CN108956814B - Method for directly constructing gasoline molecular composition model and property prediction method - Google Patents

Abstract

The invention provides a method for directly constructing a gasoline molecular composition model and a property prediction method, wherein the construction method comprises the following steps: (1) performing monomer hydrocarbon analysis on the gas chromatography detection result of the gasoline sample to identify molecules possibly contained in each peak in the gas chromatogram, and calculating the relative fraction of each peak; (2) classifying chromatographic peaks according to the analysis result of the monomer hydrocarbon and the types of the chromatographic peaks; (3) analyzing the classified chromatographic peaks according to the chromatographic peak types and the molecular types to obtain a complete monomer hydrocarbon molecule composition result; (4) the results are composed of the complete monomeric hydrocarbon molecules to directly generate a compositional model of the gasoline molecules. The method provided by the invention is based on a gas chromatography detection result, the molecular composition is reconstructed by using a peak regulation algorithm based on statistical distribution, then a gasoline molecular composition model is established, and the macroscopic property of gasoline is predicted.

Description

A method for directly constructing a gasoline molecular composition model and a method for predicting properties

technical field

The invention relates to a method for directly constructing a gasoline molecular composition model and a method for predicting properties, in particular to a method for directly constructing a gasoline molecular composition model from gas chromatography results and a method for predicting properties, belonging to the technical field of oil composition analysis.

Background technique

The processing and blending of gasoline is very demanding on its quality and composition, and the traditional lumped model is increasingly unable to meet the needs of a modern refinery. In order to improve refinery efficiency and improve product quality standards, it is increasingly important to develop molecular-level models and implement molecular-level management of light oil processing and blending processes. The first problem to be solved in establishing a molecular-level model is to obtain the molecular composition of the oil.

Existing methods for obtaining the molecular composition of gasoline can be mainly divided into experimental methods and methods for reconstructing molecular composition with computer assistance. Traditional analytical instruments, such as gas chromatography-hydrogen flame ionization detector (GC-FID), contain many co-evolution peaks and unidentifiable peaks in the analysis results. Existing computer-aided reconstruction methods of gasoline molecular composition usually derive the molecular composition from the macroscopic properties of gasoline samples. However, there are uncertainties in such methods, which will affect the subsequent processing and reconciliation simulation process. In addition, the predicted value of gasoline properties obtained by this method is greatly affected by the input experimental value, and some of these methods also rely heavily on the quality of the molecular composition database and the method of training data.

Therefore, it has become an urgent technical problem to be solved in the art to provide a method for directly constructing a gasoline molecular composition model and a method for predicting properties.

SUMMARY OF THE INVENTION

In order to solve the above shortcomings and deficiencies, the purpose of the present invention is to provide a method for directly constructing a gasoline molecular composition model.

Another object of the present invention is to provide a system for directly constructing a gasoline molecular composition model.

Another object of the present invention is to provide a method and system for predicting the macroscopic properties of gasoline. The technical scheme provided by the present invention adopts the peak adjustment algorithm based on statistical distribution to reconstruct the molecular results obtained by gas chromatography, and obtains a complete molecular composition, which is used to establish a gasoline molecular composition model and predict its properties. The method combines experimental methods and computer reconstruction methods to split or speculate the co-escape peaks and unidentified peaks in the gas chromatographic results according to a certain statistical distribution, providing accurate and stable molecular composition results, without the need to establish molecular Composing the database and associating training overcomes the shortcomings of existing methods. The macroscopic properties of the predicted molecular composition based on this method are not affected by the experimental values and have more reference value.

In order to achieve the above object, on the one hand, the present invention provides a method for directly constructing a gasoline molecular composition model, wherein the method comprises the following steps:

(1) carry out monomer hydrocarbon analysis to the gas chromatographic detection result of gasoline sample, to identify the molecules that each peak may contain in the gas chromatogram, and calculate the relative fraction of each peak;

(2) According to the analysis results of monomer hydrocarbons, and classify the chromatographic peaks according to their chromatographic peak types;

(3) analyze the classified chromatographic peaks according to the chromatographic peak type and molecular type, and obtain the complete monomer hydrocarbon molecular composition result;

(4) A composition model of gasoline molecules is directly generated from the complete monomeric hydrocarbon molecule composition results.

According to the method of the present invention, preferably, the gasoline sample includes catalytically cracked gasoline, catalytically reformed gasoline, straight-run gasoline, catalytically cracked gasoline, hydrogenated gasoline or coker gasoline.

According to the method of the present invention, preferably, the chromatographic peak types include known peaks, co-elution peaks and unidentified peaks.

According to the method of the present invention, preferably, the analysis in step (3) is to use a statistical distribution-based peak adjustment algorithm to analyze the classified chromatographic peaks according to chromatographic peak types and molecular types.

According to the method of the present invention, preferably, the peak adjustment algorithm based on statistical distribution specifically includes the following steps:

Fit the distribution of known peaks: classify known peaks according to molecular type and carbon number, and fit the classified data according to statistical distribution;

Separation of co-escape peaks: Split the co-escape peaks, assuming the relative content of each component in the co-escape peaks, test all the hypotheses in turn, and select the appropriate hypothesis to accept; repeat the above process continuously until all the The co-evolution peaks are all split;

Infer unidentified peaks: Assume the molecular type and carbon number of the components contained in the unidentified peaks, test all hypotheses in turn, and select the appropriate hypothesis to accept; repeat the above process continuously until all unidentified peaks are inferred.

According to the method of the present invention, wherein, fitting the distribution of known peaks specifically includes the following steps: classifying known peaks by molecular type and carbon number, and dividing the content of all molecules with the same molecular type and carbon number Perform the summation to obtain a data matrix; and then fit the data of each molecular series relative to the carbon number in the data matrix according to the statistical distribution.

According to the method of the present invention, preferably, the statistical distribution includes a gamma distribution.

According to the method of the present invention, preferably, the molecular types in step (3) include normal paraffins (NP), isoparaffins (IP), olefins (O), cycloparaffins (NC) and aromatic hydrocarbons (A) . Wherein, the isoparaffins also include single-branched isoparaffins (MP), double-branched isoparaffins (DP), and three-branched isoparaffins (TP); the olefins also include straight-chain olefins (NO) , isomerized olefins (BO).

According to the method of the present invention, preferably, the direct generation of the gasoline molecular composition model in step (4) includes: after obtaining the complete monomeric hydrocarbon molecular composition results, establishing a molecule for each of the molecules. object, the molecular object is used to perform the operation of querying molecular properties and molecular content;

A gasoline object is then created, the gasoline object includes the molecular object, and the gasoline object is used to perform the operation of calculating the macroscopic properties of gasoline.

In another aspect, the present invention also provides a system for directly constructing a gasoline molecular composition model, wherein the system includes:

The first unit, the first unit is used to analyze the monomer hydrocarbons on the gas chromatographic detection result of the gasoline sample, to identify the molecules that each peak in the gas chromatogram may contain, and to calculate the relative fraction of each peak;

a second unit for classifying chromatographic peaks by chromatographic peak type according to the results of the monomeric hydrocarbon analysis;

The third unit, the third unit is used to analyze the classified chromatographic peaks according to the chromatographic peak types and molecular types, and obtain a complete monomer hydrocarbon molecular composition result;

A fourth unit for directly generating a composition model of gasoline molecules from the complete monomeric hydrocarbon molecular composition results.

According to the system of the present invention, preferably, in the third unit, the analysis is to use a statistical distribution-based peak adjustment module to analyze the classified chromatographic peaks according to chromatographic peak types and molecular types.

According to the system of the present invention, preferably, the statistical distribution-based peak adjustment module specifically includes:

The first module, the first module is used to fit the distribution of known peaks: the known peaks are classified according to molecular type and carbon number, and each series of data after classification is fitted according to statistical distribution;

The second module, the second module is used for the splitting of the co-evolution peaks: splitting the co-evolution peaks, assuming the relative content of each component in the co-evolution peaks, and checking all the assumptions in turn, and selecting the appropriate one Assume acceptance; repeat the above process until all co-evolution peaks are resolved;

The third module is used to infer unidentified peaks: assume the molecular type and carbon number of the components contained in the unidentified peak, test all the hypotheses in turn, and select the appropriate hypothesis to accept; repeat the above process continuously until All unidentified peaks were inferred.

In another aspect, the present invention also provides a method for predicting macroscopic properties of gasoline, wherein the method includes the following steps:

Construct the composition model of gasoline molecules according to the method for directly constructing the gasoline molecular composition model, so as to obtain the molecular composition of gasoline and the properties of each molecule;

According to the molecular composition of gasoline and the properties of each molecule, the macroscopic properties of gasoline are predicted through the corresponding macro-property mixing rules.

According to the method for predicting macroscopic properties of gasoline according to the present invention, preferably, the macroscopic properties include density, distillation range, octane number, refractive index, molecular weight and Reid vapor pressure.

In yet another aspect, the present invention also provides a system for predicting macroscopic properties of gasoline, wherein the system includes:

The first unit, the first unit is used to construct a composition model of gasoline molecules according to the method for directly constructing a gasoline molecular composition model, so as to obtain the molecular composition of gasoline and the properties of each molecule;

The second unit is used for predicting the macroscopic properties of gasoline according to the molecular composition of gasoline and the properties of each molecule through corresponding macroscopic property mixing rules.

The method provided by the present invention is based on the detection results of gas chromatography, uses the peak adjustment algorithm (SPT algorithm) based on statistical distribution to reconstruct the molecular composition, then establishes a gasoline molecular composition model and predicts the macroscopic properties of gasoline, and the method can be used for gasoline processing. and reconciliation to provide accurate data support.

Compared with the existing method, the method provided by the present invention has the following advantages:

1. The method provided by the present invention combines the advantages of the experimental method and the computer reconstruction method, and provides a complete and stable molecular composition result;

2. The method does not need to measure the spectrum and physical properties of a large number of samples for correlation training, the workload is small, the cost is low, and manpower and material resources are saved;

3. This method can directly predict the properties of gasoline from the molecular composition, which is not affected by the experimental values of macroscopic properties, and has more reference value;

4. The spectrum fine-tuning algorithm provided by this method is directly based on statistical distribution, and does not require macroscopic properties to participate in the correction, and can be realized only by using gas chromatography results.

Description of drawings

1 is a schematic flowchart of a method for directly constructing a gasoline molecular composition model and a method for predicting properties provided in Embodiment 1 of the present invention;

Fig. 2 is the example of the gas chromatogram of gasoline and three types of peaks (known peak, co-evolution peak and unidentified peak) in the embodiment of the present invention 1;

Fig. 3 is the flow chart of SPT algorithm step 1;

Fig. 4 is the flow chart of SPT algorithm step 2;

Fig. 5 is the flow chart of SPT algorithm step 3;

6 is an example diagram of the processing result of the SPT algorithm for the molecular content of n-alkane series (NP) in Example 1 of the present invention;

7 is an example diagram of the processing result of the SPT algorithm for the molecular content of single-branched paraffin series (MP) in Example 1 of the present invention;

8 is an example diagram of the processing result of the SPT algorithm for the molecular content of the double-branched alkane series (DP) in Example 1 of the present invention;

9 is an example diagram of the processing result of the SPT algorithm for the molecular content of three-branched paraffin series (TP) in Example 1 of the present invention;

10 is an example diagram of the processing result of the SPT algorithm for the molecular content of linear olefin series (NO) in Example 1 of the present invention;

11 is an example diagram of the processing result of the SPT algorithm for the molecular content of branched olefin series (BO) in Example 1 of the present invention;

12 is an example diagram of the processing result of the SPT algorithm for the molecular content of naphthenic hydrocarbon series (NC) in Example 1 of the present invention;

13 is an example diagram of the processing result of the SPT algorithm for the molecular content of the aromatic hydrocarbon series (A) in Example 1 of the present invention;

14 is a comparison diagram between the predicted properties of the gasoline sample obtained in Example 1 of the present invention and the experimental values of the properties of the gasoline sample.

Detailed ways

In order to have a clearer understanding of the technical features, purposes and beneficial effects of the present invention, the technical solutions of the present invention are now described in detail below with reference to the following specific examples, but should not be construed as limiting the scope of the present invention.

Example 1

This embodiment provides a method for directly constructing a gasoline molecular composition model from gas chromatography results and a method for predicting macroscopic properties of gasoline. The schematic flowchart of the method is shown in FIG. 1 , and it can be seen from FIG. 1 that it includes the following steps :

Gas chromatographic detection is performed on the catalytically cracked gasoline sample, and then monomer hydrocarbon analysis is performed on the obtained detection results (the petrochemical industry standard of the People's Republic of China SH/T 0714-2002), in order to identify the molecules that each peak in the gas chromatogram may contain, and Calculate the relative fraction of each peak;

Wherein, the gas chromatographic detection described in this example is performed by an Agilent 7890B gas chromatograph from Agilent in the United States, and the gas chromatograph is equipped with a flame ionization detector (FID). Among them, the chromatographic column is an elastic quartz capillary column specially used for PONA analysis, the stationary phase is 100% methyl silicone, the column length is 50 m, the inner diameter is 0.2 mm, and the liquid film thickness is 0.2 μm; the pre-column pressure is 86 KPa; the chromatographic heating program The initial temperature was 35 °C, held for 5 minutes, the heating rate was 2 °C/min, the final temperature was 200 °C, and the final temperature residence time was 10 minutes; the injector temperature was 250 °C, the split ratio was 150:1, and the injection The volume is 0.5 μL; the detector temperature is 250°C, the gas is hydrogen, the flow rate is 35mL/min, the auxiliary gas is air, the flow rate is 350mL/min, the compensation gas is nitrogen, and the flow rate is 35mL/min; the carrier gas is nitrogen, and the average The linear velocity is 12 cm/s, and the gas chromatogram of the catalytically cracked gasoline sample is shown in Figure 2.

(2) According to the analysis results of monomer hydrocarbons, and according to the type of chromatographic peaks, the chromatographic peaks are classified into known peaks, co-evolution peaks and unidentified peaks; examples of each type of chromatographic peaks are shown in Figure 2.

(3) Using the peak adjustment algorithm (SPT) based on gamma distribution to analyze the classified chromatographic peaks according to the chromatographic peak type and molecular type, and obtain the complete monomer hydrocarbon molecular composition results. The specific flow chart of the peak adjustment algorithm is shown in Figure 3-5, and the complete monomer hydrocarbon molecular composition results are shown in Figure 6-13;

In step (3), the molecular types include normal paraffins (NP), isoparaffins, alkenes, naphthenes (NC) and aromatic hydrocarbons (A); wherein the isoparaffins include monobranched isoparaffins ( MP), double-branched isoparaffins (DP), tri-branched isoparaffins (TP); the olefins also include straight-chain olefins (NO), iso-olefins (branched olefins, BO).

The classified monomeric hydrocarbons (known peaks) were processed according to step 1 of Figure 3 to obtain a data matrix. The carbon number of each series in the data matrix and the content fraction under the corresponding carbon number are the data used for fitting. In this example, the data of each series are fitted according to the gamma distribution, and the parameters and root mean square error of the molecular content distribution of each series can be obtained.

Split the co-elution peaks according to step 2 in Figure 4: first assume the relative content of the molecules contained in the co-elution peaks. For example, a co-escape peak contains two components, molecule A and molecule B, it may be assumed that the relative content of molecule A is 1 and the relative content of molecule B is 0 in the co-escape peak. There can be many such hypotheses, and we can build a list of hypotheses and test the hypotheses individually.

After testing all the hypotheses, select the appropriate hypothesis, accept the relative content of the co-escape components, classify and update the data matrix. Perform a gamma fit on the updated data to obtain new parameters and RMSE.

Infer unidentified peaks as in Step 3 of Figure 5: First assume that the unidentified peak contains only one component. Then make assumptions about the possible carbon number of the component, and the possible molecular type. For example, suppose that an unidentified peak contains a molecule of type Aromatic with 10 carbons. Likewise, we group the various hypotheses together, build a list of hypotheses, and test the hypotheses individually. After testing all hypotheses, select the appropriate hypothesis, accept the molecular type and carbon number in this hypothesis, categorize and update the data matrix. Perform a gamma fit on the updated data to obtain new parameters and RMSE.

Figures 6 to 13 respectively show schematic diagrams of changes in the content distribution data of each molecular series after being processed by steps 1 to 3 of the peak adjustment algorithm. In this example, the molecular types will be divided into 8 categories, namely NP, MP, DP, TP, NO, BO, NC, A. The processing results of these 8 types correspond to Figures 6-13 respectively. Each graph contains three subgraphs Step 1 - Step 3, corresponding to the three steps of the peak adjustment algorithm. The abscissas of these subgraphs are all carbon numbers, and the ordinates are all mass fractions.

Taking Fig. 8 corresponding to the double-branched paraffin data as an example, the results of the three steps of the peak adjustment algorithm are presented. For the convenience of description, the three sub-graphs in FIG. 8 are referred to as FIG. 8-1, FIG. 8-2 and FIG. 8-3, respectively. The scatter in Figure 8-1 is the molecular content of the double-branched alkane series in the data matrix obtained after step 1 of the peak adjustment algorithm is completed; the dotted line is the gamma distribution represented by the parameters Parameters _{1, DP} fitted in step 1. Similar to Figure 8-1, the scattered points in Figure 8-2 and Figure 8-3 are the molecular content of the double-branched alkane series in the data matrix obtained after Step 2 and Step 3 in the peak adjustment algorithm, respectively, and the blue dotted lines are The gamma distribution represented by Parameters _{2, DP} and Parameters _{3, DP} . It can be seen that the deviation of the scatter points from the dashed line in Figure 8-1 is relatively large, especially the scatter point located at C7 is significantly lower than the dashed line. This is because some bibranched alkane molecules exist in the form of co-evolution peaks or unidentified peaks, so the content of these molecules is not counted into the data matrix of step 1. As can be seen from Figure 8-2, when step 2 of the peak adjustment algorithm, that is, after the splitting of the co-escape peaks, the distribution of scatter points is much closer to the dotted line.

Figure 8-3 is not significantly improved relative to Figure 8-2, for two reasons. One is that the content fraction of unidentified peaks is too low to make the change not obvious; the other is that the content distribution of other molecular types deviates further from the gamma distribution, so that the peak adjustment algorithm can be used in the calculation process of step 3. , more likely to infer unidentified peaks as molecules of this type. This situation can be seen in Figure 9. Figure 9-2 shows that even after the coevolution peaks are resolved, the content distribution of three-branched paraffins is still far from the ideal gamma distribution. Therefore, the algorithm would prefer to infer unidentified peaks at step 3 to be molecules of type three-branched alkanes. As a result, the distribution of scatter points in Figure 9-3 has been significantly improved.

Look again at Figure 13 for the corresponding aromatics data. Figure 13-1 shows that the aromatics content at C7 is zero. This is because toluene always co-elutes with 2,3,3-trimethylpentane in the GC-FID experimental environment. After processing by the peak adjustment algorithm, a relatively reasonable toluene content can be obtained. In addition, it is worth mentioning that all 11 n-alkanes in the molecular library can usually be identified by GC-FID without co-elution. Therefore, all images in Figure 6 corresponding to the n-alkane data are the same.

It can be seen that the distribution curve of the gasoline molecular composition reconstructed by the peak adjustment algorithm is more reasonable.

(4) Using software (property prediction module, more specifically a molecular composition model) to directly generate a composition model of gasoline molecules from the complete monomeric hydrocarbon molecule composition results, which specifically includes: reading molecular composition information, instantiating gasoline object, each molecule in the molecular composition is instantiated as a molecular object contained in the gasoline object;

(5) obtain the molecular composition of gasoline and the properties of each molecule according to the composition model of the gasoline molecules, and then predict the macroscopic properties of gasoline by mixing rules from the molecular composition of the obtained gasoline and the properties of each molecule; it specifically includes: The gasoline molecular composition model is to establish gasoline objects through molecular composition. The objects include molecular objects established by each molecule. The molecular objects can perform a series of operations such as querying molecular properties and relative content of molecules in gasoline. The properties of each molecule were obtained from the NIST database. The gasoline object can perform a series of operations such as calculating the macroscopic properties of gasoline. The macroscopic properties are calculated by the molecular properties of each molecule, the relative content, and the corresponding macroscopic property mixing rules.

Among them, the predicted values are shown in Table 1. Figure 14 shows the comparison between the predicted properties of gasoline samples and the experimental values of gasoline sample properties (measured by conventional methods in the art).

Table 1

Note: The experimental values of the macroscopic properties of the catalytically cracked gasoline samples in Table 1 are all measured by conventional methods in the art, specifically, the specific gravity is measured according to GBT 1884-2000, and the distillation range is measured according to GBT 6536 - Measured in 2010, the research octane number is measured according to GBT 5487-1995, the hydrocarbon content is measured according to GBT 11132-2008, and the Reid vapor pressure (KPa) is measured according to GBT 8017-2012.

At present, the existing molecular-level composition models of gasoline in the field usually adjust the composition according to the experimental values of macroscopic properties, and then predict the macroscopic properties of gasoline from the obtained composition. Therefore, the model is greatly affected by the input experimental values of macroscopic properties. If the experimental errors of the macroscopic properties of gasoline are large, the predicted values of this model will also have large errors. Different from this model, the model provided by this application is only based on the results obtained after GC-FID analysis and peak adjustment algorithm processing when predicting gasoline properties. Therefore, it is not affected by the experimental values of gasoline properties, so it has more reference. value.

At present, the prediction of the initial boiling point and final boiling point of the ASTM D86 distillation curve is usually difficult. As can be seen from Table 1, it can also obtain good results in the model provided in this application; the existing olefins, naphthenes, The experimental value of the volume fraction of aromatic hydrocarbons is measured by fluorescence method, but this method has large errors and poor reproducibility, while the volume fraction predicted by the model in this application is based on the monomer obtained by GC-FID analysis The results obtained from the hydrocarbon mass fraction are more accurate and reproducible. Therefore, the volume fraction predicted by the model in this application should be more credible; in addition, it can be seen from Table 1 that the octane number and Reid vapor pressure obtained by the method provided by the present invention are also important properties. Has a good prediction effect.

Claims (9)

1. a method for directly constructing gasoline molecular composition model, is characterized in that, described method may further comprise the steps: (1) carry out monomer hydrocarbon analysis to the gas chromatographic detection result of gasoline sample, to identify the molecules that each peak may contain in the gas chromatogram, and calculate the relative fraction of each peak; (2) according to the monomer hydrocarbon analysis result, and classify the chromatographic peaks according to the chromatographic peak types; the chromatographic peak types include known peaks, co-evolution peaks and unidentified peaks; (3) analyze the classified chromatographic peaks according to the chromatographic peak type and molecular type, and obtain the complete monomer hydrocarbon molecular composition result; The analysis described in the step (3) is to adopt the peak adjustment algorithm based on statistical distribution to analyze the chromatographic peak after the classification according to the chromatographic peak type and the molecular type; The statistical distribution-based peak adjustment algorithm specifically includes the following steps: Fit the distribution of known peaks: classify known peaks according to molecular type and carbon number, and fit the classified data according to statistical distribution; Separation of co-escape peaks: Split the co-escape peaks, assuming the relative content of each component in the co-escape peaks, test all the hypotheses in turn, and select the appropriate hypothesis to accept; repeat the above process continuously until all the The co-evolution peaks are all split; Infer unidentified peaks: Assume the molecular type and carbon number of the components contained in the unidentified peaks, test all the hypotheses in turn, and select the appropriate hypothesis to accept; repeat the above process continuously until all unidentified peaks are inferred; (4) directly generating the composition model of gasoline molecules from the complete monomer hydrocarbon molecule composition results; the directly generating gasoline molecular composition model described in step (4) includes: obtaining the complete monomer hydrocarbon molecule composition results Then, create a molecular object for each of the molecules, and the molecular object is used to perform the operation of querying molecular properties and molecular content; A gasoline object is then created, the gasoline object includes the molecular object, and the gasoline object is used to perform the operation of calculating the macroscopic properties of gasoline. 2. The method according to claim 1, wherein the gasoline sample comprises catalytically cracked gasoline, catalytically reformed gasoline, straight-run gasoline, catalytically cracked gasoline, hydrogenated gasoline or coker gasoline. 3. The method according to claim 1 or 2, wherein the statistical distribution comprises a gamma distribution. 4 . The method according to claim 1 or 2 , wherein the molecular types in step (3) include normal paraffins, isoparaffins, olefins, naphthenes and aromatic hydrocarbons. 5 . 5 . The method according to claim 3 , wherein the molecular types in step (3) include normal paraffins, isoparaffins, olefins, naphthenes and aromatic hydrocarbons. 6 . 6. A system for directly constructing a gasoline molecular composition model, wherein the system comprises: The first unit, the first unit is used to analyze the monomer hydrocarbons on the gas chromatographic detection result of the gasoline sample, to identify the molecules that each peak in the gas chromatogram may contain, and to calculate the relative fraction of each peak; The second unit, the second unit is used to classify the chromatographic peaks according to the chromatographic peak types according to the monomer hydrocarbon analysis results; the chromatographic peak types include known peaks, co-elution peaks and unidentified peaks; The third unit, the third unit is used to analyze the classified chromatographic peaks according to the chromatographic peak types and molecular types to obtain the complete monomer hydrocarbon molecular composition results; in the third unit, the analysis is based on The peak adjustment module of statistical distribution analyzes the classified chromatographic peaks according to the chromatographic peak type and molecular type; The statistical distribution-based peak adjustment module specifically includes: The first module, the first module is used to fit the distribution of known peaks: the known peaks are classified according to molecular type and carbon number, and each series of data after classification is fitted according to statistical distribution; The second module, the second module is used for the splitting of the co-evolution peaks: splitting the co-evolution peaks, assuming the relative content of each component in the co-evolution peaks, and checking all the assumptions in turn, and selecting the appropriate one Assume acceptance; repeat the above process until all co-evolution peaks are resolved; The third module is used to infer unidentified peaks: assume the molecular type and carbon number of the components contained in the unidentified peak, test all the hypotheses in turn, and select the appropriate hypothesis to accept; repeat the above process continuously until All unidentified peaks were inferred; The fourth unit, the fourth unit is used to directly generate the composition model of gasoline molecules from the complete monomeric hydrocarbon molecule composition result; the fourth unit is specifically used for: obtaining the complete monomeric hydrocarbon molecule composition result Then, create a molecular object for each of the molecules, and the molecular object is used to perform the operation of querying molecular properties and molecular content; A gasoline object is then created, the gasoline object includes the molecular object, and the gasoline object is used to perform the operation of calculating the macroscopic properties of gasoline. 7. A method for predicting macroscopic properties of gasoline, characterized in that the method comprises the following steps: Build a composition model of gasoline molecules according to the method for directly constructing a gasoline molecular composition model according to any one of claims 1-5 to obtain the molecular composition of gasoline and the properties of each molecule; According to the molecular composition of gasoline and the properties of each molecule, the macroscopic properties of gasoline are predicted through the corresponding macro-property mixing rules. 8. The method of claim 7, wherein the macroscopic properties include density, distillation range, octane number, refractive index, molecular weight, and Reid vapor pressure. 9. A system for predicting macroscopic properties of gasoline, wherein the system comprises: The first unit, the first unit is used to construct a composition model of gasoline molecules according to the method for directly constructing a gasoline molecular composition model according to any one of claims 1-5, to obtain the molecular composition of gasoline and the properties of each molecule; The second unit is used for predicting the macroscopic properties of gasoline according to the molecular composition of gasoline and the properties of each molecule through corresponding macroscopic property mixing rules.

Images