AU2020103635A4

AU2020103635A4 - Method for monitoring source of atmospheric particulate matters in real time

Info

Publication number: AU2020103635A4
Application number: AU2020103635A
Authority: AU
Inventors: Zhengxu HUANG; Lei Li; Mei Li; Zhen Zhou
Original assignee: Jinan University; University of Jinan
Current assignee: Jinan University
Priority date: 2020-11-24
Filing date: 2020-11-24
Publication date: 2021-02-04
Anticipated expiration: 2028-11-24

Abstract

The present disclosure discloses a method for monitoring a source of atmospheric particulate matters in real time. The method includes: detecting fine particles in the atmosphere to obtain a feature spectrum of the fine particles; performing normalization processing on an eigenvector of the obtained feature spectrum, to obtain a normalized eigenvector corresponding to the feature spectrum; and multiplying the normalized eigenvector by each eigenvector pre-stored in a source feature spectrum database, and classifying the fine particles based on a multiplication result, to implement source monitoring of the atmospheric particulate matters. The method of the present disclosure has simple steps and is easy to implement, facilitating real-time monitoring of sources of atmospheric particulate matters. In addition, this method does not involve too many human operations, thereby reducing human errors and improving monitoring accuracy. The method for monitoring a source of atmospheric particulate matters in real time according to the present disclosure is widely used in the field of atmospheric monitoring. DRAWINGS Detect fine particles in the atmosphere by using a single particle aerosol mass spectrometer to obtain a feature spectrum of the fine particles Perform normalization processing on an eigenvector of the obtained feature spectrum, to obtain a normalized eigenvector Multiply the obtained normalized eigenvector by each pre-stored eigenvector, and classify the fine particles based on a multiplication result, to implement source monitoring FIG. 1 I

Description

DRAWINGS

Detect fine particles in the atmosphere by using a single particle aerosol mass spectrometer to obtain a feature spectrum of the fine particles

Perform normalization processing on an eigenvector of the obtained feature spectrum, to obtain a normalized eigenvector

Multiply the obtained normalized eigenvector by each pre-stored eigenvector, and classify the fine particles based on a multiplication result, to implement source monitoring

FIG. 1

I

METHOD FOR MONITORING SOURCE OF ATMOSPHERIC PARTICULATE MATTERS IN REAL TIME TECHNICAL FIELD The present disclosure relates to an information processing method, and in particular, to a method for monitoring a source of atmospheric particulate matters in real time. BACKGROUND Atmospheric particulate matters are a multiphase dispersion system formed by the interaction of solids and liquids suspended in gas, which directly affect human health, regional air quality, and global climate. At present, the pollution of atmospheric particulate matters has become one of the major problems in China, and need to be tackled urgently. To effectively control the pollution of atmospheric particulate matters, it is necessary to monitor sources of the atmospheric particulate matters. A chemical mass balance (CMB) method or a positive matrix factorization (PMF) method are conventionally used to monitor a source of atmospheric particulate matters usually use . However, these traditional monitoring methods have many defects. For example, 1. The processing steps are complex and time-consuming, leading to poor real-time performance; 2. Too many human operations are involved, which may introduce many human errors and reduce monitoring accuracy. SUMMARY To resolve the above technical problems, the present disclosure provides a simple, accurate, and easy-to-implement method for monitoring a source of atmospheric particulate matters in real time. The technical solution adopted by the present disclosure is: a method for monitoring a source of atmospheric particulate matters in real time. The method includes: B. detecting fine particles in the atmosphere by using a single-particle aerosol mass spectrometer to obtain a feature spectrum of the fine particles; C. performing normalization processing on an eigenvector of the obtained feature spectrum, to obtain a normalized eigenvector corresponding to the feature spectrum; and D. multiplying the normalized eigenvector obtained in step C by each eigenvector pre-stored in a source feature spectrum database, and classifying the fine particles based on a multiplication result, to implement source monitoring of the atmospheric particulate matters. Further, step A may be provided before step B, and step A may be: establishing the source feature spectrum database. Further, step C may include: C1. discretizing the obtained feature spectrum and converting it into a vector, where each element in the vector represents a peak area of each mass-to-charge ratio ion peak in the feature spectrum; and

C2. dividing each element in the vector obtained in step C1 by the Euclidean norm of the vector to obtain the normalized eigenvector corresponding to the feature spectrum. Further, an expression of the normalized eigenvector may be:

[j Vi' ... V ) = kI A VAi 2+AiT22

where Vi,n represents the nth normalized eigenvector value in the ith feature spectrum, and Ai,, represents a peak area value of the nth mass-to-charge ratio ion peak in the it featurespectrum. Further, step D may include: D1. performing a dot-product operation on the normalized eigenvector obtained in step C and each eigenvector pre-stored in the source feature spectrum database, to obtain a plurality of dot product values; and D2. determining whether the plurality of dot product values are all less than a specified threshold; and if yes, determining a category of the fine particles as others; if no, determining a category of the fine particles as a category corresponding to a maximum dot product value, so as to implement monitoring of the sources of the atmospheric particulate matters. Further, the feature spectrum of the fine particles may include positive and negative ion spectra. The beneficial effects of the present disclosure are as follows: By using the method of the present disclosure, the sources of atmospheric particulate matters can be monitored online in real time. In addition, the method of the present disclosure is simple and easy to implement, thereby achieving a high real-time performance. Furthermore, since the monitoring method of the present disclosure does not involve too many human operations, man-made errors can be greatly reduced, and accuracy of monitoring is improved. BRIEF DESCRIPTION OF DRAWINGS The specific implementations of the present disclosure are described in more detail with reference to the accompanying drawings. FIG. 1 is a flowchart of steps of a method for monitoring a source of atmospheric particulate matters in real time according to the present disclosure. DETAILED DESCRIPTION As shown in FIG. 1, a method for monitoring a source of atmospheric particulate matters in real time includes: B. detecting fine particles in the atmosphere by using a single-particle aerosol mass spectrometer to obtain a feature spectrum of the fine particles; C. performing normalization processing on an eigenvector of the obtained feature spectrum, to obtain a normalized eigenvector corresponding to the feature spectrum; and

D. multiplying the normalized eigenvector obtained in step C by each eigenvector pre-stored in a source feature spectrum database, and classifying the fine particles based on a multiplication result, to implement source monitoring of the atmospheric particulate matters, where for the eigenvectors pre-stored in the source feature spectrum database, one pre-stored eigenvector represents one source category of fine particles, and one source category of fine particles may correspond to a plurality of different pre-stored eigenvectors. Further, in a preferred implementation, step A may be provided before step B, and step A may be: establishing the source feature spectrum database. Further, in a preferred implementation, step C may include: C1. discretizing the obtained feature spectrum and converting it into a vector, where each element in the vector represents a peak area of each mass-to-charge ratio ion peak in the feature spectrum; and C2. dividing each element in the vector obtained in step C1 by the Euclidean norm of the vector to obtain the normalized eigenvector corresponding to the feature spectrum. Further, in a preferred implementation, an expression of the normalized eigenvector may be:

k A A A +A2 + +A

where Vi,n represents thent' normalized eigenvector value in the ith feature spectrum, and Ai,, represents a peak area value of the n'f mass-to-charge ratio ion peak in the ith featurespectrum. Further, in a preferred implementation, step D may include: D1. performing a dot-product operation on the normalized eigenvector obtained in step C and each eigenvector pre-stored in the source feature spectrum database, to obtain a plurality of dot product values; and D2. determining whether the plurality of dot product values are all less than a specified threshold; and if yes, determining a category of the fine particles as others; if no, determining a category of the fine particles as a category corresponding to a maximum dot product value, so as to implement monitoring of the sources of the atmospheric particulate matters. Further, in a preferred implementation, the feature spectrum of the fine particles may include positive and negative ion spectra. Detailed description of the present disclosure A method for monitoring a source of atmospheric particulate matters in real time is as follows: Si. Establish a source feature spectrum database, where the source feature spectrum database is used to pre-store eigenvectors corresponding to source categories of fine particles, that is, one pre-stored eigenvector in the source feature spectrum database represents one source category of fine particles, and one source category of fine particles may correspond to a plurality of different pre-stored eigenvectors. S2. Detect fine particles in the atmosphere by using a single-particle aerosol mass spectrometer to obtain a feature spectrum of the fine particles. S3. Discretize the obtained feature spectrum and convert it into a vector, where each element in the vector represents a peak area of each mass-to-charge ratio ion peak in the feature spectrum. S4. Divide each element in the vector obtained in step S3 by the Euclidean norm of the vector to obtain the normalized eigenvector corresponding to the feature spectrum, where an expression of the normalized eigenvector is: where Vi,n represents the nth normalized eigenvector value in the ith feature spectrum, and Ai,, represents a peak area value of the nth mass-to-charge ratio ion peak in the ith featurespectrum. S5. Perform a dot-product operation on the normalized eigenvector obtained in step S4 and each eigenvector pre-stored in the source feature spectrum database, that is, compare a similarity between a feature spectrum of the detected fine particles and a feature spectrum pre-stored in the source feature spectrum database, to obtain a plurality of dot product values. S6. Determine whether the plurality of dot product values obtained are all less than a specified threshold; and if yes, determine a category of the detected fine particles as others; if no, determine a category of the detected fine particles as a category corresponding to a maximum dot product value, that is, when some or all of the dot product values are greater than or equal to the specified threshold, select the maximum dot product value from these dot product values, and then assign the source category corresponding to this maximum dot product value to the detected fine particles, so as to implement monitoring of the sources of the atmospheric particulate matters. Step Sl of establishing a source feature spectrum database includes the following specific processing steps: Step 1. Use a single-particle aerosol mass spectrometer to detect the fine particles of different common emission sources simulated in the laboratory and the fine particles of typical industrial emission sources in the region, to obtain a series of feature spectra for each emission source, where each feature spectrum includes positive and negative ion spectra. One category of emission source may correspond to a plurality of feature spectra. Step 2. Discretize the obtained feature spectrum obtained in step 1 and convert it into a vector, where each element in each vector represents a peak area of each mass-to-charge ratio ion peak in one feature spectrum.

Step 3. Perform normalization processing on the vector obtained in step 2, that is, divide each element in one vector obtained by the Euclidean norm of the vector to obtain a normalized eigenvector corresponding to the feature spectrum, and then combine eigenvectors to establish the source feature spectrum database in the present disclosure. For step S5 of performing a dot-product operation on the normalized eigenvector obtained in step S4 and each eigenvector pre-stored in the source feature spectrum database, an expression of the operation is as follows:

DX =VxWM, +V 2x WM 2 +...+V xWM

where DPij represents a dot product value of two eigenvectors, Vi, represents the nth normalized eigenvector value in the it feature spectrum, and WMj,n represents the nth normalized eigenvector in the jth pre-stored eigenvector, that is, in the jth pre-stored feature spectrum. It can be learned from the above that the steps of the method of the present disclosure is simple and easy to implement, thereby achieving a high real-time performance of monitoring. In addition, since the monitoring method of the present disclosure does not involve too many human operations, man-made errors can be greatly reduced, and accuracy of monitoring is improved. Furthermore, because the method of the present disclosure is implemented by using a database, the staff can conveniently update and maintain data in the database, further improving the reliability, stability and accuracy of source monitoring. The above merely describes specific embodiments of the present disclosure, but the present disclosure is not limited thereto. A person skilled in the art can make modifications or replacements without departing from the spirit of the present disclosure, and these modifications or replacements shall fall within the protection scope of the claims of the present disclosure.

Claims

What is claimed is: 1. A method for monitoring a source of atmospheric particulate matters in real time, wherein the method comprises: B. detecting fine particles in the atmosphere by using a single-particle aerosol mass spectrometer to obtain a feature spectrum of the fine particles; C. performing normalization processing on an eigenvector of the obtained feature spectrum, to obtain a normalized eigenvector corresponding to the feature spectrum; and D. multiplying the normalized eigenvector obtained in step C by each eigenvector pre-stored in a source feature spectrum database, and classifying the fine particles based on a multiplication result, to implement source monitoring of the atmospheric particulate matters.
2. The method for monitoring a source of atmospheric particulate matters in real time according to claim 1, wherein step A is provided before step B, and step A comprises: establishing the source feature spectrum database.
3. The method for monitoring a source of atmospheric particulate matters in real time according to claim 1, wherein step C comprises: C1. discretizing the obtained feature spectrum and converting it into a vector, wherein each element in the vector represents a peak area of each mass-to-charge ratio ion peak in the feature spectrum; and C2. dividing each element in the vector obtained in step C1 by the Euclidean norm of the vector to obtain the normalized eigenvector corresponding to the feature spectrum; wherein an expression of the normalized eigenvector is:

VA, 2 +IA 2 2+i + Ai2 wherein Vi,n represents the nth normalized eigenvectr value in the ith feature spectrum, and Ai,n represents a peak area value of the nth mass-to-charge ratio ion peak in the ith featurespectrum.
4. The method for monitoring a source of atmospheric particulate matters in real time according to claim 1, wherein step D comprises: D1. performing a dot-product operation on the normalized eigenvector obtained in step C and each eigenvector pre-stored in the source feature spectrum database, to obtain a plurality of dot product values; and D2. determining whether the plurality of dot product values are all less than a specified threshold; and if yes, determining a category of the fine particles as others; if no, determining a category of the fine particles as a category corresponding to a maximum dot product value, so as to implement monitoring of the sources of the atmospheric particulate matters.
5. The method for monitoring a source of atmospheric particulate matters in real time according to claim 1, wherein the feature spectrum of the fine particles comprises positive and negative ion spectra.

FIG. 1 DRAWINGS