CN102338743A - Mid-infrared spectrum method for identifying engine fuel type and brand - Google Patents
Mid-infrared spectrum method for identifying engine fuel type and brand Download PDFInfo
- Publication number
- CN102338743A CN102338743A CN2011101397274A CN201110139727A CN102338743A CN 102338743 A CN102338743 A CN 102338743A CN 2011101397274 A CN2011101397274 A CN 2011101397274A CN 201110139727 A CN201110139727 A CN 201110139727A CN 102338743 A CN102338743 A CN 102338743A
- Authority
- CN
- China
- Prior art keywords
- sample
- model
- fuel
- diesel oil
- type
- 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.)
- Granted
Links
Images
Landscapes
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention discloses a mid-infrared spectrum method for identifying engine fuel type and brand. The method comprises the following steps: collecting a certain number of samples of each type as a training set; measuring the mid-infrared adsorption spectrum in the samples of the training set; assigning a type value to the samples of the training set; and establishing a model of the each fuel type value and the infrared spectrum by use of a PLS (partial least squares) method, and determining an identification rule. To identify an unknown engine fuel sample, a user just needs to measure the infrared adsorption spectrum of the sample, and then a computer determines the type value based on the established model by use of the spectrum and identifies the type and brand according to the identification rule. By utilizing the method disclosed by the invention, the engine type can be judged and the brand can be determined quickly in a few minutes, the fuel anti-counterfeit capability is improved, and the loss of the user caused by using a fuel by mistake or using the adulterated fuel is avoided.
Description
Technical field
The present invention relates to the mid-infrared light spectral method of a kind of quick identification motor fuel kind and the trade mark; Specifically; Relate to a kind of employing mid-infrared light spectral technology; Fast unknown motor fuel is identified as diesel oil, gasoline and jet fuel in conjunction with least square (PLS) method, and confirms the method for its trade mark.
Background technology
The engine purposes is different, causes its motor fuel of a great variety.Such as, diesel motor uses diesel fuel, petrol engine to use petrol power fuel, and aircraft engine uses jet fuel.All kinds of motor fuels are divided into the different trades mark according to service condition.Such as, different according to serviceability temperature, diesel oil is divided into the trades mark such as No. 10, No. 0 ,-No. 10 and-No. 35.According to the capability of antidetonance, gasoline is divided into the trades mark such as No. 90, No. 93 and No. 97.Various types of and various trade mark fuel can not be used with, otherwise engine can not normally move, and cause fault to take place.Such as, under cryogenic conditions, using high trade mark diesel oil, diesel oil at low-temperature is mobile poor, causes engine normally to start; Low trade mark gasoline is during as high trade mark gasoline, and pinking can take place engine, even damages engine.In actual use, exist the motor fuel kind and the trade mark to use with and the quality adulteration, objectionable impurities is pretended to be high-grade gasoline such as in cheap low grade gasoline, mixing.Existing quality testing means can't satisfy the quick identification demand, therefore, press for motor fuel kind and trade mark quick identification technology, and accurate instruction improves the ability of cracking down on counterfeit goods with oil, avoids misapplying the loss that oil plant brings.
Because near-infrared spectrum technique has quick, many property analysis and is fit to online nondestructive analysis characteristics, the increasing at present petrochemical industry that is applied to has irreplaceable effect in fuel quality detection with aspect cracking down on counterfeit goods.Because near infrared spectrum belongs to molecular vibration spectrum, wavelength coverage is 700-2500nm, and for molecule frequency multiplication and combination absorb frequently, peak shape is a broad peak; A little less than the characteristic, only X-H group (X is C, O etc.) there is absorption; A little less than the signal, antijamming capability is strong.Have These characteristics just because of near infrared spectrum, thus must be by the Chemical Measurement and the computer technology in modern times, and near-infrared spectrum technique could well be used.Along with the fast development and the near infrared spectroscopy instrument adaptive capacity to environment of Chemical Measurement and computing machine are strong, low price has promoted the near-infrared spectrum technique fast development, has become hydro carbons (C-H) fuel mass fast detecting and online process monitoring technique at present.
Middle infrared spectrum also belongs to molecular vibration spectrum, is the fundamental frequency absorption of molecular vibration, and wave-number range is 400-4000cm
-1(wavelength is 2.5 microns～25 microns).Define from it, middle infrared spectrum and near infrared spectrum are wavelength coverage.From the essence of its generation, can well find that the two spectrum shape characteristic has obvious difference, thereby the technology and the application that cause finally being adopted there is very big difference.Compare with near infrared spectrum, the middle infrared spectrum peak shape is a spike, and characteristic is strong, and all there is tangible absorption in all kinds of functional groups (comprising the non-X-H of X-H and other functional group); Signal is strong, and microcomponent or adjuvant all have absorption, identifies so the mid-infrared light spectral technology is mainly used in the unknown materials functional group in analytical chemistry field, belongs to one of " four compose greatly ", is used for the unknown materials chemical constitution and identifies.Because a little less than the adaptive capacity to environment of middle infrared spectrum instrument, instrument is expensive, discern with kind so should technology also be difficult to the quality testing of petrochemical industry at present.
Analyze theoretically; Middle infrared spectrum not only has absorption to the main hydrocarbon composition of fuel; And a small amount of non-hydrocarbons component and trace mineral supplement component all there is response; If combine so the mid-infrared light spectral technology learns a skill with stoichiometry, will realize the irrealizable function of near-infrared spectrum technique, measure such as manganese type additive level.Manganese type adjuvant mainly improves the gasoline capability of antidetonance, directly influences the trade mark of gasoline.The present invention adopts the mid-infrared light spectral technology, by means of Chemical Measurement, and the kind of identification of Engine fuel and the trade mark.
Summary of the invention
The object of the invention provides the mid-infrared light spectral method of a kind of identification of Engine fuel type and the trade mark; This method is through the sample infrared spectrum; According to identifying schemes, fast unknown motor fuel is identified as diesel oil, gasoline and jet fuel in conjunction with the PLS method, and confirms the method for its trade mark.
The present invention is to provide the mid-infrared light spectral method of a kind of identification of Engine fuel type and the trade mark, comprise the steps:
The first step: each fuel collection some sample is as training set;
Second step: the mid infrared absorption spectrum of measuring the training set sample;
The 3rd step: confirm identifying schemes: adopt the identification of a plurality of model proceed step by step motor fuel kinds and the trade mark;
The 4th step: utilize the training set sample, adopt PLS to set up and verify the model of identifying schemes;
The 5th step: for the identification of unknown motor fuel sample, at first measure its middle infrared spectrum, utilize model of cognition and the identifying schemes set up to carry out kind and trade mark identification then.
Said motor fuel comprises gasoline, diesel oil and jet fuel; Wherein, gasoline comprises No. 90, No. 93, No. 97 motor petrol and No. 90, No. 93, No. 97 ethanol petrol; Diesel oil comprises No. 0 ,-No. 10 ,-No. 35 light diesel fuel and-No. 10 ,-No. 35 military diesel oil; Jet fuel is No. 3 jet fuels.
Said the 3rd step identifying schemes is described below:
Adopt 7 models to discern successively;
At first adopt model 1, unknown fuel is identified as gasoline, No. 3 jet fuels and diesel oil three major types fuel;
If model 1 recognition result of unknown fuel is a diesel oil, then adopt model 2 that diesel oil further is identified as military diesel oil and light diesel fuel; Adopt model 4 that military diesel oil is identified as-No. 10 military diesel oil and-No. 35 military diesel oil at last respectively; Adopt model 5 that light diesel fuel is identified as No. 0 light diesel fuel ,-No. 10 light diesel fuels and-No. 35 light diesel fuels;
If model 1 recognition result of unknown fuel is a gasoline, then adopt model 3 that gasoline further is identified as ethanol petrol and motor petrol; Adopt model 6 that ethanol petrol is identified as No. 90 ethanol petrols, No. 93 ethanol petrols and No. 97 ethanol petrols at last respectively; Adopt model 7 that motor petrol is identified as No. 90 motor petrol, No. 93 motor petrol and No. 97 motor petrol.
The PLS method model of cognition in said the 4th step is set up and verification step is:
(1) compose with corresponding type numerical value for each training set sample, the type numerical value of same type of sample is identical, and the type numerical value of inhomogeneity sample is different;
(2) training set is divided into calibration set and checking collection, calibration set is used to set up model, and the checking collection is used for testing model;
(3) calibration set sample infrared spectrum is carried out pre-service, select the spectroscopic data of suitable interval, combine the PLS method to set up the quantitative model of infrared spectrum and type numerical value, and confirm that according to the model determination result type identification is regular;
(4) utilize the type numerical value of the institute's model prediction of setting up checking collection sample, confirm sample type according to the type identification rule, and compare the accuracy of verification model with true type.
(5) for the kind identification of unknown sample, at first measure mid infrared absorption spectrum, utilize institute's established model then, measure its type numerical value,, confirm type according to recognition rule.
7 model divisions of said identifying schemes are following:
In model 1, the type numerical value of No. 3 jet fuels, gasoline and diesel oil is respectively 1,500 and 1000; Recognition rule: when types of models numerical evaluation result less than 300, then this sample is No. 3 jet fuels; Measuring the result is 300～700, and then this sample is a gasoline; Measure the result greater than 700 o'clock, this sample is a diesel oil.
In model 2, military diesel oil, light diesel fuel type numerical value are respectively 1 and 100; Recognition rule: when types of models numerical evaluation result less than 55, then this sample is military diesel oil; Measure the result greater than 55, then this sample is a light diesel fuel.
In model 3, motor petrol, ethanol petrol type numerical value are respectively 0 and 1; Recognition rule: when types of models numerical evaluation result less than 0.5, then this sample is a motor petrol; Measure the result greater than 0.5, then this sample is an ethanol petrol.
In model 4 ,-No. 10 military diesel oil ,-No. 35 military diesel oil numerical value are respectively 0 and 1; Recognition rule: when types of models numerical evaluation result less than 0.5, then this sample is-No. 10 military diesel oil; Measure the result greater than 0.5, then this sample is-No. 35 military diesel oil.
In model 5, No. 0 light diesel fuel ,-No. 10 light diesel fuels and-No. 35 light diesel fuel numerical value are respectively 1,500 and 1000; Recognition rule: when types of models numerical evaluation result less than 250, then this sample is No. 0 light diesel fuel; Measuring the result is 250～750, and then this sample is-No. 10 light diesel fuels, measures the result greater than 750 o'clock, and then this sample is-No. 35 light diesel fuels.
In model 6, No. 90 ethanol petrols, No. 93 ethanol petrols and No. 97 ethanol petrol numerical value are respectively 1,500 and 1000; Recognition rule: when types of models numerical evaluation result less than 250, then this sample is No. 90 ethanol petrols; Measuring the result is 250～750, and then this sample is No. 93 ethanol petrols, measures the result greater than 750 o'clock, and then this sample is No. 97 ethanol petrols.
In model 7, No. 90 motor petrol, No. 93 motor petrol and No. 97 motor petrol numerical value are respectively 1,500 and 1000; Recognition rule: when types of models numerical evaluation result less than 250, then this sample is No. 90 motor petrol; Measuring the result is 250～750, and then this sample is No. 93 motor petrol, measures the result greater than 750 o'clock, and then this sample is No. 97 motor petrol.
The present invention utilizes PLS to set up and checking fuel type numerical value and mid-infrared spectral quantitative model, and definite recognition rule.For testing sample, measure its middle infrared spectrum earlier, utilize Model Calculation to go out the type numerical value of unknown sample then, confirm the kind and the trade mark according to recognition rule then.The user only need carry out spectroscopic assay, and all the other steps are accomplished by computing machine automatically, and therefore simple to operate, speed is fast, can in several minutes, successfully unknown motor fuel be identified as gasoline, diesel oil and jet fuel, and can further be identified as the different trades mark.Can guides user correctly use motor fuel on the one hand, avoid misapplying the loss that fuel brings, can be used for quality supervised department on the other hand, on-the-spot quick check fuel improves the ability of cracking down on counterfeit goods, and avoids the use of and mingles the loss that fuel brings.
Description of drawings
The identifying schemes of Fig. 1 the inventive method;
Fig. 2 all samples mid infrared absorption spectrum figure;
Fig. 3 motor fuel kind model tuning collection type numerical value match value and true value relation;
Fig. 4 motor fuel kind modelling verification collection number of types value prediction value and true value relation;
Fig. 5 light diesel fuel and military diesel oil model of cognition calibration set type numerical value match value and true value relation;
Fig. 6 light diesel fuel and military diesel oil model of cognition checking collection number of types value prediction value and true value relation;
Fig. 7 ethanol petrol and motor petrol model of cognition calibration set type numerical value match value and true value relation;
Fig. 8 ethanol petrol and motor petrol model of cognition checking collection number of types value prediction value and true value relation;
The training set type numerical value match value of the military diesel oil trade mark of Fig. 9 model of cognition and true value relation;
The military diesel oil trade mark of Figure 10 model of cognition training set cross-verification number of types value prediction value and true value relation;
Figure 11 light diesel fuel trade mark model of cognition calibration set type numerical value match value and true value relation;
Figure 12 light diesel fuel trade mark model of cognition checking collection number of types value prediction value and true value relation;
The training set type numerical value match value of Figure 13 ethanol petrol trade mark model of cognition and true value relation;
Figure 14 ethanol petrol trade mark model of cognition training set cross-verification number of types value prediction value and true value relation;
Figure 15 motor petrol trade mark model of cognition calibration set type numerical value match value and true value relation;
Figure 16 motor petrol trade mark model of cognition checking collection number of types value prediction value and true value relation.
Embodiment
The implementation step of the inventive method is following:
The first step: each type collected some samples as training set;
Second step: the mid infrared absorption spectrum of measuring the training set sample;
The 3rd step: according to the identifying schemes of Fig. 1, proceed step by step motor fuel kind and trade mark identification;
The 4th step: utilize the training set sample, adopt PLS (PLS) to set up and verify 7 model of cognition;
The 5th step: for the identification of unknown motor fuel sample, at first measure its middle infrared spectrum, utilize model of cognition and the identifying schemes proceed step by step kind and the trade mark identification of setting up then.
The identifying schemes of Fig. 1 is described below:
At first adopt model 1, unknown fuel is identified as gasoline, No. 3 jet fuels and diesel oil three major types fuel;
If model 1 recognition result of unknown fuel is a diesel oil, then adopt model 2 that diesel oil further is identified as military diesel oil and light diesel fuel; Adopt model 4 that military diesel oil is identified as-No. 10 military diesel oil and-No. 35 military diesel oil at last respectively; Adopt model 5 that light diesel fuel is identified as No. 0 light diesel fuel ,-No. 10 light diesel fuels and-No. 35 light diesel fuels;
If model 1 recognition result of unknown fuel is a gasoline, then adopt model 3 that gasoline further is identified as ethanol petrol and motor petrol; Adopt model 6 that ethanol petrol is identified as No. 90 ethanol petrols, No. 93 ethanol petrols and No. 97 ethanol petrols at last respectively; Adopt model 7 that motor petrol is identified as No. 90 motor petrol, No. 93 motor petrol and No. 97 motor petrol.
PLS method model of cognition is set up and verification step:
(1) compose with corresponding type numerical value for each training set sample, the type numerical value of same type of sample is identical, and the type numerical value of inhomogeneity sample is different;
(2) training set is divided into calibration set and checking collection, calibration set is used to set up model, and the checking collection is used for testing model;
(3) calibration set sample infrared spectrum is carried out pre-service, select the spectroscopic data of suitable interval, combine the PLS method to set up the quantitative model of infrared spectrum and type numerical value, and the model determination result confirms that type identification is regular;
(4) utilize the type numerical value of the institute's model prediction of setting up checking collection sample, confirm sample type according to the type identification rule, and compare the accuracy of verification model with true type.
(5) for the kind identification of unknown sample, at first measure mid infrared absorption spectrum, utilize institute's established model then, measure its type numerical value,, confirm type according to recognition rule.
7 model divisions of said identifying schemes are following:
In model 1, the type numerical value of No. 3 jet fuels, gasoline and diesel oil is respectively 1,500 and 1000; Recognition rule: when types of models numerical evaluation result less than 300, then this sample is No. 3 jet fuels; Measuring the result is 300～700, and then this sample is a gasoline; Measure the result greater than 700 o'clock, this sample is a diesel oil.
In model 2, military diesel oil, light diesel fuel type numerical value are respectively 1 and 100; Recognition rule: when types of models numerical evaluation result less than 55, then this sample is military diesel oil; Measure the result greater than 55, then this sample is a light diesel fuel.
In model 3, motor petrol, ethanol petrol type numerical value are respectively 0 and 1; Recognition rule: when types of models numerical evaluation result less than 0.5, then this sample is a motor petrol; Measure the result greater than 0.5, then this sample is an ethanol petrol.
In model 4 ,-No. 10 military diesel oil ,-No. 35 military diesel oil numerical value are respectively 0 and 1; Recognition rule: when types of models numerical evaluation result less than 0.5, then this sample is-No. 10 military diesel oil; Measure the result greater than 0.5, then this sample is-No. 35 military diesel oil.
In model 5, No. 0 light diesel fuel ,-No. 10 light diesel fuels and-No. 35 light diesel fuel numerical value are respectively 1,500 and 1000; Recognition rule: when types of models numerical evaluation result less than 250, then this sample is No. 0 light diesel fuel; Measuring the result is 250～750, and then this sample is-No. 10 light diesel fuels, measures the result greater than 750 o'clock, and then this sample is-No. 35 light diesel fuels.
In model 6, No. 90 ethanol petrols, No. 93 ethanol petrols and No. 97 ethanol petrol numerical value are respectively 1,500 and 1000; Recognition rule: when types of models numerical evaluation result less than 250, then this sample is No. 90 ethanol petrols; Measuring the result is 250～750, and then this sample is No. 93 ethanol petrols, measures the result greater than 750 o'clock, and then this sample is No. 97 ethanol petrols.
In model 7, No. 90 motor petrol, No. 93 motor petrol and No. 97 motor petrol numerical value are respectively 1,500 and 1000; Recognition rule: when types of models numerical evaluation result less than 250, then this sample is No. 90 motor petrol; Measuring the result is 250～750, and then this sample is No. 93 motor petrol, measures the result greater than 750 o'clock, and then this sample is No. 97 motor petrol.
The user can adopt the inventive method, according to the fuel type of reality, can select targetedly or upgrade, and realizes the identification of the fuel type and the trade mark.
Pass through example in detail the present invention below, but the present invention is not limited to this.
Instance 1: the foundation and the checking of motor fuel kind model of cognition
1) collects the training set sample
Collect gasoline, diesel oil and No. 3 jet fuels totally 884 samples from each refinery of China, wherein gasoline sample is 546,238 of diesel samples, 100 of No. 3 jet fuels.
2) measure training set sample infrared spectrum
Adopt TENSOR 27 mid-infrared light spectrometers to measure training set sample mid infrared absorption spectrum, spectral range: 600cm
-1～4000cm
-1The transmission sample pond, the 0.1mm light path, spectrogram is seen Fig. 2.
3) the type numerical value of jet fuel, gasoline and diesel oil is respectively 1,500 and 1000.Just training set is divided into calibration set and checking collection, and number is respectively 530 and 354.Utilize the calibration set sample to combine the PLS method to set up the model of sample type numerical value and spectrum; Modeling parameters is seen table 1; Utilize the checking collection to test then, the match value of calibration set type numerical value and true value relation are seen Fig. 3, and checking collection number of types value prediction value and true value relation are seen Fig. 4.Confirm recognition rule according to Fig. 3 result: when types of models numerical evaluation result less than 300, then this sample is No. 3 jet fuels; Measuring the result is 300～700, and then this sample is a gasoline; Measure the result greater than 700 o'clock, this sample is a diesel oil.Utilize the sample type of this rule identification checking collection sample.Recognition result is seen table 1, and calibration set is 100% with checking collection discrimination, and this model can use.
Table 1 motor fuel kind model parameter and recognition result
Instance 2: the foundation and the checking of light diesel fuel and military diesel oil model of cognition
1) collects the training set sample
Collect light diesel fuel and military diesel oil totally 258 samples, 58 of wherein military diesel oil, 200 of light diesel fuels from each refinery of China.
2) measure training set sample infrared spectrum
Adopt TENSOR 27 mid-infrared light spectrometers to measure training set sample mid infrared absorption spectrum, spectral range: 600cm
-1～4000cm
-1The transmission sample pond, the 0.1mm light path.
3) the type numerical value of military diesel oil and light diesel fuel is respectively 1 and 100.Training set is divided into calibration set and checking collection, and number is followed successively by 131 and 127.Utilize calibration set samples using PLS method to set up the model of sample type numerical value and spectrum, modeling parameters is seen table 2; Utilize the checking collection to test, the match value of calibration set type numerical value and true value relation are seen Fig. 5, and checking collection number of types value prediction value and true value relation are seen Fig. 6.Confirm recognition rule according to Fig. 5 result: when types of models numerical evaluation result less than 55, then this sample is military diesel oil; Measure the result greater than 55 o'clock, then this sample is a light diesel fuel.Utilize the sample type of this rule identification checking collection sample, recognition result is seen table 2, and calibration set is 100% with checking collection discrimination, and this model can use.
Table 2 light diesel fuel and military diesel oil model of cognition parameter and recognition result
Instance 3: ethanol petrol and motor petrol identification
1) collects the training set sample
Collect ethanol petrol and motor petrol totally 430 samples from each refinery of China, wherein motor petrol is 369,61 of ethanol petrols.
2) measure training set sample infrared spectrum
Adopt TENSOR 27 mid-infrared light spectrometers to measure training set sample mid infrared absorption spectrum, spectral range: 600cm
-1～4000cm
-1The transmission sample pond, the 0.1mm light path.
3) the type numerical value of ethanol petrol and motor petrol is respectively 0 and 1.Training set is divided into calibration set and checking collection, and number is followed successively by 216 and 214.Utilize calibration set samples using PLS method to set up the model of sample type numerical value and spectrum, modeling parameters is seen table 3; Utilize the checking collection to test then, the match value of calibration set type numerical value and true value relation are seen Fig. 7, and checking collection number of types value prediction value and true value relation are seen Fig. 8.Confirm recognition rule according to Fig. 7 result: when types of models numerical evaluation result less than 0.5, then this sample is a motor petrol; Measure the result greater than 0.5 o'clock, then this sample is an ethanol petrol.Utilize the sample type of this rule identification checking collection sample, recognition result is seen table 3, and calibration set is 100% with checking collection discrimination, and this model can use.
Table 3 motor petrol and ethanol petrol model of cognition parameter and recognition result
Instance 4: military diesel oil trade mark model of cognition is set up and checking
1) collects the training set sample
Collect-No. 10 military diesel oil and-No. 35 military diesel oil totally 59 samples from each refinery of China, wherein-No. 10 military diesel oil is 45,14 of-No. 35 military diesel oil.
2) measure training set sample infrared spectrum
Adopt TENSOR 27 mid-infrared light spectrometers to measure training set sample mid infrared absorption spectrum, spectral range: 600cm
-1～4000cm
-1The transmission sample pond, the 0.1mm light path.
3)-the type numerical value of No. 10 military diesel oil and-No. 35 military diesel oil is respectively 0 and 1.Training set samples using PLS method is set up the model of sample type numerical value and spectrum, and modeling parameters is seen table 4; Utilize the cross-verification mode to test then, the match value of training set type numerical value and true value relation are seen Fig. 9, and training set cross-verification number of types value prediction value and true value relation are seen Figure 10.Confirm recognition rule according to Fig. 9 result: when types of models numerical evaluation result less than 0.5, then this sample is-No. 10 military diesel oil; Measure the result greater than 0.5 o'clock, then this sample is-No. 35 military diesel oil.Utilize the sample type of this rule identification checking collection sample, recognition result is seen table 4, and calibration set is 100% with checking collection discrimination, and this model can use.
The military diesel oil trade mark of table 4 model of cognition parameter and recognition result
Instance 5: light diesel fuel trade mark identification
1) collects the training set sample
Collect No. 0 light diesel fuel ,-No. 10 light diesel fuels and-No. 35 light diesel fuels totally 197 samples from each refinery of China, wherein No. 0 light diesel fuel is 122,44 of-No. 10 light diesel fuels, 31 of-No. 35 light diesel fuels.
2) measure training set sample infrared spectrum
Adopt TENSOR 27 mid-infrared light spectrometers to measure training set sample mid infrared absorption spectrum, spectral range: 600cm
-1～4000cm
-1The transmission sample pond, the 0.1mm light path.
3) the type numerical value of No. 0 light diesel fuel ,-No. 10 light diesel fuels and-No. 35 light diesel fuels is respectively 0,500 and 1000.Training set is divided into calibration set and checking collection, and the sample number is respectively 99 and 98.Calibration set sample samples using PLS method is set up the model of sample type numerical value and spectrum, and parameter is seen table 5; Utilize checking collection sample to test then, the match value of calibration set type numerical value and true value relation are seen Figure 11, and checking collection number of types value prediction value and true value relation are seen Figure 12.Confirm recognition rule according to Figure 11 result: when types of models numerical evaluation result less than 250, then this sample is No. 0 light diesel fuel; Measuring the result is 250～750 o'clock, and then this sample is-No. 10 light diesel fuels; Measure the result greater than 750 o'clock, this sample is-No. 35 light diesel fuels.Utilize the sample type of this rule identification checking collection sample, recognition result is seen table 5, and calibration set is 100% with checking collection discrimination, and this model can use.
Table 5 light diesel fuel trade mark model of cognition parameter and recognition result
| Pre-service | Wave number (cm -1) scope | Main gene | Calibration set/checking collection discrimination |
| Second derivative | 3996.43315.5 | 10 | 100％/100％ |
Instance 6: ethanol petrol trade mark model of cognition is set up and checking
1) collects the training set sample
Collect No. 90 ethanol petrols, No. 93 ethanol petrols and 59 samples of 97 ethanol petrols from each refinery of China, wherein No. 90 ethanol petrols are 5,40 of No. 93 ethanol petrols, 14 of No. 97 ethanol petrols.
2) measure training set sample infrared spectrum
Adopt TENSOR 27 mid-infrared light spectrometers to measure training set sample mid infrared absorption spectrum, spectral range: 600cm
-1～4000cm
-1The transmission sample pond, the 0.1mm light path.
3) the type numerical value of No. 90 ethanol petrols, No. 93 ethanol petrols and No. 97 ethanol petrols is respectively 0,500 and 1000.Training set samples using PLS method is set up the model of sample type numerical value and spectrum, and parameter is seen table 6; Utilize the cross-verification mode to test then, the match value of training set type numerical value and true value relation are seen Figure 13, and training set cross-verification number of types value prediction value and true value relation are seen Figure 14.Confirm recognition rule according to Figure 13 result: when types of models numerical evaluation result less than 250, then this sample is No. 90 ethanol petrols; Measuring the result is 250～750 o'clock, and then this sample is No. 93 ethanol petrols; Measure the result greater than 750 o'clock, this sample is No. 97 ethanol petrols.Utilize the sample type of this rule identification checking collection sample, recognition result is seen table 6, and calibration set is 100% with checking collection discrimination, and this model can use.
Table 6 ethanol petrol trade mark model of cognition parameter and recognition result
Instance 7: the foundation and the checking of the identification of the motor petrol trade mark
1) collects the training set sample
Collect No. 90 motor petrol, No. 93 motor petrol and No. 97 motor petrol totally 197 samples from each refinery of China, wherein No. 90 motor petrol are 57,219 of No. 93 motor petrol, 74 of No. 97 motor petrol.
2) measure training set sample infrared spectrum
Adopt TENSOR 27 mid-infrared light spectrometers to measure training set sample mid infrared absorption spectrum, spectral range: 600cm
-1～4000cm
-1The transmission sample pond, the 0.1mm light path.
3) the type numerical value of No. 90 motor petrol, No. 93 motor petrol and No. 97 motor petrol is respectively 0,500 and 1000, and training set is divided into calibration set and checking collection, and the sample number is respectively 175 and 175.Calibration set sample samples using PLS method is set up the model of sample type numerical value and spectrum, and parameter is seen table 7; Utilize checking collection sample to test then, the match value of calibration set type numerical value and true value relation are seen Figure 15, and checking collection number of types value prediction value and true value relation are seen Figure 16.Confirm recognition rule according to Figure 15 result: when types of models numerical evaluation result less than 250, then this sample is No. 90 motor petrol; Measuring the result is 250～750 o'clock, and then this sample is No. 93 motor petrol; Measure the result greater than 750 o'clock, this sample is No. 97 motor petrol.Utilize the sample type of this rule identification checking collection sample, recognition result is seen table 7, and calibration set is 100% with checking collection discrimination, and this model can use.
Table 7 motor petrol trade mark model of cognition parameter and recognition result
Instance 8: the kind of unknown motor fuel and trade mark identification
Select the motor fuel sample of 5 known types and the trade mark, measure mid infrared absorption spectrum respectively, discern according to the identifying schemes of Fig. 1.
At first, adopt model 1 to discern, the recognition result of 1,2,3,4,5 sample is followed successively by No. 3 jet fuels, diesel oil, diesel oil, gasoline, gasoline.
For No. 1 sample, need not further identification, the result is No. 3 jet fuels.
For No. 2 and No. 3 samples, adopt model 2 to discern the result: No. 2 is military diesel oil, and No. 3 is light diesel fuel.Adopt 4 pairs of No. 2 samples of model further to discern, the result is-No. 10 military diesel oil; Adopt 5 pairs of No. 3 samples of model further to discern, the result is No. 0 light diesel fuel;
For No. 4 and No. 5 samples, adopt model 3 to discern the result: No. 4 is ethanol petrol, and No. 5 is motor petrol; Adopt 6 pairs of No. 4 samples of model further to discern, the result is No. 97 ethanol petrols; Adopt 7 pairs of No. 5 samples of model further to discern, the result is No. 93 motor petrol;
According to The above results, 1～No. 5 sample type is followed successively by: No. 3 jet fuels ,-No. 10 military diesel oil, No. 0 light diesel fuel, No. 97 ethanol petrols and No. 93 motor petrol, and in full accord with its actual type, explain that the inventive method is effective.
Claims (10)
1. the mid-infrared light spectral method of the identification of Engine fuel type and the trade mark comprises the steps:
The first step: each fuel collection some sample is as training set;
Second step: the mid infrared absorption spectrum of measuring the training set sample;
The 3rd step: confirm identifying schemes: adopt the identification of a plurality of model proceed step by step motor fuel kinds and the trade mark;
The 4th step: utilize the training set sample, adopt PLS to set up and verify the model of identifying schemes;
The 5th step: for the identification of unknown motor fuel sample, at first measure its middle infrared spectrum, utilize model of cognition and the identifying schemes set up to carry out kind and trade mark identification then.
2. method according to claim 1 is characterized in that said motor fuel comprises gasoline, diesel oil and jet fuel; Wherein, gasoline comprises No. 90, No. 93, No. 97 motor petrol and No. 90, No. 93, No. 97 ethanol petrol; Diesel oil comprises No. 0 ,-No. 10 ,-No. 35 light diesel fuel and-No. 10 ,-No. 35 military diesel oil; Jet fuel is No. 3 jet fuels.
3. method according to claim 1 is characterized in that said the 3rd step identifying schemes is following:
Adopt 7 models to discern successively;
At first adopt model 1, unknown fuel is identified as gasoline, No. 3 jet fuels and diesel oil three major types fuel;
If model 1 recognition result of unknown fuel is a diesel oil, then adopt model 2 that diesel oil further is identified as military diesel oil and light diesel fuel; Adopt model 4 that military diesel oil is identified as-No. 10 military diesel oil and-No. 35 military diesel oil at last respectively; Adopt model 5 that light diesel fuel is identified as No. 0 light diesel fuel ,-No. 10 light diesel fuels and-No. 35 light diesel fuels;
If model 1 recognition result of unknown fuel is a gasoline, then adopt model 3 that gasoline further is identified as ethanol petrol and motor petrol; Adopt model 6 that ethanol petrol is identified as No. 90 ethanol petrols, No. 93 ethanol petrols and No. 97 ethanol petrols at last respectively; Adopt model 7 that motor petrol is identified as No. 90 motor petrol, No. 93 motor petrol and No. 97 motor petrol.
4. method according to claim 1 is characterized in that the foundation of least square method model of cognition and the verification step in said the 4th step is following:
(1) compose with corresponding type numerical value for each training set sample, the type numerical value of same type of sample is identical, and the type numerical value of inhomogeneity sample is different;
(2) training set is divided into calibration set and checking collection, calibration set is used to set up model, and the checking collection is used for testing model;
(3) calibration set sample infrared spectrum is carried out pre-service, select the spectroscopic data of suitable interval, combine the PLS method to set up the quantitative model of infrared spectrum and type numerical value, and confirm that according to the model determination result type identification is regular;
(4) utilize the type numerical value of the institute's model prediction of setting up checking collection sample, confirm sample type according to the type identification rule, and compare the accuracy of verification model with true type.
(5) for the kind identification of unknown sample, at first measure mid infrared absorption spectrum, utilize institute's established model then, measure its type numerical value,, confirm type according to recognition rule.
5. method according to claim 3 is characterized in that: in the said model 1, the type numerical value of No. 3 jet fuels, gasoline and diesel oil is respectively 1,500 and 1000; Recognition rule is: when types of models numerical evaluation result less than 300, then this sample is No. 3 jet fuels; Measuring the result is 300～700, and then this sample is a gasoline; Measure the result greater than 700 o'clock, this sample is a diesel oil.
6. method according to claim 3 is characterized in that: in the said model 2, military diesel oil, light diesel fuel type numerical value are respectively 1 and 100; Recognition rule is: when types of models numerical evaluation result less than 55, then this sample is military diesel oil; Measure the result greater than 55, then this sample is a light diesel fuel.
7. method according to claim 3 is characterized in that: in the said model 3, motor petrol, ethanol petrol type numerical value are respectively 0 and 1; Recognition rule is: when types of models numerical evaluation result less than 0.5, then this sample is a motor petrol; Measure the result greater than 0.5, then this sample is an ethanol petrol.
8. method according to claim 3 is characterized in that: in the said model 4 ,-No. 10 military diesel oil ,-No. 35 military diesel oil numerical value are respectively 0 and 1; Recognition rule is: when types of models numerical evaluation result less than 0.5, then this sample is-No. 10 military diesel oil; Measure the result greater than 0.5, then this sample is-No. 35 military diesel oil.
9. method according to claim 3 is characterized in that: in the said model 5, No. 0 light diesel fuel ,-No. 10 light diesel fuels and-No. 35 light diesel fuel numerical value are respectively 1,500 and 1000; Recognition rule is: when types of models numerical evaluation result less than 250, then this sample is No. 0 light diesel fuel; Measuring the result is 250～750, and then this sample is-No. 10 light diesel fuels, measures the result greater than 750 o'clock, and then this sample is-No. 35 light diesel fuels.
10. method according to claim 3 is characterized in that: in the said model 6, No. 90 ethanol petrols, No. 93 ethanol petrols and No. 97 ethanol petrol numerical value are respectively 1,500 and 1000; Recognition rule is: when types of models numerical evaluation result less than 250, then this sample is No. 90 ethanol petrols; Measuring the result is 250～750, and then this sample is No. 93 ethanol petrols, measures the result greater than 750 o'clock, and then this sample is No. 97 ethanol petrols; In the said model 7, No. 90 motor petrol, No. 93 motor petrol and No. 97 motor petrol numerical value are respectively 1,500 and 1000; Recognition rule is: when types of models numerical evaluation result less than 250, then this sample is No. 90 motor petrol; Measuring the result is 250～750, and then this sample is No. 93 motor petrol, measures the result greater than 750 o'clock, and then this sample is No. 97 motor petrol.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201110139727 CN102338743B (en) | 2011-05-27 | 2011-05-27 | Mid-infrared spectrum method for identifying engine fuel type and brand |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201110139727 CN102338743B (en) | 2011-05-27 | 2011-05-27 | Mid-infrared spectrum method for identifying engine fuel type and brand |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102338743A true CN102338743A (en) | 2012-02-01 |
| CN102338743B CN102338743B (en) | 2013-04-17 |
Family
ID=45514582
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 201110139727 Expired - Fee Related CN102338743B (en) | 2011-05-27 | 2011-05-27 | Mid-infrared spectrum method for identifying engine fuel type and brand |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102338743B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103439311A (en) * | 2013-09-02 | 2013-12-11 | 中国检验检疫科学研究院 | Raman spectrum method for rapidly recognizing gasoline brands |
| CN103983582A (en) * | 2014-05-09 | 2014-08-13 | 中国人民解放军空军勤务学院 | Jet fuel anti-freezing additive on-line detection system |
| CN107831135A (en) * | 2017-10-23 | 2018-03-23 | 大连理工大学 | It is a kind of to establish two-dimentional qualitative analysis model using near infrared spectroscopy to differentiate the method in the fresh extra large stichopus japonicus place of production |
| CN108388224A (en) * | 2018-04-03 | 2018-08-10 | 江南大学 | A kind of Desalting and Dewatering from Crude Oil process operation method for monitoring state based near infrared spectrum |
| CN109374565A (en) * | 2018-09-30 | 2019-02-22 | 华东交通大学 | A kind of methanol gasoline ethanol petrol differentiates and content assaying method |
| CN112833955A (en) * | 2021-01-08 | 2021-05-25 | 三一汽车起重机械有限公司 | Method for establishing hydraulic oil quality monitoring model, monitoring method, device and system |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5349189A (en) * | 1990-04-09 | 1994-09-20 | Ashland Oil, Inc. | Process and apparatus for analysis of hydrocarbons by near-infrared spectroscopy |
| CN101782511A (en) * | 2010-03-31 | 2010-07-21 | 中国人民解放军总后勤部油料研究所 | Method for rapidly identifying type and grade of lubricating oil |
| CN101806729A (en) * | 2010-03-31 | 2010-08-18 | 中国人民解放军总后勤部油料研究所 | In-use lubricating oil quality rapid testing method |
-
2011
- 2011-05-27 CN CN 201110139727 patent/CN102338743B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5349189A (en) * | 1990-04-09 | 1994-09-20 | Ashland Oil, Inc. | Process and apparatus for analysis of hydrocarbons by near-infrared spectroscopy |
| CN101782511A (en) * | 2010-03-31 | 2010-07-21 | 中国人民解放军总后勤部油料研究所 | Method for rapidly identifying type and grade of lubricating oil |
| CN101806729A (en) * | 2010-03-31 | 2010-08-18 | 中国人民解放军总后勤部油料研究所 | In-use lubricating oil quality rapid testing method |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103439311A (en) * | 2013-09-02 | 2013-12-11 | 中国检验检疫科学研究院 | Raman spectrum method for rapidly recognizing gasoline brands |
| CN103983582A (en) * | 2014-05-09 | 2014-08-13 | 中国人民解放军空军勤务学院 | Jet fuel anti-freezing additive on-line detection system |
| CN103983582B (en) * | 2014-05-09 | 2016-04-20 | 中国人民解放军空军勤务学院 | Jet fuel antifreeze additive on-line detecting system |
| CN107831135A (en) * | 2017-10-23 | 2018-03-23 | 大连理工大学 | It is a kind of to establish two-dimentional qualitative analysis model using near infrared spectroscopy to differentiate the method in the fresh extra large stichopus japonicus place of production |
| CN108388224A (en) * | 2018-04-03 | 2018-08-10 | 江南大学 | A kind of Desalting and Dewatering from Crude Oil process operation method for monitoring state based near infrared spectrum |
| CN109374565A (en) * | 2018-09-30 | 2019-02-22 | 华东交通大学 | A kind of methanol gasoline ethanol petrol differentiates and content assaying method |
| CN112833955A (en) * | 2021-01-08 | 2021-05-25 | 三一汽车起重机械有限公司 | Method for establishing hydraulic oil quality monitoring model, monitoring method, device and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102338743B (en) | 2013-04-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102338743B (en) | Mid-infrared spectrum method for identifying engine fuel type and brand | |
| CN101893561B (en) | Near infrared spectrum quick test method of new oil quality of lubricating oil | |
| CN102288573A (en) | Method for fast recognizing fuel type and designation of engine by use of mid-infrared spectrum technique | |
| CN101339150B (en) | Method for determining octane number based on dielectric spectra technology | |
| CN102735642B (en) | Method for quickly and losslessly identifying virgin olive oil and olive-residue oil | |
| CN101806729A (en) | In-use lubricating oil quality rapid testing method | |
| CN103018195B (en) | Method for determination of PCTFE content in PBX explosive by near infrared spectrum | |
| CN102778442A (en) | Method for rapidly identifying type of balsam material liquid for cigarette | |
| CN104990894A (en) | Detection method of gasoline properties based on weighted absorbance and similar samples | |
| CN103115889B (en) | Method by transmitted infrared light spectrum prediction sulfur content in crude oil | |
| CN102841069B (en) | Method for rapidly identifying types of crude oil by using mid-infrared spectrum | |
| CN102221534B (en) | Mid-infrared spectrum method for quickly identifying engine fuel type | |
| CN105319179A (en) | Method of predicting content of hydrogen sulfide in desulfuration amine liquid through mid-infrared spectroscopy | |
| CN109991206B (en) | Method for measuring total alcohol content of alcohol gasoline based on partial least square method | |
| CN108072627A (en) | It is a kind of that the method for amino-acid nitrogen and total acid content in soy sauce is quickly detected with middle infrared spectrum | |
| CN103063599B (en) | The method of prediction oil density is composed by transmitted infrared light | |
| CN102323235B (en) | Method for measuring quality indexes of engine fuel by mid-infrared spectroscopy technology | |
| CN102954946A (en) | Method for determining sulfur content of crude oil through infrared spectrometry | |
| CN102252994A (en) | Method for rapid detection of quality index of motor fuel | |
| CN103134764B (en) | The method of prediction true boiling point curve of crude oil is composed by transmitted infrared light | |
| CN105954228A (en) | Method for measuring content of sodium metal in oil sand based on near infrared spectrum | |
| CN100425975C (en) | Method for measuring character data of gasoline from near infrared light spectrum | |
| CN101021495A (en) | Method for distinguishing petroleum products variety | |
| CN107402191A (en) | A kind of oil product of gasoline quality determining method for evading disturbing factor | |
| CN102954945B (en) | A kind of method by infrared spectrum measurement acid value for crude oil |
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 | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20130417 Termination date: 20140527 |