CN113869578A - Intelligent prediction and diagnosis method for salt content of crude oil after removal of electric desalting system of atmospheric and vacuum distillation unit - Google Patents
Intelligent prediction and diagnosis method for salt content of crude oil after removal of electric desalting system of atmospheric and vacuum distillation unit Download PDFInfo
- Publication number
- CN113869578A CN113869578A CN202111127300.2A CN202111127300A CN113869578A CN 113869578 A CN113869578 A CN 113869578A CN 202111127300 A CN202111127300 A CN 202111127300A CN 113869578 A CN113869578 A CN 113869578A
- Authority
- CN
- China
- Prior art keywords
- crude oil
- desalting
- atmospheric
- salt content
- tank
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000011033 desalting Methods 0.000 title claims abstract description 205
- 239000010779 crude oil Substances 0.000 title claims abstract description 145
- 150000003839 salts Chemical class 0.000 title claims abstract description 97
- 238000003745 diagnosis Methods 0.000 title claims abstract description 69
- 238000000034 method Methods 0.000 title claims abstract description 45
- 238000005292 vacuum distillation Methods 0.000 title claims abstract description 13
- 238000007637 random forest analysis Methods 0.000 claims abstract description 25
- 238000005070 sampling Methods 0.000 claims abstract description 14
- 238000007781 pre-processing Methods 0.000 claims abstract description 9
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 39
- 238000003066 decision tree Methods 0.000 claims description 32
- 238000012549 training Methods 0.000 claims description 23
- 238000012360 testing method Methods 0.000 claims description 18
- 230000008569 process Effects 0.000 claims description 12
- 230000000694 effects Effects 0.000 claims description 7
- NINIDFKCEFEMDL-UHFFFAOYSA-N Sulfur Chemical compound [S] NINIDFKCEFEMDL-UHFFFAOYSA-N 0.000 claims description 6
- 230000018044 dehydration Effects 0.000 claims description 6
- 238000006297 dehydration reaction Methods 0.000 claims description 6
- 239000008213 purified water Substances 0.000 claims description 6
- 229910052717 sulfur Inorganic materials 0.000 claims description 6
- 239000011593 sulfur Substances 0.000 claims description 6
- 239000002253 acid Substances 0.000 claims description 5
- 238000002347 injection Methods 0.000 claims description 4
- 239000007924 injection Substances 0.000 claims description 4
- 239000007788 liquid Substances 0.000 claims description 4
- 230000007246 mechanism Effects 0.000 claims description 4
- 238000013501 data transformation Methods 0.000 claims description 3
- 230000010354 integration Effects 0.000 claims description 3
- 238000012544 monitoring process Methods 0.000 claims description 3
- 238000004140 cleaning Methods 0.000 claims description 2
- 238000005192 partition Methods 0.000 claims description 2
- 238000009825 accumulation Methods 0.000 claims 2
- 101100314150 Caenorhabditis elegans tank-1 gene Proteins 0.000 claims 1
- 238000004422 calculation algorithm Methods 0.000 abstract description 12
- 238000004458 analytical method Methods 0.000 abstract description 10
- 238000001514 detection method Methods 0.000 abstract description 2
- 238000010612 desalination reaction Methods 0.000 description 8
- 241000196324 Embryophyta Species 0.000 description 4
- 238000007599 discharging Methods 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 239000003921 oil Substances 0.000 description 3
- 239000002994 raw material Substances 0.000 description 3
- 238000012935 Averaging Methods 0.000 description 2
- 238000003556 assay Methods 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 239000003054 catalyst Substances 0.000 description 2
- 230000007797 corrosion Effects 0.000 description 2
- 238000005260 corrosion Methods 0.000 description 2
- 230000003009 desulfurizing effect Effects 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 238000007670 refining Methods 0.000 description 2
- 235000008331 Pinus X rigitaeda Nutrition 0.000 description 1
- 235000011613 Pinus brutia Nutrition 0.000 description 1
- 241000018646 Pinus brutia Species 0.000 description 1
- FAPWRFPIFSIZLT-UHFFFAOYSA-M Sodium chloride Chemical compound [Na+].[Cl-] FAPWRFPIFSIZLT-UHFFFAOYSA-M 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000003795 chemical substances by application Substances 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000005684 electric field Effects 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000014509 gene expression Effects 0.000 description 1
- 229910001385 heavy metal Inorganic materials 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000004989 laser desorption mass spectroscopy Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 231100000572 poisoning Toxicity 0.000 description 1
- 230000000607 poisoning effect Effects 0.000 description 1
- 238000000611 regression analysis Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 239000000243 solution Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/067—Enterprise or organisation modelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/04—Manufacturing
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Economics (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Development Economics (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Data Mining & Analysis (AREA)
- Game Theory and Decision Science (AREA)
- Educational Administration (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Manufacturing & Machinery (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Housings And Mounting Of Transformers (AREA)
Abstract
The invention relates to the technical field of petrochemical devices, in particular to an intelligent prediction and diagnosis method for the salt content of crude oil after being removed by an electric desalting system of an atmospheric and vacuum distillation unit. The method comprises the steps of constructing an index system influencing the content of crude oil salt after an electric desalting system of an atmospheric and vacuum device is subjected to desalting, and obtaining sample data according to indexes in the system; preprocessing the sample data; constructing an intelligent prediction diagnosis model of the content of crude oil salt after the removal of the electric desalting system of the atmospheric and vacuum device based on the random forest by utilizing the preprocessed data; carrying out real-time prediction diagnosis on the electric desalting system through the constructed intelligent prediction diagnosis model; and comparing the prediction diagnosis result with the actual value, and judging and optimizing the intelligent prediction diagnosis model of the crude oil salt content after the electric desalting system is removed. According to the method, the intelligent early warning is carried out on the situation that the salt content of the removed crude oil exceeds the standard by adopting a random forest algorithm to construct a prediction model, the cost of the electro-desalting sampling analysis is effectively reduced, and the efficiency of the detection of the salt content of the removed crude oil of the electro-desalting system is improved.
Description
Technical Field
The invention relates to the technical field of petrochemical device detection, in particular to an intelligent prediction and diagnosis method for the salt content of crude oil after being removed by an electric desalting system of an atmospheric and vacuum device.
Background
The atmospheric and vacuum distillation device is the first process of oil refining enterprises and provides qualified and high-quality raw materials for a plurality of downstream secondary processing devices. The electric desalting system is a 'tap' of an atmospheric and vacuum device and is a crude oil pretreatment process which is necessary for providing high-quality raw materials for subsequent devices. The salt content in the crude oil after the electric desalting is increased, so that the scaling, the blockage and the corrosion of equipment and pipelines of the device are easily caused, and the heavy metal contained in the crude oil also easily causes the catalyst poisoning in the subsequent processing process and possibly influences the product quality of downstream devices. Therefore, the electric desalting is an important process for reducing energy consumption, reducing equipment corrosion and scaling, reducing catalyst loss, improving product quality and prolonging the operation period of the device in oil refining enterprises, and is also a process unit which is focused on by device personnel.
In recent years, with the rapid increase of the processing scale of petrochemical enterprises and the diversification of oil sources, the properties of crude oil tend to be heavy and inferior, the salt content, the sulfur content and the acid content of the crude oil increase year by year, the difficulty of the electric desalting oil-water separation continuously increases, the raw material with high salt content seriously restricts the long-period operation of a subsequent device, the processing cost is increased, and the economic benefit is reduced.
In the traditional method for monitoring the salinity of the crude oil subjected to electric desalting and desalting, a manual sampling method is generally adopted, and the salinity of the crude oil subjected to desalting is obtained by regular (generally 1-2 times/day) analysis and assay for guiding production. Although a certain technical means is introduced in the sampling and analyzing process, the accuracy of the analysis result depends on the experience and the operation level of personnel to a certain extent, and meanwhile, if the sampling and the testing are frequently carried out, huge cost is brought. Because the working condition is always in fluctuation in the operation process of the device, the influence factor of the desalting effect is complex, and for a large number of working conditions which are not subjected to online sampling, the salt content of the crude oil after being desalted is difficult to accurately analyze and predict through artificial experience, so that the salt content after being desalted in the operation of the actual device is very common.
Disclosure of Invention
The invention provides an intelligent prediction diagnosis method for the crude oil salt content after the crude oil is removed in an electric desalting system based on random forests, aiming at overcoming the problems of long analysis and test time, influence of personnel level and low efficiency in the crude oil salt content prediction after the crude oil is removed by the traditional manual sampling analysis method.
In order to achieve the above purpose, the invention provides the following technical scheme: an intelligent prediction and diagnosis method for the salt content of crude oil after being removed by an electric desalting system of an atmospheric and vacuum device comprises the following steps:
s1, constructing an index system influencing the content of crude oil salt after an electric desalting system of an atmospheric and vacuum device is removed, and acquiring historical operation and real-time operation data of the electric desalting system of the atmospheric and vacuum device as sample data according to indexes in the system;
s2, preprocessing the sample data;
s3, constructing an intelligent prediction diagnosis model of the content of crude oil salt after the electric desalting system of the atmospheric and vacuum device based on the random forest by utilizing the preprocessed data;
s4, carrying out real-time prediction diagnosis on the electric desalting system through a constructed intelligent crude oil salt content prediction diagnosis model after the electric desalting system based on random forests is subjected to dehydration;
s5, comparing the predicted diagnosis result with the true value according to S4, and if the predicted diagnosis result is within the error range, indicating that the predicted diagnosis effect of the intelligent crude oil salt content prediction diagnosis model after the electric desalting system is removed is ideal; otherwise, further optimizing the intelligent crude oil salt content prediction diagnosis model after the electric desalting system is removed according to the result.
Preferably, the index includes 46 parameters, which are respectively: the sulfur content of crude oil before removal, the acid value of crude oil before removal, the water mass content of crude oil before removal, the density of crude oil before removal and the salt content of crude oil before removal; primary desalination tank transformers a1 and a2 and A3 voltage display values, primary desalination tank transformers a1 and a2 and A3 current display values, primary desalination tank transformers B1 and B2 and B3 voltage display values, primary desalination tank transformers B1 and B2 and B3 current display values; adjusting the boundary level of a first desalting tank to-1, adjusting the boundary level of a first desalting tank to-2, indicating the boundary level of the first desalting tank, adjusting the boundary level of a second desalting tank to-1, adjusting the boundary level of the second desalting tank to-2, indicating the boundary level of the second desalting tank, adjusting the front-back differential pressure indication of a mixer in front of the first desalting tank, adjusting the front-back differential pressure indication of a mixer in front of the second desalting tank to the pressure indication of crude oil in the first desalting tank, indicating the pressure indication of the first desalting tank and the pressure indication of the second desalting tank, indicating the temperature of crude oil in the first desalting tank, indicating the temperature of crude oil in the outlet of the second desalting tank, indicating and adjusting the flow rate of a water pump of the first desalting tank, indicating the purified water flow rate from a system to the second desalting tank, adjusting the liquid level of the second desalting tank, desalting drainage, desalting water temperature of the shell pass of a desalting water heat exchanger and desalting water temperature, and desulfurizing and purifying water flow rate to the second desalting tank, The desalting and draining temperature of the pipe pass of the desalting and draining cooler; accumulating the flow of desalted and drained water of the discharging device, indicating the temperature of desalted and drained water of the discharging device and leading the temperature to a demulsifier of a crude oil line before dehydration; the crude oil flow before the removal from the tank area is accumulated to 1 to 6.
Preferably, the preprocessing includes data cleansing, data integration, and data transformation.
Preferably, in S3, an intelligent prediction diagnosis model of the content of crude oil salt after being removed by the electric desalting system of the atmospheric and vacuum distillation plant based on the random forest is constructed, and the specific method is as follows:
s31, dividing the preprocessed data into a training sample set and a testing sample set according to a set proportion;
s32, generating N sample subsets from the training sample set by using a replaceable random sampling mode, wherein the number of samples in each sample subset is the same as that of the samples in the training sample set;
and S33, generating N decision trees by using the N sample subsets, randomly selecting 1 parameter from 46 parameters on each node of each decision tree as a branch, selecting a partition point based on an average error MSE minimization principle, and constructing the decision tree.
The decision Tree is generated by using a CART (Classification and Regression Tree) algorithm, the mean division error MSE is minimized to be a common feature measurement division mode in the CART algorithm, and the contents of li-hang statistical learning method, qinghua university press, chapter 5.5 can be referred to specifically.
And S34, according to the N decision trees generated in the step S33, predicting the sample according to a certain mechanism by integrating the test result of each tree, wherein the process is an intelligent prediction diagnosis model of the crude oil salt content after the removal of the electric desalting system of the atmospheric and vacuum device based on the random forest.
The set ratio is 2/3-4/5 training samples to total pre-processed data.
Preferably, in S32, the pre-processed data in the unselected training sample set is used as the out-of-bag data, and the out-of-bag data is used as the test data of the training sample set.
Preferably, the specific step of randomly selecting 1 parameter from 46 parameters is: randomly selecting m parameters from 46 parameters, wherein m is smaller than 46; and then selecting one parameter from the m parameters to branch according to the average error MSE minimization principle.
Preferably, the certain mechanism is to average the prediction results of all decision trees.
Preferably, the optimization in S5 is: and analyzing and evaluating the model result, if the prediction result precision of the model cannot meet the prediction requirement, adjusting the number of decision trees generated by the sample subsets in the S33, generating N 'decision trees by using N sample subsets, wherein N' is more than N, and then constructing and diagnosing an intelligent prediction diagnosis model of the crude oil salt content after the electric desalting system of the atmospheric and vacuum distillation plant based on the random forest in the same way.
The invention has the beneficial effects that:
aiming at the problems of prediction and standard exceeding intelligent diagnosis of the crude oil salt content after being removed by the electric desalting system, the invention constructs an intelligent prediction model of the crude oil salt content after being removed by the electric desalting system, adopts a random forest algorithm to train and learn historical data of the electric desalting system, excavates characteristics influencing the crude oil salt content after being removed by the electric desalting system, and determines comprehensive characteristic expressions of the characteristics when the electric desalting system exceeds the standard, thereby performing online intelligent prediction on the electric desalting system according to real-time operation data of the electric desalting system, performing intelligent early warning on the condition that the crude oil salt content after being removed exceeds the standard, effectively reducing the labor cost of electric desalting sampling analysis, and improving the efficiency of detecting the crude oil salt content after being removed by the electric desalting system.
Because the reason for the situation that the salt content of the crude oil after being desalted by the electric desalting system exceeds the standard is various, the existing research usually determines whether the electric desalting system sends out alarm information that the salt content of the crude oil after being desalted exceeds the standard through manual sampling and assay, and the method is time-consuming and labor-consuming and has certain defects. The characteristics of the salt content of the crude oil after being removed are usually not obvious, and a single characteristic variable capable of effectively representing the electric desalting effect does not exist, the indexes of the electric desalting system are not independently researched, a random forest algorithm is adopted, the integral characteristics of a plurality of influence factors are considered, the potential correlation among the plurality of influence factors is comprehensively considered, the output result is comprehensively judged, and the prediction efficiency and the prediction precision of the existing electric desalting system on the salt content of the crude oil after being removed can be greatly improved.
Drawings
FIG. 1 is a flow chart of a method for intelligently predicting and diagnosing the salt content of crude oil after being removed by an electric desalting system of an atmospheric and vacuum distillation plant based on a random forest according to an embodiment of the invention;
FIG. 2 is a flow chart of the post-removal crude oil salt content intelligent prediction based on the random forest algorithm, which is constructed in the method for predicting the post-removal crude oil salt content of the electric desalting system of the atmospheric and vacuum distillation plant according to the embodiment of the invention;
FIG. 3 is a scatter plot of the predicted values and the true post-stripping crude oil salt content of the test samples in the example of the present invention; FIG. 3A is a scatter plot of the predicted values and the true post-stripping crude oil salt content of the test samples in the example of the present invention; fig. 3B is a scatter plot of the model provided in the embodiment of the present invention, after training for a period of time, the operation data, the predicted value of the model, and the real salt content of the crude oil after stripping;
FIG. 4 shows the misjudgment ratio of the number of decision trees and the number of times of salt content exceeding standard (>3mg/l) according to an embodiment of the present invention;
FIG. 5 is a graph of the number of decision trees versus the mean square error MSE of the model prediction results provided by an embodiment of the present invention;
FIG. 6 is a graph showing the importance of the indicator variables to the salt content after removal, according to an embodiment of the present invention.
Detailed Description
The technical scheme of the invention is described in more detail by combining the embodiment and the attached drawings:
example 1
As shown in fig. 1, an intelligent prediction and diagnosis method for the salt content of crude oil after being removed by an electric desalting system of an atmospheric and vacuum device comprises the following steps:
s1, constructing an index system influencing the content of crude oil salt after an electric desalting system of an atmospheric and vacuum device is removed, and acquiring historical operation and real-time operation data of the electric desalting system of the atmospheric and vacuum device as sample data according to indexes in the system;
in this embodiment, data of the electric desalting system is obtained from the LIMS system and the sampling analysis system of the atmospheric and vacuum device as sample data; in the acquisition process, because some required index data in the system has incomplete data, in the selection process, data with higher data integrity degree is selected as sample data as much as possible.
The index system comprises 46 parameters such as an electric desalting process, operation data, crude oil analysis sampling data and the like, and the parameters are respectively as follows:
crude oil property parameters 5: the sulfur content of crude oil before removal, the acid value of crude oil before removal, the water mass content of crude oil before removal, the density of crude oil before removal and the salt content of crude oil before removal;
electric field intensity is 12: primary desalination tank transformers a1 and a2 and A3 voltage display values, primary desalination tank transformers a1 and a2 and A3 current display values, primary desalination tank transformers B1 and B2 and B3 voltage display values, primary desalination tank transformers B1 and B2 and B3 current display values;
relevant parameters of the electric desalting tank are 22: adjusting the boundary level of a first desalting tank to-1, adjusting the boundary level of a first desalting tank to-2, indicating the boundary level of the first desalting tank, adjusting the boundary level of a second desalting tank to-1, adjusting the boundary level of the second desalting tank to-2, indicating the boundary level of the second desalting tank, adjusting the front-back differential pressure indication of a mixer in front of the first desalting tank, adjusting the front-back differential pressure indication of a mixer in front of the second desalting tank to the pressure indication of crude oil in the first desalting tank, indicating the pressure indication of the first desalting tank and the pressure indication of the second desalting tank, indicating the temperature of crude oil in the first desalting tank, indicating the temperature of crude oil in the outlet of the second desalting tank, indicating and adjusting the flow rate of a water pump of the first desalting tank, indicating the purified water flow rate from a system to the second desalting tank, adjusting the liquid level of the second desalting tank, desalting drainage, desalting water temperature of the shell pass of a desalting water heat exchanger and desalting water temperature, and desulfurizing and purifying water flow rate to the second desalting tank, The desalting and draining temperature of the pipe pass of the desalting and draining cooler; accumulating the flow of the desalted and drained water of the outlet device and indicating the temperature of the desalted and drained water of the outlet device;
the injection amount of the demulsifier is 1: crude oil line demulsifier before dehydration;
crude oil throughput 6: the crude oil flow before the removal from the tank area is accumulated to 1 to 6.
The parameters belong to common parameters of an electric desalting system of an atmospheric and vacuum distillation device, have relatively specific meanings, and can be generally understood by a person skilled in the art or refer to Jiapeng forest, Rouyima pine, Chu happiness, crude oil electric desalting and dewatering technology [ M ] China petrochemical press, 2010, and other corresponding books.
S2, preprocessing the selected sample data; the pretreatment comprises data cleaning, data integration and data transformation, obvious abnormal data are removed, and the data characteristic conditions of normal and standard exceeding of the desalting effect of the electric desalting system are preliminarily and qualitatively analyzed.
S3, constructing an intelligent prediction diagnosis model of the content of crude oil salt after the electric desalting system of the atmospheric and vacuum device based on the random forest by using the preprocessed data, wherein the intelligent prediction diagnosis model comprises the following steps:
s31, dividing the preprocessed data into a training sample set and a test sample set according to the proportion of the training sample set in the preprocessed data 2/3-4/5;
s32, generating N sample subsets from the training sample set by using a replaceable random sampling mode, wherein the number of samples in each sample subset is the same as that of the samples in the training sample set; and the pre-processing data in the unselected training sample set is used as the data outside the bag, and the data outside the bag is used as the test data in the training sample set.
S33, generating N decision trees by using the N sample subsets, wherein m parameters are randomly selected from 46 parameters on each node of each decision tree, and m is smaller than 46; and then selecting one parameter from the m parameters to branch according to the average error MSE minimization principle.
The decision Tree is generated by using a CART (Classification and Regression Tree) algorithm, the mean division error MSE is minimized to be a common feature measurement division mode in the CART algorithm, and the contents of li-hang statistical learning method, qinghua university press, chapter 5.5 can be referred to specifically.
And S34, according to the N decision trees generated in the S33, integrating the test result of each tree, and then averaging the prediction results of all the decision trees.
S4, carrying out real-time prediction diagnosis on the electric desalting system through a constructed intelligent crude oil salt content prediction diagnosis model after the electric desalting system based on random forests is subjected to dehydration;
according to the intelligent prediction and diagnosis model of the salt content of the crude oil after the electric desalting system is removed, obtained in S4, relevant data of the electric desalting system are input into the model, monitoring and diagnosis are carried out on operation data of the electric desalting system in real time, and a predicted value of the salt content of the crude oil after the removal is given; and when the diagnosis result is that the salt content of the dehydrated crude oil exceeds the standard, sending alarm information to the electric desalting system and giving important parameter indexes for priority suggestion adjustment.
S5, comparing the predicted diagnosis result with the true value according to S4, and if the predicted diagnosis result is within the error range, indicating that the predicted diagnosis effect of the intelligent crude oil salt content prediction diagnosis model after the electric desalting system is removed is ideal; otherwise, further optimizing the intelligent crude oil salt content prediction diagnosis model after the electric desalting system is removed according to the result.
In this embodiment, the intelligent prediction and diagnosis model for the salt content of crude oil after being removed by the constructed electric desalting system is trained and feedback-adjusted. The method has the advantages that more removed crude oil salt content exceeds the standard and normal operation data of the electric desalting system are collected, so that the model covers all working condition characteristics as far as possible, the model is updated in time when the model is trained regularly, particularly the crude oil quality changes and the device treatment capacity changes, and the accuracy of the model is continuously improved.
Example 2
An intelligent prediction diagnosis model of the content of crude oil salt after being removed by an atmospheric and vacuum device electric desalting system based on random forests is constructed, and the steps are as follows:
s31, preprocessing data according to the ratio of 0.8: 0.2 dividing the test sample set into a training sample set and a testing sample set;
and S32, generating N sample subsets from the training sample set by using a replaceable random sampling mode, wherein the number N of the decision trees selected in the embodiment is 100 by considering the running time of the model and the precision of the model. The number of samples in each of the subsets of samples is the same as the number of samples in the set of training samples.
And S33, when the random forest algorithm is used for regression analysis, selecting the division characteristics according to the minimum mean error (MSE) to select parameters, and then constructing the whole decision tree and the random forest by adopting a recursive method. Generating 100 decision trees by using 100 sample subsets, wherein m parameters are randomly selected from 46 parameters on each node of each decision tree, and m is smaller than 46; and then selecting one parameter from the m parameters to branch based on the principle of mean error minimum MSE.
The minimization of the square error (MSE) is a common feature metric division mode in a CART algorithm used in a decision tree generation algorithm, and parameters and division points are selected by using the principle of minimizing the MSE by the average error, which can be specifically carried out by referring to the contents of Li aviation, statistical learning method, Qing Hua university Press, chapter 5.5.
And S34, predicting the test sample set according to the 100 generated decision trees, and then averaging the prediction results of all the decision trees to obtain a final prediction result.
The random forest algorithm can well utilize the randomness thereof, including randomly generating a sub-sample set and randomly selecting sub-features, minimize the correlation among trees and improve the overall performance. Meanwhile, because the generation time of each tree is very short, and the forest can be parallelized, the random forest regression speed is very fast.
Assume that random forests have many CART decision tree regression models:
h(X；θk)，i＝1,2,…,K
a combined regression model formed by the method obtains a regression model prediction sequence through K rounds of training:
{h(X；θ1)，h(X；θ2)，…，h(X；θk)}
and then a combined prediction model is formed by the sequences, and the final prediction result is as follows:
in this embodiment, taking the situation that the salt content of the crude oil after the control of a certain enterprise requires 3mg/L as an example, the obtained model results are as follows:
for the test data set, fig. 3A, 3B show scatter plots of true and predicted salt content of the crude oil after stripping. As can be seen from the figure, for the predicted value of the desalted salt content of the electric desalting system and the real sampled and analyzed value of the desalted crude oil, most cases have certain coincidence and basically keep the consistency on the trend. The method shows that the intelligent prediction and diagnosis model for the salt content of the crude oil after the electric desalting system is removed has certain accuracy on the prediction of the salt content of the removed crude oil. As can be seen from fig. 4 and 5, for the case where the salt content after being desalted exceeds the standard (>, 3mg/L Nacl), the output results of the intelligent prediction and diagnosis model for the salt content of crude oil after being desalted by the electric desalting system are shown in the following table 1:
table 1 model diagnosis result of whether salt content after dehydration exceeds standard
From the table above, it can be seen that, for the situation that the salt content of the crude oil after being desalted by the electric desalting system exceeds the standard, the error rate of verifying whether the salt content after being desalted exceeds the standard based on the test data is 0.29%, which indicates that the model is relatively ideal for the prediction and diagnosis result of whether the salt content of the crude oil after being desalted exceeds the standard.
Based on the analysis of the intelligent prediction and diagnosis model for the salt content of the desalted crude oil of the electric desalting system, from the angle of data analysis, the relative importance of the influence factors on the salt content of the desalted crude oil is also pointed out, as shown in fig. 6, the abscissa number of fig. 6 is an index number, and the abscissa sequentially represents the sulfur content of the crude oil before desalting, the acid value of the crude oil before desalting, the water content of the crude oil before desalting, the density of the crude oil before desalting, the salt content of the crude oil before desalting, the voltage display values of the primary desalting tank transformers A1, A2 and A3, the current display values of the primary desalting tank transformers A1, A2 and A3, the voltage display values of the primary desalting tank transformers B1, B2 and B3, the current display values of the primary desalting tank transformers B1, B2 and B3, the boundary position of the primary desalting tank is adjusted to be-1, the boundary position of the primary desalting tank is adjusted to be-2, the boundary position of the primary desalting tank is indicated, and the secondary desalting tank is adjusted to be-1, Adjusting the interface level of a secondary desalting tank-2, indicating the interface level of a secondary desalting tank, adjusting the front and back differential pressure indication of a mixer in front of a primary desalting tank, adjusting the front and back differential pressure indication of a mixer in front of a secondary desalting tank, indicating the crude oil pressure of the primary desalting tank, indicating the pressure of the secondary desalting tank, indicating the crude oil temperature of the primary desalting tank, indicating the crude oil temperature of an outlet of the secondary desalting tank, indicating the flow of a water injection pump outlet of the primary desalting tank, adjusting the purified water flow from a system to the secondary desalting tank, adjusting the liquid level of the secondary desalting tank, adjusting and indicating the desalted water-desalted water injection heat exchanger shell pass outlet desalted water temperature, the desulfurized and purified water flow of the secondary desalting discharge water cooler and tube pass desalted water temperature; accumulating the flow of desalted and drained water of the discharging device, indicating the temperature of desalted and drained water of the discharging device, accumulating the flow of crude oil before dewatering to a crude oil line demulsifier before dewatering and accumulating the flow of crude oil before dewatering from a tank area by 1-6. The key influence on the salt content of the crude oil after stripping can be seen in the ranking in FIG. 6: and indicating the front and back differential pressure in front of the primary desalting tank by the salt content of the crude oil, the density of the crude oil, the boundary position of the electric desalting tank, the sulfur content of the crude oil, the outlet temperature of the primary electric desalting tank, the voltage of the electric desalting tank and the demulsifying agent amount of the crude oil before desalting.
And in addition, analyzing and evaluating the results of the intelligent prediction diagnosis model of the crude oil salt content after the electric desalting system is removed, and determining whether to further optimize the intelligent prediction diagnosis model of the crude oil salt content after the electric desalting system is removed according to the analysis and evaluation results.
If the prediction diagnosis result is within the error range, the prediction diagnosis effect of the intelligent prediction diagnosis model of the crude oil salt content after the electric desalting system is removed is ideal; otherwise, further optimizing the intelligent prediction diagnosis model of the crude oil salt content after the electric desalting system is desalted according to the result, optimizing the number of the decision trees generated by mainly adjusting the sample subsets, and generating N 'decision trees by using N sample subsets, wherein N' is more than N. By collecting more removed crude oil salt content exceeding standard and normal operation data of the electric desalting system, the model covers all working condition characteristics as far as possible, and the accuracy of the model is continuously improved.
The above embodiments are only used to illustrate the technical solutions of the present invention, and do not limit the present invention; the documents cited above are only intended as supplementary illustrations of the prior art and do not limit the invention; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those skilled in the art that: any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. An intelligent prediction and diagnosis method for the salt content of crude oil after being removed by an electric desalting system of an atmospheric and vacuum device is characterized by comprising the following steps:
s1, constructing an index system influencing the content of crude oil salt after an electric desalting system of an atmospheric and vacuum device is removed, and acquiring historical operation and real-time operation data of the electric desalting system of the atmospheric and vacuum device as sample data according to indexes in the system;
s2, preprocessing the sample data;
s3, constructing an intelligent prediction diagnosis model of the content of crude oil salt after the electric desalting system of the atmospheric and vacuum device based on the random forest by utilizing the preprocessed data;
s4, carrying out real-time prediction diagnosis on the electric desalting system through a constructed intelligent crude oil salt content prediction diagnosis model after the electric desalting system based on random forests is subjected to dehydration;
s5, comparing the predicted diagnosis result with the true value according to S4, and if the predicted diagnosis result is within the error range, indicating that the predicted diagnosis effect of the intelligent crude oil salt content prediction diagnosis model after the electric desalting system is removed is ideal; otherwise, further optimizing the intelligent crude oil salt content prediction diagnosis model after the electric desalting system is removed according to the result.
2. The method for intelligently predicting and diagnosing the salt content of the crude oil after being desalted by the electric desalting system of the atmospheric and vacuum distillation unit as claimed in claim 1, wherein the index comprises 46 parameters which are respectively: the crude oil before desalting comprises the sulfur content of crude oil before desalting, the acid value of crude oil before desalting, the water mass content of crude oil before desalting, the density of crude oil before desalting, the salt content of crude oil before desalting, voltage display values of a primary desalting tank transformer A1, A2 and A3, current display values of a primary desalting tank transformer A1, A2 and A3, voltage display values of a primary desalting tank transformer B1, B2 and B3, current display values of a primary desalting tank transformer B1, B2 and B3, boundary position adjustment of a primary desalting tank-1, boundary position adjustment of a primary desalting tank-2, boundary position indication of a secondary desalting tank, boundary position adjustment of a primary desalting tank-front mixer-rear differential pressure adjustment of a primary desalting tank-rear mixer, differential pressure adjustment of a secondary desalting tank-front mixer-rear differential pressure adjustment of a secondary desalting tank, pressure indication of crude oil to the primary desalting tank, pressure indication of a secondary desalting tank, Crude oil temperature to a first-stage desalting tank, crude oil temperature indication at an outlet of the first-stage desalting tank, crude oil temperature indication at an outlet of a second-stage desalting tank, flow indication adjustment at an outlet of a water injection pump of the first-stage desalting tank, purified water flow from a system to the second-stage desalting tank, liquid level adjustment indication of the second-stage desalting tank, desalted and drained water temperature at an outlet of a shell pass of a desalted and drained water heat exchanger, desulfurized and purified water flow to the second-stage desalting tank, desalted and drained water temperature of a desalted and drained water cooler tube pass, desalted and drained water flow accumulation of an outlet device, desalted and drained water temperature indication of the outlet device, demulsifier of a crude oil line before desalting, and crude oil flow accumulation before desalting from a tank area is 1-6.
3. The intelligent prediction and diagnosis method for the salt content of the crude oil after the electric desalting system of the atmospheric and vacuum distillation unit as claimed in claim 1, wherein the preprocessing comprises data cleaning, data integration and data transformation.
4. The intelligent prediction and diagnosis method for the content of the crude oil salt after the electric desalting system of the atmospheric and vacuum device as claimed in claim 1, wherein in S3, an intelligent prediction and diagnosis model for the content of the crude oil salt after the electric desalting system of the atmospheric and vacuum device based on random forests is constructed, and the specific method comprises the following steps:
s31, dividing the preprocessed data into a training sample set and a testing sample set according to a set proportion;
s32, generating N sample subsets from the training sample set by using a replaceable random sampling mode, wherein the number of samples in each sample subset is the same as that of the samples in the training sample set;
and S33, generating N decision trees by using the N sample subsets, randomly selecting 1 parameter from 46 parameters on each node of each decision tree as a branch, selecting a partition point based on an average error MSE minimization principle, and constructing the decision tree.
And S34, according to the generated N decision trees, predicting a sample according to a certain mechanism by integrating the test result of each tree, wherein the process is an intelligent prediction diagnosis model of the salt content of the crude oil after the crude oil is removed by the electric desalting system of the atmospheric and vacuum device based on the random forest.
5. The method as claimed in claim 4, wherein the predetermined ratio is 2/3-4/5 training samples to total preprocessing data.
6. The method as claimed in claim 4, wherein in step S32, the pre-processed data in the unselected training sample set is used as the off-bag data, and the off-bag data is used as the test data in the training sample set.
7. The intelligent prediction and diagnosis method for the salt content of the crude oil after the removal of the electric desalting system of the atmospheric and vacuum distillation unit as claimed in claim 4, wherein the specific step of randomly selecting 1 parameter from 46 parameters is as follows: randomly selecting m parameters from 46 parameters, wherein m is smaller than 46; and then selecting one parameter from the m parameters to branch according to the average error minimum MSE principle.
8. The intelligent prediction and diagnosis method for the salt content of crude oil after being removed by the electric desalting system of an atmospheric and vacuum device as claimed in claim 4, wherein the certain mechanism is to take the average value of all the decision tree prediction results.
9. The intelligent prediction and diagnosis method for the salt content of the crude oil after the electric desalting system of the atmospheric and vacuum distillation unit as claimed in claim 1, wherein in S4, the real-time prediction and diagnosis specifically comprises: inputting indexes in an index system of the electric desalting system into an intelligent prediction and diagnosis model of the salt content of the crude oil after the electric desalting system of the atmospheric and vacuum device is removed, monitoring and diagnosing operation data of the electric desalting system in real time, and giving a predicted value of the salt content of the crude oil after the removal; and when the diagnosis result is found to be that the salt content of the dehydrated crude oil exceeds the standard, sending alarm information to the electric desalting system and giving important parameter indexes for priority suggestion adjustment.
10. The method for intelligently predicting and diagnosing the content of the crude oil salt after the electric desalting system of the atmospheric and vacuum device as claimed in claim 4, wherein in the step S5, the number of decision trees generated by the sample subsets in the step S33 is adjusted, N 'decision trees are generated by using N sample subsets, N' > N, and then an intelligent prediction and diagnosis model of the content of the crude oil salt after the electric desalting system of the atmospheric and vacuum device based on the random forest is constructed in the same manner.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111127300.2A CN113869578A (en) | 2021-09-26 | 2021-09-26 | Intelligent prediction and diagnosis method for salt content of crude oil after removal of electric desalting system of atmospheric and vacuum distillation unit |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111127300.2A CN113869578A (en) | 2021-09-26 | 2021-09-26 | Intelligent prediction and diagnosis method for salt content of crude oil after removal of electric desalting system of atmospheric and vacuum distillation unit |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113869578A true CN113869578A (en) | 2021-12-31 |
Family
ID=78994305
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111127300.2A Pending CN113869578A (en) | 2021-09-26 | 2021-09-26 | Intelligent prediction and diagnosis method for salt content of crude oil after removal of electric desalting system of atmospheric and vacuum distillation unit |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113869578A (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107179503A (en) * | 2017-04-21 | 2017-09-19 | 美林数据技术股份有限公司 | The method of Wind turbines intelligent fault diagnosis early warning based on random forest |
| CN110428113A (en) * | 2019-08-09 | 2019-11-08 | 云南电网有限责任公司电力科学研究院 | A kind of predicting model for dissolved gas in transformer oil method based on random forest |
| CN111985796A (en) * | 2020-08-07 | 2020-11-24 | 华中科技大学 | Method for predicting concrete structure durability based on random forest and intelligent algorithm |
-
2021
- 2021-09-26 CN CN202111127300.2A patent/CN113869578A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107179503A (en) * | 2017-04-21 | 2017-09-19 | 美林数据技术股份有限公司 | The method of Wind turbines intelligent fault diagnosis early warning based on random forest |
| CN110428113A (en) * | 2019-08-09 | 2019-11-08 | 云南电网有限责任公司电力科学研究院 | A kind of predicting model for dissolved gas in transformer oil method based on random forest |
| CN111985796A (en) * | 2020-08-07 | 2020-11-24 | 华中科技大学 | Method for predicting concrete structure durability based on random forest and intelligent algorithm |
Non-Patent Citations (1)
| Title |
|---|
| 朱建新等: "基于操作参数大数据分析的电脱盐系统运行优化技术研究", 《石油化工自动化》, vol. 57, no. 5, 20 September 2021 (2021-09-20), pages 6 - 10 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106845796B (en) | One kind is hydrocracked flow product quality on-line prediction method | |
| CN109814513B (en) | Catalytic cracking unit optimization method based on data model | |
| CN109408774B (en) | Method for predicting sewage effluent index based on random forest and gradient lifting tree | |
| CN101501593B (en) | Method to analyze economics of asset management solutions for nuclear steam generators | |
| CN113724800A (en) | Catalytic cracking unit simulation prediction method based on molecular level mechanism model and big data technology | |
| CN110015754B (en) | Online monitoring system and method for distributed type sewage effluent quality standard exceeding early warning | |
| CN111160776A (en) | Method for detecting abnormal working condition in sewage treatment process by utilizing block principal component analysis | |
| CN101673096B (en) | Soft-measuring method for density in concentration process of salvia miltiorrhiza injection production | |
| CN111191855B (en) | Water quality abnormal event identification and early warning method based on pipe network multi-element water quality time sequence data | |
| CN116595881A (en) | Method for predicting temperature field of crude oil storage tank and evaluating oil storage safety | |
| Yang et al. | Teacher–Student Uncertainty Autoencoder for the Process-Relevant and Quality-Relevant Fault Detection in the Industrial Process | |
| CN114510467A (en) | Intelligent water affair data abnormity identification method | |
| CN117421616B (en) | Mine shaft detection system and method | |
| CN113869578A (en) | Intelligent prediction and diagnosis method for salt content of crude oil after removal of electric desalting system of atmospheric and vacuum distillation unit | |
| CN115879379B (en) | Intelligent corrosion monitoring and early warning method and system for equipment | |
| CN117312972A (en) | Method for identifying health state of scraper conveyor speed reducer | |
| CN106770861B (en) | The evaluation method of oil-filled transformer on-line monitoring availability of data | |
| CN111126694A (en) | Time series data prediction method, system, medium and device | |
| CN116432856A (en) | Pipeline dynamic early warning method and device based on CNN-GLSTM model | |
| CN115983534A (en) | Method and system for evaluating state of sewage treatment process | |
| Hilgefort | Big data analysis using Bayesian network modeling: a case study with WG-ICDA of a gas storage field | |
| CN108388224A (en) | A kind of Desalting and Dewatering from Crude Oil process operation method for monitoring state based near infrared spectrum | |
| CN109273058A (en) | A kind of composite algorism for the exceeded early warning of anaerobic processes volatile fatty acid | |
| Al-Jlibawi et al. | The efficiency of soft sensors modelling in advanced control systems in oil refinery through the application of hybrid intelligent data mining techniques | |
| TR202014625A1 (en) | Prediction system and method of nitrogen in the raw material for hydrocracker reactor temperature optimization. |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |