Method and apparatus for matching particulate matter pollution sources based on single-particle mass spectrometry

Abstract

This invention provides a method and apparatus for matching particulate matter pollution sources based on single-particle mass spectrometry. A pollution source library is established by pre-collecting single-particle mass spectrometry data from multiple known pollution source categories. During the detection phase, positive and / or negative ion mass spectra of the particulate matter to be identified are acquired and subjected to mass-to-charge ratio alignment, noise removal, and normalization to obtain the feature vector of the ion to be identified. The matching discrepancy coefficient between the feature vector of the ion to be identified and the reference source spectra of each known pollution source category is calculated, and the candidate pollution source category is determined based on the matching discrepancy coefficient. Furthermore, the feature vector of the ion to be identified is matched with the source-specific fingerprint ion information corresponding to the candidate pollution source category to obtain the fingerprint ion matching result, and the pollution source matching result is output accordingly. This invention can improve the objectivity, accuracy, and interpretability of particulate matter pollution source matching.

Description

Technical Field

[0001] This invention relates to the field of atmospheric particulate matter analysis and testing technology, and more specifically, to a method and apparatus for matching particulate matter pollution sources based on single-particle mass spectrometry. Background Technology

[0002] Single-particle aerosol mass spectrometry (SPAMS) enables real-time online detection of the particle size and chemical composition of individual aerosol particles. It offers advantages such as high temporal resolution, low sample pretreatment requirements, and the ability to obtain chemical information at the single-particle level, making it a crucial technique for analyzing the sources of atmospheric particulate matter and characterizing pollution emissions. This technique typically involves focusing aerosol particles using an aerodynamic lens, measuring particle velocity and calculating particle size using a dual-laser velocimetry system, and then desorbing and ionizing the particles using a pulsed ultraviolet laser. Finally, bipolar time-of-flight mass spectrometry (BIFS) is used to obtain the positive and negative ion mass spectra for each individual particle. Thus, each detected particle can generate a single-particle mass spectrum containing multiple mass-to-charge ratio positions and ion peak intensities, providing a data foundation for subsequent particulate matter source determination.

[0003] In existing technologies, pollution source apportionment based on SPAMS data typically requires the prior collection of particulate matter samples from known pollution sources such as coal combustion, vehicle exhaust, biomass burning, dust, and industrial emissions. Single-particle mass spectrometry data from known sources are then classified using clustering methods such as the ART-2a adaptive resonance theory neural network to identify different particle types. Average spectra or characteristic ions for each particle type are extracted, and reference source spectra or pollution source spectral libraries corresponding to different pollution source categories are constructed. In the application phase, after obtaining single-particle mass spectra from environmental receptor samples or aerosol samples to be detected, the mass spectrometry characteristics of the particulate matter to be identified can be compared with the reference source spectra or characteristic ions in the pollution source spectral library to determine the possible pollution source category of the particulate matter.

[0004] However, existing SPAMS-based methods for detecting particulate matter pollution sources still have shortcomings. Current methods often rely primarily on the presence of characteristic ions, manual comparison, or similarity assessments of clustering results when determining the source of particulate matter to be identified. They lack a quantitative indicator that can directly measure the overall difference between the mass spectrum of a single particle and the reference spectra of known pollution sources. Because the frequency and relative intensity of peaks with different mass-to-charge ratios vary significantly in single-particle mass spectra, relying solely on the presence or absence of a few characteristic ions can easily overlook compositional differences across the entire spectrum, leading to insufficient objectivity and stability in pollution source classification.

[0005] Furthermore, existing technologies typically focus more on frequently occurring characteristic ions, using high-frequency ions as a key basis for pollution source identification. However, ion peaks with lower frequencies but source-specific characteristics are often filtered or weakened during feature screening. In reality, due to the complex chemical composition of particulate matter emitted from different pollution sources, some low-frequency ions, while not consistently present in all particulate matter, may only appear in specific pollution source categories, thus possessing strong source indicative significance. Current methods lack a detection process that combines overall spectral matching results with source-specific fingerprint ion matching results, making it difficult to establish an effective link between overall similarity assessment and local feature confirmation.

[0006] Furthermore, when detecting pollution sources for particulate matter to be identified, existing methods often only provide qualitative conclusions that a certain type of particulate matter is "similar" or "potentially related" to a certain type of pollution source, making it difficult to simultaneously output results such as candidate pollution source categories, degree of matching differences, fingerprint ion matching, and matching confidence levels. Therefore, when multiple pollution source categories have similar mass spectrometry characteristics, existing technologies struggle to objectively determine which pollution source category the particulate matter to be identified should belong to, and also find it difficult to identify whether it belongs to particulate matter of unknown origin, particulate matter of mixed origin, or particulate matter requiring verification.

[0007] Therefore, there is an urgent need for a particulate matter pollution source matching method based on single-particle mass spectra. This method should be able to preprocess the single-particle mass spectra of the particles to be identified to form a feature vector of the ions to be identified, and call a pre-established pollution source spectral library to calculate the matching divergence coefficient between the feature vector of the ions to be identified and the reference source spectra corresponding to each known pollution source category, so as to quantitatively determine the candidate pollution source category. At the same time, the method should be combined with the source-specific fingerprint ion information corresponding to the candidate pollution source category for matching confirmation, so as to output the pollution source matching result including the candidate pollution source category, the matching divergence coefficient, the fingerprint ion matching result and / or the matching confidence level, thereby improving the objectivity, accuracy and interpretability of pollution source detection at the single-particle level. Summary of the Invention

[0008] In view of the shortcomings of the existing technology, the purpose of this invention is to provide a method and apparatus for matching particulate matter pollution sources based on single particle mass spectrometry, so as to overcome the shortcomings of the existing technology.

[0009] The above-mentioned technical objective of the present invention is achieved through the following technical solution: Firstly, a method for matching particulate matter pollution sources based on single-particle mass spectrometry, comprising: Obtain a single-particle mass spectrum of the particulate matter to be identified, wherein the single-particle mass spectrum includes a positive ion mass spectrum and / or a negative ion mass spectrum of the particulate matter to be identified. The mass spectrum of the single particle is preprocessed to obtain the feature vector of the ion to be identified of the particle to be identified. The feature vector of the ion to be identified includes multiple mass-to-charge ratio positions and the relative intensities corresponding to each mass-to-charge ratio position. The system invokes a pre-established pollution source spectrum library, which includes reference source spectra corresponding to multiple known pollution source categories and source-specific fingerprint ion information associated with each reference source spectrum. Calculate the matching divergence coefficient between the feature vector of the ion to be identified and the reference source spectrum corresponding to each known pollution source category; Based on each matching divergence coefficient, candidate pollution source categories are determined from the plurality of known pollution source categories; The feature vector of the ion to be identified is matched with the source-specific fingerprint ion information corresponding to the candidate pollution source category to obtain the fingerprint ion matching result; Based on the matching divergence coefficients corresponding to the candidate pollution source categories and the fingerprint ion matching results, the pollution source matching results of the particulate matter to be identified are determined and output.

[0010] In one embodiment, the reference source spectrum includes: a weighted average source spectrum corresponding to a known pollution source category; The weighted average source spectrum includes: multiple mass-to-charge ratio locations and the reference relative intensity corresponding to each mass-to-charge ratio location; The calculation of the matching divergence coefficient between the feature vector of the ion to be identified and the reference source spectrum corresponding to each known pollution source category includes: At the same mass-to-charge ratio position, the matching divergence coefficient between the feature vector of the ion to be identified and the reference relative intensity in the reference source spectrum is calculated based on the relative intensity in the feature vector of the ion to be identified and the reference relative intensity in the reference source spectrum.

[0011] In one embodiment, calculating the matching divergence coefficient between the feature vector of the ion to be identified and the reference source spectrum based on the relative intensity in the feature vector of the ion to be identified and the reference relative intensity in the reference source spectrum includes: At multiple mass-to-charge ratio positions involved in the calculation, the relative intensity in the feature vector of the ion to be identified is compared with the reference relative intensity in the reference source spectrum. For each mass-to-charge ratio position, the ion peak splitting term at that mass-to-charge ratio position is determined by the ratio of the difference between the relative intensity and the reference relative intensity to the sum of the two. The matching divergence coefficient is obtained by performing mean square statistical processing on the divergence terms of ion peaks at multiple mass-to-charge ratio positions.

[0012] In one embodiment, determining candidate pollution source categories from the plurality of known pollution source categories based on each matching divergence coefficient includes: Determine the minimum matching divergence coefficient from multiple matching divergence coefficients; The known pollution source category corresponding to the minimum matching divergence coefficient is determined as the candidate pollution source category; Determine whether the minimum matching divergence coefficient is less than a preset matching threshold; If the minimum matching divergence coefficient is less than the preset matching threshold, then the candidate pollution source category is retained; If the minimum matching divergence coefficient is not less than the preset matching threshold, then the particulate matter to be identified is marked as particulate matter of unknown origin.

[0013] In one embodiment, the source-specific fingerprint ion information includes the mass-to-charge ratio position of the source-specific fingerprint ions; The step of matching the feature vector of the ion to be identified with the source-specific fingerprint ion information corresponding to the candidate pollution source category to obtain the fingerprint ion matching result includes: Determine whether there is an ion peak in the feature vector of the ion to be identified that has the same mass-to-charge ratio position as the source-specific fingerprint ion corresponding to the candidate pollution source category; If present, a fingerprint ion matching result is generated to support the candidate pollution source category; If it does not exist, a fingerprint ion matching result is generated to reduce the confidence of the candidate pollution source category.

[0014] In one embodiment, the source-specific fingerprint ion information further includes a reference relative intensity range of the source-specific fingerprint ions; After determining that there is an ion peak in the feature vector of the ion to be identified that has the same mass-to-charge ratio position as the source-specific fingerprint ion, it is further determined whether the relative intensity of the ion peak falls within the reference relative intensity range. When the relative intensity of the ion peak falls within the reference relative intensity range, a fingerprint ion matching result is generated to support the candidate pollution source category.

[0015] In one embodiment, the pollution source library is pre-established through the following steps: Obtain single-particle mass spectra of particulate matter from known sources corresponding to multiple known pollution source categories; The single-particle mass spectra of the known source particles are preprocessed to obtain the characteristic vectors of the known source ions; Clustering the feature vectors of ions from multiple known sources under the same known pollution source category yields multiple particle types; Calculate the average source spectrum for each particle type; Based on the proportion of each particle type in the corresponding known pollution source category, a weighted average source spectrum of multiple particle types is calculated to obtain the weighted average source spectrum corresponding to the known pollution source category. The weighted average source spectrum is stored in the pollution source spectrum library as a reference source spectrum corresponding to the known pollution source category.

[0016] In one embodiment, the pollution source library further establishes classification criteria among known pollution source categories through the following steps: Calculate the source spectrum divergence coefficient between the reference source spectra corresponding to different known pollution source categories; When the source spectrum divergence coefficient between two known pollution source categories is less than the first divergence coefficient threshold, the two known pollution source categories are marked as highly similar source categories; When the source spectrum divergence coefficient between two known pollution source categories is between the first divergence coefficient threshold and the second divergence coefficient threshold, the two known pollution source categories are marked as moderately similar source categories. When the source spectrum divergence coefficient between two known pollution source categories is greater than the second divergence coefficient threshold, the two known pollution source categories are marked as highly dissimilar source categories. Wherein, the first divergence coefficient threshold is less than the second divergence coefficient threshold.

[0017] In one embodiment, the pollution source library further establishes source-specific fingerprint ion information through the following steps: Calculate the frequency of occurrence of ion peaks corresponding to each mass-to-charge ratio position in each known pollution source category; When the frequency of the ion peak corresponding to the target mass-to-charge ratio position in the target known pollution source category is greater than the first frequency threshold, and the frequency of its occurrence in other known pollution source categories is less than the second frequency threshold, the ion peak corresponding to the target mass-to-charge ratio position is determined as the source-specific fingerprint ion of the target known pollution source category. The mass-to-charge ratio position of the source-specific fingerprint ions is associated with the reference source spectrum corresponding to the known pollution source category of the target and stored in the pollution source spectrum library.

[0018] Secondly, a particulate matter pollution source matching device based on single-particle mass spectrometry includes: The spectrum acquisition module is used to acquire the single-particle mass spectrum of the particulate matter to be identified, wherein the single-particle mass spectrum includes the positive ion mass spectrum and / or negative ion mass spectrum of the particulate matter to be identified. The preprocessing module is used to preprocess the mass spectrum of the single particle to obtain the feature vector of the ion to be identified of the particle to be identified. The feature vector of the ion to be identified includes multiple mass-to-charge ratio positions and the relative intensities corresponding to each mass-to-charge ratio position. The spectral library calling module is used to call a pre-established pollution source spectral library, which includes reference source spectra corresponding to multiple known pollution source categories and source-specific fingerprint ion information associated with each reference source spectrum; The divergence coefficient calculation module is used to calculate the matching divergence coefficient between the feature vector of the ion to be identified and the reference source spectrum corresponding to each known pollution source category; The candidate pollution source category determination module is used to determine candidate pollution source categories from the plurality of known pollution source categories based on each matching divergence coefficient; The fingerprint ion matching module is used to match the feature vector of the ion to be identified with the source-specific fingerprint ion information corresponding to the candidate pollution source category to obtain the fingerprint ion matching result. The pollution source matching result output module is used to determine and output the pollution source matching result of the particulate matter to be identified based on the matching divergence coefficient corresponding to the candidate pollution source category and the fingerprint ion matching result.

[0019] Thirdly, a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in the first aspect.

[0020] Fourthly, a computer device includes a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of the method described in the first aspect.

[0021] In summary, the present invention has the following beneficial effects: by matching the divergence coefficient to quantify the overall difference between the particulate matter to be identified and the reference source spectrum of each pollution source, and by combining the source-specific fingerprint ions for confirmation, the present invention can reduce the error of human experience judgment and improve the objectivity, accuracy and interpretability of particulate matter pollution source matching. Attached Figure Description

[0022] Figure 1 This is a flowchart of the particulate matter pollution source matching method based on single particle mass spectrometry of the present invention; Figure 2 This is a structural diagram of the particulate matter pollution source matching device based on single-particle mass spectrometry in an embodiment of the present invention; Figure 3 This is an internal structural diagram of the computer device in an embodiment of the present invention; Figure 4 This refers to the change in the number of retained ions in the detectable ions under different minimum occurrence frequency thresholds in Embodiment 5 of the present invention. Figure 5 This is the variation trend of the average divergence coefficient CD between different industrial source spectra under different minFO thresholds in Embodiment 5 of the present invention; Figure 6 This is the CD matrix among 14 industrial source spectrum objects calculated based on full spectrum data in Embodiment 5 of the present invention.

[0023] The diagram shows: 1. Spectrum acquisition module; 2. Preprocessing module; 3. Spectrum library retrieval module; 4. Divergence coefficient calculation module; 5. Candidate pollution source category determination module; 6. Fingerprint ion matching module; 7. Pollution source matching result output module. Detailed Implementation

[0024] To make the objectives, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Several embodiments of the present invention are shown in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein.

[0025] Example 1 To address the aforementioned problems, this invention provides a particulate matter pollution source matching method based on single-particle mass spectrometry. Before identifying particulate matter, it is first necessary to classify each pollution source and then establish a pollution source spectral library. This pollution source spectral library serves as a reference for subsequent pollution source matching of the particulate matter to be identified.

[0026] The pollution source library includes at least reference source spectra corresponding to multiple known pollution source categories, source-specific fingerprint ion information for each reference source spectrum, source spectrum divergence coefficient information between different known pollution source categories, and source category merging or differentiation rules established based on the source spectrum divergence coefficients. The known pollution source categories may include coal combustion sources, vehicle exhaust sources, biomass combustion sources, dust sources, industrial furnace sources, iron and steel smelting sources, cement production sources, and waste incineration sources, etc. Particulate matter samples from different known pollution source categories are detected using a single-particle aerosol mass spectrometer to obtain corresponding single-particle mass spectrometry data, and a pollution source library is established based on this single-particle mass spectrometry data.

[0027] Constructing a pollution source library first requires collecting particulate matter samples from multiple known pollution source categories. During sampling, pre-vacuum-sealed glass bottles are used as sampling containers, and a diffusion drying tube is placed at the front end of the sampling pipeline to reduce the relative humidity in the flue gas or aerosol samples. For example, the relative humidity of the samples is controlled below 50% to reduce the impact of moisture on particulate matter transport, ionization, and mass spectrometry signal stability. After sampling, the samples are stored at 4°C in the dark, and single-particle aerosol mass spectrometry analysis is performed within a preset time, such as 4 hours, to reduce adsorption loss of particulate matter on the container walls and chemical changes in the samples during storage. In this step, by controlling the sampling humidity, storage temperature, and analysis time, the original chemical composition of the known pollution source particulate matter can be maintained as much as possible, making the subsequently established pollution source library closer to the actual emission characteristics and reducing non-source-specific errors introduced by the sampling and storage process.

[0028] Then, particulate matter samples corresponding to each known pollution source category were introduced into a single-particle aerosol mass spectrometer. The single-particle aerosol mass spectrometer focuses the particulate matter using an aerodynamic lens, measures the particle velocity and calculates the particle size using a dual-laser system, desorbs and ionizes the particulate matter using an ultraviolet pulsed laser, and finally acquires the positive and negative ion mass spectra of individual particles using a bipolar time-of-flight mass spectrometry module. In this embodiment, the single-particle aerosol mass spectrometer uses a 266nm Nd:YAG laser for desorption and ionization, and the acquired mass-to-charge ratio range is m / z -250 to m / z +250. For any known pollution source category... The collected first A single-particle mass spectrum of a known source of particulate matter can be represented as: ;in, Indicates the first Under the known pollution source category, the first Single-particle mass spectra of the particles. Indicates the first Mass-to-charge ratio position This indicates the particulate matter's position relative to its mass-to-charge ratio. The intensity of the original ion peak at that location. This indicates the number of mass-to-charge ratio locations involved in the analysis. This embodiment utilizes single-particle aerosol mass spectrometry to simultaneously obtain particle size and chemical composition information at the single-particle level, avoiding the problem of mixed signals from multiple particles in traditional offline filter membrane sampling. This provides a data foundation for establishing a pollution source spectral library with particle type resolution capabilities.

[0029] After obtaining the single-particle mass spectrum of particulate matter from a known source, the positive ion mass spectrum and the negative ion mass spectrum are preprocessed respectively. The preprocessing may include mass-to-charge ratio alignment, noise peak removal, invalid spectrum removal, and intensity normalization. Intensity normalization is performed as follows: for the same particulate matter's positive or negative ion mass spectrum, the highest intensity ion peak in the spectrum is taken as the reference peak, and its intensity is set to 100%. Other ion peaks are converted according to the ratio to the reference peak. The normalization formula can be expressed as: ; in, Indicates the first Among the known pollution source categories, the first The particle at the mass-to-charge ratio position Normalized intensity at that point This indicates the intensity of the original ion peak at that mass-to-charge ratio position. This represents the maximum ion peak intensity in the corresponding positive or negative ion mass spectrum of the particulate matter. After normalization, each particulate matter from a known source can form a characteristic vector of ions from a known source: ;in, , indicating the first Under the known pollution source category, the first The first particulate matter in the... The relative intensity at each mass-to-charge ratio position. Specifically, this embodiment reduces the influence of factors such as the absolute signal strength of particulate matter, laser energy fluctuations, and changes in sample concentration on spectral comparison through normalization, so that subsequent clustering, source spectrum construction, and divergence coefficient calculation mainly reflect the relative chemical composition differences between different particulate matter.

[0030] After normalization, for multiple known source ion feature vectors under the same known pollution source category, the Adaptive Resonance Theory Neural Network (ART-2a) is used for clustering to classify particulate matter with similar mass spectrometry features into the same particle type. In a preferred embodiment, the ART-2a clustering parameters can be set as follows: a warning value of 0.65, a learning rate of 0.05, and 20 iterations. After clustering, the number of particles corresponding to each particle type is counted, and particle types with a particle count ratio of less than 1% are removed and regarded as accidental particle types or noise classes. Specifically, particulate matter emitted from the same pollution source is usually not a single type, but contains multiple particle groups with different chemical compositions. Through ART-2a clustering, the diversity of particulate matter composition within the pollution source can be preserved, so that the pollution source spectral library can not only reflect the overall average characteristics of a pollution source, but also reflect the compositional structure of different particle types within it.

[0031] For each retained particle type, the average normalized intensity of all particles within that particle type at each mass-to-charge ratio position is calculated to obtain the average source spectrum of the particle type. Among the known pollution source categories, the first The average source spectrum of each particle type can be expressed as: Among them, the first The average normalized intensity at each mass-to-charge ratio position is: ;in, Indicates the first Among the known pollution source categories, the first The particle type in the first The average relative intensity at each mass-to-charge ratio position This represents a collection of particulate matter belonging to this particle type. Indicates the first particle of this particle type The first particulate matter in the... Normalized intensity at each mass-to-charge ratio position Indicates the first Of the known pollution source categories, the first The number of particles contained in a particle type, i.e., the set The number of particulate matter. Specifically, this embodiment reduces the impact of random fluctuations in individual particulate matter by calculating the average source spectrum of particle type, highlighting the stable mass spectrometry characteristics of the same particle type, and providing a reference for subsequent construction of weighted average source spectra and fine classification matching of the particulate matter to be identified.

[0032] After obtaining the average source spectrum for each particulate type, a weighted average is calculated based on the proportion of each particulate type in the corresponding known pollution source category to obtain the weighted average source spectrum corresponding to that known pollution source category. The weighted average source spectrum of known pollution source categories can be expressed as: Expanding to each mass-to-charge ratio position, it can be expressed as: ;in, Indicates the first Weighted average source spectrum corresponding to a known pollution source category Indicates the first The total number of particulate types retained in each known pollution source category. Indicates the first Among the known pollution source categories, the first The percentage of particles of each particle type, i.e. , Indicates the first Average source spectrum of particle type for each particle type Indicates the first The known pollution source categories in the first Reference relative strength at each mass-to-charge ratio location No. Among the known pollution source categories, the first The particle type in the first The average relative intensity at each mass-to-charge ratio position. Specifically, the weighted average source spectrum calculated in this embodiment is not simply an average of all particulate matter, but rather retains the particle type structure first, and then weights it according to the proportion of particle type. Therefore, it can take into account the contributions of the main and secondary particle types within the pollution source, making the obtained reference source spectrum more representative of the overall emission characteristics of the pollution source.

[0033] To further establish source-specific fingerprint ion information, the frequency of occurrence of the ion peak corresponding to each mass-to-charge ratio position in each known pollution source category was calculated.

[0034] For the The known pollution source categories, the first The frequency of occurrence of ion peaks corresponding to each mass-to-charge ratio position can be expressed as: ; in, Indicates the first The ion peak at the mass-to-charge ratio position is at the [missing information]. Frequency of occurrence among known pollution source categories Indicates the first Among the known pollution source categories, the first... The number of particles in the ion peak at each mass-to-charge ratio position. Indicates the first Total effective particle count for each known pollution source category.

[0035] To illustrate the frequency calculation process described above, let's take an example. Assume there are three known pollution sources: A: a cement plant; B: a coal-fired boiler; and C: a steel plant. We measure 1000 particles from each pollution source and count several... m / z The presence of ions can be obtained from Table 1 below.

[0036] Table 1: m / z The presence of ions

[0037] Furthermore, for each mass-to-charge ratio location, its minimum frequency of occurrence among all known pollution source categories can be calculated: ; in, This represents the total number of known pollution source categories. Indicates the first The mass-to-charge ratio position corresponds to the minimum frequency of occurrence of an ion peak across all known pollution source categories. Specifically, the frequency of occurrence reflects the stability of a particular ion peak within a specific pollution source category. This embodiment, by calculating the frequency of occurrence in different pollution source categories, can distinguish between ubiquitous common ions and specific ions that appear only in specific pollution sources, providing a quantitative basis for subsequent screening of source-specific fingerprint ions.

[0038] For example, This refers to the minimum frequency of a particular ion among all pollution sources. m / z Taking 40 as an example, cement plants account for 85%; coal-fired boilers for 78%; and steel plants for 72%. Therefore, we can conclude: This shows m / z 40 is a relatively stable ion found in all pollution sources. Similarly, let's look at... m / z 56: Cement plants 5%; coal-fired boilers 8%; steel plants 68%. Therefore: This shows m / z 56 does not consistently appear in all pollution sources, but is mainly concentrated in steel plants. It may not be a common ion, but it could be a characteristic ion of steel plants. In summary, we can conclude that... It refers to the quantity of a certain ion present in a certain pollution source; It depends on whether this ion appears stably in all pollution sources.

[0039] To assess the contribution of ions in different frequency ranges to the differences in the source spectrum, we can... Multiple occurrence frequency thresholds are set, and ion sets corresponding to different thresholds are formed. For example, ;in, It can represent the entire spectrum of ions. For any threshold The corresponding set of non-full-spectrum ions can be represented as: ;in, This indicates that the minimum occurrence frequency is greater than the threshold. The set of mass-to-charge ratio positions, when using a full-spectrum ion set, is defined as follows: For the set of all detected effective mass-to-charge ratio locations, unaffected by The limitations are as follows. Specifically, this embodiment uses multi-level threshold settings to examine the effects of high-frequency ions, mid-frequency ions, low-frequency ions, and full-spectrum ions on source spectrum differences, thereby avoiding the loss of low-frequency ion information that has source identification value when constructing source spectra using only high-frequency ions.

[0040] Specifically, the purpose of this step is to separate the ions into layers according to their level of stability. Only retain ions that are stable in all sources; Retaining relatively stable ions; : Retains moderately stable ions; : Retains ions with lower frequencies but still a certain frequency of occurrence; Full spectrum: All ions are included, including low-frequency ions, source-specific ions and auxiliary marker ions.

[0041] Taking the contents of Table 1 as an example, At that time, only m / z 12. m / z 40. These are ions commonly found in various pollution sources; they are relatively stable, but may not have a strong ability to distinguish between them. At that time, it may be added again. m / z 50. At this point, we can see whether the difference in source spectrum between mid-frequency ions is helpful. For the full spectrum, [something] will be added. m / z 56. m / z 97. m / z Low-frequency or source-specific ions, such as 115, are also present. If the differences between different pollution sources become significantly larger, it indicates that these low-frequency ions are likely the key to distinguishing the pollution sources.

[0042] When establishing a pollution source spectrum library, the source spectrum divergence coefficient between different known pollution source categories can be calculated to assess the overall degree of difference between different reference source spectra. For the first... The known pollution source categories and the A known category of pollution source, in the ion set The source spectrum divergence coefficient can be expressed as: ; in, Indicates the frequency threshold Under the corresponding ion set, the first The known pollution source categories and the Source spectrum divergence coefficient among known pollution source categories; Indicates the first The known pollution source categories in the first Reference relative strength at each mass-to-charge ratio location; Indicates the first The known pollution source categories in the first Reference relative strength at each mass-to-charge ratio location; This indicates the number of locations involved in the mass-to-charge ratio calculation. When When this happens, the bifurcation term corresponding to the mass-to-charge ratio position can be set to 0, or a preset non-zero correction term can be added to the denominator to avoid the zero denominator problem.

[0043] Furthermore, for the same threshold It can calculate the mean and standard deviation of the source spectrum divergence coefficients between all known pollution source categories: ; ; in, Indicates threshold The average of the source spectrum divergence coefficients among all known pollution source categories. This represents the corresponding standard deviation.

[0044] Specifically, the divergence coefficient (CD) represents the degree of difference between the source spectra of two pollution sources using a value between 0 and 1. The closer CD is to 0, the more similar the two pollution sources are; the closer CD is to 1, the greater the difference between the two pollution sources.

[0045] After obtaining the source spectrum divergence coefficients at multiple occurrence frequency thresholds, the differences between the different thresholds can be compared. This is to determine whether the source spectrum differences between different known pollution source categories significantly increase with the gradual inclusion of low-frequency ions. For example, it can be calculated from... The increment of the source spectrum divergence coefficient when the full spectrum of ions is reached: ;in, This represents the average source spectrum divergence coefficient under the full-spectrum ion set. express Average source spectrum divergence coefficient at the threshold.

[0046] Furthermore, the contribution of low-frequency ions to the source spectrum differences can be calculated: ; in, express The average source spectrum divergence coefficient below the threshold. It should be noted that when using the above contribution formula, the denominator represents the total source spectrum difference increment from the high-frequency ion set to the full-spectrum ion set, and the numerator represents the difference from... The contribution of low-frequency ions to the increase in the source spectrum ion set. Therefore, this formula can characterize the proportion of low-frequency ions in the increase in source spectrum difference.

[0047] Specifically, following the previous example, let's assume we calculate the CD values ​​for cement plants, coal-fired boilers, and steel plants separately: the CD value for cement plants and coal-fired boilers is 0.75, indicating a significant difference between them; the CD value for cement plants and steel plants is 0.82, indicating a very large difference between them; and the CD value for coal-fired boilers and steel plants is 0.48, indicating a moderate difference between them. This allows us to objectively determine that cement plants and steel plants have the greatest difference; coal-fired boilers and steel plants have some similarities; and if a later unknown particulate matter is closer to the source spectrum of a steel plant, we can provide a matching basis based on the CD values.

[0048] Furthermore, since the ions have been divided into different sets in the aforementioned steps, the CD between pollution sources needs to be calculated separately using these sets when calculating the divergence coefficient. Assuming we need to calculate the divergence coefficient between a cement plant and a steel plant, we need to calculate the divergence coefficient within each threshold range according to the previous level thresholds, resulting in Table 2.

[0049] Table 2: Disagreement coefficients between cement plants and steel plants at various levels

[0050] As shown in Table 2, if only high-frequency ions are considered, cement plants and steel plants may appear similar; however, once low-frequency and specific ions are also included, the differences become significant. This indicates that low-frequency ions are not useless noise, but play a crucial role in distinguishing pollution sources.

[0051] To establish source-specific fingerprint ion information corresponding to each known pollution source category, candidate characteristic ions can be screened first. For the first... The first of the known pollution source categories The average relative intensity of an ion can be calculated from the mass-to-charge ratio position: ; in, Indicates the first The ion peak at the mass-to-charge ratio position is at the [missing information]. The average relative intensity among known pollution source categories Indicates the first A collection of particulate matter containing this ion peak from a known pollution source category. Indicates the first The first particulate matter in the... Normalized intensity at each mass-to-charge ratio position.

[0052] In a preferred embodiment, when a certain ion peak satisfies and When this is the case, it can be identified as a candidate characteristic ion.

[0053] Specifically, the screening of candidate characteristic ions considers both the frequency of occurrence and the average relative intensity, which avoids misjudging occasional weak ions as characteristic ions and avoids ignoring their occurrence stability based solely on intensity.

[0054] After obtaining candidate characteristic ions, the frequency differences of these candidate characteristic ions among different known pollution source categories are further compared. When the ion peak corresponding to a certain mass-to-charge ratio position meets the following condition, it is identified as the source-specific fingerprint ion of the target known pollution source category: And for any other known pollution source category ,satisfy: ,in, This indicates the frequency of the ion peak within the target known pollution source category. This indicates the frequency of the ion peak in other known pollution source categories. In this embodiment, 33% is used as the first frequency threshold and 10% is used as the second frequency threshold.

[0055] For ion peaks that occur less than 33% of the time but are consistently present only in known target pollution source categories and have a clear chemical classification, they can also be stored as auxiliary marker ions in the pollution source spectral library. For example, such ions may not be used as primary fingerprint ions, but can be used to increase or decrease the matching confidence of candidate pollution source categories during subsequent particulate matter matching.

[0056] Based on the source spectrum divergence coefficient under the full-spectrum ion set, classification rules between known pollution source categories can be established. Specifically, for any two known pollution source categories... and Based on its full-spectrum source spectrum divergence coefficient Categorize: when When two known pollution source categories are determined to be highly similar, they can be merged in the receptor model or subsequent matching process. when When two known pollution sources are classified as moderately similar, further auxiliary judgment is needed when merging or distinguishing them, based on source-specific fingerprint ion information. when When two known pollution source categories are determined to be highly dissimilar, they should be treated as independent source categories.

[0057] After the above steps, the weighted average source spectrum, particle type average source spectrum, source-specific fingerprint ion information, auxiliary marker ion information, source spectrum divergence coefficient matrix, and source spectrum classification rules corresponding to each known pollution source category are associated and stored to form a pollution source spectrum library.

[0058] After establishing the pollution source spectrum library, this embodiment further provides a particulate pollution source matching method based on single-particle mass spectra, used to match the source of the particulate matter to be identified in the aerosol sample to be tested. The pollution source spectrum library pre-stores reference source spectra corresponding to multiple known pollution source categories, as well as source-specific fingerprint ion information associated with each reference source spectrum. Therefore, after obtaining the single-particle mass spectrum of the particulate matter to be identified, the mass spectrometric characteristics of the particulate matter to be identified can be compared with the reference source spectra in the pollution source spectrum library to determine the candidate pollution source category that is closer to the particulate matter to be identified, and the source-specific fingerprint ion information is further used to confirm the candidate pollution source category.

[0059] In this embodiment, the known pollution source categories may include coal combustion sources, vehicle exhaust sources, biomass combustion sources, dust sources, industrial furnaces and kilns, iron and steel smelting sources, cement production sources, and waste incineration sources. The reference source spectrum is preferably a weighted average source spectrum obtained during the aforementioned spectral library establishment process. This weighted average source spectrum includes multiple mass-to-charge ratio positions and corresponding reference relative intensities for each mass-to-charge ratio position. The source-specific fingerprint ion information includes the mass-to-charge ratio positions of the source-specific fingerprint ions, and may further include the reference relative intensity range, frequency of occurrence, average relative intensity, corresponding pollution source category, and auxiliary marker ion information for the source-specific fingerprint ions.

[0060] S1. First, the aerosol sample to be tested is introduced into the single-particle aerosol mass spectrometer. The aerosol sample to be tested can be derived from ambient air receptor samples, industrial emission samples, motor vehicle exhaust samples, combustion emission samples, or other particulate matter samples that require pollution source identification.

[0061] A single-particle aerosol mass spectrometer detects particulate matter in an aerosol sample individually. Specifically, after entering the instrument, aerosol particles are first focused by an aerodynamic lens to form a concentrated particle beam. Then, the particles pass through a dual-laser velocimetry region, where the instrument calculates the particle velocity based on the time difference between the two velocimetry laser beams and determines the aerodynamic particle size based on the velocity. Next, the particles enter a desorption / ionization region, where individual particles are desorbed and ionized by a UV pulsed laser, causing the chemical components in the particles to form positive and / or negative ions. Finally, a bipolar time-of-flight mass spectrometry module acquires the positive and negative ion mass spectrometry signals respectively to obtain the single-particle mass spectrum of the particle to be identified. In a preferred embodiment, the single-particle aerosol mass spectrometer uses a 266nm Nd:YAG laser to desorb and ionize the particles, and the acquired mass-to-charge ratio range is... m / z -250 to m / z +250. For a particulate matter to be identified. Its single-particle mass spectrum can be represented as: ; in, Indicates particulate matter to be identified Single-particle mass spectrum, Indicates the first Mass-to-charge ratio position Indicates particulate matter to be identified At the mass-to-charge ratio position The intensity of the original ion peak at that location. This indicates the number of mass-to-charge ratio locations involved in the analysis.

[0062] S2. After obtaining the single-particle mass spectrum of the particulate matter to be identified, the single-particle mass spectrum is preprocessed to obtain the feature vector of the ion to be identified. The preprocessing includes one or more of the following: mass-to-charge ratio alignment, noise peak removal, invalid spectrum removal, and intensity normalization. Specifically, mass-to-charge ratio alignment refers to unifying the single-particle mass spectrum of the particulate matter to be identified with the reference source spectrum in the pollution source spectral library to the same set of mass-to-charge ratio positions. For example, for the mass-to-charge ratio position set already established in the pollution source spectral library. The ion peaks of the particles to be identified are mapped to positions with the same mass-to-charge ratio. If no effective ion peak is detected at a certain mass-to-charge ratio position, the relative intensity corresponding to that mass-to-charge ratio position can be set to 0; if multiple adjacent peaks or peak position shifts occur, peak positions can be merged according to a preset mass tolerance range.

[0063] Noise peak removal refers to removing ion peaks with low signal-to-noise ratios, peak intensities below preset noise thresholds, or those that do not meet the criteria for valid peaks. Invalid spectrum removal refers to marking particulate matter as invalid when the number of valid ion peaks, total ion intensity, or abnormal mass spectrometry signals in the positive and negative ion mass spectra of the particulate matter to be identified are too few, or the total ion intensity is too low, or the mass spectrometry signal is abnormal. This invalid particulate matter will not participate in subsequent pollution source matching.

[0064] Intensity normalization refers to using the maximum ion peak intensity in the same positive ion mass spectrum or the same negative ion mass spectrum as the reference intensity, and converting the intensities of other ion peaks into relative intensities relative to the reference intensity. Normalization can be expressed as: ,in, Indicates particulate matter to be identified At the mass-to-charge ratio position Normalized intensity at that point This indicates the intensity of the original ion peak at that mass-to-charge ratio position. Indicates particulate matter to be identified The maximum ion peak intensity corresponding to the positive or negative ion mass spectrum. After preprocessing, the particulate matter to be identified... Forming the feature vector of the ion to be identified: ,in, , indicating the particulate matter to be identified In the The relative intensity at a mass-to-charge ratio location.

[0065] By aligning the mass-to-charge ratio, we can ensure that the feature vector of the ion to be identified is compared with the reference source spectrum in the pollution source spectrum library at the same mass-to-charge ratio position. By eliminating noise peaks, we can reduce the influence of random noise on the matching divergence coefficient and fingerprint ion matching results. By normalizing the intensity, we can reduce the interference of the absolute signal strength of particulate matter, laser energy fluctuations and changes in injection concentration on subsequent comparisons, so that the matching results mainly reflect the relative chemical composition differences between the particulate matter to be identified and the reference source spectrum.

[0066] S3. After obtaining the feature vector of the ion to be identified, a pre-established pollution source spectral library is invoked. The pollution source spectral library can be obtained using the spectral library establishment method described in the preceding embodiments, and includes at least multiple reference source spectra corresponding to known pollution source categories, as well as source-specific fingerprint ion information associated with each reference source spectrum. In one specific embodiment, the first... The reference source spectrum corresponding to each known pollution source category is a weighted average source spectrum, which can be expressed as: ;in, Indicates the first Reference source spectra corresponding to each known pollution source category Indicates the first The known pollution source categories in the first The reference relative intensity at each mass-to-charge ratio position, where n represents the number of mass-to-charge ratio positions participating in the comparison.

[0067] The pollution source library can also store the first... The source-specific fingerprint ion set corresponding to each known pollution source category: ;in, Indicates the first A set of source-specific fingerprint ion information corresponding to a known pollution source category. This indicates the mass-to-charge ratio position of source-specific fingerprint ions. and These represent the source-specific fingerprint ions at the [number]th [time]. Reference relative intensity range for a known pollution source category This indicates the frequency of occurrence of source-specific fingerprint ions. This indicates the average relative intensity of source-specific fingerprint ions. Indicates the first The number of source-specific fingerprint ions corresponding to each known pollution source category.

[0068] Among them, the reference source spectrum Used for overall spectral difference calculation, source-specific fingerprint ion set Used to confirm the local characteristics of candidate pollution source categories. Specifically, the reference source spectrum is mainly used to determine the similarity of the overall ion distribution, while the source-specific fingerprint ion information is mainly used to determine whether there are markers unique to that pollution source.

[0069] S4. After calling the pollution source spectral library, the feature vector of the ion to be identified is compared with the reference source spectrum corresponding to each known pollution source category, and the matching divergence coefficient is calculated. The matching divergence coefficient is used to characterize the degree of overall spectral difference between the single particle mass spectrum of the particulate matter to be identified and the corresponding reference source spectrum at multiple mass-to-charge ratio positions.

[0070] Specifically, at positions with the same mass-to-charge ratio, the relative intensity in the feature vector of the ion to be identified is obtained. and the Reference relative intensity in the reference source spectrum corresponding to each known pollution source category For each mass-to-charge ratio position, the ion peak divergence term between the particle to be identified and the reference source spectrum at that mass-to-charge ratio position is calculated. The ion peak divergence term can be determined based on the intensity difference between the relative intensity and the reference relative intensity, as well as the sum of their intensities.

[0071] In a preferred embodiment, the particulate matter to be identified With the The matching divergence coefficient between the reference source spectra corresponding to each known pollution source category can be expressed as: ;in, Indicates particulate matter to be identified With the The matching divergence coefficient between the reference source spectrum corresponding to each known pollution source category Indicates particulate matter to be identified In the The relative strength at each mass-to-charge ratio position Indicates the first The reference source spectrum corresponding to the known pollution source categories is in the [number]th ...]. Reference relative strength at each mass-to-charge ratio location This indicates the number of mass-to-charge ratio locations involved in the calculation.

[0072] when When this happens, the ion peak splitting term corresponding to the mass-to-charge ratio position can be set to 0, or a preset non-zero correction term can be added to the denominator to avoid the zero denominator problem.

[0073] In another implementation, the mass-to-charge ratio (MTR) positions used in the calculation may not be all MTR positions, but rather selected based on a pre-defined ion set in the pollution source spectral library. For example, a full-spectrum ion set, a set of characteristic ions related to candidate pollution source categories, a set of ions with a frequency greater than a preset threshold, or a valid ion set excluding noise ions can be selected. Regardless of the ion set used, it should be ensured that the feature vector of the ion to be identified and the reference source spectrum are calculated correspondingly at the same MTR position.

[0074] Specifically, by matching the divergence coefficient, the overall spectral differences between the particulate matter to be identified and multiple known pollution source categories can be converted into comparable numerical values. The smaller the matching divergence coefficient, the more similar the particulate matter to be identified is to the reference source spectrum corresponding to that known pollution source category; the larger the matching divergence coefficient, the more significant the difference in their overall chemical composition.

[0075] S5. Obtain the particulate matter to be identified After determining the matching divergence coefficients between the reference source spectra and the known pollution source categories, the minimum matching divergence coefficient is determined from the multiple matching divergence coefficients: ,in, This represents the total number of known pollution source categories in the pollution source library. Indicates particulate matter to be identified The corresponding minimum matching divergence coefficient. The known pollution source category corresponding to the minimum matching divergence coefficient is determined as the candidate pollution source category: ;in, Indicates particulate matter to be identified The corresponding candidate pollution source categories.

[0076] Then, it is determined whether the minimum matching divergence coefficient is less than a preset matching threshold. If the minimum matching divergence coefficient is less than the preset matching threshold, it is considered that the particulate matter to be identified has sufficient overall spectral similarity with the candidate pollution source category, and the candidate pollution source category is retained; if the minimum matching divergence coefficient is not less than the preset matching threshold, it indicates that the particulate matter to be identified has significant differences from the existing reference source spectra in the pollution source spectral library, and the particulate matter to be identified can be marked as particulate matter of unknown origin.

[0077] The preset matching threshold can be determined based on the matching distribution of homologous particles obtained during the pollution source spectral library establishment phase, the source spectrum divergence coefficient distribution among known pollution source categories, or the results of manual calibration. For example, the mean of the matching divergence coefficients between known source particles and their reference source spectra within the same known pollution source category can be added to a preset multiple standard deviation as the preset matching threshold; alternatively, the first divergence coefficient threshold or the second divergence coefficient threshold can be set as a reference threshold according to the spectral library classification rules.

[0078] In a further embodiment, if the discrepancy coefficients between the particulate matter to be identified and two or more known pollution source categories are all small, and the difference between different discrepancy coefficients is less than a preset difference threshold, then the particulate matter to be identified can be marked as suspected mixed source particulate matter or particulate matter to be verified. For example, if the discrepancy coefficient between the particulate matter to be identified and a coal-fired source is 0.31, and the discrepancy coefficient between it and an industrial furnace source is 0.34, and the difference between the two is small, then two candidate pollution source categories can be output simultaneously, and further confirmation can be made subsequently by combining source-specific fingerprint ion information.

[0079] Specifically, this embodiment can achieve objective ranking of multiple known pollution source categories by using the minimum matching divergence coefficient; by setting a preset matching threshold, it can avoid forcibly classifying particulate matter that is dissimilar to all reference source spectra in the spectral library into a certain known pollution source category, thereby improving the reliability of pollution source matching results.

[0080] S6. After determining the candidate pollution source category, the feature vector of the ion to be identified is further matched with the source-specific fingerprint ion information corresponding to the candidate pollution source category to obtain the fingerprint ion matching result.

[0081] Specifically, assuming the candidate pollution source category is number 1 For each known pollution source category, the corresponding source-specific fingerprint ion set is as follows: ;in, This represents the source-specific fingerprint ion set corresponding to the candidate pollution source category. Indicates the first The mass-to-charge ratio position of source-specific fingerprint ions and These represent the reference relative intensity range of the source-specific fingerprint ion in the candidate pollution source category. In this embodiment, the reference relative intensity range can be determined based on the normalized intensity distribution of all peak-containing particles of the source-specific fingerprint ion in the target known pollution source category. For example, a preset quantile interval of the normalized intensity distribution can be taken, or a range formed by the upper and lower preset proportions of the average relative intensity can be taken. This represents the number of source-specific fingerprint ions corresponding to the candidate pollution source category. For any given source-specific fingerprint ion... Determine the feature vector of the ion to be identified. Does the sample contain an ion peak with the same mass-to-charge ratio position as the fingerprint ion from that source? If the feature vector of the ion to be identified contains an ion peak with a mass-to-charge ratio position... If the relative intensity of the ion peak is greater than the preset effective peak intensity threshold, then the particulate matter to be identified is considered to have hit the source-specific fingerprint ion.

[0082] When the source-specific fingerprint ion information also includes a reference relative intensity range, after determining that there are ion peaks at positions with the same mass-to-charge ratio, it is further determined that the particulate matter to be identified is present in the source-specific fingerprint ion... Relative strength at the position corresponding to the mass-to-charge ratio Does it fall within the reference relative intensity range? If the above conditions are met, a fingerprint ion matching result is generated to support the candidate pollution source category; if the source-specific fingerprint ion is not present in the particulate matter to be identified, or the relative intensity of the ion peak does not fall within the reference relative intensity range, a fingerprint ion matching result is generated to reduce the confidence of the candidate pollution source category.

[0083] Specifically, the matching divergence coefficient can reflect the overall spectral similarity, but the spectra of individual particles may exhibit random fluctuations. Therefore, relying solely on the overall matching results may not be sufficient to determine the source. By further matching source-specific fingerprint ions, it can be confirmed whether the particles to be identified possess characteristic ion markers of candidate pollution source categories, thereby improving the interpretability and reliability of the pollution source matching results.

[0084] S7. After obtaining the matching divergence coefficients and fingerprint ion matching results corresponding to the candidate pollution source categories, the pollution source matching results of the particulate matter to be identified are determined by combining the two.

[0085] In one specific implementation, the pollution source matching result includes at least the candidate pollution source category, the matching divergence coefficient corresponding to the candidate pollution source category, and the fingerprint ion matching result. For example, the pollution source matching result may include: the particulate matter number to be identified, the aerodynamic particle size of the particulate matter, the candidate pollution source category, the matching divergence coefficient corresponding to the candidate pollution source category, whether the source-specific fingerprint ion is matched, the mass-to-charge ratio position of the matched source-specific fingerprint ion, the fingerprint ion matching rate or the weighted fingerprint ion matching score, and the pollution source matching conclusion, etc.

[0086] In a preferred embodiment, the matching confidence level can be determined based on the matching divergence coefficient and fingerprint ion matching result corresponding to the candidate pollution source category. For example, when the matching divergence coefficient corresponding to the candidate pollution source category is less than a preset matching threshold, and the particulate matter to be identified matches one or more source-specific fingerprint ions corresponding to the candidate pollution source category, the candidate pollution source category is determined as the pollution source matching result of the particulate matter to be identified and given a high confidence level; when the matching divergence coefficient corresponding to the candidate pollution source category is less than the preset matching threshold, but does not match the source-specific fingerprint ions corresponding to the candidate pollution source category, the candidate pollution source category can be retained, but its matching confidence level can be reduced, or the particulate matter to be identified can be marked as particulate matter to be verified; when the matching divergence coefficient corresponding to the candidate pollution source category is not less than the preset matching threshold, the particulate matter to be identified is marked as particulate matter of unknown origin.

[0087] Furthermore, this embodiment can output pollution source matching results according to the following rules: when If the fingerprint ion matching result is less than the preset matching threshold and supports the candidate pollution source category, output: "The particulate matter to be identified has been matched to the candidate pollution source category". when If the fingerprint ion matching result is less than the preset matching threshold but does not support the candidate pollution source category, the output will be: "The particulate matter to be identified is suspected to be matched to the candidate pollution source category, and further confirmation is required". when If the value is not less than the preset matching threshold, output: "The particulate matter to be identified is particulate matter from an unknown source"; When the matching divergence coefficients of multiple known pollution source categories all meet the matching conditions, and the particulate matter to be identified simultaneously hits the source-specific fingerprint ions corresponding to multiple known pollution source categories, the output "the particulate matter to be identified is a suspected mixed source particulate matter" is output.

[0088] This embodiment combines the matching divergence coefficient at the overall spectral level with the source-specific fingerprint ion matching results at the local feature level, which can avoid relying on a single indicator to determine the source and improve the accuracy and interpretability of the pollution source matching results.

[0089] Example 2 Please see Figure 2 A particulate matter pollution source matching device based on single-particle mass spectrometry, the particulate matter pollution source matching device based on single-particle mass spectrometry includes: The spectrum acquisition module 1 is used to acquire the single-particle mass spectrum of the particulate matter to be identified, wherein the single-particle mass spectrum includes the positive ion mass spectrum and / or negative ion mass spectrum of the particulate matter to be identified. Preprocessing module 2 is used to preprocess the mass spectrum of the single particle to obtain the feature vector of the ion to be identified of the particle to be identified. The feature vector of the ion to be identified includes multiple mass-to-charge ratio positions and the relative intensities corresponding to each mass-to-charge ratio position. The spectral library calling module 3 is used to call a pre-established pollution source spectral library, which includes reference source spectra corresponding to multiple known pollution source categories and source-specific fingerprint ion information associated with each reference source spectrum. The divergence coefficient calculation module 4 is used to calculate the matching divergence coefficient between the feature vector of the ion to be identified and the reference source spectrum corresponding to each known pollution source category; Candidate pollution source category determination module 5 is used to determine candidate pollution source categories from the plurality of known pollution source categories based on each matching divergence coefficient; The fingerprint ion matching module 6 is used to match the feature vector of the ion to be identified with the source-specific fingerprint ion information corresponding to the candidate pollution source category to obtain the fingerprint ion matching result. The pollution source matching result output module 7 is used to determine and output the pollution source matching result of the particulate matter to be identified based on the matching divergence coefficient corresponding to the candidate pollution source category and the fingerprint ion matching result.

[0090] Specific limitations regarding the particulate matter pollution source matching device based on single-particle mass spectrometry can be found in the limitations of the particulate matter pollution source matching method based on single-particle mass spectrometry mentioned above, and will not be repeated here. Each module in the aforementioned particulate matter pollution source matching device based on single-particle mass spectrometry can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0091] Those skilled in the art will understand that Figure 2 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the present application. Specific particulate matter pollution source matching devices based on single-particle mass spectrometry may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0092] Example 3 A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the particulate matter pollution source matching method based on single-particle mass spectrometry as described in Example 1.

[0093] Example 4 In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 3 As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system, computer programs, and the database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage medium. When the computer program is executed by the processor, it implements a particulate matter pollution source matching method based on single-particle mass spectrometry.

[0094] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0095] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps: S1. Obtain the single-particle mass spectrum of the particulate matter to be identified, wherein the single-particle mass spectrum includes the positive ion mass spectrum and / or negative ion mass spectrum of the particulate matter to be identified. S2. Preprocess the mass spectrum of the single particle to obtain the feature vector of the ion to be identified of the particle to be identified. The feature vector of the ion to be identified includes multiple mass-to-charge ratio positions and the relative intensities corresponding to each mass-to-charge ratio position. S3. Call the pre-established pollution source spectrum library, which includes reference source spectra corresponding to multiple known pollution source categories and source-specific fingerprint ion information associated with each reference source spectrum; S4. Calculate the matching divergence coefficient between the feature vector of the ion to be identified and the reference source spectrum corresponding to each known pollution source category; S5. Based on each matching divergence coefficient, determine the candidate pollution source category from the plurality of known pollution source categories; S6. Match the feature vector of the ion to be identified with the source-specific fingerprint ion information corresponding to the candidate pollution source category to obtain the fingerprint ion matching result; S7. Based on the matching divergence coefficients corresponding to the candidate pollution source categories and the fingerprint ion matching results, determine and output the pollution source matching results of the particulate matter to be identified.

[0096] Example 5 Furthermore, this application provides an embodiment for constructing a pollution source spectral library, used to illustrate the process of establishing a pollution source spectral library in the particulate matter pollution source matching method based on single-particle mass spectrometry described in this invention, and the role of low-frequency ions in identifying differences in the spectra of different pollution sources. This embodiment selects several typical industrial facilities as known pollution source categories, samples and performs single-particle mass spectrometry analysis on the particulate matter emitted by each industrial facility, and constructs an industrial source pollution source spectral library based on the obtained single-particle mass spectrometry data.

[0097] The sampling targets in this embodiment cover several typical industrial sectors, including non-metallic mineral products, chemical raw materials and chemical products, ferrous metal smelting and rolling processing, electricity and heat production and supply, metal products, automobile manufacturing, and transportation equipment manufacturing. Specific industrial facilities sampled are shown in the table below.

[0098] Table 3: Overview of Sampling Industrial Facilities

[0099] As shown in Table 3, the industrial facilities in this embodiment cover various industrial processes such as building materials, glass, cement, chemicals, fertilizers, steel manufacturing, power generation, container manufacturing, auto parts manufacturing, and shipbuilding. Because different industrial processes differ in their raw material composition, combustion conditions, production processes, and emission characteristics, their emitted particulate matter typically exhibits different ion peak combinations and relative intensity characteristics in single-particle mass spectra, making them suitable as known source samples for constructing pollution source spectral libraries.

[0100] After sampling, single-particle aerosol mass spectrometry was used to detect particulate matter emitted from various industrial facilities, obtaining positive and negative ion mass spectra for each particle. Subsequently, the acquired single-particle mass spectra underwent mass-to-charge ratio alignment, noise peak removal, invalid spectrum removal, and normalization to obtain the ion characteristic vectors of particulate matter from known sources. This preprocessing reduces the impact of instrument signal strength, variations in sample concentration, and fluctuations in particulate ionization efficiency on subsequent source spectrum construction, ensuring that comparisons between different pollution sources primarily reflect differences in particulate chemical composition.

[0101] Subsequently, clustering was performed on the feature vectors of multiple known source ions under the same industrial facility. Particulate matter with similar mass spectrometry characteristics was classified into the same particle type, and incidental or noise types with a low proportion of particles were removed. For the retained particle types, the average source spectrum of each particle type was calculated. Then, a weighted average was calculated based on the proportion of particles of each particle type in the corresponding industrial facility to obtain the weighted average source spectrum for each industrial facility. The weighted average source spectrum is stored in the pollution source spectrum library as the reference source spectrum for the corresponding industrial facility.

[0102] Through the above processing, this embodiment generates multiple industrial source spectrum objects. For facilities with similar sources or processes, merged source spectrum objects can be formed based on source spectrum similarity. For example, N2-N3 can be considered as the same non-metallic mineral product source spectrum object, and T1-T2 can be considered as the same shipbuilding source spectrum object. Thus, this embodiment obtains an industrial source reference source spectrum set for source spectrum difference analysis and source class differentiation.

[0103] In the source spectrum construction results, different industrial sources exhibit different chemical fingerprint characteristics. For example, Na in non-metallic mineral product sources... + Ca + Si + The plasma peaks are prominent and can reflect the characteristics of particulate matter related to glass, cement, or building materials; Fe in steel manufacturing sources + Mn + Cr + Metal-related ions have strong indicative significance; coal-fired power sources contain carbon ions, K... +Sulfate-related ions exhibit combustion emission characteristics; Cl in chemical or fertilizer manufacturing sources - The results show that some organic fragment ions are representative of the sources; while the sources from metal product manufacturing exhibit a mixed characteristic of coexistence of metal ions and organic fragment ions. These results indicate that particulate matter emitted from different industrial facilities has distinguishable source spectral characteristics at the single-particle mass spectrometry level.

[0104] Furthermore, such as Figure 4 As shown, the detectable ions in all industrial source spectra were statistically analyzed. In this embodiment, a total of 172 detectable ions were detected, and statistical analysis was performed stratified according to the minimum occurrence frequency of each ion in different industrial sources. Specifically, when the minimum occurrence frequency was greater than 0.65, only 3 ions met this condition, while 169 ions were below this threshold; when the minimum occurrence frequency was greater than 0.50, only 5 ions met this condition; when the minimum occurrence frequency was greater than 0.33, only 8 ions met this condition; when the minimum occurrence frequency was greater than 0.20, only 11 ions met this condition; and when the minimum occurrence frequency was greater than 0.10, 30 ions met this condition, while 142 ions were below this threshold, accounting for 82.6% of all detectable ions. When using a full-spectrum ion set, all 172 detectable ions were included in the analysis.

[0105] The above results indicate that low-frequency ions constitute a significant proportion of single-particle mass spectrometry data from industrial sources. If only high-frequency ions are retained during source spectrum construction or feature screening, a large number of low-frequency ions that may be source-specific will be excluded, resulting in the loss of crucial difference information in the pollution source library. Therefore, this embodiment did not simply remove low-frequency ions as noise during the construction of the pollution source library, but further analyzed their contribution to the differences in the source spectra.

[0106] Subsequently, based on the ion sets corresponding to different minimum occurrence frequency thresholds, the source spectrum differences between different industrial source spectra were analyzed. The results are as follows: Figure 5 As shown, when using ion sets with minimum occurrence frequencies greater than 0.65, 0.50, 0.33, 0.20, and 0.10, the average divergence coefficients between different industrial source spectra are generally at a moderate level, indicating that the distinguishability between different industrial source spectra is limited when relying solely on ions with higher occurrence frequencies. When using a full-spectrum ion set, the average divergence coefficient increases significantly, from approximately 0.41 when the minimum occurrence frequency is greater than 0.10 to approximately 0.75, an increase of approximately 0.34.

[0107] These results demonstrate that incorporating low-frequency ions into source spectrum difference analysis significantly increases the overall differences between different industrial source spectra. In other words, although low-frequency ions may occur in a low proportion within a specific industrial source particle population, they may only appear under specific industrial source or process conditions, thus providing source difference information that is difficult to reflect with high-frequency common ions. Therefore, low-frequency ions are not all random noise, but rather make a substantial contribution to distinguishing different pollution source categories.

[0108] Further as Figure 6 As shown, pairwise comparisons were performed on each industrial source spectrum object under the full-spectrum ion set to obtain the industrial source spectrum divergence coefficient matrix. The matrix results show that there are generally significant differences between different industrial source spectra, with the divergence coefficients of many cross-industry industrial source spectrum pairs reaching above 0.70, and some source spectrum pairs reaching above 0.80. For example, the source spectrum divergence coefficients between steel manufacturing sources and some non-metallic mineral product sources, chemical manufacturing sources, metal product sources, automobile manufacturing sources, or shipbuilding sources are at a high level, indicating that their source spectrum differences are significant and suitable for storage as independent pollution source categories in the pollution source spectrum library. For some industrial facilities within the same industry or with similar processes, their divergence coefficients are relatively low, indicating that the source spectra of these facilities have a certain similarity, and source-specific fingerprint ion information can be combined to further determine whether they should be merged into the same pollution source category.

[0109] In terms of source-specific fingerprint ion identification, this embodiment uses the frequency and relative intensity characteristics of each ion in different industrial sources to screen out multiple ions with source indication significance. For example, Cr in steel manufacturing sources + Fe + Mn + Metal-related ions can serve as important indicators of particulate matter in steel manufacturing; Na in non-metallic mineral products sources + Ca + Si + Plasma can serve as a characteristic ion source for glass, cement, or building materials; coal-fired power sources contain carbon ions, K+, and other ions. + Sulfate-related ions can also serve as characteristic features of coal combustion emissions; Cl in chemical or fertilizer manufacturing sources - Some organic fragment ions can be used as auxiliary identification criteria. For some ions that occur less frequently but are stable only in specific industrial sources and have a clear chemical classification, they can also be stored as auxiliary marker ions in the pollution source spectral library to enhance the ability to distinguish between pollution source categories.

[0110] Based on the above results, this embodiment associates and stores the weighted average source spectrum, particle type average source spectrum, detectable ion frequency, source-specific fingerprint ion information, auxiliary marker ion information, and source spectrum divergence coefficient matrix corresponding to each industrial source, forming an industrial source pollution spectrum library. This pollution source spectrum library not only records the overall reference source spectrum of different industrial sources, but also retains low-frequency ion information and source-specific fingerprint ion information that can reflect the differences between pollution sources, enabling the pollution source spectrum library to more fully express the differences in particulate matter chemical composition between different industrial facilities.

[0111] The results of this embodiment demonstrate that low-frequency ions play a crucial role in the construction of industrial source pollution spectral libraries. Specifically, low-frequency ions constitute a significant proportion of detectable ions, and when they are included in the full-spectrum source spectrum difference analysis, the average divergence coefficient between different industrial sources increases significantly, indicating that low-frequency ions can significantly enhance the distinguishability between industrial source spectra. Compared to methods that only retain high-frequency ions for source spectrum construction, this embodiment, by retaining and analyzing low-frequency ions, allows the pollution source spectral library to contain more source-specific chemical information, thereby providing a more objective and quantifiable basis for subsequent merging, splitting, and independent classification of pollution source categories.

[0112] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0113] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0114] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A method for matching a particle pollution source based on single particle mass spectrometric patterns, characterized by, include: Obtain a single-particle mass spectrum of the particulate matter to be identified, wherein the single-particle mass spectrum includes a positive ion mass spectrum and / or a negative ion mass spectrum of the particulate matter to be identified. The mass spectrum of the single particle is preprocessed to obtain the feature vector of the ion to be identified of the particle to be identified. The feature vector of the ion to be identified includes multiple mass-to-charge ratio positions and the relative intensities corresponding to each mass-to-charge ratio position. The system invokes a pre-established pollution source spectrum library, which includes reference source spectra corresponding to multiple known pollution source categories and source-specific fingerprint ion information associated with each reference source spectrum. The pollution source spectral library is established through the following steps: Obtain single-particle mass spectra of particulate matter from known sources corresponding to multiple known pollution source categories; The single-particle mass spectra of the known source particles are preprocessed to obtain the characteristic vectors of the known source ions; Clustering the feature vectors of ions from multiple known sources under the same known pollution source category yields multiple particle types; Calculate the average source spectrum for each particle type; Based on the proportion of each particle type in the corresponding known pollution source category, a weighted average source spectrum of multiple particle types is calculated to obtain the weighted average source spectrum corresponding to the known pollution source category. The weighted average source spectrum is stored in the pollution source spectrum library as a reference source spectrum corresponding to the known pollution source category; The source-specific fingerprint ion information is established through the following steps: Calculate the frequency of occurrence of ion peaks corresponding to each mass-to-charge ratio position in each known pollution source category; When the frequency of the ion peak corresponding to the target mass-to-charge ratio position in the target known pollution source category is greater than the first frequency threshold, and the frequency of its occurrence in other known pollution source categories is less than the second frequency threshold, the ion peak corresponding to the target mass-to-charge ratio position is determined as the source-specific fingerprint ion of the target known pollution source category. The mass-to-charge ratio position of the source-specific fingerprint ions is associated with the reference source spectrum corresponding to the known pollution source category of the target and stored in the pollution source spectrum library; Calculate the matching divergence coefficient between the feature vector of the ion to be identified and the reference source spectrum corresponding to each known pollution source category; Based on each matching divergence coefficient, candidate pollution source categories are determined from the plurality of known pollution source categories; The feature vector of the ion to be identified is matched with the source-specific fingerprint ion information corresponding to the candidate pollution source category to obtain the fingerprint ion matching result; Based on the matching divergence coefficients corresponding to the candidate pollution source categories and the fingerprint ion matching results, the pollution source matching results of the particulate matter to be identified are determined and output.

2. The single particle mass spectrometric pattern-based particulate pollution source matching method according to claim 1, characterized by, The reference source spectrum includes: a weighted average source spectrum corresponding to known pollution source categories; The weighted average source spectrum includes: multiple mass-to-charge ratio locations and the reference relative intensity corresponding to each mass-to-charge ratio location; The calculation of the matching divergence coefficient between the feature vector of the ion to be identified and the reference source spectrum corresponding to each known pollution source category includes: At the same mass-to-charge ratio position, the matching divergence coefficient between the feature vector of the ion to be identified and the reference relative intensity in the reference source spectrum is calculated based on the relative intensity in the feature vector of the ion to be identified and the reference relative intensity in the reference source spectrum.

3. The single particle mass spectrometric pattern-based particulate pollution source matching method according to claim 2, characterized by, The step of calculating the matching divergence coefficient between the feature vector of the ion to be identified and the reference relative intensity in the reference source spectrum, based on the relative intensity in the feature vector of the ion to be identified and the reference relative intensity in the reference source spectrum, includes: At multiple mass-to-charge ratio positions involved in the calculation, the relative intensity in the feature vector of the ion to be identified is compared with the reference relative intensity in the reference source spectrum. For each mass-to-charge ratio position, the ion peak splitting term at that mass-to-charge ratio position is determined by the ratio of the difference between the relative intensity and the reference relative intensity to the sum of the two. The matching divergence coefficient is obtained by performing mean square statistical processing on the divergence terms of ion peaks at multiple mass-to-charge ratio positions.

4. The single particle mass spectrometric pattern-based particulate pollution source matching method according to claim 3, characterized by, The step of determining candidate pollution source categories from the plurality of known pollution source categories based on each matching divergence coefficient includes: Determine the minimum matching divergence coefficient from multiple matching divergence coefficients; The known pollution source category corresponding to the minimum matching divergence coefficient is determined as the candidate pollution source category; Determine whether the minimum matching divergence coefficient is less than a preset matching threshold; If the minimum matching divergence coefficient is less than the preset matching threshold, then the candidate pollution source category is retained; If the minimum matching divergence coefficient is not less than the preset matching threshold, then the particulate matter to be identified is marked as particulate matter of unknown origin.

5. The particulate matter pollution source matching method based on single-particle mass spectrometry according to claim 1, characterized in that, The source-specific fingerprint ion information includes the mass-to-charge ratio and position of the source-specific fingerprint ions; The step of matching the feature vector of the ion to be identified with the source-specific fingerprint ion information corresponding to the candidate pollution source category to obtain the fingerprint ion matching result includes: Determine whether there is an ion peak in the feature vector of the ion to be identified that has the same mass-to-charge ratio position as the source-specific fingerprint ion corresponding to the candidate pollution source category; If present, a fingerprint ion matching result is generated to support the candidate pollution source category; If it does not exist, a fingerprint ion matching result is generated to reduce the confidence of the candidate pollution source category.

6. The particulate matter pollution source matching method based on single-particle mass spectrometry according to claim 5, characterized in that, The source-specific fingerprint ion information also includes a reference relative intensity range of the source-specific fingerprint ions; After determining that there is an ion peak in the feature vector of the ion to be identified that has the same mass-to-charge ratio position as the source-specific fingerprint ion, it is further determined whether the relative intensity of the ion peak falls within the reference relative intensity range. When the relative intensity of the ion peak falls within the reference relative intensity range, a fingerprint ion matching result is generated to support the candidate pollution source category.

7. The particulate matter pollution source matching method based on single-particle mass spectrometry according to claim 1, characterized in that, The pollution source library also establishes classification criteria among known pollution source categories through the following steps: Calculate the source spectrum divergence coefficient between the reference source spectra corresponding to different known pollution source categories; When the source spectrum divergence coefficient between two known pollution source categories is less than the first divergence coefficient threshold, the two known pollution source categories are marked as highly similar source categories; When the source spectrum divergence coefficient between two known pollution source categories is between the first divergence coefficient threshold and the second divergence coefficient threshold, the two known pollution source categories are marked as moderately similar source categories. When the source spectrum divergence coefficient between two known pollution source categories is greater than the second divergence coefficient threshold, the two known pollution source categories are marked as highly dissimilar source categories; wherein, the first divergence coefficient threshold is less than the second divergence coefficient threshold.

8. A particulate matter pollution source matching device based on single-particle mass spectrometry, characterized in that, include: The spectrum acquisition module is used to acquire the single-particle mass spectrum of the particulate matter to be identified, wherein the single-particle mass spectrum includes the positive ion mass spectrum and / or negative ion mass spectrum of the particulate matter to be identified. The preprocessing module is used to preprocess the mass spectrum of the single particle to obtain the feature vector of the ion to be identified of the particle to be identified. The feature vector of the ion to be identified includes multiple mass-to-charge ratio positions and the relative intensities corresponding to each mass-to-charge ratio position. The spectral library calling module is used to call a pre-established pollution source spectral library, which includes reference source spectra corresponding to multiple known pollution source categories and source-specific fingerprint ion information associated with each reference source spectrum; The pollution source spectral library is established through the following steps: Obtain single-particle mass spectra of particulate matter from known sources corresponding to multiple known pollution source categories; The single-particle mass spectra of the known source particles are preprocessed to obtain the characteristic vectors of the known source ions; Clustering the feature vectors of ions from multiple known sources under the same known pollution source category yields multiple particle types; Calculate the average source spectrum for each particle type; Based on the proportion of each particle type in the corresponding known pollution source category, a weighted average source spectrum of multiple particle types is calculated to obtain the weighted average source spectrum corresponding to the known pollution source category. The weighted average source spectrum is stored in the pollution source spectrum library as a reference source spectrum corresponding to the known pollution source category; The source-specific fingerprint ion information is established through the following steps: Calculate the frequency of occurrence of ion peaks corresponding to each mass-to-charge ratio position in each known pollution source category; When the frequency of the ion peak corresponding to the target mass-to-charge ratio position in the target known pollution source category is greater than the first frequency threshold, and the frequency of its occurrence in other known pollution source categories is less than the second frequency threshold, the ion peak corresponding to the target mass-to-charge ratio position is determined as the source-specific fingerprint ion of the target known pollution source category. The mass-to-charge ratio position of the source-specific fingerprint ions is associated with the reference source spectrum corresponding to the known pollution source category of the target and stored in the pollution source spectrum library; The divergence coefficient calculation module is used to calculate the matching divergence coefficient between the feature vector of the ion to be identified and the reference source spectrum corresponding to each known pollution source category; The candidate pollution source category determination module is used to determine candidate pollution source categories from the plurality of known pollution source categories based on each matching divergence coefficient; The fingerprint ion matching module is used to match the feature vector of the ion to be identified with the source-specific fingerprint ion information corresponding to the candidate pollution source category to obtain the fingerprint ion matching result. The pollution source matching result output module is used to determine and output the pollution source matching result of the particulate matter to be identified based on the matching divergence coefficient corresponding to the candidate pollution source category and the fingerprint ion matching result.

Images