Accurate traceability analysis method and device

Abstract

The present invention provides a precise source tracing and analysis method and device, the method comprising: obtaining meteorological data, atmospheric pollutant monitoring data and comprehensive source spectrum of pollution sources corresponding to each production device in the area corresponding to the production device; analyzing the atmospheric pollutant monitoring data to obtain the corresponding atmospheric pollutant source analysis results; determining the target production device according to the meteorological data corresponding to the appearance time of the abnormal value in the atmospheric pollutant source analysis results, and obtaining the pollution contribution rate corresponding to the target production device based on the atmospheric pollutant monitoring data corresponding to the appearance time of the abnormal value and the comprehensive source spectrum of the pollution source corresponding to the target production device; analyzing the change of the pollution contribution rate, and determining the production device to be checked from the target production device. The method of the present invention can shorten the investigation time, has high traceability accuracy, and can quickly and accurately locate the pollution source.

Description

Technical Field

[0001] The present invention relates to the technical field of atmospheric pollutant monitoring and source tracing, and in particular to an accurate source tracing and analysis method and an accurate source tracing and analysis device. Background Art

[0002] The petrochemical industry is a key industry for VOCs emissions, accounting for about 18-40% of the country's industrial VOCs emissions. Comprehensive remediation of air pollutants and odor control are imminent.

[0003] Currently, LDAR technology (Leak Detection And Repair) has been proposed to target VOCs emissions from petrochemical enterprises in order to control VOCs leaks from process equipment and pipelines in the petrochemical industry at the source. However, LDAR technology requires a lot of work in actual work, such as taking photos and manually entering data, resulting in low LDAR efficiency, high error rate, and inability to monitor in real time. It also takes a long time to use LDAR to identify possible leaks in a production unit, resulting in uneven results in the application of LDAR technology.

[0004] With the emergence of urban refineries, there are sensitive points such as external enterprises, residents, and schools around the plant area. The internal equipment areas of the refinery are closely arranged, and there are many types of pollutants with complex components. It is difficult to accurately grasp the pollution status of the refinery itself, the source of pollution, the impact of the pollution source on the surrounding environment, and how to monitor, control, and trace the source. The traceability methods that are currently being studied are basically aimed at chemical parks, and can only be traced to enterprises in the park. The components of these enterprises are quite different, and it is easier to trace the source. To trace the source to a specific device or a specific leakage point in the enterprise, due to the small difference in components between the devices, there is a lack of grid monitoring technology for VOCs leakage traceability in petrochemical enterprises, which makes it difficult to further judge and identify. Therefore, it is very necessary to study a refined traceability system for atmospheric pollutants. On the one hand, it can monitor the pollution contribution rate of possible pollution sources in key monitoring areas in real time online, and on the other hand, it can perceive possible abnormalities or emergencies as early as possible, and then realize early warning to avoid greater risks caused by abnormal accidents. Summary of the invention

[0005] The purpose of the embodiments of the present invention is to provide an accurate tracing and analysis method to at least solve the above-mentioned problems of difficulty in accurately grasping the pollution status of the refinery itself, inability to accurately determine the source of pollution, and how to monitor, control and accurately and quickly trace the source.

[0006] In order to achieve the above-mentioned object, the first aspect of the present invention provides a precise source tracing and analysis method, comprising:

[0007] Obtain meteorological data, atmospheric pollutant monitoring data, and comprehensive source spectrum of pollution sources corresponding to each production device in the area corresponding to the production device;

[0008] Analyze the air pollutant monitoring data to obtain corresponding air pollutant source analysis results;

[0009] If there are abnormal values ​​in the atmospheric pollutant source analysis results, the target production device is determined according to the meteorological data corresponding to the time when the abnormal value occurs, and the pollution contribution rate corresponding to the target production device is obtained based on the atmospheric pollutant monitoring data corresponding to the time when the abnormal value occurs and the comprehensive source spectrum of the pollution source corresponding to the target production device;

[0010] Analyze the change of the pollution contribution rate and determine the production equipment that needs to be checked from the target production equipment.

[0011] Optionally, the analyzing the air pollutant monitoring data to obtain corresponding air pollutant source analysis results includes:

[0012] The principal component analysis model, the positive definite matrix factorization model, the ME2 model or the UNMIX model is called, and the atmospheric pollutant monitoring data is used as input to obtain corresponding atmospheric pollutant source analysis results.

[0013] Optionally, obtaining the pollution contribution rate corresponding to the target production device based on the atmospheric pollutant monitoring data corresponding to the occurrence time of the abnormal value and the comprehensive source spectrum of the pollution source corresponding to the target production device includes:

[0014] The chemical mass balance model is called, and the atmospheric pollutant monitoring data corresponding to the occurrence time of the abnormal value and the comprehensive source spectrum of the pollution source corresponding to the target production device are used as input to obtain the pollution contribution rate corresponding to the target production device.

[0015] Optionally, obtaining a comprehensive source spectrum of pollution sources corresponding to each production device includes:

[0016] For each production unit:

[0017] Collecting concentration data of each gas sample of the production device;

[0018] Calculation steps: Calculate the average component concentration of the gas sample;

[0019] Elimination step: eliminating the gas sample whose concentration data has the lowest similarity with the average component concentration in the gas sample;

[0020] For the remaining gas sample, the calculation step and the elimination step are repeatedly performed until the similarity between the concentration data of the remaining gas sample and the average component concentration of the remaining gas sample is greater than a preset value, and the average component concentration of the remaining gas sample is used as the characterizing component concentration of the gas sample of the production device;

[0021] A comprehensive source spectrum of the pollution sources corresponding to each production device is prepared based on the concentration of the characteristic components of the gas samples of each production device.

[0022] Optionally, the comprehensive source spectrum of the pollution sources may be updated multiple times.

[0023] Optionally, the method further includes:

[0024] The air pollutant monitoring data is processed as follows:

[0025] Eliminate air pollutant monitoring data that are higher than the preset pollutant photochemical reaction activity standard and lower than the preset air pollutant detection rate standard;

[0026] The atmospheric pollutant monitoring data after data elimination are sorted into pollutant concentration files corresponding to the principal component analysis model, the positive definite matrix factor decomposition model, the ME2 model or the UNMIX model as inputs of the corresponding models.

[0027] Optionally, the atmospheric pollutant source analysis result is the atmospheric pollutant concentration change value and the contribution rate of different pollution factors distributed in time series; the abnormal value is determined by the atmospheric pollutant concentration change value;

[0028] The method further comprises:

[0029] The contribution rates of different pollution factors in a certain time period and the meteorological data corresponding to the time period are used as input, and a conditional probability distribution diagram is obtained using a conditional probability function. The conditional probability distribution diagram is used to show the source direction of different pollution factors.

[0030] Optionally, the method further includes:

[0031] Eliminate abnormal data from the abnormal values, and use the atmospheric pollutant monitoring data corresponding to the occurrence time of the eliminated abnormal values ​​as the input of the chemical mass balance model, wherein the abnormal data includes monitoring data obtained when the atmospheric pollutant monitoring equipment does not work properly and abnormal meteorological data;

[0032] The determining of the target production device according to the meteorological data corresponding to the occurrence time of the abnormal value includes: determining the production device in the upwind position as the target production device according to the dominant wind direction in the meteorological data.

[0033] Optionally, the method further includes:

[0034] A portable TVOC measuring instrument is used to check for pollution leakage points in the production equipment that needs to be checked.

[0035] A second aspect of the present invention provides a precise source tracing and analysis device, the device comprising:

[0036] An acquisition module is used to acquire meteorological data of the area corresponding to the production device, atmospheric pollutant monitoring data, and a comprehensive source spectrum of pollution sources corresponding to each production device;

[0037] A first analysis module is used to analyze the air pollutant monitoring data to obtain corresponding air pollutant source analysis results;

[0038] A second analysis module is used to determine the target production device according to the meteorological data corresponding to the time when the abnormal value appears in the atmospheric pollutant source analysis result, and obtain the pollution contribution rate corresponding to the target production device based on the atmospheric pollutant monitoring data corresponding to the time when the abnormal value appears and the comprehensive source spectrum of the pollution source corresponding to the target production device;

[0039] A determination module is used to analyze the change of the pollution contribution rate and determine the production equipment that needs to be checked from the target production equipment.

[0040] Optionally, the first analysis module is used to: call a principal component analysis model, a positive definite matrix factor decomposition model, an ME2 model or an UNMIX model, and use the atmospheric pollutant monitoring data as input to obtain corresponding atmospheric pollutant source analysis results.

[0041] Optionally, the second analysis module is used to: call the chemical mass balance model, take the atmospheric pollutant monitoring data corresponding to the occurrence time of the abnormal value and the comprehensive source spectrum of the pollution source corresponding to the target production device as input, and obtain the pollution contribution rate corresponding to the target production device.

[0042] Optionally, the acquisition module is used to:

[0043] For each production unit:

[0044] Collecting concentration data of each gas sample of the production device;

[0045] Calculation steps: Calculate the average component concentration of the gas sample;

[0046] Elimination step: eliminating the gas sample whose concentration data has the lowest similarity with the average component concentration in the gas sample;

[0047] For the remaining gas sample, the calculation step and the elimination step are repeatedly performed until the similarity between the concentration data of the remaining gas sample and the average component concentration of the remaining gas sample is greater than a preset value, and the average component concentration of the remaining gas sample is used as the characterizing component concentration of the gas sample of the production device;

[0048] A comprehensive source spectrum of the pollution sources corresponding to each production device is prepared based on the concentration of the characteristic components of the gas samples of each production device.

[0049] Optionally, the comprehensive source spectrum of the pollution sources may be updated multiple times.

[0050] Optionally, the first parsing module is further used to process the air pollutant monitoring data as follows:

[0051] Eliminate air pollutant monitoring data that are higher than the preset pollutant photochemical reaction activity standard and lower than the preset air pollutant detection rate standard;

[0052] The atmospheric pollutant monitoring data after data elimination are sorted into pollutant concentration files corresponding to the principal component analysis model, the positive definite matrix factor decomposition model, the ME2 model or the UNMIX model as inputs of the corresponding models.

[0053] Optionally, the second parsing module is further used for:

[0054] Eliminate abnormal data from the abnormal values, and use the atmospheric pollutant monitoring data corresponding to the occurrence time of the eliminated abnormal values ​​as the input of the chemical mass balance model, wherein the abnormal data includes monitoring data obtained when the atmospheric pollutant monitoring equipment does not work properly and abnormal meteorological data;

[0055] The second analysis module is used to determine the production device in the upwind position as the target production device according to the dominant wind direction in the meteorological data.

[0056] On the other hand, the present invention provides a machine-readable storage medium having instructions stored thereon, the instructions being used to enable a machine to execute the precise source tracing and parsing method described above in the present application.

[0057] Through the above technical scheme, the atmospheric pollutants outside the production equipment of the enterprise are measured and analyzed, and the source direction of the atmospheric pollutants is obtained by combining the conditional probability distribution diagram, and finally the results of atmospheric pollutant source analysis are obtained, and the abnormal values ​​in the atmospheric pollutant source analysis results are analyzed in detail, and the pollution source is further located to the production equipment, and then the portable TVOC measuring instrument is used to further check the production equipment area, and the pollution leakage point is quickly tracked and located. By conducting rapid monitoring in the corresponding device area, it is possible to screen points where abnormal conditions may exist, determine possible abnormal conditions, and monitor the pollution contribution rate of possible pollution sources in key monitoring areas in real time online. On the other hand, it can perceive possible abnormalities or emergencies as early as possible, and can avoid further deterioration of abnormal conditions or emergencies leading to serious consequences. It also provides a basis for targeted, rapid and accurate guidance to solve abnormal problems, and can save manual workload, shorten the investigation time, and has high traceability accuracy.

[0058] Other features and advantages of the embodiments of the present invention will be described in detail in the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the following specific embodiments, they are used to explain the embodiments of the present invention, but do not constitute a limitation on the embodiments of the present invention. In the accompanying drawings:

[0060] Figure 1 It is a flow chart of the precise source tracing and analysis method provided by the present invention;

[0061] Figure 2 It is the technical roadmap of the precise source tracing and analysis method provided by the present invention;

[0062] Figure 3 It is a structural schematic diagram of the precise source tracing and analysis device provided by the present invention;

[0063] Figure 4 It is a schematic diagram of the calculation results of the VOCs component similarity between the comprehensive source spectrum of the wax oil hydrogenation device and each sampling point provided in Example 1 of the present invention;

[0064] Figure 5 It is a schematic diagram of the calculation results of the similarity of VOCs components between the comprehensive source spectrum of the catalytic cracking unit and each sampling point provided in Example 2 of the present invention;

[0065] Figure 6 is a schematic diagram of similarity calculation results of comprehensive source spectra of different devices provided in Example 3 of the present invention;

[0066] Figure 7 Schematic diagram of the location distribution of the oil refining device and the online analysis and monitoring points provided in Example 4 of the present invention;

[0067] Figure 8 is a conditional probability distribution diagram of the main pollution factors and the pollutant compounds in each factor provided in Example 4 of the present invention;

[0068] Fig. 9 Schematic diagram of the pollution contribution rate of each device to the online analysis monitoring point provided by Example 4 of the present invention;

[0069] Fig.10 It is a time series diagram of changes in atmospheric pollutant concentrations provided in Example 5 of the present invention.

[0070] Description of Reference Numerals

[0071] 10 - acquisition module; 20 - first analysis module; 30 - second analysis module; 40 - determination module. DETAILED DESCRIPTION

[0072] The specific implementation of the present invention is described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation described here is only used to illustrate and explain the present invention, and is not used to limit the present invention.

[0073] Figure 1 This is a flow chart of a precise source tracing and analysis method provided by one embodiment of the present invention. Figure 1 As shown, an embodiment of the present invention provides an accurate source tracing and analysis method, the method comprising:

[0074] Step 101, obtaining meteorological data, atmospheric pollutant monitoring data, and a comprehensive source spectrum of pollution sources corresponding to each production device in the area corresponding to the production device;

[0075] Pollutant monitoring focuses on areas with dense pollution emission devices, complex pollution components, and a large number of enterprises outside the factory. In order to understand the pollution emission conditions of internal enterprises and enterprise devices, online analysis monitoring points (i.e., online analysis huts) can be set outside the production device. The online analysis monitoring points are used to obtain atmospheric pollutant monitoring data and meteorological data outside the production device.

[0076] Step 102: Analyze the air pollutant monitoring data to obtain corresponding air pollutant source analysis results;

[0077] Step 103: If there is an abnormal value in the atmospheric pollutant source analysis result, determine the target production device according to the meteorological data corresponding to the time when the abnormal value appears, and obtain the pollution contribution rate corresponding to the target production device based on the atmospheric pollutant monitoring data corresponding to the time when the abnormal value appears and the comprehensive source spectrum of the pollution source corresponding to the target production device;

[0078] Step 104: Analyze the change of the pollution contribution rate and determine the production equipment that needs to be checked from the target production equipment.

[0079] Among them, before obtaining the meteorological data and atmospheric pollutant monitoring data outside the production device, the pollutant emission device area is determined and an online analysis monitoring point is set; and the target production device determined according to the meteorological data corresponding to the time of occurrence of the abnormal value can be one or more, and analyzing the change in the pollution contribution rate is to analyze the change in the pollution contribution rate of each pollution source to the online analysis monitoring point; the production device that needs to be checked from the target production device can be checked with a portable TVOC meter, so as to determine the pollution leakage point, which can reduce the workload and shorten the investigation time, and avoid greater risks due to abnormal accidents.

[0080] Furthermore, the air pollutant monitoring data is analyzed to obtain corresponding air pollutant source analysis results, including:

[0081] The principal component analysis model, the positive definite matrix factorization model, the ME2 model or the UNMIX model is called, and the atmospheric pollutant monitoring data is used as input to obtain corresponding atmospheric pollutant source analysis results.

[0082] In this implementation, a PMF model (Positive Matrix Factorization) is used as an example for key analysis.

[0083] Furthermore, the air pollutant monitoring data corresponding to the occurrence time of the abnormal value and the comprehensive source spectrum of the pollution source corresponding to the target production device are used to obtain the pollution contribution rate corresponding to the target production device, including:

[0084] The CMB model (Chemical Mass Balance) is called, and the atmospheric pollutant monitoring data corresponding to the occurrence time of the abnormal value and the comprehensive source spectrum of the pollution source corresponding to the target production device are used as input to obtain the pollution contribution rate corresponding to the target production device. The CMB model can also be replaced by an improved CMB model.

[0085] Figure 2 This is the technical roadmap of the precise source tracing and analysis method provided by the present invention. Figure 2 As shown, the general technical roadmap of the precise traceability analysis method is presented.

[0086] According to one embodiment, volatile organic compounds can be measured by arranging an online VOCs monitoring system. The concentration of volatile organic compounds can be measured by GC-MS / FID (Gas Chromatography-Mass Spectrometer) or GC-TOFMS (Gas Chromatography-Mass Spectrometer-Time Of Flight Mass Spectrometry). The monitored organic matter includes alkanes, olefins, aromatic hydrocarbons, halogenated hydrocarbons, oxygen-containing and sulfur-containing organic matter, etc. The collection time interval of the pollutant monitoring data and the meteorological data can be set to n hours (n≥1). For example, the collection time can be set to collect an atmospheric sample once every hour, and the online test results can be displayed in real time. The long-term monitoring data is combined with the meteorological data to comprehensively monitor the pollution status of the area, and the source of pollutants can be dynamically traced in real time.

[0087] Furthermore, the step of obtaining a comprehensive source spectrum of pollution sources corresponding to each production device includes:

[0088] For each production unit:

[0089] Collecting concentration data of each gas sample of the production device;

[0090] Calculation steps: Calculate the average component concentration of the gas sample;

[0091] Elimination step: eliminating the gas sample whose concentration data has the lowest similarity with the average component concentration in the gas sample;

[0092] For the remaining gas sample, the calculation step and the elimination step are repeatedly performed until the similarity between the concentration data of the remaining gas sample and the average component concentration of the remaining gas sample is greater than a preset value, and the average component concentration of the remaining gas sample is used as the characterizing component concentration of the gas sample of the production device;

[0093] A comprehensive source spectrum of the pollution sources corresponding to each production device is prepared based on the concentration of the characteristic components of the gas samples of each production device.

[0094] Since the same production device may include different production units or different process sections, and there are differences in the concentration of pollutants generated by different production units or different process sections, in order to ensure the accuracy of the source spectrum information of the device, one or more sampling points can be arranged according to the raw materials, products, and processing technology characteristics of each device. The sampling points can be arranged according to the actual distance based on the divided process sections. In the concentrated area where the light component materials flow, the number of sampling points should be as large as possible to ensure that the samples are distributed in all areas of the device.

[0095] Specifically, in the actual operation process, each production unit or process section is set up with no less than 10 points, and the sample collection is carried out under the condition of low wind speed, and the wind speed is taken as <2m / s, so as to reduce the influence of diffusion transmission between the devices and avoid sampling in rainy and snowy weather; gas sample sampling is carried out using 3.2L Suma tank, with a sampling flow rate of 50ml / min, a sampling height of 1.5m, and a collection interval of 1 hour. At the same time, sampling is carried out under the condition of low wind speed (<2m / s), and the diffusion influence between devices can be ignored. Before use, the Suma tank was cleaned 3 times with high-purity N2 and evacuated, and the vacuum degree was checked before sample collection. For the collected gas samples, the TO14 and TO15 standards recommended by US EPA were used, and the GC-MS / FID or GC-TOF / MS coupling instrument was used to perform full spectrum analysis on the sampled samples. The standard gas volume of 20nmol / mol was 600mL, and 15mL, 30mL, 60mL, 120mL, 240mL, and 480mL of standard gas were injected for testing. The corresponding sample concentrations (sampling volume was 600mL) were 0.5nmol / mol, 1nmol / mol, 2nmol / mol, 4nmol / mol, 8nmol / mol, and 16nmol / mol. The calibration curve was drawn by the peak area of ​​the concentration and the quantitative ion to determine the linear relationship. In order to meet the test requirements, the linear correlation coefficient of the established standard curve must be greater than 0.99. The analysis process mainly analyzes 107 volatile organic compounds in the C2-C12 range in PAMS+TO15, covering multiple types of substances such as alkanes, alkenes, benzene series, halogenated hydrocarbons, oxygen-containing compounds, etc. The sample is qualitatively analyzed by the retention time of each organic compound and the standard mass spectrum search in the spectral library, and the internal standard method is used for quantification. The internal standards include bromochloromethane, 1,4-difluorobenzene, chlorobenzene-d5 and p-bromofluorobenzene.

[0096] After full spectrum analysis of the gas sample, the component concentration analysis data corresponding to the gas sample is obtained. By using MATLAB programming software and combining the RMS algorithm for iterative calculation, the RMS value X corresponding to the sample components at different sampling points of the same device can be obtained. rms , as the average component concentration of multiple VOCs samples of the device, that is, the comprehensive source spectrum of VOCs emission sources of different production devices.

[0097] The root mean square value X rms The calculation expression is as follows:

[0098]

[0099] where N is the number of sampling points, and X i is the content of compound X at different sampling points of the same device.

[0100] During the process of obtaining the comprehensive source spectrum of the production device, standard averaging is performed. However, after standard averaging, the difference between different components may be reduced. Therefore, the root mean square value algorithm is used here to fully consider the characteristic differences between components, so as to obtain data closer to the true components of VOCs in the production device.

[0101] Then, the similarity between the average component concentration obtained by the root mean square algorithm and the sample component concentration at the corresponding sampling point is obtained by using the similarity calculation method. If the similarity value between the comprehensive source spectrum of the device and the sample component concentrations at each sampling point is lower than the preset similarity judgment standard, the sample concentration analysis data corresponding to the lowest similarity value among the similarity values lower than the preset similarity judgment standard and the gas sample corresponding to the sample concentration analysis data are excluded, and the average component concentration of the remaining gas samples is calculated again by using the root mean square algorithm, as well as the similarity value between the average component concentration of the remaining gas samples calculated by using the similarity calculation method and the component concentrations of the remaining samples, until the similarity value is above the preset similarity judgment standard. Finally, the average component concentration of the remaining gas samples of the device is used as the comprehensive source spectrum of the device. The similarity value above the preset similarity judgment standard ensures that the comprehensive source spectrum can represent the intrinsic information of the components of each production device in the petrochemical enterprise, and at the same time can distinguish between different devices, achieving the effect of classification and recognition.

[0102] Specifically, the similarity calculation method adopts the cosine similarity calculation method or the Pearson correlation coefficient calculation method. Among them, the form of the cosine similarity calculation method is:

[0103]

[0104] where 0 < cosθ < 1. The more similar the two vectors are, the smaller the angle between the vectors. The cosine value ranges from 0 to 1, and the closer it is to 1, the more similar they are.

[0105] Furthermore, the comprehensive source spectrum of the pollution source can be updated multiple times. Since there are certain differences in the pollutants emitted by different devices under different production conditions, in order to ensure the accuracy of the source spectrum information of different devices, multiple measurements can be carried out within a period of time to effectively update the device source spectrum, so as to ensure that the comprehensive source spectra of different pollution sources are accurate and effective.

[0106] Furthermore, the method further comprises:

[0107] The air pollutant monitoring data is processed as follows:

[0108] The atmospheric pollutant monitoring data with a value higher than a preset pollutant photochemical reaction activity standard and lower than a preset atmospheric pollutant detection rate standard are eliminated, and the atmospheric pollutant monitoring data after the elimination process is used as the input of the positive definite matrix factor decomposition model.

[0109] In order to ensure the accuracy of air pollutant monitoring data, the photochemical reaction activity standards of pollutants and the detection rate standards of air pollutants are preset, and the air pollutants within a period of time (for example, one year) are eliminated according to the photochemical reaction activity of organic pollutants and the detection rate of air pollutants. The eliminated pollutants include two categories: one is pollutants with high photochemical reaction activity and short environmental life. Such substances are prone to photochemical reactions during the transmission process from the pollution source to the online analysis and monitoring point, resulting in changes in the types of air pollutants and the concentrations of pollutants, affecting the accuracy of traceability; the other is pollutants that have not been detected for a long time or whose detection rate is lower than the preset ratio (for example, 60%). After obtaining the air pollutants that need to be eliminated, the corresponding air pollutants within the time period are eliminated according to the photochemical reaction activity of organic pollutants and the detection rate of air pollutants to ensure the accuracy of traceability.

[0110] There are two main benchmarks for setting the photochemical reactivity standards for pollutants. One is to refer to the K OH The other is to refer to the K value of 1,2,3-trimethylbenzene. OH value, but due to the use of toluene K OH The value used as the standard will eliminate more pollutants of high concern, so 1,2,3-trimethylbenzene is preferentially selected as the benchmark to screen characteristic compounds.

[0111] Arrange the atmospheric pollutant monitoring data after data elimination into a pollutant concentration file corresponding to the positive definite matrix factor decomposition model;

[0112] Calculating the uncertainty of the concentration of air pollutants according to the air pollutant monitoring data after the elimination process to obtain an uncertainty file;

[0113] The pollutant concentration file and the uncertainty file are used as inputs to the positive definite matrix factorization model.

[0114] Among them, the atmospheric pollutant monitoring data can be collected in real time as the data to be analyzed, or the atmospheric pollutant monitoring data of a certain period of time can be used as the data to be analyzed, and the atmospheric pollutant monitoring data can be organized into a pollutant concentration file corresponding to the positive definite matrix factor decomposition model, and the uncertainty of the atmospheric pollutant concentration of the time period is calculated according to the atmospheric pollutant monitoring data to obtain an uncertainty file, and the pollutant concentration file and the uncertainty file are used as inputs of the positive definite matrix factor decomposition model. The pollutant concentration file includes the concentration value of VOCs in the input atmospheric pollutant monitoring data, each row represents a sample of a certain period of time, the first column is time, and the 2nd to mth columns are the concentrations of each pollutant respectively. In the uncertainty file, the naming and order of the column headers and row headers must be consistent with the table format of the pollutant concentration file.

[0115] Furthermore, before importing the determined pollutant concentration file and uncertainty file into the PMF model, it is necessary to determine the number of traceability characteristic factors of the positive definite matrix factor decomposition model, and input the number of traceability characteristic factors into the positive definite matrix factor decomposition model. According to the model fitting results of each pollutant, the level of each pollutant is divided into strong, weak, and eliminated. The fitting results are evaluated by the residual matrix and the fitting coefficient of the model for the predicted value and calculated value of the compound. The residual matrix is ​​kept between -3 and 3, and the residual matrix is ​​kept as small as possible. The PMF model obtains time series related data on the contribution of atmospheric pollutant sources. For example, if you want to obtain the hourly changes in the sources of atmospheric pollutants, you can input the hourly change data for a period of time into the PMF model for analysis. If you want to obtain the changes in pollution sources under conditions such as day, month, season, year, and different wind directions, you can input the corresponding day, month, season, year, and different wind directions. Conditional change data can be input into the PMF model for analysis.

[0116] Usually, the number of effective receptor samples in the PMF model should be no less than 100. The number of traceability characteristic factors is an integer between 1 and 50. The number of traceability factors depends on the collected atmospheric pollutant monitoring data, the time resolution of the collection, etc., and is determined according to actual conditions. The traceability characteristic factors include: pollution sources generated by production equipment within the enterprise, pollution sources generated by public transportation, or pollution sources generated by other man-made or natural factors.

[0117] Furthermore, the atmospheric pollutant source analysis results are atmospheric pollutant concentration change values ​​and contribution rates of different pollution factors distributed in time series; the abnormal values ​​are determined by the atmospheric pollutant concentration change values;

[0118] The method further comprises:

[0119] The contribution rates of different pollution factors in a certain time period and the meteorological data corresponding to the time period are used as input, and a conditional probability distribution diagram is obtained using a conditional probability function. The conditional probability distribution diagram is used to show the source direction of different pollution factors.

[0120] Since the source analyzed by the PMF model has no specific physical meaning, it is necessary to manually judge the possible physical meaning. Because VOCs components will overlap, the analytical factors of the PMF model alone may not accurately correspond to the production unit. Therefore, we added a conditional probability distribution diagram to judge the possible pollution source direction and further determine the pollution source. Based on the large amount of atmospheric pollutant monitoring data accumulated in the early stage, we chose to use the PMF model to obtain the contribution rate of the pollution factors of each production unit near the online analysis monitoring point. However, due to the large number of external enterprises around a certain refinery, many enterprises' processing technology is related to benzene, toluene, xylene, and alkane compounds, which overlaps with the components of some refining production units. Only using the PMF model to trace the source and analyze the possible pollution source requires experience for manual judgment, which will lead to certain errors. In response to this problem, it is proposed to use the conditional probability function to obtain the conditional probability distribution diagram combined with the results of the PMF model for judgment, and use the distribution relationship diagram between the pollution factor and the meteorological parameter to judge the source direction of each pollution factor, and then combine the process characteristics of the possible pollution source to determine the actual meaning of the factor, so that the pollution source can be determined more accurately.

[0121] Specifically, the specific steps to obtain the conditional probability distribution diagram are as follows: atmospheric pollutant monitoring data usually takes data collected within a certain period of time as input, and the contribution rate of each pollution factor in the period of time obtained by the PMF model calculation is combined with meteorological data, including: wind speed and wind direction data as the input of the conditional probability function, by calculating the percentage of the number of samples in each wind direction that is higher than a preset value to the total number of samples in the same wind direction, and then calculating the percentage of the number of samples in all wind directions that are higher than a preset value to the total number of samples in the same wind direction, thereby obtaining a conditional probability distribution diagram. In the process of determining the target production device according to the meteorological data corresponding to the time of occurrence of the abnormal value, in addition to using meteorological data for determination, the target production device can also be determined in combination with the conditional probability distribution diagram. The high pollution characteristics of the pollution factor in a certain direction indicate that there may be an emission source of the pollution factor in that direction. Through the conditional probability distribution diagram, the source direction of the pollutant and the conditional probability of the pollutant can be intuitively obtained.

[0122] Furthermore, the method further comprises:

[0123] Eliminating abnormal data from the abnormal values, wherein the abnormal data include monitoring data obtained when the atmospheric pollutant monitoring equipment does not work properly and abnormal meteorological data, and using the atmospheric pollutant monitoring data corresponding to the occurrence time of the eliminated abnormal values ​​as input to the chemical mass balance model;

[0124] The determining of the target production device according to the meteorological data corresponding to the occurrence time of the abnormal value includes: determining the production device in the upwind position as the target production device according to the dominant wind direction in the meteorological data.

[0125] Specifically, there are abnormal values ​​(outliers) in the time series of the atmospheric pollutant source analysis results. At this time, the abnormal values ​​may be data anomalies caused by changes in atmospheric pollutant concentrations, or may be data anomalies caused by force majeure factors such as failure of the collection instrument itself, abnormal weather, etc. Therefore, first determine whether the collection device is operating normally at the time when the data anomaly occurs, and determine whether there is abnormal weather such as strong winds and heavy rains. If the data anomalies caused by force majeure factors are excluded, the chemical mass balance model is used to specifically analyze the monitoring data and meteorological data of the online analysis hut at the time when the abnormal value occurs, and determine the possible pollution sources in the upwind position according to the hourly dominant wind direction. According to the established pollution source spectrum information and the monitoring data of the online analysis monitoring point (i.e., the online analysis hut), the chemical mass balance model is called to calculate the atmospheric pollutants corresponding to the time of appearance of the abnormal value after the abnormal value is eliminated. The monitoring data and the comprehensive source spectrum of the pollution sources corresponding to the target production device are used as input, and the contribution rate of each pollution source to the online analysis monitoring point is calculated. Through this method, the contribution concentration of the pollution source to the online analysis monitoring point can be obtained at the same time. By comparing the changes in the contribution rate of each pollution source to the online analysis monitoring point, if the contribution rate of the pollution source to the online analysis monitoring point becomes larger, it means that this pollution source has produced an abnormality compared with the normal state, and the pollutant emission has increased. According to the comprehensive source spectrum information of different pollution sources, the pollution source is located to a specific device area, associated with the production device, and promptly fed back to the corresponding on-site investigation personnel. A portable TVOCs meter or other related detection equipment can be used to conduct rapid monitoring in the corresponding device area to screen points where abnormal conditions may exist, such as whether valves or flanges are leaking, etc., to determine possible abnormal conditions, avoid the serious consequences that may be caused by further deterioration of abnormal conditions, and provide a basis for targeted, rapid and accurate guidance to solve abnormal problems.

[0126] Among them, the accuracy of the CMB model is judged based on parameters such as the residual sum of squares (the value of the residual sum of squares is less than 2, indicating that the result is acceptable), the regression coefficient R2 (R2 is greater than 0.8, indicating a good degree of fit), the percentage mass PM (between 0.8-1.2, indicating acceptable), and the sensitivity matrix (materials with absolute values ​​between 0.5-1 have the greatest impact).

[0127] Figure 3 Schematic diagram of the structure of a precise source tracing and analysis device provided by one embodiment of the present invention. Figure 3 As shown, an embodiment of the present invention provides a precise source tracing and analysis device, comprising:

[0128] An acquisition module 10 is used to acquire meteorological data of the area corresponding to the production device, atmospheric pollutant monitoring data, and a comprehensive source spectrum of pollution sources corresponding to each production device;

[0129] A first analysis module 20, used to analyze the air pollutant monitoring data to obtain corresponding air pollutant source analysis results;

[0130] The second analysis module 30 is used to determine the target production device according to the meteorological data corresponding to the time when the abnormal value appears in the atmospheric pollutant source analysis result, and obtain the pollution contribution rate corresponding to the target production device based on the atmospheric pollutant monitoring data corresponding to the time when the abnormal value appears and the comprehensive source spectrum of the pollution source corresponding to the target production device;

[0131] The determination module 40 is used to analyze the change of the pollution contribution rate and determine the production equipment that needs to be checked from the target production equipment.

[0132] An embodiment of the present invention further provides a machine-readable storage medium having computer program instructions stored thereon, wherein the computer program instructions, when executed by a processor, implement the above-mentioned precise source tracing and analysis method.

[0133] Common units in refineries include atmospheric and vacuum units, catalytic cracking, catalytic reforming, catalytic hydrogenation, delayed coking, sewage treatment plants, chemical production, storage tank areas, etc. Different production units are divided into multiple process sections. Different production units are divided according to production raw materials, processing technology, product requirements, etc. to ensure that enough VOCs samples from different units are collected, covering the range of different production process sections. The composition of VOCs samples is determined by GC-MS / FID analysis instruments, which includes 107 compounds such as alkanes, alkenes, alkynes, aromatics, halogenated hydrocarbons, oxygen-containing organic matter, sulfides, etc.

[0134] Embodiment 1:

[0135] Taking the wax oil hydrogenation unit as an example, the wax oil hydrogenation unit includes the reaction part, the distillation part, the circulating hydrogen desulfurization and low-fraction gas desulfurization, PSA hydrogen purification and other parts. In order to make the comprehensive source spectrum more comprehensive, a total of 16 sampling points are set outside the wax oil hydrogenation unit, distributed in various process sections of wax oil hydrogenation; Sample collection: Sampling is carried out under low wind speed (<2m / s) to reduce the influence of diffusion transmission between various devices. When sampling, a 3.2L Suma tank is used for collection, the sampling flow rate is 50ml / min, the sampling height is 1.5m, and the collection time is 1 hour. The TO14 and TO15 methods recommended by the US EPA are used, and the pre-concentration GC-MS / FID analysis instrument is used in conjunction for full spectrum analysis.

[0136] In this example, the preset similarity judgment standard is set to 0.85. After obtaining the gas sample component concentrations at multiple sampling points of the device, MATLAB is used in combination with the root mean square algorithm to process the average component concentration of the gas samples of the production device. The angle cosine method is used to calculate the similarity between the comprehensive source spectrum and the gas sample composition at different sampling points. Finally, there are 12 samples with a similarity greater than 0.85, as shown in the following figure. Figure 4 The schematic diagram of the calculation results of the VOCs component similarity between the comprehensive source spectrum of the wax oil hydrogenation unit and each sampling point is shown in Figure 4 It can be concluded that the similarity between the comprehensive source spectrum and the gas sample components at different points is mostly greater than 0.9, indicating that the comprehensive source spectrum is very similar to the composition of the samples at each point, and the comprehensive source spectrum of the device can represent the concentration distribution of VOCs composition at different points of the device.

[0137] Embodiment 2:

[0138] Taking the catalytic cracking unit as an example, a comprehensive source spectrum of VOCs in the catalytic cracking unit is established.

[0139] Using the VOCs sampling and analysis method in Example 1, 20 atmospheric VOCs samples from catalytic cracking units were obtained. The comprehensive source spectrum of the VOCs samples of the unit was obtained by using MATLAB combined with the root mean square algorithm. The difference between the comprehensive source spectrum and the VOCs composition at different points was calculated using the angle cosine method. Finally, there were 15 samples with a similarity greater than 0.85, and the following was obtained: Figure 5 The schematic diagram of the calculation results of the VOCs component similarity between the comprehensive source spectrum of the catalytic cracking unit and each sampling point is shown; Figure 5 It can be seen from the data that the similarity between the comprehensive source spectrum and the composition at different points is mostly greater than 0.9, indicating that the comprehensive source spectrum is very similar to the composition at each point. The comprehensive source spectrum of the device can represent the concentration distribution of VOCs composition at different points of the device.

[0140] Embodiment 3:

[0141] Similarity calculation between comprehensive source spectra of different devices.

[0142] In this embodiment, samples were collected from common production unit areas of refineries such as catalytic reforming, wax oil hydrogenation, residue oil hydrogenation, atmospheric and vacuum units, catalytic cracking, delayed coking, and sewage treatment plants. The VOCs sampling and analysis method in Example 1 was used, and MATLAB was used in combination with the root mean square algorithm to obtain the comprehensive source spectrum of the VOCs samples of the different production units. The angle cosine method was used to calculate the difference between the comprehensive source spectrum and the VOCs composition at different points, and the following was obtained: Figure 6 Schematic diagram of the calculation results of the similarity of the comprehensive source spectra of different devices shown; Figure 6 It can be seen that the comprehensive source spectra of different devices are very different, with similarities below 0.85, and most of them below 0.5. This shows that the obtained comprehensive source spectrum can represent the composition of the production equipment of petrochemical enterprises, and at the same time can distinguish different devices to achieve the effect of classification and identification.

[0143] Embodiment 4:

[0144] In this embodiment, an oil refinery is taken as the analysis object. Figure 7 As shown, Figure 7 This is a schematic diagram of the location distribution of the main refining units and online analysis monitoring points in the refinery. The online analysis data collected from July 2019 to September 2019 at the online analysis monitoring point was traced and analyzed. The data was analyzed on a daily basis. First, the data was organized into an input file corresponding to the PMF model. The residual matrix analysis and the correlation coefficient between the model prediction value and the measured value were used to determine the model fitting effect. The correlation coefficients of the strongly correlated compounds in the established model were all above 0.9, and the residual matrix was basically maintained between -3 and 3, and was normally distributed, indicating that the model fitting effect was very good. Combined with the production environment, after multiple debugging, a total of 6 major pollution factors were determined, and the 6 major pollution factors were named Factor 1, Factor 2, Factor 3, Factor 4, Factor 5, and Factor 6 respectively; Figure 8 The conditional probability distribution diagram of the main pollution factors and the pollutant compounds in each factor is shown, showing the conditional probability distribution diagram of the concentrations of the six main pollution factors and the pollutant compounds in each factor from July 2019 to September 2019. Figure 8 The content in the figure shows the types of 41 pollutant compounds in factors 1, 2, 3, 4, 5, and 6, the percentage of the 41 pollutant compounds in the total compounds, and the source direction of the pollution source. The source direction and conditional probability of the pollution source are represented by the gradient of color. The darker the color, the higher the pollutant concentration, and the more likely it is that the direction is the source direction of the pollution source. Taking factor 1 as an example, from Figure 8It can be concluded that the main pollutant compounds of factor 1 are n-hexane, 2,3-dimethylbutane, 2-methylpentane, 3-methylpentane, etc. From the perspective of the main pollutant compounds, they may come from device 5 in the west and device 4 in the northeast. However, further judgment using the conditional probability distribution diagram shows that the main pollution source comes from the westerly direction of the online analysis cabin, so factor 1 has a greater probability of being device 5; similarly, this method can be used to determine that the five pollution factors of factor 2, factor 3, factor 4, factor 5, and factor 6 correspond to device 2, device 5, device 3, device 1, and device 6, respectively, and the contribution rates during this period are 15.8%, 19.1%, 7.9%, 12.5%, 25.8%, and 19.0%, respectively. Fig. 9 From the schematic diagram of the pollution contribution rate of each device to the online analysis monitoring point, it can be found that the main sources of pollution may be device 1, device 5, device 2, and device 6.

[0145] Embodiment 5:

[0146] In this embodiment, the refinery is taken as the analysis object. There are outliers in the source analysis results of the PMF model. The CMB model is used to focus on analyzing the monitoring data and meteorological data of the online analysis monitoring points at the same time point. The PMF model is used to obtain the following: Fig.10 The time series diagram of the total concentration of atmospheric pollutants shown in Fig.10 It can be concluded that the observed concentration values ​​and predicted concentration values ​​before September 30, 2019 are relatively average. There are abnormal values ​​in the total concentration of atmospheric pollutants on September 30, 2019. The observed concentration values ​​and predicted concentration values ​​on September 30 are significantly higher than those on other days in September. First of all, we need to rule out abnormal data caused by force majeure such as instrument failure. After investigation, it is found that the above force majeure reasons did not occur. Therefore, it is necessary to analyze whether the working conditions of the contributing sources are normal, and further calculate the contribution rate changes of each pollution source for this data. The CMB model is used to call the monitoring data and meteorological data of the online analysis monitoring point at this time point, and the possible pollution sources in the upwind position are determined according to the hourly dominant wind direction. According to the established pollution source source spectrum information and the monitoring data of the online analysis monitoring point, the contribution rate of each pollution source to the online analysis monitoring point is calculated. According to the meteorological information, the main wind direction on September 30 is westerly. According to Example 4 Figure 7The schematic diagram of the location distribution of major refining units and online analysis monitoring points in the refinery is shown. The possible pollution sources are judged to be unit 1, unit 2, and unit 5. After calculation and comparison, it is found that the pollution contribution rate of unit 1 and unit 2 has no obvious change when the abnormal value occurs compared with the normal operating conditions. The pollution contribution rate of unit 5 increases by 20% when the abnormal value occurs compared with the normal operating conditions. Further communication with the on-site duty personnel determines that there are tanker loading and unloading operations in the unit area corresponding to unit 5 during this period, which may lead to a sudden increase in the contribution rate of pollutants at that time. The problem will be fed back to the corresponding management personnel later to further improve the VOCs treatment and control measures in the tanker loading and unloading process.

[0147] Those skilled in the art will understand that all or part of the steps in the method for implementing the above-mentioned embodiments can be completed by instructing the relevant hardware through a program, and the program is stored in a storage medium, including several instructions for making a single-chip microcomputer, a chip or a processor (processor) perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0148] The optional embodiments of the present invention are described in detail above in conjunction with the accompanying drawings. However, the embodiments of the present invention are not limited to the specific details in the above embodiments. Within the technical concept of the embodiments of the present invention, the technical scheme of the embodiments of the present invention can be subjected to a variety of simple modifications, and these simple modifications all belong to the protection scope of the embodiments of the present invention. It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. In order to avoid unnecessary repetition, the embodiments of the present invention will not further describe various possible combinations.

[0149] In addition, various embodiments of the present invention may be arbitrarily combined, and as long as they do not violate the concept of the embodiments of the present invention, they should also be regarded as the contents disclosed in the embodiments of the present invention.

Claims

1. A precise traceability analysis method, characterized in that: The method comprises: Obtain meteorological data, atmospheric pollutant monitoring data, and comprehensive source spectrum of pollution sources corresponding to each production device in the area corresponding to the production device; The air pollutant monitoring data is processed as follows: Eliminate air pollutant monitoring data that are higher than the preset pollutant photochemical reaction activity standard and lower than the preset air pollutant detection rate standard; The atmospheric pollutant monitoring data after data elimination are sorted into pollutant concentration files corresponding to the principal component analysis model, the positive definite matrix factor decomposition model, the ME2 model or the UNMIX model as the input of the corresponding model; Call the principal component analysis model, positive definite matrix factorization model, ME2 model or UNMIX model, and use the pollutant concentration file as input to obtain the corresponding atmospheric pollutant source analysis results; If there are abnormal values ​​in the atmospheric pollutant source analysis results, the production unit located in the upwind direction is determined as the target production unit according to the dominant wind direction in the meteorological data corresponding to the time when the abnormal value occurs, and the pollution contribution rate corresponding to the target production unit is obtained based on the atmospheric pollutant monitoring data corresponding to the time when the abnormal value occurs and the comprehensive source spectrum of the pollution source corresponding to the target production unit; Analyze the change of the pollution contribution rate and determine the production equipment to be checked from the target production equipment; The step of obtaining a comprehensive source spectrum of pollution sources corresponding to each production device includes: For each production unit: Collecting concentration data of each gas sample of the production device; Calculation steps: Calculate the average component concentration of the gas sample; Elimination step: eliminating the gas sample whose concentration data has the lowest similarity with the average component concentration in the gas sample; For the remaining gas sample, the calculation step and the elimination step are repeatedly performed until the similarity between the concentration data of the remaining gas sample and the average component concentration of the remaining gas sample is greater than a preset value, and the average component concentration of the remaining gas sample is used as the characterizing component concentration of the gas sample of the production device; According to the concentration of the characteristic components of the gas samples of each production device, a comprehensive source spectrum of the pollution sources corresponding to each production device is prepared; Based on the atmospheric pollutant monitoring data corresponding to the occurrence time of the abnormal value and the comprehensive source spectrum of the pollution source corresponding to the target production device, the pollution contribution rate corresponding to the target production device is obtained, including: Eliminate abnormal data from the abnormal values, and use the atmospheric pollutant monitoring data corresponding to the occurrence time of the eliminated abnormal values ​​as the input of the chemical mass balance model, wherein the abnormal data includes monitoring data obtained when the atmospheric pollutant monitoring equipment does not work properly and abnormal meteorological data; Calling a chemical mass balance model, taking the atmospheric pollutant monitoring data corresponding to the appearance time of the removed abnormal values ​​and the comprehensive source spectrum of the pollution sources corresponding to the target production device as input, to obtain the pollution contribution rate corresponding to the target production device; The atmospheric pollutant source analysis results are the atmospheric pollutant concentration change values ​​and the contribution rates of different pollution factors distributed in time series; the abnormal values ​​are determined by the atmospheric pollutant concentration change values; The method further comprises: The contribution rates of different pollution factors in a certain time period and the meteorological data corresponding to the time period are used as input, and a conditional probability distribution diagram is obtained using a conditional probability function. The conditional probability distribution diagram is used to show the source direction of different pollution factors.

2. The precise source tracing analysis method according to claim 1 is characterized in that: The comprehensive source spectrum of the pollution sources may be updated multiple times.

3. The precise source tracing analysis method according to claim 1 is characterized in that: The method further comprises: A portable TVOC measuring instrument is used to check for pollution leakage points in the production equipment that needs to be checked.

4. A precise traceability analysis device, applied to the precise traceability analysis method according to any one of claims 1 to 3, characterized in that: The device comprises: An acquisition module is used to acquire meteorological data of the area corresponding to the production device, atmospheric pollutant monitoring data, and a comprehensive source spectrum of pollution sources corresponding to each production device; The first parsing module is used to process the air pollutant monitoring data as follows: Eliminate air pollutant monitoring data that are higher than the preset pollutant photochemical reaction activity standard and lower than the preset air pollutant detection rate standard; Arrange the atmospheric pollutant monitoring data after data elimination into a pollutant concentration file corresponding to the principal component analysis model, the positive definite matrix factor decomposition model, the ME2 model or the UNMIX model as the input of the corresponding model; Call the principal component analysis model, positive definite matrix factorization model, ME2 model or UNMIX model, and use the pollutant concentration file as input to obtain the corresponding atmospheric pollutant source analysis results; The second analysis module is used to determine the production device in the upwind position as the target production device according to the dominant wind direction in the meteorological data corresponding to the time when the abnormal value appears in the atmospheric pollutant source analysis result, and obtain the pollution contribution rate corresponding to the target production device based on the atmospheric pollutant monitoring data corresponding to the time when the abnormal value appears and the comprehensive source spectrum of the pollution source corresponding to the target production device; A determination module, used to analyze the change of the pollution contribution rate and determine the production equipment to be checked from the target production equipment; The acquisition module is also used for: For each production unit: Collecting concentration data of each gas sample of the production device; Calculation steps: Calculate the average component concentration of the gas sample; Elimination step: eliminating the gas sample whose concentration data has the lowest similarity with the average component concentration in the gas sample; For the remaining gas sample, the calculation step and the elimination step are repeatedly performed until the similarity between the concentration data of the remaining gas sample and the average component concentration of the remaining gas sample is greater than a preset value, and the average component concentration of the remaining gas sample is used as the characterizing component concentration of the gas sample of the production device; According to the concentration of the characteristic components of the gas samples of each production device, a comprehensive source spectrum of the pollution sources corresponding to each production device is prepared; The second parsing module is also used for: Eliminate abnormal data from the abnormal values, and use the atmospheric pollutant monitoring data corresponding to the occurrence time of the eliminated abnormal values ​​as the input of the chemical mass balance model, wherein the abnormal data includes monitoring data obtained when the atmospheric pollutant monitoring equipment does not work properly and abnormal meteorological data; Calling a chemical mass balance model, taking the atmospheric pollutant monitoring data corresponding to the appearance time of the removed abnormal values ​​and the comprehensive source spectrum of the pollution sources corresponding to the target production device as input, to obtain the pollution contribution rate corresponding to the target production device; The atmospheric pollutant source analysis results are the atmospheric pollutant concentration change values ​​and the contribution rates of different pollution factors distributed in time series; the abnormal values ​​are determined by the atmospheric pollutant concentration change values; The determining module is also used for: The contribution rates of different pollution factors in a certain time period and the meteorological data corresponding to the time period are used as input, and a conditional probability distribution diagram is obtained using a conditional probability function. The conditional probability distribution diagram is used to show the source direction of different pollution factors.

5. The precise source tracing and analysis device according to claim 4, characterized in that: The comprehensive source spectrum of the pollution sources may be updated multiple times.

6. A machine-readable storage medium having instructions stored thereon, the instructions being used to enable a machine to execute the precise traceability and analysis method described in any one of claims 1 to 3.

Images