CN113447783B - Voltage transformer insulation fault identification model construction method and device - Google Patents
Voltage transformer insulation fault identification model construction method and device Download PDFInfo
- Publication number
- CN113447783B CN113447783B CN202111007070.6A CN202111007070A CN113447783B CN 113447783 B CN113447783 B CN 113447783B CN 202111007070 A CN202111007070 A CN 202111007070A CN 113447783 B CN113447783 B CN 113447783B
- Authority
- CN
- China
- Prior art keywords
- fault
- identification model
- data set
- voltage transformer
- voltage
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/12—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention relates to a method and a device for constructing an insulation fault identification model of a voltage transformer, which are characterized by firstly collecting secondary voltage data of the voltage transformer in an insulation fault simulation experiment and constructing an insulation fault experiment data set; secondly, performing feature extraction on the insulation fault experiment data set by using a primary voltage consistent relation of a plurality of voltage transformers in phase, eliminating primary voltage fluctuation influence, and acquiring a laboratory fault feature data set by using a principal component analysis method; then, training the laboratory fault feature data set by using a supervised learning algorithm to obtain a fault identification model; and finally, migrating the fault identification model to the actual substation working condition by using a migration learning algorithm to obtain an improved fault identification model. The fault identification model constructed by the invention can realize online real-time monitoring of the insulation state of the voltage transformer without power failure maintenance, grasp the running state of the voltage transformer in real time and reduce the risk that the fault voltage transformer serves an electric power system.
Description
Technical Field
The invention relates to the technical field of power equipment state evaluation methods, in particular to a method and a device for constructing an insulation fault identification model of a voltage transformer.
Background
The voltage transformer is one of key basic devices widely applied to electric data acquisition in an electric power system, and specifically comprises an electromagnetic voltage transformer, a capacitor voltage transformer and an electronic voltage transformer.
The voltage transformer is susceptible to the influence of external environment factors in the operation process to generate a degradation phenomenon, so that the accuracy of electrical data measurement is influenced, even the accurate action of relay protection is influenced in severe cases, and the safe and stable operation of a power system is threatened. Therefore, it is very important to monitor the operation state of the voltage transformer and maintain the fault voltage transformer in time to ensure that the operation state of the voltage transformer meets the relevant regulations.
The traditional voltage transformer detection method is a fixed-period off-line detection method, the method needs planned power failure matching, the detection period is fixed, so that the problems of 'under repair' and 'over repair' coexist, and the operation state of the voltage transformer cannot be mastered in time. To deal with the above-mentioned not enough, utility model patent that application number is CN201820315588.3 provides an insulating on-line monitoring device of capacitive voltage transformer, monitors the insulating state of capacitive voltage transformer body and tail end through external equipment, and the real-time is showing the improvement than traditional method, and this method only qualitative analysis trouble classification, has the limitation of unable quantitative analysis trouble severity nevertheless. In order to solve the problems of incomplete detection amount and low sensitivity of the traditional power failure detection method, the invention patent with the application number of CN201911125797.7 provides an insulation detection method and device for a capacitor voltage transformer, the secondary side current of the capacitor voltage transformer is obtained through an external detection device, and abnormity is judged through the longitudinal difference value of the transformer, and the two methods both need to additionally install equipment on the primary side of the transformer and have the limitation of influencing the physical characteristics of a primary loop.
In the prior art, a method for evaluating the insulation state in the capacitor voltage transformer by using a fault simulation test exists, however, the method only verifies the effectiveness on a fault simulation test platform, and the field application of a substation field is not considered. The voltage transformers on the substation site are different from laboratory models in the aspects of model, parameters and application environment, so that the fault characteristics of the voltage transformers are different between a laboratory and the substation site. Therefore, the precondition for implementing the method is to simulate the fault experiment again on the voltage transformer of the transformer substation site, and the transformer substation site does not have the condition for carrying out the fault simulation experiment, so that the practical applicability of the method is limited. The model established in the laboratory is applied to the transformer substation site in a low-cost and feasible mode, the applicability and the accuracy of the model are improved, and the key problem is solved by the invention.
Disclosure of Invention
The invention provides a method and a device for constructing an insulation fault identification model of a voltage transformer, aiming at the technical problems in the prior art, and the accuracy of identifying the faults of the voltage transformer in different running states is realized by constructing the identification model.
The technical scheme for solving the technical problems is as follows:
on one hand, the invention provides a voltage transformer insulation fault identification model construction method, which comprises the following steps:
s1, collecting secondary voltage data of the voltage transformer in an insulation fault simulation experiment, and constructing an insulation fault experiment data set;
s2, performing feature extraction on the insulation fault experimental data set by using the consistent relation of primary voltages of a plurality of in-phase voltage transformers, eliminating the influence of primary voltage fluctuation, and acquiring a laboratory fault feature data set by using a principal component analysis method;
s3, training the laboratory fault feature data set by using a supervised learning algorithm to obtain a fault identification model;
and S4, migrating the fault identification model to the actual substation working condition by using a migration learning algorithm to obtain an improved fault identification model.
Further, the voltage transformer secondary voltage data includes: fault-free operating state data and fault state simulation data; the fault-free operation state data is used as reference data for feature extraction; the fault state simulation data are based on insulation fault simulation experiments and comprise voltage transformer interlayer breakdown, turn-to-turn breakdown, capacitive voltage divider capacitive breakdown and capacitor dielectric loss abnormity.
Further, the obtaining of the laboratory fault feature data set by using the principal component analysis method includes:
analyzing the secondary voltage data of the voltage transformer in a fault-free operation state by adopting a principal component analysis method, and solving a transfer matrix of residual components;
and solving residual components of the secondary voltage data of the voltage transformer in the fault operation state according to the transfer matrix to form a laboratory fault characteristic data set.
Further, step S3 trains the laboratory fault feature data set by using a random forest classification algorithm, and adjusts model parameters according to the out-of-bag estimation score, thereby obtaining a fault identification model.
Further, the step S4 includes:
constructing a substation field fault characteristic data set;
and fitting the laboratory fault characteristic data set and the substation field fault characteristic data set by adopting a migration component analysis algorithm, obtaining a mapping relation between the laboratory fault characteristic data set and the substation field fault characteristic data set, and migrating the fault identification model to the actual substation working condition to obtain an improved fault identification model.
Further, the constructing of the substation field fault feature data set includes:
acquiring secondary voltage data of a voltage transformer in a normal state in a transformer substation, and solving a transfer matrix of residual components by using a principal component analysis method;
and (4) deducing secondary voltage data under the fault state of the simulation voltage transformer according to a theory, and calculating residual components by using the transfer matrix to form a substation field fault characteristic data set.
On the other hand, the invention provides a voltage transformer insulation fault identification model construction device, which comprises:
the data acquisition module is used for acquiring secondary voltage data of the voltage transformer in an insulation fault simulation experiment and constructing an insulation fault experiment data set;
the fault feature extraction module is used for extracting features of the insulation fault experimental data set by utilizing the consistent relation of primary voltages of the in-phase voltage transformers, eliminating the influence of primary voltage fluctuation and acquiring a laboratory fault feature data set by utilizing a principal component analysis method;
the model training module is used for training the laboratory fault feature data set by using a supervised learning algorithm to obtain a fault identification model;
and the migration module migrates the fault identification model to the actual substation working condition by using a migration learning algorithm to obtain an improved fault identification model.
Further, the obtaining of the laboratory fault feature data set by using the principal component analysis method includes:
analyzing the secondary voltage data of the voltage transformer in a fault-free operation state by adopting a principal component analysis method, and solving a transfer matrix of residual components;
and solving residual components of the secondary voltage data of the voltage transformer in the fault operation state according to the transfer matrix to form a laboratory fault characteristic data set.
Further, the model training module trains the laboratory fault feature data set by adopting a random forest classification algorithm, and adjusts model parameters according to the out-of-bag estimation score, so as to obtain a fault identification model.
Furthermore, the migration module adopts a migration component analysis algorithm to fit the laboratory fault characteristic data set and the substation field fault characteristic data set, obtains a mapping relation between the laboratory fault characteristic data set and the substation field fault characteristic data set, and migrates the fault identification model to an actual substation working condition to obtain an improved fault identification model.
The invention has the beneficial effects that: the fault identification model constructed by the invention can realize online real-time monitoring of the insulation state of the voltage transformer without power failure maintenance, grasp the running state of the voltage transformer in real time, is not influenced by primary load change and active voltage regulation of an electric power system, provides guidance for online insulation fault identification and running maintenance work of the voltage transformer, and reduces the risk that the fault voltage transformer serves the electric power system.
The method disclosed by the invention integrates professional knowledge of the primary side electrical topological structure of the voltage transformer and secondary voltage data to carry out online insulation fault identification, eliminates the influence of primary load change and active voltage regulation of the power system, and has high identification sensitivity and accuracy.
The fault identification model constructed in the laboratory environment is migrated to the transformer substation for field application by means of migration learning, the model is high in generalization performance, can be widely applied to fault identification of multiple voltage transformer groups in the same phase, provides guidance for operation and maintenance of the voltage transformers, reduces labor and time cost brought by periodic maintenance, and greatly reduces the risk of the fault voltage transformers serving a power system.
Drawings
Fig. 1 is a topological diagram of a wiring structure of a plurality of n groups of voltage transformers on the same bus according to an embodiment of the present invention;
fig. 2 is a schematic flow chart of a voltage transformer insulation fault identification model construction method provided in an embodiment of the present invention;
fig. 3 is a schematic structural diagram of a voltage transformer insulation fault identification model construction device according to an embodiment of the present invention.
Detailed Description
The principles and features of this invention are described below in conjunction with the following drawings, which are set forth by way of illustration only and are not intended to limit the scope of the invention.
The embodiment of the invention provides a method for constructing an insulation fault identification model of a voltage transformer. The wiring structure topology of n groups of voltage transformers on the same bus is shown in figure 1.
Specifically, the method comprises the following steps:
s1, carrying out an insulation fault simulation experiment in a laboratory environment, collecting secondary voltage data of a voltage transformer in the insulation fault simulation experiment, and constructing an insulation fault experiment data set;
s2, performing feature extraction on the insulation fault experimental data set by using the consistent relation of primary voltages of a plurality of in-phase voltage transformers, eliminating the influence of primary voltage fluctuation, and acquiring a laboratory fault feature data set by using a principal component analysis method;
s3, training the laboratory fault feature data set by using a supervised learning algorithm to obtain a fault identification model;
and S4, migrating the fault identification model to the actual substation working condition by using a migration learning algorithm to obtain an improved fault identification model.
As shown in the flowchart of fig. 2, in step S1, an insulation fault simulation experiment is designed, secondary voltage data of the voltage transformer is collected, and an insulation fault experiment data set is constructed. The data set includes fault-free operating condition data and fault condition simulation data. Step S2 corresponds to the feature extraction part of fig. 2, and uses the non-fault state data for principal component analysis to obtain a residual component transfer matrix, and converts the fault state simulation data into fault feature data according to the residual component transfer matrix, thereby eliminating the influence of primary voltage fluctuation and constructing a fault feature data set for establishing a fault identification model. Step S3 is to train the fault feature data set by using a supervised learning classification algorithm to construct an insulation fault identification model corresponding to the fault identification model part shown in fig. 2. Step S4 corresponds to the migration learning model building part shown in fig. 2, and is configured to collect a substation field data set of the voltage transformer in a normal operation state, obtain insulation fault data of the voltage transformer through numerical simulation, obtain a substation field fault feature data set through feature extraction, and build a migration learning model between the data set and a fault feature data set built in a laboratory environment, where the migration learning model is used to improve a fault identification model to obtain a final model.
Specifically, the secondary voltage data of the voltage transformer includes: fault-free operating state data and fault state simulation data; the fault-free operation state data is used as reference data for feature extraction; the fault state simulation data are based on insulation fault simulation experiments and comprise voltage transformer interlayer breakdown, turn-to-turn breakdown, capacitive voltage divider capacitive breakdown and capacitor dielectric loss abnormity.
The method for acquiring the laboratory fault characteristic data set by using the principal component analysis method comprises the following steps:
analyzing the secondary voltage data of the voltage transformer in a fault-free operation state by adopting a principal component analysis method, and solving a transfer matrix of residual components;
and solving residual components of the secondary voltage data of the voltage transformer in the fault operation state according to the transfer matrix to form a laboratory fault characteristic data set.
Step S3, training the laboratory fault feature data set by adopting a random forest classification algorithm, adjusting model parameters including the number of base learners and parameters of a decision tree according to the out-of-bag estimation score, and improving the generalization performance of the model so as to obtain a fault identification model.
Further, the step S4 includes:
acquiring secondary voltage data of a voltage transformer in a normal state in a transformer substation, and solving a transfer matrix of residual components by using a principal component analysis method; secondary voltage data under the fault state of the voltage transformer is simulated and deduced according to theory, wherein the secondary voltage data comprises interlayer breakdown, turn-to-turn breakdown, capacitive breakdown of a capacitive voltage divider and dielectric loss abnormity of a capacitor, residual components are calculated by utilizing a transfer matrix, and a transformer substation field fault characteristic data set is formed;
and fitting the laboratory fault characteristic data set and the substation field fault characteristic data set by adopting a migration component analysis algorithm, obtaining a mapping relation between the laboratory fault characteristic data set and the substation field fault characteristic data set, and migrating the fault identification model to the actual substation working condition to obtain an improved fault identification model.
The present embodiment will now be described with reference to specific application examples.
The voltage data of 5 groups of Capacitor Voltage Transformers (CVT) on the same bus of a certain 220kV transformer substation are selected, and the online identification method for the insulation fault of the voltage transformers is implemented. The topological diagram of the wiring structure of a plurality of groups of voltage transformers on the same bus is shown in figure 1.
The implementation steps of the method of the invention are shown in figure 2:
1) the CVT fault simulation experiment is carried out based on a voltage transformer simulation experiment platform in a laboratory environment, and experimental CVT secondary voltage data are collected, wherein the data comprise fault-free operation state data and fault state simulation data. The fault-free operation state data is used as reference data for feature extraction. The fault state simulation data is based on an insulation fault simulation experiment and comprises 1 breakdown of a CVT high-voltage capacitor, 2 breakdown of a high-voltage capacitor, 1 breakdown of a medium-voltage capacitor, 2 breakdown of a medium-voltage capacitor, abnormal high-voltage capacitor dielectric loss and abnormal medium-voltage capacitor dielectric loss. Taking an a-phase CVT capacitance breakdown fault as an example, the CVT secondary voltage data set is shown in table 1, and the fault line number, the fault state of the test item, and the number of samples collected correspondingly are listed in table 1. Meanwhile, in order to facilitate the processing of the supervised learning algorithm and to more clearly and intuitively represent the fault items, the fault states are encoded in the text, and an encoded label corresponding to each fault state is also listed in table 1. The five numbers of the coded label represent the fault condition of 5 sets of a-phase CVTs, respectively, where 0 represents the normal condition, 1 and 2 represent 1 and 2 breakdowns of the high voltage capacitor, respectively, and-1 and-2 represent 1 and 2 breakdowns of the medium voltage capacitor, respectively.
TABLE 1 CVT Fault simulation data
2) As shown in fig. 1, the primary voltage amplitude measured by the in-phase CVT remains the same, as shown by:
the relationship between the amplitudes of the secondary voltages of the in-phase CVT is given by:
whereinε A1，ε B1，ε C1，ε A2，ε B2，ε C2，…，ε An ，ε Bn ，ε Cn Respectively, the ratio errors of the n sets of CVT. From the equation (2), it can be seen that the in-phase relationship of the CVT includes information of the primary voltage and error information of the CVT, and if the characteristic information of the primary voltage and the characteristic information of the error can be decoupled from each other according to the in-phase relationship, the influence of primary voltage fluctuation can be eliminated in the process of establishing the insulation fault identification model, and the fault can be amplifiedAnd the caused error changes information, so that the fault identification model can effectively learn the fault characteristics.
And training the fault-free running state data by adopting a principal component analysis method. In consideration of the consistency of primary voltages of the in-phase CVT, the number of principal components is selected to be 1, and a transfer matrix P of residual components is obtained through principal component analysiseTherefore, the fault characteristic data is obtained, and the calculation method of the fault characteristic data is shown as the following formula:
and E is fault characteristic data representing the fault subjected to characteristic extraction, and X is secondary voltage data of the voltage transformer subjected to standardization. And when the reconstructed data containing the primary side wiring structure knowledge is eliminated from the original data set, a residual error component E representing fault characteristic information is obtained, and the residual error component is used as a data set of a supervised learning model to amplify the fault characteristic information and improve the performance of the model.
3) And fitting the fault feature data set E by adopting a supervised learning classification algorithm random forest, adjusting algorithm parameters including the number of base learners, the maximum depth of a decision tree, the minimum sample number required by node division, the minimum sample number of leaf nodes, the maximum feature number of node division, a node division criterion and the like, and improving the generalization performance of the model. The model generalization performance is determined by the out-of-bag estimation score, and the larger the value, the stronger the model generalization performance is.
The adjustment sequence of model parameter optimization is carried out from high to low according to the influence degree of parameters on the model complexity, the number of the base learners is adjusted firstly, the model test is carried out by taking the initial value as 10 and the step length as 10, and the test result is as follows: when the number of the basis learners is 10, the out-of-bag estimation score is 98.79%, and after the number of the basis learners is greater than or equal to 20, the out-of-bag parameters are all 1.000, so that the number of the basis learners with 20 models is selected, and the model complexity is high and the generalization capability is weak due to the excessively high number of the basis learners. Then, according to the above-mentioned thinking, the maximum depth of the decision tree, the minimum sample number required by node division, the minimum sample number of leaf nodes, the maximum feature number of node division, the node division criterion and the like are adjusted in sequence. The minimum sample number required by node division and the minimum sample number of leaf nodes are correlated with each other, so that the adjustment is performed by adopting a grid search method.
4) Collecting CVT secondary voltage data in a 220kV transformer substation under a normal state, and solving a transfer matrix of residual components by using a principal component analysis method. And (3) according to theory, deducing and simulating secondary voltage data in the fault state of the CVT, wherein the secondary voltage data comprises 1 breakdown of a high-voltage capacitor, 2 breakdown of a high-voltage capacitor, 1 breakdown of a medium-voltage capacitor, 2 breakdown of a medium-voltage capacitor, abnormal dielectric loss of the high-voltage capacitor and abnormal dielectric loss of the medium-voltage capacitor, and calculating residual components by using a transfer matrix. And the residual components form a substation field fault characteristic data set. Fitting a laboratory environment fault characteristic data set and a transformer substation field fault characteristic data set by adopting a migration component analysis algorithm, mapping the two data sets to a space with smaller maximum mean difference through a matrix W, and obtaining a mapped fault characteristic data set as shown in the following formula:
wherein E is1And E2Respectively a source domain data set and a target domain data set, namely a laboratory environment fault characteristic data set and a transformer substation field fault characteristic data set; eN1And EN2Respectively a laboratory environment fault characteristic data set and a transformer substation field fault characteristic data set which are subjected to migration component analysis and mapping. Fitting E by random forest algorithmN1And adjusting the model parameters according to the sequence in the step 3) to obtain the optimal model parameters with the highest out-of-bag estimation score as shown in the table 2.
Using the above model pair EN2The insulation fault identification accuracy of the obtained model is 88.3 percent through prediction, namely the model can effectively identify the same-phase power in the transformer substationInsulation failure of the voltage transformer.
Based on the above method, an embodiment of the present invention further provides a voltage transformer insulation fault identification model building apparatus, as shown in fig. 3, including:
the data acquisition module is used for acquiring secondary voltage data of the voltage transformer in an insulation fault simulation experiment and constructing an insulation fault experiment data set;
the fault feature extraction module is used for extracting features of the insulation fault experimental data set by utilizing the consistent relation of primary voltages of the in-phase voltage transformers, eliminating the influence of primary voltage fluctuation and acquiring a laboratory fault feature data set by utilizing a principal component analysis method;
the model training module is used for training the laboratory fault feature data set by using a supervised learning algorithm to obtain a fault identification model;
and the migration module migrates the fault identification model to the actual substation working condition by using a migration learning algorithm to obtain an improved fault identification model.
Further, the obtaining of the laboratory fault feature data set by using the principal component analysis method includes:
analyzing the secondary voltage data of the voltage transformer in a fault-free operation state by adopting a principal component analysis method, and solving a transfer matrix of residual components;
and solving residual components of the secondary voltage data of the voltage transformer in the fault operation state according to the transfer matrix to form a laboratory fault characteristic data set.
Further, the model training module trains the laboratory fault feature data set by adopting a random forest classification algorithm, and adjusts model parameters according to the out-of-bag estimation score, so as to obtain a fault identification model.
Furthermore, the migration module adopts a migration component analysis algorithm to fit the laboratory fault characteristic data set and the substation field fault characteristic data set, obtains a mapping relation between the laboratory fault characteristic data set and the substation field fault characteristic data set, and migrates the fault identification model to an actual substation working condition to obtain an improved fault identification model.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (8)
1. A voltage transformer insulation fault identification model construction method is characterized by comprising the following steps:
s1, collecting secondary voltage data of the voltage transformer in an insulation fault simulation experiment, and constructing an insulation fault experiment data set; the insulation fault experimental data set comprises fault-free operation state data and fault state simulation data;
s2, using the fault-free state data for principal component analysis to obtain a residual component transfer matrix, converting the fault state simulation data into fault characteristic data according to the residual component transfer matrix, eliminating the influence of primary voltage fluctuation, and constructing a laboratory fault characteristic data set for establishing a fault identification model;
s3, training the laboratory fault feature data set by using a supervised learning algorithm to obtain a fault identification model;
and S4, migrating the fault identification model to the actual substation working condition by using a migration learning algorithm to obtain an improved fault identification model.
2. The method for constructing the insulation fault identification model of the voltage transformer according to claim 1, wherein the secondary voltage data of the voltage transformer comprises: fault-free operating state data and fault state simulation data; the fault-free operation state data is used as reference data for feature extraction; the fault state simulation data are based on insulation fault simulation experiments and comprise voltage transformer interlayer breakdown, turn-to-turn breakdown, capacitive voltage divider capacitive breakdown and capacitor dielectric loss abnormity.
3. The method for constructing the insulation fault identification model of the voltage transformer according to claim 1, wherein step S3 is implemented by training the laboratory fault feature data set by using a random forest classification algorithm, and adjusting model parameters according to the out-of-bag estimation score to obtain the fault identification model.
4. The method for constructing the insulation fault identification model of the voltage transformer according to claim 1, wherein the step S4 comprises:
constructing a substation field fault characteristic data set;
and fitting the laboratory fault characteristic data set and the substation field fault characteristic data set by adopting a migration component analysis algorithm, obtaining a mapping relation between the laboratory fault characteristic data set and the substation field fault characteristic data set, and migrating the fault identification model to the actual substation working condition to obtain an improved fault identification model.
5. The method for constructing the insulation fault identification model of the voltage transformer according to claim 4, wherein the constructing of the field fault feature data set of the substation comprises the following steps:
acquiring secondary voltage data of a voltage transformer in a normal state in a transformer substation, and solving a transfer matrix of residual components by using a principal component analysis method;
and (4) deducing secondary voltage data under the fault state of the simulation voltage transformer according to a theory, and calculating residual components by using the transfer matrix to form a substation field fault characteristic data set.
6. The utility model provides a voltage transformer insulation fault identification model construction device which characterized in that includes:
the data acquisition module is used for acquiring secondary voltage data of the voltage transformer in an insulation fault simulation experiment and constructing an insulation fault experiment data set;
the fault characteristic extraction module is used for using the fault-free state data for principal component analysis to obtain a residual component transfer matrix, converting the fault state simulation data into fault characteristic data according to the residual component transfer matrix, eliminating the influence of primary voltage fluctuation, and constructing a laboratory fault characteristic data set for establishing a fault identification model;
the model training module is used for training the laboratory fault feature data set by using a supervised learning algorithm to obtain a fault identification model;
and the migration module migrates the fault identification model to the actual substation working condition by using a migration learning algorithm to obtain an improved fault identification model.
7. The device for constructing the insulation fault identification model of the voltage transformer according to claim 6, wherein the model training module trains the laboratory fault feature data set by adopting a random forest classification algorithm, and adjusts model parameters according to the out-of-bag estimation score to obtain the fault identification model.
8. The device for constructing the insulation fault identification model of the voltage transformer according to claim 6, wherein the migration module fits the laboratory fault feature data set and the substation field fault feature data set by adopting a migration component analysis algorithm, obtains a mapping relation between the laboratory fault feature data set and the substation field fault feature data set, and migrates the fault identification model to an actual substation working condition to obtain an improved fault identification model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111007070.6A CN113447783B (en) | 2021-08-30 | 2021-08-30 | Voltage transformer insulation fault identification model construction method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111007070.6A CN113447783B (en) | 2021-08-30 | 2021-08-30 | Voltage transformer insulation fault identification model construction method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113447783A CN113447783A (en) | 2021-09-28 |
| CN113447783B true CN113447783B (en) | 2021-11-26 |
Family
ID=77819013
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111007070.6A Active CN113447783B (en) | 2021-08-30 | 2021-08-30 | Voltage transformer insulation fault identification model construction method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113447783B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113884839A (en) * | 2021-11-04 | 2022-01-04 | 国网四川省电力公司电力科学研究院 | Multi-parameter insulation state evaluation method and system for capacitor voltage transformer |
| CN114358092B (en) * | 2022-03-10 | 2022-05-27 | 武汉格蓝若智能技术有限公司 | Method and system for online diagnosis of internal insulation performance of capacitor voltage transformer |
| CN114814501B (en) * | 2022-06-30 | 2022-09-16 | 国网湖北省电力有限公司营销服务中心(计量中心) | On-line diagnosis method for capacitor breakdown fault of capacitor voltage transformer |
| CN116304846B (en) * | 2023-05-24 | 2023-09-12 | 国网江西省电力有限公司电力科学研究院 | CVT internal insulation abnormality online assessment method based on self-supervision learning |
| CN116702630B (en) * | 2023-08-07 | 2023-11-10 | 国网安徽省电力有限公司营销服务中心 | Error iterative evaluation method for capacitive voltage transformer CVT |
| CN117192371B (en) * | 2023-11-03 | 2024-01-30 | 南通清浪智能科技有限公司 | Test method and system for motor driver of new energy automobile |
| CN117849558B (en) * | 2024-03-06 | 2024-05-24 | 武汉格蓝若智能技术股份有限公司 | Method and device for on-line monitoring breakdown fault of main capacitor of capacitive voltage transformer |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2001268780A (en) * | 2000-03-16 | 2001-09-28 | Kawasaki Steel Corp | Transformer and detecting method for its insulation failure |
| KR101413788B1 (en) * | 2012-12-27 | 2014-06-30 | 주식회사 효성 | Method and apparratus of malfunction detection of transformer |
| CN106096076A (en) * | 2016-05-26 | 2016-11-09 | 国网江苏省电力公司检修分公司 | Capacitance type potential transformer operation troubles analogy method based on PSCAD |
| CN109597396A (en) * | 2018-11-26 | 2019-04-09 | 国网湖北省电力有限公司电力科学研究院 | A kind of distribution transforming on-line fault diagnosis method based on high amount of traffic and transfer learning |
| CN109752685A (en) * | 2019-03-12 | 2019-05-14 | 中国电力科学研究院有限公司 | A kind of appraisal procedure, the system of distribution mutual inductor insulation performance |
| CN109902735A (en) * | 2019-02-22 | 2019-06-18 | 武汉格蓝若智能技术有限公司 | A kind of two norm Independent Component Analysis for realizing that threephase potential transformer kinematic error is assessed in substation |
| CN112329821A (en) * | 2020-10-23 | 2021-02-05 | 苏州鑫睿益荣信息技术有限公司 | Intelligent diagnosis system for power transformer and diagnosis method based on decision tree classification |
| CN112485556A (en) * | 2020-11-03 | 2021-03-12 | 国电南瑞南京控制系统有限公司 | CVT fault detection method and system based on transformer substation monitoring system and storage medium |
| CN112884070A (en) * | 2021-03-17 | 2021-06-01 | 云南电网有限责任公司电力科学研究院 | High-voltage switch fault diagnosis method based on transfer learning |
-
2021
- 2021-08-30 CN CN202111007070.6A patent/CN113447783B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2001268780A (en) * | 2000-03-16 | 2001-09-28 | Kawasaki Steel Corp | Transformer and detecting method for its insulation failure |
| KR101413788B1 (en) * | 2012-12-27 | 2014-06-30 | 주식회사 효성 | Method and apparratus of malfunction detection of transformer |
| CN106096076A (en) * | 2016-05-26 | 2016-11-09 | 国网江苏省电力公司检修分公司 | Capacitance type potential transformer operation troubles analogy method based on PSCAD |
| CN109597396A (en) * | 2018-11-26 | 2019-04-09 | 国网湖北省电力有限公司电力科学研究院 | A kind of distribution transforming on-line fault diagnosis method based on high amount of traffic and transfer learning |
| CN109902735A (en) * | 2019-02-22 | 2019-06-18 | 武汉格蓝若智能技术有限公司 | A kind of two norm Independent Component Analysis for realizing that threephase potential transformer kinematic error is assessed in substation |
| CN109752685A (en) * | 2019-03-12 | 2019-05-14 | 中国电力科学研究院有限公司 | A kind of appraisal procedure, the system of distribution mutual inductor insulation performance |
| CN112329821A (en) * | 2020-10-23 | 2021-02-05 | 苏州鑫睿益荣信息技术有限公司 | Intelligent diagnosis system for power transformer and diagnosis method based on decision tree classification |
| CN112485556A (en) * | 2020-11-03 | 2021-03-12 | 国电南瑞南京控制系统有限公司 | CVT fault detection method and system based on transformer substation monitoring system and storage medium |
| CN112884070A (en) * | 2021-03-17 | 2021-06-01 | 云南电网有限责任公司电力科学研究院 | High-voltage switch fault diagnosis method based on transfer learning |
Non-Patent Citations (1)
| Title |
|---|
| 深度迁移学习算法及其应用研究;付家慧;《中国优秀博硕士学位论文全文数据库（硕士）·信息科技辑》;20210215(第02期);正文第1-52页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113447783A (en) | 2021-09-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113447783B (en) | Voltage transformer insulation fault identification model construction method and device | |
| Setayeshmehr et al. | On-line monitoring and diagnoses of power transformer bushings | |
| Niţu et al. | Power transformer fault diagnosis using fuzzy logic technique based on dissolved gas analysis and furan analysis | |
| CN111929527B (en) | Adaptability evaluation system and method for equipment fault characteristic regularization knowledge representation | |
| Abbas et al. | Single line to ground fault detection and location in medium voltage distribution system network based on neural network | |
| Coban et al. | Artificial neural network based fault location on 230 kv transmission line using voltage and current signals | |
| CN117668751B (en) | High-low voltage power system fault diagnosis method and device | |
| Rajamani et al. | Identification of simultaneously occurring dynamic disc-to-disc insulation failures in transformer winding under impulse excitation | |
| CN117131786B (en) | Voltage transformer insulation fault online identification method | |
| Yanming et al. | DGA based insulation diagnosis of power transformer via ANN | |
| CN105741184A (en) | Transformer state evaluation method and apparatus | |
| Gao et al. | Research on the calculation method for the parameters of the simplified Schavemaker AC arc model | |
| Tippannavar et al. | Smart transformer-An analysis of recent technologies for monitoring transformer | |
| CN115951292A (en) | Electronic voltage transformer error state online evaluation method and device | |
| Tran | Integration of neural network and distance relay to improve the fault localization on transmission lines | |
| CN114779029A (en) | CVT internal insulation online evaluation method fusing group redundancy association and structural parameters | |
| BEYZA et al. | Fault type estimation in power systems | |
| Ding et al. | On-line transformer winding's fault monitoring and condition assessment | |
| Ansari et al. | Moisture evaluation of OIP bushings using artificial neural networks and dielectric frequency response | |
| Qian et al. | Fault Diagnosis of On-Load Tap-Changer Based on the Parameter-adaptive VMD and SA-ELM | |
| Ma et al. | Fault Identification in Power Distribution Systems based on Domain Adaptation Probabilistic Learning | |
| Talib et al. | Fault Identification of In-Service Power Transformer using Depolarization Current Analysis. | |
| Li et al. | Fault diagnosis of power cables based on wavelet transform and CNN-informer | |
| Wang et al. | Study on fault diagnosis method of transformer using multi-neural network and evidence theory | |
| Moniri et al. | MODELING THE FREQUENCY RESPONSE MOVEMENT IN POWER TRANSFORMERS FOR PREDICTING PURPOSES |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP01 | Change in the name or title of a patent holder | ||
| CP01 | Change in the name or title of a patent holder |
Address after: Room 1803-1805, building 2-07, guanggu.core center, 303 Guanggu Avenue, Donghu New Technology Development Zone, Wuhan City, Hubei Province, 430000 Patentee after: Wuhan Gelanruo Intelligent Technology Co.,Ltd. Address before: Room 1803-1805, building 2-07, guanggu.core center, 303 Guanggu Avenue, Donghu New Technology Development Zone, Wuhan City, Hubei Province, 430000 Patentee before: WUHAN GLORY ROAD INTELLIGENT TECHNOLOGY Co.,Ltd. |