CN117783795A - Comprehensive analysis method and system for insulation state of converter transformer valve side sleeve by edge analysis - Google Patents
Comprehensive analysis method and system for insulation state of converter transformer valve side sleeve by edge analysis Download PDFInfo
- Publication number
- CN117783795A CN117783795A CN202410212001.6A CN202410212001A CN117783795A CN 117783795 A CN117783795 A CN 117783795A CN 202410212001 A CN202410212001 A CN 202410212001A CN 117783795 A CN117783795 A CN 117783795A
- Authority
- CN
- China
- Prior art keywords
- data
- edge
- valve side
- converter transformer
- transformer valve
- 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.)
- Granted
Links
- 238000004458 analytical method Methods 0.000 title claims abstract description 99
- 238000009413 insulation Methods 0.000 title claims abstract description 47
- 238000004364 calculation method Methods 0.000 claims abstract description 41
- 238000000034 method Methods 0.000 claims abstract description 22
- 238000007405 data analysis Methods 0.000 claims abstract description 17
- 230000004044 response Effects 0.000 claims abstract description 15
- 230000002159 abnormal effect Effects 0.000 claims description 58
- 238000004422 calculation algorithm Methods 0.000 claims description 37
- 238000003066 decision tree Methods 0.000 claims description 18
- 238000009826 distribution Methods 0.000 claims description 15
- 230000005856 abnormality Effects 0.000 claims description 14
- 238000013528 artificial neural network Methods 0.000 claims description 13
- 238000012545 processing Methods 0.000 claims description 9
- 238000004590 computer program Methods 0.000 claims description 8
- 230000008569 process Effects 0.000 claims description 8
- 230000000694 effects Effects 0.000 claims description 6
- 230000004927 fusion Effects 0.000 claims description 6
- 238000012549 training Methods 0.000 claims description 6
- 230000008447 perception Effects 0.000 claims description 4
- 230000009471 action Effects 0.000 claims description 3
- 230000001276 controlling effect Effects 0.000 claims description 3
- 238000013135 deep learning Methods 0.000 claims description 3
- 230000002708 enhancing effect Effects 0.000 claims description 3
- 239000000284 extract Substances 0.000 claims description 3
- 238000001914 filtration Methods 0.000 claims description 3
- 238000009499 grossing Methods 0.000 claims description 3
- 230000002452 interceptive effect Effects 0.000 claims description 3
- 230000007246 mechanism Effects 0.000 claims description 3
- 230000001105 regulatory effect Effects 0.000 claims description 3
- 238000012423 maintenance Methods 0.000 abstract description 6
- 238000012544 monitoring process Methods 0.000 abstract description 6
- 230000008901 benefit Effects 0.000 description 5
- 230000008859 change Effects 0.000 description 4
- 230000035882 stress Effects 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 238000001514 detection method Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 230000006872 improvement Effects 0.000 description 3
- 230000007774 longterm Effects 0.000 description 3
- 238000000540 analysis of variance Methods 0.000 description 2
- 238000003491 array Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 238000003745 diagnosis Methods 0.000 description 2
- 230000004069 differentiation Effects 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000012098 association analyses Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000011217 control strategy Methods 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 239000006185 dispersion Substances 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000008713 feedback mechanism Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 239000011810 insulating material Substances 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 238000011002 quantification Methods 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
Landscapes
- Testing Relating To Insulation (AREA)
Abstract
The invention relates to the technical field of insulation state analysis of a converter transformer valve side sleeve, and discloses a comprehensive analysis method and a comprehensive analysis system of insulation state of a converter transformer valve side sleeve by edge analysis, wherein the comprehensive analysis method comprises the following steps: an edge calculating unit is configured at a sensor for collecting data on the converter transformer valve side sleeve; the edge computing unit is used for carrying out data analysis on the data acquired by the respective sensors; collecting the data analysis result of each edge calculation unit to a central calculation unit, wherein the central calculation unit performs comprehensive analysis; and triggering a response instruction in the central computing unit according to the comprehensive analysis result. The method for comprehensively analyzing the insulation state of the converter transformer valve side sleeve is efficient, accurate, safe and high in cost effectiveness, is suitable for real-time monitoring and maintenance of a modern power system, and has remarkable practical value and wide application prospect.
Description
Technical Field
The invention relates to the technical field of insulation state analysis of a converter transformer valve side sleeve, in particular to a comprehensive insulation state analysis method and system of a converter transformer valve side sleeve for edge analysis.
Background
In modern power systems, the reliability of the converter valves is of paramount importance. They play a central role in the hvdc transmission system, responsible for the conversion between ac and dc. The insulation state of the converter transformer valve side sleeve is a key factor for guaranteeing the stable operation of the power system. Poor insulation conditions may lead to system failure or reduced efficiency and even serious safety accidents.
Traditionally, condition monitoring of converter valve side bushings relies on periodic inspection and manual diagnostics, which is time consuming and labor intensive, but also difficult to implement in real time. Existing automated monitoring systems typically rely on centralized data processing, which can lead to data processing bottlenecks, reducing the response speed and accuracy of the system.
With the development of edge computing technology, more and more data processing is being transferred to the source of data generation. The method reduces the burden of the central server and improves the speed and efficiency of data processing. The application of edge computing in electrical power systems, in particular in state monitoring and fault diagnosis, presents great potential.
Disclosure of Invention
The present invention has been made in view of the above-described problems.
Therefore, the technical problems solved by the invention are as follows: the existing analysis method for the insulation state of the sleeve on the converter transformer valve side has the problems of high cost, low calculation speed, low response speed and the like.
In order to solve the technical problems, the invention provides the following technical scheme: the comprehensive analysis method for the insulation state of the converter transformer valve side sleeve in the edge analysis comprises the following steps:
an edge calculating unit is configured at a sensor for collecting data on the converter transformer valve side sleeve;
the edge computing unit is used for carrying out data analysis on the data acquired by the respective sensors;
collecting the data analysis result of each edge calculation unit to a central calculation unit, wherein the central calculation unit performs comprehensive analysis;
and triggering a response instruction in the central computing unit according to the comprehensive analysis result.
As a preferable scheme of the comprehensive analysis method for the insulation state of the converter transformer valve side sleeve for edge analysis, the invention comprises the following steps: the sensor comprises an electric parameter sensor, a temperature sensor, a mechanical stress sensor, an acoustic emission sensor and a vibration sensor.
As a preferable scheme of the comprehensive analysis method for the insulation state of the converter transformer valve side sleeve for edge analysis, the invention comprises the following steps: the data output end of the sensor is connected with an edge computing unit with data analysis capability;
and embedding an identification algorithm and a general algorithm which can analyze access data into the edge computing unit, computing the data by the general algorithm, and analyzing the processed data by the identification algorithm.
As a preferable scheme of the comprehensive analysis method for the insulation state of the converter transformer valve side sleeve for edge analysis, the invention comprises the following steps: the general algorithm includes adaptively filtering the raw data while enhancing key features:
Y(t)=β·X(t)+(1-β)·Y(t-1)+δ·dX(t)/dt;
wherein X (t) represents the raw data at time t; y (t) represents the processed data at time t; beta represents a smoothing coefficient for controlling the influence of the history data; delta represents an enhancement factor for adjusting the contribution of the derivative of the data to the result; dX (t)/dt represents the derivative of the raw data at time t;
capturing short-term dependencies of the data using a local autoregressive model;
;
wherein Z (t) represents the local autoregressive model output, a, at time t i The coefficients that represent the autoregressive model,representing the error term at time t, n representing the order of the autoregressive model;
each data updating period is used for carrying out resource sharing on the database of the identification algorithm in the edge computing unit of the same sensor for processing the same physical quantity, and the identification algorithm in the edge computing unit is updated by learning the database shared by other edge computing units; correcting the recognition algorithm through a local database of the self edge computing unit;
the shared database comprises a data set for recording the output results of the general algorithm and the abnormal insulation state of the sleeve at the converter transformer valve side in all edge computing units; the local database comprises a data set for recording the output result of the general algorithm in a local edge computing unit and the abnormal insulation state of the converter transformer valve side sleeve;
the identification algorithm comprises the steps of optimizing a decision tree branching process of a tree by using the combination of information gain and a data distribution sensing mechanism according to a shared database;
the combined information gain is expressed as:
IG’(D,A)=IG(D,A)+α·P(D,A);
where IG' (D, a) represents the enhanced information gain; IG (D, a) represents a conventional information gain; p (D, A) represents a data distribution perception factor, and information gain is adjusted by considering the distribution characteristic of data; alpha represents a regulatory factor; d represents a dataset; a represents attribute characteristics and represents abnormal occurrence;
according to the local database, using a neural network to adjust and optimize branch decisions of the decision tree; DT (DT) optimized =NDTO(EDT,X local ,θ)
Wherein DT is optimized Represents the optimized decision tree, NDTO represents the neural network optimizer, EDT represents the initial enhanced decision tree, X local Representing the input of the local database, θ represents a parameter of the neural network.
As a preferable scheme of the comprehensive analysis method for the insulation state of the converter transformer valve side sleeve for edge analysis, the invention comprises the following steps: the central computing unit takes the analysis result of the edge computing unit as abnormal evidence of edge analysis of the insulating state of the sleeve on the converter transformer valve side, and the central unit receives the abnormal evidence so as to analyze the insulating state of the sleeve on the converter transformer valve side;
each edge of the same converter transformer valve side sleeveThe abnormal condition output by the edge calculation unit is assigned and is expressed as L i The method comprises the steps of carrying out a first treatment on the surface of the Normal represents 0, with 1 added sequentially with the degree of abnormality, where the most serious abnormality is represented as c;
calculating the average value of the abnormal grades output by all the edge calculation unitsAnd standard deviation σ (L):
;
wherein n represents the total number of edge calculation units;
by setting a standard deviation threshold value sigma, if sigma (L) is less than or equal to sigma, the output of each edge calculating unit is judged to be consistent enough, and the judgment result is outputThe corresponding abnormal situation; if sigma (L)>Sigma, judging that the difference between evidences is overlarge, and not outputting a comprehensive abnormal judgment result;
wherein,representing no less than +.>Is a minimum integer of (a).
As a preferable scheme of the comprehensive analysis method for the insulation state of the converter transformer valve side sleeve for edge analysis, the invention comprises the following steps: when the analysis result of the insulation state of the converter transformer valve side sleeve is that the comprehensive abnormality judgment result is not output, the central computing unit extracts and restores the original data used by the edge computing unit on the converter transformer valve side sleeve, and re-analyzes the insulation state of the converter transformer valve side sleeve according to the original data;
training to obtain an identifier capable of identifying an abnormal state according to historical data of a sensor of the converter valve side sleeve which does not output a comprehensive abnormal judgment result; inputting real-time abnormal data into the identifier, and outputting an identified abnormal analysis result by the identifier;
the training process of the recognizer comprises the following steps of interactive feature fusion with dynamic weight adjustment:
F int =σ((W f (t)⊙R(t))⋅[X stft ;H dl ]+b f (t));
wherein W is f (t) represents the fusion layer weight at time step t; b f (t) represents the bias at time step t; x is X stft Representing time-frequency features extracted from the STFT; h dl Representing an output of the deep learning feature extractor; r (t) represents a dynamically adjusted attention weight vector, and is adaptively adjusted according to the learning effect of the previous moment; the ";
constructing a graph attention network, wherein nodes represent different time-frequency characteristics, and the relationship between the different characteristics is captured by using GAT;
H GAT =GAT(F int ,E);
wherein E represents edges between features, which are relationships between features;
classification of abnormal state for GAT output using fully connected layer:
Y pred =softmax(W y ·H GAT +b y );
wherein W is y Representing the weights of the classification layer, b y Representing the bias of the classification layer.
As a preferable scheme of the comprehensive analysis method for the insulation state of the converter transformer valve side sleeve for edge analysis, the invention comprises the following steps: the response instructions comprise three control instructions, and abnormal conditions corresponding to the three control instructions are respectively arranged in each converter transformer valve side sleeve;
when the abnormal condition of the comprehensive analysis result of the central computing unit triggers a control instruction, the execution module controls according to the instruction content;
the three control instructions comprise that the first control instruction is to stop the machine after completing the task list and send out an alarm; the second control instruction is that stopping the machine after completing the task of the round and giving an alarm; and the third control instruction is to immediately stop the operation and send out an early warning.
The invention relates to a comprehensive analysis system for insulation state of a converter transformer valve side sleeve by adopting edge analysis of the method, which is characterized in that:
the edge analysis module is used for configuring an edge calculation unit at a sensor for acquiring data on the converter transformer valve side sleeve; the edge computing unit is used for carrying out data analysis on the data acquired by the respective sensors;
the central analysis module is used for collecting the data analysis result of each edge calculation unit to a central calculation unit, and the central calculation unit performs comprehensive analysis; triggering a response instruction in a central computing unit according to the comprehensive analysis result;
and the execution module executes the control action according to the instruction of the central computing unit.
A computer device, comprising: a memory and a processor; the memory stores a computer program characterized in that: the processor, when executing the computer program, implements the steps of the method of any of the present invention.
A computer-readable storage medium having stored thereon a computer program, characterized by: the computer program, when executed by a processor, implements the steps of the method of any of the present invention.
The invention has the beneficial effects that: according to the comprehensive analysis method for the insulation state of the converter transformer valve side sleeve provided by the invention, the delay of data transmission and processing is greatly reduced by the application of edge calculation, so that a system can process a large amount of sensor data in real time and respond quickly. When an abnormal condition is detected, the central computing unit can immediately trigger a corresponding response instruction, such as a shutdown alarm or maintenance notification, thereby reducing potential risks and losses. By identifying and addressing potential problems in advance, unexpected downtime and emergency maintenance work is facilitated to be reduced. The method for comprehensively analyzing the insulation state of the converter transformer valve side sleeve is efficient, accurate, safe and high in cost effectiveness, is suitable for real-time monitoring and maintenance of a modern power system, and has remarkable practical value and wide application prospect.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is an overall flowchart of a comprehensive analysis method for insulation state of a sleeve on a converter transformer valve side by edge analysis according to a first embodiment of the present invention.
Detailed Description
So that the manner in which the above recited objects, features and advantages of the present invention can be understood in detail, a more particular description of the invention, briefly summarized above, may be had by reference to the embodiments, some of which are illustrated in the appended drawings. All other embodiments, which can be made by one of ordinary skill in the art based on the embodiments of the present invention without making any inventive effort, shall fall within the scope of the present invention.
Example 1
Referring to fig. 1, for one embodiment of the present invention, there is provided a comprehensive analysis method of insulation state of a converter transformer valve side bushing for edge analysis, including:
s1: an edge calculating unit is configured at a sensor for collecting data on the converter transformer valve side sleeve; and carrying out data analysis on the data acquired by the respective sensors by utilizing the edge computing unit.
Further, the sensor comprises an electric parameter sensor, a temperature sensor, a mechanical stress sensor, an acoustic emission sensor and a vibration sensor. The electric parameter sensor can monitor electric parameters such as current, voltage and the like so as to evaluate the efficiency and stability of power transmission. The temperature sensor is able to monitor the temperature change of the bushing and excessive temperatures may indicate aging of the insulating material or overload of the system. Mechanical stress sensors are capable of detecting structural stresses, which is critical to ensure the physical integrity of the casing during long-term operation. The acoustic emission sensor is capable of capturing acoustic waves generated by defects such as cracks or breaks, which is an effective early failure detection means. Vibration sensors are capable of monitoring vibrations caused by unbalance, misalignment or other mechanical faults, which are often early signs of fault development.
Still further, the data output end of the sensor is connected to an edge computing unit with data analysis capability; and embedding an identification algorithm and a general algorithm which can analyze access data into the edge computing unit, computing the data by the general algorithm, and analyzing the processed data by the identification algorithm.
It is to be noted that the general algorithm includes adaptive filtering of the raw data while enhancing key features:
Y(t)=β·X(t)+(1-β)·Y(t-1)+δ·dX(t)/dt;
wherein X (t) represents the raw data at time t; y (t) represents the processed data at time t; beta represents a smoothing coefficient for controlling the influence of the history data; delta represents an enhancement factor for adjusting the contribution of the derivative of the data to the result; dX (t)/dt represents the derivative of the raw data at time t; by introducing the derivative, besides the abnormal influence of the data value, the influence of the excessively rapid increase of the numerical value is also considered, if the electric leakage abnormality occurs, the change rate of the waveform is obviously different from that of the conventional waveform, and the key characteristic can be effectively considered by capturing the characteristic through the derivative.
Capturing short-term dependencies of the data using a local autoregressive model;
;
wherein Z (t) represents the local autoregressive model output, a, at time t i The coefficients that represent the autoregressive model,representing the error term at time t, n represents the order of the autoregressive model.
Each data updating period is used for carrying out resource sharing on the database of the identification algorithm in the edge computing unit of the same sensor for processing the same physical quantity, and the identification algorithm in the edge computing unit is updated by learning the database shared by other edge computing units; and correcting the recognition algorithm through a local database of the self edge computing unit.
The shared database comprises a data set for recording the output results of the general algorithm and the abnormal insulation state of the sleeve at the converter transformer valve side in all edge computing units; the local database comprises a data set for recording the output result of the general algorithm in the local edge computing unit and the abnormal insulation state of the converter transformer valve side sleeve. By sharing the data of the different edge computing units, the recognition algorithm can obtain a more diversified data input, which helps capture and learn a wider range of data patterns and anomalies. This cross-cell learning allows the algorithm to adapt to various operating conditions and possible abnormal conditions, thereby improving its accuracy and generalization ability. The feedback mechanism of the local database ensures that the algorithm can be corrected and corrected according to actual conditions, which is helpful to maintain the long-term stability and reliability of the algorithm, so that the calculation process of each edge calculation unit has own characteristics, and the performance of each converter transformer valve side sleeve and the data acquired by the sensor are slightly different, and the detection of the equipment can be adapted to through the correction of the local database.
The identification algorithm comprises the steps of optimizing a decision tree branching process of a tree by using the combination of information gain and a data distribution sensing mechanism according to a shared database; the combined information gain is expressed as:
IG’(D,A)=IG(D,A)+α·P(D,A);
where IG' (D, a) represents the enhanced information gain; IG (D, a) represents a conventional information gain; p (D, A) represents a data distribution perception factor, and information gain is adjusted by considering the distribution characteristic of data; alpha represents a regulatory factor; d represents a dataset; a represents attribute characteristics and represents occurrence of abnormality.
Quantification of the data distribution perception factor P (D, a) includes distribution consistency, which is a study of whether the distributions of different classes (e.g., normal and abnormal) in the data set D are consistent under attribute a, and differentiation of attributes. This can be quantified by comparing the differences in the distribution of the categories under attribute a. The degree of differentiation of the attributes is the effectiveness of the measurement attribute a in distinguishing between different categories (e.g., normal and abnormal). Statistical methods (such as analysis of variance) may be used for evaluation. For example:
p (D, a) =γΣ agreement measure (D, a) +δΣ discrimination measure (D, a);
where γ and δ are coefficients that trade off these two aspects for adjusting the effect of consistency and discrimination in P (D, a). The consistency metric may use a distance-based metric (e.g., KL divergence) to compare the data distribution differences for different categories under attribute a. The discrimination metric may employ analysis of variance to calculate the variance ratio of attribute a when discriminating between different categories.
By considering the distribution, the enhancement of the data signal in the region can be realized, so that the data of the abnormal region can be more obvious, and the abnormal region can be found.
According to the local database, using a neural network to adjust and optimize branch decisions of the decision tree;
DT optimized =NDTO(EDT,X local ,θ);
wherein DT is optimized Represents the optimized decision tree, NDTO represents the neural network optimizer, EDT represents the initial enhanced decision tree, X local Representing the input of the local database, θ represents a parameter of the neural network.
Decision trees are optimized through neural networks because decision trees are formed using shared data, which only characterizes the data. Each device may have a different characteristic index, for example, in a multiple relationship or by a different value. Adjusting this decision tree by means of a neural network enables his decision scheme to be more consistent with the current device. Each device may have unique operating characteristics and environmental conditions and decision trees built based on shared data may not fully capture these personalized differences. Each edge computing unit is trained by utilizing local data, so that the computing capacity of each edge computing unit can be ensured to accord with the equipment characteristics of the edge computing unit. In addition, over time and with changing environments, the operating characteristics of the device may change. The neural network can continually learn from the latest local data, ensuring that the decision tree dynamically adjusts as the environment changes. Such a calculation mode also represents an advantage of edge analysis.
The neural network is utilized to optimize the decision tree, and the individualized adjustment is carried out according to the specific condition of each converter transformer valve side sleeve, so that the method is technically feasible, the accuracy and the adaptability of the decision can be obviously improved, and the method has important significance for improving the reliability and the efficiency of a power system.
S2: and collecting the data analysis result of each edge calculation unit to a central calculation unit, wherein the central calculation unit performs comprehensive analysis.
The central computing unit is used for taking the analysis result of the edge computing unit as abnormal evidence of edge analysis of the insulating state of the sleeve on the converter transformer valve side, and the central computing unit is used for receiving the abnormal evidence so as to analyze the insulating state of the sleeve on the converter transformer valve side. Assigning the abnormal condition output by each edge computing unit of the same converter transformer valve side sleeve, and representing the abnormal condition as L i The method comprises the steps of carrying out a first treatment on the surface of the Normal represents 0, with 1 added sequentially with the degree of abnormality, where the most serious abnormality is represented as c. For example: normal is 0, slight anomaly is 1, moderate anomaly is 2, severe anomaly is 3, and anomaly level can be set according to the requirements of technicians.
Calculating the average value of the abnormal grades output by all the edge calculation unitsAnd standard deviation σ (L):
;
where n represents the total number of edge calculation units.
By setting a standard deviation threshold value sigma, if sigma (L) is less than or equal to sigma, the output of each edge calculating unit is judged to be consistent enough, and the judgment result is outputThe corresponding abnormal situation; if sigma (L)>Sigma, judging that the difference between evidences is overlarge, and not outputting a comprehensive abnormal judgment result; wherein (1)>Representing no less than +.>Is a minimum integer of (a). By calculating the standard deviation of the anomaly level of all the edge calculation units, we can quantify the consistency of the judgment between the different edge calculation units. The standard deviation is a statistical index for measuring the fluctuation or dispersion degree of the data, and the lower standard deviation means that the judgment of each edge calculation unit is more consistent, so that the obtained comprehensive result can be more reliably considered to be accurate. Greater than the threshold value means that the judgment of the edge calculation units differs greatly, which may be due to data noise, sensor failure, or complexity of the actual environment. In this case, erroneous judgment based on inconsistent data can be avoided without outputting the comprehensive abnormality judgment result, and at this time, comprehensive analysis is required to be performed on the data of the plurality of data sources.
When the analysis result of the insulation state of the converter transformer valve side sleeve is that the comprehensive abnormality judgment result is not output, the central computing unit extracts and restores the original data used by the edge computing unit on the converter transformer valve side sleeve, and re-analyzes the insulation state of the converter transformer valve side sleeve according to the original data. Training to obtain an identifier capable of identifying an abnormal state according to historical data of a sensor of the converter valve side sleeve which does not output a comprehensive abnormal judgment result; and inputting real-time abnormal data into the identifier, and outputting an identified abnormal analysis result by the identifier. That is, when the evidence contradiction is too large, the evidence cannot indicate the insulation state, and at this time, the original data on the sleeve on the converter valve side is all called out, and the original data is used for the comprehensive analysis again. Thereby realizing the recognition of the abnormality.
The training process of the recognizer comprises the following steps of interactive feature fusion with dynamic weight adjustment:
F int =σ((W f (t)⊙R(t))⋅[X stft ;H dl ]+b f (t));
wherein W is f (t) represents the fusion layer weight at time step t; b f (t) represents the bias at time step t; x is X stft Representing time-frequency features extracted from the STFT; h dl Representing an output of the deep learning feature extractor; r (t) represents a dynamically adjusted attention weight vector, and is adaptively adjusted according to the learning effect of the previous moment; as indicated by the letter, ".
Where R (t) is dynamically changing, e.g. a larger contribution to a certain parametric model at a previous moment, it is explained that he may have a larger influence on a later moment, this effect being increased by R (t). Thereby realizing the data association analysis effect on the time-frequency domain.
Constructing a graph attention network, wherein nodes represent different time-frequency characteristics, and the relationship between the different characteristics is captured by using GAT:
H GAT =GAT(F int ,E);
where E represents the edges between features and is the relationship between features.
Classification of abnormal state for GAT output using fully connected layer:
Y pred =softmax(W y ·H GAT +b y );
wherein W is y Representing the weights of the classification layer, b y Representing the bias of the classification layer.
S3: and triggering a response instruction in the central computing unit according to the comprehensive analysis result.
Further, the response instruction comprises setting three control instructions, and setting abnormal conditions corresponding to the three control instructions in each converter transformer valve side sleeve respectively.
It is known that, in consideration of significant differences in fault tolerance and reliability of different devices (such as the degree of freshness, the historical maintenance record, the use intensity and the like), setting control instructions for the devices respectively can ensure that the response strategy is matched with the actual conditions of the devices. For example: new devices may be able to survive for some time in the event of some anomaly, while some devices are older and should be shut down immediately upon the discovery of the anomaly. For this case, the abnormal conditions corresponding to the three instructions are set for each device, respectively. For example, for a new device, the first instruction is executed in both normal and slightly abnormal situations. For new equipment or equipment with stronger fault tolerance, the operation is allowed to continue under the condition of slight abnormality until the task list is completed, so that the operation efficiency can be maximized, and the unnecessary downtime is reduced. The second instruction is executed for the old device in a slight exception. For equipment which is aged or is more prone to faults, stricter control measures are adopted when slight anomalies are detected, for example, the equipment is stopped after the round of tasks is completed, potential faults are prevented, and influences caused by the faults are reduced. As equipment conditions change (e.g., new equipment ages), the control strategy may also be adjusted accordingly to maintain optimal operating conditions and highest safety standards.
When the abnormal condition of the comprehensive analysis result of the central computing unit triggers a control instruction, the execution module controls according to the instruction content. The three control instructions comprise that the first control instruction is to stop the machine after completing the task list and send out an alarm; the second control instruction is that stopping the machine after completing the task of the round and giving an alarm; and the third control instruction is to immediately stop the operation and send out an early warning.
On the other hand, this embodiment also provides a comprehensive analysis system of insulation state of the converter transformer valve side sleeve of edge analysis, which comprises:
the edge analysis module is used for configuring an edge calculation unit at a sensor for acquiring data on the converter transformer valve side sleeve; and carrying out data analysis on the data acquired by the respective sensors by utilizing the edge computing unit.
The central analysis module is used for collecting the data analysis result of each edge calculation unit to a central calculation unit, and the central calculation unit performs comprehensive analysis; and triggering a response instruction in the central computing unit according to the comprehensive analysis result.
And the execution module executes the control action according to the instruction of the central computing unit.
The above functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a usb disk, a removable hard disk, a Read-only memory (ROM), a random access memory (RAM, randomAccessMemory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
Logic and/or steps represented in the flowcharts or otherwise described herein, e.g., a ordered listing of executable instructions for implementing logical functions, can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor-containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions. For the purposes of this description, a "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device.
More specific examples (a non-exhaustive list) of the computer-readable medium would include the following: an electrical connection (electronic device) having one or more wires, a portable computer diskette (magnetic device), a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). In addition, the computer readable medium may even be paper or other suitable medium on which the program is printed, as the program may be electronically captured, via, for instance, optical scanning of the paper or other medium, then compiled, interpreted or otherwise processed in a suitable manner, if necessary, and then stored in a computer memory.
It is to be understood that portions of the present invention may be implemented in hardware, software, firmware, or a combination thereof. In the above-described embodiments, the various steps or methods may be implemented in software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, may be implemented using any one or combination of the following techniques, as is well known in the art: discrete logic circuits having logic gates for implementing logic functions on data signals, application specific integrated circuits having suitable combinational logic gates, programmable Gate Arrays (PGAs), field Programmable Gate Arrays (FPGAs), and the like.
Example 2
In the following, for one embodiment of the invention, a comprehensive analysis method of insulation state of a converter transformer valve side sleeve by edge analysis is provided, and in order to verify the beneficial effects of the invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.
Performance index comparison results were obtained by using the test experiments before and after the present invention, as shown in table 1.
TABLE 1 Performance index Table
As can be seen from table 1, the present invention exhibits significant improvements in a number of key performance indicators, including increased accuracy in fault detection, reduced system response time, reduced maintenance costs, increased equipment life, increased safety, and increased energy efficiency. These improvements demonstrate significant advantages of the present invention in improving efficiency, reliability and safety of power system operation, while reducing operating costs and improving overall use.
Table 2 is obtained by comparison of the present invention and conventional methods through simulation test experiments.
Table 2 data comparison table
Table 2 provides a more comprehensive view to demonstrate the improvement of the present invention over the conventional approach. The present invention shows significant advantages in particular in terms of data processing efficiency, fault diagnosis time and long-term operating costs, which further demonstrates its effectiveness in improving the operating efficiency of the power system and reducing costs.
It should be noted that the above embodiments are only for illustrating the technical solution of the present invention and not for limiting the same, and although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solution of the present invention may be modified or substituted without departing from the spirit and scope of the technical solution of the present invention, which is intended to be covered in the scope of the claims of the present invention.
Claims (10)
1. The comprehensive analysis method for the insulation state of the converter transformer valve side sleeve by edge analysis is characterized by comprising the following steps of:
an edge calculating unit is configured at a sensor for collecting data on the converter transformer valve side sleeve;
the edge computing unit is used for carrying out data analysis on the data acquired by the respective sensors;
collecting the data analysis result of each edge calculation unit to a central calculation unit, wherein the central calculation unit performs comprehensive analysis;
and triggering a response instruction in the central computing unit according to the comprehensive analysis result.
2. The comprehensive analysis method for insulation state of converter transformer valve side sleeve by edge analysis according to claim 1, wherein the comprehensive analysis method comprises the following steps: the sensor comprises an electric parameter sensor, a temperature sensor, a mechanical stress sensor, an acoustic emission sensor and a vibration sensor.
3. The comprehensive analysis method for insulation state of converter transformer valve side sleeve by edge analysis according to claim 2, wherein: the data output end of the sensor is connected with an edge computing unit with data analysis capability;
and embedding an identification algorithm and a general algorithm which can analyze access data into the edge computing unit, computing the data by the general algorithm, and analyzing the processed data by the identification algorithm.
4. The comprehensive analysis method for insulation state of converter transformer valve side sleeve according to claim 3, wherein: the general algorithm includes adaptively filtering the raw data while enhancing key features:
Y(t)=β·X(t)+(1-β)·Y(t-1)+δ·dX(t)/dt;
wherein X (t) represents the raw data at time t; y (t) represents the processed data at time t; beta represents a smoothing coefficient for controlling the influence of the history data; delta represents an enhancement factor for adjusting the contribution of the derivative of the data to the result; dX (t)/dt represents the derivative of the raw data at time t;
capturing short-term dependencies of the data using a local autoregressive model;
;
wherein Z (t) represents the local autoregressive model output, a, at time t i The coefficients that represent the autoregressive model,representing the error term at time t, n representing the order of the autoregressive model;
each data updating period is used for carrying out resource sharing on the database of the identification algorithm in the edge computing unit of the same sensor for processing the same physical quantity, and the identification algorithm in the edge computing unit is updated by learning the database shared by other edge computing units; correcting the recognition algorithm through a local database of the self edge computing unit;
the shared database comprises a data set for recording the output results of the general algorithm and the abnormal insulation state of the sleeve at the converter transformer valve side in all edge computing units; the local database comprises a data set for recording the output result of the general algorithm in a local edge computing unit and the abnormal insulation state of the converter transformer valve side sleeve;
the identification algorithm comprises the steps of optimizing a decision tree branching process of a tree by using the combination of information gain and a data distribution sensing mechanism according to a shared database;
the combined information gain is expressed as:
IG’(D,A)=IG(D,A)+α·P(D,A);
where IG' (D, a) represents the enhanced information gain; IG (D, a) represents a conventional information gain; p (D, A) represents a data distribution perception factor, and information gain is adjusted by considering the distribution characteristic of data; alpha represents a regulatory factor; d represents a dataset; a represents attribute characteristics and represents abnormal occurrence;
according to the local database, using a neural network to adjust and optimize branch decisions of the decision tree; DT (DT) optimized =NDTO(EDT,X local ,θ)
Wherein DT is optimized Represents the optimized decision tree, NDTO represents the neural network optimizer, EDT represents the initial enhanced decision tree, X local Representing the input of the local database, θ represents a parameter of the neural network.
5. The comprehensive analysis method for the insulation state of the converter transformer valve side sleeve by edge analysis according to claim 4, wherein the comprehensive analysis method comprises the following steps: the central computing unit takes the analysis result of the edge computing unit as abnormal evidence of edge analysis of the insulating state of the sleeve on the converter transformer valve side, and the central unit receives the abnormal evidence so as to analyze the insulating state of the sleeve on the converter transformer valve side;
assigning the abnormal condition output by each edge computing unit of the same converter transformer valve side sleeve, and representing the abnormal condition as L i The method comprises the steps of carrying out a first treatment on the surface of the Normal represents 0, with 1 added sequentially with the degree of abnormality, where the most serious abnormality is represented as c;
calculating the average value of the abnormal grades output by all the edge calculation unitsAnd standard deviation σ (L):
;
wherein n represents the total number of edge calculation units;
by setting a standard deviation threshold value sigma, if sigma (L) is less than or equal to sigma, the output of each edge calculating unit is judged to be consistent enough, and the judgment result is outputThe corresponding abnormal situation; if sigma (L)>Sigma, judging that the difference between evidences is overlarge, and not outputting a comprehensive abnormal judgment result;
wherein,representing no less than +.>Is a minimum integer of (a).
6. The comprehensive analysis method for the insulation state of the converter transformer valve side sleeve by edge analysis according to claim 5, wherein the comprehensive analysis method comprises the following steps: when the analysis result of the insulation state of the converter transformer valve side sleeve is that the comprehensive abnormality judgment result is not output, the central computing unit extracts and restores the original data used by the edge computing unit on the converter transformer valve side sleeve, and re-analyzes the insulation state of the converter transformer valve side sleeve according to the original data;
training to obtain an identifier capable of identifying an abnormal state according to historical data of a sensor of the converter valve side sleeve which does not output a comprehensive abnormal judgment result; inputting real-time abnormal data into the identifier, and outputting an identified abnormal analysis result by the identifier;
the training process of the recognizer comprises the following steps of interactive feature fusion with dynamic weight adjustment:
F int =σ((W f (t)⊙R(t))⋅[X stft ;H dl ]+b f (t));
wherein W is f (t) represents the fusion layer weight at time step t; b f (t) represents the bias at time step t; x is X stft Representing time-frequency features extracted from the STFT; h dl Representing an output of the deep learning feature extractor; r (t) represents a dynamically adjusted attention weight vector, and is adaptively adjusted according to the learning effect of the previous moment; the ";
constructing a graph attention network, wherein nodes represent different time-frequency characteristics, and the relationship between the different characteristics is captured by using GAT;
H GAT =GAT(F int ,E);
wherein E represents edges between features, which are relationships between features;
classification of abnormal state for GAT output using fully connected layer:
Y pred =softmax(W y ·H GAT +b y );
wherein W is y Representing the weights of the classification layer, b y Representing the bias of the classification layer.
7. The comprehensive analysis method for the insulation state of the converter transformer valve side sleeve by edge analysis according to claim 6, wherein the comprehensive analysis method comprises the following steps: the response instructions comprise three control instructions, and abnormal conditions corresponding to the three control instructions are respectively arranged in each converter transformer valve side sleeve;
when the abnormal condition of the comprehensive analysis result of the central computing unit triggers a control instruction, the execution module controls according to the instruction content;
the three control instructions comprise that the first control instruction is to stop the machine after completing the task list and send out an alarm; the second control instruction is that stopping the machine after completing the task of the round and giving an alarm; and the third control instruction is to immediately stop the operation and send out an early warning.
8. A converter transformer valve side bushing insulation state comprehensive analysis system employing the edge analysis of any one of claims 1-7, characterized in that:
the edge analysis module is used for configuring an edge calculation unit at a sensor for acquiring data on the converter transformer valve side sleeve; the edge computing unit is used for carrying out data analysis on the data acquired by the respective sensors;
the central analysis module is used for collecting the data analysis result of each edge calculation unit to a central calculation unit, and the central calculation unit performs comprehensive analysis; triggering a response instruction in a central computing unit according to the comprehensive analysis result;
and the execution module executes the control action according to the instruction of the central computing unit.
9. A computer device, comprising: a memory and a processor; the memory stores a computer program characterized in that: and the processor executes the computer program to realize the comprehensive analysis method of the insulation state of the converter transformer valve side sleeve of the edge analysis.
10. A computer-readable storage medium having stored thereon a computer program, characterized by: and the computer program is executed by a processor to realize the comprehensive analysis method of the insulation state of the converter transformer valve side sleeve of edge analysis.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202410212001.6A CN117783795B (en) | 2024-02-27 | 2024-02-27 | Comprehensive analysis method and system for insulation state of converter transformer valve side sleeve by edge analysis |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202410212001.6A CN117783795B (en) | 2024-02-27 | 2024-02-27 | Comprehensive analysis method and system for insulation state of converter transformer valve side sleeve by edge analysis |
Publications (2)
Publication Number | Publication Date |
---|---|
CN117783795A true CN117783795A (en) | 2024-03-29 |
CN117783795B CN117783795B (en) | 2024-06-28 |
Family
ID=90393103
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202410212001.6A Active CN117783795B (en) | 2024-02-27 | 2024-02-27 | Comprehensive analysis method and system for insulation state of converter transformer valve side sleeve by edge analysis |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN117783795B (en) |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107015145A (en) * | 2017-06-01 | 2017-08-04 | 湖州知维技术服务有限公司 | Primary cut-out is monitored on-line and state evaluation system |
CN110927537A (en) * | 2019-11-27 | 2020-03-27 | 国网江苏省电力有限公司电力科学研究院 | Partial discharge monitoring device and method based on Internet of things edge calculation |
CN113589096A (en) * | 2021-06-30 | 2021-11-02 | 国网电力科学研究院武汉南瑞有限责任公司 | Edge calculation system and method for multi-state-quantity configurable power transformation equipment |
WO2022227105A1 (en) * | 2021-04-30 | 2022-11-03 | 广州中国科学院软件应用技术研究所 | Cooperative control system and method based on device-cloud fusion |
CN115566696A (en) * | 2022-10-27 | 2023-01-03 | 云南电网有限责任公司 | Cloud-edge cooperative power grid frequency online monitoring and analyzing method and system |
US20230012038A1 (en) * | 2019-11-27 | 2023-01-12 | Sentient Technology Holdings, LLC | Systems and methods for automated detection of switch capacitor operation |
CN117235443A (en) * | 2023-09-19 | 2023-12-15 | 南方电网数字平台科技(广东)有限公司 | Electric power operation safety monitoring method and system based on edge AI |
CN117271981A (en) * | 2023-11-21 | 2023-12-22 | 湖南嘉创信息科技发展有限公司 | Artificial intelligence management system based on cross-platform data interaction |
CN117452889A (en) * | 2023-10-23 | 2024-01-26 | 青科驭远(青岛)智能科技有限公司 | Intelligent tire equipment platform system and control method |
-
2024
- 2024-02-27 CN CN202410212001.6A patent/CN117783795B/en active Active
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107015145A (en) * | 2017-06-01 | 2017-08-04 | 湖州知维技术服务有限公司 | Primary cut-out is monitored on-line and state evaluation system |
CN110927537A (en) * | 2019-11-27 | 2020-03-27 | 国网江苏省电力有限公司电力科学研究院 | Partial discharge monitoring device and method based on Internet of things edge calculation |
US20230012038A1 (en) * | 2019-11-27 | 2023-01-12 | Sentient Technology Holdings, LLC | Systems and methods for automated detection of switch capacitor operation |
WO2022227105A1 (en) * | 2021-04-30 | 2022-11-03 | 广州中国科学院软件应用技术研究所 | Cooperative control system and method based on device-cloud fusion |
CN113589096A (en) * | 2021-06-30 | 2021-11-02 | 国网电力科学研究院武汉南瑞有限责任公司 | Edge calculation system and method for multi-state-quantity configurable power transformation equipment |
CN115566696A (en) * | 2022-10-27 | 2023-01-03 | 云南电网有限责任公司 | Cloud-edge cooperative power grid frequency online monitoring and analyzing method and system |
CN117235443A (en) * | 2023-09-19 | 2023-12-15 | 南方电网数字平台科技(广东)有限公司 | Electric power operation safety monitoring method and system based on edge AI |
CN117452889A (en) * | 2023-10-23 | 2024-01-26 | 青科驭远(青岛)智能科技有限公司 | Intelligent tire equipment platform system and control method |
CN117271981A (en) * | 2023-11-21 | 2023-12-22 | 湖南嘉创信息科技发展有限公司 | Artificial intelligence management system based on cross-platform data interaction |
Non-Patent Citations (3)
Title |
---|
孟亦凡等: "基于LSTM边缘计算与随机森林雾决策的矿工状态监测设备", 煤矿机械, no. 11, 15 November 2018 (2018-11-15) * |
毕昊林等: "基于边缘计算的柱上开关状态监测研究", 中国新技术新产品, no. 6, 25 March 2020 (2020-03-25) * |
鲁东生等: "对基于深度学习方式实现无线电干扰智能化监测的思考", 中国无线电, 26 August 2022 (2022-08-26) * |
Also Published As
Publication number | Publication date |
---|---|
CN117783795B (en) | 2024-06-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109186813B (en) | Temperature sensor self-checking device and method | |
CN112284440B (en) | Sensor data deviation self-adaptive correction method | |
US8130112B2 (en) | Method of alarm mask generation and condition monitoring of wind turbines | |
CN112629905A (en) | Equipment anomaly detection method and system based on deep learning and computer medium | |
CN114862275B (en) | Storage logistics system reliability assessment method and system based on digital twin model | |
CN111582542A (en) | Power load prediction method and system based on abnormal restoration | |
CN117973902A (en) | Intelligent decision method and system based on level-keeping table fault handling | |
CN117910999A (en) | Intelligent power plant equipment maintenance method and system | |
CN118013439A (en) | Early warning method and system for fan fault detection | |
CN117706943B (en) | Self-adaptive control method and system for converter transformer valve side sleeve end screen voltage divider | |
CN118365307A (en) | Power plant equipment state overhaul decision support method and system | |
CN117783795B (en) | Comprehensive analysis method and system for insulation state of converter transformer valve side sleeve by edge analysis | |
CN117786578A (en) | Method, device, equipment and storage medium for detecting electric power running state | |
CN117574264A (en) | Transformer fault diagnosis method and system based on knowledge constraint neural network | |
CN117114248A (en) | Health evaluation method for water quality monitoring system, computer equipment and medium | |
JP7062505B2 (en) | Equipment management support system | |
CN116363843A (en) | Laboratory equipment early warning system | |
CN115456041A (en) | Equipment fault early warning method and device, computing equipment and storage medium | |
CN113552444A (en) | Online setting method and device for leakage current characteristic threshold | |
CN113627088B (en) | Machine performance degradation evaluation method and system based on gene programming and data fusion | |
Strackeljan et al. | Smart adaptive monitoring and diagnostic systems | |
CN118428741B (en) | Intelligent anti-misoperation processing method suitable for special operation and emergency operation | |
CN117639282B (en) | Converter transformer valve side sleeve end screen voltage divider frequency domain response processing method and system | |
CN118549780B (en) | Transformer partial discharge sensing system | |
CN118033344A (en) | On-line monitoring method and system for insulation state of converter transformer valve side sleeve |
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 |