A method and apparatus for identifying crude oil by near infrared spectroscopy

Abstract

The application provides a method and device for identifying crude oil by near-infrared spectrum, which comprises the following steps: determining the near-infrared spectrum of a crude oil sample to be detected and performing second-order differential processing, selecting the absorbance of a characteristic spectral region to form a spectrum vector x; for a first physical property, solving X=T×P by a PLS algorithm to obtain a score vector of the spectrum vector x, for a second physical property, solving X=T×P by a PLS algorithm to obtain a score vector of the spectrum vector x, splicing the two score vectors at the head and tail to obtain a score combination vector k; calculating each moving correlation coefficient of the score combination vector k and each sample in a crude oil near-infrared score database, calculating an identification parameter Q of each database sample i ; calculating a threshold value Q t ; if Q i >Q t , and each moving correlation coefficient of the i sample is not less than a predetermined value, then the crude oil to be identified is identical to the i sample. The scheme can effectively improve the accuracy of crude oil identification and improve the identification calculation speed.

Description

Technical Field

[0001] This invention relates to a method for identifying crude oil types, specifically a method and apparatus for identifying crude oil types using near-infrared spectroscopy. Background Technology

[0002] Crude oil evaluation plays a vital role in crude oil extraction, trading, and processing. Although a relatively complete set of crude oil evaluation methods has been established, these methods are time-consuming, labor-intensive, and costly, far from meeting the needs of practical applications. Therefore, large petrochemical companies both domestically and internationally are currently developing rapid crude oil evaluation technologies based on various modern instrumental analysis methods, including GC-MS, NMR, NIR, and IR. Among these, the NIR method is highly favored due to its convenient measurement, speed, and applicability to on-site or online analysis.

[0003] Unlike NIR spectroscopy for other petroleum products such as gasoline and diesel, crude oil has numerous evaluation indicators. For example, there are dozens of general properties of crude oil alone, and hundreds if the properties of each fraction are included. Using traditional factor analysis methods such as partial least squares (PLS) to establish calibration models for each property is obviously too cumbersome. Combining near-infrared spectroscopy with a crude oil property database is one of the better technical approaches to solving this problem. This involves identifying the crude oil under test using NIR spectra, determining its type from the NIR spectral library, and then retrieving its evaluation data from the existing crude oil property database. This achieves rapid crude oil evaluation and provides a simple method for obtaining timely evaluation data to determine crude oil processing schemes and optimize production decisions.

[0004] CN101995389A discloses "A Method for Rapidly Identifying Crude Oil Types Using Near-Infrared Spectroscopy," which proposes a method for rapid identification of crude oil using near-infrared spectroscopy based on the moving window concept combined with the traditional correlation coefficient method—the moving window correlation coefficient method. This method can accurately identify crude oil types and, combined with a crude oil evaluation database, can quickly provide property data of the crude oil to be tested. It is a simple and reliable method for rapid prediction of crude oil evaluation and analysis data. However, in practical applications, the chemical composition of the same crude oil extracted at different times often varies, or other types of crude oil may be mixed in during storage and transportation, causing the properties of the crude oil to change accordingly. Therefore, the method of CN101995389A often fails to identify the crude oil type that is completely consistent with the crude oil to be tested from the crude oil near-infrared spectral library, greatly limiting the application scope of this rapid identification technology. Moreover, the algorithm requires identification of all samples in the database one by one, and the large number of identification variables leads to a long calculation time. Summary of the Invention

[0005] The purpose of this invention is to provide a method and apparatus for identifying crude oil using near-infrared spectroscopy, which can effectively improve the accuracy of crude oil identification and increase the speed of identification calculation.

[0006] To achieve the above objectives, embodiments of the present invention provide a method for identifying crude oil types using near-infrared spectroscopy, comprising the following steps:

[0007] The near-infrared spectrum of the crude oil sample to be tested was measured and processed by second-order differentiation. The absorbance of the characteristic spectral region was selected to form the spectral vector x.

[0008] For the first property, the score vector of the spectral vector x is obtained by solving X = T × P using the PLS algorithm. For the second property, the score vector of the spectral vector x is obtained by solving X = T × P using the PLS algorithm. The two score vectors are then concatenated to obtain the score combination vector k.

[0009] Calculate the correlation coefficient between the score combination vector k and each movement of each sample in the crude oil near-infrared score database, and calculate the identification parameter Qi for each database sample according to equation (I).

[0010]

[0011] In equation (Ⅰ), r ji denoted as , where i is the sample number in the crude oil near-infrared scoring database, j is the moving window number, n is the total number of moving windows, and m is the total number of samples in the crude oil near-infrared scoring database.

[0012] Calculate the threshold Qt, Q t = (dw-0.15), where d is the number of data points and w is the width of the moving window;

[0013] If Q i If Qt > 1, and each moving correlation coefficient of sample i is not less than a predetermined value, then the crude oil to be identified is the same as sample i in the crude oil near-infrared score database.

[0014] The crude oil near-infrared score database is obtained through the following operations:

[0015] Various types of crude oil samples were collected, and their first and second physical property data for the first and second physical properties were measured. Their near-infrared spectra were also measured, and second-order differential processing was performed. The absorbance of the characteristic spectral regions and the corresponding physical property data of the crude oil samples were selected to establish a near-infrared spectral database X for crude oil samples.

[0016] For the first physical property, the score matrix T1 of the near-infrared spectral database X of the crude oil sample is obtained by solving X = T × P using the PLS algorithm. For the second physical property, the score matrix T2 of the near-infrared spectral database X of the crude oil sample is obtained by solving X = T × P using the PLS algorithm. The score matrix T1 and the score matrix T2 are concatenated end to end to obtain a new score combination matrix K, which constitutes the near-infrared score database of crude oil.

[0017] The first and second physical property data are sulfur content data and acid value data, respectively.

[0018] The characteristic spectral region is 4628–4000 cm⁻¹. -1 and 6076~5556cm -1 .

[0019] The predetermined value is 0.9920.

[0020] If the identification parameters of all samples in the crude oil near-infrared scoring database are not greater than Qt, or if the moving correlation coefficient of any sample is greater than the predetermined value, then there is no sample in the crude oil near-infrared scoring database that is the same as the crude oil to be identified.

[0021] The sampling point interval for the absorbance of the near-infrared spectrum is 1 to 10 wavenumbers.

[0022] The moving correlation coefficient is determined using a moving window.

[0023] The width of the moving window is 3 to 15 sampling points.

[0024] The method for determining the moving correlation coefficient is as follows: In the data interval, select the width of a moving window from a fixed direction (forward or backward), calculate the correlation coefficient between the crude oil sample to be identified in the window and the score of each sample in the database, and then move the moving window forward by several sampling points to form the next moving window. Calculate the correlation coefficient between the crude oil sample to be identified in this moving window and the score of each sample in the database. Continuously move the moving window in the above method and calculate the correlation coefficient between the crude oil sample to be identified in each moving window and the score of each sample in the database.

[0025] The distance of each movement of the window is 1 to 10 sampling points.

[0026] The number of data points refers to the principal components of the PLS, which is 30-120.

[0027] The method further includes: determining the properties of the crude oil to be identified based on the identified sample i that is identical to the crude oil to be identified.

[0028] On the other hand, the present invention provides an apparatus for identifying crude oil types by near-infrared spectroscopy, the apparatus comprising: a memory; and a processor configured to utilize the above-described method for identifying crude oil types by near-infrared spectroscopy.

[0029] On the other hand, the present invention provides a machine-readable storage medium storing instructions, characterized in that, when executed by a processor, the instructions cause the processor to be configured to perform the above-described method for identifying crude oil types by near-infrared spectroscopy.

[0030] Principal Component Analysis (PCA) is a classic algorithm in chemometrics. Its basic model is X = T × P, which decomposes the spectral matrix X into the product of a T matrix (score matrix) and a P matrix (loading matrix). Generally, PCA only decomposes the X matrix (spectral matrix) into the product of the T matrix (score matrix) and the P matrix (loading matrix). The score matrix is ​​characterized by the fact that the first few variables essentially cover most of the information in the original matrix, thus greatly compressing the amount of data required for calculation. Therefore, in this case, the score is used to replace the original spectral matrix for calculation. However, the score here does not guarantee the highest correlation with the sample's physical properties and composition (y matrix). Therefore, this case uses the PLS method to extract principal components, because the PLS decomposition of the X matrix must consider maintaining the highest correlation with the sample's physical properties and composition. The resulting score variables then have a high correlation with the sample's physical properties and composition. For example, the score matrix T1 has a high correlation with the first physical property data (e.g., sulfur content data), and the score matrix T2 has a high correlation with the second physical property data (acid value data). The purpose of this processing is to incorporate the physical properties of the sample for identification calculations. Principal component analysis (PCA) is characterized by its ability to significantly compress data, a general conclusion based on algorithmic principles. Computational speed inevitably increases after data compression. For example, the original spectrum might have thousands of data points, while compression might only require a few dozen. Therefore, the above technical solution can improve the speed of identification calculations.

[0031] Furthermore, compared with the existing proprietary moving window correlation coefficient method, this method, by incorporating crude oil property information, yields more accurate and reliable identification results.

[0032] In addition, the present invention identifies samples that are identical to the crude oil to be tested, and obtains the properties of the crude oil to be tested from the properties of the crude oil corresponding to the sample, because the spectrum reflects the vibrational information of the molecular skeleton, and if the spectra are consistent, the physical properties are consistent.

[0033] Other features and advantages of the embodiments of the present invention will be described in detail in the following detailed description section. Attached Figure Description

[0034] The accompanying drawings are provided to further illustrate embodiments of the present invention and form part of the specification. They are used together with the following detailed description to explain the embodiments of the present invention, but do not constitute a limitation thereof. In the drawings:

[0035] Figure 1 This is a flowchart of the method for identifying crude oil using near-infrared spectroscopy provided by the present invention. Detailed Implementation

[0036] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the scope of the present invention.

[0037] To achieve the above objectives, embodiments of the present invention provide a method for identifying crude oil types using near-infrared spectroscopy, comprising the following steps:

[0038] S110: Measure the near-infrared spectrum of the crude oil sample to be tested and perform second-order differential processing, and select the absorbance of the characteristic spectral region to form the spectral vector x;

[0039] S120: For the first physical property, the score vector of the spectral vector x is obtained by solving X = T × P using the PLS algorithm. For the second physical property, the score vector of the spectral vector x is obtained by solving X = T × P using the PLS algorithm. The two score vectors are concatenated end to end to obtain the score combination vector k.

[0040] Here, the PLS method is used to calculate X = T × P, decomposing the spectral matrix X into a T matrix (term: score matrix) and a P matrix (term: load matrix). The decomposition process considers Y to ensure that it is highly correlated with Y (physical properties or composition). The P matrix here can be understood as a new coordinate axis. The score vector p of the sample is obtained by directly projecting the spectral vector x of the sample onto the new coordinate axis (P).

[0041] S130: Calculate the correlation coefficient between the score combination vector k and each movement of each sample in the crude oil near-infrared score database, and calculate the identification parameter Qi of each database sample according to equation (Ⅰ).

[0042]

[0043] In equation (Ⅰ), r ji denoted as , where i is the sample number in the crude oil near-infrared scoring database, j is the moving window number, n is the total number of moving windows, and m is the total number of samples in the crude oil near-infrared scoring database.

[0044] S140: Calculate the threshold Qt, Q t= (dw-0.15), where d is the number of data points and w is the width of the moving window;

[0045] S150: If Q i Q t If each movement correlation coefficient of sample i is not less than a predetermined value, then the crude oil to be identified is the same as sample i in the crude oil near-infrared score database.

[0046] The crude oil near-infrared score database is obtained through the following operations:

[0047] Various types of crude oil samples were collected, and their first and second physical property data for the first and second physical properties were measured. Their near-infrared spectra were also measured, and second-order differential processing was performed. The absorbance of the characteristic spectral regions and the corresponding physical property data of the crude oil samples were selected to establish a near-infrared spectral database X for crude oil samples.

[0048] For the first physical property, the score matrix T1 of the near-infrared spectral database X of the crude oil sample is obtained by solving X = T × P using the PLS algorithm. For the second physical property, the score matrix T2 of the near-infrared spectral database X of the crude oil sample is obtained by solving X = T × P using the PLS algorithm. The score matrix T1 and the score matrix T2 are concatenated end to end to obtain a new score combination matrix K, which constitutes the near-infrared score database of crude oil.

[0049] It should be noted that the aforementioned crude oil near-infrared score database can be established in advance before the above method is executed.

[0050] The technical solution of the present invention will now be described in conjunction with specific implementation details. However, the present invention is not limited to these descriptions, and other implementations that can be foreseen by those skilled in the art after reading this application are also feasible. Furthermore, in this application, the terms "physical property," "nature," and "composition" are used interchangeably.

[0051] The method of this invention uses a moving window in the characteristic spectral region of 4628–4000 cm⁻¹ -1 and 6076~5556cm -1A spectral database of the crude oil sample to be tested and a database of samples is established. Using the near-infrared spectral database X and sulfur content data, the score matrix T1 is obtained by solving X = T × P using the PLS algorithm. Similarly, the score matrix T2 is obtained by solving X = T × P using the near-infrared spectral database X and acid value data. The scores T1 and T2 from the near-infrared spectral database X are concatenated to obtain a new score combination matrix K. The spectra of the crude oil sample to be tested undergo the same processing to obtain the score combination vector k. Using the moving correlation coefficient method, the identification parameters of the sample to be tested and each sample in the database are calculated from the moving correlation coefficient of each window during the moving process. Database samples with identification parameters greater than a threshold are selected. If each moving correlation coefficient is not less than 0.9920, the sample to be identified is considered to be the same type of crude oil as the sample in the database. This method can accurately and quickly identify the same crude oil type as the sample to be tested in the spectral database. Combined with a crude oil evaluation database, it can quickly provide the property data of the crude oil to be tested, resulting in a simple and reliable method for rapid prediction of crude oil evaluation and analysis data. Compared with the existing patented moving window correlation coefficient method, this method, by incorporating crude oil property information, yields more accurate and reliable identification results.

[0052] The method of the present invention includes the following steps:

[0053] (1) To establish a near-infrared spectral database of crude oil samples, the number of representative crude oil samples collected is preferably 200 to 800. Various required physical properties of the crude oil samples are determined using standard methods, and their characteristic spectral regions are processed by second-order differential analysis, selecting a range of 4628–4000 cm⁻¹. -1 and 6076~5556cm -1 The absorbance in the spectral region was used to establish a near-infrared spectral database X for crude oil samples.

[0054] (2) The crude oil sample to be tested is processed in the same way as described above.

[0055] (3) The next step is to construct a score combination matrix for identification. The method is to use the near-infrared spectral database X and sulfur content data to solve X = T × P using the PLS algorithm to obtain its score matrix T1, and use the near-infrared spectral database X and acid value data to solve X = T × P using the PLS algorithm to obtain its score matrix T2. The score T1 and the score matrix T2 of the near-infrared spectral database X are concatenated to obtain a new score combination matrix K, thereby constructing a near-infrared score database for crude oil samples; in addition, the same processing is performed on the spectrum of the crude oil sample to be tested to obtain the score combination vector k.

[0056] (4) The next step is to identify the crude oil sample to be tested in the established near-infrared score database of crude oil samples and calculate its identification parameter Q relative to each sample in the database. i .

[0057] (5) The threshold is calculated, and the threshold is used as one of the criteria for judging successful identification. The number of data points is the principal component number of PLS, which is approximately 30-120.

[0058] The near-infrared spectrum described in this invention refers to the absorbance corresponding to each sampling point within the scanning wavenumber range. The sampling point interval is 1 to 10 wavenumbers, which is determined by the resolution of the infrared spectrometer.

[0059] The moving correlation coefficient is determined using a moving window, with a width of 3 to 15 sampling points.

[0060] This invention uses the sum of moving correlation coefficients as an identification parameter, which serves as one of the conditions for determining whether the crude oil sample to be tested is the same as the crude oil sample in the spectral database.

[0061] Traditional correlation coefficients are often used to compare the similarity between two spectra. However, in this case, the correlation coefficient is used to compare the similarity between two concatenated score vectors. The spectral variables mentioned below are all converted into score data. During the calculation, all spectral variables participate in the operation, and finally a correlation coefficient value is obtained. The calculation formula is shown in Equation (II):

[0062]

[0063] In formula (II), Let be the mean absorbance of all sampling points in the i-th and j-th spectra, respectively, where d is the number of sampling points, k is the sampling point index, and x is the average absorbance of all sampling points in the i-th and j-th spectra, respectively. ik Let x be the absorbance at the k-th sampling point of the i-th spectrum. jk Let be the absorbance at the k-th sampling point of the j-th spectrum. The closer two spectra are, the closer their similarity coefficient is to 1 or -1.

[0064] The moving correlation coefficient described in this invention refers to the calculation of the moving correlation coefficient for each moving window interval of two concatenated score vectors to be compared, using the traditional correlation coefficient formula (II). This yields a series of moving correlation coefficients for sub-intervals. The sub-wavenumber interval is the width of a moving window.

[0065] The moving window is a spectral window of width w, selected and moved from the first wavenumber sampling point of the entire spectrum. The movement distance is defined as the sampling interval of one or more wavenumbers, continuing until the last wavenumber sampling point. The moving window can move in either direction, from a sampling point with a lower wavenumber to one with a higher wavenumber, or vice versa. The preferred distance for each movement is 1 to 10 sampling points.

[0066] The preferred method for determining the moving correlation coefficient using a moving window is as follows: within the characteristic spectral range, i.e., 4628–4000 cm⁻¹. -1 and 6076~5556cm -1 The spectral range and score data are combined. A window movement direction is arbitrarily set, and a window width is selected. The correlation coefficient between the crude oil sample to be identified within this window and the score of each sample in the database is calculated. The window is then moved several more sampling points to form the next moving window. The correlation coefficient between the crude oil sample to be identified within this moving window and the absorbance of each sample in the database is calculated. This process is repeated continuously, calculating the correlation coefficient between the crude oil sample to be identified and each sample in the database within each moving window. This yields the moving correlation coefficient for each moving window. The obtained correlation coefficient values ​​are plotted against the starting position of the corresponding moving window to obtain the moving correlation coefficient graph. This graph easily shows the similarity between two spectra. If two spectra are completely identical, the moving correlation coefficient value across the entire spectral range is 1. If the two spectra differ only in a certain interval, the correlation coefficient value in that interval will decrease significantly. Clearly, compared to the traditional correlation coefficient based on the entire spectrum, the moving correlation coefficient can distinguish between two spectra with subtle differences, improving the accuracy of spectral identification and facilitating the extraction of implicit information.

[0067] In calculating the moving correlation coefficient, the width of the moving window should be a fixed value. A moving window that is too small, while helpful for identifying detailed information, carries the risk of inaccurately identifying the same type of crude oil. A moving window that is too large, while able to eliminate the influence of external testing conditions such as temperature and humidity, carries the risk of misidentification.

[0068] The method for identifying crude oil types using the identification parameters described in this invention is as follows: Calculate the moving correlation coefficient between the scores of all samples in the spectral database and the score of the crude oil sample to be identified; sum the moving correlation coefficients of all moving windows to obtain the identification parameter Q for each database sample. Then, Q... i With threshold Q t In contrast, if there exists a value greater than Q t If the database samples are selected, they are first filtered out. Then, it is determined whether the shift correlation coefficient values ​​of the spectra of all samples in the database are not less than 0.9920. If all are not less than 0.9920, it can be determined that the crude oil sample to be tested is of the same type as the sample selected from the database. If the identification parameters of all samples in the database are not greater than Q... t If no sample has a moving correlation coefficient greater than 0.9920, then there is no sample type in the database that is the same as the crude oil to be identified.

[0069] The method of this invention is applicable to the rapid identification of unknown crude oil samples and known crude oil samples. It can quickly determine whether unknown crude oil and known crude oil are of the same type of crude oil through near-infrared spectroscopy, so as to quickly evaluate the properties of unknown crude oil by using the properties of known crude oil.

[0070] The present invention will be described in detail below with examples, but the present invention is not limited thereto.

[0071] In this example, the instrument used to determine the near-infrared spectrum of crude oil was a Thermo Antaris II Fourier transform near-infrared spectrometer, with a spectral range of 3800–10000 cm⁻¹. -1 8cm resolution -1 The cumulative number of scans was 64, using transmission measurement method.

[0072] The conventional methods for determining the physical properties of crude oil samples are as follows:

[0073] Density: GB / T 13377 Determination of density or relative density of crude oil and liquid or solid petroleum products.

[0074] Acid value: GB / T 7304 Determination of acid value of petroleum products by potentiometric titration.

[0075] Carbon residue: GB / T 17144 Determination of carbon residue in petroleum products.

[0076] Sulfur content: GB / T17040 Determination of Sulfur Content in Petroleum Products

[0077] Nitrogen content: GB / T 17674 Determination of nitrogen content in crude oil,

[0078] Wax content: SY / T 0537 Determination of wax content in crude oil,

[0079] Asphaltene and gum content: SY / T 7550 Determination of wax, gum and asphaltene content in crude oil

[0080] True boiling point distillation data: GB / T 17280 Standard Test Method for Crude Oil Distillation.

[0081] Example 1

[0082] (1) Establish a near-infrared spectral database of crude oil samples

[0083] Seven hundred representative crude oil samples were collected, covering most of the world's major crude oil producing regions. The near-infrared spectra of the crude oil samples were measured, and their second-order derivatives were applied, selecting a range of 4628.0–4000.0 cm⁻¹. -1 and 6076.0~5556.0cm -1The absorbance in the spectral region was used to establish a near-infrared spectral data matrix X for crude oil samples. The dimension of X is 700×289, where 700 represents the number of crude oil samples collected and 289 represents the number of sampling points for near-infrared spectral absorbance.

[0084] The crude oil property matrix Y is composed of 31 property data from these 700 crude oil samples, including density, acid value, carbon residue, sulfur content, nitrogen content, wax content, gum content, asphaltene content, and true boiling point distillation data (TBP, cumulative mass yield at 23 temperature points: 65℃, 80℃, 100℃, 120℃, 140℃, 165℃, 180℃, 200℃, 220℃, 240℃, 260℃, 280℃, 300℃, 320℃, 350℃, 380℃, 400℃, 425℃, 450℃, 470℃, 500℃, 520℃, 540℃). The dimension of Y is 700×31, where 700 is the number of crude oil samples collected and 31 is the number of crude oil properties.

[0085] A near-infrared spectral database of crude oil samples is established using the near-infrared spectral matrix X of the crude oil sample and the corresponding property matrix Y of the crude oil sample. That is, in matrices X and Y, the same index represents the absorbance and property data of the same crude oil sample.

[0086] (2) Establish the absorbance vector of the crude oil to be identified.

[0087] The near-infrared spectrum of unknown crude oil A was measured under the same conditions as those used to establish the spectral database. The second derivative of its near-infrared spectrum was then performed, and the values ​​in the range of 4628.0–4000.0 cm⁻¹ were calculated. -1 and 6076.0~5556.0cm -1 The absorbance in the spectral region constitutes the spectral vector x. A Its dimension is 1×289.

[0088] (3) Construct the score combination matrix K

[0089] The score matrix T1 is obtained by solving X = T × P using the near-infrared spectral database X and sulfur content data through the PLS algorithm. The scores of the first 50 principal components of X are then taken. The score matrix T2 is obtained by solving X = T × P using the near-infrared spectral database X and acid value data through the PLS algorithm. The scores of the first 50 principal components of X are then taken. The score matrix T1 obtained by solving X = T × P using the near-infrared spectral database X and sulfur content data through the PLS algorithm and the score matrix T2 obtained by solving X = T × P using the acid value data through the PLS algorithm are then concatenated to obtain a new score combination matrix K with a dimension of 700 × 100. The spectrum of the crude oil sample to be tested is processed in the same way to obtain the score combination vector k with a dimension of 1 × 100.

[0090] (4) Identification of crude oil samples to be tested

[0091] For the score combination matrix K and the score combination vector k, a corresponding data interval is selected starting from the lowest wavelength point as a moving window. The moving window has an interval of 8 sampling points, i.e., the width of the moving window is 9 sampling points. Within the first moving window, the correlation coefficient between the absorbance of each score combination in K and the score combination vector k within this characteristic range is calculated, denoted as r. 1,i Let i = 1, 2, ..., 700. Then, move the moving window by one sampling point, that is, calculate the correlation coefficient between the absorbance of each score combination in K and the score combination vector k in this feature range within the next interval, denoted as r. 2,i Let i = 1, 2, ..., 700. Continue this process until the correlation coefficient of the last moving window is calculated, denoted as r. 92,i , i = 1, 2, ..., 700.

[0092] Calculate the identification parameter Q for each library spectral pair k in the spectral vector X using the following formula:

[0093]

[0094] Threshold Q t =dw-0.15=100-9-0.25=90.75.

[0095] For the unknown crude oil sample A, the spectrum was calculated to be consistent with that of crude oil samples No. 6, 11, 20, 77 and 108 in the spectral database X, and the moving correlation coefficient value of each sample was not less than 0.9920, and the calculation time was short.

[0096] Table 1

[0097]

[0098] (6) Predicting crude oil properties

[0099] The evaluation data of successfully identified crude oil samples are the evaluation data of the samples to be tested. The prediction results are shown in Table 2.

[0100] Table 2

[0101]

[0102] Comparative Example 1

[0103] The crude oil in the same storage tank as the crude oil to be tested, A, was identified from the crude oil spectral database using the method in CN101995389A.8. The results are shown in Table 3. The identification results are the same as in Example 1, but the time taken is longer than in Example 1.

[0104] Table 3

[0105]

[0106] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0107] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0108] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0109] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0110] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0111] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0112] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0113] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0114] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for identifying crude oil types using near-infrared spectroscopy, comprising the following steps: The near-infrared spectrum of the crude oil sample was measured and subjected to second-order differential processing. The absorbance in the characteristic spectral region was selected to construct the spectral vector. x ; For the first physical property, the PLS algorithm is used to solve it. X=T×P The spectral vector is obtained x The score vector, for the second property, is solved using the PLS algorithm. X=T×P The spectral vector is obtained x The score vector is obtained by concatenating the first and second score vectors obtained for the first and second physical properties respectively, resulting in a combined score vector. k , Calculate the score combination vector k The identification parameters for each sample in the crude oil near-infrared score database are calculated using Equation (Ⅰ) based on the correlation coefficient between each sample and each movement. Q i , Q i ＝ ， i =1,2,…, m （Ⅰ） In formula (Ⅰ), r ji This is the moving correlation coefficient; i This refers to the sample number in the crude oil near-infrared scoring database. j This is the sequence number of the moved window. n The total number of movable windows, m The total number of samples in the crude oil near-infrared score database. Calculate threshold Q t , Q t = ( dw -0.15), where d For the number of data points, w To move the window width; like Q i > Q t ,and i If the correlation coefficient of each movement of the sample is not less than a predetermined value, then the crude oil to be identified is compared with the crude oil near-infrared score database. i The samples are the same. The crude oil near-infrared score database is obtained through the following operations: Various types of crude oil samples were collected, and their first and second physical property data corresponding to the first and second physical properties were measured. Near-infrared spectra were also measured, and second-order differential processing was performed. A near-infrared spectral database of crude oil samples was established by selecting the absorbance of the characteristic spectral regions and the corresponding physical property data. X , For the first physical property, the PLS algorithm is used to solve it. X=T×P Obtain the near-infrared spectral database of the crude oil sample. X Score matrix T1 For the second physical property, the PLS algorithm is used to solve it. X=T×P Obtain the near-infrared spectral database of the crude oil sample. X Score matrix T2 The score matrix T1 With the score matrix T2 By concatenating the first and last parts, a new scoring combination matrix is ​​obtained. K， This constitutes the crude oil near-infrared score database.

2. The method according to claim 1, characterized in that, The first and second physical property data are sulfur content data and acid value data, respectively.

3. The method according to claim 1, characterized in that, The characteristic spectral region is 4628~4000 cm⁻¹. -1 and 6076~5556cm -1 .

4. The method according to claim 1, characterized in that, The predetermined value is 0.9920.

5. The method according to claim 1, characterized in that, If the identification parameters of all samples in the crude oil near-infrared scoring database are not greater than Q t If no sample has a motion correlation coefficient greater than the predetermined value, then there is no sample in the crude oil near-infrared score database that is the same as the crude oil to be identified.

6. The method according to claim 1, characterized in that, The sampling point interval for the absorbance of the near-infrared spectrum is 1 to 10 wavenumbers.

7. The method according to claim 1, characterized in that, The moving correlation coefficient was determined using a moving window.

8. The method according to claim 1 or 7, characterized in that, The width of the moving window is 3 to 15 sampling points.

9. The method according to claim 1 or 7, characterized in that, The method for determining the moving correlation coefficient is as follows: In the data interval, select the width of a moving window from a fixed direction, calculate the correlation coefficient between the crude oil sample to be identified in the window and the score of each sample in the database, and then move the moving window forward by several sampling points to form the next moving window. Calculate the correlation coefficient between the crude oil sample to be identified in this moving window and the score of each sample in the database. Continuously move the moving window in the above method and calculate the correlation coefficient between the crude oil sample to be identified in each moving window and the score of each sample in the database.

10. The method according to claim 9, characterized in that, Each time the window is moved, the distance is 1 to 10 sampling points.

11. The method according to claim 1, characterized in that, The number of data points is the principal component number of PLS, which is 30-120.

12. The method according to claim 1, characterized in that, The method also includes: based on the identified crude oil that is identical to the crude oil to be identified. i The sample was used to determine the properties of the crude oil to be identified.

13. An apparatus for identifying crude oil types using near-infrared spectroscopy, the apparatus comprising: Memory; as well as A processor configured to utilize the method for identifying crude oil types by near-infrared spectroscopy as described in any one of claims 1-12.

14. A machine-readable storage medium storing instructions thereon, characterized in that, When executed by a processor, this instruction causes the processor to be configured to perform the method for identifying crude oil types by near-infrared spectroscopy as described in any one of claims 1-12.

Images