Online oil chromatography reliability early warning system

Abstract

The invention discloses an on-line oil chromatography reliability early warning system. The system comprises a data acquisition unit used for acquiring and storing monitoring data to form a historical database; the data processing unit is used for uniformly processing the data; the abnormal data extraction unit is used for counting abnormal working data of the monitoring device; the abnormal data analysis unit is used for performing clustering analysis on the abnormal data; the abnormal data identification unit is used for carrying out abnormal data identification on the current monitoring data; the comparison and judgment unit is used for comparing the current abnormal data with the abnormal characteristics; the early warning unit is used for giving out early warning on the result; the abnormal data can be extracted and analyzed, the abnormal data feature library can be constructed, the current monitoring data can be identified and can be compared and analyzed with the abnormal data feature library, the abnormal data can be quickly found, reliability early warning can be given out, and the accuracy and reliability of the monitoring device can be effectively improved.

Description

Technical Field

[0001] This invention belongs to the field of online oil chromatography technology, and specifically relates to an online oil chromatography reliability early warning system. Background Technology

[0002] In power systems, oil-immersed equipment such as transformers and instrument transformers are important components for ensuring stable power transmission and grid operation. As key equipment in the power grid, the operating status of oil-immersed transformers and instrument transformers is directly related to the safety and reliability of the entire power system. As the service life of these devices continues to extend, problems such as aging and decreased insulation performance gradually emerge, further increasing the risk of failure.

[0003] Dissolved gas analysis (DGA) technology is one of the most widely used methods for fault diagnosis of transformers and instrument transformers. By detecting changes in the composition and content of dissolved gases in transformer and instrument transformer oil, potential faults such as insulation aging, local overheating, and partial discharge within the equipment can be effectively identified. Online DGA monitoring devices mainly include oil-gas separation modules, gas separation modules, gas detection modules, and data analysis modules. In practical use, the online monitoring device itself may be affected by environmental conditions and aging over time, which can easily lead to malfunctions and a large amount of abnormal data. Data distortion caused by these malfunctions can interfere with transformer anomaly identification and fault diagnosis, significantly reducing the accuracy of anomaly identification and potentially leading to misjudgments of the main equipment status or malfunctions in equipment protection systems, ultimately causing power grid safety accidents.

[0004] Therefore, it is necessary to develop an online oil chromatography reliability early warning system to solve the above problems. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide an online oil chromatography reliability early warning system. This system extracts and analyzes abnormal data, constructs an abnormal data feature library, identifies the current monitoring data and compares and analyzes it with the abnormal data feature library, thereby enabling rapid detection of abnormal data and issuing reliability early warnings, effectively improving the accuracy and reliability of the monitoring device.

[0006] The objective of this invention is achieved as follows: an online oil chromatography reliability early warning system, comprising a data acquisition unit, a data processing unit, an abnormal data extraction unit, an abnormal data analysis unit, an abnormal data identification unit, a comparison and judgment unit, and an early warning unit;

[0007] The data acquisition unit is used to collect and store the online monitoring data of the online oil chromatography monitoring device in real time to form a historical database.

[0008] The data processing unit is used to perform unified processing on the data in the historical database, including data standardization.

[0009] The abnormal data extraction unit is used to combine historical databases and the monitoring principle of online oil chromatography devices to statistically analyze abnormal operating data of the monitoring devices and form an abnormal device database.

[0010] The abnormal data analysis unit is used to perform cluster analysis on abnormal data based on the device abnormal database and construct an abnormal data feature library;

[0011] The abnormal data identification unit is used to identify abnormal data in the current monitoring data of the online oil chromatography monitoring device.

[0012] The comparison and judgment unit is used to compare the abnormal data in the current data identified by the abnormal data identification unit with the abnormal feature library. When the abnormal data matches a certain type of abnormal feature in the abnormal feature library, the monitoring device is judged to be abnormal; otherwise, the transformer is faulty.

[0013] The early warning unit is used to issue early warnings based on the judgment results of the comparison and judgment unit, including reliability early warnings and transformer fault early warnings. The reliability early warning is due to a reliability failure of the monitoring device, or an abnormal monitoring status of the monitoring device causing data distortion. The transformer fault early warning is due to a normal monitoring status of the monitoring device, or a transformer failure.

[0014] Furthermore, the data processing unit performs data standardization processing, specifically in the following manner: Assuming m characteristic gases, and each characteristic gas sample data contains n sample data, a multidimensional data matrix is ​​obtained. The standardized data is then represented as: In the formula: , These represent the raw data and the standardized data, respectively. express The mean, express The variance is then used to obtain the standardized sample data matrix. .

[0015] Furthermore, the anomaly data analysis unit specifically employs the K-means clustering algorithm to perform cluster analysis on the anomaly data, which includes the following steps: based on the anomaly data feature library, constructing an anomaly data set D, and dividing the n data points in the data set D into K clusters. ① Select K initial cluster centers from the dataset D; ② Calculate the Euclidean distance from each data point in the dataset to the cluster centers, expressed as: In the formula: Let n represent two data points in space, respectively. express 1. Calculate the Euclidean distance between the data points; 2. Select the nearest cluster center and include it in the cluster; 3. After calculating all data points, recalculate the cluster center for each cluster; 4. Determine if the cluster center has changed. If it has changed, return to step 2. If it has not changed, output the clustering result.

[0016] Furthermore, the optimization objective of the K-means clustering algorithm in the abnormal data analysis unit is represented by minimizing the squared error of the clusters: In the formula: Representing a cluster Cluster centers.

[0017] Furthermore, the abnormal data in the abnormal data extraction unit and the abnormal data identification unit includes abnormal monitoring values, abnormal coefficients of variation, and abnormal data offset rates; wherein, the abnormal monitoring values ​​include singular values, null values, and negative values; the abnormal coefficient of variation is used to identify abnormal data fluctuations, specifically expressed as follows: In the formula: Represents the coefficient of variation. Indicates the standard deviation of the outlier interval. This represents the mean of the abnormal interval, with the anomaly coefficient threshold set to 30%. The data offset rate anomaly is used to identify the offset rate of abnormal intervals in the detection data of each characteristic gas, specifically expressed as follows: In the formula: Indicates the first Characteristic gas Sky offset rate, Indicates the first Characteristic gas Daily monitoring values Indicates the first Characteristic gases Daily average.

[0018] Furthermore, the comparison and judgment unit uses a specific correlation analysis method to compare abnormal data. The abnormal data sequence identified in the abnormal data identification unit is used as the comparison sequence, and the data sequence in the abnormal feature library is used as the reference sequence. The correlation between the comparison sequence and the reference sequence is then expressed as: , The amount of data in the reference sequence and the comparison sequence. Represents the correlation coefficient. Indicates the first The first group of comparison sequences with the reference sequence The difference between the data points express The minimum value, express The maximum value, This represents the resolution coefficient, which is set to 0.5 here.

[0019] Due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0020] (1) Real-time collection and storage of online monitoring data from the online oil chromatography monitoring device to form a historical database. Combining the historical database and the monitoring principle of the online oil chromatography device, statistical analysis of abnormal working data of the monitoring device is performed to form an abnormal device database. Abnormal data in the database are then clustered and analyzed to extract abnormal features for easy identification.

[0021] (2) By identifying abnormal data in the current online monitoring data and comparing and analyzing the abnormal features in the device abnormality database, device abnormalities can be quickly identified, early warnings can be issued, misjudgments can be avoided, and the reliability and accuracy of the monitoring device can be effectively improved. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of the principle of the present invention.

[0023] Figure 2 This is a flowchart of the present invention. Detailed Implementation

[0024] The technical solution of the present invention will be further described in detail below through embodiments and in conjunction with the accompanying drawings.

[0025] like Figure 1 , Figure 2 As shown, an online oil chromatography reliability early warning system includes a data acquisition unit, a data processing unit, an abnormal data extraction unit, an abnormal data analysis unit, an abnormal data identification unit, a comparison and judgment unit, and an early warning unit.

[0026] The data acquisition unit is used to collect and store the online monitoring data of the online oil chromatography monitoring device in real time to form a historical database.

[0027] The data processing unit is used to perform unified processing on the data in the historical database, including data standardization.

[0028] Preferably, the data processing unit performs data standardization processing in the following way: Assuming m characteristic gases, and each characteristic gas sample data contains n sample data, a multidimensional data matrix is ​​obtained. The standardized data is then represented as: In the formula: , These represent the raw data and the standardized data, respectively. express The mean, express The variance is then used to obtain the standardized sample data matrix. .

[0029] The abnormal data extraction unit is used to combine historical databases and the monitoring principle of online oil chromatography devices to statistically analyze abnormal operating data of the monitoring devices and form an abnormal device database.

[0030] The abnormal data analysis unit is used to perform cluster analysis on abnormal data based on the device abnormal database and construct an abnormal data feature library.

[0031] Preferably, the abnormal data analysis unit specifically employs the K-means clustering algorithm to perform cluster analysis on the abnormal data, which includes the following steps: constructing an abnormal data set D based on the abnormal data feature library, and dividing the n data points in the data set D into K clusters. ① Select K initial cluster centers from the dataset D; ② Calculate the Euclidean distance from each data point in the dataset to the cluster centers, expressed as: In the formula: Let n represent two data points in space, respectively. express 1. Calculate the Euclidean distance between the data points; 2. Select the nearest cluster center and include it in the cluster; 3. After calculating all data points, recalculate the cluster center for each cluster; 4. Determine if the cluster center has changed. If it has changed, return to step 2. If it has not changed, output the clustering result.

[0032] Preferably, the optimization objective of the K-means clustering algorithm in the abnormal data analysis unit is represented by minimizing the squared error of the clusters: In the formula: Representing a cluster Cluster centers.

[0033] The abnormal data identification unit is used to identify abnormal data in the current monitoring data of the online oil chromatography monitoring device.

[0034] Preferably, the abnormal data in the abnormal data extraction unit and the abnormal data identification unit includes abnormal monitoring values, abnormal coefficient of variation, and abnormal data offset rate; wherein, the abnormal monitoring values ​​include singular values, null values, and negative values; the abnormal coefficient of variation is used to identify abnormal data fluctuations, specifically expressed as follows: In the formula: Represents the coefficient of variation. Indicates the standard deviation of the outlier interval. This represents the mean of the abnormal interval, with the anomaly coefficient threshold set to 30%. The data offset rate anomaly is used to identify the offset rate of abnormal intervals in the detection data of each characteristic gas, specifically expressed as follows: In the formula: Indicates the first Characteristic gas Sky offset rate, Indicates the first Characteristic gas Daily monitoring values Indicates the first Characteristic gases Daily average.

[0035] The comparison and judgment unit is used to compare the abnormal data in the current data identified by the abnormal data identification unit with the abnormal feature library. When the abnormal data matches a certain type of abnormal feature in the abnormal feature library, the monitoring device is judged to be abnormal; otherwise, the transformer is faulty.

[0036] Preferably, the comparison and judgment unit uses a specific correlation analysis method to compare abnormal data, using the abnormal data sequence identified in the abnormal data identification unit as the comparison sequence and the data sequence in the abnormal feature library as the reference sequence. The correlation between the comparison sequence and the reference sequence is then expressed as: , The amount of data in the reference sequence and the comparison sequence. Represents the correlation coefficient. Indicates the first The first group of comparison sequences with the reference sequence The difference between the data points express The minimum value, express The maximum value, This represents the resolution coefficient, which is set to 0.5 here.

[0037] The early warning unit is used to issue early warnings based on the judgment results of the comparison and judgment unit, including reliability early warnings and transformer fault early warnings. The reliability early warning is due to a reliability failure of the monitoring device, or an abnormal monitoring status of the monitoring device causing data distortion. The transformer fault early warning is due to a normal monitoring status of the monitoring device, or a transformer failure.

[0038] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of the claims of the present invention.

Claims

1. An online oil chromatography reliability early warning system, characterized in that: It includes a data acquisition unit, a data processing unit, an abnormal data extraction unit, an abnormal data analysis unit, an abnormal data identification unit, a comparison and judgment unit, and an early warning unit; The data acquisition unit is used to collect and store the online monitoring data of the online oil chromatography monitoring device in real time to form a historical database. The data processing unit is used to perform unified processing on the data in the historical database, including data standardization. The abnormal data extraction unit is used to combine historical databases and the monitoring principle of online oil chromatography devices to statistically analyze abnormal operating data of the monitoring devices and form an abnormal device database. The abnormal data analysis unit is used to perform cluster analysis on abnormal data based on the device abnormal database and construct an abnormal data feature library; The abnormal data identification unit is used to identify abnormal data in the current monitoring data of the online oil chromatography monitoring device. The comparison and judgment unit is used to compare the abnormal data in the current monitoring data identified by the abnormal data identification unit with the abnormal feature library. When the abnormal data matches a certain type of abnormal feature in the abnormal feature library, the monitoring device is judged to be abnormal; otherwise, the transformer is faulty. The early warning unit is used to issue early warnings based on the judgment results of the comparison and judgment unit, including reliability early warnings and transformer fault early warnings. The reliability early warning is due to a reliability failure of the monitoring device, or an abnormal monitoring status of the monitoring device causing data distortion. The transformer fault early warning is due to a normal monitoring status of the monitoring device, or a transformer failure.

2. The online oil chromatography reliability early warning system according to claim 1, characterized in that: The data processing unit performs data standardization in the following manner: Assuming m characteristic gases, and each characteristic gas sample data contains n sample data, a multidimensional data matrix is ​​obtained. The standardized data is then represented as: In the formula: , These represent the raw data and the standardized data, respectively. express The mean, express The variance is then used to obtain the standardized sample data matrix. .

3. The online oil chromatography reliability early warning system according to claim 1, characterized in that: The anomaly data analysis unit specifically employs the K-means clustering algorithm to perform cluster analysis on the anomaly data, including the following steps: Based on the anomaly data feature library, an anomaly data set D is constructed, and the n data points in the data set D are divided into K clusters. ① Select K initial cluster centers from the dataset D; ② Calculate the Euclidean distance from each data point in the dataset to the cluster centers, expressed as: In the formula: Let n represent two data points in space, respectively. express 1. Calculate the Euclidean distance between the data points; 2. Select the nearest cluster center and include it in the cluster; 3. After calculating all data points, recalculate the cluster center for each cluster; 4. Determine if the cluster center has changed. If it has changed, return to step 2. If it has not changed, output the clustering result.

4. The online oil chromatography reliability early warning system according to claim 3, characterized in that: The optimization objective of the K-means clustering algorithm in the anomaly data analysis unit is represented by minimizing the squared error of the clusters: In the formula: Representing a cluster Cluster centers.

5. The online oil chromatography reliability early warning system according to claim 1, characterized in that: The abnormal data in the abnormal data extraction unit and the abnormal data identification unit include abnormal monitoring values, abnormal coefficients of variation, and abnormal data offset rates; wherein, the abnormal monitoring values ​​include singular values, null values, and negative values; the abnormal coefficient of variation is used to identify abnormal data fluctuations, specifically expressed as follows: In the formula: Represents the coefficient of variation. Indicates the standard deviation of the outlier interval. This represents the mean of the abnormal interval, with the anomaly coefficient threshold set to 30%. The data offset rate anomaly is used to identify the offset rate of abnormal intervals in the detection data of each characteristic gas, specifically expressed as follows: In the formula: Indicates the first Characteristic gas Sky offset rate, Indicates the first Characteristic gas Daily monitoring values Indicates the first Characteristic gases Daily average.

6. The online oil chromatography reliability early warning system according to claim 1, characterized in that: The comparison and judgment unit uses a specific correlation analysis method to compare abnormal data. The abnormal data sequence identified in the abnormal data identification unit is used as the comparison sequence, and the data sequence in the abnormal feature library is used as the reference sequence. The correlation between the comparison sequence and the reference sequence is then expressed as: , The amount of data in the reference sequence and the comparison sequence. Represents the correlation coefficient. Indicates the first The first group of comparison sequences with the reference sequence The difference between the data points express The minimum value, express The maximum value, This represents the resolution coefficient, which is set to 0.5 here.

Images