Oil and gas valve safety control and monitoring system
Through the data acquisition, abnormal detection and correlation analysis module of the oil and gas valve safety control and monitoring system, the inaccuracy of the evaluation of load pressure in the oil and gas pipeline is solved, and in-depth evaluation and early warning of rust and corrosion are achieved, and the safety of oil and gas valves is improved. Control ability.
Patent Information
- Application Number
- CN202410779772.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-06-17
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2044-06-17
AI Technical Summary
In the prior art, there are few researches on the load pressure load of oil and gas pipelines, and the abnormal evaluation method of a single parameter is inaccurate, so it is impossible to effectively predict and detect load pressure load pressure accidents caused by rust and corrosion, which affects the safety control of oil and gas valves.
A safety control and monitoring system for oil and gas valves is designed, including data acquisition module, abnormality detection module, association analysis module and evaluation module. By obtaining corrosion monitoring parameters in multiple evaluation dimensions, abnormality detection and association analysis are performed, and abnormality data is calculated. Associate relationships and credibility, and input them into the exception evaluation function to generate exception state scores to evaluate the severity of pipeline corrosion.
The system can more accurately identify the corrosion status of the pipeline, provide quantitative data support, reduce false alarms and missed reports, improve the safety control and early warning capabilities of oil and gas valves, and ensure the safe operation of the pipeline.
Smart Images

Figure CN118568635B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of oil and gas pipeline data evaluation, and in particular to an oil and gas valve safety control and monitoring system. Background Art
[0002] The study found that there are many risk factors in the transportation of oil and gas pipelines, and oil and gas pipeline managers have taken a lot of measures to control them. After evaluating the risks, quantifying the risks will help analyze the pipeline safety level and guide pipeline risk management. It is necessary to conduct a safety assessment of the corrosion defects of the in-service oil and gas pipeline system.
[0003] Oil and gas pipelines are corroded during use, which affects the quality and life of use and also brings great safety hazards. At present, there are many studies on the failure of oil and gas pipelines and the loss of pressure in the pipelines after failure, but there are few studies on the problem of pipeline load pressure. Pipeline pressure loss and pipeline load pressure are two completely different accident situations, and the accidents caused by pipeline load pressure are more serious. Therefore, it is crucial to predict and detect the problem of pipeline load pressure more carefully. Studies have found that stress corrosion (especially rust corrosion) in oil and gas pipelines is an important factor causing load pressure; therefore, how to accurately realize the safety assessment of the rust corrosion of the detected oil and gas pipelines is an important factor affecting the safety control of oil and gas valves. If the rust corrosion problems that may exist in the oil and gas pipelines in the future cannot be prepared in time, it will affect the control of the oil and gas valves (at the same time, the existing technology only focuses on the abnormality of a single parameter for the safety assessment and feedback control of load pressure, which is inaccurate and unreasonable).
[0004] Further research found that there are many factors that affect stress corrosion in oil and gas pipelines, and they often have a certain correlation. However, how to correlate and analyze the close relationship between the influencing factors is the key to more accurately assessing the overall corrosion state of the pipeline and ensuring the safe control of its oil and gas valves. Summary of the invention
[0005] The purpose of the present invention is to provide an oil and gas valve safety control and monitoring system to solve the above technical problems pointed out in the prior art.
[0006] The present invention provides an oil and gas valve safety control and monitoring system, including a data acquisition module, an abnormality detection module, a correlation analysis module and an evaluation module;
[0007] Wherein, the data acquisition module is used to acquire a pipeline corrosion monitoring data information set of the pipeline corrosion monitoring data to be measured, wherein the pipeline corrosion monitoring data information set includes a plurality of pipeline corrosion monitoring parameters of multiple evaluation dimensions of the pipeline corrosion monitoring data to be measured;
[0008] The anomaly detection module is used to perform anomaly detection using the pipeline corrosion monitoring data information set to obtain multiple abnormal pipeline corrosion monitoring data sets;
[0009] The association analysis module is used to perform association analysis on each of the pipeline corrosion monitoring parameters in the abnormal pipeline corrosion monitoring data set to obtain the abnormal data association relationship of the pipeline corrosion monitoring data to be tested, and the credibility corresponding to the abnormal data association relationship;
[0010] The evaluation module is used to substitute the abnormal data association relationship and the credibility into the abnormal evaluation function to obtain the abnormal state score of the pipeline corrosion monitoring data to be tested.
[0011] Preferably, the calculation method of the abnormality assessment function is:
[0012]
[0013] In the formula, R(a) is the abnormal state score, h(a s ) is the abnormal scoring function of the sth pipeline corrosion monitoring parameter of the pipeline corrosion monitoring data to be tested, and h(a s )>0, N is the number of pipeline corrosion monitoring parameter types of pipeline corrosion monitoring data to be tested, k(a t ) is the credibility function;
[0014] Among them, the credibility k(a t ) is calculated as:
[0015]
[0016] In the formula, for a s with a t There is a correlation relationship, and the pipeline corrosion monitoring data to be tested is abnormal pipeline corrosion monitoring data; f t is the credibility, 0≤f t ≤1.
[0017] Preferably, the multiple pipeline corrosion monitoring parameters of the multiple evaluation dimensions refer to at least two evaluation dimensions, and each evaluation dimension includes at least two pipeline corrosion monitoring parameters; the pipeline corrosion monitoring data information set is the data obtained by detecting the pipeline corrosion monitoring data to be tested.
[0018] Preferably, the credibility represents the credibility of the abnormal data association relationship.
[0019] Preferably, the association analysis module includes a format unification submodule, an abnormal item set establishment submodule, a first analysis submodule and a second analysis submodule;
[0020] The format unification submodule is used to extract each abnormal pipeline corrosion monitoring data set to obtain abnormal data points; perform unified format setting on each abnormal data point to obtain corresponding first abnormal data points respectively;
[0021] The abnormal item set establishment submodule is used to establish an abnormal item set for all the first abnormal data points in each of the abnormal pipeline corrosion monitoring data sets;
[0022] The first analysis submodule is used to analyze each of the abnormal item sets to obtain support and confidence;
[0023] The second analysis submodule is used to calculate and obtain the credibility according to the support and the confidence; and to analyze and obtain the abnormal data association relationship according to the credibility.
[0024] Preferably, the credibility is calculated as follows:
[0025]
[0026] Where β is the learning rate, which is used to balance the influence of support and confidence; λ is the control parameter, which is used to adjust the influence of learning rate on fitness.
[0027] Preferably, the format unification submodule, when executing the process of "setting the formats of the abnormal data points in a unified manner to obtain the first abnormal data points respectively", specifically includes a eigenvector calculation submodule, a covariance calculation submodule, an eigenvalue calculation submodule, and an abnormal data point calculation submodule;
[0028] The feature vector calculation submodule is used to construct the dimension data point set U from the abnormal data points corresponding to each evaluation dimension. v = {u v 1,u v 2, u v 3, ..., u v o}; among them, u v o represents the oth abnormal data point of the vth evaluation dimension; constructing a covariance matrix M based on the set of dimensional data points; calculating and obtaining the eigenvalues of each of the abnormal data points in the covariance matrix and the eigenvectors corresponding to the eigenvalues;
[0029] The covariance calculation submodule is used for each element in the covariance matrix M to represent the covariance between two corresponding abnormal data points; the elements on the diagonal represent the variance of each abnormal data point itself;
[0030] The eigenvalue calculation submodule is used to arrange the eigenvalues from small to large to obtain an eigenvalue sequence set; and obtain a target eigenvalue K according to the eigenvalue sequence set;
[0031] The abnormal data point calculation submodule is used to calculate and obtain the first abnormal data point P corresponding to each of the abnormal data points according to the covariance matrix M and the target eigenvalue K; the calculation method of the first abnormal data point P is: P=M×K.
[0032] Preferably, the first abnormal data point P represents a data matrix, in which each element represents the first abnormal data point in the same dimension obtained by dimensionality reduction of various abnormal data points in multiple dimensions.
[0033] Preferably, when the abnormal data point calculation submodule executes “obtaining the target eigenvalue K according to the eigenvalue sequence set”, it specifically includes an overall eigenvalue calculation submodule, a characteristic contribution rate calculation submodule, a sum calculation submodule and a judgment processing submodule;
[0034] The overall eigenvalue calculation submodule is used to calculate and obtain the overall eigenvalue based on all the eigenvalues;
[0035] The characteristic contribution rate calculation submodule is used to calculate and obtain the characteristic contribution rate corresponding to each characteristic value according to the overall characteristic value;
[0036] The summation calculation submodule is used to start from the eigenvalue with the highest eigenvalue in the eigenvalue sequence set, and accumulate and calculate in order from high to low to obtain the sum of eigenvalues; and calculate and obtain the sum of eigencontribution rates according to the sum of eigenvalues;
[0037] The judgment processing submodule is used to judge whether the ratio of the feature contribution rate and the overall feature value is greater than or equal to the preset feature contribution rate and the minimum threshold; if so, the feature contribution rate and the corresponding multiple feature values are selected as the target feature value K; if not, the above operation is returned until the ratio of the feature contribution rate and the overall feature value is greater than or equal to the preset feature contribution rate and the minimum threshold.
[0038] The present invention provides an oil and gas valve safety control and monitoring method, comprising the following operating steps:
[0039] Acquire a pipeline corrosion monitoring data information set of the pipeline corrosion monitoring data to be measured, wherein the pipeline corrosion monitoring data information set includes a plurality of pipeline corrosion monitoring parameters of a plurality of evaluation dimensions of the pipeline corrosion monitoring data to be measured;
[0040] Using the pipeline corrosion monitoring data information set to perform anomaly detection to obtain multiple abnormal pipeline corrosion monitoring data sets;
[0041] Performing correlation analysis on each of the pipeline corrosion monitoring parameters in the abnormal pipeline corrosion monitoring data set to obtain an abnormal data correlation relationship of the pipeline corrosion monitoring data to be tested, and a credibility corresponding to the abnormal data correlation relationship;
[0042] Substitute the abnormal data association relationship and the credibility into the abnormal evaluation function to obtain the abnormal state score of the pipeline corrosion monitoring data to be tested.
[0043] Compared with the prior art, the embodiments of the present invention have at least the following technical advantages:
[0044] By analyzing the above-mentioned oil and gas valve safety control and monitoring system provided by the present invention, it can be known that in specific applications, firstly, by recording in detail the pipeline corrosion monitoring parameters of various evaluation dimensions of each pipeline corrosion monitoring data information set, the comprehensiveness and meticulousness of the data can be ensured, providing a rich source of information for subsequent steps; further, scattered anomalies are identified by mapping and analyzing scatter plots. This process makes anomaly detection more intuitive and easier to understand, effectively distinguishes normal and abnormal pipeline corrosion monitoring data, and provides a clear goal and basis for further analysis; in-depth analysis of pipeline corrosion monitoring data parameters helps to understand the complex relationship between pipeline corrosion monitoring states, and provides more accurate data support for oil and gas transportation safety prevention and maintenance; finally, the correlation relationship and credibility are substituted into the abnormal state score calculated by the abnormal evaluation function, providing technical engineers with quantitative data reference, which helps to accurately diagnose the pipeline corrosion monitoring state. BRIEF DESCRIPTION OF THE DRAWINGS
[0045] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0046] Figure 1 A schematic diagram of the overall operation steps of an oil and gas valve safety control and monitoring system provided in the first embodiment of the present invention;
[0047] Figure 2 A schematic diagram of a scatter plot simulation in an oil and gas valve safety control and monitoring system provided in the first embodiment of the present invention;
[0048] Figure 3 A schematic diagram of a simulation of correlation relationships in an oil and gas valve safety control and monitoring system provided in the first embodiment of the present invention;
[0049] Figure 4A schematic diagram of the operation steps for analyzing and obtaining abnormal data association and credibility in an oil and gas valve safety control and monitoring system provided in the first embodiment of the present invention;
[0050] Figure 5 A schematic diagram of the operation steps for obtaining a first abnormal data point in an oil and gas valve safety control and monitoring system provided in the first embodiment of the present invention;
[0051] Figure 6 A schematic diagram of the operation steps for obtaining a target characteristic value K in an oil and gas valve safety control and monitoring system provided in the first embodiment of the present invention;
[0052] Figure 7 A schematic diagram of the overall architecture of an oil and gas valve safety control and monitoring system provided in Embodiment 2 of the present invention.
[0053] Reference numerals: data acquisition module 10 , anomaly detection module 20 , association analysis module 30 , evaluation module 40 , format unification submodule 31 , abnormal item set establishment submodule 32 , first analysis submodule 33 , second analysis submodule 34 . DETAILED DESCRIPTION
[0054] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0055] The present invention is further described in detail below through specific embodiments in conjunction with the accompanying drawings.
[0056] Embodiment 1
[0057] like Figure 1 As shown, the present invention proposes an oil and gas valve safety control and monitoring system and its implementation processing method, including the following operating steps:
[0058] Step S10: obtaining a pipeline corrosion monitoring data information set of the pipeline corrosion monitoring data to be measured, wherein the pipeline corrosion monitoring data information set includes a plurality of pipeline corrosion monitoring parameters of a plurality of evaluation dimensions of the pipeline corrosion monitoring data to be measured;
[0059] It should be noted that the multiple pipeline corrosion monitoring parameters of the above-mentioned multiple evaluation dimensions refer to at least two or more evaluation dimensions, and each evaluation dimension includes at least two pipeline corrosion monitoring parameters. The pipeline corrosion monitoring data information set is the data obtained by detecting the pipeline corrosion monitoring data to be tested. For example, during the implementation process, professional equipment such as high-definition electron microscopes are used to detect pipeline corrosion monitoring rust sheets (i.e., rust sample sheets peeled off somewhere) to obtain pipeline corrosion monitoring data; the above-mentioned evaluation dimensions include morphological characteristics of the pipeline corrosion monitoring data to be tested, corrosion thickness loss, rust surface roughness characteristics and other evaluation dimensions; accordingly, the above-mentioned pipeline corrosion monitoring parameters refer to the parameter values of the above-mentioned evaluation dimensions, such as morphological characteristic values, corrosion thickness loss values, etc.
[0060] Analyzing the corrosion surface characteristics of the outer surface of the oil and gas pipeline, it is found that only by preparing a multi-faceted assessment can its surface characteristics be accurately reflected. Therefore, the collection and analysis of surface morphology (morphological characteristics, corrosion thickness loss, and rust surface roughness characteristics) can reveal the characteristics of rust pits and lay the foundation for subsequent safety analysis.
[0061] Generally speaking, if the roughness of the rusted surface of the pipeline is obvious, that is, the roughness of the rusted surface is higher, it proves that the rust corrosion is more serious. On the contrary, if the roughness of the rusted surface of the pipeline is relatively smooth, that is, the roughness of the rusted surface is lower, it proves that the rust corrosion is not very serious.
[0062] Step S20: using the pipeline corrosion monitoring data information set to perform abnormality detection to obtain multiple abnormal pipeline corrosion monitoring data sets;
[0063] It should be noted that the above anomaly detection refers to obtaining a scatter plot (such as Figure 2 As shown), the pipeline corrosion monitoring data information corresponding to the scattered abnormal scatter plot is obtained; the above-mentioned scattered abnormality refers to the scattered scatter plot when the distribution of each pipeline corrosion monitoring parameter is outside the preset normal scattered area;
[0064] Step S30: performing correlation analysis on each of the pipeline corrosion monitoring parameters in the abnormal pipeline corrosion monitoring data set to obtain the abnormal data correlation relationship of the pipeline corrosion monitoring data to be tested, and the credibility corresponding to the abnormal data correlation relationship;
[0065] It should be noted that the above correlation analysis refers to the analysis of the correlation between each pipeline corrosion monitoring parameter in each abnormal pipeline corrosion monitoring data set, such as the correlation between the morphological feature value and the corrosion thickness loss value (such as Figure 3As shown in the figure), we can have a deeper understanding of the relationship between different pipeline corrosion monitoring states, thereby affecting the abnormal state score of the subsequent pipeline corrosion data to be tested; the above credibility refers to the credibility of the abnormal data association relationship, that is, the degree of association between abnormal item sets. This credibility reflects the reliability or confidence level of the abnormal data association relationship (different from the confidence described below, the credibility here is calculated by combining confidence and support as influencing parameters. For details, see the following operation. It is to combine support and confidence to obtain a comprehensive indicator to measure the credibility of the abnormal data association relationship);
[0066] Step S40: Substitute the abnormal data association relationship and the credibility (ie, substitute them into the following calculation formula as parameter factors respectively) into the abnormal evaluation function to obtain the abnormal state score of the pipeline corrosion monitoring data to be tested.
[0067] The calculation method of the anomaly assessment function is:
[0068]
[0069] In the formula, R(a) is the abnormal state score, h(a s ) is the abnormal scoring function of the sth pipeline corrosion monitoring parameter of the pipeline corrosion monitoring data to be tested, and h(a s )>0, N is the number of pipeline corrosion monitoring parameter types of pipeline corrosion monitoring data to be tested, k(a t ) is the credibility function;
[0070] Among them, the credibility k(a t ) is calculated as:
[0071]
[0072] In the formula, a s ⊕a t for a s with a t There is a correlation relationship, and the pipeline corrosion monitoring data to be tested is abnormal pipeline corrosion monitoring data; f t is the credibility, 0≤f t ≤1.
[0073] It should be noted that the above-mentioned embodiment of the present application firstly records in detail the pipeline corrosion monitoring parameters of various evaluation dimensions of each pipeline corrosion monitoring data information set, so as to ensure the comprehensiveness and meticulousness of the data and provide a rich source of information for subsequent steps; further, scattered anomalies are identified by mapping and analyzing scatter plots. This process makes anomaly detection more intuitive and easier to understand, effectively distinguishes normal and abnormal pipeline corrosion monitoring data, and provides a clear goal and basis for further analysis; in-depth analysis of pipeline corrosion monitoring data parameters helps to understand the complex relationship between pipeline corrosion monitoring states, and provides more accurate data support for safety prevention and maintenance of oil and gas transportation;
[0074] Finally, the correlation and credibility are substituted into the abnormal status score calculated by the abnormal assessment function, providing technical engineers with quantitative data reference, which is helpful to accurately diagnose the pipeline corrosion monitoring status and ultimately ensure early warning and safety control of oil and gas valves. For example, if subsequent detection predicts that the current safety corrosion assessment result is relatively serious, the oil and gas valves can be partially closed in advance or some pipeline sections can be depressurized to transport to the spare pipeline to reduce the possibility of subsequent accidents.
[0075] The above-mentioned correlation analysis refers to analyzing the correlation between each pipeline corrosion monitoring parameter in each abnormal pipeline corrosion monitoring data set, such as the correlation between the morphological feature value and the corrosion thickness loss value; for example, through correlation analysis, the interaction and influence between different pipeline corrosion monitoring parameters can be fully understood. This means that the system of this embodiment not only focuses on the abnormality of a single parameter, but can identify the complex relationship between multiple parameters, so as to more accurately evaluate the overall corrosion status of the pipeline.
[0076] The above correlation analysis can help to deeply understand the relationship between different parameters, such as the correlation between the morphological characteristic values and the corrosion thickness loss values, which helps to reveal the development trend of pipeline corrosion. By analyzing the correlation between parameters, the patterns and laws of abnormal data can be identified. In some cases, the abnormality of a single parameter may not be enough to cause concern, but when multiple parameters show abnormal correlation at the same time, this may indicate a more serious corrosion problem. That is, the above parameter factors are referenced, and the correlation between the above identification parameters is also referenced, and finally the abnormal state score is output, which can more accurately predict and evaluate the corrosion state of the pipeline, thereby reducing false alarms and missed alarms.
[0077] The subsequent execution of "abnormal data association relationship, and the credibility corresponding to the abnormal data association relationship" is obtained in the following manner;
[0078] Specifically, Figure 4As shown, in step S30, the abnormal pipeline corrosion monitoring data set is subjected to association analysis to obtain the abnormal data association relationship of the pipeline corrosion monitoring data to be tested and the credibility corresponding to the abnormal data association relationship, including the following operation steps:
[0079] Step S31: extracting each abnormal pipeline corrosion monitoring data set to obtain abnormal data points; performing unified format setting on each abnormal data point to obtain corresponding first abnormal data points respectively;
[0080] Step S32: establishing an abnormal item set for all the first abnormal data points in each of the abnormal pipeline corrosion monitoring data sets;
[0081] Step S33: Analyze each of the abnormal item sets to obtain support and confidence;
[0082] It should be noted that the above abnormal data association relationship is obtained by analyzing the support and confidence of each pipeline corrosion monitoring parameter in the abnormal item set. The calculation of the support is illustrated as follows: there are currently 10 abnormal item sets, of which 5 evaluation dimensions include morphological features, corrosion thickness loss, rust surface roughness features, pipeline corrosion monitoring rust sheet area, and pipeline corrosion monitoring rust sheet texture; among them, in the morphological features, there are 4 pipeline corrosion monitoring data parameters with a morphological feature value of 2 in each abnormal item set, and the support of the morphological feature is 4÷10=0.4;
[0083] Furthermore, the calculation method of confidence is exemplified as follows: Continuing with the above example, in the 10 abnormal item sets, the evaluation values of each evaluation dimension (i.e., the above morphological feature values, corrosion thickness loss values, rust surface roughness feature values, pipeline corrosion monitoring rust sheet area values, pipeline corrosion monitoring rust sheet texture values) will have a pre-set abnormal value interval in the scattered point diagram. The confidence is calculated by judging that each evaluation dimension is within the abnormal value interval. Each pipeline corrosion monitoring parameter is taken as a sub-item set. The probability of another sub-item set appearing when each sub-item set appears is the above confidence. For example, the confidence of the morphological feature value is calculated, and the number of times E when the morphological feature value of the sub-item set is 2 and the number of times Q when the morphological feature value of the sub-item set and the corrosion thickness loss value are 4 in each abnormal item set are counted.
[0084] The above step S33 quantifies the strength and reliability of the association between abnormal data points by calculating support and confidence. Among them, support reflects the universality of abnormal data points, while confidence reflects the prediction accuracy of abnormal data points under specific conditions. Through the learning rate β and control parameter λ introduced in step S34, the system can balance the influence of support and confidence, making the calculation of credibility highly adaptable, and finally outputting the final calculation result of credibility.
[0085] Furthermore, the confidence level when the corrosion loss value is 4 when the morphological feature value of the sub-item set is 2 is calculated as: Q÷E.
[0086] Step S34: Calculate the credibility based on the support and the confidence; and analyze the abnormal data association relationship based on the credibility;
[0087] The credibility is calculated as follows:
[0088] In the formula, β is the learning rate, which is used to balance the influence of support and confidence; λ is the control parameter, which is used to adjust the influence of learning rate on fitness;
[0089] It should be noted that the above embodiment of the present application is described by way of example: Assume that there is an abnormal data set, which contains the following two abnormal item sets:
[0090] Abnormal item set X: {morphological feature value, corrosion thickness loss value}, which means that if the morphological feature value (for example, 2) appears, the corrosion thickness loss value (for example, 4) appears;
[0091] Abnormal item set Y: {corrosion thickness loss value, pipeline corrosion monitoring rust sheet area value}, which means that if the corrosion thickness loss value (for example, 1) appears, the pipeline corrosion monitoring rust sheet area value (for example, 6) appears;
[0092] The support and confidence of these two abnormal item sets have now been calculated:
[0093] Support: the frequency of occurrence of abnormal item sets, that is, the probability of including these items in the data set at the same time;
[0094] Confidence level: The probability that a conclusion (such as the value of corrosion loss thickness) is valid when the premise (such as the value of morphological characteristics) is met.
[0095] Assume that the support of abnormal item set X is 0.2 and the confidence is 0.8; the support of abnormal item set Y is 0.3 and the confidence is 0.6;
[0096] Now we want to calculate the credibility of these two abnormal item sets; assume that the weighted combination method adopted is to take the average value of support and confidence (that is, the above-mentioned learning rate β), that is, β = 0.5.
[0097] The following formula is used to calculate the credibility:
[0098] Assuming λ=1, the credibility of the abnormal item set X is calculated as follows:
[0099]
[0100] The credibility of the abnormal item set Y is calculated as follows:
[0101]
[0102] Therefore, the credibility of the abnormal item set X is about 0.3846, and the credibility of the abnormal item set Y is about 0.3913;
[0103] Furthermore, the abnormal data association relationship obtained by analyzing the credibility is based on the preset credibility minimum threshold. When the calculated credibility is greater than or equal to the credibility threshold, the abnormal item set is considered to be an existing association relationship. For example, if the preset credibility minimum threshold is 0.3, the abnormal item set X: {morphological feature value, corrosion thickness loss value} and the abnormal item set Y: {corrosion thickness loss value, pipeline corrosion monitoring rust sheet area value} are obviously greater than the credibility minimum threshold of 0.3, so they are determined to be reasonable association relationships; unreasonable association relationships are deleted and discarded;
[0104] During the specific implementation of the above-mentioned embodiment of the present application, the technicians found that in the specific operation process of uniformly formatting each of the abnormal data points and obtaining the first abnormal data points respectively, due to the large deviation of the data dimensions, it is impossible to accurately obtain a data format setting standard with similar dimensions, so that the error of the obtained first abnormal data point is large, thereby affecting the determination of the subsequent association relationship;
[0105] Therefore, it is also necessary to analyze the unified format setting of the abnormal data points to obtain an accurate format setting method, so that the first abnormal data point obtained can accurately express the meaning of the abnormal data parameters of each dimension, so that the subsequent determination of the association relationship is accurate. See the execution process of steps S311-S314 for details. At the same time, the above operation steps are conducive to realizing the dimensionality reduction and simplified analysis and processing of complex data structures.
[0106] Specifically, Figure 5 As shown, in step S31, a unified format is set for each of the abnormal data points to obtain a first abnormal data point respectively, including the following operation steps:
[0107] Step S311: construct the abnormal data points corresponding to each evaluation dimension to obtain a dimension data point set U v = {u v 1,u v 2,u v 3, ..., u v o}; among them, u v o represents the oth abnormal data point in the vth evaluation dimension;
[0108] Step S312: constructing a covariance matrix M based on the set of dimensional data points; calculating and obtaining the eigenvalues of each of the abnormal data points in the covariance matrix and the eigenvectors corresponding to the eigenvalues;
[0109] Each element in the covariance matrix M represents the covariance between the corresponding two abnormal data points (such as the covariance between two corrosion loss values); the elements on the diagonal represent the variance of each abnormal data point itself; the diagonal refers to the elements along the main diagonal (the diagonal from the upper left corner to the lower right corner) in the matrix (covariance matrix). In the covariance matrix, the elements on the diagonal represent the variance of the abnormal data point itself. The above covariance is used to measure the strength of the linear relationship between two abnormal data points.
[0110] It should be noted that the above eigenvalues refer to the relationship between different evaluation dimensions of each abnormal data point in the covariance matrix; through the above calculation and analysis process, the information that can be retained in each direction (direction is eigenvector) (retained information is eigenvalue), so as to obtain the direction (eigenvector) that can retain the most information (eigenvalue); therefore, extracting each abnormal pipeline corrosion monitoring data set to obtain abnormal data points is an important data analysis process; first of all, it is preferred to set a unified format for each abnormal data point, and obtain the first abnormal data point respectively; which involves constructing a set of dimensional data points--constructing a covariance matrix--characterization analysis of elements in the matrix--finally calculating the corresponding first abnormal data point P;
[0111] Step S313: Arrange the eigenvalues from small to large to obtain an eigenvalue sequence set; obtain a target eigenvalue K according to the eigenvalue sequence set;
[0112] Step S314: Calculate and obtain the first abnormal data point P corresponding to each of the abnormal data points according to the covariance matrix M and the target eigenvalue K;
[0113] The calculation method of the first abnormal data point P is:
[0114] P = M × K;
[0115] It should be noted that the calculation of the first abnormal data point P above results in a data matrix, in which each element represents the first abnormal data point in the same dimension obtained after dimensionality reduction of various abnormal data points in multiple dimensions;
[0116] In step S314, the covariance matrix M and the target eigenvalue K are used to calculate the first abnormal data point P. This calculation process is actually a dimensionality reduction process, which projects the original multidimensional data points into a new space defined by the eigenvector to highlight the most important features in the data. The final P is a data matrix, each element of which represents the representation of the original multidimensional abnormal data point in the new feature space after feature analysis and dimensionality reduction. These data points are now located in the same dimension, making them easier to compare and analyze. The above-mentioned embodiment of the present application reduces the dimensionality of the multidimensional abnormal data points by means of principal component analysis to obtain the first abnormal data point of one dimension. The subsequent judgment of the association relationship is performed by the first abnormal data point in a unified format, which can make the result of the judgment analysis more accurate.
[0117] In general, this process is a method of data dimensionality reduction and feature extraction, which identifies the most important features in the data by analyzing the covariance relationship between abnormal data points and reduces its dimensionality to a form that is easier to handle and analyze. This helps to more clearly identify and understand patterns and anomalies in the data in subsequent data analysis.
[0118] Furthermore, during the specific implementation of the above-mentioned embodiments of the present application, technical personnel also found that the first abnormal data point obtained by the dimensionality reduction processing method of principal component analysis will cause the loss of characteristic information of some data, thereby affecting the subsequent judgment of the association relationship. Therefore, further analysis and processing is needed to reduce the loss of characteristic information to a greater extent and improve the accuracy of the subsequent judgment of the association relationship.
[0119] Specifically, Figure 6 As shown, in step S313, obtaining the target eigenvalue K according to the eigenvalue sequence set includes the following steps:
[0120] Step S3131: Obtaining an overall eigenvalue based on all the eigenvalues;
[0121] It should be noted that the above eigenvalues reflect the degree of variation of each abnormal data point on the corresponding eigenvector, while the overall eigenvalue reflects the overall degree of variation;
[0122] Step S3132: Calculate and obtain the characteristic contribution rate corresponding to each characteristic value according to the overall characteristic value;
[0123] It should be noted that the above-mentioned characteristic contribution rate refers to the proportion of each characteristic value in the overall characteristic value;
[0124] Step S3133: starting from the eigenvalue with the highest eigenvalue in the eigenvalue sequence set, accumulating and calculating in order from high to low to obtain the sum of eigenvalues; and calculating and obtaining the sum of eigencontribution rates according to the sum of eigenvalues;
[0125] It should be noted that the above-mentioned feature contribution rate represents the feature value and the sum of the feature contribution rates corresponding to the corresponding multiple feature values.
[0126] Step S3134: Determine whether the ratio of the feature contribution rate and the overall eigenvalue is greater than or equal to the preset feature contribution rate and the minimum threshold; if so, select the feature contribution rate and the corresponding multiple eigenvalues as the target eigenvalue K; if not, return to the above operation until the ratio of the feature contribution rate and the overall eigenvalue is greater than or equal to the preset feature contribution rate and the minimum threshold.
[0127] It should be noted that the above-mentioned embodiment of the present application sorts the calculated eigenvalues from large to small according to size, because a large eigenvalue indicates that the principal component occupies a larger variance in the data, that is, more information (but directly selecting an eigenvalue with the largest eigenvalue as the target eigenvalue will cause the loss of feature information for the first abnormal data point obtained by the subsequent unified format setting), and then calculates the contribution rate corresponding to each eigenvalue to obtain the feature contribution rate, and further selects the feature contribution rate and the corresponding multiple eigenvalues and as the target eigenvalue K through the judgment of the feature contribution rate and. This can retain data information to a greater extent and reduce the loss of data information, thereby improving the accuracy of subsequent association relationship judgment.
[0128] Embodiment 2
[0129] like Figure 7 As shown, accordingly, the present invention also proposes an oil and gas valve safety control and monitoring system, including a data acquisition module 10, an abnormality detection module 20, a correlation analysis module 30 and an evaluation module 40;
[0130] The data acquisition module 10 is used to acquire a pipeline corrosion monitoring data information set of the pipeline corrosion monitoring data to be measured, wherein the pipeline corrosion monitoring data information set includes a plurality of pipeline corrosion monitoring parameters of multiple evaluation dimensions of the pipeline corrosion monitoring data to be measured;
[0131] The multiple pipeline corrosion monitoring parameters of the multiple evaluation dimensions refer to at least two evaluation dimensions, and each evaluation dimension includes at least two pipeline corrosion monitoring parameters; the pipeline corrosion monitoring data information set is the data obtained by detecting the pipeline corrosion monitoring data to be tested;
[0132] The anomaly detection module 20 is used to perform anomaly detection using the pipeline corrosion monitoring data information set to obtain multiple abnormal pipeline corrosion monitoring data sets;
[0133] The association analysis module 30 is used to perform association analysis on each of the pipeline corrosion monitoring parameters in the abnormal pipeline corrosion monitoring data set to obtain the abnormal data association relationship of the pipeline corrosion monitoring data to be tested, and the credibility corresponding to the abnormal data association relationship;
[0134] The credibility represents the credibility of the abnormal data association relationship;
[0135] The evaluation module 40 is used to substitute the abnormal data association relationship and the credibility into the abnormal evaluation function to obtain the abnormal state score of the pipeline corrosion monitoring data to be tested.
[0136] The calculation method of the anomaly assessment function is:
[0137]
[0138] In the formula, R(a) is the abnormal state score, h(a s ) is the abnormal scoring function of the sth pipeline corrosion monitoring parameter of the pipeline corrosion monitoring data to be tested, and h(a s )>0, N is the number of pipeline corrosion monitoring parameter types of pipeline corrosion monitoring data to be tested, k(a t ) is the credibility function;
[0139] Among them, the credibility k(a t ) is calculated as:
[0140]
[0141] In the formula, for a s with a t There is a correlation relationship, and the pipeline corrosion monitoring data to be tested is abnormal pipeline corrosion monitoring data; f t is the credibility, 0≤f t ≤1.
[0142] The present invention also proposes an oil and gas valve safety control and monitoring system, comprising a data acquisition module 10, an abnormality detection module 20, a correlation analysis module 30 and an evaluation module 40;
[0143] Data acquisition module 10: responsible for collecting and acquiring the corrosion monitoring data of the pipeline to be tested, which includes multiple pipeline corrosion monitoring parameters in multiple evaluation dimensions.
[0144] Abnormal detection module 20: Use the pipeline corrosion monitoring data information set for analysis, identify abnormal conditions in the data, and form an abnormal pipeline corrosion monitoring data set. The most important is the association analysis module 30; the association analysis module 30 is used to conduct in-depth analysis of the parameters in the abnormal pipeline corrosion monitoring data set, identify the associations between the parameters, and evaluate the credibility of these associations. Through association analysis, the system can understand the interactions and dependencies between different corrosion monitoring parameters, and then evaluate the credibility of abnormal data, providing important association information for the evaluation of abnormal conditions.
[0145] The evaluation module 40 output unit is used to take the abnormal data association relationship and credibility as input, and calculate the abnormal state score of the pipeline corrosion monitoring data to be tested through the abnormal evaluation function. The evaluation module integrates the results of the association analysis and provides a quantitative abnormal state score, which can intuitively reflect the severity and possible risks of pipeline corrosion.
[0146] The above four modules together constitute a complete monitoring and evaluation process, from data collection to abnormal identification, to correlation analysis and final evaluation, ensuring the systematic and effective safety control and monitoring of oil and gas valves. Through this systematic approach, the state of pipeline corrosion can be more accurately predicted and evaluated, and maintenance and repair measures can be taken in a timely manner, thereby ensuring the safe operation of oil and gas pipelines.
[0147] In the specific implementation process, relying on the above-mentioned oil and gas valve safety control and monitoring system, the processing method implemented is realized. Through in-depth analysis of pipeline corrosion monitoring data parameters, it is helpful to understand the complex relationship between pipeline corrosion monitoring status; finally, the correlation relationship and credibility are substituted into the abnormal state score calculated by the abnormal evaluation function to provide technical engineers with quantitative data reference, which is helpful to accurately diagnose the pipeline corrosion monitoring status, and finally ensure the early warning and safety control of oil and gas valves; for example, if subsequent detection predicts that the current safety corrosion assessment result is relatively serious, it is entirely possible to partially close the oil and gas valves in advance or relieve pressure on some pipeline sections to transport to the spare pipeline to reduce the possibility of subsequent accidents.
[0148] The above-mentioned correlation analysis refers to analyzing the correlation between each pipeline corrosion monitoring parameter in each abnormal pipeline corrosion monitoring data set, such as the correlation between the morphological feature value and the corrosion thickness loss value; for example, through correlation analysis, the interaction and influence between different pipeline corrosion monitoring parameters can be fully understood. This means that the system of this embodiment not only focuses on the abnormality of a single parameter, but can identify the complex relationship between multiple parameters, so as to more accurately evaluate the overall corrosion status of the pipeline.
[0149] The above correlation analysis can help to deeply understand the relationship between different parameters, such as the correlation between the morphological characteristic values and the corrosion thickness loss values, which helps to reveal the development trend of pipeline corrosion. By analyzing the correlation between parameters, the patterns and laws of abnormal data can be identified. In some cases, the abnormality of a single parameter may not be enough to cause concern, but when multiple parameters show abnormal correlation at the same time, this may indicate a more serious corrosion problem. That is, the above parameter factors are referenced, and the correlation between the above identification parameters is also referenced, and finally the abnormal state score is output, which can more accurately predict and evaluate the corrosion state of the pipeline, thereby reducing false alarms and missed alarms.
[0150] Preferably, the association analysis module 30 includes a format unification submodule 31, an abnormal item set establishment submodule 32, a first analysis submodule 33 and a second analysis submodule 34;
[0151] The format unification submodule 31 is used to extract each abnormal pipeline corrosion monitoring data set to obtain abnormal data points; perform unified format setting on each abnormal data point to obtain corresponding first abnormal data points respectively;
[0152] The abnormal item set establishing submodule 32 is used to establish an abnormal item set for all the first abnormal data points in each of the abnormal pipeline corrosion monitoring data sets;
[0153] The first analysis submodule 33 is used to analyze each of the abnormal item sets to obtain support and confidence;
[0154] The second analysis submodule 34 is used to calculate and obtain the credibility according to the support and the confidence; and to analyze and obtain the abnormal data association relationship according to the credibility.
[0155] The credibility is calculated as follows:
[0156]
[0157] Where β is the learning rate, which is used to balance the influence of support and confidence; λ is the control parameter, which is used to adjust the influence of learning rate on fitness.
[0158] Preferably, the format unification submodule, when executing the process of "setting the formats of the abnormal data points in a unified manner to obtain the first abnormal data points respectively", specifically includes a eigenvector calculation submodule, a covariance calculation submodule, an eigenvalue calculation submodule, and an abnormal data point calculation submodule;
[0159] The feature vector calculation submodule is used to construct the dimension data point set U from the abnormal data points corresponding to each evaluation dimension. v = {uv 1,u v 2, u v 3, ..., u v o}; among them, u v o represents the oth abnormal data point of the vth evaluation dimension; constructing a covariance matrix M based on the set of dimensional data points; calculating and obtaining the eigenvalues of each of the abnormal data points in the covariance matrix and the eigenvectors corresponding to the eigenvalues;
[0160] The covariance calculation submodule is used for each element in the covariance matrix M to represent the covariance between two corresponding abnormal data points; the elements on the diagonal represent the variance of each abnormal data point itself;
[0161] The eigenvalue calculation submodule is used to arrange the eigenvalues from small to large to obtain an eigenvalue sequence set; and obtain a target eigenvalue K according to the eigenvalue sequence set;
[0162] The abnormal data point calculation submodule is used to calculate and obtain the first abnormal data point P corresponding to each of the abnormal data points according to the covariance matrix M and the target eigenvalue K; the calculation method of the first abnormal data point P is: P=M×K.
[0163] Preferably, the first abnormal data point P represents a data matrix, in which each element represents the first abnormal data point in the same dimension obtained by dimensionality reduction of various abnormal data points in multiple dimensions.
[0164] Preferably, when the abnormal data point calculation submodule executes “obtaining the target eigenvalue K according to the eigenvalue sequence set”, it specifically includes an overall eigenvalue calculation submodule, a characteristic contribution rate calculation submodule, a sum calculation submodule and a judgment processing submodule;
[0165] The overall eigenvalue calculation submodule is used to calculate and obtain the overall eigenvalue according to all the eigenvalues;
[0166] The characteristic contribution rate calculation submodule is used to calculate and obtain the characteristic contribution rate corresponding to each characteristic value according to the overall characteristic value;
[0167] The summation calculation submodule is used to start from the eigenvalue with the highest eigenvalue in the eigenvalue sequence set, and accumulate and calculate in order from high to low to obtain the sum of eigenvalues; and calculate and obtain the sum of eigencontribution rates according to the sum of eigenvalues;
[0168] The judgment processing submodule is used to judge whether the ratio of the feature contribution rate and the overall feature value is greater than or equal to the preset feature contribution rate and the minimum threshold; if so, the feature contribution rate and the corresponding multiple feature values are selected as the target feature value K; if not, the above operation is returned until the ratio of the feature contribution rate and the overall feature value is greater than or equal to the preset feature contribution rate and the minimum threshold.
[0169] In summary, the oil and gas valve safety control and monitoring system proposed in the example of the present invention can ensure the comprehensiveness and meticulousness of the data by first recording in detail the pipeline corrosion monitoring parameters of various evaluation dimensions of each pipeline corrosion monitoring data information set, thereby providing a rich information source for subsequent steps;
[0170] At the same time, it further identifies scattered anomalies by mapping and analyzing scatter plots. This process makes anomaly detection more intuitive and easier to understand, effectively distinguishing normal and abnormal pipeline corrosion monitoring data, and providing a clear goal and basis for further analysis; in-depth analysis of pipeline corrosion monitoring data parameters helps to understand the complex relationship between pipeline corrosion monitoring states, and provides more accurate data support for quality defect prevention and maintenance of oil and gas pipelines;
[0171] Finally, the association relationship and credibility are substituted into the abnormal state score calculated by the abnormal evaluation function to provide a quantitative data reference for technical engineers, which is helpful for accurately diagnosing the pipeline corrosion monitoring status; further, when analyzing the association of pipeline corrosion monitoring data parameters, the multi-dimensional abnormal data points are reduced in dimension by principal component analysis to obtain the first abnormal data point of one dimension. The subsequent association relationship judgment is performed by the first abnormal data point in a unified format, which can make the judgment analysis result more accurate; the calculated eigenvalues are sorted from large to small according to size, because a large eigenvalue indicates that the principal component has a larger variance in the data, that is, more information, and then the contribution rate corresponding to each eigenvalue is calculated to obtain the feature contribution rate. Further, through the judgment of the feature contribution rate and the sum, the multiple eigenvalues corresponding to the feature contribution rate and the sum are selected as the target eigenvalue K, which can retain data information to a greater extent and reduce the loss of data information, thereby improving the accuracy of subsequent association relationship judgment;
[0172] The credibility of the association relationship is further calculated and analyzed by analyzing the confidence and support obtained from the abnormal item set established by the first abnormal data point, so as to analyze the association relationship of each abnormal data parameter.
[0173] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. A person skilled in the art may modify the technical solutions described in the above embodiments, or replace part or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.
Claims
1. An oil and gas valve safety control and monitoring system, characterized in that: It includes data acquisition module, anomaly detection module, correlation analysis module and evaluation module; Wherein, the data acquisition module is used to acquire a pipeline corrosion monitoring data information set of the pipeline corrosion monitoring data to be measured, wherein the pipeline corrosion monitoring data information set includes a plurality of pipeline corrosion monitoring parameters of multiple evaluation dimensions of the pipeline corrosion monitoring data to be measured; The anomaly detection module is used to perform anomaly detection using the pipeline corrosion monitoring data information set to obtain multiple abnormal pipeline corrosion monitoring data sets; The association analysis module is used to perform association analysis on each of the pipeline corrosion monitoring parameters in the abnormal pipeline corrosion monitoring data set to obtain the abnormal data association relationship of the pipeline corrosion monitoring data to be tested, and the credibility corresponding to the abnormal data association relationship; The evaluation module is used to substitute the abnormal data association relationship and the credibility into the abnormal evaluation function to obtain the abnormal state score of the pipeline corrosion monitoring data to be tested; The association analysis module includes a format unification submodule, an abnormal item set establishment submodule, a first analysis submodule and a second analysis submodule; The format unification submodule is used to extract each abnormal pipeline corrosion monitoring data set to obtain abnormal data points; perform unified format setting on each abnormal data point to obtain corresponding first abnormal data points respectively; The abnormal item set establishment submodule is used to establish an abnormal item set for all the first abnormal data points in each of the abnormal pipeline corrosion monitoring data sets; The first analysis submodule is used to analyze each of the abnormal item sets to obtain support and confidence; Among them, the calculation of support: there are currently 10 abnormal item sets, of which 5 are evaluated dimensions, including morphological characteristics, corrosion thickness loss, rust surface roughness characteristics, pipeline corrosion monitoring rust sheet area, pipeline corrosion monitoring rust sheet texture; among them, in the morphological characteristics, there are 4 pipeline corrosion monitoring data parameters with a morphological feature value of 2 in each abnormal item set, and the support of the morphological feature is 4÷10=0.4; Among them, the calculation of confidence is as follows: in the 10 abnormal item sets, the evaluation values of each evaluation dimension will have a pre-set abnormal value interval in the scatter plot. The confidence is calculated by judging the evaluation dimensions whose evaluation values are within the abnormal value interval. Each pipeline corrosion monitoring parameter is taken as a sub-item set. The probability of another sub-item set appearing when each sub-item set appears is recorded as the above confidence: the confidence of the morphological feature value is calculated, and the number of times E when the morphological feature value of the sub-item set in each abnormal item set appears and the number of times Q when the morphological feature value of the sub-item set is 2 and the corrosion thickness loss value is 4 appear at the same time are counted; The confidence level of corrosion loss value of 4 when the value of the morphological feature of the sub-item set is 2 is calculated as: Q÷E; The second analysis submodule is used to calculate and obtain the credibility according to the support and the confidence; and to analyze and obtain the abnormal data association relationship according to the credibility; The format unification submodule, when executing the process of "setting the formats of the abnormal data points in a unified manner to obtain the first abnormal data points respectively", specifically includes a eigenvector calculation submodule, a covariance calculation submodule, an eigenvalue calculation submodule, and an abnormal data point calculation submodule; The feature vector calculation submodule is used to construct the dimension data point set U from the abnormal data points corresponding to each evaluation dimension. v = {u v 1,u v 2,u v 3, ..., u v o}; among them, u v o represents the oth abnormal data point of the vth evaluation dimension; constructing a covariance matrix M based on the set of dimensional data points; calculating and obtaining the eigenvalues of each of the abnormal data points in the covariance matrix and the eigenvectors corresponding to the eigenvalues; The covariance calculation submodule is used for each element in the covariance matrix M to represent the covariance between two corresponding abnormal data points; the elements on the diagonal represent the variance of each abnormal data point itself; The eigenvalue calculation submodule is used to arrange the eigenvalues from small to large to obtain an eigenvalue sequence set; and obtain a target eigenvalue K according to the eigenvalue sequence set; The abnormal data point calculation submodule is used to calculate and obtain the first abnormal data point P corresponding to each abnormal data point according to the covariance matrix M and the target eigenvalue K; the calculation method of the first abnormal data point P is: P=M×K; When the abnormal data point calculation submodule executes "obtaining the target eigenvalue K according to the eigenvalue sequence set", it specifically includes an overall eigenvalue calculation submodule, a characteristic contribution rate calculation submodule, a sum calculation submodule and a judgment processing submodule; The overall eigenvalue calculation submodule is used to calculate and obtain the overall eigenvalue according to all the eigenvalues; The characteristic contribution rate calculation submodule is used to calculate and obtain the characteristic contribution rate corresponding to each characteristic value according to the overall characteristic value; The summation calculation submodule is used to start from the eigenvalue with the highest eigenvalue in the eigenvalue sequence set, and accumulate and calculate in order from high to low to obtain the sum of eigenvalues; and calculate and obtain the sum of eigencontribution rates according to the sum of eigenvalues; The judgment processing submodule is used to judge whether the ratio of the sum of the feature contribution rate and the overall eigenvalue is greater than or equal to the preset feature contribution rate and the minimum threshold; if so, the feature contribution rate and the corresponding multiple eigenvalues are selected as the target eigenvalue K; if not, the operation of the above-mentioned sum calculation submodule is returned until the ratio of the sum of the feature contribution rate and the overall eigenvalue is greater than or equal to the preset feature contribution rate and the minimum threshold; The credibility is calculated as follows: Where β is the learning rate, which is used to balance the influence of support and confidence; λ is the control parameter, which is used to adjust the influence of learning rate on fitness.
2. The oil and gas valve safety control and monitoring system according to claim 1 is characterized in that: The calculation method of the anomaly assessment function is: In the formula, R(a) is the abnormal state score, h(a s ) is the abnormal scoring function of the sth pipeline corrosion monitoring parameter of the pipeline corrosion monitoring data to be tested, and h(a s )>0, N is the number of pipeline corrosion monitoring parameter types of pipeline corrosion monitoring data to be tested, k(a t ) is the credibility function; Wherein, the credibility function k(a t ) is calculated as: In the formula, for a s with a t There is a correlation relationship, and the pipeline corrosion monitoring data to be tested is abnormal pipeline corrosion monitoring data; f t is the credibility, 0≤f t ≤1.
3. The oil and gas valve safety control and monitoring system according to claim 2 is characterized in that: The multiple pipeline corrosion monitoring parameters of the multiple evaluation dimensions refer to at least two evaluation dimensions, and each evaluation dimension includes at least two pipeline corrosion monitoring parameters; the pipeline corrosion monitoring data information set is the data obtained by detecting the pipeline corrosion monitoring data to be tested.
4. The oil and gas valve safety control and monitoring system according to claim 3 is characterized in that: The credibility represents the credibility of the abnormal data association relationship.
5. The oil and gas valve safety control and monitoring system according to claim 4, characterized in that: The first abnormal data point P represents a data matrix, in which each element represents a first abnormal data point in the same dimension obtained by dimensionality reduction of various abnormal data points in multiple dimensions.
