Crude oil property prediction method and system based on self-organizing map network

Abstract

The present application relates to the technical field of petroleum chemical production, more particularly to a crude oil property prediction method and system based on a self-organizing mapping network. The method comprises: collecting a batch of crude oil samples as reference benchmark crude oil samples, and dividing the obtained physicochemical property values into easy-to-measure vectors and difficult-to-measure vectors; collecting a crude oil sample to be measured, obtaining the corresponding easy-to-measure vectors, inputting the easy-to-measure vector data of the crude oil sample to be measured and the reference benchmark crude oil samples into a self-organizing mapping network for clustering, and obtaining a plurality of groups of crude oil vector clusters; screening reference benchmark crude oil samples under the same cluster as the crude oil sample to be measured; calculating the similarity of the screened reference benchmark crude oil samples and the crude oil sample to be measured, assigning a weight value to the difficult-to-measure vectors of the reference benchmark crude oil samples; calculating the difficult-to-measure vectors of the crude oil sample to be measured, and realizing the prediction of the crude oil properties of the crude oil sample to be measured. The present application provides reliable support for the rapid analysis of the overall physicochemical properties of crude oil.

Description

Technical Field

[0001] This invention relates to the field of petrochemical production technology, and more specifically, to a method and system for predicting crude oil properties based on self-organizing mapping networks. Background Technology

[0002] Crude oil is a viscous, oily liquid with a distinctive odor. It is a mixture of various liquid hydrocarbons, including alkanes, cycloalkanes, aromatics, and alkenes. The physicochemical properties of crude oil vary significantly depending on the region of origin. Petrochemical companies employ complex methods for processing crude oil, and the operation of their internal production units is closely linked to its physicochemical properties. Fluctuations in these properties can disrupt the stable operation of the units, negatively impacting product yield and energy waste. Therefore, petrochemical companies urgently need a computational method that can predict the required physicochemical properties of crude oil in a timely manner, laying a solid foundation for the stable operation of refinery production units.

[0003] Existing methods, while comprehensive and detailed laboratory analysis of crude oil can provide a complete picture of its physicochemical properties, are time-consuming and costly due to the diverse properties of crude oil and the complexity of experimental procedures. Furthermore, they cannot provide timely and effective data support for production processes. Although some large refineries have deployed online rapid evaluation systems for crude oil to monitor its physicochemical properties in real time, these systems can only provide data on a limited range of crude oil properties, failing to meet the data requirements for optimizing atmospheric and vacuum distillation unit operations. Therefore, developing methods for calculating crude oil properties, enabling the prediction and extrapolation of other crude oil properties based on a subset of properties, is of significant practical importance for refineries in formulating processing plans and guiding actual production.

[0004] Chinese invention patent CN104155245B discloses a method for detecting multiple properties of oil products based on pattern recognition and spectral mapping. This method projects the near-infrared spectrum of an unknown sample onto a feature plane, correlates the location of the region with a Naive Bayes classifier, and establishes a mathematical model to predict the properties of unknown crude oil. However, this method is limited by the principles of near-infrared spectroscopy analysis, requiring a stable spectrometer and a large number of crude oil spectra for modeling, resulting in a long processing time.

[0005] Chinese invention patent CN110987866A discloses a method for evaluating the properties of gasoline. This method uses a near-infrared spectrometer to measure the near-infrared spectra of known and unknown gasoline samples, classifies the unknown samples, and predicts their properties based on a property prediction model for each category. However, this method requires high repeatability from the near-infrared spectrometer.

[0006] Chinese invention patent CN113139155A discloses a method for optimizing crude oil blending over a wide range. This method uses principal component analysis and cluster analysis on near-infrared spectral data from a crude oil property spectral database to exhaustively explore all possible crude oil blending scenarios, thereby selecting the optimal blending ratio and obtaining its various properties. However, this method is highly dependent on the spectrometer, and the exhaustive algorithm has low efficiency.

[0007] Therefore, there is an urgent need for a crude oil property prediction method and system to improve and address the shortcomings of existing crude oil property calculation methods. Summary of the Invention

[0008] The purpose of this invention is to provide a method and system for predicting crude oil properties based on self-organizing mapping networks, which solves the problem that existing crude oil property calculation methods are time-consuming and labor-intensive when obtaining specific physicochemical properties of crude oil.

[0009] To achieve the above objectives, this invention provides a crude oil property prediction method based on self-organizing mapping networks, comprising the following steps:

[0010] Step S1: Collect a batch of crude oil samples as reference crude oil samples, obtain the corresponding physicochemical properties, and divide the physicochemical property values ​​into easy-to-measurable vectors and difficult-to-measurable vectors according to the ease of obtaining them;

[0011] Step S2: Collect crude oil samples to be tested, obtain the corresponding easy-to-measurable vectors, input the easy-to-measurable vector data of the crude oil samples to be tested and the reference crude oil samples into a self-organizing map network for clustering, and obtain multiple crude oil vector clusters.

[0012] Step S3: Select the reference crude oil samples from Step S1 that are in the same cluster as the crude oil samples to be tested;

[0013] Step S4: Calculate the similarity between the reference crude oil sample and the crude oil sample to be tested obtained in step S3, and assign weights to the difficult-to-measure vector of the reference crude oil sample.

[0014] Step S5: Calculate the unmeasurable vector of the crude oil sample to be tested based on the unmeasurable vector and corresponding weights of the reference crude oil sample, thereby predicting the crude oil properties of the crude oil sample to be tested.

[0015] In one embodiment, the easily measurable vectors in step S1 include density, sulfur content, chlorine content, carbon residue, naphtha yield, diesel yield, wax oil yield, and residue oil yield.

[0016] The difficult-to-measure vectors in step S1 include wax content, asphaltene, nitrogen content, resin, and acid value.

[0017] In one embodiment, the method for obtaining the physicochemical properties in step S1 includes near-infrared crude oil rapid evaluation, nuclear magnetic resonance crude oil rapid evaluation, and crude oil evaluation standard test methods.

[0018] In one embodiment, step S2 further includes:

[0019] Step S21: Collect the crude oil to be tested and obtain the corresponding easy-to-test vector. Combine the obtained easy-to-test vector with the easy-to-test vector obtained in step S1 to form a complete dataset.

[0020] Step S22: Standardize or normalize the complete dataset to obtain the processed easily measurable vectors;

[0021] Step S23: Perform principal component analysis on the processed easily measurable vectors to obtain the principal component analysis results;

[0022] Step S24: Input the principal component analysis results into a self-organizing map network for clustering to obtain multiple crude oil vector clusters.

[0023] In one embodiment, the standardized calculation formula in step S22 is as follows:

[0024] a′=(aa avg ) / σ

[0025] Where a′ is the standardized physicochemical property value of crude oil, a is the original physicochemical property value of crude oil, and a avg σ represents the average value of the physicochemical properties of the crude oil across all samples, and σ represents the standard deviation of the physicochemical properties of the crude oil across all samples.

[0026] In one embodiment, the normalization process in step S22 includes linear normalization and mean normalization;

[0027] The formula for calculating linear normalization is as follows:

[0028] a′=(aa min ) / (a max -a min );

[0029] Where a′ is the linearly normalized value of the crude oil's physicochemical properties, a is the original value of the crude oil's physicochemical properties, and a max a is the maximum value of the physicochemical properties of the crude oil among all samples. min This represents the minimum value of the crude oil's physicochemical properties among all samples.

[0030] The formula for calculating the mean normalization is as follows:

[0031] a′=(aa avg ) / (a max -a min);

[0032] Where a′ is the mean-normalized physicochemical property value of crude oil, a is the original physicochemical property value of crude oil, and a avg a is the average value of the physicochemical properties of the crude oil across all samples. max a is the maximum value of the physicochemical properties of the crude oil among all samples. min This represents the minimum value of the physicochemical properties of crude oil among all samples.

[0033] In one embodiment, the cluster index of the reference crude oil sample obtained in step S3 satisfies the following expression:

[0034] Index input =Index i

[0035] Among them, Index input The index is the cluster index corresponding to the clustered crude oil sample to be tested. i This is the cluster index corresponding to the i-th reference crude oil sample selected in step S1.

[0036] In one embodiment, step S4 further includes calculating the degree of similarity using a distance-based similarity criterion method;

[0037] The distance-based similarity criterion method further includes Euclidean distance, cosine distance, and Mahalanobis distance.

[0038] In one embodiment, in step S4, weights are assigned to the reference crude oil sample obtained in step S3 for the physicochemical properties of the crude oil sample to be tested, expressed as mass fractions.

[0039] The weighting methods used include linear normalization, logarithmic normalization, and inverse cotangent normalization.

[0040] In one embodiment, in step S4, for the physicochemical properties of the crude oil sample to be tested that are not expressed as mass fractions, one group of crude oil samples from the reference crude oil samples screened in step S3 that are most similar to the physicochemical properties is selected, and the physicochemical property value corresponding to the group of crude oil samples is used as the predicted value of the crude oil sample to be tested in terms of the physicochemical property.

[0041] To achieve the above objectives, the present invention provides a crude oil property prediction system based on a self-organizing mapping network, comprising:

[0042] Memory is used to store instructions that can be executed by the processor;

[0043] A processor for executing the instructions to implement the method as described in any of the preceding descriptions.

[0044] To achieve the above objectives, the present invention provides a computer-readable medium having computer instructions stored thereon, wherein when the computer instructions are executed by a processor, the method described in any of the preceding claims is performed.

[0045] This invention provides a crude oil property prediction method and system based on self-organizing mapping networks. Employing a distance-based similarity criterion, it can rapidly predict difficult-to-obtain physicochemical properties of crude oil samples based on readily available physicochemical properties. This solves the problems of time-consuming and labor-intensive collection and high analysis difficulty for certain crude oil properties, thus improving the efficiency of crude oil analysis and providing reliable support for the rapid analysis of comprehensive physicochemical properties of crude oil. It also strongly promotes the improvement of refinery production efficiency and the growth of economic benefits. Attached Figure Description

[0046] The above and other features, properties and advantages of the present invention will become more apparent from the following description taken in conjunction with the accompanying drawings and embodiments, in which the same reference numerals always denote the same features, wherein:

[0047] Figure 1 A flowchart of a crude oil property prediction method based on a self-organizing mapping network according to an embodiment of the present invention is disclosed;

[0048] Figure 2 A block diagram illustrating the principle of a crude oil property prediction system based on a self-organizing mapping network according to an embodiment of the present invention is disclosed. Detailed Implementation

[0049] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the invention.

[0050] To address the problems of high difficulty and time-consuming crude oil property analysis in existing technologies, this invention provides a method and system for rapid calculation and prediction of crude oil properties, which can be widely applied in the field of intelligent crude oil dispatching.

[0051] Figure 1 A flowchart of a crude oil property prediction method based on a self-organizing mapping network according to an embodiment of the present invention is disclosed, as follows: Figure 1 As shown, the crude oil property prediction method based on self-organizing mapping networks proposed in this invention specifically includes the following steps:

[0052] Step S1: Collect a batch of crude oil samples as reference crude oil samples, obtain the corresponding physicochemical properties, and divide the physicochemical property values ​​into easy-to-measurable vectors and difficult-to-measurable vectors according to the ease of obtaining them;

[0053] Step S2: Collect the crude oil sample to be tested, obtain the corresponding easy-to-measurable vector, input the easy-to-measurable vector data of the crude oil sample to be tested and the reference crude oil sample into the self-organizing map network for clustering, and obtain multiple crude oil vector clusters.

[0054] Step S3: Select the reference crude oil samples from Step S1 that are in the same cluster as the crude oil samples to be tested;

[0055] Step S4: Calculate the similarity between the reference crude oil sample and the crude oil sample to be tested obtained in step S3, and assign weights to the difficult-to-measure vector of the reference crude oil sample.

[0056] Step S5: Calculate the unmeasurable vector of the crude oil sample to be tested based on the unmeasurable vector and corresponding weights of the reference crude oil sample, thereby predicting the crude oil properties of the crude oil sample to be tested.

[0057] These steps will be described in detail below. It should be understood that, within the scope of this invention, the above-described technical features of this invention and the technical features specifically described below (such as in the embodiments) can be combined and related to each other to form preferred technical solutions.

[0058] Step S1: Collect a batch of crude oil samples as reference crude oil samples, obtain the corresponding physicochemical properties, and divide the physicochemical property values ​​into easy-to-measure vectors and difficult-to-measure vectors according to the ease of obtaining them.

[0059] A batch of crude oil samples was collected, and several of their physicochemical properties were obtained. The easily obtainable physicochemical property values ​​were combined to form an easily measurable vector, and the more difficult-to-obtain physicochemical property values ​​were combined to form a difficult-to-measurable vector. This batch of crude oil samples can be used as a reference benchmark.

[0060] In this embodiment, the easily measurable vectors of crude oil include, but are not limited to, physicochemical properties such as density, sulfur content, chlorine content, carbon residue, naphtha yield, diesel yield, wax oil yield, and residue oil yield.

[0061] In this embodiment, the difficult-to-measure vectors of crude oil include, but are not limited to, physicochemical properties such as wax content, asphaltene, nitrogen content, gum content, and acid value;

[0062] In this embodiment, the methods for obtaining the physicochemical properties of crude oil include, but are not limited to, near-infrared crude oil rapid evaluation, nuclear magnetic resonance crude oil rapid evaluation, crude oil evaluation standard test methods, or other crude oil evaluation methods;

[0063] Measurable vectors of collected crude oil samples The corresponding expression is:

[0064]

[0065] Where a1 is the first readily available physicochemical property value of crude oil, a2 is the second readily available physicochemical property value of crude oil, and a n Let be the nth readily available physicochemical property value of crude oil.

[0066] Difficult-to-measure vector of the collected crude oil sample The corresponding expression is:

[0067]

[0068] Where b1 represents the first type of physicochemical property of crude oil that is difficult to obtain, b2 represents the second type of physicochemical property of crude oil that is difficult to obtain, and b... n This represents the nth physicochemical property value of crude oil, which is relatively difficult to obtain.

[0069] The terms "measurable properties" and "difficult-to-measurable properties" are general terms for the physicochemical properties of crude oil. When using this method, those skilled in the art can classify the physicochemical properties of crude oil into measurable properties and difficult-to-measurable properties based on factors such as the calculation requirements for crude oil properties, laboratory analytical capabilities, historical data of crude oil, and computational complexity.

[0070] Step S2: Collect the crude oil sample to be tested, obtain the corresponding easy-to-measurable vector, input the easy-to-measurable vector data of the crude oil sample to be tested and the reference crude oil sample into the self-organizing map network for clustering, and obtain multiple sets of crude oil vector clusters.

[0071] A new batch of crude oil samples is obtained, and the easily measurable vectors of this batch are obtained experimentally. These vectors, along with the easily measurable vector data from step S1, are preprocessed and then fed into a self-organizing map network for clustering to obtain multiple crude oil vector clusters. This new batch of crude oil samples is the crude oil sample to be tested.

[0072] In this embodiment, step S2 further includes:

[0073] Step S21: Collect the crude oil to be tested and obtain the corresponding easy-to-test vector. Combine the obtained easy-to-test vector with the easy-to-test vector obtained in step S1 to form a complete dataset.

[0074] A new batch of crude oil samples is obtained as the crude oil samples to be tested, and the measurable vector of this batch of crude oil samples is obtained through experiments.

[0075] Step S22: Standardize or normalize the complete dataset to obtain the processed easily measurable vectors;

[0076] Standardizing or normalizing the dataset can eliminate the order-of-magnitude differences and / or dimensional differences between different physicochemical properties of crude oil, resulting in better calculation results.

[0077] The standardization or normalization formulas applicable to this method include, but are not limited to, the following.

[0078] The standardized calculation formula is as follows:

[0079] a′=(aa avg ) / σ

[0080] Where a′ is the standardized physicochemical property value of crude oil, a is the original physicochemical property value of crude oil, and a avg σ represents the average value of the physicochemical properties of the crude oil across all samples, and σ represents the standard deviation of the physicochemical properties of the crude oil across all samples.

[0081] The formula for linear normalization is as follows:

[0082] a′=(aa min ) / (a max -a min );

[0083] Where a′ is the linearly normalized value of the crude oil's physicochemical properties, a is the original value of the crude oil's physicochemical properties, and a max a is the maximum value of the physicochemical properties of the crude oil among all samples. min This is the minimum value of the physicochemical properties of the crude oil among all samples.

[0084] The formula for calculating mean normalization is as follows:

[0085] a′=(aa avg ) / (a max -a min );

[0086] Where a′ is the mean-normalized physicochemical property value of crude oil, a is the original physicochemical property value of crude oil, and a avg a is the average value of the physicochemical properties of the crude oil across all samples. max a is the maximum value of the physicochemical properties of the crude oil among all samples. min This represents the minimum value of the physicochemical properties of crude oil among all samples.

[0087] Obtain the standardized or normalized easily measurable vector The corresponding expression is:

[0088]

[0089] Step S23: Perform principal component analysis on the processed easily measurable vectors to obtain the principal component analysis results;

[0090] The easily measurable vector obtained in S22 Principal component analysis can yield different results by setting different variance contribution rates.

[0091] Principal Component Analysis (PCA) is a statistical method that uses orthogonal transformations to convert a set of potentially correlated variables into a set of linearly uncorrelated variables; the transformed set of variables is called the principal components.

[0092] Step S24: Input the principal component analysis results into a self-organizing map network for clustering to obtain multiple crude oil vector clusters;

[0093] The principal component analysis results obtained in S23 are put into a self-organizing map network for clustering to obtain multiple crude oil vector clusters.

[0094] The self-organizing map network achieves clustering by mapping a high-dimensional input space to a low-dimensional output space. It consists of a two-dimensional grid, with each grid cell called a node.

[0095] Each node has a weight vector that is compared with the sample vectors in the input space.

[0096] During clustering, input samples are gradually assigned to the crude oil vector cluster represented by the winning node that is most similar to them. The shape and size of the crude oil vector cluster depend on the features of the input data and the parameter settings of the clustering algorithm. Nodes that are close together will form the same crude oil vector cluster, while nodes that are far apart will represent different crude oil vector clusters.

[0097] Step S3: Select the reference crude oil samples from Step S1 that are in the same cluster as the crude oil samples to be tested.

[0098] Find the crude oil samples in step S1 that belong to the same cluster as the crude oil sample to be tested, and the cluster index of the selected crude oil samples in step S1 should satisfy the following expression:

[0099] Index input =Index i (1)

[0100] Among them, Index input The index is the cluster index corresponding to the clustered crude oil sample to be tested. i This is the cluster index corresponding to the i-th reference crude oil sample selected in step S1.

[0101] Step S4: Calculate the similarity between the reference crude oil sample and the crude oil sample to be tested obtained from step S3, and assign weights to the unmeasurable vector of the reference crude oil sample.

[0102] The similarity between the crude oil samples selected in step S3 and the new batch of crude oil samples is taken into account, and different weights are assigned to the former.

[0103] In this embodiment, step S4 further includes: calculating the degree of similarity using a distance-based similarity criterion method.

[0104] Similarity criteria based on distance include, but are not limited to, Euclidean distance, cosine distance, and Mahalanobis distance.

[0105] In this embodiment, the formula for calculating Euclidean distance is:

[0106]

[0107] Where EucD is the Euclidean distance value, the smaller the value, the smaller the distance, indicating that the reference crude oil sample and the crude oil sample to be tested are more similar in step S1.

[0108] a′ i，new Let a′ be the i-th physicochemical property of the easily measurable vector after standardization of the crude oil sample to be tested. i The i-th physicochemical property is the easily measurable vector after standardization of the reference crude oil sample in step S1.

[0109] In this embodiment, the formula for calculating the cosine distance is:

[0110]

[0111] Where cosD is the cosine distance, the smaller the value, the smaller the distance, indicating that the reference crude oil sample and the crude oil sample to be tested are more similar in step S1.

[0112] a′ i，new Let a′ be the i-th physicochemical property of the easily measurable vector after standardization of the crude oil sample to be tested. i The i-th physicochemical property is the easily measurable vector after standardization of the reference crude oil sample in step S1.

[0113] In this embodiment, the formula for calculating the Mahalanobis distance is as follows:

[0114]

[0115] in, This is the standardized, easily measurable vector of the crude oil sample to be tested. S is the standardized, easily measurable vector of the reference crude oil sample in step S1. and The covariance matrix is ​​given by , where mahD is the Mahalanobis distance. The smaller the value, the smaller the distance, indicating that the reference crude oil sample and the crude oil sample to be tested are more similar.

[0116] In step S4, the units of the physicochemical properties of the crude oil sample to be tested can be divided into mass fraction and non-mass fraction. Based on the units of the physicochemical properties, the crude oil sample to be tested is subjected to the following different treatments.

[0117] A distance-based similarity criterion method is used to calculate the similarity between the reference crude oil sample selected in step S3 and the crude oil sample to be tested.

[0118] For the physicochemical properties of the crude oil sample to be tested, expressed as mass fraction, each of the i reference crude oil samples selected in step S3 is assigned a weight. i .

[0119] The methods used to assign weights include, but are not limited to, linear normalization, logarithmic normalization, and inverse cotangent normalization.

[0120] For the physicochemical properties of the crude oil sample to be tested that are not expressed as mass fractions, select one group of crude oil samples from the reference crude oil samples obtained in step S3 that are most similar to the physicochemical properties, and use the physicochemical property values ​​corresponding to this group of crude oil samples as the predicted values ​​of the crude oil sample to be tested in terms of these physicochemical properties.

[0121] Step S5: Calculate the unmeasurable vector of the crude oil sample to be tested based on the unmeasurable vector and corresponding weights of the reference crude oil sample, thereby predicting the crude oil properties of the crude oil sample to be tested.

[0122] Based on the unmeasurable vectors and their weights of the reference crude oil samples selected in steps S3 and S4, the unmeasurable vectors of the crude oil samples to be tested are calculated, thus realizing the calculation and prediction of the crude oil properties of the crude oil samples to be tested.

[0123] For the physicochemical properties of the crude oil sample to be tested, expressed as mass fraction, the indeterminate vector of the crude oil sample to be tested is calculated based on the indeterminate vector and its weights of the reference crude oil samples selected in steps S3 and S4. This achieves the calculation of the crude oil properties. The specific calculation formula is as follows:

[0124]

[0125] in, The unmeasurable vector of the crude oil sample to be tested. For the difficult-to-measure vector of the nth reference crude oil sample selected in step S3, Weight n Its corresponding weight.

[0126] Although the methods described above are illustrated and depicted as a series of actions for the sake of simplicity, it should be understood and appreciated that these methods are not limited by the order of the actions, as some actions may occur in a different order and / or concurrently with other actions from the illustrations and descriptions herein or not illustrated and described herein but which may be understood by those skilled in the art, according to one or more embodiments.

[0127] The following section uses a set of collected crude oil samples and their corresponding physicochemical properties as experimental subjects to provide a detailed explanation of the crude oil property prediction method based on self-organizing mapping networks proposed in this invention.

[0128] Step S1: Collect a batch of crude oil samples as reference crude oil samples and obtain several of their physicochemical properties.

[0129] In this embodiment, the easily measurable vectors of crude oil include density, sulfur content, carbon residue, naphtha yield, diesel yield, wax oil yield, and residue oil yield, while the difficult-to-measurable vectors of crude oil include asphaltenes, nitrogen content, gum content, and acid value.

[0130] Generate easily measurable vectors for crude oil samples and unpredictable vectors

[0131] Step S2: Obtain a new batch of crude oil samples as the crude oil samples to be tested, obtain the easily measurable vectors of this batch of crude oil samples, and after standardizing and performing principal component analysis on the easily measurable vector data in step S1, input them together into a self-organizing map network for clustering to obtain 650 crude oil vector clusters.

[0132] Step S3: Identify the crude oil samples from Step S1 that are in the same cluster as the new batch of crude oil samples.

[0133] Step S4: Using the Mahalanobis distance similarity criterion, the physicochemical properties in the difficult-to-measure vector, such as asphaltene, nitrogen content, resin, and acid value, are calculated using linear normalization to obtain the 22 crude oil samples screened in Step S3 and assigned corresponding weights Weighti.

[0134] Step S5: Calculate the unpredictable vector of a new batch of crude oil samples based on the unpredictable vectors and their weights of the crude oil samples selected in steps S3 and S4, thus realizing the prediction of crude oil properties of the new batch of crude oil samples.

[0135] Based on the above method, the difficult-to-measure vectors of 26 sets of crude oil were calculated.

[0136] The prediction results of crude oil asphalt are shown in Table 1. The reference value is the crude oil property data value obtained by the near-infrared crude oil rapid evaluation method, and the predicted value is the data value calculated by this method.

[0137] Table 1: Prediction Results of Crude Oil Asphaltene

[0138]

[0139] The predicted results of crude oil nitrogen content are shown in Table 2.

[0140] Table 2: Prediction Results of Crude Oil Nitrogen Content

[0141]

[0142]

[0143] The prediction results for crude oil gum content are shown in Table 3.

[0144] Table 3: Prediction Results of Crude Oil Rubber

[0145]

[0146]

[0147] The predicted acid value of crude oil is shown in Table 4.

[0148] Table 4: Prediction Results of Crude Oil Acid Value

[0149] 1 0.196 0.129 -33.824 2 0.217 0.157 -27.623 3 0.220 0.150 -32.134 4 0.130 0.149 15.079 5 0.175 0.187 6.489 6 0.203 0.149 -26.331 7 0.179 0.148 -17.298 8 0.167 0.143 -14.486 9 0.185 0.152 -17.932 10 0.175 0.150 -14.396 11 0.213 0.204 -4.516 12 0.223 0.218 -2.285 13 0.226 0.208 -8.012 14 0.221 0.220 -0.292 15 0.168 0.169 0.229 16 0.221 0.146 -33.906 17 0.209 0.149 -28.670 18 0.226 0.197 -12.692 19 0.197 0.207 5.227 20 0.194 0.212 9.273 21 0.204 0.198 -3.026 22 0.205 0.173 -15.289 23 0.230 0.229 -0.453 24 0.259 0.191 -26.077 25 0.228 0.199 -12.556 26 0.213 0.225 5.671

[0150] As shown in Tables 1, 2, 3, and 4, the average absolute values ​​of the relative errors corresponding to asphaltene, nitrogen content, resin content, and acid value are 14.10%, 10.99%, 18.59%, and 14.37%, respectively. This indicates that by using a self-organizing mapping network to obtain multiple crude oil vector clusters, and then using a similarity criterion method to find known similar crude oils within the same cluster and assigning them weights, the specified physicochemical properties of crude oil can be calculated.

[0151] Clearly, the dimension of the crude oil easy-to-measurable vector directly affects the distribution of the crude oil vector cluster. The higher the dimension, the more accurate the vector cluster separation, but the longer it takes.

[0152] The distribution, variance, mean, and other characteristics of the values ​​of each dimension of the crude oil easy-to-measurable vector will have varying degrees of influence on the distribution results of the crude oil vector cluster. When those skilled in the art use this method to calculate and predict the physicochemical properties of crude oil, they can adjust the variance contribution rate in the principal component analysis according to actual production needs to improve the prediction efficiency or accuracy.

[0153] Figure 2 A block diagram illustrating the principle of a crude oil property prediction system based on a self-organizing mapping network according to an embodiment of the present invention is disclosed. The crude oil property prediction system based on a self-organizing mapping network may include an internal communication bus 201, a processor 202, a read-only memory (ROM) 203, a random access memory (RAM) 204, a communication port 205, and a hard disk 207. The internal communication bus 201 enables data communication between components of the crude oil property prediction system based on the self-organizing mapping network. The processor 202 can perform judgments and issue prompts. In some embodiments, the processor 202 may consist of one or more processors.

[0154] Communication port 205 enables data transmission and communication between the crude oil property prediction system based on self-organizing mapping networks and external input / output devices. In some embodiments, the crude oil property prediction system based on self-organizing mapping networks can send and receive information and data from the network through communication port 205. In some embodiments, the crude oil property prediction system based on self-organizing mapping networks can transmit data and communicate with external input / output devices in a wired manner through input / output port 206.

[0155] The crude oil property prediction system based on self-organizing mapping networks can also include different forms of program storage units and data storage units, such as hard disk 207, read-only memory (ROM) 203, and random access memory (RAM) 204, capable of storing various data files used for computer processing and / or communication, as well as possible program instructions executed by processor 202. Processor 202 executes these instructions to implement the main part of the method. The results processed by processor 202 are transmitted to an external output device via communication port 205 and displayed on the user interface of the output device.

[0156] For example, the implementation process file of the above-mentioned crude oil property prediction method based on self-organizing mapping network can be a computer program, stored in hard disk 207, and can be loaded into processor 202 for execution to implement the method of this application.

[0157] When the implementation process document of the crude oil property prediction method based on self-organizing mapping networks is a computer program, it can also be stored as an article of manufacture in a computer-readable storage medium. For example, computer-readable storage media can include, but are not limited to, magnetic storage devices (e.g., hard disks, floppy disks, magnetic stripes), optical discs (e.g., compact discs (CDs), digital multifunction discs (DVDs)), smart cards, and flash memory devices (e.g., electrically erasable programmable read-only memory (EPROM), cards, sticks, key drives). Furthermore, the various storage media described herein can represent one or more devices and / or other machine-readable media used for storing information. The term "machine-readable medium" can include, but is not limited to, wireless channels and various other media (and / or storage media) capable of storing, containing, and / or carrying code and / or instructions and / or data.

[0158] Compared with existing technologies, the crude oil property prediction method and system based on self-organizing mapping networks provided by this invention have the following advantages:

[0159] 1) It eliminates the need for expert experience to infer the physicochemical properties of crude oil, reducing the difficulty of technology application and providing a method for rapid detection of the physicochemical properties of crude oil;

[0160] 2) It has multiple similarity criteria methods, which ensures the accuracy of the calculation results;

[0161] 3) It is not constrained by the units of the physical and chemical properties of crude oil, which improves the universality of the calculation method.

[0162] Although the methods described above are illustrated and depicted as a series of actions for the sake of simplicity, it should be understood and appreciated that these methods are not limited by the order of the actions, as some actions may occur in a different order and / or concurrently with other actions from the illustrations and descriptions herein or not illustrated and described herein but which may be understood by those skilled in the art, according to one or more embodiments.

[0163] As indicated in this application and claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" are not specifically singular and may include plural forms. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of explicitly identified steps and elements, which do not constitute an exclusive list, and the method or apparatus may also include other steps or elements.

[0164] Those skilled in the art will understand that information, signals, and data can be represented using any of a variety of different techniques and arts. For example, the data, instructions, commands, information, signals, bits, symbols, and chips described throughout the above description can be represented by voltage, current, electromagnetic waves, magnetic fields or magnetic particles, light fields or optical particles, or any combination thereof.

[0165] Those skilled in the art will further appreciate that the various illustrative logic blocks, modules, circuits, and algorithm steps described in conjunction with the embodiments disclosed herein can be implemented as electronic hardware, computer software, or a combination of both. To clearly illustrate this interchangeability between hardware and software, the various illustrative components, blocks, modules, circuits, and steps are described above in a generalized manner in terms of their functionality. Whether such functionality is implemented as hardware or software depends on the specific application and the design constraints imposed on the overall system. Those skilled in the art may implement the described functionality in different ways for each specific application, but such implementation decisions should not be construed as departing from the scope of the invention.

[0166] The various illustrative logic modules and circuits described in conjunction with the embodiments disclosed herein may be implemented or performed using a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. The general-purpose processor may be a microprocessor, but in alternatives, it may be any conventional processor, controller, microcontroller, or state machine. The processor may also be implemented as a combination of computing devices, such as a combination of a DSP and a microprocessor, multiple microprocessors, one or more microprocessors cooperating with a DSP core, or any other such configuration.

[0167] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of both. The software module may reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art. An exemplary storage medium is coupled to a processor such that the processor can read and write information to / from the storage medium. In an alternative, the storage medium may be integrated into the processor. The processor and storage medium may reside in an ASIC. The ASIC may reside in a user terminal. In an alternative, the processor and storage medium may reside as discrete components in the user terminal.

[0168] The above embodiments are provided for those skilled in the art to implement or use the present invention. Those skilled in the art can make various modifications or changes to the above embodiments without departing from the inventive concept of the present invention. Therefore, the protection scope of the present invention is not limited to the above embodiments, but should be the maximum scope that conforms to the innovative features mentioned in the claims.

Claims

1. A method for predicting crude oil properties based on self-organizing mapping networks, characterized in that, Includes the following steps: Step S1: Collect a batch of crude oil samples as reference crude oil samples, obtain the corresponding physicochemical properties, and divide the physicochemical property values ​​into easy-to-measurable vectors and difficult-to-measurable vectors according to the ease of obtaining them; Step S2: Collect crude oil samples to be tested, obtain the corresponding easy-to-measurable vectors, input the easy-to-measurable vector data of the crude oil samples to be tested and the reference crude oil samples into a self-organizing map network for clustering, and obtain multiple crude oil vector clusters. Step S3: Select the reference crude oil samples from Step S1 that are in the same cluster as the crude oil samples to be tested; Step S4: Calculate the similarity between the reference crude oil sample and the crude oil sample to be tested obtained in step S3, and assign weights to the difficult-to-measure vector of the reference crude oil sample. Step S5: Calculate the unmeasurable vector of the crude oil sample to be tested based on the unmeasurable vector and corresponding weights of the reference crude oil sample, thereby predicting the crude oil properties of the crude oil sample to be tested. The easily measurable vectors in step S1 include density, sulfur content, chlorine content, carbon residue, naphtha yield, diesel yield, wax oil yield, and residue oil yield. The difficult-to-measure vectors in step S1 include wax content, asphaltene, nitrogen content, resin, and acid value. Step S2 further includes: Step S21: Collect the crude oil to be tested and obtain the corresponding easy-to-test vector. Combine the obtained easy-to-test vector with the easy-to-test vector obtained in step S1 to form a complete dataset. Step S22: Standardize or normalize the complete dataset to obtain the processed easily measurable vectors; Step S23: Perform principal component analysis on the processed easily measurable vectors to obtain the principal component analysis results; Step S24: Input the principal component analysis results into a self-organizing map network for clustering to obtain multiple crude oil vector clusters.

2. The crude oil property prediction method based on self-organizing mapping networks according to claim 1, characterized in that, The methods for obtaining physicochemical properties in step S1 include near-infrared crude oil rapid evaluation, nuclear magnetic resonance crude oil rapid evaluation, and crude oil evaluation standard test methods.

3. The crude oil property prediction method based on self-organizing mapping networks according to claim 1, characterized in that, The standardized calculation formula in step S22 is as follows: in, These are the standardized physicochemical properties of crude oil. These are the original physicochemical properties of the crude oil. This represents the average of the physicochemical properties of the crude oil across all samples. This represents the standard deviation of the physicochemical properties of the crude oil across all samples.

4. The crude oil property prediction method based on self-organizing mapping networks according to claim 1, characterized in that, The normalization process in step S22 includes linear normalization and mean normalization; The formula for calculating linear normalization is as follows: ； in, These are the linearly normalized values ​​of the crude oil's physicochemical properties. These are the original physicochemical properties of the crude oil. This represents the maximum value of the crude oil's physicochemical properties among all samples. This represents the minimum value of the crude oil's physicochemical properties among all samples. The formula for calculating the mean normalization is as follows: ； in, These are the mean-normalized values ​​of the crude oil's physicochemical properties. These are the original physicochemical properties of the crude oil. This represents the average value of the physicochemical properties of the crude oil across all samples. This represents the maximum value of the crude oil's physicochemical properties among all samples. This represents the minimum value of the physicochemical properties of crude oil among all samples.

5. The crude oil property prediction method based on self-organizing mapping networks according to claim 1, characterized in that, The cluster index of the reference crude oil sample obtained in step S3 satisfies the following expression: in, This is the cluster index corresponding to the clustered crude oil sample to be tested. For the selected step S1, the first... i The cluster index corresponding to each reference crude oil sample.

6. The crude oil property prediction method based on self-organizing mapping networks according to claim 1, characterized in that, Step S4 further includes calculating the degree of similarity using a distance-based similarity criterion method; The distance-based similarity criterion method further includes Euclidean distance, cosine distance, and Mahalanobis distance.

7. The crude oil property prediction method based on self-organizing mapping networks according to claim 6, characterized in that, In step S4, for the physicochemical properties of the crude oil sample to be tested, expressed in terms of mass fraction, a weight is assigned to the reference crude oil sample obtained in step S3. The weighting methods used include linear normalization, logarithmic normalization, and inverse cotangent normalization.

8. The crude oil property prediction method based on self-organizing mapping networks according to claim 1, characterized in that, In step S4, for the physicochemical properties of the crude oil sample to be tested that are not expressed as mass fractions, one group of crude oil samples from the reference crude oil samples screened in step S3 that are most similar to the physicochemical properties is selected, and the physicochemical property values ​​corresponding to the group of crude oil samples are used as the predicted values ​​of the crude oil sample to be tested in terms of the physicochemical properties.

9. A crude oil property prediction system based on self-organizing mapping networks, comprising: Memory is used to store instructions that can be executed by the processor; A processor for executing the instructions to implement the method as described in any one of claims 1-8.

10. A computer-readable medium having stored thereon computer instructions, wherein when the computer instructions are executed by a processor, the method of any one of claims 1-8 is performed.

Images