CN112880750A - Transformer multidimensional comprehensive online monitoring intelligent diagnosis system - Google Patents
Transformer multidimensional comprehensive online monitoring intelligent diagnosis system Download PDFInfo
- Publication number
- CN112880750A CN112880750A CN202110324633.8A CN202110324633A CN112880750A CN 112880750 A CN112880750 A CN 112880750A CN 202110324633 A CN202110324633 A CN 202110324633A CN 112880750 A CN112880750 A CN 112880750A
- Authority
- CN
- China
- Prior art keywords
- transformer
- layer
- diagnosis
- edge calculation
- sensor
- 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.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01D—MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
- G01D21/00—Measuring or testing not otherwise provided for
- G01D21/02—Measuring two or more variables by means not covered by a single other subclass
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Testing Relating To Insulation (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention provides a transformer multi-dimensional comprehensive online monitoring intelligent diagnosis system which comprises a perception sensing layer, an edge calculation layer, a deep learning layer, a cloud feedback layer and a user control layer, wherein the perception sensing layer is used for sensing the position of a transformer; the sensing layer comprises a plurality of intelligent sensors and is used for acquiring a plurality of dimensional parameters of the transformer; the edge calculation layer comprises an edge calculation terminal, and the sensing layer sends the plurality of dimensional data to the edge calculation terminal to execute edge calculation; the deep learning layer performs summary analysis on a plurality of edge calculation results of the edge calculation terminal, and determines whether to provide a feedback signal to the cloud feedback layer or not based on the summary analysis result; the cloud feedback layer updates an edge calculation model of the edge calculation terminal based on the feedback signal; and the user control layer is used for controlling the sensing state of the sensing layer and identifying the state of the transformer based on the edge calculation result. The invention can comprehensively and intelligently diagnose based on multidimensional data and update the edge model in a self-learning way.
Description
Technical Field
The invention belongs to the technical field of intelligent monitoring of transformers, and particularly relates to a multi-dimensional comprehensive online monitoring intelligent diagnosis system for a transformer.
Background
The reliability of the power transformer, which is one of the most important devices in the power grid, is directly related to whether the power grid can be operated safely, efficiently and economically. Reducing transformer faults means increasing the economic efficiency of the grid. Since the transformer is continuously operated in the power grid for a long time, various faults and accidents inevitably occur. Analysis and monitoring of the causes of these faults and accidents have been a hot issue for many years for transformer design, operation and maintenance personnel.
The common transformer faults have various testing and diagnosing measures, such as measurement of direct current resistance, chromatographic analysis, partial discharge test of insulation performance test, measurement of insulation resistance and absorption ratio, measurement of dielectric loss tangent value, measurement of direct current leakage current, low-voltage short-circuit impedance test of winding deformation test and the like, and basically go through three stages of sensory diagnosis, experimental test and intelligent detection.
For example, high pressure and high temperature may decompose insulating oil of a transformer to generate various low molecular hydrocarbons and gases such as hydrogen, carbon dioxide, and carbon monoxide; when latent overheating or discharge failure occurs, the generation speed of the gases is rapidly accelerated, the decomposed gases are dissolved in the oil, and the composition and the content of the decomposed gases have certain relations with the type, the property and the severity of the failure; for example, the sound emitted from the transformer body varies under different operating conditions of the transformer, and the sound variation generated during the transformer fault operation is used to diagnose the fault condition of the transformer, which is called an acoustic diagnosis method.
The Chinese invention patent application with the application number of CN202011209201.4 provides a system and a method for identifying the fault mode of a dry-type transformer with multi-information edge calculation, wherein the system comprises an intelligent sensing unit, a network transmission unit, a data processing unit, a data storage unit and an intelligent application unit; according to the method, the mass data of the edge equipment is processed by using the edge computing method, the computing amount of a cloud computing center is reduced, the difficulty in information fusion of multiple information and large data volume is solved, the fault mode of the dry-type transformer is more accurately identified, the understanding of the running state and the fault state of the dry-type transformer is enhanced, and the overhaul and maintenance work is effectively guided.
The Chinese patent application with the application number of CN202010280306.2 provides a transformer fault diagnosis method based on a deep support vector machine, which comprises the following steps: (1) analyzing the collected dissolved gas in the power transformer oil to obtain original data of the volume fraction of the characteristic gas components, taking 70% of the original data as a training set, and taking the remaining 30% as a test set, wherein the original data is used for testing the accuracy of the obtained fault diagnosis model; (2) carrying out normalization processing on the original data; (3) and (3) optimizing the support vector machine parameters by utilizing the gray wolf algorithm according to the normalized data obtained in the step (2) to obtain optimal parameters for transformer fault diagnosis, and training a support vector machine model Mr through a limited Boltzmann machine in deep learning to obtain an optimal model for transformer fault diagnosis. The method can further improve the efficiency of transformer fault diagnosis and effectively increase the accuracy of fault diagnosis.
However, the various diagnostic protocols mentioned in the prior art have certain drawbacks. Firstly, judging by only adopting a single variable, wherein the result has deviation; in addition, the adopted model is a static model and cannot be updated along with the actual situation; finally, the actual precedence and fault dependency are not considered in the process of testing or acquiring signals, higher hardware cost is needed, energy consumption is higher, and economic benefit is reduced.
Disclosure of Invention
In order to solve the technical problems, the invention provides a transformer multi-dimensional comprehensive online monitoring intelligent diagnosis system which comprises a perception sensing layer, an edge calculation layer, a deep learning layer, a cloud feedback layer and a user control layer; the sensing layer comprises a plurality of intelligent sensors and is used for acquiring a plurality of dimensional parameters of the transformer; the edge calculation layer comprises an edge calculation terminal, and the sensing layer sends the plurality of dimensional data to the edge calculation terminal to execute edge calculation; the deep learning layer performs summary analysis on a plurality of edge calculation results of the edge calculation terminal, and determines whether to provide a feedback signal to the cloud feedback layer or not based on the summary analysis result; the cloud feedback layer updates an edge calculation model of the edge calculation terminal based on the feedback signal; and the user control layer is used for controlling the sensing state of the sensing layer and identifying the state of the transformer based on the edge calculation result. The invention can comprehensively and intelligently diagnose based on multidimensional data and update the edge model in a self-learning way.
As a summary, the system comprises a perception sensing layer, an edge calculation layer, a deep learning layer, a cloud feedback layer and a user control layer;
the sensing device comprises a sensing layer, an edge calculation layer and a sensing layer, wherein the sensing layer is used for sensing a plurality of signals of the transformer;
the deep learning layer collects a plurality of edge calculation results of the edge calculation layer through a collection link and communicates with the cloud feedback layer;
the user control layer may be in communication with the cloud feedback layer and regulate the perception sensing layer.
In a specific implementation, the sensing layer comprises a plurality of intelligent sensors, and the intelligent sensors are used for acquiring a plurality of dimensional parameters of the transformer; the intelligent sensor comprises an acoustic sensor, an electric energy sensor and an oil-gas combined sensor;
the edge computing layer comprises an edge computing terminal, the edge computing terminal is communicated with the sensing layer, and the sensing layer sends the plurality of dimensional data to the edge computing terminal to execute edge computing;
the deep learning layer carries out summary analysis on a plurality of edge calculation results of the edge calculation terminal, and determines whether to provide a feedback signal to the cloud feedback layer or not based on the summary analysis result;
the cloud feedback layer updates an edge calculation model of the edge calculation terminal based on the feedback signal;
and the user control layer is used for controlling the sensing state of the sensing layer and identifying the state of the transformer based on the edge calculation result.
A first transformer diagnosis model and a second transformer diagnosis model are arranged in the edge computing terminal;
the input semaphore of the first transformer diagnostic model comprises a plurality of characteristic sound signals;
the input signal quantity of the second transformer diagnosis model comprises a plurality of characteristic gas signals.
The acoustic sensor acquires a plurality of characteristic sound signals of the transformer according to a first preset period and sends the plurality of characteristic signals to the edge computing terminal;
the edge computing terminal inputs the characteristic sound signal into a first transformer diagnosis model after preprocessing the characteristic sound signal to obtain a first diagnosis result;
after the oil-gas combined sensor is activated, continuously collecting various characteristic gas signals in the transformer oil, and sending the various characteristic gas signals to the edge computing terminal;
and the edge computing terminal inputs the characteristic gas signal into a second transformer diagnosis model after preprocessing the characteristic gas signal to obtain a second diagnosis result.
And if the difference degree of the first diagnosis result and the second diagnosis result obtained in two continuous periods exceeds a preset value, providing a feedback signal to the cloud feedback layer.
The cloud feedback layer updates the first transformer diagnosis model or updates the second transformer diagnosis model based on the feedback signal.
Compared with the prior art, the intelligent transformer state monitoring system adopts various data with different dimensions, and realizes intelligent transformer state monitoring;
in addition, the invention does not simply combine the single diagnosis models or single diagnosis indexes in the prior art as multi-dimension, but endows different precedence priorities;
more importantly, the invention uses the local edge computing layer, so that the data interaction between the field and the cloud is greatly reduced;
meanwhile, in order to avoid the failure or error of the local model, even if the diagnostic model used by the edge computing terminal is used, the self-learning update can be continuously obtained on the basis of the real-time or periodic computing result;
in addition, most of the main sensors (the electric energy sensor and the oil-gas combined sensor) are in a dormant state when not receiving an activation signal, so that the loss can be reduced, and the cost can be saved.
Further advantages of the invention will be apparent in the detailed description section in conjunction with the drawings attached hereto.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without creative efforts.
FIG. 1 is a diagram of the actual equipment components of the transformer multi-dimensional integrated online monitoring intelligent diagnosis system according to an embodiment of the present invention
FIG. 2 is an abstract data hierarchy diagram of a transformer multi-dimensional integrated online monitoring intelligent diagnosis system according to an embodiment of the present invention
FIG. 3 is a flow chart of specific data processing that may be further implemented based on the hierarchy diagram of FIG. 1
FIG. 4 is a flow chart of the system of FIG. 1 for performing a diagnostic procedure on a primary signature sound signal
FIG. 5 is a schematic diagram of a second transformer diagnostic model implemented by the system of FIG. 1
Detailed Description
The invention is further described with reference to the following drawings and detailed description.
Referring to fig. 1, a diagram of an actual device of a transformer multi-dimensional integrated online monitoring intelligent diagnosis system according to an embodiment of the present invention is shown.
In fig. 1, a plurality of sensing devices of the sensing layer communicate with the transformer, perform sensing monitoring on various signals of the transformer, and transmit sensing monitoring data to the edge calculation layer;
the deep learning layer collects a plurality of edge calculation results of the edge calculation layer through a collection link and communicates with the cloud feedback layer;
the user control layer may be in communication with the cloud feedback layer and regulate the perception sensing layer.
Based on fig. 1, fig. 2 shows an abstract data hierarchy architecture diagram of the transformer multi-dimensional comprehensive online monitoring intelligent diagnosis system.
In fig. 2, the system includes a sensing layer, an edge calculation layer, a deep learning layer, a cloud feedback layer, and a user control layer;
the perception sensing layer sends the dimensional data to the edge calculation layer to execute edge calculation;
the deep learning layer carries out summary analysis on a plurality of edge calculation results of the edge calculation layer, and determines whether to provide a feedback signal to the cloud feedback layer or not based on the summary analysis result;
more specifically, the edge computation layer includes an edge computation terminal that includes an edge computation model.
The edge calculation is an open platform fusing network, calculation, storage and application core capabilities at the edge side of a network close to an object or a data source, so that edge intelligent interconnection services are provided nearby, and the key requirements of industry digitization on aspects of agile connection, real-time service, data optimization, application intelligence, safety, privacy protection and the like are met.
The edge calculation terminal is a terminal device that can perform the edge calculation.
The deep learning layer carries out summary analysis on a plurality of edge calculation results of the edge calculation layer, analyzes whether the edge calculation model is abnormal or not based on the summary analysis result, and provides a feedback signal for the cloud feedback layer if the edge calculation model is abnormal.
On the basis of the edge calculation results, whether the diagnosis result is normal or not can be judged, and if the diagnosis result is normal, a normal diagnosis signal is sent to the user control layer; and the user control layer adjusts a first preset period of the perception sensing layer based on the diagnosis normal signal.
On the basis of fig. 1-2, referring to fig. 3, fig. 3 shows a specific framework and a data flow diagram of the technical solution of the present invention.
Based on fig. 1 to fig. 3, the following embodiments of the present application can be described in detail:
the sensing layer comprises a plurality of intelligent sensors, and the intelligent sensors are used for acquiring a plurality of dimensional parameters of the transformer;
as shown in fig. 3, the smart sensor includes an acoustic sensor, an electric power sensor, and an oil-gas combination sensor.
The edge computing layer comprises a plurality of edge computing terminals, the edge computing terminals are communicated with the sensing layer, and the sensing layer sends the plurality of dimensional data to the edge computing terminals to execute edge computing;
the deep learning layer carries out summary analysis on a plurality of edge calculation results of the edge calculation terminal, and determines whether to provide a feedback signal to the cloud feedback layer or not based on the summary analysis result;
the cloud feedback layer updates an edge calculation model of the edge calculation terminal based on the feedback signal;
and the user control layer is used for controlling the sensing state of the sensing layer and identifying the state of the transformer based on the edge calculation result.
More specifically, the acoustic sensor is used for acquiring a characteristic sound signal of the transformer;
the electric energy sensor is used for acquiring electric energy parameters of the transformer, and the electric energy parameters comprise three-phase voltage values and/or three-phase current values of the transformer;
the oil-gas combined sensor comprises an oil level and oil temperature sensor and a characteristic gas sensor, wherein the oil level and oil temperature sensor is used for acquiring the oil level value and the oil temperature value of the transformer oil, and the characteristic gas sensor is used for acquiring various characteristic gas signals in the transformer oil.
In various embodiments of the present invention, the characteristic sound and the characteristic gas refer to a sound wave with a specific frequency/waveband closely related to an abnormal state of the transformer and a specific trace gas element.
As a general understanding, technicians with a lot of experience in transformer overhaul operation work for many years can judge the operation state of the transformer according to the sound emitted by the transformer. The operation state of the transformer is judged by listening to the sound emitted from the inside of the transformer by the ear close to the box body of the transformer.
The detection of internal faults of transformers by using acoustic signals emitted by the transformers has become a hot point of research. The method is characterized in that a plurality of methods are applied, namely an ultrasonic analysis method, a noise analysis method and a vibration analysis method, the methods can detect sound waves with specific frequency/wave band, and then the transformer sound signal of the transformer substation which normally operates is actually measured on site, so that a foundation is provided for identifying the characteristics of the fault sound signal, then a model of typical spark discharge fault inside the transformer is built, and a mathematical model of the spark discharge sound signal is built, namely whether the fault exists or not can be judged or a diagnosis result can be obtained by inputting the detected sound waves with specific frequency/wave band.
Since one of the improvements of the present invention is to use multi-dimensional data instead of a separate acoustic wave detection model itself, the specific construction of the above acoustic wave-fault detection model is not specifically developed, and those skilled in the art can adopt various acoustic wave detection models for transformer fault detection or condition diagnosis known in the art, including the following prior arts:
Bartoletti C,Desiderio M,Di Carlo D,et al.Vibro-Acoustic Techniques To Diagnose Power Transformers[J].IEEE Transactions On Power Delivery.2004,19(1):221-229.
Berler Z,Golubev A,Patterson C.Vibro-Acoustic Method of Transformer Clamping Pressure Monitoring[Z].Anaheim CA USA:2000.
Mullerova E.Acoustic Method in Diagnostics of ransformer Insulation[Z].Potsdam Germany:2010。
similarly, when the inside of the transformer is abnormal, the insulating oil generates abnormal gas, and the analysis of the gas can predict and diagnose the fault. Dissolved gas in transformer oil is used as characteristic gas as a model input variable for monitoring the state of the transformer on line, and the prior art has a plurality of records, namely, a diagnosis model using various characteristic gases as input variables is established, so that the normal or abnormal state of the transformer can be correspondingly detected.
As an illustrative example, the characteristic gas signal can be the content values of the characteristic gas ratios C2H2/C2H4, CH2/H2, C2H4/C2H6 in the transformer oil as characteristic values.
Further examples can be found in the following prior art:
yondon qui, analysis and diagnosis of gases in transformer oil, journal of the Hubei press 1987;
guide rule for analyzing and judging dissolved gas in standard transformer oil in power industry of people's republic of China
DL/T722-2000 is implemented by the national economic and trade Committee of the people's republic of China 2000-11-03 standard 2001-01-01;
All M et al Measuring and understanding the aging of Kraft insulating paper in power transformers IEEE Trans EI 12(3)1996,p605-p608。
Yukinori Suzuki,Self-Organizing QRS-Wave Recognition in ECG Using Neural Networks,IEEE Trans on NN,Vol.6,No.6,1995,p1469-p1477。
continuing with FIG. 3, the following is presented:
a first transformer diagnosis model and a second transformer diagnosis model are arranged in the edge computing terminal; the input semaphore of the first transformer diagnostic model comprises a plurality of characteristic sound signals; the input signal quantity of the second transformer diagnosis model comprises a plurality of characteristic gas signals.
The input semaphore of the first transformer diagnostic model comprises a plurality of characteristic sound signals, which is the first-stage model of the multi-dimensional comprehensive detection of the invention.
Figure 4 shows the basic process flow of such a model.
In this embodiment, a plurality of characteristic sound signals of the transformer are periodically acquired by the acoustic sensor.
More specifically, the acoustic sensor acquires a plurality of characteristic sound signals of the transformer according to a first preset period, and sends the plurality of characteristic signals to the edge computing terminal;
the edge computing terminal inputs the characteristic sound signal into a first transformer diagnosis model after preprocessing the characteristic sound signal to obtain a first diagnosis result;
when the first diagnosis result meets a first preset condition, the edge computing terminal sends a first activation signal, and the first activation signal activates the electric energy sensor and/or the oil-gas combination sensor.
It should be noted that the edge calculation focuses on the analysis of real-time and short-period data, so as to better support the real-time intelligent processing and execution of local services.
Fig. 4 shows a part of the preprocessing steps, including extracting characteristic signals, physical quantity processing, singularity detection, characteristic signal identification, etc., and then inputting to the first transformer diagnosis model.
The first diagnosis result output after being input into the first transformer diagnosis model comprises normal and abnormal results,
if normal, the detection is continued.
In the present embodiment, the satisfaction of the first predetermined condition may be that the diagnosis result is abnormal;
when the first diagnosis result does not meet a first preset condition, sending a normal diagnosis signal to the user control layer;
the user control layer controls the sensing state of the sensing layer, and the method comprises the following steps:
and the user control layer adjusts the first preset period based on the diagnosis normal signal.
More specifically, the first preset period may be increased.
It is noted that, in various embodiments of the present invention, the electric power sensor and the oil-gas combination sensor are in a sleep state when not receiving an activation signal, and the edge computing terminal transmits a first activation signal that activates the electric power sensor and/or the oil-gas combination sensor only when the first diagnostic result satisfies a first predetermined condition.
Therefore, most of the main sensors (the electric energy sensor and the oil-gas combined sensor) are in a dormant state when not receiving an activation signal, so that the loss can be reduced, and the cost can be saved.
After the oil-gas combined sensor is activated, continuously collecting various characteristic gas signals in the transformer oil, and sending the various characteristic gas signals to the edge computing terminal;
and the edge computing terminal inputs the characteristic gas signal into a second transformer diagnosis model after preprocessing the characteristic gas signal to obtain a second diagnosis result.
As mentioned above, the input of the second diagnostic model may be a plurality of characteristic gas signals, including, for example, the characteristic gas ratios C2H2/C2H4, CH2/H2, C2H4/C2H6 in transformer oil.
Preferably, the second diagnostic model described in this embodiment is a probabilistic neural network model (PNN), and fig. 5 shows a basic structure thereof.
The probabilistic neural network PNN is a neural network which can be used for pattern classification, and the essence of the probabilistic neural network PNN is a parallel algorithm developed based on Bayes minimum risk criterion.
FIG. 5 shows a PNN structure with input samples divided into two classes, and the hierarchical model of the PNN is a forward network with four layers of PNN, namely an input layer, a mode layer, an accumulation layer and an output layer, which are proposed by Specht according to Bayesian classification rules and a probability density function of Parzen. When the network works, a sample X to be identified is directly sent to each node of a mode layer from an input layer; each mode layer node performs weighted summation on the transmitted input signals, and then transmits the input signals into an accumulation layer after a nonlinear operator operation.
In the above-described embodiment, as another advantage of embodying the improvement of the present invention, the plurality of edge calculation results include the first diagnosis result and the second diagnosis result;
the deep learning layer collects and analyzes a plurality of edge calculation results of the edge calculation terminal, and determines whether to provide a feedback signal to the cloud feedback layer based on the collected and analyzed results, and the method specifically comprises the following steps:
and if the difference degree of the first diagnosis result and the second diagnosis result obtained in two continuous periods exceeds a preset value, providing a feedback signal to the cloud feedback layer.
The cloud feedback layer updates the first transformer diagnosis model or updates the second transformer diagnosis model based on the feedback signal.
Therefore, in the technical scheme of the invention, even the diagnostic model used by the edge computing terminal can be continuously updated by self-learning based on real-time or periodic computing results.
On the basis, the user control layer identifies the state of the transformer based on the first diagnosis result, the second diagnosis result and the electric energy parameter of the transformer acquired by the electric energy sensor.
The identification can be expert intervention identification based on manual intervention or automatic alarm identification based on rules.
For example, the rule may be set to determine that there is an abnormality in the transformer when the degree of reliability of the first diagnostic result exceeds a first predetermined value and at least two classification values of the second diagnostic result are determined to be the occurrence of the abnormality.
Of course, other rules may be preset, and the present invention is not limited to this.
Therefore, the intelligent transformer state monitoring system adopts various data with different dimensions, and realizes intelligent transformer state monitoring; in addition, the invention does not simply combine the single diagnosis models or single diagnosis indexes in the prior art as multi-dimension, but endows different precedence priorities; more importantly, the invention uses the local edge computing layer, so that the data interaction between the field and the cloud is greatly reduced; meanwhile, in order to avoid the failure or error of the local model, even the diagnosis model used by the edge computing terminal can be continuously updated by self-learning based on the real-time or periodic computing result.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (10)
1. A transformer multi-dimensional comprehensive online monitoring intelligent diagnosis system comprises a perception sensing layer, an edge calculation layer, a deep learning layer, a cloud feedback layer and a user control layer;
the method is characterized in that:
the sensing layer comprises a plurality of intelligent sensors, and the intelligent sensors are used for acquiring a plurality of dimensional parameters of the transformer; the intelligent sensor comprises an acoustic sensor, an electric energy sensor and an oil-gas combined sensor;
the edge computing layer comprises an edge computing terminal, the edge computing terminal is communicated with the sensing layer, and the sensing layer sends the plurality of dimensional data to the edge computing terminal to execute edge computing;
the deep learning layer carries out summary analysis on a plurality of edge calculation results of the edge calculation terminal, and determines whether to provide a feedback signal to the cloud feedback layer or not based on the summary analysis result;
the cloud feedback layer updates an edge calculation model of the edge calculation terminal based on the feedback signal;
and the user control layer is used for controlling the sensing state of the sensing layer and identifying the state of the transformer based on the edge calculation result.
2. The transformer multidimensional integrated online monitoring intelligent diagnosis system of claim 1, wherein:
the acoustic sensor is used for acquiring a characteristic sound signal of the transformer;
the electric energy sensor is used for acquiring electric energy parameters of the transformer, and the electric energy parameters comprise three-phase voltage values and/or three-phase current values of the transformer;
the oil-gas combined sensor comprises an oil level and oil temperature sensor and a characteristic gas sensor, wherein the oil level and oil temperature sensor is used for acquiring the oil level value and the oil temperature value of the transformer oil, and the characteristic gas sensor is used for acquiring various characteristic gas signals in the transformer oil.
3. The transformer multidimensional integrated online monitoring intelligent diagnosis system of claim 1, wherein:
a first transformer diagnosis model and a second transformer diagnosis model are arranged in the edge computing terminal;
the input semaphore of the first transformer diagnostic model comprises a plurality of characteristic sound signals;
the input signal quantity of the second transformer diagnosis model comprises a plurality of characteristic gas signals.
4. The transformer multidimensional integrated online monitoring intelligent diagnosis system as claimed in claim 1, 2 or 3, wherein:
the acoustic sensor periodically collects a plurality of characteristic sound signals of the transformer;
the electric energy sensor and the oil-gas combined sensor are in a dormant state when not receiving an activation signal.
5. The transformer multidimensional integrated online monitoring intelligent diagnosis system as claimed in claim 1, 2 or 3, wherein:
the acoustic sensor acquires a plurality of characteristic sound signals of the transformer according to a first preset period and sends the plurality of characteristic signals to the edge computing terminal;
the edge computing terminal inputs the characteristic sound signal into a first transformer diagnosis model after preprocessing the characteristic sound signal to obtain a first diagnosis result;
when the first diagnosis result meets a first preset condition, the edge computing terminal sends a first activation signal, and the first activation signal activates the electric energy sensor and/or the oil-gas combination sensor.
6. The transformer multidimensional integrated online monitoring intelligent diagnosis system of claim 5, wherein:
after the oil-gas combined sensor is activated, continuously collecting various characteristic gas signals in the transformer oil, and sending the various characteristic gas signals to the edge computing terminal;
and the edge computing terminal inputs the characteristic gas signal into a second transformer diagnosis model after preprocessing the characteristic gas signal to obtain a second diagnosis result.
7. The transformer multidimensional integrated online monitoring intelligent diagnosis system of claim 6, wherein:
the plurality of edge calculation results includes the first diagnostic result and the second diagnostic result;
the deep learning layer collects and analyzes a plurality of edge calculation results of the edge calculation terminal, and determines whether to provide a feedback signal to the cloud feedback layer based on the collected and analyzed results, and the method specifically comprises the following steps:
and if the difference degree of the first diagnosis result and the second diagnosis result obtained in two continuous periods exceeds a preset value, providing a feedback signal to the cloud feedback layer.
8. The transformer multidimensional integrated online monitoring intelligent diagnosis system as claimed in claim 6 or 7, wherein:
the edge calculation model of the edge calculation terminal comprises a first transformer diagnosis model and a second transformer diagnosis model;
the cloud feedback layer updates the first transformer diagnosis model or updates the second transformer diagnosis model based on the feedback signal.
9. The transformer multidimensional integrated online monitoring intelligent diagnosis system of claim 5, wherein:
when the first diagnosis result does not meet a first preset condition, sending a normal diagnosis signal to the user control layer;
the user control layer controls the sensing state of the sensing layer, and the method comprises the following steps:
and the user control layer adjusts the first preset period based on the diagnosis normal signal.
10. The transformer multidimensional integrated online monitoring intelligent diagnosis system of claim 6, wherein:
the identifying, by the user control layer, the state of the transformer based on the edge calculation result specifically includes:
the user control layer identifies the state of the transformer based on the first diagnosis result, the second diagnosis result and the electric energy parameter of the transformer acquired by the electric energy sensor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110324633.8A CN112880750A (en) | 2021-03-26 | 2021-03-26 | Transformer multidimensional comprehensive online monitoring intelligent diagnosis system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110324633.8A CN112880750A (en) | 2021-03-26 | 2021-03-26 | Transformer multidimensional comprehensive online monitoring intelligent diagnosis system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112880750A true CN112880750A (en) | 2021-06-01 |
Family
ID=76042593
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110324633.8A Pending CN112880750A (en) | 2021-03-26 | 2021-03-26 | Transformer multidimensional comprehensive online monitoring intelligent diagnosis system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112880750A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113422835A (en) * | 2021-07-01 | 2021-09-21 | 河南医学高等专科学校 | Multidimensional data monitoring, evaluating and early warning system in home-based old-age care environment |
| CN113758890A (en) * | 2021-08-18 | 2021-12-07 | 清华大学 | Gas concentration calculation method, device, equipment and storage medium |
| CN114184232A (en) * | 2021-12-10 | 2022-03-15 | 重庆大学 | A multi-parameter integrated monitoring system for transmission lines |
| CN114252110A (en) * | 2022-03-02 | 2022-03-29 | 山东和兑智能科技有限公司 | Intelligent evaluation system and evaluation method for power transformation equipment |
| CN114594712A (en) * | 2022-03-15 | 2022-06-07 | 胜斗士(上海)科技技术发展有限公司 | Sensor-based management apparatus and method thereof |
| CN114814642A (en) * | 2022-04-15 | 2022-07-29 | 国网山东省电力公司莱芜供电公司 | Method and system for diagnosing corrosion state of substation grounding grid based on sequential probability ratio |
| WO2023077862A1 (en) * | 2021-11-06 | 2023-05-11 | 国网山西省电力公司电力科学研究院 | Transformer oil conservator defect monitoring device and detection method based on edge computing |
| CN117109730A (en) * | 2023-10-23 | 2023-11-24 | 山东和兑智能科技有限公司 | Transformer fault diagnosis system and method based on artificial intelligence |
| US12359947B2 (en) | 2021-11-06 | 2025-07-15 | State Grid Shanxi Electric Power Research Institute | Defect monitoring apparatus and detection method for transformer oil conservator based on edge computing |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109840548A (en) * | 2019-01-02 | 2019-06-04 | 广东工业大学 | One kind being based on BP neural network Diagnosis Method of Transformer Faults |
| CN110401262A (en) * | 2019-06-17 | 2019-11-01 | 北京许继电气有限公司 | GIS equipment status intelligent monitoring system and method based on edge computing technology |
| CN210270089U (en) * | 2018-07-31 | 2020-04-07 | Abb瑞士股份有限公司 | Motor condition monitoring device |
| CN111784026A (en) * | 2020-05-28 | 2020-10-16 | 国网信通亿力科技有限责任公司 | Comprehensive physical examination system for substation electrical equipment based on cloud-edge-terminal collaborative perception |
| CN112395959A (en) * | 2020-10-30 | 2021-02-23 | 天合云能源互联网技术(杭州)有限公司 | Power transformer fault prediction and diagnosis method and system based on audio features |
-
2021
- 2021-03-26 CN CN202110324633.8A patent/CN112880750A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN210270089U (en) * | 2018-07-31 | 2020-04-07 | Abb瑞士股份有限公司 | Motor condition monitoring device |
| CN109840548A (en) * | 2019-01-02 | 2019-06-04 | 广东工业大学 | One kind being based on BP neural network Diagnosis Method of Transformer Faults |
| CN110401262A (en) * | 2019-06-17 | 2019-11-01 | 北京许继电气有限公司 | GIS equipment status intelligent monitoring system and method based on edge computing technology |
| CN111784026A (en) * | 2020-05-28 | 2020-10-16 | 国网信通亿力科技有限责任公司 | Comprehensive physical examination system for substation electrical equipment based on cloud-edge-terminal collaborative perception |
| CN112395959A (en) * | 2020-10-30 | 2021-02-23 | 天合云能源互联网技术(杭州)有限公司 | Power transformer fault prediction and diagnosis method and system based on audio features |
Non-Patent Citations (1)
| Title |
|---|
| 侯一鸣: "基于物联网和工业云的选矿设备状态监控系统", 《计算机集成制造系统》 * |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113422835A (en) * | 2021-07-01 | 2021-09-21 | 河南医学高等专科学校 | Multidimensional data monitoring, evaluating and early warning system in home-based old-age care environment |
| CN113422835B (en) * | 2021-07-01 | 2023-10-13 | 河南医学高等专科学校 | Multidimensional data monitoring, evaluating and early warning system in home care environment |
| CN113758890A (en) * | 2021-08-18 | 2021-12-07 | 清华大学 | Gas concentration calculation method, device, equipment and storage medium |
| WO2023077862A1 (en) * | 2021-11-06 | 2023-05-11 | 国网山西省电力公司电力科学研究院 | Transformer oil conservator defect monitoring device and detection method based on edge computing |
| US12359947B2 (en) | 2021-11-06 | 2025-07-15 | State Grid Shanxi Electric Power Research Institute | Defect monitoring apparatus and detection method for transformer oil conservator based on edge computing |
| CN114184232A (en) * | 2021-12-10 | 2022-03-15 | 重庆大学 | A multi-parameter integrated monitoring system for transmission lines |
| CN114252110A (en) * | 2022-03-02 | 2022-03-29 | 山东和兑智能科技有限公司 | Intelligent evaluation system and evaluation method for power transformation equipment |
| CN114594712A (en) * | 2022-03-15 | 2022-06-07 | 胜斗士(上海)科技技术发展有限公司 | Sensor-based management apparatus and method thereof |
| CN114814642A (en) * | 2022-04-15 | 2022-07-29 | 国网山东省电力公司莱芜供电公司 | Method and system for diagnosing corrosion state of substation grounding grid based on sequential probability ratio |
| CN117109730A (en) * | 2023-10-23 | 2023-11-24 | 山东和兑智能科技有限公司 | Transformer fault diagnosis system and method based on artificial intelligence |
| CN117109730B (en) * | 2023-10-23 | 2024-01-12 | 山东和兑智能科技有限公司 | Transformer fault diagnosis system and method based on artificial intelligence |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112880750A (en) | Transformer multidimensional comprehensive online monitoring intelligent diagnosis system | |
| KR102604708B1 (en) | Switchboard diagnosis system based on artificial intelligence and switchboard diagnosis method based on artificial intelligence | |
| CN111239672B (en) | Gradual fault prediction method of optical fiber current transformer based on machine learning algorithm | |
| Dukarm | Transformer oil diagnosis using fuzzy logic and neural networks | |
| CN111650451B (en) | Converter transformer fault reason identification method and system | |
| CN102779230B (en) | State analysis and maintenance decision judging method of power transformer system | |
| CN118757347B (en) | Intelligent identification method and system for wind generating set | |
| CN118536048A (en) | Switchgear insulation status monitoring and management system | |
| CN118885944B (en) | Multi-mode fault diagnosis and early warning method | |
| CN116449256A (en) | A transformer state fault diagnosis system and method based on voiceprint sensing | |
| CN119622550A (en) | Power product operation fault detection system and method based on big data | |
| CN117992776B (en) | Real-time prediction method for health state of power grid equipment based on artificial intelligence | |
| CN117371207A (en) | Extra-high voltage converter valve state evaluation method, medium and system | |
| CN117933447A (en) | A fault prediction method and system based on data analysis | |
| CN116884432A (en) | VMD-JS divergence-based power transformer fault voiceprint diagnosis method | |
| CN117153193A (en) | Power equipment fault voiceprint recognition method integrating physical characteristics and data diagnosis | |
| CN118152741A (en) | A temperature situation awareness auxiliary control method for substation equipment | |
| CN110186557A (en) | A kind of Reactor Fault diagnostic method | |
| CN119313506B (en) | Intelligent safety power distribution platform management system based on Internet of things technology | |
| CN119716407A (en) | Online monitoring platform for cable insulation status in distribution network based on equivalent electrical characteristics | |
| CN118884333A (en) | A method and system for evaluating the health status of a capacitor voltage transformer | |
| CN117630594A (en) | Transformer operating status evaluation system and method | |
| CN117665105A (en) | A foreign matter detection system in GIS busbar barrel | |
| CN113722659B (en) | Fault mode-based substation main equipment diagnosis method and system | |
| CN116070140A (en) | A system and method for monitoring the safe operation status of a power distribution substation |
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 |