CN103926215B - A kind of method for quick of Asphalt Penetration - Google Patents

Abstract

A kind of method for quick of Asphalt Penetration, the near infrared spectrum data storehouse based on Colophonium for the method, by combining near infrared technology and topology technology, realize the quick detection to Asphalt Penetration；First, using near infrared spectrometer, several Colophonium sample is scanned obtaining with the near infrared spectrum of each Colophonium, obtains sample spectrum data, set up Colophonium near infrared spectrum data storehouse；Using near infrared spectrometer, Colophonium sample to be measured is scanned, obtains spectrum vector；Calculate the distance between each sample spectrum in spectrum to be measured and data base, search in library of spectra and be less than the Asphalt Penetration data of a certain threshold value with the spectrum intervals of this spectrum to be measured, calculate penetration of bitumens to be measured.The present invention is based near infrared detection technology and topology modeling technique, measures Asphalt Penetration in order to quick, meanwhile, can improve detection adaptability by being processed to increase sample point in library of spectra to available sample in Colophonium library of spectra.

Description

A kind of method for quick of Asphalt Penetration

Technical field

The present invention relates to the Asphalt Penetration detection technique of petrochemical industry, especially Asphalt Penetration Fast Evaluation skill Art.

Background technology

In petroleum refining process, the normal operation of decompressor is important step, after light oil is extracted, flows out under vacuum tower Be straight asphalt or residual oil.The JTG F40-2004 of Ministry of Communications's revision《Standard specification for construction and acceptance of highway asphalt pavement》Middle by pin As a new index, in order to reflect the quality of Colophonium, therefore analyse whether accurately can be to Colophonium for needle penetration for in-degree index Classification causes to directly affect.That the detection of Asphalt Penetration is commonly used at present is GB/T4509-2010《Asphalt Penetration measures Method》, enter the depth representing of bitumen samples in certain load, time and the sagging direct puncture of temperature conditionss with standard pin.The method exists There is more influence factor, for example standard pin weight, injection time and test operation maneuver etc. are all to Colophonium pin in continuous mode The mensure of in-degree affects greatly, and in addition, its detection time is longer, is difficult to meet in the course of processing to Asphalt Penetration The requirement of analysis real-time.In order to solve this problem, the present invention proposes a kind of Asphalt Penetration method for quick, adopts Near infrared detection technology combines topology modeling technique, realizes the quick detection of Asphalt Penetration.

Content of the invention

The present invention seeks to time-consuming for traditional Asphalt Penetration detection method, artifical influence factor is more, be difficult to Meet the requirement that Petrochemical Enterprises analyze real-time and accuracy to Asphalt Penetration, propose a kind of suitable commercial Application, quick, accurate True and eco-friendly Asphalt Penetration method for quick.The method is based on Colophonium near infrared spectrum and Asphalt Penetration data Storehouse, using topology predicting means it is achieved that quick detection to Asphalt Penetration.The method can measure Colophonium in 10 minutes Needle penetration, has good real-time and economy.

Method proposed by the present invention comprises the following steps：

A kind of method for quick of Asphalt Penetration, the near infrared spectrum data storehouse based on Colophonium for the method, by inciting somebody to action Near infrared technology and topology technology combine, and realize the quick detection to Asphalt Penetration；

The method of the present invention comprises the following steps：

(1), using near infrared spectrometer, several Colophonium sample is scanned obtaining the near infrared spectrum of each Colophonium, obtains The spectrum vector that absorbance under different characteristic wave number for the various Colophoniumes is formed is taken to be sample spectrum data；Set up Colophonium closely red External spectrum data base, records sample names, sample spectrum data and the sample needle penetration attribute of aforementioned several Colophonium；

(2), it is scanned obtaining the near infrared spectrum of this Colophonium using near infrared spectrometer to Colophonium sample to be measured, obtain The spectrum vector that absorbance under different characteristic wave number for this Colophonium is formed；

(3), calculate the distance between each sample spectrum in spectrum to be measured and Colophonium near infrared spectrum data storehouse, search Colophonium near It is less than the Asphalt Penetration data of a certain threshold value with the spectrum intervals of this spectrum to be measured in infrared spectrum storehouse, calculate Colophonium to be measured Needle penetration.

Total number of samples in the Colophonium near infrared spectrum data storehouse of the present invention should be not less than 150 groups, Colophonium near infrared spectrum Sample needle penetration in data base is obtained using traditional evaluation method.

In the step (3) of the present invention,

The formula calculating distance between each sample spectrum in spectrum to be measured and Colophonium near infrared spectrum data storehouse is as follows：

S is X_i ^TAnd Y_j ^TCovariance matrix, its element s_ijRepresent：

In formula, i represents the numbering of Colophonium to be measured, and j represents the Colophonium numbering in Colophonium near infrared spectrum data storehouse, d_ijFor treating Survey the spectrum intervals of a certain Colophonium in Colophonium and near infrared spectrum data storehouse；X_iAnd Y_jIt is respectively Colophonium to be measured and near infrared spectrum The spectrum vector that absorbance under different characteristic wave number for the library of spectra sample being compared in data base is formed；X_i ^TAnd Y_j ^TPoint Wei not vector X_iAnd Y_jTransposition；S is X_i ^TAnd Y_jCovariance matrix, S be p row p column matrix；P is the characteristic waves participating in calculating Sum；A is that Colophonium spectrum sample characteristic waves are numbered, a=1,2,3 ... p；For spectroscopic data x_1i, x_2i, x_3i…x_piEqual Value；For spectroscopic data y_1j, y_2j, y_3j…y_pjAverage；(X_i-Y_j)^TFor matrix X_i-Y_jTransposed matrix；S^-1For matrix S's Inverse matrix.

In the step (3) of the present invention, a certain when being less than with the spectrum intervals of this spectrum to be measured in Colophonium Near-infrared spectrum database When the Colophonium sample number of threshold value is more than the preset ratio of library of spectra total number of samples, Asphalt Penetration to be measured is calculated by following formula：

In formula, i represents the numbering of Colophonium to be measured, and j represents the Colophonium numbering in Colophonium near infrared spectrum data storehouse, and i, j take Integer more than zero；d_ijSpectrum intervals for a certain Colophonium in Colophonium to be measured and near infrared spectrum data storehouse；PI_iFor Colophonium to be measured Needle penetration；PI_jNeedle penetration for jth kind Colophonium sample；M is less than threshold value in Colophonium Near-infrared spectrum database with Colophonium distance to be measured Colophonium total sample number.

In the step (3) of the present invention, a certain when being less than with the spectrum intervals of this spectrum to be measured in Colophonium Near-infrared spectrum database The Colophonium sample number of threshold value be less than library of spectra total number of samples preset ratio when it is believed that predict the outcome unreliable, increase Colophonium near Sample size in infrared spectrum storehouse, repeat step (3).

In the present invention, threshold value is 5, and preset ratio is 5%.

In the increase Colophonium Near-infrared spectrum database of the present invention, the method for sample size is as follows, using linear plus and new to calculate Increase the spectroscopic data of sample, and the needle penetration newly increasing sample is calculated using blending rule, wherein, needle penetration mixing rule As follows：

lg(C_D)=C_1f×lg(C_1p)+C_2f×lg(C_2p)+…+C_qf×lg(C_qp)

In formula, C_DFor newly increasing mix asphalt sample needle penetration, q is the library of spectra sample number participating in hybrid, for big Integer in 1 and less than total sample number in Colophonium near infrared spectrum data storehouse, C₁, C₂…C_qFor participating in the library of spectra of hybrid Sample；C_1p,C_2p…C_qpFor participating in the library of spectra sample C of hybrid₁, C₂…C_qNeedle penetration；C_1f,C_2f…C_qfMixed for participating in Total library of spectra sample C calculating₁, C₂…C_qMixed proportion.

The invention has the beneficial effects as follows：

The present invention proposes a kind of Asphalt Penetration method for quick, and the method is based near infrared detection technology and topology Learn modeling technique, measure Asphalt Penetration in order to quick, meanwhile, can be by being processed to available sample in Colophonium library of spectra locating Reason, to increase sample point in library of spectra, improves detection adaptability.

In recent years, near infrared spectrum is widely used to the light-end products property analysis process such as crude oil and product oil.Closely red External spectrum mainly reflects the characteristic information of hydric group, and near infrared spectrum region, various hydric groups have certain times Spectral band belongs to, and therefore, the present invention is surveyed with structural information to the composition of hydrocarbon compound in Colophonium using near infrared technology Fixed, combination model is calculated Asphalt Penetration, and the method has the advantages that quick, accurate, reproducible, pollution-free.

For Asphalt Penetration, the matching being difficult to find out a property and spectrogram dependency is very high and the scope of application is very wide is public Formula.Therefore, present invention introduces topology modeling, with topological method research structure and associating between property, building topology model, Thus reaching the purpose of predicting of substance property.In recent years, topological development and its to chemical field infiltration be knot The research of structure property relationship provides strong help.

Brief description

Fig. 1 is Asphalt Penetration quick detection process overall procedure block diagram.

Fig. 2 is the Colophonium spectrogram to be measured of embodiment one.

Fig. 3 is the Colophonium spectrogram to be measured of embodiment two.

Specific implementation process

Below in conjunction with the accompanying drawings and concrete example, detailed calculating process and concrete operations flow process are provided, to do to the present invention Further illustrate.Implement the Colophonium near infrared spectrum data storehouse in example, include 150 groups of different pitches spectroscopic datas.This enforcement Example is implemented under premised on technical solution of the present invention, but protection scope of the present invention is not limited to following enforcements and calculates Example.

Embodiment 1

Specific implementation process is as follows：

1) gather a Colophonium sample to be measured, through near infrared spectrometer scanning, obtain Colophonium spectrogram to be measured, as shown in Figure 2.Horizontal seat It is designated as spectrum wave number, scope is from 4000cm^-1To 4800cm^-1；Vertical coordinate is absorbance.

2) above-mentioned Colophonium near infrared spectrum to be measured is compared one by one with the spectroscopic data in data base, the light between calculating two-by-two Spectrum distance is from spectrum intervals calculating process is as follows：

Reflect the size of Asphalt Penetration difference ability according to characteristic waves, choose and select characteristic waves in 4000～4500 models Absorbance in enclosing participates in spectrum intervals and calculates, and characteristic waves are spaced apart 25, and characteristic waves sum is 21.With Colophonium in library of spectra As a example sample spectrum Sphalt_001, calculate the Colophonium spectrum Sphalt_dc to be measured that numbering is 1 and sample spectrum in library of spectra The spectrum intervals of Sphalt_001.Absorbance such as table 1 institute under Colophonium sample characteristics wave number in Colophonium sample to be measured and library of spectra Show：

Absorbance under Colophonium sample (Sphalt_001) characteristic waves in table 1. Colophonium sample to be measured and library of spectra

SampleID	4000	4025	4050	4075	4100
						Sphalt_dc	0.03268563	0.03802782	0.04971625	0.05482151	0.05214073
Sphalt_001	0.03450295	0.03874355	0.04889621	0.05299251	0.0515429
						SampleID	4125	4150	4175	4200	4225
Sphalt_dc	0.05376484	0.05502375	0.05699742	0.05544399	0.05716617
						Sphalt_001	0.05419647	0.05687382	0.05841867	0.05723312	0.05882526
SampleID	4250	4275	4300	4325	4350
						Sphalt_dc	0.07496396	0.06748891	0.06300468	0.09211791	0.06530404
Sphalt_001	0.07665316	0.06595442	0.06124319	0.09341353	0.06682604
						SampleID	4375	4400	4425	4450	4475
Sphalt_dc	0.05051042	0.04377439	0.01988578	0.00851298	0.00502607
						Sphalt_001	0.04892384	0.03994466	0.01889637	0.00819711	0.00458475
SampleID	4500
						Sphalt_dc	0.00362275
Sphalt_001	0.0031375

The spectrum vector that absorbance under different characteristic wave number for the Colophonium sample Sphalt_dc to be measured is formed adopts X₁Table It is shown as：

X₁=[0.03268563 0.03802782 0.04971625 0.05482151 0.05214073 0.05376484 0.05502375

0.05699742 0.05544399 0.05716617 0.07496396 0.06748891 0.06300468

0.09211791 0.06530404 0.05051042 0.04377439 0.01988578 0.00851298

0.00502607]^T

The spectrum arrow that in library of spectra, absorbance under different characteristic wave number for the Colophonium sample spectrum Sphalt_001 is formed Amount adopts Y₁It is expressed as：

Y₁=[0.03450295 0.03874355 0.04889621 0.05299251 0.0515429 0.05419647 0.05687382

0.05841867 0.05723312 0.05882526 0.07665316 0.06595442 0.06124319

0.09341353 0.06682604 0.04892384 0.03994466 0.01889637 0.00819711

0.00458475]^T

Therefore, (X₁-Y₁)^T

=[- 0.00181732-0.00071572 0.00082005 0.00182899 0.00059783- 0.00043163 -0.00185007

-0.00142125 -0.00178912 -0.00165909 -0.00168920 0.00153448 0.00176150 -0.00129562

-0.00152200 0.00158659 0.00382974 0.00098941 0.00031588 0.00044133 0.00048525] MATLAB is adopted to calculate X₁ ^TAnd Y₁ ^TCovariance matrix S, method is as follows：

A=(X_i ^T；Y₁ ^T)

S=cov (A)

It is calculated the symmetrical matrix that S is 21 row 21 row, as shown in table 2.

Table 2 covariance matrix S

Using formulaCalculate in Colophonium spectrum Sphalt_dc to be measured and library of spectra The spectrum intervals of sample spectrum Sphalt_001, as follows using MATLAB calculation method：

It is calculated：d₁₁=3.6616

With reference to above-mentioned calculating process, calculate the light spectrum distance of Colophonium near infrared spectrum and Colophonium spectrum to be measured in library of spectra one by one From, and interpretation is less than the sample point quantity of effective spectrum intervals.Calculated by above formula, 10 can be found in library of spectra and (be more than The 5% of library of spectra total number of samples) with Colophonium sample spectrum distance to be measured less than threshold value 5 sample point.As shown in table 3：

Table 3 and Colophonium sample spectrum distance to be measured are less than the library of spectra sample point of threshold value

Colophonium sample number	Spectrum intervals	Colophonium sample number	Spectrum intervals
				Sphalt_001	3.6616	Sphalt_145	0.7240
Sphalt_010	2.2166	Sphalt_121	1.1162
				Sphalt_012	1.2721	Sphalt_105	2.7909
Sphalt_025	1.3287	Sphalt_029	1.6706
				Sphalt_038	1.6567	Sphalt_069	2.0568

Therefore, there is no need to increase sample, can directly pass through above-mentioned Colophonium sample, calculate Asphalt Penetration numerical value to be measured.

3) it is calculated as follows Asphalt Penetration to be measured：

Calculating process is as follows：

Above-mentioned 10 kinds of penetration of bitumens are as shown in table 4：

Table 4 and Colophonium sample spectrum distance to be measured are less than the library of spectra sample point needle penetration of threshold value

Colophonium sample number	Needle penetration	Colophonium sample number	Needle penetration
				Sphalt_001	95	Sphalt_145	87
Sphalt_010	99	Sphalt_121	88
				Sphalt_012	96	Sphalt_105	91
Sphalt_025	90	Sphalt_029	85
				Sphalt_038	98	Sphalt_069	87

Therefore, penetration of bitumens to be measured is：

Embodiment 2：

1) collection one Colophonium sample to be measured is numbered is 2.Through near infrared spectrometer scanning, obtain Colophonium spectrogram to be measured, as Fig. 3 institute Show.Abscissa is spectrum wave number, and scope is from 4000cm^-1To 4800cm^-1；Vertical coordinate is absorbance.

2) Colophonium spectrum to be measured is compared one by one with the spectroscopic data near infrared spectrum data storehouse, the light between calculating two-by-two Spectrum distance is from (calculating process is identical with the step 2 in embodiment 1).By calculating, library of spectra is only able to find 5 (less than light The 5% of spectrum storehouse total number of samples) sample point less than 5 with Colophonium sample spectrum threshold value to be measured.As shown in table 5：

Table 5 and Colophonium sample spectrum distance to be measured are less than the library of spectra sample point of threshold value

Colophonium sample number	Spectrum intervals	Colophonium sample number	Spectrum intervals
				Sphalt_015	1.6848	Sphalt_009	0.9641
Sphalt_103	3.6471	Sphalt_075	2.3156
				Sphalt_066	3.0245

Therefore, predict the outcome unreliable, sample size in library of spectra need to be increased；

3) existing Colophonium spectrum samples in library of spectra are increased, using linear plus with calculating the spectrum number newly increasing sample According to, and to calculate, using following blending rules, the needle penetration newly increasing sample.

lg(C_D)=C_1f×lg(C_1p)+C_2f×lg(C_2p)+…+C_qf×lg(C_qp)

Press 0.2,0.6,0.2 with tri- Colophonium samples of Sphalt_008, Sphalt_015, Sphalt_20 in library of spectra As a example mixed proportion is mixed, list calculating detailed process, as follows：

The spectrum vector that in library of spectra, absorbance under different characteristic wave number for the Colophonium sample Sphalt_008 is formed is adopted Use Y₈It is expressed as：

Y₈=[0.03454343 0.03876472 0.04891891 0.05301028 0.05154569 0.05419553

0.05688159 0.0584288 0.05724228 0.05881272 0.07665679 0.06594097

0.0612307 0.09339508 0.06681824 0.04888785 0.03992409 0.01888973

0.00819002 0.00458349 0.00313908]^T

The spectrum vector that in library of spectra, absorbance under different characteristic wave number for the Colophonium sample Sphalt_015 is formed is adopted Use Y₁₅It is expressed as：

Y₁₅=[0.03310126 0.03769716 0.04845486 0.05510251 0.05115978 0.05264836 0.05442039

0.05720746 0.05579442 0.05723187 0.07437372 0.06706857 0.06288219

0.09373545 0.06700707 0.0504078 0.04385249 0.02015411 0.00870699

0.00522539 0.00376815]^T

The spectrum vector that in library of spectra, absorbance under different characteristic wave number for the Colophonium sample Sphalt_20 is formed is used

Y₂₀It is expressed as：

Y₂₀=[0.03309589 0.03764875 0.04838402 0.05504122 0.05117614 0.05267337 0.05446153

0.05724254 0.05582455 0.05725101 0.07439601 0.06708532 0.06288583

0.09377839 0.06701277 0.05039995 0.043845 0.02013723 0.00869096 0.00521371

0.0037558]^T

In library of spectra, Colophonium sample Sphalt_008, Sphalt_015, Sphalt_020 press 0.2,0.6,0.2 mixing respectively After ratio mixing, the spectrum vector that the mixing sample Sphalt_d01 that the obtains absorbance under different characteristic wave number is formed is used Y_d01Calculated using following equation：

Y_d01=0.2 Y₈+0.6·Y₁₅+0.2·Y₂₀

=[0.03338862 0.037900993 0.048533504 0.054671805 0.051240234 0.052962796

0.054920859 0.057458744 0.056090019 0.057551865 0.074834793 0.066846401

0.062552622 0.093675964 0.066970443 0.050102241 0.043065312 0.019897856

0.008600392 0.005094674 0.003639863]^T

The needle penetration of library of spectra Colophonium sample Sphalt_008, Sphalt_015, Sphalt_020 is respectively 87,89,91, Sample Sphalt_008, Sphalt_015, Sphalt_020 carry out after proportioning by 0.2,0.6,0.2 mixed proportion respectively, newly The needle penetration calculating process increasing sample Sphalt_d01 is as follows：

lg(C_D)=C_1f×lg(C_1p)+C_2f×lg(C_2p)+…+C_qf×lg(C_qp)

lg(PI_d01)=0.2 × lg87+0.6 × lg89+0.2 × lg91=1.9493

It is calculated：PI_d01=88.9910

Equally, in this way the sample in library of spectra is calculated with different ratio, increased Colophonium sample in library of spectra This number increases to 200 groups.

4) Colophonium spectrum to be measured is compared one by one with the spectroscopic data in the near infrared spectrum data storehouse increasing after sample point, Spectrum intervals between calculating two-by-two.By calculate, can find in library of spectra 12 (more than the 5% of library of spectra total number of samples) with The sample point that Colophonium sample spectrum threshold value to be measured is less than 5.As shown in table 6：

Therefore, sample size increases, and predict the outcome reliability, can pass through above-mentioned Colophonium sample, calculates Asphalt Penetration to be measured Numerical value.

5) it is calculated as follows Asphalt Penetration to be measured：

Calculating process is as follows：

Above-mentioned 12 kinds of penetration of bitumens are as shown in table 7：

Table 7 and Colophonium sample spectrum distance to be measured are less than the library of spectra sample point needle penetration of threshold value

Colophonium sample number	Needle penetration	Colophonium sample number	Needle penetration
				Sphalt_015	89	Sphalt_009	95
Sphalt_103	90	Sphalt_075	93
				Sphalt_066	88	Sphalt_d07	87
Sphalt_d28	90	Sphalt_d176	91
				Sphalt_d141	92	Sphalt_d322	90
Sphalt_d264	87	Sphalt_d227	86

Therefore, penetration of bitumens to be measured is：

Claims (4)

1. a kind of method for quick of Asphalt Penetration it is characterised in that the method be based on Colophonium near infrared spectrum data storehouse, By combining near infrared technology and topology technology, realize the quick detection to Asphalt Penetration；

The method comprises the following steps：

(1), using near infrared spectrometer, several Colophonium sample is scanned obtaining with the near infrared spectrum of each Colophonium, obtains each The spectrum vector that kind absorbance under different characteristic wave number for the Colophonium is formed is sample spectrum data；Set up Colophonium near infrared light Modal data storehouse, records sample names, sample spectrum data and the sample needle penetration attribute of aforementioned several Colophonium；

(2), it is scanned obtaining the near infrared spectrum of this Colophonium using near infrared spectrometer to Colophonium sample to be measured, obtain this drip The spectrum vector that blue or green absorbance under different characteristic wave number is formed；

(3), the distance between each sample spectrum in computation and measurement spectrum and Colophonium near infrared spectrum data storehouse, searches Colophonium near-infrared It is less than the Asphalt Penetration data of a certain threshold value with the spectrum intervals of this measure spectrum, the pin calculating Colophonium to be measured enters in library of spectra Degree；

In step (3), when in Colophonium Near-infrared spectrum database, the spectrum intervals with this measure spectrum is less than the Colophonium sample of a certain threshold value When this number is more than the preset ratio of library of spectra total number of samples, Asphalt Penetration to be measured is calculated by following formula：

$\lg \left({\mathrm{PI}}_i\right)=\underset{j=1}{\overset{m}{\Σ}}(\frac{1}{d_{ij}}\×\lg ({\mathrm{PI}}_j)/\underset{j=1}{\overset{m}{\Σ}}\frac{1}{d_{ij}})$

In formula, i represents the numbering of Colophonium to be measured, and j represents the Colophonium numbering in Colophonium near infrared spectrum data storehouse, and i, j take and be more than Zero integer；d_ijSpectrum intervals for a certain Colophonium in Colophonium to be measured and Colophonium near infrared spectrum data storehouse；PI_iFor Colophonium to be measured Needle penetration；PI_jNeedle penetration for jth kind Colophonium sample；M is less than threshold value in Colophonium Near-infrared spectrum database with Colophonium distance to be measured Colophonium total sample number；

In step (3), when in Colophonium Near-infrared spectrum database, the spectrum intervals with this measure spectrum is less than the Colophonium sample of a certain threshold value This number be less than library of spectra total number of samples preset ratio when it is believed that predict the outcome unreliable, increase Colophonium Near-infrared spectrum database in Sample size, repeat step (3)；

The method increasing sample size in Colophonium Near-infrared spectrum database is as follows, and the ratio according to blending rule adopts linear plus sum Mode newly increases the spectroscopic data of sample to calculate, and wherein, needle penetration blending rule is as follows：

lg(C_D)=C_1f×lg(C_1p)+C_2f×lg(C_2p)+…+C_qf×lg(C_qp)

In formula, C_DFor newly increasing the needle penetration of mix asphalt sample, q is the library of spectra sample number participating in hybrid, is more than 1 And the integer less than total sample number in Colophonium near infrared spectrum data storehouse, C₁, C₂…C_qFor participating in the library of spectra sample of hybrid This；C_1p,C_2p…C_qpFor participating in the library of spectra sample C of hybrid₁, C₂…C_qNeedle penetration；C_1f,C_2f…C_qfFor participating in mixing The library of spectra sample C calculating₁, C₂…C_qMixed proportion.

2. a kind of method for quick of Asphalt Penetration according to claim 1, is characterized in that Colophonium near infrared spectrum Total number of samples in data base should be not less than 150 groups, and the sample needle penetration in Colophonium near infrared spectrum data storehouse is commented using tradition Valency mode obtains.

3. the method for quick of a kind of Asphalt Penetration according to claim 1, is characterized in that in step (3), calculates In measure spectrum and Colophonium near infrared spectrum data storehouse, between each sample spectrum, the formula of distance is as follows：

$d_{ij}=\sqrt{{(X_i-Y_j)}^T\·S^{-1}\·(X_i-Y_j)}$

$\begin{array}{cc}{X}_i=\left[\begin{array}{c}{x}_{1i}\\ x_{2i}\\ \·\\ \·\\ \·\\ x_{pi}\end{array}\right]& Y_j=\left[\begin{array}{c}{y}_{1j}\\ y_{2j}\\ \·\\ \·\\ \·\\ y_{pj}\end{array}\right]\end{array}$

S is X_i ^TAnd Y_j ^TCovariance matrix, its element s_ijRepresent：

$s_{ij}=\frac{\underset{k=1}{\overset{p}{\Σ}}(x_{ki}-\stackrel{\‾}{x_i})(y_{kj}-\stackrel{\‾}{y_j})}{p-1}$

$\begin{array}{cc}\stackrel{\‾}{x_i}=\frac{1}{p}\underset{a=1}{\overset{p}{\Σ}}{x}_{ai}& \stackrel{\‾}{y_j}=\frac{1}{p}\underset{a=1}{\overset{p}{\Σ}}{y}_{aj}\end{array}$

In formula, i represents the numbering of Colophonium to be measured, and j represents the Colophonium numbering in Colophonium near infrared spectrum data storehouse, d_ijFor drip to be measured The blue or green spectrum intervals with a certain Colophonium in Colophonium near infrared spectrum data storehouse；X_iAnd Y_jIt is respectively Colophonium to be measured and Colophonium near-infrared The spectrum vector that absorbance under different characteristic wave number for the library of spectra sample being compared in spectra database is formed；X_i ^TWith Y_j ^TIt is respectively vector X_iAnd Y_jTransposition；S is X_i ^TAnd Y_j ^TCovariance matrix, S be p row p column matrix；P is the spy participating in calculating Levy the sum of wave number；A is that Colophonium spectrum samples characteristic waves are numbered, a=1,2,3 ... p；For spectroscopic data x_1i, x_2i, x_3i… x_piAverage；For spectroscopic data y_1j, y_2j, y_3j…y_pjAverage；(X_i-Y_j)^TFor matrix X_i-Y_jTransposed matrix；S^-1For square The inverse matrix of battle array S.

4. the method for quick of a kind of Asphalt Penetration according to claim 1, is characterized in that threshold value is 5, presets ratio Example is 5%.