CN100498293C - Method for measuring content of dialkene in C10-C13 positive formation hydrocarbon through spectrum of infrared light - Google Patents

Method for measuring content of dialkene in C10-C13 positive formation hydrocarbon through spectrum of infrared light Download PDF

Info

Publication number
CN100498293C
CN100498293C CN 200410102808 CN200410102808A CN100498293C CN 100498293 C CN100498293 C CN 100498293C CN 200410102808 CN200410102808 CN 200410102808 CN 200410102808 A CN200410102808 A CN 200410102808A CN 100498293 C CN100498293 C CN 100498293C
Authority
CN
China
Prior art keywords
spectrum
sample
diene content
infrared
model
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN 200410102808
Other languages
Chinese (zh)
Other versions
CN1796980A (en
Inventor
张继忠
袁洪福
王京华
褚小立
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Sinopec Research Institute of Petroleum Processing
China Petroleum and Chemical Corp
Original Assignee
Sinopec Research Institute of Petroleum Processing
China Petroleum and Chemical Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Sinopec Research Institute of Petroleum Processing, China Petroleum and Chemical Corp filed Critical Sinopec Research Institute of Petroleum Processing
Priority to CN 200410102808 priority Critical patent/CN100498293C/en
Publication of CN1796980A publication Critical patent/CN1796980A/en
Application granted granted Critical
Publication of CN100498293C publication Critical patent/CN100498293C/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Abstract

The present invention relates to a method for measuring diene content in C10 to C13 normal hydrocarbon by an infrared spectrum, which comprises the following steps that representative C10 to C13 normaThe invention is a method for determining the diolefin contents of the C10~C13 normal hydrocarbons with infrared spectrum, comprising the steps of: selecting representative C10~C13 normal hydrocarbon l hydrocarbon samples are selected, and the diene content thereof is measured by a standard method; then, the medium infrared or long-wave near-infrared spectrum which each sample corresponds to, 920 samples, determining their diolefin contents by standard method; then determining a medium infrared or long wave near infrared spectrum of each sample, selecting 920~890cm<-1>, 1000~980cm<-1> and 800~to 890cm<-1>, 1000 to 980cm<-1> and 800 to 1100cm<-1> of the medium infrared spectrum or 6000 to 6300cm<-1> of long-wave near-infrared spectrum is selected as a characteristic spectrum region, and the1100cm<-1> wave bands of the medium infrared spectrums or a 6000~6300cm<-1> wave band of the long wave near infrared spectrum as characteristic spectrum regions and relating the response values of the response value of characteristic spectrum region is correlated with the diene content of the samples measured by the standard method to establish a correction equation or model; and finally, the dien characteristic spectrum regions with the diolefin contents of the samples determined by standard method to establish a correcting equation or model; And finally forecasting the diolefin content of ane content of an unknown sample is forecast with the infrared spectrum of the unknown sample by the correction equation or model. The method is suitable for measuring the diene content in the C10 to C1 unknown sample with its infrared spectrum through the correcting equation or model. The method is convenient to operation, and accurate and quick to analyze. 3 normal hydrocarbon, and has the advantages of rapid and exact analysis and easy operation.

Description

Infrared spectrum measurement C 10~C 13The method of diene content in the positive structure hydrocarbon
Technical field
The present invention is a kind of method with diene content in the infrared spectrum measurement potpourri, specifically, is a kind of infrared spectrum measurement C that uses 10~C 13The method of diene content in the positive structure hydrocarbon.
Background technology
Development along with computer technology and Chemical Measurement, the technical characterstic that Infrared Spectrum Technology is fast with its analysis speed, cost is low, do not destroy and consumed sample begins to be applied at petrochemical industry, has brought into play positive role in the optimal control of petroleum refining and processing.Will in infrared and near infrared spectrum be used for octane number, arene content, olefin(e) centent, benzene content, the research of mensuration such as diesel cetane-number, density, boiling range, solidifying point, flash-point has reported in literature successively, but infrared spectrum is used for C 10~C 13The mensuration of diolefin does not still have report in the N-alkanes alkene.
In the washing agent chemical plant, the raw material of producing alkyl benzene is C 10~C 13N-alkane, this n-alkane generates corresponding alkene after dehydrogenation, contain 10~13% alkene in the dehydrogenation product approximately.This dehydrogenation afterproduct is carried out alkylation, make wherein alkene and benzene reaction generate alkyl benzene, isolated alkane loops back dehydrogenation unit from product.Above-mentioned normal paraffin dehydrogenation reaction product also comprises 1.0~1.5% C 10~C 13Diolefin, wherein conjugated diene can generate accessory substances such as indane, tetralin, hexichol alkane in alkyl plant, influences the alkyl benzene quality.Therefore, before alkylation, to carry out selective hydrogenation and handle, remove conjugated diene wherein, remain about 0.2~0.5% non-conjugated diene hydrocarbon dehydrogenation feed.Therefore, will measure in the selective hydrogenation device diene content in the material in time, exactly in production control, to monitor in real time, the control diolefin is converted into the efficient of mono-olefin.
Existing mensuration C 10~C 13The analytical approach of diene content is UOP-869 in the long-chain alkane alkene, i.e. high performance liquid chromatography (HPLC) method.This method adopts two silicagel column polyphones, and moving phase is isooctane, and the differential detecting device utilizes the area normalization method of band correction factor to calculate diene content.But the service time of chromatographic column and the minor amount of water in the moving phase are bigger to the influence of analysis result precision in this method, and analysis time is also longer.Therefore can not be in time, provide the diene content data for production control exactly.
Summary of the invention
The purpose of this invention is to provide a kind of infrared spectrum measurement C that uses 10~C 13The method of diene content in the positive structure hydrocarbon, this method is measured accurately, and is easy and simple to handle, quick.
Infrared spectrum measurement C provided by the invention 10~C 13The method of diene content in the positive structure hydrocarbon comprises:
(1) sets up correction equation or model: choose representational C 10~C 13Positive structure hydrocarbon samples is measured its diene content with standard method, and then measure each sample correspondence in infrared or long wavelength near infrared spectrometry, select 920~890cm in the middle infrared spectrum -1With 1000~980cm -1, 800~1100cm -1Or 6000~6300cm of long wavelength near infrared spectrometry -1Be characteristic spectrum area, the diene content that the response of described characteristic spectrum area and sample standard method are measured is associated and sets up correction equation or model,
(2) measure the unknown sample diene content: under the test condition same with setting up correction equation or model the test unknown sample in infrared or long wavelength near infrared spectrometry, with middle infrared spectrum 920~890cm -1With 1000~980cm -1, 800~1100cm -1Or long wavelength near infrared spectrometry 6000~6300cm -1The corresponding correction equation of response substitution or the model in spectrum district obtain the diene content in the unknown sample.
The inventive method is at the selected expression of middle infrared range or long wave near-infrared spectra district C 10~C 13The characteristic spectrum of the diene that contains in the positive structure hydrocarbon is set up correction equation or model with the metrology method, has obtained predicting the outcome accurately.Compare with the standard testing method, shortened the analytical test time of sample greatly, step also simplifies the operation simultaneously.
Description of drawings
Fig. 1 is the correlogram that infrared ATR spectrum and UOP-869 method are measured diolefin in the validation-cross process.
Fig. 2 is the correlogram that long wavelength near infrared spectrometry and UOP-869 method are measured diolefin in the validation-cross process.
Embodiment
The inventive method adopts the infrared spectrum prediction C of sample 10~C 13Diene content in the positive structure hydrocarbon, the spectrum that is used to predict can be middle infrared spectrum or long wavelength near infrared spectrometry, response by characteristic spectrum area relevant with diene content in selected spectrum district is set up predicted correction equation or model, and then the diene content of prediction unknown sample.Method is easy and simple to handle, analysis is rapid, accurately.
Infrared spectrum is that the vibration-rotational energy level transition owing to molecule produces, normal position of using wave number to represent absorption band in the Infrared spectroscopy, and wave number is the number of included ripple in the wavelength of every cm distance.Traditionally, be wavelength 2500~25000nm (wave number 4000~400cm often -1) the spectrum district be called in infrared (being called for short infrared) district, be wavelength 780~2500nm (wave number 12820~4000cm -1) the spectrum district be called the near-infrared region.The fundamental frequency of most organic compounds and many mineral compound molecular vibrations all appears at region of ultra-red, and this qualitative analysis and micro constitutent analysis for organic constitution is very effective.And occur more between the near-infrared region is the frequency multiplication and the sum of fundamental frequencies of molecular chemistry key chattering, and it is much weak that the relative fundamental frequency of intensity is wanted, and spectrum that this zone forms is called near infrared spectrum.Usually with this scope division again near infrared shortwave district (780~1100nm, wave number 12820~9090cm -1) and near infrared long-wavelength region (1100~2500nm, wave number 9090~4000cm -1).Because the near infrared spectrum extinction coefficient is significantly low than mid-infrared spectral, thus macro-analysis only be applicable to, but near infrared ray is comparatively easy, is applicable to that more the on-line analysis in the production run is used.
C 10~C 13The variation of diene content in the N-alkanes alkene, make the infrared spectrum of sample change thereupon, C=C key stretching vibration wherein and=stretching vibration of c h bond, deformation vibration in absorption intensity generation respective change in the infrared and near infrared spectrum, thereby can in the zone that changes of infrared and near infrared spectrum generation absorption intensity be characteristic spectrum area, by the diene content of response prediction sample in this spectrum district.
Characteristic spectrums infrared and two spectrum districts of long wave near infrared were predicted the content of sample diolefin during the present invention selected for use.When adopting middle infrared spectrum prediction sample diene content, can adopt two kinds of method predictions of multiple linear regression and partial least square method diene content.
Adopt the method for multiple linear regression method prediction diene content to be: with 920~890cm in the described middle infrared spectrum -1With 1000~980cm -1The spectrum district is characteristic spectrum area, and spectrum area or the second-order differential spectrum area with described two characteristic spectrum areas is response respectively, and the diene content of measuring with the sample standard method is associated, and sets up the multiple linear regression correction equation.
Adopt the method for partial least square method method prediction diene content to be: with 800~1100cm -1During for characteristic spectrum area, be response with the absorbance of handling through second-order differential in this spectrum district, the diene content of measuring with the sample standard method is associated, and sets up calibration model with partial least square method.
Because C 10~C 13Positive structure hydrocarbon is stronger in the absorption of mid infrared region, when measuring its middle infrared spectrum, should adopt multiple attenuated total reflection annex (ART) to measure infrared spectrum, i.e. the present invention predicts the preferred multiple attenuated total reflection spectrum of the used middle infrared spectrum of the diene content of sample.ATR is one of mid-infrared light spectrometry annex of using always, is used for the measurement of liquid, powder and film more.Its measuring principle is: after a branch of radiant light that is sent by light source entered the crystal with high index, if incident angle is greater than critical angle on the interface of crystal and sample, then this light reflected on the interface almost completely thus.If the sample that contacts with crystal, then penetrates the following radiation beam of reflecting surface to the selective absorption of radiation because of being weakened by absorption of sample, thereby makes the total reflection wire harness be attenuated at the energy of these wavelength, the gained reflectance spectrum also has the characteristics of absorption spectra.The energy variation of a total reflection is little, and the absorption band of gained spectrum is very weak.If increase the number of times of total reflection, then can make the absorption band of reflection strengthen multiple attenuated total reflection that Here it is.
In the inventive method, when adopting long wavelength near infrared spectrometry prediction sample diene content, because of the absorbance in the spectrum district a little less than, should adopt partial least square method to set up calibration model.The method of setting up calibration model is: with long wavelength near infrared spectrometry 6000~6300cm -1The spectrum district is characteristic spectrum area, is response with the absorbance of handling through second-order differential in this spectrum district, and the diene content of measuring with the sample standard method is associated, and sets up calibration model with partial least square method.
The standard method of the inventive method working sample diene content is a high performance liquid chromatography.The diene content of working sample is 0.1~1.7 quality %.
Be the accuracy of testing model, generally will be divided into calibration set and checking collection with the sample that diene content is measured in standard method.The calibration set sample size is more, and representative, and promptly the diene content of calibration set sample should be contained all pre-diene contents of measuring.The checking collection is then randomly drawed, and its sample as unknown sample, is verified the accuracy of calibration model.Checking collection sample size is less, is about about 1/3 of specimen total quantity.
The above-mentioned basic skills of setting up the used multiple linear regression of correction equation is as follows:
Selected p independent variable, m measurement result are set up following system of equations:
y 1=b 0+b 1X 11+b 2X 12+...+b pX 1p
y 2=b 0+b 1X 21+b 2X 22+...+b pX 2p
.......................................
Y m=b 0+b 1X m1+b 2X m2+...+b pX mp
Write above system of equations as following matrix form:
Y=XB
B=(X ' X) then -1X ' Y
In the formula, X ' is an X transpose of a matrix matrix, (X ' X) -1Be (the inverse matrix of X ' X).
Column vector wherein
B = b 0 b 1 &CenterDot; &CenterDot; &CenterDot; b p
Be the regression coefficient of regression equation, just can set up multiple linear regression ten thousand journeys by regression coefficient.
In the inventive method, law of Beer is expressed as C=PA λ+ E sets up the regression equation of predicting concentration of component with characteristic spectrum with this.In the described law of Beer expression formula, C is a concentration of component, A λBe the absorbance of af at wavelength lambda, also can represent the area of certain SPECTRAL REGION, E is the residual error between predicted value and actual value.Concentration of component with m mensuration is set up the Y matrix, sets up the X matrix with the area of characteristic spectrum area, can set up regression equation with above-mentioned multiple linear regression analysis method.
The method that described partial least square method (PLS) is set up calibration model is summarized as follows:
(1) law of Beer is expressed as C=PA+E, sets up the relation of absorbance matrix and concentration of component matrix thus, it is as follows to make matrix decomposition:
A n×p=T n×fP f×p+E A
C n×m=U n×fQ f×m+E C
Wherein, T and U are respectively the sub matrix that gets of spectrum matrix and concentration matrix, and P and Q are respectively load (the being major component) matrix of spectrum matrix and concentration matrix, E AAnd E CThe residual matrix of introducing when being respectively with PLS match A and C, n is a sample number, the p number of wavelengths, m is a number of components, f is a number of principal components.
(2) T and U are done linear regression
U n×f=T n×fB f×f
Wherein, B is the correlation coefficient matrix.
Try to achieve the B matrix value, can try to achieve its concentration by the spectrum of unknown sample through the following steps.
At first by the spectrum matrix A of sample UnknownThe P that obtains with correction Proofread and correctObtain the T of sample spectra Unknown, ask the concentration of unknown sample again by following formula
C Unknown=T UnknownBQ (1)
During actual computation, data matrix decomposed and return and do a step, the decomposition that is spectroscopic data and concentration data is carried out simultaneously, and concentration information is incorporated in the spectroscopic data decomposable process, before new major component of every calculating, spectrum score and concentration score are exchanged, make the spectrum major component spectrum that obtains directly relevant with analyzed concentration of component.
The PLS algorithm is summarized as follows:
(1) modelling (correction)
In the computation process, the transposed matrix of certain matrix D of D ' expression, D -1Represent its inverse matrix.Earlier spectroscopic data and concentration data are all carried out the centralization processing.Descend column count then set by step:
1. row of getting the concentration Matrix C are designated as c, give the decomposition vector of matrix U as the primary iteration vector assignment of decomposing:
u=c
2. replace the decomposition vector t of T matrix to calculate the decomposition vector v of P matrix with u
v′=uA(u′u) -1
One row of above-mentioned two step working concentration Matrix C decompose as the quadrature that the primary iteration vector carries out the A matrix, have introduced concentration information when promptly decomposing the A matrix.
3. standardization: v '=v ' | v ' | -1
4. by v ' compute vectors t:t=Av (v ' v) -1
5. replace the decomposition vector u of U matrix to calculate the decomposition vector q of Q matrix with t
q′=t′C(t′t) -1
This step uses A matrix decomposition vector t that the C matrix is decomposed, and has promptly introduced spectral information when the C matrix decomposition.
6. standardization: q '=q ' | q ' | -1
7. by the decomposition vector u of q ' calculating U matrix:
u=Cq(q′q) -1
8. judge whether s restrains, i.e. ‖ t Previous round-t Take turns back one‖ (t Previous round-t Take turns back oneNorm) whether less than a certain constant θ.If convergence then to the 9. step, otherwise get back to the 2. step, proceed to decompose iteration with the u that 7. obtains.
9. calculate the internal relations between t and the u, carry out regressing calculation: b=u ' t (t ' t) -1
The above-mentioned vector that respectively decomposes is designated as t all corresponding to first principal component 1, v 1, u 1, q 1, b 1Calculate the residual error battle array then:
E A=A-t 1v′
E C=C-u 1q 1′=C-b 1t 1q 1
Then with E A, E CReplace matrix A and C respectively, return above-mentioned next major component of 1. calculating, obtain corresponding vector t 2, v 2, u 2, q 2, b 2, and the like, till the whole major components that calculate A and C.
2, prediction and calculation
To unknown sample, spectrum is A Unknown, according to concerning A Unknown=T UnknownThe P ' that P ' and timing store calculates T Unknown, the B vector sum matrix Q that stores according to timing can calculate C by formula (1) again Unknown
3, definite method of best major component
The above-mentioned spectrum of rebuilding with the major component and the match of must assigning to, Utopian is that the spectrum of rebuilding comprises whole spectral informations of sample and do not contain any noise.It is very few to get number of principal components, and not enough to rebuilding spectrum generation match, number of principal components is too much, then produces overfitting to rebuilding spectrum, introduces noise.Cross verification commonly used is determined number of principal components.Promptly the sample actual value of being measured by standard method is formed the concentration Matrix C, with calibration model predicted value predicted composition concentration Matrix C pThen its residual sum of squares (RSS) (PRESS) is defined as follows:
PRESS = &Sigma; i = 1 n &Sigma; j = 1 m ( C p , j - C i , j ) 2
In the formula, n is the correcting sample number, and m is analyzed number of components.
The number of principal components f of minimum PRESS correspondence *Be not best number of principal components, often cause overfitting.Usually determine number of principal components with F statistic law or manual analysis.The F statistic law is as follows:
F(f)=PRESS(f)/PRESS(f *)
Best f compares f *Little, f should be as far as possible little and satisfies F (f)<F α, β, β(α=0.25, β is a degree of freedom).
With the calibration model that multivariate calibration methods is set up, evaluation statistical parameter commonly used is as follows:
1, model tuning evaluating
SEC = &Sigma; i = 1 n ( y i - y ^ i ) 2 / ( n - 1 )
R = 1 - ( &Sigma; i = 1 n ( y i - y ^ i ) 2 / &Sigma; i = 1 n ( y i - y &OverBar; ) 2 )
In the formula:
SEC---calibration set sample standard deviation;
R---calibration set sample measured value and validation-cross predicted value related coefficient;
y i---the measured value of i sample in the calibration set;
Figure C200410102808D0008153431QIETU
---the predicted value of i sample in the calibration set;
Y---the mean value of calibration set all samples measured value;
N---calibration set sample number.
2, modelling verification evaluating
SEP = &Sigma; i = 1 m ( y i - y ^ i ) 2 / ( m - 1 )
R = 1 - ( &Sigma; i = 1 n ( y i - y ^ i ) 2 / &Sigma; i = 1 n ( y i - y &OverBar; ) 2 )
In the formula:
SEP---checking collection sample standard deviation;
R---checking collection sample measured value and model predication value related coefficient;
y i---the measured value of i sample is concentrated in checking;
Figure C200410102808D0008153431QIETU
---the predicted value of i sample is concentrated in checking;
Y---the mean value of checking collection all samples measured value;
M---checking collection sample number.
3, test of hypothesis
Suppose system error-free between spectroscopic analysis methods result and the standard method result (being that the result that infrared spectrum analysis obtains is accurately), do not have marked difference between the difference mean value d and 0 between the two analytical approach measured values so, i.e. d=0.Calculate the t test statistics:
t = d &OverBar; - 0 S d / m
In the formula: S d---the standard deviation of checking collection sample residual; Described residual error is the poor of sample determination value and predicted value;
D---checking collection sample residual average;
M---checking collection number of samples.
To certain confidence level α, have | t|<t (α, m-1), illustrate that hypothesis is correct, two method measurement result unanimities.
4, repeatability is calculated
The standard deviation of replication value: &sigma; i = &Sigma; j = 1 l i ( y ^ ij - y ^ &OverBar; i ) 2 / ( l i - 1 )
Mean standard deviation: &sigma; = &Sigma; i = 1 k l i &sigma; i 2 / l
In the formula:
Figure C200410102808D00101
---i sample l iI measurement result in the inferior duplicate measurements;
Figure C200410102808D00102
---the mean value of i sample measurement value;
l i---the mensuration number of times of i sample;
L---always measure number of times;
K---sample number.
When all variances of duplicate measurements belonged to same parent, the mean standard deviation of available duplicate measurements was as the repeatability of the repeated deviation calculation method of method, that is:
r = t &times; 2 &sigma;
In the formula: r---repeated result;
T---under the replication degree of freedom, the critical value that the t of bilateral 95% degree of confidence distributes.
The inventive method is applicable to measures C 10~C 13Diene content in the long-chain normal alkane alkene can be used middle infrared spectrum during test, also available long wavelength near infrared spectrometry.Middle infrared spectrum is suitable for the experimental determination diene content, and long wavelength near infrared spectrometry is suitable for carrying out in the production run on-line analysis.
Below by example in detail the present invention, but the present invention is not limited to this.
The standard method of working sample diene content is the UOP-869 method in the example, promptly uses high performance liquid chromatography (HPLC) to measure C 10~C 13Diene content in the alkane olefin liquid that dehydrogenation of long-chain alkane generates.Adopt two silicagel column polyphones during this method, moving phase is isooctane, and the differential detecting device utilizes the area normalization method of band correction factor to calculate diene content, and represents with massfraction.
The Nicolet 560 mid-infrared light spectrometers that the middle infrared spectrum of sample adopts U.S. Nicolet company to produce, the multiple attenuated total reflection of 0002-105 type (ATR) annex: ZnSe crystal, 45 ° of incident angles.Assay method is as follows: with dropper sample is dripped on the atr crystal, make it cover the whole of crystal, cover lid scans its spectrum, scanning times 16 times fast rapidly.Adopt spectral limit: 4000~650cm -1Resolution: 4cm -1
The Bomem 160 long wavelength near infrared spectrometry instrument that the long wavelength near infrared spectrometry of sample adopts Canadian Bomem company to produce.Assay method is: testing sample is poured in the quartz sample pool that light path is 5mm, and putting into the sample cell frame was that reference carries out spectral scan with the air after 2 minutes, and scanning times is 16 times.Adopt spectral limit: 10000~5000cm -1Resolution: 4cm -1
Example 1
This example is set up regression equation with multiple linear regression method, is used to predict the diene content of sample.
Choose the C of detergent factory 10~C 13120 in N-alkanes alkene sample is measured its diene content with the UOP-869 method, gets 82 representational samples and forms calibration set, is used to set up regression equation, and remaining 38 samples are formed the checking collection, are used for the check of regression equation forecasting accuracy.
(1) sets up multiple regression equation
Respectively with 920~890cm in the infrared ATR spectrum in the calibration set sample -1With 1000~980cm -1The spectrum area and the second-order differential spectrum area in spectrum district are independent variable, are that dependent variable is set up multiple regression equation with the UOP-869 method measured value of sample.
With the spectrum area is that the multiple regression equation that independent variable is set up is:
Y 2=8.377148X 10+30.23895X 11+0.462525 (2)
In the formula: Y 2---diene content, quality %,
X 10---1000~980cm -1Interval spectrum area,
X 11---920~890cm -1Interval spectrum area.
With second-order differential spectrum area is that the multiple regression equation that independent variable is set up is:
Y 2=-168.39973X 12-48.288308X 13-0.052902 (3)
In the formula: X 12---1000~980cm -1Interval second-order differential spectrum area,
X 13---920~890cm -1Interval second-order differential spectrum area.
(2) checking regression equation accuracy
Measure infrared ATR spectrum in the checking collection sample, determine that it is at 920~890cm -1With 1000~980cm -1The spectrum area and the second-order differential spectrum area in spectrum district, substitution formula (2) and (3) respectively, the comparative result of the sample diene content predicted value that is obtained by regression equation and the measured value of UOP-869 method sees Table 1.
As shown in Table 1, all less by predicted value deviation average (d) and standard deviation (SEP) that spectrum area and the multiple regression of second-order differential spectrum area obtain, illustrate that the predicted value of two kinds of methods and UOP-869 method measured value are all comparatively approaching.
Example 2
This example partial least square method, with in the absorbance in infrared special card spectrum district and the measured value of UOP-869 method set up calibration model, be used to predict the diene content of sample.
(1) sets up calibration model
Get calibration set sample in 63 examples 1, measure wherein infrared ATR spectrum, get 800~1100cm in the spectrogram -1The absorbance that the spectrum district handles through second-order differential is associated with the measured value of UOP-869 method, sets up calibration model with partial least square method, and the model major parameter sees Table 2, and the validation-cross correlogram is seen Fig. 1.
(2) checking calibration model accuracy
Measure checking collection sample in the example 1 in infrared ATR spectrum, determine 800~1100cm in the spectrogram -1The absorbance that the spectrum district handles through second-order differential, the substitution calibration model obtains the predicted value of sample in olefin(e) centent.Model is mainly proofreaied and correct and be the results are shown in Table 2, and checking the results are shown in Table 3.Table 3 shows, is calculated by checking collection sample diene content | the t| value is less than the critical value t of t under the given level of significance (α=0.05) 0 05(n), promptly in there was no significant difference between infrared ATR spectroscopic analysis methods and the UOP-869 method, illustrate that the diene content that infrared ATR spectrographic technique is measured in adopting is accurately.
(3) replica test
2 samples (a and b) that any selection has different diolefin levels, infrared ATR spectrum correction model carries out repeatedly repetition measurement in the employing, the results are shown in Table 4.As shown in Table 4, the repeatability of the inventive method predicted value is better than the UOP-869 method.
Example 3
This example partial least square method is set up calibration model with the absorbance in the special card spectrum of long wave near infrared district and the measured value of UOP-869 method, is used to predict the diene content of sample.
(1) sets up calibration model
Get calibration set sample in 63 examples 1, measure its long wavelength near infrared spectrometry, get 6000~6300cm in the spectrogram -1The absorbance that the spectrum district handles through second-order differential is associated with the measured value of UOP-869 method, sets up calibration model with partial least square method, and the model major parameter sees Table 5, and the validation-cross correlogram is seen Fig. 2.
(2) checking calibration model accuracy
Measure the long wavelength near infrared spectrometry of checking collection sample in the example 1, determine 6000~6300cm in the spectrogram -1The absorbance that the spectrum district handles through second-order differential, the substitution calibration model obtains the predicted value of sample in olefin(e) centent.Model is mainly proofreaied and correct and be the results are shown in Table 5, and checking the results are shown in Table 6.Table 6 result shows, is calculated by checking collection sample diolefin | the t| value is all descended the critical value (t of t less than given level of significance (α=0.05) 0 05(n)=2.02), promptly there was no significant difference between long wavelength near infrared spectrometry analytical approach and the UOP-869 method illustrates that the diene content that adopts long wavelength near infrared spectrometry to measure is accurately.
(3) replica test
2 samples (a and b) that any selection has different diolefin levels adopt the infrared spectrum calibration model of long wave to carry out repeatedly repetition measurement, the results are shown in Table 7.As shown in Table 7, the repeatability of the inventive method predicted value is better than the UOP-869 method.
Table 1
Table 2
Figure C200410102808D00141
Table 3
Figure C200410102808D00142
Table 4
Figure C200410102808D00151
Table 5
Figure C200410102808D00152
Table 6
Figure C200410102808D00161
Table 7
Figure C200410102808D00171

Claims (4)

1, a kind of infrared spectrum measurement C 10~C 13The method of diene content in the positive structure hydrocarbon comprises:
(1) sets up correction equation or model: choose representational C 10~C 13Positive structure hydrocarbon samples, with its diene content of high effective liquid chromatography for measuring, and then measure each sample correspondence in infrared or long wavelength near infrared spectrometry, select 920~890cm in the middle infrared spectrum -1With 1000~980cm -1, 800~1100cm -1Or 6000~6300cm of long wavelength near infrared spectrometry -1Be characteristic spectrum area, the response of described characteristic spectrum area and sample be associated with the diene content of high effective liquid chromatography for measuring set up correction equation or model, described middle infrared spectrum is multiple attenuated total reflection spectrum, and the diene content of described sample is 0.1~1.7 quality %
(2) measure the unknown sample diene content: under the test condition same with setting up correction equation or model the test unknown sample in infrared or long wavelength near infrared spectrometry, with middle infrared spectrum 920~890cm -1With 1000~980cm -1, 800~1100cm -1Or long wavelength near infrared spectrometry 6000~6300cm -1The corresponding correction equation of response substitution or the model in spectrum district obtain the diene content in the unknown sample.
2, in accordance with the method for claim 1, it is characterized in that with 920~890cm -1With 1000~980cm -1During for characteristic spectrum area, spectrum area or the second-order differential spectrum area with described two characteristic spectrum areas is response respectively, is associated with the diene content of sample with high effective liquid chromatography for measuring, sets up the multiple linear regression correction equation.
3, in accordance with the method for claim 1, it is characterized in that with 800~1100cm -1Spectrum district when being characteristic spectrum area, be response with the absorbance of handling through second-order differential in this spectrum district, be associated with the diene content of sample with high effective liquid chromatography for measuring, set up calibration model with partial least square method.
4, in accordance with the method for claim 1, it is characterized in that with long wavelength near infrared spectrometry 6000~6300cm -1When the spectrum district is characteristic spectrum area, be response, be associated, set up calibration model with partial least square method with the diene content of sample with high effective liquid chromatography for measuring with the absorbance of handling through second-order differential in this spectrum district.
CN 200410102808 2004-12-28 2004-12-28 Method for measuring content of dialkene in C10-C13 positive formation hydrocarbon through spectrum of infrared light Active CN100498293C (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN 200410102808 CN100498293C (en) 2004-12-28 2004-12-28 Method for measuring content of dialkene in C10-C13 positive formation hydrocarbon through spectrum of infrared light

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN 200410102808 CN100498293C (en) 2004-12-28 2004-12-28 Method for measuring content of dialkene in C10-C13 positive formation hydrocarbon through spectrum of infrared light

Publications (2)

Publication Number Publication Date
CN1796980A CN1796980A (en) 2006-07-05
CN100498293C true CN100498293C (en) 2009-06-10

Family

ID=36818190

Family Applications (1)

Application Number Title Priority Date Filing Date
CN 200410102808 Active CN100498293C (en) 2004-12-28 2004-12-28 Method for measuring content of dialkene in C10-C13 positive formation hydrocarbon through spectrum of infrared light

Country Status (1)

Country Link
CN (1) CN100498293C (en)

Families Citing this family (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102954946B (en) * 2011-08-30 2016-01-20 中国石油化工股份有限公司 By the method for infrared spectrum measurement sulfur content in crude oil
CN105021747B (en) * 2014-04-18 2018-03-20 中国石油化工股份有限公司 The method being made up of proton nmr spectra prediction diesel oil race
CN108872135A (en) * 2018-08-31 2018-11-23 天津科技大学 A method of utilizing polymalic acid content near infrared ray fermentation liquid
CN109324012A (en) * 2018-09-14 2019-02-12 中国海洋石油集团有限公司 A method of for light hydrocarbon component quantitative analysis in crude oil
CN111650324A (en) * 2019-03-04 2020-09-11 内蒙古伊泰煤基新材料研究院有限公司 Online detection method for hydrocarbon content
CN111647424A (en) * 2019-03-04 2020-09-11 内蒙古伊泰煤基新材料研究院有限公司 Linear olefin production method capable of being monitored on line
CN111829976A (en) * 2019-04-18 2020-10-27 中国石油化工股份有限公司 Method for predicting composition of gasoline fraction hydrocarbon group of crude oil by near infrared spectrum
CN112185470B (en) * 2020-09-17 2021-07-06 长江存储科技有限责任公司 Data processing method

Non-Patent Citations (6)

* Cited by examiner, † Cited by third party
Title
红外光谱透射法测定裂解汽油二烯值. 张继忠,袁洪福,王京华.石油化工,第33卷第8期. 2004
红外光谱透射法测定裂解汽油二烯值. 张继忠,袁洪福,王京华.石油化工,第33卷第8期. 2004 *
近红外光谱快速测定成品罐汽油的有关性质. 邵波,黄小英,王京华.石油化工,第31卷第10期. 2002
近红外光谱快速测定成品罐汽油的有关性质. 邵波,黄小英,王京华.石油化工,第31卷第10期. 2002 *
采用相关分析解析石油产品近红外光谱. 许育鹏,袁洪福,陆婉珍.现代仪器,第6期. 2002
采用相关分析解析石油产品近红外光谱. 许育鹏,袁洪福,陆婉珍.现代仪器,第6期. 2002 *

Also Published As

Publication number Publication date
CN1796980A (en) 2006-07-05

Similar Documents

Publication Publication Date Title
AU2005287020B2 (en) Method of assaying a hydrocarbon-containing feedstock
EP3519811B1 (en) Measurement of hydrocarbon fuel gas composition and properties
EP0552300B1 (en) Spectral data measurement and correction
CN105319198B (en) Benzene content in gasoline Forecasting Methodology based on Raman spectrum analytic technique
JPH07301599A (en) Method to analyze hydrocarbon by near-infrared spectroscopy analysis
EP1463937A1 (en) Method for analyzing an unknown material as a blend of known materials calculated so as to match certain analytical data and predicting properties of the unknown based on the calculated blend
EP0706049A1 (en) Cracking property determination
ZA200702715B (en) Method of assaying a hydrocarbon-containing feedstock
CN100498293C (en) Method for measuring content of dialkene in C10-C13 positive formation hydrocarbon through spectrum of infrared light
CN100470235C (en) Method for measuring content of dialkene in gasoline through spectrum of near infrared light
CN107966499B (en) Method for predicting crude oil carbon number distribution by near infrared spectrum
CN102954946A (en) Method for determining sulfur content of crude oil through infrared spectrometry
CN102998276A (en) Method for determining crude oil true boiling point distillation curve by using infrared spectroscopy
CN1116878A (en) Method for prediction of physical property data of hydrocarbon products
EP3861320B1 (en) Systems and methods for implicit chemical resolution of vacuum gas oils and fit quality determination
CN1979131A (en) Method for detecting hydrgenated tail-oil paraffin composition using near-infrared spectrum
CN103134764B (en) The method of prediction true boiling point curve of crude oil is composed by transmitted infrared light
CN103134763A (en) Method for predicting crude oil density by infrared spectroscopy
CN107356556A (en) A kind of double integrated modelling approach of Near-Infrared Spectra for Quantitative Analysis
BR9307250A (en) Process for predicting the cetane index of gas oils
CN115060681A (en) SF for FTIR in conjunction with CARS-ELM 6 Method for detecting decomposition product
CN111829976A (en) Method for predicting composition of gasoline fraction hydrocarbon group of crude oil by near infrared spectrum
Pavoni et al. FT‐IR Spectroscopy and Chemometrics as a Useful Approach for Determining Chemical‐Physical Properties of Gasoline, by Minimizing Analytical Times and Sample Handling
CN100451615C (en) Method for detecting hydrogenated tail-oil cyclanes and arene composition using near infrared spectrum
Delaney et al. Statistical prediction of file searching results for vapor phase infrared spectrometric identification of gas chromatographic peaks

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant