Intelligent substation relay protection fault diagnosis information acquisition and preprocessing method

Abstract

The application discloses a smart substation relay protection fault diagnosis information acquisition and preprocessing method, and belongs to the technical field of smart substations, and comprises the following steps: acquiring smart substation relay protection multi-source data based on multiple sensors in real time, and performing consistency checking and correction on the acquired smart substation relay protection multi-source data; and performing preprocessing, safe storage and analysis and mining on the acquired smart substation relay protection multi-source data, and generating smart substation relay protection fault diagnosis characteristic data. The application solves the problem that the existing method cannot realize effective acquisition and preprocessing of smart substation relay protection fault diagnosis information, and reduces the operation and maintenance efficiency, reliability and safety of the smart substation. The application can realize effective acquisition and preprocessing of smart substation relay protection fault diagnosis information, and the accuracy and reliability of the smart substation relay protection fault diagnosis result are high, and the operation and maintenance efficiency, reliability and safety of the smart substation can be improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent substation, in particular to an intelligent substation relay protection fault diagnosis information acquisition and preprocessing method. BACKGROUND

[0002] The intelligent substation adopts reliable, economical, integrated, low-carbon, and environmentally friendly equipment and design, with full-station information digitization, communication platform networking, information sharing standardization, system function integration, compact structure design, high-voltage equipment intelligence, and operation state visualization as basic requirements, which can support real-time online analysis and control decision of the power grid, thereby improving the reliability and economy of the entire power grid operation.

[0003] Among them, the intelligent substation relay protection is an important measure to detect faults or abnormal conditions in the power system, thereby issuing an alarm signal, or directly isolating and removing the fault part, to reduce or avoid damage to equipment and impact on adjacent areas.

[0004] The existing intelligent substation cannot effectively acquire and preprocess the intelligent substation relay protection fault diagnosis information during operation, resulting in low accuracy and reliability of the intelligent substation relay protection fault diagnosis result, and reducing the operation and maintenance efficiency, reliability, and safety of the intelligent substation. SUMMARY

[0005] The purpose of the present application is to provide an intelligent substation relay protection fault diagnosis information acquisition and preprocessing method, which can effectively acquire and preprocess the intelligent substation relay protection fault diagnosis information, make the accuracy and reliability of the intelligent substation relay protection fault diagnosis result higher, improve the operation and maintenance efficiency, reliability, and safety of the intelligent substation, and solve the problems raised in the background.

[0006] To achieve the above purpose, the present application provides the following technical scheme:

[0007] The intelligent substation relay protection fault diagnosis information acquisition and preprocessing method comprises:

[0008] Based on multiple sensors, the operation of the intelligent substation is monitored, and based on the communication protocol, the intelligent substation relay protection multi-source data is acquired in real time, and the acquired intelligent substation relay protection multi-source data is subjected to consistency verification and correction;

[0009] The acquired intelligent substation relay protection multi-source data is preprocessed, securely stored, and analyzed and mined to generate intelligent substation relay protection fault diagnosis feature data, which is used for fault diagnosis of the intelligent substation relay protection, to ensure the accuracy of the intelligent substation fault diagnosis and the reliability of the safe operation.

[0010] Preferably, the operation of the smart substation is monitored based on multiple sensors, and the multiple-source data of the relay protection of the smart substation is acquired in real time based on a communication protocol, and the following operations are performed:

[0011] High-precision intelligent sensors are deployed at key nodes of the smart substation.

[0012] The voltage, current, power, and temperature during the operation of the smart substation are monitored in real time based on the high-precision intelligent sensors, and the data between the high-precision intelligent sensors is synchronously transmitted based on the communication protocol of the IEC61850 standard, and the multiple-source data of the relay protection of the smart substation is acquired in real time.

[0013] Preferably, the consistency of the acquired multiple-source data of the relay protection of the smart substation is verified, and the following operations are performed:

[0014] The consistency of the multiple-source data of the relay protection of the smart substation is verified based on timestamps, and whether the timestamps of the multiple-source data of the relay protection of the smart substation are consistent is checked to ensure the time synchronization of the acquisition of the multiple-source data of the relay protection of the smart substation.

[0015] The acquired multiple-source data of the relay protection of the smart substation is compared based on the transmitted multiple-source data of the relay protection of the smart substation, the matching degree between the transmitted multiple-source data of the relay protection of the smart substation and the acquired multiple-source data of the relay protection of the smart substation is analyzed, and whether the acquired multiple-source data of the relay protection of the smart substation has consistency is determined.

[0016] When the transmitted multiple-source data of the relay protection of the smart substation matches the acquired multiple-source data of the relay protection of the smart substation, the acquired multiple-source data of the relay protection of the smart substation has consistency.

[0017] When the transmitted multiple-source data of the relay protection of the smart substation does not match the acquired multiple-source data of the relay protection of the smart substation, the acquired multiple-source data of the relay protection of the smart substation does not have consistency.

[0018] Preferably, the acquired multiple-source data of the relay protection of the smart substation is corrected, and the following operations are performed:

[0019] When the acquired multiple-source data of the relay protection of the smart substation does not have consistency, the acquired multiple-source data of the relay protection of the smart substation is checked one by one based on the transmitted multiple-source data of the relay protection of the smart substation, and missing data in the acquired multiple-source data of the relay protection of the smart substation is found.

[0020] The missing data found is filled into the obtained smart substation relay protection multi-source data, and consistency check is performed on the smart substation relay protection multi-source data after the missing data is filled in, so as to evaluate whether the missing data is correctly filled into the obtained smart substation relay protection multi-source data, and ensure that the obtained smart substation relay protection multi-source data has consistency.

[0021] Preferably, when the high-precision intelligent sensor is used to monitor the voltage, current, power and temperature in the operation process of the smart substation in real time, the sampling frequency is dynamically adjusted. In the case of rapid dynamic change of the signal, the sampling frequency is increased to capture more details. In the case of slow dynamic change of the signal, the sampling frequency is reduced to reduce data storage and processing operations.

[0022] Preferably, the obtained smart substation relay protection multi-source data is preprocessed, and the following operations are performed:

[0023] The obtained smart substation relay protection multi-source data is cleaned to remove null values in the smart substation relay protection multi-source data, identify missing or abnormal data in the smart substation relay protection multi-source data, and perform interpolation processing on the missing or abnormal data in the smart substation relay protection multi-source data to ensure the continuity of the smart substation relay protection multi-source data.

[0024] The smart substation relay protection multi-source data is filtered based on wavelet transform to eliminate harmonic noise in the smart substation relay protection multi-source data, and Kalman filter is used to compensate for drift errors of the high-precision intelligent sensor.

[0025] Preferably, the obtained smart substation relay protection multi-source data is securely stored, and the following operations are performed:

[0026] The obtained smart substation relay protection multi-source data is normalized to unify different dimensions of the smart substation relay protection multi-source data to the same scale, eliminate the dimensional differences between the smart substation relay protection multi-source data, and form standardized smart substation relay protection multi-source data.

[0027] The standardized smart substation relay protection multi-source data is integrated to integrate smart substation relay protection multi-source data from different sources into a unified data view, and the integrated smart substation relay protection multi-source data is subjected to integrity verification. After passing the verification, the integrated smart substation relay protection multi-source data is securely stored and managed using distributed storage technology.

[0028] Preferably, the obtained smart substation relay protection multi-source data is analyzed and mined to generate smart substation relay protection fault diagnosis feature data, and the following operations are performed:

[0029] The multi-source data of the intelligent substation relay protection is analyzed and mined based on principal component analysis, feature vectors related to intelligent substation relay protection fault diagnosis are extracted from the multi-source data of the intelligent substation relay protection, and intelligent substation relay protection fault diagnosis feature data are determined, including the fundamental wave effective value of the current and voltage before and after the fault, the positive sequence, negative sequence and zero sequence components of the current and voltage, the total harmonic distortion rate and the specific harmonic content;

[0030] According to the intelligent substation relay protection fault diagnosis feature data, the intelligent substation relay protection is diagnosed to ensure the accuracy of the intelligent substation fault diagnosis and the reliability of the safe operation.

[0031] Preferably, the abnormal data in the intelligent substation relay protection multi-source data is identified, including:

[0032] The intelligent substation relay protection multi-source data is taken as the to-be-identified data;

[0033] A device-data binary correlation weighted graph is constructed based on the to-be-identified data; wherein the nodes include relay protection device nodes and data nodes; the edges are the ownership correlation of the devices and data and the logical correlation between the data; the edge weight is determined by the product of the basic weight and the dynamic weight;

[0034] A first step size and a second step size are preset; the second step size is greater than the first step size;

[0035] For each data node, random walk is performed based on the first step size and the second step size, respectively, wherein the walk path of the first step size is to perform single-step walk starting from the data node, and the walk path of the second step size is to perform two-step walk starting from the data node;

[0036] The random walk is performed for each data node for several times to determine the multi-step random walk sequence of each data node;

[0037] The single-step loss function is calculated for each data node;

[0038] The loss functions of the first step size and the second step size are respectively assigned weights, and the weighted sum is obtained to obtain the joint loss function of each data node;

[0039] The joint loss functions of all data nodes are summed to obtain a target loss function; when the target loss function meets the requirements, the vectors of each data node are taken as low-dimensional embedding vectors;

[0040] The dynamic disturbance factor and the data correlation risk degree corresponding to the to-be-identified data are calculated;

[0041] The comprehensive abnormal index value corresponding to the to-be-identified data is determined based on the low-dimensional embedding vector, the dynamic disturbance factor and the data correlation risk degree;

[0042] The local reachable density of the neighborhood of each data node is calculated, and a ratio of a mean value of the local reachable density of the neighborhood nodes to the local reachable density of the neighborhood is taken as an abnormal evaluation value; if the abnormal evaluation value is greater than or equal to a preset abnormal evaluation threshold, the data node is determined as abnormal data, and all the abnormal data constitute a to-be-cleaned abnormal data set.

[0043] Preferably, after the consistency check and correction of the acquired smart substation relay protection multi-source data, the method further comprises: performing data quality evaluation on the smart substation relay protection multi-source data after the consistency check and correction, and determining a data quality evaluation value; comparing the data quality evaluation value with a preset data quality evaluation threshold, and when it is determined that the data quality evaluation value is less than the preset data quality evaluation threshold, reacquiring the smart substation relay protection multi-source data.

[0044] The data quality evaluation on the smart substation relay protection multi-source data after the consistency check and correction comprises:

[0045] The acquired smart substation relay protection multi-source data is divided based on time sequence, and smart substation relay protection multi-source sub-data corresponding to each time point in the time sequence is taken as a data point.

[0046] The reliability of each data point in different dimensions is calculated.

[0047] ;

[0048] The reliability of each data point in different dimensions is calculated. represents the reliability of the i th data point in the j th quality dimension; represents the normalized measurement value of the i th data point in the j th dimension; represents the average reference value of the same type data as the i th data point in the j th dimension; represents the value of the k th neighborhood data point of the i th data point in the j th dimension; represents the standard deviation of the j th dimension; N represents the total number of neighborhood data points.

[0049] Based on the reliability of each data point in different dimensions and the dimension weight, the comprehensive quality of each data point is calculated.

[0050] ;

[0051] The reliability of each data point in different dimensions is calculated. represents the comprehensive data quality score of the i th data point; represents the weight of the j th quality dimension;

[0052] Based on the comprehensive quality of each data point, the data quality evaluation value of the smart substation relay protection multi-source data is calculated.

[0053] ;

[0054] Wherein, The data quality evaluation value of the multi-source data of the intelligent substation relay protection is represented by M, and M represents the total number of data points. The data integrity coverage rate of the multi-source data of the intelligent substation relay protection is represented by M.

[0055] Compared with the prior art, the beneficial effects of the present application are:

[0056] The present application monitors the operation of the intelligent substation through multiple sensors, obtains the multi-source data of the intelligent substation relay protection in real time based on the communication protocol, and performs consistency checking and correction on the obtained multi-source data of the intelligent substation relay protection. The multi-source data of the intelligent substation relay protection is preprocessed, securely stored and analyzed and mined to generate intelligent substation relay protection fault diagnosis feature data for fault diagnosis of the intelligent substation relay protection, ensuring the accuracy and safety of the intelligent substation fault diagnosis and the reliability of the intelligent substation operation. The present application can realize effective acquisition and preprocessing of intelligent substation relay protection fault diagnosis information, and can improve the operation efficiency, reliability and safety of the intelligent substation. BRIEF DESCRIPTION OF DRAWINGS

[0057] Figure 1 The flowchart of the intelligent substation relay protection fault diagnosis information acquisition and preprocessing method of the present application. DETAILED DESCRIPTION

[0058] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0059] In order to solve the problem that the existing intelligent substation cannot realize effective acquisition and preprocessing of intelligent substation relay protection fault diagnosis information during operation, resulting in low accuracy and reliability of intelligent substation relay protection fault diagnosis results, and reducing the operation efficiency, reliability and safety of the intelligent substation, please refer to Figure 1 The present embodiment provides the following technical solutions:

[0060] The intelligent substation relay protection fault diagnosis information acquisition and preprocessing method comprises:

[0061] The operation of the smart substation is monitored based on multiple sensors, and the multiple-source data of the relay protection of the smart substation is obtained in real time based on a communication protocol, and consistency checking and correction are performed on the obtained multiple-source data of the relay protection of the smart substation.

[0062] In this embodiment, the operation of the smart substation is monitored based on multiple sensors, the multiple-source data of the relay protection of the smart substation is obtained in real time based on a communication protocol, and the following operations are performed:

[0063] High-precision intelligent sensors are deployed at key nodes of the smart substation.

[0064] The voltage, current, power and temperature during the operation of the smart substation are monitored in real time based on high-precision intelligent sensors, and the data between the high-precision intelligent sensors are synchronized and transmitted based on the communication protocol of the IEC61850 standard, so that the multiple-source data of the relay protection of the smart substation are obtained in real time, and the real-time and accurate collection of the multiple-source data of the relay protection of the smart substation is realized.

[0065] Specifically, the voltage, current, power and temperature during the operation of the smart substation are monitored in real time based on high-precision intelligent sensors, and the obtained multiple-source data of the relay protection of the smart substation are as shown in Table 1:

[0066] Table 1: Multiple-source data of relay protection of smart substation

[0067]

[0068] Therefore, by monitoring the operation of the smart substation through multiple sensors, the multiple-source data of the relay protection of the smart substation can be obtained, which facilitates better diagnosis of the relay protection fault of the smart substation.

[0069] In this embodiment, in order to ensure the reliability and integrity of the multiple-source data of the relay protection of the smart substation, consistency checking and correction processing are performed on the obtained multiple-source data of the relay protection of the smart substation, and the following operations are performed:

[0070] The synchronization of the multiple-source data of the relay protection of the smart substation is verified based on the time stamp, the time stamp of the multiple-source data of the relay protection of the smart substation is checked to ensure the time synchronization of the collection of the multiple-source data of the relay protection of the smart substation;

[0071] According to the transmitted multiple-source data of the relay protection of the smart substation, the obtained multiple-source data of the relay protection of the smart substation are compared, the matching degree between the transmitted multiple-source data of the relay protection of the smart substation and the obtained multiple-source data of the relay protection of the smart substation is analyzed, and it is judged whether the obtained multiple-source data of the relay protection of the smart substation has consistency;

[0072] When the transmitted smart substation relay protection multi-source data matches the acquired smart substation relay protection multi-source data, the acquired smart substation relay protection multi-source data has consistency;

[0073] When the transmitted smart substation relay protection multi-source data does not match the acquired smart substation relay protection multi-source data, the acquired smart substation relay protection multi-source data does not have consistency;

[0074] When the acquired smart substation relay protection multi-source data does not have consistency, the acquired smart substation relay protection multi-source data is checked one by one based on the transmitted smart substation relay protection multi-source data to find missing data in the acquired smart substation relay protection multi-source data;

[0075] The found missing data is supplemented into the acquired smart substation relay protection multi-source data, and the smart substation relay protection multi-source data after the missing data is supplemented is subjected to consistency verification to evaluate whether the missing data is correctly supplemented into the acquired smart substation relay protection multi-source data, so as to ensure that the acquired smart substation relay protection multi-source data has consistency.

[0076] Specifically, the acquired smart substation relay protection multi-source data is subjected to consistency verification and correction processing, and the consistency verification and correction processing conditions are shown in Table 2:

[0077] Table 2: Consistency verification and correction processing conditions

[0078]

[0079] Therefore, by subjecting the acquired smart substation relay protection multi-source data to consistency verification and correction processing, the reliability and integrity of the smart substation relay protection multi-source data can be ensured.

[0080] In this embodiment, when the voltage, current, power and temperature conditions in the operation process of the smart substation are monitored in real time based on the high-precision intelligent sensor, the sampling frequency is dynamically adjusted, the sampling frequency is increased in the case of fast dynamic change of the signal, and more details are captured, and the sampling frequency is reduced in the case of slow dynamic change of the signal, and data storage and processing operations are reduced.

[0081] The acquired smart substation relay protection multi-source data is preprocessed, securely stored and analyzed and mined to generate smart substation relay protection fault diagnosis feature data, which is used for fault diagnosis of the smart substation relay protection, and ensures the fault diagnosis accuracy and safe operation reliability of the smart substation.

[0082] In this embodiment, the acquired smart substation relay protection multi-source data is preprocessed, and the following operations are performed:

[0083] The acquired smart substation relay protection multi-source data is cleaned, the null values in the smart substation relay protection multi-source data are removed, the missing or abnormal data in the smart substation relay protection multi-source data is identified, and the missing or abnormal data in the smart substation relay protection multi-source data is interpolated, to ensure the continuity of the smart substation relay protection multi-source data.

[0084] The smart substation relay protection multi-source data is filtered based on wavelet transform, to eliminate harmonic noise in the smart substation relay protection multi-source data, and Kalman filter is used to compensate the drift error of high-precision sensors.

[0085] In this embodiment, the acquired smart substation relay protection multi-source data is stored safely, and the following operations are performed:

[0086] The acquired smart substation relay protection multi-source data is normalized, to unify the smart substation relay protection multi-source data of different dimensions to the same scale, eliminate the dimensional differences between the smart substation relay protection multi-source data, and form standardized smart substation relay protection multi-source data.

[0087] The standardized smart substation relay protection multi-source data is integrated, to integrate the smart substation relay protection multi-source data of different sources into a unified data view, and perform integrity verification on the integrated smart substation relay protection multi-source data. After passing the verification, the integrated smart substation relay protection multi-source data is stored and managed safely by using distributed storage technology.

[0088] In this embodiment, the acquired smart substation relay protection multi-source data is analyzed and mined, to generate smart substation relay protection fault diagnosis feature data, and the following operations are performed:

[0089] The smart substation relay protection multi-source data is analyzed and mined based on principal component analysis, to extract feature vectors related to smart substation relay protection fault diagnosis from the smart substation relay protection multi-source data, and determine the smart substation relay protection fault diagnosis feature data, including the fundamental effective value of current and voltage before and after the fault, positive sequence, negative sequence and zero sequence components of current and voltage, total harmonic distortion rate, and specific harmonic content.

[0090] Specifically, the fundamental effective values of the current and voltage before and after the fault are the most basic features for judging the severity and type of the fault; the positive sequence, negative sequence and zero sequence components of the current and voltage are the decisive features for distinguishing the fault type, for example, the zero sequence current is the key to judging the ground fault; the total harmonic distortion rate is the distortion degree of the current and voltage, for example, the arc fault will cause the harmonic distortion rate to rise significantly; the specific harmonic content is the amplitude of a specific harmonic, and the extracted feature vector can effectively diagnose the fault of the intelligent substation relay protection.

[0091] According to the intelligent substation relay protection fault diagnosis feature data, the intelligent substation relay protection is diagnosed, and the fault diagnosis accuracy and safe operation reliability of the intelligent substation are ensured.

[0092] In this embodiment, the abnormal data in the multi-source data of the intelligent substation relay protection is identified, including:

[0093] The multi-source data of the intelligent substation relay protection is taken as the to-be-identified data;

[0094] A device-data binary correlation weighted graph is constructed based on the to-be-identified data; wherein the nodes include relay protection device nodes and data nodes; the edges are the ownership correlation of the devices and data and the logical correlation between the data; the edge weight is determined by the product of the basic weight and the dynamic weight;

[0095] A first step size and a second step size are preset; the second step size is greater than the first step size;

[0096] For each data node, random walk is performed based on the first step size and the second step size, respectively, wherein the walk path of the first step size is to perform single-step walk starting from the data node, and the walk path of the second step size is to perform two-step walk starting from the data node;

[0097] The random walk is performed for each data node for several times to determine the multi-step random walk sequence of each data node;

[0098] The single-step loss function is calculated for each data node;

[0099] The loss functions of the first step size and the second step size are respectively assigned weights, and the weighted sum is obtained to obtain the joint loss function of each data node;

[0100] The joint loss functions of all data nodes are summed to obtain a target loss function; when the target loss function meets the requirements, the vectors of the data nodes are taken as low-dimensional embedding vectors;

[0101] The dynamic disturbance factor and the data correlation risk degree corresponding to the to-be-identified data are calculated;

[0102] The comprehensive abnormal index value corresponding to the to-be-identified data is determined based on the low-dimensional embedding vector, the dynamic disturbance factor and the data correlation risk degree.

[0103] The local reachable density of each data node is calculated, and the ratio of the average of the local reachable density of the neighborhood nodes to the local reachable density of the neighborhood nodes is taken as the abnormal evaluation value. If the abnormal evaluation value is greater than or equal to a preset abnormal evaluation threshold, the data node is determined to be abnormal data, and all abnormal data constitute a set of abnormal data to be cleaned.

[0104] In this embodiment, the relay protection device node includes a protection device, a mutual inductor, and a switch; the data node includes a sampling value, a state signal, a message data, and an environmental data in the data to be identified; the ownership association of the device and the data is, for example, the protection device and the corresponding sampling value; the logical association between the data is, for example, the closing signal and the closing GOOSE message; the edge weight W is determined by the product of the basic weight and the dynamic weight: in the basic weight, the weight corresponding to the sampling value and the state signal is 0.8, and the weight corresponding to the message data and the environmental data is 0.3; the dynamic weight is the historical association integrity of the data and the device, and the dynamic weight is increased by 0.1 for each 10% reduction in the data loss rate, and the weight value range is [0.5, 1.0].

[0105] In this embodiment, a single-step loss function is calculated for each data node, ; (k = 1, 2) is the step length, u is the starting data node, is the kth step walk sequence node set, is the walk probability; the walk probability of each step is positively correlated with the edge weight.

[0106] In this embodiment, the dynamic disturbance factor of multi-source data is calculated ; represents the standard deviation within a period of a stationary sampling sequence; represents the standard deviation within a period of a stationary sampling sequence under normal working conditions; represents the transmission delay of effective message data; represents the message delay threshold, ; represents the normalized temperature.

[0107] In this embodiment, the data association risk degree is calculated ; wherein represents the risk weight corresponding to the intersection event; represents the number of intersection events of real-time data and historical abnormal data; represents the number of union events of real-time data and historical abnormal data; n represents the total number of events associated with the historical abnormal data;

[0108] In this embodiment, a comprehensive abnormal index R is constructed, ; wherein S is the average cosine similarity of the low-dimensional embedding vector of the data node and the low-dimensional embedding vector of the historical abnormal data node.

[0109] In this embodiment, the risk-weighted distance is defined as ; represents the low-dimensional embedding vector of data nodes u, v; The comprehensive anomaly index of data nodes u, v; 1.2 is the risk weight coefficient, and the risk-weighted distance is used as the distance measure when calculating the local density reachable.

[0110] The working principle and beneficial effects of the above technical solution are: the multi-source data of the intelligent substation relay protection has the characteristics of multi-source heterogeneity, strong correlation, high noise and dynamic time variation. The present scheme maps the heterogeneous data into a graph structure through the backup-data binary correlation weighted graph, and the edge weight is fused with the basic weight and the dynamic weight. The sampling value / state signal and the message / environment data are differentiated modeling, avoiding the high sampling rate data from drowning the key logical information, and the dynamic weight is self-adaptive to the working condition change, ensuring that the normal fluctuation data is not misjudged. The traditional method needs to manually design multiple sets of rules to process different data types, and the present scheme realizes one-time modeling and unified processing through the graph structure, reducing the cleaning complexity. The random walk path is designed as data node-association node, which excavates the deep association of equipment-data-data. The risk-weighted distance integrates the comprehensive anomaly index R into the distance measure, and strengthens the influence of the neighborhood of high-risk data. The associated abnormality which is easily missed by the traditional method is successfully identified, and the false positive rate is reduced. The detection rate of the logical break abnormality is improved. The problem that the multi-source data correlation abnormality accounts for a large proportion of the relay protection faults but cannot be identified is solved. The present scheme realizes the associated chain detection through the graph random walk, improving the detection rate of abnormal data identification. The dynamic disturbance factor quantifies the comprehensive influence of data fluctuation, communication delay and environmental disturbance. The comprehensive anomaly index dynamically weights the similarity, disturbance and risk, and adapts to the working condition drift. In the load mutation scene, the false deletion rate is reduced. The data offset caused by the change of environmental temperature is automatically corrected, and the cleaning stability is improved. The present scheme realizes adaptive cleaning and reduces the operation and maintenance manpower investment, solving the problem that the traditional method needs frequent manual parameter adjustment due to the strong dynamic nature of the substation environment.

[0111] In this embodiment, after the consistency check and correction of the obtained multi-source data of the intelligent substation relay protection, the method further comprises: performing data quality evaluation on the consistency checked and corrected multi-source data of the intelligent substation relay protection, determining a data quality evaluation value; comparing the data quality evaluation value with a preset data quality evaluation threshold value, and when it is determined that the data quality evaluation value is less than the preset data quality evaluation threshold value, reacquiring the multi-source data of the intelligent substation relay protection;

[0112] The data quality evaluation on the consistency checked and corrected multi-source data of the intelligent substation relay protection comprises:

[0113] The obtained smart substation relay protection multi-source data is divided based on time sequence, and the smart substation relay protection multi-source data corresponding to each time point in the time sequence is taken as a data point;

[0114] The credibility of each data point in different dimensions is calculated;

[0115] ;

[0116] wherein, represents the credibility of the i th data point in the j th quality dimension; represents the normalized measurement value of the i th data point in the j th dimension; represents the average reference value of the data of the same type as the i th data point in the j th dimension; represents the value of the k th neighbor data point of the i th data point in the j th dimension; represents the standard deviation of the j th dimension; N represents the total number of neighbor data points;

[0117] Based on the credibility of each data point in different dimensions and the dimension weight, the comprehensive quality of each data point is calculated;

[0118] ;

[0119] wherein, represents the comprehensive data quality score of the i th data point; represents the weight of the j th quality dimension;

[0120] Based on the comprehensive quality of each data point, the data quality evaluation value of the smart substation relay protection multi-source data is calculated;

[0121] ;

[0122] wherein, represents the data quality evaluation value of the smart substation relay protection multi-source data; M represents the total number of data points; represents the data integrity coverage rate of the smart substation relay protection multi-source data.

[0123] In this embodiment, the data quality evaluation dimensions of the smart substation relay protection multi-source data include integrity dimension, accuracy dimension, timeliness dimension and consistency dimension; the integrity dimension ; the accuracy dimension ; the timeliness dimension ; the consistency dimension ; all are mapped to [0, 1] through normalization; 0 represents the worst quality, and 1 represents the best quality.

[0124] In this embodiment, .

[0125] The working principle and beneficial effects of the above technical solution are: the comprehensive quality is calculated based on the credibility of each data point in different dimensions and the dimension weight, the information of multiple quality dimensions is comprehensively considered, and the limitation of single dimension evaluation is avoided; different quality dimensions may reflect the characteristics of data in different aspects, such as integrity, accuracy, timeliness, etc., and through comprehensive calculation, the quality of data points can be more comprehensively evaluated, so that the evaluation of data reliability is more accurate; the data quality evaluation value is calculated in combination with the comprehensive quality of all data points and the data integrity coverage rate, which further measures the quality of the multi-source data of the intelligent substation relay protection from the whole; the introduction of the data integrity coverage rate ensures the integrity of the data as a whole, so that the evaluation result is more reliable; the data quality evaluation value is compared with the preset threshold value, and the data is reacquired when the evaluation value is less than the threshold value. This dynamic monitoring mechanism can timely find the situation of data quality decline and take corresponding measures to correct it, so as to ensure that the data used by the intelligent substation relay protection system always has high quality.

[0126] In summary, by monitoring the operation of the intelligent substation through multiple sensors, real-time acquisition of the multi-source data of the intelligent substation relay protection based on the communication protocol, consistency verification and correction of the acquired multi-source data of the intelligent substation relay protection, pre-processing, safe storage and analysis and mining of the acquired multi-source data of the intelligent substation relay protection, generation of intelligent substation relay protection fault diagnosis feature data for fault diagnosis of the intelligent substation relay protection, and ensuring the accuracy of fault diagnosis and the reliability of safe operation of the intelligent substation, the effective acquisition and preprocessing of intelligent substation relay protection fault diagnosis information can be realized, the accuracy and reliability of the intelligent substation relay protection fault diagnosis result are high, and the operation and maintenance efficiency, reliability and safety of the intelligent substation can be improved.

[0127] It should be noted that in this paper, relationship terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between the entities or operations. Moreover, the terms "include", "contain" or any other variant thereof are intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or equipment.

[0128] While embodiments of the application have been shown and described, it is to be understood that the embodiments described are merely exemplary of the principles and application of the present application. Numerous modifications and adaptions can be effected without departing from the spirit and scope of the present application, which is not limited to the exact construction and arrangement described. It is intended, therefore, to cover all modifications and adaptions that fall within the scope of the claims and their equivalents.

Claims

1. A method for acquiring and preprocessing fault diagnosis information of relay protection in intelligent substations, characterized in that, include: The system monitors the operation of smart substations using multiple sensors, acquires multi-source data on smart substation relay protection in real time based on communication protocols, and performs consistency verification and correction on the acquired multi-source data on smart substation relay protection. The acquired multi-source data of intelligent substation relay protection is preprocessed, securely stored, and analyzed to generate fault diagnosis feature data of intelligent substation relay protection, which is used for fault diagnosis of intelligent substation relay protection. After performing consistency verification and correction on the acquired multi-source data of smart substation relay protection, the method further includes: performing data quality assessment on the consistency verification and correction data of smart substation relay protection, and determining the data quality assessment value; comparing the data quality assessment value with a preset data quality assessment threshold, and when it is determined that the data quality assessment value is less than the preset data quality assessment threshold, re-acquiring the multi-source data of smart substation relay protection. The data quality assessment of the multi-source data for smart substation relay protection after consistency verification and correction includes: The acquired multi-source data of smart substation relay protection is divided based on time sequence, and the multi-source sub-data of smart substation relay protection corresponding to each time point in the time sequence is taken as a data point; Calculate the credibility of each data point across different dimensions; ； in, This represents the confidence level of the i-th data point in the j-th quality dimension. This represents the normalized measurement value of the i-th data point in the j-th dimension; This represents the average reference value of data of the same type as the i-th data point in the j-th dimension; This represents the value of the k-th neighboring data point of the i-th data point in the j-th dimension; The j-th dimension represents the standard deviation; N represents the total number of neighboring data points. The overall quality of each data point is calculated based on its credibility and dimensional weights across different dimensions. ； in, This represents the overall data quality score for the i-th data point; This represents the weight of the j-th quality dimension; The data quality assessment value of multi-source data for relay protection in smart substations is calculated based on the comprehensive quality of each data point. ； Wherein, DQI represents the data quality assessment value of multi-source data for relay protection in smart substations; M represents the total number of data points; This indicates the data integrity coverage rate of multi-source data for relay protection in intelligent substations.

2. The method for acquiring and preprocessing fault diagnosis information of intelligent substation relay protection according to claim 1, characterized in that, Based on multi-sensor monitoring of the operation of smart substations, and based on communication protocols, real-time acquisition of multi-source data from smart substation relay protection is performed to carry out the following operations: Deploy high-precision intelligent sensors at key nodes in smart substations; Based on high-precision intelligent sensors, the voltage, current, power and temperature of the intelligent substation are monitored in real time. Based on the communication protocol of IEC61850 standard, data synchronous transmission between high-precision intelligent sensors is realized, and multi-source data of relay protection of intelligent substation are acquired in real time.

3. The method for acquiring and preprocessing fault diagnosis information of intelligent substation relay protection according to claim 2, characterized in that, Perform consistency verification on the acquired multi-source data of smart substation relay protection, and perform the following operations: The synchronization of multi-source data for relay protection in smart substations is verified based on timestamps. This checks whether the timestamps of the multi-source data for relay protection in smart substations are consistent, ensuring the time synchronization of the multi-source data acquisition for relay protection in smart substations. Based on the transmitted multi-source data of intelligent substation relay protection, the acquired multi-source data of intelligent substation relay protection is compared, the degree of matching between the transmitted and acquired multi-source data of intelligent substation relay protection is analyzed, and the consistency of the acquired multi-source data of intelligent substation relay protection is determined. When the transmitted multi-source data of intelligent substation relay protection matches the acquired multi-source data of intelligent substation relay protection, the acquired multi-source data of intelligent substation relay protection is consistent. When the transmitted multi-source data of intelligent substation relay protection does not match the acquired multi-source data of intelligent substation relay protection, the acquired multi-source data of intelligent substation relay protection is inconsistent.

4. The method for acquiring and preprocessing fault diagnosis information of intelligent substation relay protection according to claim 3, characterized in that, To correct the acquired multi-source data of smart substation relay protection, perform the following operations: When the acquired multi-source data of smart substation relay protection is inconsistent, the acquired multi-source data of smart substation relay protection is checked one by one based on the transmitted multi-source data of smart substation relay protection to find the missing data in the acquired multi-source data of smart substation relay protection. The missing data is then added to the acquired multi-source data of smart substation relay protection. A consistency check is performed on the multi-source data of smart substation relay protection after the missing data is added to evaluate whether the missing data has been correctly added to the acquired multi-source data of smart substation relay protection.

5. The method for acquiring and preprocessing fault diagnosis information of intelligent substation relay protection according to claim 4, characterized in that, When using high-precision intelligent sensors to monitor voltage, current, power, and temperature during the operation of intelligent substations in real time, the sampling frequency is dynamically adjusted.

6. The method for acquiring and preprocessing fault diagnosis information of intelligent substation relay protection according to claim 5, characterized in that, The acquired multi-source data of smart substation relay protection is preprocessed by performing the following operations: The acquired multi-source data of intelligent substation relay protection is cleaned to remove null values, identify missing or abnormal data, and perform interpolation processing on the missing or abnormal data to ensure the continuity of the multi-source data of intelligent substation relay protection. Wavelet transform is used to filter multi-source data of relay protection in smart substations to eliminate harmonic noise, and Kalman filtering is used to compensate for drift error of high-precision smart sensors.

7. The method for acquiring and preprocessing fault diagnosis information of intelligent substation relay protection according to claim 6, characterized in that, Securely store the acquired multi-source data from smart substation relay protection and perform the following operations: The acquired multi-source data of intelligent substation relay protection is normalized to unify the multi-source data of intelligent substation relay protection with different dimensions to the same scale, eliminate the difference in dimensions between the multi-source data of intelligent substation relay protection, and form standardized multi-source data of intelligent substation relay protection. Standardized multi-source data for smart substation relay protection is integrated, combining multi-source data from different sources into a unified data view. The integrity of the integrated multi-source data for smart substation relay protection is verified. After successful verification, distributed storage technology is used to securely store and manage the integrated multi-source data for smart substation relay protection.

8. The method for acquiring and preprocessing fault diagnosis information of intelligent substation relay protection according to claim 7, characterized in that, The acquired multi-source data on relay protection in smart substations is analyzed and mined to generate fault diagnosis feature data for relay protection in smart substations. The following operations are performed: Principal component analysis is used to analyze and mine multi-source data of relay protection in smart substations. Feature vectors related to fault diagnosis of relay protection in smart substations are extracted from the multi-source data of relay protection in smart substations. The characteristic data of fault diagnosis of relay protection in smart substations are determined, including the fundamental effective value of current and voltage before and after the fault, the positive sequence, negative sequence and zero sequence components of current and voltage, total harmonic distortion rate, and specific harmonic content. Fault diagnosis of relay protection in intelligent substations is performed based on fault diagnosis feature data of relay protection in intelligent substations.

9. The method for acquiring and preprocessing fault diagnosis information of intelligent substation relay protection according to claim 6, characterized in that, Identify abnormal data in the multi-source data of relay protection in smart substations, including: The multi-source data of relay protection in intelligent substations are used as the data to be identified. A device-data binary association weighted graph is constructed based on the data to be identified; wherein, the nodes include relay protection device nodes and data nodes; the edges represent the attribution association between devices and data and the logical association between data; the edge weights are determined by the product of the basic weights and the dynamic weights. The first step length and the second step length are preset; the second step length is greater than the first step length. For each data node, a random walk is performed based on the first step length and the second step length, respectively. The walk path with the first step length is a single-step walk starting from the data node, and the walk path with the second step length is a two-step walk starting from the data node. Perform several random walks on each data node to determine the multi-step random walk sequence for each data node; Calculate the single-step loss function for each data node; Weights are assigned to the loss functions of the first and second steps respectively, and the weighted sum is used to obtain the joint loss function for each data node; The target loss function is obtained by summing the joint loss functions of all data nodes; when the target loss function meets the requirements, the vectors of each data node are used as low-dimensional embedding vectors. Calculate the dynamic disturbance factor and data association risk level corresponding to the data to be identified; The comprehensive anomaly index value corresponding to the data to be identified is determined based on low-dimensional embedding vectors, dynamic perturbation factors, and data association risk. Calculate the local reachability density of each data node in its neighborhood. Use the ratio of the mean of the local reachability density of the neighborhood nodes to the local reachability density of the neighborhood nodes as the anomaly evaluation value. If the anomaly evaluation value is greater than or equal to the preset anomaly evaluation threshold, the data node is determined to be an anomaly data. All anomaly data constitute the set of anomaly data to be cleaned.

Images