Comprehensive ozone source apportionment method based on OBM and EBM multi-mode

Abstract

The present invention discloses an ozone comprehensive source analysis method based on OBM and EBM multi-modes, and relates to the technical field of air pollution prevention and control; this ozone comprehensive source analysis method based on OBM and EBM multi-modes integrates the analysis results of multiple models, and can avoid the deviation and uncertainty of the analysis results of a single model to a certain extent, improve the accuracy and reliability of the final analysis results, have high operability and promotion value, and provide reliable technical support for ozone pollution control by environmental management departments. The calculated weight coefficient has a reference value for evaluating the accuracy of the analysis results of a single model, and has guiding significance for guiding the optimization of single model parameters; at the same time, this method makes up for some problems existing in the use of a single model, and has certain scientific value for improving the refined source analysis methods of ozone and particulate matter.

Description

Technical Field

[0001] The present invention relates to the technical field of air pollution prevention and control, and in particular to an ozone comprehensive source analysis method based on OBM and EBM multi-modes. Background Art

[0002] Source apportionment is a method to qualitatively or quantitatively identify the sources of ozone pollutants in environmental receptors through chemical, physical, mathematical and other methods. Currently, ozone source apportionment methods include observation-based modeling (OBM) and source emission-based modeling (EBM).

[0003] Emission-based modeling (EBM) refers to a technical method based on local high-temporal and spatial resolution emission inventory data and meteorological field data, quantitatively describing the atmospheric physical and chemical processes that ozone precursors undergo from emission to ozone generation using numerical model methods at different scales, and quantitatively estimating the contribution of emissions from different regions and different types of pollution sources to the formation of ozone and its precursors in ambient air. Commonly used EBM analysis methods include the secondary pollutant source analysis model NAQPMS-OSAM based on the nested grid air quality forecast model system and the ozone source analysis model CAMx-OSAT / PSAT based on the open source air quality model. Due to the large uncertainty of emission factors, lack of data on the level of human pollution activities, and difficulty in accurately counting emissions from various sources, there are large errors in the source list. At the same time, due to the complex migration and transformation process of pollutants in the atmospheric environment, it is difficult to establish the relationship between pollution sources and receptors, and the analysis results often have large uncertainties. Therefore, whether it is NAQPMS-OSAM or CAMx-OSAT / PSAT, the analysis results are obviously affected by the errors in the pollution source list and the complex migration and transformation mechanism, and the uncertainty is large.

[0004] The observation-based analysis method (OBM) refers to a technical method that uses data analysis and modeling methods to determine the sensitivity of ozone formation based on actual observation data of ozone and precursors, and uses receptor models to quantitatively analyze the contribution of various VOCs pollution sources to the formation of ozone and its precursors. Commonly used source analysis methods include the chemical mass balance model (CMB) and the orthogonal matrix factor analysis model (PMF).

[0005] Chemical mass balance (CMB) is a source apportionment method that constructs a set of linear equations based on the principle of mass conservation and calculates the contribution concentration of various emission sources to the receptors by comparing the receptor concentration of each chemical component with the content of this chemical component in the composition spectrum of various emission sources. The application of chemical mass balance method requires the following assumptions: (1) The chemical composition of each pollution emission source is relatively stable during the source sample collection and the transmission from the source to the receptor; (2) There is no interaction between the chemical components of the particulate matter emitted by each source type, that is, linear addition is satisfied; (3) The contribution of each source type to the receptor pollutants is certain, and the chemical composition of the pollutants emitted is significantly different; (4) The composition spectra of each source are relatively independent and there is no collinearity; (5) The measurement uncertainty is independent and random, and obeys the normal distribution; (6) The number of chemical components in the source must be greater than or equal to the number of sources. In practical applications, it is very difficult to meet all six conditions at the same time, so there is a certain deviation in the analysis results of the CMB model. At the same time, the source composition spectrum has a great influence on the accuracy of the results of the chemical mass balance model. Due to the lack of local pollution source composition spectrum, there is a certain degree of subjectivity and experience in the selection of emission sources, which also affects the analysis results. In addition, the analysis results of CMB often have negative values, which makes the analysis results lack actual physical meaning.

[0006] Positive definite matrix factorization (PMF) is a factor analysis method based on the least squares method, non-negative constraints on the decomposition matrix, and optimization using data deviation. Compared with traditional factor analysis methods, the positive definite matrix factorization method imposes non-negative constraints on both factor scores and factor loadings during the solution process, avoiding negative values ​​in the matrix decomposition results, making the source component spectrum and source contribution rate interpretable and with clear physical meaning. The positive definite matrix factorization method does not require the measurement of the source component spectrum, and uses error estimation for each individual data point to make data processing more accurate. The positive definite matrix factorization model also has some shortcomings. Although the model can ensure that the contribution rate of the pollution source is positive, it does not take into account the uncertainty of the pollution source emissions; at the same time, the VOC characteristic components of different source types are not unique to this type of source, which also affects the results of source analysis.

[0007] Therefore, integrating the multi-mode source apportionment results of observation-based models (OBM) and emissions-based models (EBM) is conducive to reducing the deviation of source apportionment results of different models, improving the accuracy and reliability of overall source apportionment results, and providing better technical support for environmental management. But in fact, there is no "standard answer" for source apportionment results. Unlike the atmospheric diffusion model that simulates the diffusion of pollutants, the actual monitored environmental pollutant concentrations can be used to determine the accuracy of the simulation results. Therefore, it is very difficult to integrate the analysis results of different models. Summary of the invention

[0008] In view of the deficiencies in the prior art, the present invention provides an ozone comprehensive source analysis method based on OBM and EBM multi-modes, which solves the problems raised in the above-mentioned background technology.

[0009] To achieve the above objectives, the present invention is implemented by the following technical scheme: a comprehensive ozone source analysis method based on OBM and EBM multi-modes, comprising the following steps:

[0010] 1. A comprehensive ozone source analysis method based on OBM and EBM multi-modes, characterized by comprising the following steps:

[0011] S1: The local VOCs emission sources are uniformly divided into Cn sources (n = 1, 2, 3...);

[0012] S2: Establish a local VOCs source composition spectrum based on a unified source classification;

[0013] S3: Calculate the ozone generation potential of each VOCs component of each pollution source;

[0014] S4: Establish a local VOCs emission source inventory based on a unified source classification;

[0015] S5: Monitor the concentration of VOCs components in the local atmospheric environment;

[0016] S6: Convert the environmental concentration of each VOCs component into ozone generation potential;

[0017] S7: importing the obtained VOCs component ozone generation potential data into a positive definite matrix factorization model (PMF);

[0018] S8: importing the obtained ozone generation potential data of VOCs components and the obtained OFP source component spectrum of VOCs into a chemical mass balance model (CMB);

[0019] S9: Use the local source inventory and WRF meteorological simulation data to import into the NAQPMS-OSAM model;

[0020] S10: Calculate the correlation coefficients between the ozone source analysis results of the above three models of Cn pollution sources respectively;

[0021] S11: Calculate the average value R of the correlation coefficient between the PMF model of Cn pollution sources and the source analysis results of other models PMF-Cn , the average correlation coefficient between the CMB model and other model source analysis results R CMB-Cn , the average value of the correlation coefficient between the NAQPMS model and the source apportionment results of other models R NP-Cn ;

[0022] S12: The obtained RPMF-Cn , R CMB-Cn , R NP-Cn Perform normalization;

[0023] S13: Analyze the obtained weight coefficients to determine the reliability of the analysis results of a single model and determine whether the model needs further optimization;

[0024] S14: Calculate the comprehensive contribution ratio of ozone sources of various pollution sources Cn.

[0025] S15: Optimize and adjust the single source analysis model and re-run it according to the above procedure.

[0026] Optionally, S1 uniformly classifies local VOCs into C1, C2, C3, ... C7, representing natural sources, fossil fuel combustion sources, road traffic sources, solvent use sources, process sources, biomass combustion sources and restaurant fume sources, respectively;

[0027] Optionally, S2 calculates the ozone generation potential (OFP) of multiple components of each pollution source according to the following formula to obtain the ozone generation potential (OFP) of each component of VOCs, and then normalizes it to generate the OFP source component spectrum of VOCs;

[0028] OFPi=MIRi×IRi;

[0029] Wherein, i is the number of components, i=1,2,3…n, n is an integer; OFPi is the ozone generation potential of VOCs component i; MIRi is the maximum incremental reaction activity coefficient of VOCs component i, and IRi is the concentration value of VOCs component i.

[0030] Optionally, when S3 is monitoring, the hourly environmental concentration of VOCs components (the components include at least 56 types of PAMS) in a specific period of time and the method detection limit of each component are obtained.

[0031] Optionally, S7 selects representative component data from the VOCs component ozone generation potential data obtained by S6 and imports them into a positive definite matrix factorization model (PMF) to calculate the average value M of the ozone source contribution of each pollution source in a specific period of time. Cn . Where Cn is the pollution source category, n = 1, 2, 3, 4, 5, 6, 7.

[0032] Optionally, S8 imports the ozone generation potential data of the VOCs components obtained in S6 and the OFP source component spectrum of VOCs obtained in step 3 into the chemical mass balance model (CMB), and calculates the average value C of the ozone source contribution of each pollution source in a specific period of time. Cn .

[0033] Optionally, the S9 uses the local source list and WRF meteorological simulation data (the time is consistent with the VOCs environmental monitoring time) to import into the NAQPMS-OSAM model to calculate the average value N of the ozone source contribution of each pollution source in a specific period of time. Cn .

[0034] Optionally, in S12, the analysis results of different models of Cn type pollution sources are normalized to obtain the normalized correlation coefficient P PMF-Cn , P CMB-Cn , P NP-Cn , which are the weight coefficients of various pollution sources in the three models.

[0035] Optionally, the weight coefficient P obtained in S13 PMF-Cn , P CMB-Cn , P NP-Cn Conduct analysis to determine the reliability of the analysis results of a single model and determine whether the model needs further optimization; determine whether the weight coefficient is between 0 and 1; if the weight coefficient is obviously out of range (the range can be set according to the actual situation, and the maximum allowable deviation is generally set to ±10%, that is, outside the range of -0.1 to 1.1 is obviously out of range), then the corresponding model is unreliable for the analysis results of Cn type sources and needs further optimization.

[0036] The present invention provides an ozone comprehensive source analysis method based on OBM and EBM multi-modes, which has the following beneficial effects:

[0037] This comprehensive ozone source analysis method based on OBM and EBM multi-mode integrates the analysis results of multiple models, which can avoid the deviation and uncertainty of the analysis results of a single model to a certain extent, improve the accuracy and reliability of the final analysis results, and has high operability and promotion value. It provides reliable technical support for ozone pollution control in environmental management departments. The calculated weight coefficient has a reference value for evaluating the accuracy of the analysis results of a single model and has guiding significance for guiding the optimization of single model parameters. At the same time, this method makes up for some problems existing in the use of a single model, and has certain scientific value for improving the refined source analysis method of ozone and particulate matter. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is a step diagram of the present invention;

[0039] Figure 2 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0040] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0041] See also Figure 1 to Figure 2 The present invention provides a technical solution: a comprehensive ozone source analysis method based on OBM and EBM multi-mode, comprising the following steps:

[0042] S1: The local VOCS emission sources are uniformly divided into C1, C2, C3, ... C7 sources;

[0043] S2: Establish a local VOCs source composition spectrum based on a unified source classification;

[0044] S3: Calculate the ozone generation potential of each VOCs component of each pollution source;

[0045] S4: Establish a local VOCs emission source inventory based on a unified source classification;

[0046] S5: Monitor the concentration of VOCs components in the local atmospheric environment;

[0047] S6: Convert the environmental concentration of each VOCs component into ozone generation potential;

[0048] S7: importing the obtained VOCs component ozone generation potential data into a positive definite matrix factorization model (PMF);

[0049] S8: importing the obtained ozone generation potential data of VOCs components and the obtained OFP source component spectrum of VOCs into a chemical mass balance model (CMB);

[0050] S9: Use the local source inventory and WRF meteorological simulation data to import into the NAQPMS-OSAM model;

[0051] S10: Calculate the correlation coefficient R between the ozone source contributions of the above three models for each type of pollution source Cn (n = 1, 2, 3, 4, 5, 6, 7) M-C-Cn , R M-N-Cn , R C-N-Cn； Where R M-C-Cn is the correlation coefficient between the source analysis results of the PMF model and the CMB model for Cn type pollution sources; R M-N-Cn is the correlation coefficient between the source apportionment results of the PMF model and the NAQPMS-OSAM model for Cn pollution sources; R C-N-S is the correlation coefficient between the source apportionment results of the CMB model and the NAQPMS-OSAM model for Cn type pollution sources.

[0052] S11: Calculate the average value R of the correlation coefficient between the PMF model of Cn pollution sources and the source analysis results of other models PMF-Cn , the average correlation coefficient between the CMB model and other model source analysis results R CMB-Cn , the average value of the correlation coefficient between the NAQPMS model and the source apportionment results of other models R NP-Cn ;

[0053] See the example results:

[0054]

[0055] Correlation coefficient table of ozone source analysis results of three models

[0056] S12: The obtained R PMF-Cn , R CMB-Cn , R NP-Cn Perform normalization;

[0057] S13: Analyze the obtained weight coefficients to determine the reliability of the analysis results of a single model and determine whether the model needs further optimization;

[0058] S14: Calculate the comprehensive contribution ratio of ozone sources of various pollution sources Cn.

[0059] Those skilled in the art will know that the comprehensive ozone source analysis method based on OBM and EBM multi-mode provided in this embodiment first classifies the local VOCs emission sources, and classifies the local VOCs into C1, C2, C3, ..., C7. According to the unified source classification, a local VOCs source component spectrum is established, and the ozone generation potential of multiple components of each pollution source is calculated according to OFPi=MIRi×IRi to obtain the ozone generation potential (OFP) of each VOCs component, and then normalized to generate VOCs. OFP source component spectrum, according to the unified source classification, establish a local VOCs emission source list, monitor the concentration of VOCs components in the local atmospheric environment, obtain the hourly environmental concentration of VOCs components (components include at least 56 types of PAMS) in a specific period of time and the method detection limit of each component, convert the environmental concentration of each VOCs component into ozone generation potential according to the formula OFPi=MIRi×IRi, import the VOCs component ozone generation potential data obtained in step 6, select representative VOCs component data from them, and import them into the positive definite matrix factor decomposition model (PMF) to calculate the average value M of the ozone source contribution of each pollution source in a specific period of time Cn At the same time, the ozone generation potential data of VOCs components obtained in step 6 and the OFP source component spectrum of VOCs obtained in step 3 are imported into the chemical mass balance model (CMB) to calculate the average value C of the ozone source contribution of each pollution source in a specific period of time. Cn, using the local source list and WRF meteorological simulation data, imported into the NAQPMS-OSAM model, the average value N of the ozone source contribution of each pollution source in a specific period of time was calculated Cn , respectively calculate the correlation coefficient R between the ozone source contributions of the above three models for each type of pollution source Cn (n = 1, 2, 3, 4, 5, 6, 7) M-C-Cn , R M-N-Cn , R C-N-Cn , R M-C-Cn is the correlation coefficient between the source analysis results of the PMF model and the CMB model for Cn type pollution sources; R M-N-Cn is the correlation coefficient between the source apportionment results of the PMF model and the NAQPMS-OSAM model for Cn pollution sources; R C-N-Cn is the correlation coefficient between the CMB model of Cn pollution sources and the source analysis results of the NAQPMS-OSAM model; the average value R of the correlation coefficient between the source analysis results of the PMF model of Cn pollution sources and other models is calculated respectively. PMF-Cn , the average correlation coefficient between the CMB model and other model source analysis results R CMB-Cn , the average value of the correlation coefficient between the NAQPMS model and the source apportionment results of other models R NP-Cn , respectively normalize the analytical results of different models of Cn type pollution sources, and obtain the normalized correlation coefficient P PMF-Cn , P CMB-Cn , P NP-Cn , which are the weight coefficients of various pollution sources in the three models respectively; the obtained weight coefficients are analyzed to judge the quality of the results of a single model and determine whether further optimization is needed. If it is determined that no optimization is needed, the following formula is used to calculate the comprehensive source contribution ratio of ozone of various pollution sources Cn:

[0060] P Sn =M Sn ×P PMF-Sn +C Sn ×P CMB-Sn +N Sn ×P NP-Sn

[0061] Where: P Sn is the weight coefficient of Cn type pollution source;

[0062] Therefore, the contribution of various pollution sources Cn to ozone can be obtained based on the calculated comprehensive ozone source contribution ratio.

[0063] Furthermore, S1 uniformly classifies local VOCS into C1, C2, C3, … C7, representing natural sources, fossil fuel combustion sources, road traffic sources, solvent use sources, technological process sources, biomass combustion sources, and restaurant fume sources, respectively.

[0064] Further, S2 calculates the ozone generation potential of multiple components of each pollution source according to the following formula to obtain the ozone generation potential (OFP) of each VOCs component, and then normalizes it to generate the OFP source component spectrum of VOCs;

[0065] OFPi=MIRi×IRi;

[0066] Wherein, i is the number of components, i=1,2,3…n, n is an integer; OFPi is the ozone generation potential of VOCs component i; MIRi is the maximum incremental reaction activity coefficient of VOCs component i, and IRi is the concentration value of VOCs component i.

[0067] Furthermore, when S5 is monitoring, the hourly environmental concentration of VOCs components (the components include at least 56 types of PAMS) in a specific period of time and the method detection limit of each component are obtained.

[0068] Furthermore, S7 imports the ozone generation potential data of VOCs components obtained by S6, selects representative component data, and imports them into the positive definite matrix factorization model (PMF) to calculate the average value M of the ozone source contribution of each pollution source in a specific period of time. Cn . Where Cn is the pollution source category, n = 1, 2, 3, 4, 5, 6, 7.

[0069] Furthermore, S8 imports the ozone generation potential data of VOCs components obtained in S6 and the OFP source component spectrum of VOCs obtained in step 3 into the chemical mass balance model (CMB) to calculate the average value C of the ozone source contribution of each pollution source in a specific period of time. Cn .

[0070] Furthermore, S9 uses the local source list and WRF meteorological simulation data (the time is consistent with the VOCs environmental monitoring time) and imports them into the NAQPMS-OSAM model to calculate the average value N of the ozone source contribution of each pollution source in a specific period of time. Cn .

[0071] Furthermore, in S12, the analysis results of different models of Cn pollution sources are normalized to obtain the normalized correlation coefficient P PMF-Cn , P CMB-Cn , P NP-Cn , which are the weight coefficients of various pollution sources in the three models.

[0072] See the example results:

[0073]

[0074]

[0075] Table of weight coefficients of each pollution source in three models in ozone comprehensive source analysis

[0076] Further, the weight coefficient P obtained in S13 PMF-Cn , P CMB-Cn , P NP-Cn Analyze and judge the quality of the results of a single model to determine whether further optimization is needed; determine whether the weight coefficient is between 0 and 1; the weight coefficient range of each type of pollution source in the three models is generally between 0 and 1. When the weight coefficient calculated for a certain pollution source is obviously out of range (the range can be set according to the actual situation, and the maximum allowable deviation is generally set to ±10%, that is, outside the range of -0.1 to 1.1 is obviously out of range), it is considered that the corresponding model is not reliable for the analysis results of this type of pollution source, and the corresponding model parameters should be further optimized.

[0077] S15: Optimize and adjust the single source analysis model and re-run the program.

[0078] From the correlation coefficient table of each pollution source of ozone source analysis of the three models, it can be seen that the weight coefficients of the three models for C3 and C4 pollution sources deviate significantly from (0,1). Combined with the correlation coefficient results of the correlation coefficient table of the ozone source analysis results of the three models, due to the high correlation between the PMF and NAQPMS model results, it can be determined that the CMB analysis results for C3 and C4 pollution sources are unreliable, and the model operation parameters need to be optimized or the source component spectrum needs to be reviewed. After the model is optimized and adjusted, re-run from step 7 to step 13 until the weight result range is between 0 and 1 (the maximum allowable error is ±10%); after the model is optimized and adjusted, the example results obtained by re-running are shown in:

[0079]

[0080] Table of weight coefficients of each pollution source in three models in ozone comprehensive source analysis after optimization and adjustment

[0081] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. An integrated ozone source apportionment method based on OBM and EBM multi-modes, characterized by: The following steps are involved: S1: Classify local VOCs emission sources into Cn sources, n = 1, 2, 3...; S2: Establish a local VOCs source composition spectrum based on a unified source classification; S3: Calculate the ozone generation potential of each VOCs component of each pollution source; S4: Establish a local VOCs emission source inventory based on a unified source classification; S5: Monitor the concentration of VOCs components in the local atmospheric environment; S6: Convert the ambient concentration of each VOCs component into ozone generation potential (OFP); S7: importing the obtained VOCs component ozone generation potential data into the positive definite matrix factorization model; S8: importing the obtained ozone generation potential data of VOCs components and the obtained OFP source component spectrum of VOCs into the chemical mass balance model; S9: Use local source inventory and WRF meteorological simulation data to import into NAQPMS-OSAM model; S10: Calculate the correlation coefficients between the ozone source analysis results of the above three models of Cn pollution sources respectively; S11: Calculate the average value R of the correlation coefficient between the positive definite matrix factorization model of Cn type pollution sources and the source analysis results of other models PMF-Cn , the average value of the correlation coefficient between the chemical mass balance model and the source apportionment results of other models R CMB-Cn , the average value of the correlation coefficient between the NAQPMS model and the source apportionment results of other models R NP-Cn ; S12: The obtained R PMF-Cn , R CMB-Cn , R NP-Cn Perform normalization; S13: Analyze the obtained weight coefficients to determine the reliability of the analysis results of a single model and determine whether the model needs further optimization; S14: Calculate the comprehensive contribution ratio of ozone sources of various pollution sources Cn; S15: Optimize and adjust the single source analysis model, and re-run it according to the above steps.

2. The comprehensive ozone source apportionment method based on OBM and EBM multi-mode according to claim 1, characterized in that: The S1 classifies local VOCs emission sources into C1, C2, C3, ... C7, which respectively represent natural sources, fossil fuel combustion sources, road traffic sources, solvent use sources, process sources, biomass combustion sources and restaurant fume sources.

3. The comprehensive ozone source apportionment method based on OBM and EBM multi-mode according to claim 1, characterized in that: The S3 calculates the ozone generation potential of multiple components of each pollution source according to the following formula to obtain the ozone generation potential of each component of VOCs, and then normalizes it to generate the OFP source component spectrum of VOCs; OFPi=MIRi×IRi; Wherein, i is the number of components, i=1,2,3…n, n is an integer; OFPi is the ozone generation potential of VOCs component i; MIRi is the maximum incremental reaction activity coefficient of VOCs component i, and IRi is the concentration value of VOCs component i.

4. The comprehensive ozone source apportionment method based on OBM and EBM multi-mode according to claim 1, characterized in that: When the S5 is monitoring, the hourly environmental concentration of the VOCs components in a specific period of time and the method detection limit of each component are obtained.

5. The comprehensive ozone source apportionment method based on OBM and EBM multi-mode according to claim 1 is characterized in that: S7 selects representative component data from the VOCs component ozone generation potential data obtained by S6 and imports them into the positive definite matrix factor decomposition model to calculate the average value M of the ozone source contribution of each pollution source in a specific period of time. Cn , where Cn is the pollution source category, n = 1, 2, 3,… 6. The comprehensive ozone source apportionment method based on OBM and EBM multi-mode according to claim 1, characterized in that: S8 imports the ozone generation potential data of VOCs components obtained by S6 and the OFP source component spectrum of VOCs obtained by S2 into the chemical mass balance model to calculate the average value C of the ozone source contribution of each pollution source in a specific period of time. Cn。 7. The comprehensive ozone source apportionment method based on OBM and EBM multi-mode according to claim 1 is characterized in that: The S9 uses the local source list and WRF meteorological simulation data, which are consistent with the VOCs environmental monitoring time, and imports them into the NAQPMS-OSAM model to calculate the average value N of the ozone source contribution of each pollution source in a specific period of time. Cn。 8. The comprehensive ozone source apportionment method based on OBM and EBM multi-mode according to claim 1 is characterized in that: In the S12, the analytical results of different models of Cn type pollution sources are normalized to obtain the normalized correlation coefficient P PMF-Cn , P CMB-Cn , P NP-Cn , which are the weight coefficients of various pollution sources in the three models.

9. The comprehensive ozone source apportionment method based on OBM and EBM multi-mode according to claim 1, characterized in that: The weight coefficient P obtained in S13 PMF-Cn , P CMB-Cn , P NP-Cn Conduct analysis to determine the reliability of the analysis results of a single model and determine whether the model needs further optimization; determine whether the weight coefficients of different Cn source models are between 0 and 1; if the weight coefficient is obviously beyond the range, the range is set according to the actual situation, and the maximum allowable deviation is set to ±10%, that is, outside the range of -0.1 to 1.1 is obviously beyond the range, then the corresponding model is considered unreliable for the Cn source analysis results and needs further optimization.

Images