and the eliminated monitoring sampling points are re-projected; Until the difference between the monitoring sampling point location and the physical distance within the monitoring area is less than or equal to the distance location, the initial monitoring sampling point distribution map is obtained.

8. The AMC sampling point distribution optimization method based on multi-pollution source singular point fusion according to claim 7 is characterized in that: The method of obtaining the purchase price of the monitoring sampler, calculating the application cost of the initial monitoring sampling point distribution map based on the purchase price, establishing a minimum objective function based on the application cost, and optimizing the initial distribution map to determine the final monitoring sampling points includes: Obtain the purchase price of the monitoring sampler and calculate the purchase cost of the monitoring sampler under the condition of the maximum number of monitoring sampling points. At the same time, predict the deployment, operation and maintenance, and additional electricity costs of the initial monitoring sampling point distribution map during the monitoring process. A total application cost function is constructed based on the additional cost and the acquisition cost. With the goal of minimizing the total application cost, a multi-objective function for optimizing the distribution of monitoring sampling points is constructed by combining monitoring coverage and deployment efficiency. The monitoring sampling point distribution scheme is encoded according to the multi-objective function and defined as nodes. The final monitoring sampling points are determined using the reverse learning differential evolution algorithm of non-dominated sorting.

9. The AMC sampling point distribution optimization method based on multi-pollution source singular point fusion according to claim 8 is characterized in that: The method of encoding the monitoring sampling point distribution scheme according to the multi-objective function, defining the monitoring sampling point distribution scheme as nodes, and determining the final monitoring sampling points using the reverse learning differential evolution algorithm of non-dominated sorting includes: Based on the multi-objective function, the monitoring sampling point distribution scheme is encoded into node representation, the node representation is used to monitor the configuration of the sampling points, and the differential evolution algorithm is used to perform differential mutation processing on the nodes to generate mutation vectors; Perform crossover operations on the mutation vectors according to the crossover probability to generate test individuals, calculate the fitness value of each test individual, and calculate the performance of the corresponding node under the multi-objective function based on the fitness value; Select the optimal solution from the experimental individuals based on the performance calculation results and the greedy selection strategy; According to the optimal solution, the Pareto optimal solution set is output, and the monitoring sampling point distribution scheme with the lowest total cost, the highest monitoring coverage and deployment efficiency is selected from the Pareto optimal solution set as the final monitoring sampling point.

10. The AMC sampling point distribution optimization method based on multi-pollution source singular point fusion according to claim 9 is characterized in that: The method of selecting the optimal solution from the experimental individuals based on the performance calculation results and the greedy selection strategy includes: Based on the performance results of the nodes under the multi-objective function, the test individuals are ranked according to their performance on multiple objectives, and the comprehensive performance of the test individuals under multiple objectives is compared to confirm the existence of the dominance relationship; The experimental individuals are hierarchically sorted according to the existence of dominance relationships, and the Pareto levels are divided based on the hierarchical sorting results. The greedy selection strategy is used to gradually select the experimental individuals with the best performance from the Pareto levels and output the optimal solution.

Images