Summary of the invention
In order to solve the problems of the technologies described above, the object of this invention is to provide the method for real-time in the high air fine grained source of a kind of easy realization, simple, real-time and accuracy.
The technical solution adopted in the present invention is: a kind of method of real-time for air fine grained source, and the method comprises:
B, by adopting individual particle aerosol mass spectrometer to detect the fine grained in air, thus obtain described fine grain feature spectrogram;
C, the normalized of proper vector is carried out to the feature spectrogram obtained, thus obtain the normalization characteristic vector corresponding with this feature spectrogram;
After D, the normalization characteristic vector obtained by step C are multiplied with each proper vector prestored in source characteristics spectrum library, according to the result be multiplied, origin classification are carried out to described fine grained, thus realize the monitoring in air fine grained source.
Further, be also provided with steps A before described step B, described steps A is: set up source characteristics spectrum library.
Further, described step C comprises:
C1, to obtain feature spectrogram carry out sliding-model control after be converted into vector, wherein, each element in described vector represents the peak area of each mass-to-charge ratio quasi-molecular ions in this feature spectrogram;
Each element in C2, the vector that obtained by step C1 and the Euclid norm of this vector are divided by, thus obtain the normalization characteristic vector corresponding with this feature spectrogram.
Further, the representation formula of described normalization characteristic vector is:
Wherein, V
i , nrepresent the n-th normalization characteristic vector value in i-th feature spectrogram, A
i , nrepresent the peak area value of the n-th mass-to-charge ratio quasi-molecular ions in i-th feature spectrogram.
Further, described step D comprises:
After each proper vector prestored in D1, the normalization characteristic vector obtained by step C and source characteristics spectrum library carries out point multiplication operation one by one, obtain multiple somes product value;
D2, judge whether multiple somes product value are all less than the threshold value of setting, if so, then this fine grained are judged to be other classification, otherwise, then this fine grained is judged to be the classification corresponding with maximum point product value, thus realizes the monitoring in air fine grained source.
Further, described fine grain characteristic spectrum comprises negative ions spectrogram.
The invention has the beneficial effects as follows: by adopting method of the present invention, staff can monitor the fine grain source of air real-time online, and operation is convenient, and method step of the present invention simple, be easy to realize, therefore real-time is high.Further, because monitoring method of the present invention is without the need to relating to too much manual operation, therefore, it is possible to greatly reduce the error artificially brought, thus improve the accuracy of monitoring.
Embodiment
As shown in Figure 1, a kind of method of real-time for air fine grained source, the method comprises:
B, by adopting individual particle aerosol mass spectrometer to detect the fine grained in air, thus obtain described fine grain feature spectrogram;
C, the normalized of proper vector is carried out to the feature spectrogram obtained, thus obtain the normalization characteristic vector corresponding with this feature spectrogram;
After D, the normalization characteristic vector obtained by step C are multiplied with each proper vector prestored in source characteristics spectrum library, according to the result be multiplied, origin classification are carried out to described fine grained, thus realize the monitoring in air fine grained source;
Wherein, for the proper vector prestored in described source characteristics spectrum library, a proper vector prestored represents a kind of fine grain source categories, and a kind of fine grain source categories can the proper vector that prestores of corresponding multiple difference.
Be further used as preferred embodiment, be also provided with steps A before described step B, described steps A is: set up source characteristics spectrum library.
Be further used as preferred embodiment, described step C comprises:
C1, to obtain feature spectrogram carry out sliding-model control after be converted into vector, wherein, each element in described vector represents the peak area of each mass-to-charge ratio quasi-molecular ions in this feature spectrogram;
Each element in C2, the vector that obtained by step C1 and the Euclid norm of this vector are divided by, thus obtain the normalization characteristic vector corresponding with this feature spectrogram.
Be further used as preferred embodiment, the representation formula of described normalization characteristic vector is:
Wherein, V
i , nrepresent the n-th normalization characteristic vector value in i-th feature spectrogram, A
i , nrepresent the peak area value of the n-th mass-to-charge ratio quasi-molecular ions in i-th feature spectrogram.
Be further used as preferred embodiment, described step D comprises:
After each proper vector prestored in D1, the normalization characteristic vector obtained by step C and source characteristics spectrum library carries out point multiplication operation one by one, obtain multiple somes product value;
D2, judge whether multiple somes product value are all less than the threshold value of setting, if so, then this fine grained are judged to be other classification, otherwise, then this fine grained is judged to be the classification corresponding with maximum point product value, thus realizes the monitoring in air fine grained source.
Be further used as preferred embodiment, described fine grain characteristic spectrum comprises negative ions spectrogram.
The specific embodiment of the present invention
For the method for real-time in air fine grained source, its concrete grammar is as follows:
S1, set up source characteristics spectrum library, described source characteristics spectrum library is for the proper vector corresponding with fine grained source categories that prestore, namely in source characteristics spectrum library, a proper vector prestored represents a kind of fine grain source categories, and a kind of fine grain source categories can the proper vector that prestores of corresponding multiple difference;
S2, by adopting individual particle aerosol mass spectrometer to detect the fine grained in air, thus obtain described fine grain feature spectrogram;
S3, to obtain feature spectrogram carry out sliding-model control after be converted into vector, wherein, each element in described vector represents the peak area of each mass-to-charge ratio quasi-molecular ions in this feature spectrogram;
Each element in S4, the vector that obtained by step S3 and the Euclid norm of this vector are divided by, thus obtain the normalization characteristic vector corresponding with this feature spectrogram, and the representation formula of described normalization characteristic vector is:
Wherein, V
i , nrepresent the n-th normalization characteristic vector value in i-th feature spectrogram, A
i , nrepresent the peak area value of the n-th mass-to-charge ratio quasi-molecular ions in i-th feature spectrogram;
Each proper vector prestored in S5, the normalization characteristic vector obtained by step S4 and source characteristics spectrum library carries out point multiplication operation one by one, namely contrast the similarity between the characteristic spectrum that prestores in this detected fine grain characteristic spectrum and source characteristics spectrum library, and after point multiplication operation, just obtain multiple somes product value;
S6, the multiple somes product value obtained are judged, judge whether multiple somes product value are all less than the threshold value of setting, if, then this detected fine grained is judged to be other classification, otherwise, then this detected fine grained is judged to be the classification corresponding with maximum point product value, namely when partly or entirely some product value is all more than or equal to the threshold value of setting, be more than or equal in the some product value of setting threshold value at these, the point product value that selected value is maximum, then this detected fine grained is given to by with the source categories corresponding to this maximum some product value, so just, the enforcement monitoring in air fine grained source can be realized.
For step S1, set up source characteristics spectrum library, its concrete treatment step is as follows:
Step one, employing individual particle aerosol mass spectrometer are to the fine grained of the common different emission source of experiment lab simulation, and the fine grained of the typical industrial emission source in area detects, so just, the series of features spectrogram of each classification emission source can be obtained, and each feature spectrogram contains negative ions spectrogram.Wherein, a kind emission source can corresponding multiple feature spectrogram.
Step 2, each the feature spectrogram obtained step one are converted into single vector after carrying out sliding-model control one by one, and each element in each vector then represents the peak area of each mass-to-charge ratio quasi-molecular ions in a feature spectrogram.
Step 3, the vector obtained in step 2 to be normalized, Euclid norm by each element in a vector and this vector is divided by, so just, the normalization characteristic vector corresponding with this feature spectrogram can be obtained, then, these proper vectors are combined, then establishes source characteristics spectrum library of the present invention.
For step S5, when each proper vector prestored in the normalization characteristic vector obtained by step S4 with source characteristics spectrum library carries out point multiplication operation one by one, the expression formula of this computing is as follows:
Wherein, DP
i,jrepresent the some product value of above-mentioned two proper vectors, V
i , nrepresent the n-th normalization characteristic vector value in i-th feature spectrogram, and WM
j , nbe expressed as in the individual proper vector prestored of jth, in the characteristic spectrum that namely jth prestores, the n-th normalization characteristic vector value.
From the above, method step of the present invention is simple, is easy to realize, and therefore, the real-time of its monitoring is high.In addition, because monitoring method of the present invention is without the need to relating to too much manual operation, therefore, it is possible to greatly reduce the error that manual operation brings, thus improve the accuracy of monitoring.Further, because method of the present invention have employed database to realize, therefore, it is possible to be convenient to staff to carry out updating maintenance to the data in database, thus improve reliability, stability and the accuracy of monitoring in source further.
More than that better enforcement of the present invention is illustrated, but the invention is not limited to described embodiment, those of ordinary skill in the art also can make all equivalent variations or replacement under the prerequisite without prejudice to spirit of the present invention, and these equivalent distortion or replacement are all included in the application's claim limited range.