Offshore platform injection water quality multi-parameter intelligent online monitoring method and system

Abstract

The invention discloses an offshore platform water injection quality multi-parameter intelligent online monitoring method and system, and relates to the field of intelligent online monitoring, and the method comprises the steps: determining abnormal data in a water injection sample data set of an offshore platform, and correcting the water injection sample data set; extracting a characteristic parameter set of the corrected water injection sample data set; dividing the feature parameter set to obtain n first-level feature sets; performing data fusion on the n first level feature sets to obtain p first data sets; determining m difference sets in the p first data sets; performing data fusion on the m difference sets to obtain q second data sets; determining monitoring data of the water injection sample data set according to the second data set and the first data set; predicting the water injection quality data through a pre-trained prediction model; and determining a prediction result of the injection water quality of the offshore platform according to the prediction data. The problem of inaccurate manual detection can be effectively solved.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of intelligent online monitoring, and in particular, to a multi-parameter intelligent online monitoring method and system for the water quality of water injection on an offshore platform. Background Art

[0002] In the field of offshore oil development technology, it is necessary to simultaneously conduct online monitoring of multiple parameters of the water quality of water injection on an offshore platform, such as oil in water, suspended solids, dissolved oxygen, corrosion rate, total iron, sulfide, pH, etc., in order to achieve the purpose of timely adjusting the production water treatment process of offshore oilfields and timely guiding the water injection operation of offshore oilfields.

[0003] The existing technology for monitoring the water quality of water injection on an offshore platform mainly involves sampling seawater and analyzing the sampled seawater in a laboratory. A series of tests are carried out on the seawater samples in the laboratory through manual detection methods. However, the process of sending the seawater samples to the laboratory for testing not only wastes time but also leads to inaccurate monitoring of the water quality of water injection on the offshore platform due to the diversity of real-time changes in seawater.

[0004] Therefore, there is an urgent need for a multi-parameter intelligent online monitoring method and system for the water quality of water injection on an offshore platform to solve the problem of inaccurate manual detection. Summary of the Invention

[0005] The embodiments of the present application provide a multi-parameter intelligent online monitoring method and system for the water quality of water injection on an offshore platform to solve the problem of inaccurate manual detection.

[0006] To achieve the above object, the embodiments of the present application adopt the following technical solutions: In a first aspect, a multi-parameter intelligent online monitoring method for the water quality of water injection on an offshore platform is provided. The method includes: Obtaining a water injection sample data set of the offshore platform, where the water injection sample data set includes suspended solid data, dissolved oxygen data, corrosion rate data, total iron content data, sulfide data, pH data, and oil content data; Determining abnormal data in the water injection sample data set; Correcting the water injection sample data set according to the abnormal data to obtain a corrected water injection sample data set; Extracting a feature parameter set of the corrected water injection sample data set; Dividing the feature parameter set to obtain n first-level feature sets, where n≥2 and n is an integer; Performing data fusion on the n first-level feature sets to obtain p first data sets, where p < n and p is a positive integer; Determining m difference sets in the p first data sets, where m < p and m is a positive integer; Performing data fusion on the m difference sets to obtain q second data sets, where q<m, and q is a positive integer; Determining monitoring data of the water injection sample data set according to the second data set and the first data set; Inputting the monitoring data into a pre-trained prediction model to obtain injection water quality data at a future time node; Based on the injection water quality data, a prediction result of the injection water quality of the offshore platform is determined.

[0007] In a possible implementation manner of the first aspect, determining abnormal data in the water injection sample data set includes: Acquire the water injection sample data set within a preset period; Dividing the water injection sample data set to obtain a plurality of water injection sample data subsets; generating a scatter plot according to the water injection sample data in the plurality of water injection sample data subsets; Based on a preset first clustering radius, cluster analysis is performed on the scatter plot to determine abnormal data.

[0008] In a possible implementation manner of the first aspect, the correcting the water injection sample data set according to the abnormal data to obtain a corrected water injection sample data set includes: Dividing the water injection sample data set into a first data set and a second data set, wherein the overlapping points of the first data set and the second data set are the abnormal data; Calculate the average value of each two adjacent data in the first data set to obtain a first correction data set; Calculate the average value of every two adjacent data in the second data set to obtain a second corrected data set; Inputting the first correction data set and the second correction data set into the Kalman filter equation respectively to obtain a first correction result and a second correction result; using an average value of the first correction result and the second correction result as correction data for the abnormal data; The water injection sample data set is corrected in combination with the correction data to obtain a corrected water injection sample data set.

[0009] In a possible implementation of the first aspect, the feature parameter set includes a suspended matter feature parameter set, a dissolved oxygen feature parameter set, a corrosion rate feature parameter set, a total iron content feature parameter set, a sulfide feature parameter set, a pH feature parameter set, and an oil content feature parameter set.

[0010] In a possible implementation manner of the first aspect, dividing the feature parameter set to obtain n first-level feature sets includes: All the feature parameter sets are divided into a plurality of first-level feature sets according to corresponding feature parameter categories, wherein the number of feature parameters in each first-level feature set is the same, and the number of the first-level feature sets is n.

[0011] In a possible implementation manner of the first aspect, the feature parameter categories include turbidity features, chromaticity features, pH features, oxygen demand features, conductivity features, and dissolved oxygen content features, and all of the feature parameter sets are divided into a plurality of first-level feature sets according to the corresponding feature parameter categories, including: Dividing the suspended matter characteristic parameter set according to the turbidity characteristics to obtain a plurality of suspended matter characteristic parameter subsets; Dividing the dissolved oxygen characteristic parameter set according to the dissolved oxygen content characteristics to obtain a plurality of dissolved oxygen characteristic parameter subsets; Dividing the corrosion rate characteristic parameter set according to the conductivity characteristics to obtain a plurality of corrosion rate characteristic parameter set subsets; Dividing the total iron content characteristic parameter set according to the conductivity characteristics to obtain a plurality of total iron content characteristic parameter set subsets; Dividing the sulfide characteristic parameter set according to the oxygen demand characteristics to obtain a plurality of sulfide characteristic parameter set subsets; Dividing the pH characteristic parameter set according to the acidity and alkalinity characteristics to obtain a plurality of pH characteristic parameter set subsets; Dividing the oil content characteristic parameter set according to the chromaticity feature to obtain a plurality of oil content characteristic parameter subsets; The suspended matter characteristic parameter subset, the dissolved oxygen characteristic parameter subset, the total iron content characteristic parameter subset, the sulfide characteristic parameter subset, the pH characteristic parameter subset, and the oil content characteristic parameter subset are summarized to obtain a plurality of characteristic parameter subsets.

[0012] In a possible implementation manner of the first aspect, screening the first data set to determine m difference sets of the first data set includes: Acquire the quantity and category of all the feature parameters in all the first data sets; Based on a preset number of characteristic parameters and a preset category of characteristic parameters, the number of characteristic parameters and the category of characteristic parameters in all the first data sets are respectively identified to obtain a subset of the number of characteristic parameters being abnormal, a subset of the category of characteristic parameters being abnormal, and a subset of the number and category of characteristic parameters being abnormal; Marking a subset of the characteristic parameter quantity abnormality, a subset of the characteristic parameter category abnormality, a subset of the characteristic parameter category abnormality, and a subset of the characteristic parameter quantity and category abnormality to obtain a first type of abnormality, a second type of abnormality, and a third type of abnormality; The feature parameter subsets having the first type of anomalies, the second type of anomalies, and the third type of anomalies are screened to determine m difference sets of the first data set.

[0013] In a possible implementation manner of the first aspect, determining the monitoring data of the water injection sample data set according to the second data set and the first data set includes: Classifying the second data set to determine the types and quantities of all characteristic parameters in the second data set; Performing first-category anomaly identification on all the feature parameter set subsets in the first data set to obtain first-category anomaly subsets, wherein the first-category anomaly subsets include parameter quantity missing subsets and parameter quantity duplication subsets; Identifying the number of characteristic parameters of the parameter number missing subset to obtain the number of missing parameters; Based on the number of the missing parameters and the number of all characteristic parameters in the second data set, interpolating the missing subset of the number of parameters to obtain first interpolated data; Identifying the number of characteristic parameters of the parameter number repetition subset to obtain the number of repetition parameters; Based on the number of repeated parameters, similarity calculation is performed on the repeated subset of the number of parameters to determine the repeated parameters, and the repeated parameters are screened out to obtain first screened-out data; Determine first target data based on the first interpolation data and the first screened data; Performing second-category anomaly identification on all the feature parameter set subsets in the first data set to obtain second-category anomaly subsets, where the second-category anomaly subsets include parameter type missing subsets and parameter type repeated subsets; Obtaining the locations of the missing subset of the parameter category and the repeated subset of the parameter category; Identifying the characteristic parameter category of the parameter category missing subset to obtain the category of the missing parameter; Based on the category of the missing parameter and the categories of all characteristic parameters in the second data set, interpolating the missing subset of the parameter category to obtain second interpolated data; Identifying the characteristic parameter category of the repeated subset of the parameter category to obtain the category of the repeated parameter; Based on the category of the repeated parameter, similarity calculation is performed on the repeated subset of the parameter category to determine the repeated category, and the repeated category is screened out to obtain second screened data; Determine second target data by combining the second interpolation data and the second screened data; Performing third-category anomaly identification on all the feature parameter set subsets in the first data set to obtain third-category anomaly subsets, wherein the third-category anomaly subsets include a subset in which both parameter type and quantity are missing and a subset in which both parameter type and quantity are missing; Identify the type and quantity of the characteristic parameters of the missing subset of the parameter type and quantity, and obtain the type and quantity of the missing parameters; Based on the type and quantity of the missing parameters and the type and quantity of all characteristic parameters in the second data set, interpolating the missing subset of parameter types and quantities to obtain third interpolated data; Identify the characteristic parameter types and quantities of the repeated subset of the parameter types and quantities, and obtain the types and quantities of repeated parameters; Based on the type and quantity of the repeated parameters, similarity calculation is performed on the repeated subsets of the parameter type and quantity to determine the repeated type and quantity, and the repeated type and quantity are screened out to obtain third screened-out data; Determine third target data by combining the third interpolation data and the third screened data; The monitoring data of the water injection sample data set is determined in combination with the first target data, the second target data and the third target data.

[0014] In a second aspect, the present application provides a machine-readable storage medium having instructions stored thereon, the instructions being used to enable a machine to execute the above-mentioned multi-parameter intelligent online monitoring method for water quality of offshore platform injection.

[0015] In a third aspect, the present application provides an offshore platform water injection water quality multi-parameter intelligent online monitoring system, comprising: a memory configured to store instructions; and The processor is configured to call the instructions from the memory and implement the above-mentioned multi-parameter intelligent online monitoring method for water quality of offshore platform water injection when executing the instructions.

[0016] Through the above technical solution, by obtaining the water injection sample data set of the offshore platform, determining the abnormal data in the water injection sample data set and correcting the water injection sample data set, potential problems in the data can be discovered in time, the accuracy of the sample data set can be ensured, the characteristic parameter set of the corrected water injection sample data set is extracted and the characteristic parameter set is divided to obtain the first-level feature set, and the first-level feature set is data fused. By integrating the data of multiple feature sets, it can help improve the quality and consistency of the data. Determine the difference set in the first data set, fuse the difference set, determine the second data set, and further fuse the difference data, which can not only reduce the uncertainty of the difference data, but also ensure the accuracy of the monitoring data obtained through the second data set, and effectively improve the efficiency and accuracy of online monitoring of water quality of offshore platform water injection.

[0017] Other features and advantages of the embodiments of the present application will be described in detail in the subsequent specific implementation section. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A schematic diagram of a process for a multi-parameter intelligent online monitoring method of water quality for offshore platform injection provided in an embodiment of the present application; Figure 2 A structural schematic diagram of a method for multi-parameter intelligent online monitoring of water quality for offshore platform injection provided in an embodiment of the present application. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. It should be understood that the specific implementation methods described herein are only used to illustrate and explain the embodiments of the present application, and are not used to limit the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0020] It should be noted that if the embodiments of the present application involve directional indications (such as up, down, left, right, front, back...), such directional indications are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0021] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of this application, the descriptions of "first", "second", etc. are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include at least one such feature. In addition, the technical solutions between various embodiments may be combined with each other, but it must be based on the ability of those of ordinary skill in the art to implement. When the combination of technical solutions results in contradictions or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection required by this application.

[0022] Figure 1 Schematically shows a schematic flow chart of a multi-parameter intelligent on-line monitoring method for the water injection quality of an offshore platform according to an embodiment of the present application. As Figure 1 shown, an embodiment of the present application provides a multi-parameter intelligent on-line monitoring method for the water injection quality of an offshore platform, and the method may include the following steps.

[0023] S101. Obtain a water injection sample data set of the offshore platform, where the water injection sample data set includes suspended solid data, dissolved oxygen data, corrosion rate data, total iron content data, sulfide data, pH data, and oil content data; S102. Determine abnormal data in the water injection sample data set; S103. Correct the water injection sample data set according to the abnormal data to obtain a corrected water injection sample data set; S104. Extract a feature parameter set of the corrected water injection sample data set; S105. Divide the feature parameter set to obtain n first-level feature sets, where n≥2 and n is an integer; S106. Perform data fusion on the n first-level feature sets to obtain p first data sets, where p < n and p is a positive integer; S107. Determine m difference sets in the p first data sets, where m < n and m is a positive integer; S108. Perform data fusion on the m difference sets to obtain q second data sets, where q < m and q is a positive integer; S109. Determine the monitoring data of the water injection sample data set according to the second data set and the first data set; S110. Input the monitoring data into a pre-trained prediction model to obtain the water injection quality data at future time nodes; S111. Determine the prediction result of the water injection quality of the offshore platform based on the water injection quality data.

[0024] First, obtain the water injection sample data set of the offshore platform. In this embodiment, the water injection sample data set includes suspended matter data, dissolved oxygen data, corrosion rate data, total iron content data, sulfide data, pH data, and oil content data. By monitoring the suspended matter, dissolved oxygen content, corrosion rate, total iron content, sulfide, pH value, and oil content in the water injection samples, it is used to understand and evaluate the water quality of seawater. After obtaining the water injection sample data set, determine the abnormal data in the water injection sample data set. Specifically, first, obtain the water injection sample data within a preset period. In this embodiment, the preset period can be determined according to the actual situation. The water injection sample data set is divided to obtain multiple water injection sample data subsets. Generate a scatter plot for the multiple water injection sample data subsets. A scatter plot is a graph obtained by plotting points in a rectangular coordinate system using two corresponding variable values ​​as the horizontal and vertical coordinates of a point in the graph, and is used to observe the degree and type of correlation between two variables. Based on a preset first clustering radius, perform cluster analysis on the scatter plot to determine the abnormal data. In this embodiment, the preset first clustering radius can be determined according to the actual situation. Cluster analysis refers to dividing the data set into multiple groups based on the similarities between samples, so that samples in the same subset have some similar attributes. ‌‌ After determining the abnormal data, the water injection sample data set is corrected according to the abnormal data to obtain a corrected water injection sample data set. Specifically, first, the water injection sample data set is divided into a first data set and a second data set, wherein the overlap point of the first data set and the second data set is the abnormal data. In this embodiment, the abnormal data can be suspended matter abnormal data, dissolved oxygen content abnormal data, corrosion rate abnormal data, total iron content abnormal data, sulfide abnormal data, pH abnormal data and oil content abnormal data. By calculating the average value of each adjacent two data in the first data set, a first correction data set is obtained, and by calculating the average value of each adjacent two data in the second data set, a second correction data set is obtained. The first correction data set and the second correction data set are respectively input into the Kalman filter equation to obtain a first correction result and a second correction result. The Kalman filter equation is an algorithm that uses a linear system state equation to optimally estimate the system state by inputting and outputting observation data of the system. The average value of the first correction result and the second correction result is used as the correction data of the abnormal data. The water injection sample data set is corrected by the correction data of the abnormal data to obtain a corrected water injection sample data set.

[0025] Extract the characteristic parameter set of the corrected water injection sample data set. In this embodiment, the characteristic parameter of the suspended matter data sample data set is turbidity, the characteristic parameter of the dissolved oxygen data set is dissolved oxygen content, the characteristic parameter of the corrosion rate data set is conductivity, the characteristic parameter of the total iron content data set is conductivity, the characteristic parameter of the sulfide data set is oxygen demand, the characteristic parameter of the pH data set is acidity and alkalinity, and the characteristic parameter of the oil content data set is chromaticity.

[0026] The feature parameter set is divided to obtain n first-level feature sets, where n≥2 and n is an integer. Specifically, all feature parameter sets are divided into multiple first-level feature sets according to corresponding feature parameter categories. In this embodiment, the first-level feature set is obtained by dividing the feature parameter set of the water injection sample data set after extraction and correction. The number of feature parameters in each first-level feature set is the same. In this embodiment, the corresponding feature parameter categories are turbidity feature, chromaticity feature, pH feature, oxygen demand feature, conductivity feature and dissolved oxygen content feature. The suspended matter feature parameter set is divided according to the turbidity feature, the dissolved oxygen feature parameter set is divided according to the dissolved oxygen content feature, the corrosion rate feature parameter set is divided according to the conductivity feature, the total iron content feature parameter set is divided according to the conductivity feature, the sulfide feature parameter set is divided according to the oxygen demand feature, the pH feature parameter set is divided according to the pH feature, and the oil content feature parameter set is divided according to the chromaticity feature. After the feature parameter set is divided, n first-level feature sets are obtained, where n≥2 and n is an integer, that is, the number of first-level feature sets is greater than or equal to 2 and the number is an integer.

[0027] After obtaining n first-level feature sets, the n first-level feature sets are subjected to data fusion to obtain p first data sets, where p < n and p is a positive integer. In this embodiment, the feature parameter set includes a suspended solid feature parameter set, a dissolved oxygen feature parameter set, a corrosion rate feature parameter set, a total iron content feature parameter set, a sulfide feature parameter set, a pH feature parameter set, and an oil content feature parameter set. To perform data fusion on the n first-level feature sets, specifically, the suspended solid feature parameter set, the dissolved oxygen feature parameter set, the corrosion rate feature parameter set, the total iron content feature parameter set, the sulfide feature parameter set, the pH feature parameter set, and the oil content feature parameter set are respectively subjected to data fusion to obtain p first data sets. In this embodiment, the first data set refers to multiple data sets obtained after data fusion of the suspended solid feature parameter set, the dissolved oxygen feature parameter set, the corrosion rate feature parameter set, the total iron content feature parameter set, the sulfide feature parameter set, and the pH feature parameter set. The obtained multiple data sets are used as p first data sets. After performing data fusion on the n first-level feature sets to obtain p first data sets, where p < n and p is a positive integer, that is to say, the number of first data sets is less than the number of first-level feature sets, and the number of obtained first data sets is a positive integer.

[0028] After obtaining p first data sets, m difference sets are determined from the p first data sets, where m < n and m is a positive integer. That is to say, difference sets can be determined from the first data sets, where the number of difference sets is less than the number of first data sets, and the number of difference sets is a positive integer. In this embodiment, the difference set is a set obtained by screening the first data set and having abnormal numbers and categories of feature parameters. To determine m difference sets from the p first data sets, based on a preset number of feature parameters and a preset category of feature parameters, the numbers and categories of feature parameters in all first data sets are respectively identified to obtain a subset with an abnormal number of feature parameters, a subset with an abnormal category of feature parameters, and a subset with both abnormal numbers and categories of feature parameters. In this embodiment, the preset number of feature parameters and the preset category of feature parameters can be determined according to the actual situation. The subset with an abnormal category of feature parameters is marked as the first type of abnormality, the subset with an abnormal category of feature parameters is marked as the second type of abnormality, and the subset with both abnormal numbers and categories of feature parameters is marked as the third type of abnormality. By screening the subsets with the first type of abnormality, the second type of abnormality, and the third type of abnormality, m difference sets are determined.

[0029] Data fusion is performed on the obtained difference set to obtain q second data sets, q<m, and q is a positive integer, that is, the number of obtained second data sets is less than the number of difference sets, and the number of second data sets is a positive integer. In this embodiment, the second data set is a data set obtained by data fusion of the difference set. Data fusion is performed on the obtained difference set to obtain q second data sets. Data fusion can be performed on the subsets with the first type of anomaly, the second type of anomaly, and the third type of anomaly, so as to determine q second data sets. In this embodiment, the second data set is a plurality of data sets obtained by data fusion of the subsets with the first type of anomaly, the second type of anomaly, and the third type of anomaly.

[0030] According to the second data set and the first data set, the monitoring data of the water injection sample data set is determined. Specifically, first, the first type of abnormality identification is performed on all feature parameter set subsets in the first data set to obtain the first type of abnormal subset, wherein the first type of abnormal subset includes a parameter quantity missing subset and a parameter quantity repeated subset. In this embodiment, the first type of abnormality is a subset with missing or repeated feature parameter quantities. Interpolation is performed on the subset with missing feature parameter quantities, and the repeated feature parameter quantities are screened out for the subset with repeated feature parameter quantities, thereby determining the first target data. Secondly, the subsets with missing or repeated feature parameter categories of all feature parameter set subsets in the first data set are identified to obtain the second type of abnormal subset. In this embodiment, the second type of abnormal subset includes a parameter type missing subset and a parameter type repeated subset. Interpolation is performed on the subset with missing feature parameter categories, and the repeated feature parameter categories are screened out for the subset with repeated feature parameter categories, thereby determining the second target data. Finally, the subsets of all feature parameter set subsets in the first data set whose feature parameter quantity and category are missing or repeated are identified to obtain the third type of abnormal subsets. In this embodiment, the third type of abnormal subsets include subsets with missing parameter types and quantities and subsets with repeated parameter types and quantities. The subsets with missing feature parameter categories and quantities are interpolated, and the subsets with repeated feature parameter categories and quantities are screened to remove the repeated feature parameter categories and quantities, thereby determining the third target data. The monitoring data of the water injection sample data set is determined by combining the first target data, the second target data and the third target data after interpolation or screening, that is, the monitoring data of the water injection sample data set is constituted by the first target data, the second target data and the third target data after interpolation or screening.

[0031] The monitoring data is input into the pre-trained prediction model to obtain the injection water quality data at the future time node. In this embodiment, the pre-trained prediction model can be a neural network model. The neural network model is a computing model that simulates the human nervous system. Based on the processing and learning mechanism of the human nervous system, a large number of simple processing units (neurons) are interconnected to realize the analysis and processing of the input data, and complete tasks such as classification, identification and prediction. Through the neural network model, the injection water quality data at the future time node is obtained, and based on the injection water quality data, the prediction result of the injection water quality of the offshore platform is determined. For example, the content of suspended matter obtained through the injection water quality data is high, indicating that the turbidity of seawater increases and pollutants are generated. Therefore, the content of suspended matter in seawater needs to be controlled at the future time node.

[0032] Through the above technical solution, by obtaining the water injection sample data set of the offshore platform, determining the abnormal data in the water injection sample data set and correcting the water injection sample data set, potential problems in the data can be discovered in time, the accuracy of the sample data set can be ensured, the characteristic parameter set of the corrected water injection sample data set is extracted and the characteristic parameter set is divided to obtain the first-level feature set, and the first-level feature set is data fused. By integrating the data of multiple feature sets, it can help improve the quality and consistency of the data. Determine the difference set in the first data set, fuse the difference set, determine the second data set, and further fuse the difference data, which can not only reduce the uncertainty of the difference data, but also ensure the accuracy of the monitoring data obtained through the second data set, and effectively improve the efficiency of online monitoring of water quality of offshore platform water injection.

[0033] In one implementation of this embodiment, determining abnormal data in a water injection sample data set includes: S210, obtaining a water injection sample data set within a preset period; S220, dividing the water injection sample data set to obtain multiple water injection sample data subsets; S230, generating a scatter plot according to the water injection sample data in the plurality of water injection sample data subsets; S240: Based on a preset first clustering radius, perform cluster analysis on the scatter plot to determine abnormal data.

[0034] First, obtain the water injection sample data set within a preset period. The preset period can be determined according to the actual situation. In this embodiment, the water injection sample data set includes suspended matter data, dissolved oxygen data, corrosion rate data, total iron content data, sulfide data, pH data, and oil content data. By monitoring the suspended matter, dissolved oxygen content, corrosion rate, total iron content, sulfide, pH value, and oil content in the water injection samples, the water quality of seawater can be understood and evaluated. The water injection sample data set is divided to obtain multiple water injection sample data subsets, that is, the suspended matter data set, dissolved oxygen data set, corrosion rate data set, total iron content data set, sulfide data set, pH data set, and oil content data set are divided into multiple subsets respectively, and the multiple subsets after division are used as multiple water injection sample data subsets. In this embodiment, the water injection sample data subsets are subsets obtained after dividing the water injection sample data set. For example, the suspended matter sample data set, dissolved oxygen data set, corrosion rate data set, total iron content data set, sulfide data set, pH data set, and oil content data set obtained within a preset period are divided into multiple data subsets with equal time periods within a preset period according to an equal division method of time periods. The division of the dissolved oxygen data set, corrosion rate data set, total iron content data set, sulfide data set, pH data set, and oil content data set are similarly divided into multiple data subsets with equal time periods.

[0035] After obtaining multiple water injection sample data subsets, a scatter plot is generated based on the water injection sample data in the multiple water injection sample data sets. A scatter plot is a graph obtained by plotting points in a rectangular coordinate system using two corresponding variable values ​​as the horizontal and vertical coordinates of a point in the graph, and is used to observe the degree and type of correlation between two variables. A scatter plot is generated from the obtained water injection sample data subsets. For example, the suspended matter data set is divided into multiple suspended matter data set subsets, and a scatter plot is generated based on the suspended matter data in the multiple suspended matter data subsets.

[0036] In one of the embodiments, a cluster analysis can be performed on the scatter plot based on a preset first clustering radius to determine abnormal data. In this embodiment, the preset first clustering radius can be determined according to actual conditions. Specifically, the clustering radius of the scatter plot of each subset can be first calculated in the water injection sample data set, and the data with a clustering radius greater than the preset first clustering radius can be regarded as abnormal data.

[0037] The scatter plot is clustered by using the preset first cluster radius. Specifically, the cluster centers on the scatter plot are calculated, each cluster center corresponds to a data point, and circles are drawn with the cluster center as the center and the cluster radius as the radius. Then, multiple circular ranges will appear in the scatter plot. If there is a data in the scatter plot that is not within any circular range, the data is determined to be abnormal data. For multiple water injection sample data subsets, the above cluster analysis can be performed on the scatter plots generated by each water injection sample data subset to determine and summarize the abnormal data on each scatter plot.

[0038] In another embodiment, the first cluster radius can be obtained by calculation. For each cluster center, the average distance (or maximum distance) from all points in the cluster to the cluster center can be calculated. This distance can be used as the radius of the cluster, that is, the first cluster radius.

[0039] By generating a scatter plot for a subset of water injection samples and performing cluster analysis on the scatter plot, the relationship between the water injection sample data can be intuitively displayed, which helps to find abnormal data in the water injection samples in a timely and accurate manner and ensure the accuracy of water injection sample monitoring.

[0040] In one implementation of this embodiment, the water injection sample data set is corrected according to the abnormal data to obtain a corrected water injection sample data set, including: S310, dividing the water injection sample data set into a first data set and a second data set, wherein the overlapping points of the first data set and the second data set are abnormal data; S320, calculating the average value of every two adjacent data in the first data set to obtain a first correction data set; S330, calculating the average value of every two adjacent data in the second data set to obtain a second correction data set; S340, inputting the first correction data set and the second correction data set into the Kalman filter equation respectively to obtain a first correction result and a second correction result; S350, taking the average of the first correction result and the second correction result as correction data for abnormal data; S360: Correct the water injection sample data set in combination with the calibration data to determine the first monitoring data of the water injection sample.

[0041] The water injection sample data set is divided into a first data set and a second data set, and the overlap point of the first data set and the second data set is the abnormal data, that is, in this embodiment, the first data set refers to the data set before the abnormal data and adjacent to the abnormal data, and the second data set refers to the data set after the abnormal data and adjacent to the abnormal data. The water injection sample data set is divided into a first data set and a second data set, and the midpoint of the first data set and the second data set is the abnormal data. For example, to determine the location of the abnormal data in the water injection sample data set, the first 30 data including the abnormal data can be used as the first data set, and the last 30 data including the abnormal data can be used as the second data set.

[0042] Calculate the average value of every two adjacent data in the first data set to obtain a first corrected data set. In this embodiment, the first corrected data set refers to a data set formed by calculating the average value of every two adjacent data in the first data set, that is, by calculating the average value of every two adjacent data in the data set before the abnormal data as the first corrected data set. For example, the data in the first data set are 2, 5, 3, and 4, respectively. Calculate the average value of 2 and 5, the average value of 5 and 3, and the average value of 3 and 4, respectively, and use the calculated average values ​​to form a data set as the first corrected data set.

[0043] Calculate the average value of every two adjacent data in the second data set to obtain a second corrected data set. In this embodiment, the second corrected data set refers to a data set formed by calculating the average value of every two adjacent data in the second data set. That is to say, the set of average values ​​of every two adjacent data in the second data set is used as the second corrected data set. For example, the data in the second data set are 6, 8, 8, and 1, respectively. The average values ​​of 6 and 8, 8 and 8, and 8 and 1 are calculated respectively, and the data set formed by the calculated average values ​​is used as the second corrected data set.

[0044] The first correction data set and the second correction data set are respectively input into the Kalman filter equation to obtain the first correction result and the second correction result. The Kalman filter equation is an algorithm that uses the linear system state equation to optimally estimate the system state by inputting and outputting observation data of the system. In this embodiment, the first correction data set is input into the Kalman filter equation and the output data set is used as the first correction result, and the second correction data set is input into the Kalman filter equation and the output data set is used as the second correction result. The first correction data set and the second correction data set are respectively input into the Kalman filter equation to reduce the noise and interference in the data set, and the first correction result and the second correction result are output.

[0045] After obtaining the first correction result and the second correction result, the average value of the first correction result and the second correction result is used as the correction data of the abnormal data. For example, the data in the first correction result are [2, 6, 7, 5], and the data in the second correction result are [9, 5, 7, 1]. The average values ​​of 2 in the first correction result and 9 in the second correction result are calculated respectively. Similarly, the average values ​​of 6 and 5, 7 and 7, and 5 and 1 are calculated respectively, and the calculated data sets are used as correction data.

[0046] In combination with the correction data, the water injection sample data set is corrected to determine the first monitoring data of the water injection sample. In this embodiment, the first monitoring data refers to the water injection sample data set corrected by the correction data. The sample data is corrected by the correction data, the difference between the abnormal data and the water injection sample data set is calculated, and the water injection sample data set is corrected based on the calculated difference to determine the first monitoring data of the water injection sample.

[0047] By identifying abnormal data and correcting the water injection sample data set, the reliability of the data can be improved, the accuracy of data monitoring can be guaranteed, and it will help to better monitor the water quality of offshore water injection platforms, ensure the safety of water injection quality, promote marine biodiversity, and maintain the balance and stability of the ecosystem.

[0048] In one implementation of this embodiment, the feature parameter set is divided to obtain n first-level feature sets, including: S510: Divide all feature parameter sets into a plurality of first-level feature sets according to corresponding feature parameter categories, wherein the number of feature parameters in each first-level feature set is the same, and the number of first-level feature sets is n.

[0049] All feature parameter sets are divided into multiple first-level feature sets according to corresponding feature parameter categories. In this embodiment, the first-level feature set refers to a feature parameter subset obtained by dividing the feature parameter set according to the category of the feature parameter. The feature parameter set includes a suspended matter feature parameter set, a dissolved oxygen feature parameter set, a corrosion rate feature parameter set, a total iron content feature parameter set, a sulfide feature parameter set, a pH feature parameter set, and an oil content feature parameter set. The feature parameter categories are turbidity characteristics, chromaticity characteristics, pH characteristics, oxygen demand characteristics, conductivity characteristics, and dissolved oxygen content characteristics. Divide all feature parameter sets into multiple first-level feature sets according to the corresponding feature parameter categories, that is, divide the suspended matter feature parameter set according to the turbidity feature, divide the dissolved oxygen feature parameter set according to the dissolved oxygen content feature, divide the corrosion rate feature parameter set according to the conductivity feature, divide the total iron content feature parameter set according to the conductivity feature, divide the sulfide feature parameter set according to the oxygen demand feature, divide the pH feature parameter set according to the acidity feature, and divide the oil content feature parameter set according to the chromaticity feature, and respectively obtain the suspended matter feature parameter subset, dissolved oxygen feature parameter subset, corrosion rate feature parameter set subset, total iron content feature parameter set subset, sulfide feature parameter set subset, pH feature parameter set subset, and oil content feature parameter subset. Summarize the suspended matter feature parameter subset, dissolved oxygen feature parameter subset, corrosion rate feature parameter set subset, total iron content feature parameter set subset, sulfide feature parameter set subset, pH feature parameter set subset, and oil content feature parameter subset to obtain multiple first-level feature sets.

[0050] All feature parameter sets are divided into multiple first-level feature sets according to the corresponding feature parameter categories to obtain n first-level feature sets. By dividing them according to different feature parameter categories, it is helpful to better analyze different feature parameter sets and ensure the accuracy of water quality prediction and analysis for offshore platform injection.

[0051] In one implementation of this embodiment, the feature parameter categories include turbidity features, chromaticity features, pH features, oxygen demand features, conductivity features, and dissolved oxygen content features, and all feature parameter sets are divided into a plurality of first-level feature sets according to the corresponding feature parameter categories, including: S610, dividing the suspended matter characteristic parameter set according to turbidity characteristics to obtain multiple suspended matter characteristic parameter subsets; S620, dividing the dissolved oxygen characteristic parameter set according to the dissolved oxygen content characteristics to obtain multiple dissolved oxygen characteristic parameter subsets; S630, dividing the corrosion rate characteristic parameter set according to the conductivity characteristics to obtain multiple corrosion rate characteristic parameter set subsets; S640, dividing the total iron content characteristic parameter set according to the conductivity characteristics to obtain multiple total iron content characteristic parameter set subsets; S650, dividing the sulfide characteristic parameter set according to the oxygen demand characteristics to obtain multiple sulfide characteristic parameter set subsets; S660, dividing the pH characteristic parameter set according to the acidity and alkalinity characteristics to obtain multiple pH characteristic parameter set subsets; S670, dividing the oil content characteristic parameter set according to the chromaticity feature to obtain a plurality of oil content characteristic parameter subsets; S680, summarizing the suspended matter characteristic parameter subset, dissolved oxygen characteristic parameter subset, total iron content characteristic parameter subset, sulfide characteristic parameter subset, pH characteristic parameter subset, and oil content characteristic parameter subset to obtain multiple characteristic parameter subsets.

[0052] The suspended matter characteristic parameter set is divided according to the turbidity characteristics to obtain multiple suspended matter characteristic parameter subsets. In this embodiment, the turbidity characteristics are mainly reflected in the degree of obstruction to the passage of light, including the scattering of light by suspended matter and the absorption of light by solute molecules. The suspended matter characteristic parameter set is divided according to the turbidity characteristics to obtain multiple suspended matter characteristic parameter subsets. In other words, the suspended matter characteristic parameter set is divided according to different turbidity characteristics to obtain multiple suspended matter characteristic parameter subsets. For example, the suspended matter characteristic parameter set is divided into multiple characteristic parameter subsets by the preset upper and lower limits of the turbidity characteristic parameters. In this embodiment, the preset upper and lower limits of the turbidity characteristic parameters can be determined according to actual conditions.

[0053] The dissolved oxygen characteristic parameter set is divided according to the dissolved oxygen content characteristics to obtain multiple dissolved oxygen characteristic parameter subsets, that is, the dissolved oxygen characteristic parameter set is divided by the dissolved oxygen content characteristics in the water injection platform. In this embodiment, the dissolved oxygen content characteristics refer to molecular oxygen in the air dissolved in water. The dissolved oxygen characteristic parameter set is divided according to the dissolved oxygen content characteristics to obtain multiple dissolved oxygen characteristic parameter subsets. For example, the dissolved oxygen characteristic parameter set is divided into multiple characteristic parameter subsets by the preset upper and lower limits of the dissolved oxygen content characteristic parameters to obtain multiple dissolved oxygen characteristic parameter subsets. In this embodiment, the preset upper and lower limits of the dissolved oxygen content characteristic parameters can be determined according to actual conditions.

[0054] The corrosion rate characteristic parameter set is divided according to the conductivity characteristic to obtain multiple corrosion rate characteristic parameter set subsets, that is, the change of the corrosion rate characteristic parameter is analyzed by obtaining the conductivity characteristic parameter. In this embodiment, the conductivity characteristic refers to the conductivity of seawater. The corrosion rate characteristic parameter set is divided according to the conductivity characteristic to obtain multiple corrosion rate characteristic parameter set subsets. For example, the corrosion rate characteristic parameter set is divided into multiple characteristic parameter subsets by the preset upper and lower limits of the conductivity characteristic parameter to obtain multiple conductivity characteristic parameter subsets. In this embodiment, the preset upper and lower limits of the conductivity characteristic parameter can be determined according to actual conditions.

[0055] The total iron content characteristic parameter set is divided according to the conductivity characteristic to obtain a plurality of total iron content characteristic parameter set subsets, that is, the total iron content characteristic parameter set is divided by the conductivity characteristic in the water injection platform, and in this embodiment, the conductivity characteristic refers to the conductivity of seawater. By monitoring the conductivity in seawater, the change in the total iron content in seawater can be judged. The total iron content characteristic parameter set is divided according to the conductivity characteristic to obtain a plurality of total iron content characteristic parameter set subsets, for example, by the preset upper and lower limits of the conductivity characteristic parameter, the total iron content characteristic parameter set is divided into a plurality of characteristic parameter subsets to obtain a plurality of conductivity characteristic parameter subsets, and in this embodiment, the preset upper and lower limits of the conductivity characteristic parameter can be determined according to actual conditions.

[0056] The sulfide characteristic parameter set is divided according to the oxygen demand characteristic to obtain multiple sulfide characteristic parameter set subsets. That is to say, the sulfide characteristic parameter set is divided according to the oxygen demand characteristic in the water injection platform. When the sulfide in the seawater increases, the oxygen demand will increase, because the sulfide will be oxidized by microorganisms and consume a large amount of oxygen. Therefore, the sulfide characteristic parameter set is divided according to the oxygen demand characteristic. In this embodiment, the oxygen demand characteristic refers to the amount of oxygen consumed in the process of decomposing oxygen-demanding organic matter into simple and stable inorganic substances by microorganisms. The sulfide characteristic parameter set is divided according to the oxygen demand characteristic to obtain multiple sulfide characteristic parameter set subsets. For example, the sulfide characteristic parameter set is divided into multiple characteristic parameter subsets by the preset upper and lower limits of the oxygen demand characteristic to obtain multiple sulfide characteristic parameter set subsets. In this embodiment, the preset upper and lower limits of the oxygen demand characteristic can be determined according to actual conditions. ‌ The pH characteristic parameter set is divided according to the acidity and alkalinity characteristics to obtain multiple pH characteristic parameter set subsets, that is, the pH characteristic parameter set is divided according to the acidity and alkalinity characteristics. In this embodiment, the acidity and alkalinity characteristics refer to the acidity and alkalinity of the seawater solution. The pH characteristic parameter set is divided according to the acidity and alkalinity characteristics to obtain multiple pH characteristic parameter set subsets. For example, the pH characteristic parameter set is divided into multiple characteristic parameter subsets by the preset upper and lower limits of the acidity and alkalinity characteristic parameters to obtain multiple pH characteristic parameter set subsets. In this embodiment, the preset upper and lower limits of the oxygen demand characteristic parameters can be determined according to actual conditions.

[0057] The oil content characteristic parameter set is divided according to the chromaticity feature to obtain multiple oil content characteristic parameter subsets, that is, the oil content characteristic parameter set is divided according to the chromaticity feature. In this embodiment, the chromaticity feature refers to the diversity and variability of the color of the sea surface. By analyzing the change of the sea surface color, the oil content in the seawater is judged, and the oil content characteristic parameter set is divided according to the chromaticity feature to obtain multiple oil content characteristic parameter subsets. For example, the oil content characteristic parameter set is divided into multiple characteristic parameter subsets according to the preset upper and lower limits of the chromaticity feature to obtain multiple oil content characteristic parameter subsets. In this embodiment, the preset upper and lower limits of the chromaticity feature can be determined according to actual conditions.

[0058] The suspended matter characteristic parameter subset, dissolved oxygen characteristic parameter subset, total iron content characteristic parameter subset, sulfide characteristic parameter subset, pH characteristic parameter subset, and oil content characteristic parameter subset are summarized to obtain multiple characteristic parameter subsets. That is to say, the suspended matter characteristic parameter subset, dissolved oxygen characteristic parameter subset, total iron content characteristic parameter subset, sulfide characteristic parameter subset, pH characteristic parameter subset, and oil content characteristic parameter subset are combined to obtain multiple characteristic parameter subsets to form multiple characteristic parameter subsets.

[0059] By classifying feature parameter sets according to feature categories, different data sets can be analyzed accurately, which can not only improve monitoring efficiency, but also timely discover abnormal information during the monitoring process, saving a lot of time and labor costs.

[0060] In one implementation of this embodiment, screening the first data set to determine m difference sets of the first data set includes: S710, obtaining the quantity and characteristic parameter categories of all characteristic parameters in all first data sets; S720, based on a preset number of characteristic parameters and a preset category of characteristic parameters, respectively identify the number of characteristic parameters and the category of characteristic parameters in all first data sets to obtain a subset with abnormal number of characteristic parameters, a subset with abnormal category of characteristic parameters, and a subset with abnormal number and category of characteristic parameters; S730, marking a subset of feature parameter quantity abnormalities, a subset of feature parameter category abnormalities, a subset of feature parameter category abnormalities, and a subset of feature parameter quantity and category abnormalities to obtain first-category abnormalities, second-category abnormalities, and third-category abnormalities; S740 , screening feature parameter subsets with first-category anomalies, second-category anomalies, and third-category anomalies to determine m difference sets of the first data set.

[0061] First, the number and category of all characteristic parameters in all first data sets are obtained. In this embodiment, the characteristic parameter categories include turbidity characteristics, chromaticity characteristics, pH characteristics, oxygen demand characteristics, conductivity characteristics and dissolved oxygen content characteristics. The number and category of characteristic parameters in the first data set are determined.

[0062] Based on the preset number of characteristic parameters and the preset category of characteristic parameters, in this embodiment, the preset number of characteristic parameters and the preset category of characteristic parameters can be determined according to actual conditions, and the number of characteristic parameters and the category of characteristic parameters in all first data sets are identified respectively to obtain a subset with abnormal number of characteristic parameters, a subset with abnormal category of characteristic parameters, and a subset with abnormal number and category of characteristic parameters. That is to say, in the first data set, a subset with abnormal number of characteristic parameters, a subset with abnormal category of characteristic parameters, and a subset with abnormal number and category of characteristic parameters are screened out. For example, the preset number of characteristic parameters is 4, and a subset with characteristic parameters other than 4 is screened out in the first data set. The preset number of characteristic parameters is 8, and a subset with characteristic parameter categories other than 8 is screened out.

[0063] The subsets of abnormal number of characteristic parameters, the subsets of abnormal category of characteristic parameters, and the subsets of abnormal number and category of characteristic parameters are marked to obtain the first type of abnormality, the second type of abnormality, and the third type of abnormality. In this embodiment, the first type of abnormality is the subset of abnormal number of characteristic parameters, the first type of abnormality is the subset of characteristic parameter category, and the third type of abnormality is the subset of abnormal number and category of characteristic parameters. The first type of abnormality, the second type of abnormality, and the third type of abnormality can be identified by cluster analysis. Cluster analysis refers to dividing a data set into multiple groups according to the similarity between samples, so that the samples in the same subset have some similar attributes. Through cluster analysis, the subsets with the first type of abnormality, the second type of abnormality, and the third type of abnormality are screened out, and these subsets are marked.

[0064] The feature parameter subsets with the first type of anomalies, the second type of anomalies and the third type of anomalies are screened to determine m difference sets of the first data set. In this embodiment, the difference set refers to the feature parameter subsets with the first type of anomalies, the second type of anomalies and the third type of anomalies. That is to say, the subsets with the first type of anomalies, the second type of anomalies and the third type of anomalies are screened, and the subsets obtained by the screening are determined as the difference sets.

[0065] By screening the difference set in the first data set, the accuracy of the data in the first data set can be guaranteed, which helps to ensure the accuracy of the prediction and analysis of water quality for offshore platform water injection.

[0066] In one implementation of this embodiment, determining monitoring data of the water injection sample data set according to the second data set and the first data set includes: S801, classifying the second data set to determine the types and quantities of all feature parameters in the second data set; S802, performing first-category anomaly identification on all feature parameter set subsets in the first data set to obtain first-category anomaly subsets, where the first-category anomaly subsets include parameter quantity missing subsets and parameter quantity duplication subsets; S803, identifying the number of characteristic parameters of the missing parameter subset, and obtaining the number of missing parameters; S804, based on the number of missing parameters and the number of all characteristic parameters in the second data set, interpolate the missing subset of parameters to obtain first interpolated data; S805, identifying the number of characteristic parameters of the parameter quantity repetition subset, and obtaining the number of repetition parameters; S806, based on the number of repeated parameters, similarity calculation is performed on the repeated subset of the number of parameters to determine the repeated parameters, and the repeated parameters are screened out to obtain first screened-out data; S807, determining first target data based on the first interpolation data and the first screened data; S808, performing second-category anomaly identification on all feature parameter set subsets in the first data set to obtain second-category anomaly subsets, where the second-category anomaly subsets include parameter type missing subsets and parameter type duplication subsets; S809, obtaining the locations of the missing subset of parameter categories and the repeated subset of parameter categories; S810, identifying the characteristic parameter category of the missing parameter category subset, and obtaining the category of the missing parameter; S811, based on the categories of the missing parameters and the categories of all characteristic parameters in the second data set, interpolate the missing subset of parameter categories to obtain second interpolated data; S812, identifying the characteristic parameter category of the repeated subset of parameter categories, and obtaining the category of the repeated parameters; S813, based on the categories of repeated parameters, similarity calculation is performed on repeated subsets of parameter categories to determine repeated categories, and the repeated categories are screened out to obtain second screened-out data; S814, combining the second interpolation data and the second screened data to determine the second target data; S815, performing third-category anomaly identification on all feature parameter set subsets in the first data set to obtain third-category anomaly subsets, where the third-category anomaly subsets include a subset where both parameter type and quantity are missing and a subset where both parameter type and quantity are missing; S816, identifying the type and number of characteristic parameters of the missing subset in terms of parameter type and quantity, and obtaining the type and number of the missing parameters; S817, based on the types and quantities of the missing parameters and the types and quantities of all characteristic parameters in the second data set, interpolate the missing subsets of parameter types and quantities to obtain third interpolated data; S818, identifying the characteristic parameter types and quantities of the subsets in which the parameter types and quantities are repeated, and obtaining the types and quantities of the repeated parameters; S819, based on the type and quantity of repeated parameters, similarity calculation is performed on the parameter type and quantity repeated subsets to determine the repeated type and quantity, and the repeated type and quantity are screened out to obtain third screened-out data; S820, combining the third interpolation data and the third screened data to determine third target data; S821. Determine monitoring data of the water injection sample data set by combining the first target data, the second target data, and the third target data.

[0067] First, the second data set is classified to determine the type and quantity of all characteristic parameters in the second data set, that is, the second data set is obtained by data fusion of the difference set, and the type and quantity of all characteristic parameters in the second data set are confirmed to obtain the category and quantity of all characteristic parameters in the second data set. Data fusion of the difference set refers to splitting the difference data and recombining the difference data for data fusion to obtain the second data set. Secondly, after determining the type and quantity of all characteristic parameters in the second data set, the first type of abnormal identification is performed on all characteristic parameter set subsets in the first data set to obtain the first type of abnormal subset, wherein the first type of abnormal subset includes a parameter number missing subset and a parameter number repeated subset. In this embodiment, the parameter number missing subset refers to a subset in which the number of characteristic parameters is less than the preset number of characteristic parameters, and the parameter number repeated subset refers to a subset in which the number of characteristic parameters is greater than the preset number of characteristic parameters. The number of characteristic parameters is identified for the subset in which the number of parameters is repeated, the number of characteristic parameters repeated in the subset in which the number of parameters is repeated is determined, similarity calculation is performed on the parameter number repeated subset, repeated parameters are determined, and repeated parameters are screened out to obtain the first screened data.

[0068] The similarity can be obtained by the cosine similarity, which can measure the similarity between two vectors by calculating the cosine value of the angle between them. Interpolate the subset with missing feature parameters, confirm the number of missing parameters in the subset with missing feature parameters, interpolate the feature parameters of the corresponding category in the second data set, and obtain the first interpolation data. In this embodiment, the first interpolation data refers to the feature parameter data obtained after interpolating the data subset with missing feature parameters. The first screened data refers to the feature parameter data obtained after screening the data subset with repeated feature parameters. Combine the first interpolation data and the first screened data to determine the first target data, that is, after interpolating or screening the data subset with missing or repeated feature parameters in the first data set, the first target data can be determined.

[0069] After determining the first target data, the second type of abnormality identification is performed on all feature parameter set subsets in the first data set to obtain the second type of abnormal subsets, wherein the second type of abnormal subsets include parameter category missing subsets and parameter category repeated subsets. In this embodiment, the parameter category missing subset refers to a subset in which the category of feature parameters is less than the preset feature parameter category, and the parameter category repeated subset refers to a subset in which the category of feature parameters is greater than the preset feature parameter category. The feature parameter category is identified for the subset in which the parameter category is repeated, the category of repeated feature parameters in the subset in which the parameter category is repeated is determined, the similarity calculation is performed on the parameter category repeated subset, the repeated category is determined, and the repeated parameters are screened out to obtain the second screened data. The similarity calculation can be performed by ‌‌ cosine similarity, and the cosine similarity measures their similarity by calculating the cosine value of the angle between two vectors. The subset in which the feature parameter category is missing is interpolated, the missing parameter type in the subset in which the feature parameter category is missing is confirmed, and the feature category of the corresponding category in the second data set is interpolated to obtain the second interpolated data. In this embodiment, the second interpolation data refers to the characteristic parameter data obtained after interpolation of the data subset with missing characteristic parameter categories. The second screened data refers to the characteristic parameter data obtained after screened data of the data subset with repeated characteristic parameter categories. The second target data is determined by combining the second interpolation data and the second screened data, that is, the second target data is determined after interpolation or screening of the data subset with missing or repeated characteristic parameter categories in the second data set.

[0070] After determining the second target data, perform third-category anomaly identification on all feature parameter set subsets in the first data set to obtain third-category anomaly subsets, wherein the third-category anomaly subsets include subsets with missing parameter categories and quantities and subsets with repeated parameter categories and quantities. In this embodiment, the missing parameter category and quantity subset refers to a subset in which the number of feature parameter categories is less than the preset number of feature parameter categories, and the repeated parameter category subset refers to a subset in which the number of feature parameter categories is greater than the preset number of feature parameter categories and quantities. Identify the feature parameter categories and quantities of the subsets in which the parameter categories and quantities are repeated, determine the categories and quantities of repeated feature parameters in the subsets in which the parameter categories and quantities are repeated, perform similarity calculation on the subsets in which the parameter categories and quantities are repeated, determine the repeated categories and quantities, and screen out the repeated parameters to obtain third screened-out data. The similarity calculation can be performed by ‌‌cosine similarity, which measures the similarity of two vectors by calculating the cosine value of the angle between them. Interpolate the subset where both the feature parameter category and quantity are missing, confirm the missing parameter type and quantity in the subset where both the feature parameter category and quantity are missing, interpolate the feature category and quantity of the corresponding category in the second data set, and obtain third interpolation data. In this embodiment, the third interpolation data refers to the feature parameter data obtained after interpolating the data subset where both the feature parameter category and quantity are missing. The third screened data refers to the feature parameter data obtained after screening the data subset where both the feature parameter category and quantity are repeated. Combine the third interpolation data and the third screened data to determine the third target data, that is, after interpolating or screening the data subset where the feature parameter category is missing or repeated in the third data set, the third target data is determined.

[0071] Figure 2 The schematic diagram shows a structure diagram of a multi-parameter intelligent online monitoring method for water quality of offshore platform injection, such as Figure 2 The characteristic parameter set of the water injection sample data set is divided to obtain a first-level characteristic set, and the first-level characteristic set is subjected to data fusion to obtain a first data set. The first data set is screened to determine a difference set, and the difference set is subjected to data fusion to obtain a second data set. The characteristic parameters in the second data set are interpolated to the first data set and repeated characteristic parameters in the first data set are screened to determine the monitoring data of the water injection sample data set.

[0072] The present application provides a machine-readable storage medium having instructions stored thereon, the instructions being used to enable a machine to execute the above-mentioned multi-parameter intelligent online monitoring method for water quality of offshore platform injection.

[0073] The present application provides an offshore platform water injection water quality multi-parameter intelligent online monitoring system, comprising: a memory configured to store instructions; and The processor is configured to call the instructions from the memory and implement the above-mentioned multi-parameter intelligent online monitoring method for water quality of offshore platform water injection when executing the instructions.

[0074] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0075] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0076] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0077] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0078] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0079] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0080] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0081] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0082] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A multi-parameter intelligent online monitoring method for water quality of offshore platform injection, characterized in that: Including: Obtain a water injection sample data set of an offshore platform; Determine abnormal data in the water injection sample data set; Correct the water injection sample data set according to the abnormal data to obtain a corrected water injection sample data set; Extract a set of characteristic parameters of the corrected water injection sample data set; Divide the set of characteristic parameters to obtain n first-level characteristic sets, where n≥2 and n is an integer; Perform data fusion on the n first-level characteristic sets to obtain p first data sets, where p < n and p is a positive integer; Determine m difference sets in the p first data sets, where m < p and m is a positive integer; Perform data fusion on the m difference sets to obtain q second data sets, where q < m and q is a positive integer; Determine the monitoring data of the water injection sample data set according to the second data set and the first data set; Input the monitoring data into a pre-trained prediction model to obtain water injection quality data at future time nodes; Based on the water injection quality data, determine the prediction result of the water injection quality of the offshore platform.

2. The method according to claim 1, characterized in that The determining abnormal data in the water injection sample data set includes: Obtain the water injection sample data set within a preset period; Divide the water injection sample data set to obtain multiple water injection sample data subsets; Generate a scatter plot based on the water injection sample data in the multiple water injection sample data subsets; Perform clustering analysis on the scatter plot based on a preset first clustering radius to determine abnormal data.

3. The method according to claim 1, characterized in that The correcting the water injection sample data set according to the abnormal data to obtain a corrected water injection sample data set includes: Divide the water injection sample data set into a first data set and a second data set, and the coincidence point of the first data set and the second data set is the abnormal data; Calculate the average value of every two adjacent data in the first data set to obtain a first corrected data set; Calculate the average value of every two adjacent data in the second data set to obtain a second corrected data set; Input the first corrected data set and the second corrected data set into a Kalman filter equation respectively to obtain a first correction result and a second correction result; Take the average value of the first correction result and the second correction result as the correction data of the abnormal data; Combine the correction data to correct the water injection sample data set to obtain a corrected water injection sample data set.

4. The method according to claim 1, characterized in that: The set of characteristic parameters includes a suspended solid characteristic parameter set, a dissolved oxygen characteristic parameter set, a corrosion rate characteristic parameter set, a total iron content characteristic parameter set, a sulfide characteristic parameter set, a pH characteristic parameter set, and an oil content characteristic parameter set.

5. The method according to claim 4, characterized in that The dividing the set of characteristic parameters to obtain n first-level characteristic sets includes: Divide all the characteristic parameter sets into multiple first-level characteristic sets according to the corresponding characteristic parameter categories, where the number of characteristic parameters in each first-level characteristic set is the same, and the number of the first-level characteristic sets is n.

6. The method according to claim 5, characterized in that The characteristic parameter categories include turbidity characteristics, chromaticity characteristics, pH characteristics, oxygen demand characteristics, conductivity characteristics and dissolved oxygen content characteristics. All the characteristic parameter sets are divided into a plurality of first-level characteristic sets according to the corresponding characteristic parameter categories, including: Dividing the suspended matter characteristic parameter set according to the turbidity characteristics to obtain a plurality of suspended matter characteristic parameter subsets; Dividing the dissolved oxygen characteristic parameter set according to the dissolved oxygen content characteristics to obtain a plurality of dissolved oxygen characteristic parameter subsets; Dividing the corrosion rate characteristic parameter set according to the conductivity characteristics to obtain a plurality of corrosion rate characteristic parameter set subsets; Dividing the total iron content characteristic parameter set according to the conductivity characteristics to obtain a plurality of total iron content characteristic parameter set subsets; Dividing the sulfide characteristic parameter set according to the oxygen demand characteristics to obtain a plurality of sulfide characteristic parameter set subsets; Dividing the pH characteristic parameter set according to the acidity and alkalinity characteristics to obtain a plurality of pH characteristic parameter set subsets; Dividing the oil content characteristic parameter set according to the chromaticity feature to obtain a plurality of oil content characteristic parameter subsets; The suspended matter characteristic parameter subset, the dissolved oxygen characteristic parameter subset, the total iron content characteristic parameter subset, the sulfide characteristic parameter subset, the pH characteristic parameter subset, and the oil content characteristic parameter subset are summarized to obtain a plurality of characteristic parameter subsets.

7. The method according to claim 1, characterized in that The screening of the first data set to determine m difference sets of the first data set includes: Acquire the quantity and category of all the feature parameters in all the first data sets; Based on a preset number of characteristic parameters and a preset category of characteristic parameters, the number of characteristic parameters and the category of characteristic parameters in all the first data sets are identified respectively to obtain a subset of the number of characteristic parameters being abnormal, a subset of the category of characteristic parameters being abnormal, and a subset of the number and category of characteristic parameters being abnormal; Marking a subset of the characteristic parameter quantity abnormality, a subset of the characteristic parameter category abnormality, a subset of the characteristic parameter category abnormality, and a subset of the characteristic parameter quantity and category abnormality to obtain a first type of abnormality, a second type of abnormality, and a third type of abnormality; The feature parameter subsets having the first type of anomalies, the second type of anomalies, and the third type of anomalies are screened to determine m difference sets of the first data set.

8. The method according to claim 7, characterized in that Determining the monitoring data of the water injection sample data set according to the second data set and the first data set includes: Classifying the second data set to determine the types and quantities of all characteristic parameters in the second data set; Performing first-category anomaly identification on all the feature parameter set subsets in the first data set to obtain first-category anomaly subsets, wherein the first-category anomaly subsets include parameter quantity missing subsets and parameter quantity duplication subsets; Identifying the number of characteristic parameters of the parameter number missing subset to obtain the number of missing parameters; Based on the number of the missing parameters and the number of all characteristic parameters in the second data set, interpolating the missing subset of the number of parameters to obtain first interpolated data; Identifying the number of characteristic parameters of the parameter number repetition subset to obtain the number of repetition parameters; Based on the number of repeated parameters, similarity calculation is performed on the repeated subset of the number of parameters to determine the repeated parameters, and the repeated parameters are screened out to obtain first screened-out data; Determine first target data based on the first interpolation data and the first screened data; Performing second-category anomaly identification on all the feature parameter set subsets in the first data set to obtain second-category anomaly subsets, where the second-category anomaly subsets include parameter type missing subsets and parameter type repeated subsets; Obtaining the locations of the missing subset of the parameter category and the repeated subset of the parameter category; Identifying the characteristic parameter category of the parameter category missing subset to obtain the category of the missing parameter; Based on the category of the missing parameter and the categories of all characteristic parameters in the second data set, interpolating the missing subset of the parameter category to obtain second interpolated data; Identifying the characteristic parameter category of the repeated subset of the parameter category to obtain the category of the repeated parameter; Based on the category of the repeated parameter, similarity calculation is performed on the repeated subset of the parameter category to determine the repeated category, and the repeated category is screened out to obtain second screened data; Determine second target data by combining the second interpolation data and the second screened data; Performing third-category anomaly identification on all the feature parameter set subsets in the first data set to obtain third-category anomaly subsets, wherein the third-category anomaly subsets include a subset in which both parameter type and quantity are missing and a subset in which both parameter type and quantity are missing; Identify the type and quantity of the characteristic parameters of the missing subset of the parameter type and quantity, and obtain the type and quantity of the missing parameters; Based on the type and quantity of the missing parameters and the type and quantity of all characteristic parameters in the second data set, interpolating the missing subset of parameter types and quantities to obtain third interpolated data; Identify the characteristic parameter types and quantities of the repeated subset of the parameter types and quantities, and obtain the types and quantities of repeated parameters; Based on the type and quantity of the repeated parameters, similarity calculation is performed on the repeated subsets of the parameter type and quantity to determine the repeated type and quantity, and the repeated type and quantity are screened out to obtain third screened-out data; Determine third target data by combining the third interpolation data and the third screened data; The monitoring data of the water injection sample data set is determined in combination with the first target data, the second target data and the third target data.

9. A machine-readable storage medium, characterized in that: The machine-readable storage medium stores instructions for enabling a machine to execute a multi-parameter intelligent online monitoring method for water quality of offshore platform injection water according to any one of claims 1 to 8.

10. An intelligent online multi-parameter monitoring system for water quality of offshore platform injection, characterized in that: include: a memory configured to store instructions; as well as A processor is configured to call the instruction from the memory and to implement a multi-parameter intelligent online monitoring method for water quality of offshore platform injection water according to any one of claims 1 to 8 when executing the instruction.

Images