Current transformer

Abstract

The invention discloses a current transformer, and relates to the technical field of transformer detection, and a data acquisition module in the current transformer is used for obtaining operation data of the current transformer; the data preprocessing module is used for preprocessing the operation data; the three-phase relation parameter extraction module is used for extracting three-phase relation parameters based on the operation data in combination with harmonic analysis; the abnormal element extraction module is used for extracting abnormal elements by using the target three-phase relation parameters according to baseline analysis; the abnormal component calculation module is used for performing abnormal component calculation based on the target abnormal elements and the target three-phase relation parameters; the abnormal feature extraction module is used for performing abnormal feature extraction based on the abnormal component in combination with magnetic saturation detection; and the abnormity judgment and early warning module is used for carrying out abnormity judgment on the current transformer based on the abnormal characteristics so as to judge whether early warning information needs to be sent out. According to the invention, the accuracy and efficiency of abnormal detection of the current transformer can be improved more ideally.

Description

Technical Field

[0001] The present invention mainly relates to the technical field of mutual inductor detection, and in particular to a current mutual inductor. Background Art

[0002] Current transformers are indispensable equipment in power systems, used to measure current and convert it into corresponding signals to monitor and protect the power system. Various anomalies may occur during the long-term operation of current transformers. If the anomaly of the current transformer is not discovered in a timely and accurate manner, it will not only affect the accuracy of electric energy metering, but also damage related equipment. However, the current abnormal judgment of the current transformer is usually arranged by relevant personnel to perform abnormal detection through the naked eye or simple instruments. However, this method is too dependent on the professional qualities of the relevant personnel, and it is difficult to ensure the accuracy of the abnormal judgment of the current transformer. In addition, this method cannot perform continuous abnormal detection, making the abnormal judgment of the current transformer inefficient. In response to this, some companies have integrated corresponding modules into the current transformer. Although it can improve the abnormal detection efficiency of the current transformer, it still fails to effectively improve the abnormal detection accuracy of the current transformer. Therefore, how to effectively improve the accuracy and efficiency of abnormal detection of current transformers is a difficult problem that companies have been studying. Summary of the invention

[0003] The purpose of the present invention is to overcome the deficiencies of the prior art. The present invention provides a current transformer that can replace the manual judgment mode, so that the accuracy and efficiency of abnormality detection of the current transformer can be more ideally improved.

[0004] In order to solve the above technical problems, the present invention provides a current transformer, which includes: a data acquisition module, a data preprocessing module, a three-phase relationship parameter extraction module, an abnormal element extraction module, an abnormal component calculation module, an abnormal feature extraction module and an abnormal judgment and early warning module; The data acquisition module is used to obtain the operating data of the current transformer; The data preprocessing module is used to preprocess the operation data to obtain preprocessed operation data; The three-phase relationship parameter extraction module is used to extract the three-phase relationship parameters based on the pre-processed operation data combined with harmonic analysis to obtain the target three-phase relationship parameters; The abnormal element extraction module is used to extract abnormal elements using the target three-phase relationship parameters according to the baseline analysis to obtain the target abnormal elements; The abnormal component calculation module is used to calculate the abnormal component based on the target abnormal element and the target three-phase relationship parameter to obtain the abnormal component; The abnormal feature extraction module is used to extract abnormal features based on the abnormal component combined with magnetic saturation detection to obtain target abnormal features; The abnormality judgment and early warning module is used to judge the abnormality of the current transformer based on the target abnormality characteristics, and issue an early warning message if it is judged that the current transformer is abnormal.

[0005] Optionally, the data acquisition module includes an intelligent sensor combination provided on the current transformer, and acquires operating data of the current transformer based on the intelligent sensor combination.

[0006] Optionally, the data preprocessing module includes: a data cleaning unit and a data integration unit; The data cleaning unit is used to perform data cleaning on the operation data to obtain the operation data after data cleaning; The data integration unit is used to perform data integration processing on the operation data after data cleaning processing to obtain pre-processed operation data.

[0007] Optionally, the three-phase relationship parameter extraction module includes: a multi-order difference screening unit, a three-phase current component construction unit, a three-correlation system metering parameter unit, a harmonic analysis unit and a target three-phase relationship parameter construction unit; The multi-order differential screening unit is used to perform multi-order differential screening on the operating current data extracted from the pre-processed operating data to obtain stable current data; The three-phase current component construction unit is used to construct the three-phase current components based on the steady current data to obtain the three-phase current components; The three-correlation system metering parameter unit is used to construct the three-correlation system metering parameters based on the steady current data using the model parameter matrix of the three-phase relationship; The harmonic analysis unit is used to perform harmonic analysis of three-phase current based on the operating current data to obtain harmonic analysis data; The target three-phase relationship parameter construction unit is used to construct the target three-phase relationship parameters based on the three-phase current components, three related system metering parameters and harmonic analysis data.

[0008] Optionally, the harmonic analysis unit includes: a subharmonic comparison data generation subunit, a current harmonic quantity generation subunit and a harmonic analysis data generation subunit; The subharmonic comparison data generating subunit is used to calculate the subharmonic data of the three-phase current based on the operating current data, and generate the subharmonic comparison data based on a preset time window using the subharmonic data; The current harmonic amount generating subunit is used to calculate the fundamental current component and the harmonic current component based on the operating current data, and generate the current harmonic amount based on the fundamental current component and the harmonic current component; The harmonic analysis data generating subunit is used to generate harmonic analysis data based on the subharmonic comparison data and the current harmonic amount.

[0009] Optionally, the abnormal element extraction module includes: a correlation analysis unit, a sequence mining unit, an element distance calculation unit, an abnormal boundary threshold construction unit and an abnormal element determination unit; The correlation analysis unit is used to obtain historical abnormal element record data, and perform correlation analysis on the historical abnormal element record data to obtain correlation analysis data; The sequence mining unit is used to perform sequence mining based on the association analysis data using a frequent pattern tree to obtain a target element sequence; The element distance calculation unit is used to determine the strongly correlated elements based on the target element sequence and the correlation analysis data, and calculate the target distance of the strongly correlated elements; The abnormal boundary threshold value construction unit is used to construct a three-phase statistical abnormal boundary threshold value based on the historical abnormal element record data using a kernel density function; The abnormal element determination unit is used to construct an abnormal element baseline based on the three-phase statistic abnormal boundary threshold and the target distance, and determine the target abnormal element using the abnormal element baseline based on the target three-phase relationship parameter.

[0010] Optionally, the abnormal component calculation module includes: an abnormal component amplitude calculation unit, an abnormal related component acquisition unit and an abnormal component generation unit; The abnormal component amplitude calculation unit is used to calculate a target difference between a target abnormal element and an abnormal element baseline, and determine the abnormal component amplitude based on the target difference; The abnormal related component acquisition unit is used to calculate the residual component based on the target three-phase relationship parameter, and perform decomposition and cluster analysis on the residual component to obtain the abnormal related component; The abnormal component generating unit is used to generate an abnormal component based on the abnormal component amplitude and the abnormal correlation component.

[0011] Optionally, the abnormal feature extraction module includes: a first abnormal feature extraction unit, a first magnetic saturation detection data generation unit, a second abnormal feature determination unit and a target abnormal feature generation unit; The first abnormal feature extraction unit is used to extract features of the abnormal component based on the first abnormal feature extraction model to obtain a first abnormal feature; The first magnetic saturation detection data generating unit is used to calculate the voltage amplitude conversion data and the harmonic slope conversion data based on the operation voltage data extracted from the pre-processed operation data, and determine the first magnetic saturation detection data using a preset judgment coefficient range based on the voltage amplitude conversion data and the harmonic slope conversion data; The second abnormal feature determination unit is used to determine the second magnetic saturation detection data based on the abnormal component by using the full-wave Fourier algorithm combined with the saturation identification partition, and generate a second abnormal feature based on the first magnetic saturation detection data and the second magnetic saturation detection data; The target abnormal feature generating unit is used to generate a target abnormal feature based on the first abnormal feature and the second abnormal feature.

[0012] Optionally, the abnormality judgment and early warning module includes: an abnormality recognition model building unit and an early warning judgment unit; The abnormality identification model construction unit is used to construct an abnormality knowledge graph model of the current transformer, and to construct an abnormality identification model based on the abnormality knowledge graph model combined with the abnormality classification model; The early warning judgment unit is used to input the target abnormality feature into the abnormality recognition model to perform abnormality judgment on the current transformer. If it is judged that the current transformer is abnormal, early warning information is generated based on the abnormality judgment result.

[0013] Optionally, the anomaly recognition model building unit includes: an anomaly knowledge data classification subunit, an anomaly knowledge graph model building subunit and an anomaly recognition model generation subunit; The abnormal knowledge data classification subunit is used to obtain abnormal knowledge data in the abnormal knowledge base, and classify the abnormal knowledge data to obtain abnormal knowledge data of each category; The abnormal knowledge graph model construction subunit is used to construct an abnormal knowledge graph model based on abnormal knowledge data of various categories combined with abnormal coupling impact degree analysis; The anomaly recognition model generation subunit is used to generate an anomaly recognition model based on an anomaly knowledge graph model and an anomaly classification model composed of a convolutional neural network.

[0014] In the embodiment of the present invention, a three-phase relationship parameter extraction module, an abnormal element extraction module, an abnormal component calculation module, an abnormal feature extraction module and an abnormal judgment and early warning module are integrated in the current transformer. The three-phase relationship parameter extraction module is used to extract the three-phase relationship parameters based on the pre-processed operation data combined with harmonic analysis, and can obtain more comprehensive and accurate three-phase relationship parameters. The abnormal element extraction module is used to extract abnormal elements using the target three-phase relationship parameters according to the baseline analysis, which effectively improves the reliability of abnormal element extraction. The abnormal component calculation module is used to calculate the abnormal component based on the target abnormal element and the target three-phase relationship parameter, which can avoid the abnormal component obtained from the actual situation. The abnormal feature extraction module is used to extract abnormal features based on the abnormal component combined with magnetic saturation detection, and the feature data of magnetic saturation detection is also added to the abnormal features, which can make the abnormal features obtained more comprehensive. The abnormal judgment and early warning module is used to judge the abnormality of the current transformer based on the abnormal features, which can more accurately and efficiently judge whether the current transformer is abnormal in the case of complex data, thereby replacing the manual judgment mode, so that the accuracy and efficiency of the abnormal detection of the current transformer can be more ideally improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0016] Figure 1 is a schematic diagram of the structure of a current transformer in an embodiment of the present invention; Figure 2 It is a flow chart of a method for controlling a current transformer in an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. 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.

[0018] Embodiment 1

[0019] See also Figure 1 , Figure 1It is a schematic diagram of the structural composition of the current transformer in an embodiment of the present invention, wherein the current transformer comprises: a data acquisition module 11, a data preprocessing module 12, a three-phase relationship parameter extraction module 13, an abnormal element extraction module 14, an abnormal component calculation module 15, an abnormal feature extraction module 16 and an abnormal judgment and warning module 17.

[0020] The data acquisition module 11 is used to obtain the operating data of the current transformer; In a specific implementation of the present invention, the data acquisition module 11 includes an intelligent sensor combination arranged in the current transformer, and the operation data of the current transformer is acquired based on the intelligent sensor combination.

[0021] Specifically, the intelligent sensor combination provided in the current transformer can obtain operating data of multiple dimensions such as operating current data, operating voltage data and operating temperature data of the current transformer.

[0022] The data preprocessing module 12 is used to preprocess the operation data to obtain preprocessed operation data; In the specific implementation process of the present invention, the data preprocessing module 12 includes: a data cleaning unit and a data integration unit; the data cleaning unit is used to perform data cleaning processing on the operating data to obtain the operating data after data cleaning processing; the data integration unit is used to perform data integration processing on the operating data after data cleaning processing to obtain the preprocessed operating data.

[0023] Specifically, the data cleaning unit is used to perform data cleaning processing on the operation data, remove abnormal characters and garbled characters in the operation data, and obtain the operation data after data cleaning processing. The data integration unit is used to perform data integration processing on the operation data after data cleaning processing, select each data element to be processed that needs to be processed from the operation data after data cleaning processing, perform pairing processing on each data element to be processed, obtain the matching relationship between each data element to be processed, generate a data relationship diagram based on the matching relationship using a preset topology diagram, and perform data integration processing on the operation data after data cleaning processing according to the data relationship diagram, which can effectively improve the efficiency of data integration. After completing the data integration processing, the preprocessing of the operation data is completed, and the preprocessed operation data is obtained.

[0024] The three-phase relationship parameter extraction module 13 is used to extract the three-phase relationship parameters based on the pre-processed operation data combined with harmonic analysis to obtain the target three-phase relationship parameters; In the specific implementation process of the present invention, the three-phase relationship parameter extraction module 13 includes: a multi-order differential screening unit, a three-phase current component construction unit, a three-related system metering parameter unit, a harmonic analysis unit and a target three-phase relationship parameter construction unit; the multi-order differential screening unit is used to perform multi-order differential screening on the operating current data extracted from the pre-processed operating data to obtain steady current data; the three-phase current component construction unit is used to construct three-phase current components based on the steady current data to obtain three-phase current components; the three-related system metering parameter unit is used to construct three-related system metering parameters based on the steady current data using a model parameter matrix of the three-phase relationship; the harmonic analysis unit is used to perform harmonic analysis of the three-phase current based on the operating current data to obtain harmonic analysis data; the target three-phase relationship parameter construction unit is used to construct target three-phase relationship parameters based on the three-phase current components, three-related system metering parameters and harmonic analysis data.

[0025] Furthermore, the harmonic analysis unit includes: a subharmonic comparison data generation subunit, a current harmonic quantity generation subunit and a harmonic analysis data generation subunit; the subharmonic comparison data generation subunit is used to calculate the subharmonic data of the three-phase current based on the operating current data, and generate subharmonic comparison data based on a preset time window using the subharmonic data; the current harmonic quantity generation subunit is used to calculate the fundamental current component and the harmonic current component based on the operating current data, and generate the current harmonic quantity based on the fundamental current component and the harmonic current component; the harmonic analysis data generation subunit is used to generate harmonic analysis data based on the subharmonic comparison data and the current harmonic quantity.

[0026] Specifically, the multi-order differential screening unit is used to perform multi-order differential screening on the operating current data extracted from the pre-processed operating data. The operating current data includes the current output data of each phase of the current transformer. The first-order differential and second-order differential are used to perform differential noise filtering on the extracted operating current data to filter out the data breakpoints in the operating current data, that is, to obtain stable current data. The three-phase current component construction unit is used to construct three-phase current components based on the stable current data, calculate the three-phase asymmetric current component according to the stable current data, calculate the positive sequence component, zero sequence component and negative sequence component according to the three-phase current data in the operating current data, and combine the three-phase asymmetric current component, positive sequence component, zero sequence component and negative sequence component into three-phase current components. The three-correlation system metering parameter unit is used to construct the three-correlation system metering parameters based on the steady current data using the model parameter matrix of the three-phase relationship, obtain the historical steady current data of the current transformer, extract the current amplitude, first-order difference and second-order difference of each phase current according to the historical steady current data, and construct the model parameter matrix of the three-phase relationship according to the current amplitude, first-order difference and second-order difference of each phase current using the principal component analysis method combined with the singular matrix, and calculate the statistical parameters of the three-phase current using the steady current data through the model parameter matrix, including the effective value of the current and the phase difference, so as to obtain the three-correlation system metering parameters. The subharmonic comparison data generation subunit is used to calculate the subharmonic data of the three-phase current based on the operating current data, and the subharmonics of the three-phase current can be calculated by the subharmonic calculation formula, and the subharmonic comparison data is generated based on the preset time window using the subharmonic data, and the preset time window is set according to the preset time scale range. Each subharmonic is compared in the preset time window, and the comparison can be performed by calculating the difference. The current harmonic quantity generating subunit is used to calculate the fundamental current component and the harmonic current component based on the operating current data, calculate the fundamental current component and the harmonic current component by Fourier algorithm, generate the current harmonic quantity based on the fundamental current component and the harmonic current component, calculate the harmonic restraint quantity based on the fundamental current component and the harmonic current component, and construct the current harmonic quantity based on the harmonic restraint quantity, the fundamental current component and the harmonic current component. The harmonic analysis data generating subunit is used to construct the harmonic analysis data based on the subharmonic comparison data and the current harmonic quantity. The target three-phase relationship parameter constructing unit is used to construct the target three-phase relationship parameters based on the three-phase current components, the three related system metering parameters and the harmonic analysis data, so that the obtained target three-phase relationship parameters are more comprehensive and can better reflect the operation of the current transformer.

[0027] The abnormal element extraction module 14 is used to extract abnormal elements using the target three-phase relationship parameters according to the baseline analysis to obtain the target abnormal elements; In the specific implementation process of the present invention, the abnormal element extraction module 14 includes: a correlation analysis unit, a sequence mining unit, an element distance calculation unit, an abnormal boundary threshold construction unit and an abnormal element determination unit; the correlation analysis unit is used to obtain historical abnormal element record data, and perform correlation analysis on the historical abnormal element record data to obtain correlation analysis data; the sequence mining unit is used to perform sequence mining using a frequent pattern tree based on the correlation analysis data to obtain a target element sequence; the element distance calculation unit is used to determine a strongly correlated element based on the target element sequence and the correlation analysis data, and calculate the target distance of the strongly correlated element; the abnormal boundary threshold construction unit is used to construct a three-phase statistic abnormal boundary threshold based on the historical abnormal element record data using a kernel density function; the abnormal element determination unit is used to construct an abnormal element baseline based on the three-phase statistic abnormal boundary threshold and the target distance, and determine the target abnormal element based on the target three-phase relationship parameter using the abnormal element baseline.

[0028] Specifically, the correlation analysis unit is used to obtain historical abnormal element record data, the historical abnormal element record data includes record data of related abnormal elements obtained when abnormalities occurred in the current transformer in the past, the abnormal elements include various state variable values ​​of the circuit transformer when fixed deviation abnormality, ratio deviation abnormality, failure abnormality, drift deviation abnormality and precision distortion abnormality occur, such as three-phase current component values ​​and harmonic values, and the historical abnormal element record data are subjected to correlation analysis to obtain the occurrence frequency of each abnormal state variable value in the historical abnormal element record data, calculate the abnormal trigger frequency according to the occurrence frequency, and calculate the abnormal trigger frequency according to the abnormal trigger frequency. The trigger frequency of the abnormality is calculated based on the trigger frequency of each abnormality and the trigger frequency of the abnormality, and a correlation analysis is performed according to the abnormality trigger frequency of each abnormality and the trigger frequency of the abnormality to obtain the correlation analysis data; the sequence mining unit is used to perform sequence mining based on the correlation analysis data using a frequent pattern tree. The frequent pattern tree is a data structure for efficiently mining frequent item sets, which is implemented through a tree structure. According to the correlation analysis data, the confidence analysis is used to mine the element sequence in the historical abnormal element record data to obtain the target element sequence, thereby improving the completeness of the element sequence mining; the element distance calculation unit is used to determine the strongly correlated elements based on the target element sequence and the correlation analysis data, and determine the elements of the strong correlation relationship in the target element sequence according to the correlation analysis data combined with the Pearson coefficient, that is, determine the strongly correlated elements, and calculate the target distance of the strongly correlated elements, and calculate the target distance of the strongly correlated elements through the distance calculation model; the abnormal boundary threshold construction unit is used to construct a three-phase statistical abnormal boundary threshold based on the historical abnormal element record data using a kernel density function, and construct an abnormal element feature quantity data set according to the historical abnormal element record data, and the data in the abnormal element feature quantity data set is mirrored and projected into a plane coordinate system, and a new abnormal element feature quantity data set is obtained by integration, and a boundary is set. Constraint conditions, the boundary constraints can be the lower limit of the variable value of the abnormal element. According to the boundary constraints, the density function and confidence interval analysis are combined with the new abnormal element feature data set to construct the three-phase statistical abnormal boundary threshold; the abnormal element determination unit is used to construct the abnormal element baseline based on the three-phase statistical abnormal boundary threshold and the target distance, and determine the target abnormal element based on the target three-phase relationship parameters using the abnormal element baseline, that is, compare the three-phase current components in the three-phase relationship parameters, the three-related system metering parameters and the harmonic analysis data with the abnormal element baseline. If it reaches or exceeds the abnormal element baseline, it is an abnormal element.

[0029] The abnormal component calculation module 15 is used to calculate the abnormal component based on the target abnormal element and the target three-phase relationship parameter to obtain the abnormal component; In the specific implementation process of the present invention, the abnormal component calculation module 15 includes: an abnormal component amplitude calculation unit, an abnormal related component acquisition unit and an abnormal component generation unit; the abnormal component amplitude calculation unit is used to calculate the target difference between the target abnormal element and the abnormal element baseline, and determine the abnormal component amplitude based on the target difference; the abnormal related component acquisition unit is used to calculate the residual component based on the target three-phase relationship parameter, and decompose and cluster the residual component to obtain the abnormal related component; the abnormal component generation unit is used to generate the abnormal component based on the abnormal component amplitude and the abnormal related component.

[0030] Specifically, the abnormal component amplitude calculation unit is used to calculate the target difference between the target abnormal element and the abnormal element baseline, and determine the abnormal component amplitude based on the target difference; the abnormal related component acquisition unit is used to calculate the residual component based on the target three-phase relationship parameter, decompose the target three-phase relationship parameter by principal component analysis to obtain the residual component, decompose and cluster analyze the residual component, perform empirical mode decomposition on the residual component, obtain the residual sub-component, obtain the correlation index between the current data and the residual component, the correlation index is obtained by performing correlation analysis on the current data and the residual component, and cluster analysis is performed using the residual sub-component according to the correlation index to obtain the residual component related to the abnormal type, that is, to obtain the abnormal related component; the abnormal component generation unit is used to generate the abnormal component based on the abnormal component amplitude and the abnormal related component, so as to obtain a more accurate and comprehensive abnormal component, and avoid the limitations and one-sidedness of the abnormal related component composed of a single factor.

[0031] The abnormal feature extraction module 16 is used to extract abnormal features based on the abnormal component combined with magnetic saturation detection to obtain target abnormal features; In the specific implementation process of the present invention, the abnormal feature extraction module 16 includes: a first abnormal feature extraction unit, a first magnetic saturation detection data generation unit, a second abnormal feature determination unit and a target abnormal feature generation unit; the first abnormal feature extraction unit is used to extract features of the abnormal component based on the first abnormal feature extraction model to obtain a first abnormal feature; the first magnetic saturation detection data generation unit is used to calculate voltage amplitude transformation data and harmonic slope transformation data based on the operating voltage data extracted from the preprocessed operating data, and determine the first magnetic saturation detection data based on the voltage amplitude transformation data and the harmonic slope transformation data using a preset judgment coefficient; the second abnormal feature determination unit is used to determine the second magnetic saturation detection data based on the abnormal component using a full-wave Fourier algorithm combined with a saturation identification partition, and generate a second abnormal feature based on the first magnetic saturation detection data and the second magnetic saturation detection data; the target abnormal feature generation unit is used to generate a target abnormal feature based on the first abnormal feature and the second abnormal feature.

[0032] Specifically, the first abnormal feature extraction unit is used to extract features of the abnormal component based on the first abnormal feature extraction model, and the first abnormal feature extraction model is a convergence model obtained by inputting a sample data set into a deep neural network for training to obtain a first abnormal feature. The first magnetic saturation detection data generation unit is used to calculate voltage amplitude transformation data and harmonic slope transformation data based on the operating voltage data extracted from the preprocessed operating data, the operating voltage data includes the voltage analog quantity of the secondary winding in the current transformer, the voltage analog quantity is extracted from the operating voltage data, the fundamental wave and the harmonic wave are extracted from the voltage analog quantity, and the voltage amplitude transformation quantity and the harmonic wave slope transformation quantity are calculated according to the fundamental wave and the harmonic wave, that is, the voltage amplitude transformation data and the harmonic wave slope transformation data are obtained. Based on the voltage amplitude transformation data and the harmonic slope transformation data, the first magnetic saturation detection data is determined using a preset judgment coefficient range, the first difference between the voltage amplitude transformation data and the preset amplitude transformation data is calculated, the second difference between the harmonic slope transformation data and the preset slope transformation data is calculated, the judgment coefficient is calculated according to the first difference and the second difference using their respective weight ratios, the judgment coefficient is compared with the preset judgment coefficient range, if the judgment coefficient is within the first range of the preset judgment coefficient, the core of the current transformer is not saturated, if the judgment coefficient is within the second range of the preset judgment coefficient, the core of the current transformer is slightly saturated, and if the judgment coefficient is within the third range of the preset judgment coefficient, the core of the current transformer is deeply saturated, that is, the first magnetic saturation detection data is obtained. The second abnormal feature determination unit is used to determine the second magnetic saturation detection data based on the abnormal component using the full-wave Fourier algorithm combined with the saturation identification partition. The saturation identification partition is to construct a saturation identification plane according to the relevant data of the abnormality in the zone of the current transformer, and to partition the constructed saturation identification plane using the preset partitioning rules to obtain the saturation identification partition. The full-wave Fourier algorithm is used to calculate the full-wave Fourier difference current according to the abnormal component, and the data points falling in the saturation identification partition are determined according to the full-wave Fourier difference current to determine its magnetic saturation data, that is, the second magnetic saturation detection data, and the second abnormal feature is generated based on the first magnetic saturation detection data and the second magnetic saturation detection data, so as to obtain more comprehensive and accurate abnormal feature data about magnetic saturation. The target abnormal feature generation unit is used to generate a target abnormal feature based on the first abnormal feature and the second abnormal feature, and the target abnormal feature is obtained by fusing the first abnormal feature and the second abnormal feature.

[0033] The abnormality judgment and warning module 17 is used to judge the abnormality of the current transformer based on the target abnormality characteristics, and issue a warning message if it is judged that the current transformer is abnormal.

[0034] In the specific implementation process of the present invention, the abnormal judgment and early warning module 17 includes: an abnormal identification model construction unit and an early warning judgment unit; the abnormal identification model construction unit is used to construct an abnormal knowledge graph model of the current transformer, and construct an abnormal identification model based on the abnormal knowledge graph model combined with the abnormal classification model; the early warning judgment unit is used to input the target abnormal feature into the abnormal identification model to perform abnormal judgment of the current transformer. If it is judged that the current transformer is abnormal, a warning information is generated based on the abnormal judgment result.

[0035] Furthermore, the abnormality recognition model construction unit includes: an abnormality knowledge data classification subunit, an abnormality knowledge graph model construction subunit and an abnormality recognition model generation subunit; the abnormality knowledge data classification subunit is used to obtain abnormality knowledge data in the abnormality knowledge base, and classify the abnormality knowledge data to obtain abnormal knowledge data of each category; the abnormality knowledge graph model construction subunit is used to construct an abnormality knowledge graph model based on the abnormal knowledge data of each category combined with the abnormal coupling influence degree analysis; the abnormality recognition model generation subunit is used to generate an abnormality recognition model based on the abnormality knowledge graph model and the abnormality classification model composed of a convolutional neural network.

[0036] Specifically, the abnormal knowledge data classification subunit is used to obtain abnormal knowledge data in the abnormal knowledge base, the abnormal knowledge data is related data of the abnormal type occurring in the current transformer, and classify the abnormal knowledge data, that is, classify the abnormal knowledge data according to the abnormal type, and obtain abnormal knowledge data of each category. The abnormal knowledge graph model construction subunit is used to construct an abnormal knowledge graph model based on the abnormal knowledge data of each category combined with the abnormal coupling influence degree analysis, obtain the coupling relationship and coupling strength data between the abnormal knowledge data of each category, construct an abnormal function model according to the coupling relationship and coupling strength data, obtain the abnormal coupling influence degree data corresponding to the abnormal knowledge data of each category according to the abnormal function model, construct a number of abnormal knowledge graphs according to the abnormal coupling influence degree data and the abnormal knowledge data of each category using a preset knowledge extraction strategy, establish an association relationship between each abnormal knowledge graph, and combine a number of abnormal knowledge graphs into an abnormal knowledge graph model according to the association relationship; the abnormal recognition model generation subunit is used to generate an abnormal recognition model based on the abnormal knowledge graph model and the abnormal classification model composed of a convolutional neural network. The combination of the abnormal knowledge graph model and the abnormal classification model can improve the reasoning ability and adapt to large-scale data to improve the efficiency and accuracy of abnormal recognition of current transformers. The early warning judgment unit is used to input the target abnormality feature into the abnormality recognition model to perform abnormality judgment on the current transformer. The abnormalities of the current transformer include primary side abnormality, accuracy distortion abnormality, drift deviation abnormality and error state abnormality. If it is judged that the current transformer has an abnormality, an early warning message is generated based on the abnormality judgment result. If it is judged that the power transformer has no abnormality, the operation data of the current transformer continues to be collected for data analysis to achieve uninterrupted detection of current transformer abnormalities and improve the efficiency and accuracy of current transformer abnormality detection.

[0037] In the embodiment of the present invention, a three-phase relationship parameter extraction module, an abnormal element extraction module, an abnormal component calculation module, an abnormal feature extraction module and an abnormal judgment and early warning module are integrated in the current transformer. The three-phase relationship parameter extraction module is used to extract the three-phase relationship parameters based on the pre-processed operation data combined with harmonic analysis, and can obtain more comprehensive and accurate three-phase relationship parameters. The abnormal element extraction module is used to extract abnormal elements using the target three-phase relationship parameters according to the baseline analysis, which effectively improves the reliability of abnormal element extraction. The abnormal component calculation module is used to calculate the abnormal component based on the target abnormal element and the target three-phase relationship parameter, which can avoid the abnormal component obtained from the actual situation. The abnormal feature extraction module is used to extract abnormal features based on the abnormal component combined with magnetic saturation detection, and the feature data of magnetic saturation detection is also added to the abnormal features, which can make the abnormal features obtained more comprehensive. The abnormal judgment and early warning module is used to judge the abnormality of the current transformer based on the abnormal features, which can more accurately and efficiently judge whether the current transformer is abnormal in the case of complex data, thereby replacing the manual judgment mode, so that the accuracy and efficiency of the abnormal detection of the current transformer can be more ideally improved.

[0038] Embodiment 2

[0039] The control method of the current transformer involved in the second embodiment of the present invention is implemented based on the current transformer described in the first embodiment, such as Figure 2 As shown, Figure 2 is a flow chart of a method for controlling a current transformer in an embodiment of the present invention, the method comprising: S21: The data acquisition module obtains the operating data of the current transformer; S22: a data preprocessing module preprocesses the operation data to obtain preprocessed operation data; S23: The three-phase relationship parameter extraction module extracts the three-phase relationship parameters based on the pre-processed operation data combined with harmonic analysis to obtain target three-phase relationship parameters; S24: The abnormal element extraction module extracts abnormal elements based on the baseline analysis using the target three-phase relationship parameters to obtain target abnormal elements; S25: The abnormal component calculation module calculates the abnormal component based on the target abnormal element and the target three-phase relationship parameter to obtain the abnormal component; S26: The abnormal feature extraction module extracts abnormal features based on the abnormal component combined with magnetic saturation detection to obtain target abnormal features; S27: The abnormality judgment and warning module judges the abnormality of the current transformer based on the target abnormality characteristics, and issues a warning message if it is judged that the current transformer is abnormal.

[0040] In the specific implementation process of the present invention, the specific implementation methods of the method items can refer to the implementation methods of the above embodiments, which will not be repeated here.

[0041] In an embodiment of the present invention, the three-phase relationship parameter extraction module extracts the three-phase relationship parameters based on the pre-processed operating data combined with harmonic analysis, and can obtain more comprehensive and accurate three-phase relationship parameters. The abnormal element extraction module extracts abnormal elements based on the baseline analysis using the target three-phase relationship parameters, which effectively improves the reliability of abnormal element extraction. The abnormal component calculation module calculates abnormal components based on the target abnormal elements and the target three-phase relationship parameters, which can avoid excessive deviations between the obtained abnormal components and the actual situation. The abnormal feature extraction module extracts abnormal features based on the abnormal components combined with magnetic saturation detection, and also adds the feature data of magnetic saturation detection to the abnormal features, which can make the obtained abnormal features more comprehensive. The abnormal judgment and early warning module judges the abnormality of the current transformer based on the abnormal features, which can more accurately and efficiently judge whether the current transformer is abnormal in the case of complex data, thereby replacing the manual judgment mode, so that the accuracy and efficiency of the abnormal detection of the current transformer can be more ideally improved.

[0042] A person skilled in the art may understand that all or part of the steps in the various methods of the above embodiments may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, and the storage medium may include: a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc.

[0043] In addition, the above is a detailed introduction to a current transformer provided in an embodiment of the present invention. Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A current transformer, characterized in that: The current transformer comprises: a data acquisition module, a data preprocessing module, a three-phase relationship parameter extraction module, an abnormal element extraction module, an abnormal component calculation module, an abnormal feature extraction module and an abnormal judgment and early warning module; The data acquisition module is used to obtain the operating data of the current transformer; The data preprocessing module is used to preprocess the operation data to obtain preprocessed operation data; The three-phase relationship parameter extraction module is used to extract the three-phase relationship parameters based on the pre-processed operation data combined with harmonic analysis to obtain the target three-phase relationship parameters; The abnormal element extraction module is used to extract abnormal elements using the target three-phase relationship parameters according to the baseline analysis to obtain the target abnormal elements; The abnormal component calculation module is used to calculate the abnormal component based on the target abnormal element and the target three-phase relationship parameter to obtain the abnormal component; The abnormal feature extraction module is used to extract abnormal features based on the abnormal component combined with magnetic saturation detection to obtain target abnormal features; The abnormality judgment and early warning module is used to judge the abnormality of the current transformer based on the target abnormality characteristics, and issue an early warning message if it is judged that the current transformer is abnormal.

2. The current transformer according to claim 1, characterized in that: The data acquisition module includes an intelligent sensor combination arranged on the current transformer, and acquires the operation data of the current transformer based on the intelligent sensor combination.

3. The current transformer according to claim 1, characterized in that: The data preprocessing module includes: a data cleaning unit and a data integration unit; The data cleaning unit is used to perform data cleaning on the operation data to obtain the operation data after data cleaning; The data integration unit is used to perform data integration processing on the operation data after data cleaning processing to obtain pre-processed operation data.

4. The current transformer according to claim 1, characterized in that: The three-phase relationship parameter extraction module includes: a multi-order difference screening unit, a three-phase current component construction unit, a three-correlation system metering parameter unit, a harmonic analysis unit and a target three-phase relationship parameter construction unit; The multi-order differential screening unit is used to perform multi-order differential screening on the operating current data extracted from the pre-processed operating data to obtain stable current data; The three-phase current component construction unit is used to construct the three-phase current components based on the steady current data to obtain the three-phase current components; The three-correlation system metering parameter unit is used to construct the three-correlation system metering parameters based on the steady current data using the model parameter matrix of the three-phase relationship; The harmonic analysis unit is used to perform harmonic analysis of three-phase current based on the operating current data to obtain harmonic analysis data; The target three-phase relationship parameter construction unit is used to construct the target three-phase relationship parameters based on the three-phase current components, three related system metering parameters and harmonic analysis data.

5. The current transformer according to claim 4, characterized in that: The harmonic analysis unit includes: a subharmonic comparison data generation subunit, a current harmonic quantity generation subunit and a harmonic analysis data generation subunit; The subharmonic comparison data generating subunit is used to calculate the subharmonic data of the three-phase current based on the operating current data, and generate the subharmonic comparison data based on a preset time window using the subharmonic data; The current harmonic amount generating subunit is used to calculate the fundamental current component and the harmonic current component based on the operating current data, and generate the current harmonic amount based on the fundamental current component and the harmonic current component; The harmonic analysis data generating subunit is used to generate harmonic analysis data based on the subharmonic comparison data and the current harmonic amount.

6. The current transformer according to claim 1, characterized in that: The abnormal element extraction module includes: a correlation analysis unit, a sequence mining unit, an element distance calculation unit, an abnormal boundary threshold construction unit and an abnormal element determination unit; The correlation analysis unit is used to obtain historical abnormal element record data, and perform correlation analysis on the historical abnormal element record data to obtain correlation analysis data; The sequence mining unit is used to perform sequence mining based on the association analysis data using a frequent pattern tree to obtain a target element sequence; The element distance calculation unit is used to determine the strongly correlated elements based on the target element sequence and the correlation analysis data, and calculate the target distance of the strongly correlated elements; The abnormal boundary threshold value construction unit is used to construct a three-phase statistical abnormal boundary threshold value based on the historical abnormal element record data using a kernel density function; The abnormal element determination unit is used to construct an abnormal element baseline based on the three-phase statistic abnormal boundary threshold and the target distance, and determine the target abnormal element using the abnormal element baseline based on the target three-phase relationship parameter.

7. The current transformer according to claim 1, characterized in that: The abnormal component calculation module includes: an abnormal component amplitude calculation unit, an abnormal related component acquisition unit and an abnormal component generation unit; The abnormal component amplitude calculation unit is used to calculate a target difference between a target abnormal element and an abnormal element baseline, and determine the abnormal component amplitude based on the target difference; The abnormal related component acquisition unit is used to calculate the residual component based on the target three-phase relationship parameter, and perform decomposition and cluster analysis on the residual component to obtain the abnormal related component; The abnormal component generating unit is used to generate an abnormal component based on the abnormal component amplitude and the abnormal correlation component.

8. The current transformer according to claim 1, characterized in that: The abnormal feature extraction module includes: a first abnormal feature extraction unit, a first magnetic saturation detection data generation unit, a second abnormal feature determination unit and a target abnormal feature generation unit; The first abnormal feature extraction unit is used to extract features of the abnormal component based on the first abnormal feature extraction model to obtain a first abnormal feature; The first magnetic saturation detection data generating unit is used to calculate the voltage amplitude conversion data and the harmonic slope conversion data based on the operation voltage data extracted from the pre-processed operation data, and determine the first magnetic saturation detection data using a preset judgment coefficient range based on the voltage amplitude conversion data and the harmonic slope conversion data; The second abnormal feature determination unit is used to determine the second magnetic saturation detection data based on the abnormal component by using the full-wave Fourier algorithm combined with the saturation identification partition, and generate a second abnormal feature based on the first magnetic saturation detection data and the second magnetic saturation detection data; The target abnormal feature generating unit is used to generate a target abnormal feature based on the first abnormal feature and the second abnormal feature.

9. The current transformer according to claim 1, characterized in that: The abnormality judgment and early warning module includes: an abnormality recognition model building unit and an early warning judgment unit; The abnormality identification model construction unit is used to construct an abnormality knowledge graph model of the current transformer, and to construct an abnormality identification model based on the abnormality knowledge graph model combined with the abnormality classification model; The early warning judgment unit is used to input the target abnormality feature into the abnormality recognition model to perform abnormality judgment on the current transformer. If it is judged that the current transformer is abnormal, early warning information is generated based on the abnormality judgment result.

10. The current transformer according to claim 9, characterized in that: The abnormality identification model construction unit includes: an abnormality knowledge data classification subunit, an abnormality knowledge graph model construction subunit and an abnormality identification model generation subunit; The abnormal knowledge data classification subunit is used to obtain abnormal knowledge data in the abnormal knowledge base, and classify the abnormal knowledge data to obtain abnormal knowledge data of each category; The abnormal knowledge graph model construction subunit is used to construct an abnormal knowledge graph model based on abnormal knowledge data of various categories combined with abnormal coupling impact degree analysis; The anomaly recognition model generation subunit is used to generate an anomaly recognition model based on an anomaly knowledge graph model and an anomaly classification model composed of a convolutional neural network.

Images