CN117150934B - Comprehensive monitoring system for transformer bushing state and online data processing method thereof - Google Patents

Comprehensive monitoring system for transformer bushing state and online data processing method thereof Download PDF

Info

Publication number
CN117150934B
CN117150934B CN202311417175.8A CN202311417175A CN117150934B CN 117150934 B CN117150934 B CN 117150934B CN 202311417175 A CN202311417175 A CN 202311417175A CN 117150934 B CN117150934 B CN 117150934B
Authority
CN
China
Prior art keywords
state
parameter
transformer bushing
parameters
state parameter
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202311417175.8A
Other languages
Chinese (zh)
Other versions
CN117150934A (en
Inventor
张锦程
杨铭
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nanjing Zhongxin Zhidian Technology Co ltd
Original Assignee
Nanjing Zhongxin Zhidian Technology Co ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Nanjing Zhongxin Zhidian Technology Co ltd filed Critical Nanjing Zhongxin Zhidian Technology Co ltd
Priority to CN202311417175.8A priority Critical patent/CN117150934B/en
Publication of CN117150934A publication Critical patent/CN117150934A/en
Application granted granted Critical
Publication of CN117150934B publication Critical patent/CN117150934B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/10—Pre-processing; Data cleansing
    • G06F18/15—Statistical pre-processing, e.g. techniques for normalisation or restoring missing data
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/243—Classification techniques relating to the number of classes
    • G06F18/2431—Multiple classes
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/25—Fusion techniques
    • G06F18/254—Fusion techniques of classification results, e.g. of results related to same input data
    • G06F18/256—Fusion techniques of classification results, e.g. of results related to same input data of results relating to different input data, e.g. multimodal recognition
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N7/00—Computing arrangements based on specific mathematical models
    • G06N7/02—Computing arrangements based on specific mathematical models using fuzzy logic
    • G06N7/023—Learning or tuning the parameters of a fuzzy system

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Software Systems (AREA)
  • Automation & Control Theory (AREA)
  • Fuzzy Systems (AREA)
  • Geometry (AREA)
  • Probability & Statistics with Applications (AREA)
  • Computer Hardware Design (AREA)
  • Medical Informatics (AREA)
  • Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Algebra (AREA)
  • Computational Mathematics (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Pure & Applied Mathematics (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Testing Electric Properties And Detecting Electric Faults (AREA)

Abstract

The invention relates to the technical field of transformer bushing state monitoring, and discloses a transformer bushing state comprehensive monitoring system and an online data processing method thereof, wherein the data processing method comprises the following steps: s1: establishing a state parameter set and a state grade set, and collecting measured values of all state parameters; s2: normalizing the measured value of the state parameter; s3: the weight of each state parameter is distributed; s4: calculating a relation fuzzy matrix of each state parameter pair and each state grade; s5: comprehensively evaluating the state of the transformer bushing according to the weight of each state parameter and the membership degree of each state grade of each state parameter; s6: and predicting the state change trend of the transformer bushing based on the measured values of the state parameters. The method and the device can reflect the health state of the transformer bushing on the whole, predict the state of the transformer bushing, and improve the reliability and efficiency of the state detection of the transformer bushing.

Description

Comprehensive monitoring system for transformer bushing state and online data processing method thereof
Technical Field
The invention relates to the technical field of transformer bushing state monitoring, in particular to a transformer bushing state comprehensive monitoring system and an online data processing method thereof.
Background
The sleeve serves as an important component of the transformer, plays a role of fixing the lead wires and ensuring insulation from the outside, and is also a fault-prone part. In actual operation, the sleeve often causes serious accidents such as shutdown of a transformer, even burning explosion and the like due to defects such as insulation damage or local overheating and the like. Therefore, the fault diagnosis and the state evaluation of the transformer bushing are particularly important.
The sleeve still has many problems in daily operation and maintenance, and some potential faults of the sleeve are still difficult to find by the existing maintenance method, so that the safe and reliable operation of the sleeve cannot be ensured. How to discover the potential faults of the sleeve as soon as possible and accurately, monitor and predict the running state of the sleeve accurately in real time, and ensure the safe running of the transformer sleeve and even the whole transformer is a problem to be solved. In addition, a single casing state detection method or information parameter cannot carry enough fault state information, and diagnosis and evaluation misjudgment are easy to generate.
For example, the chinese patent with the grant publication number CN112163371B discloses a transformer bushing state assessment method, which fuses a plurality of indexes reflecting the state of the transformer bushing, comprehensively reflects the state of the bushing based on a fuzzy theory and a neural network algorithm, introduces a variable weight and a neural network algorithm on the basis of fuzzy comprehensive assessment, overcomes the limitation of conventional weight and the subjectivity of artificially determining membership functions, realizes self-updating and self-adaptation of membership functions through neural network training, and enables fuzzy control to have certain self-learning capability, thereby enabling the assessment result to be more reliable.
The patent with the application publication number of CN112163371A discloses a split type transformer bushing comprehensive evaluation method, electronic equipment and storage medium, wherein the method comprises the following steps: the method comprises the steps that a scheme set is formed by three phases of a split transformer, and decision information of each phase is obtained according to a preset decision standard, wherein the decision information is expressed in a triangle fuzzy number mode; constructing a corresponding relation between the fuzzy language variable and the triangular fuzzy number, and obtaining a fuzzy language variable decision matrix according to the test result and the judgment basis; judging whether a 'fail' occurs in the fuzzy language variable decision matrix, if so, judging that the sleeve state of the corresponding phase is 'fail'; if not, based on the corresponding relation between the fuzzy language variable and the triangular fuzzy number and the fuzzy language variable decision matrix, the comprehensive state of the split transformer bushing is evaluated, an evaluation result is obtained, and the overhaul priority is determined. The method can improve the accuracy of the evaluation result, improve the timeliness of maintenance and improve the equipment management efficiency.
The problems presented in the background art exist in the above patents: it is difficult to find some potential faults of the casing and safe and reliable operation of the casing cannot be ensured.
The information disclosed in this background section is only for enhancement of understanding of the general background of the invention and should not be taken as an acknowledgement or any form of suggestion that this information forms the prior art already known to a person of ordinary skill in the art.
Disclosure of Invention
The invention aims to overcome the defects of the prior art, provides a comprehensive monitoring system for the state of a transformer bushing and an online data processing method thereof, obtains reliable conclusion from complex, large-information-quantity and multi-level information sources through an intelligent algorithm, integrally reflects the health state of the transformer bushing, predicts the state of the transformer bushing by using a historical data and probability inference method, applies a prediction result to comprehensive evaluation of the state of the bushing, improves the accuracy of evaluation, and plays an important positive role in risk evaluation and decision making.
In order to solve the technical problems, the invention provides the following technical scheme:
on one hand, the invention provides a transformer bushing state comprehensive monitoring system, which comprises a data acquisition module, a database module, a state evaluation module, a state parameter prediction module and a remote management module, wherein:
the data acquisition module is used for acquiring state parameters of the transformer bushing;
the database module is used for storing the state parameters;
the state evaluation module is used for evaluating the state grade of the transformer bushing according to the state parameters;
the state parameter prediction module is used for extracting continuous state parameters from the database module, predicting the change trend of each state parameter and obtaining a predicted value of each state parameter;
the remote management module is used for a manager to inquire the state parameters and the state grade of the transformer bushing.
As a preferable scheme of the comprehensive monitoring system for the state of the transformer bushing, the invention comprises the following steps: the data acquisition module comprises an electrical test unit and an oil chromatography test unit;
the state parameters collected by the electric test unit comprise capacitance deviation, dielectric loss angle, end screen current and partial discharge capacity;
the state parameters collected by the oil chromatography test unit comprise hydrogen content, acetylene content, methane content, ethylene content and total hydrocarbon content.
As a preferable scheme of the comprehensive monitoring system for the state of the transformer bushing, the invention comprises the following steps: the controlled acquisition mode is that a data acquisition module receives an instruction, completes the acquisition of the state parameters and transmits the state parameters to a database;
the automatic acquisition mode is that the data acquisition module completes the acquisition of the state parameters in a preset time period according to the configured acquisition mode and frequency.
As a preferable scheme of the comprehensive monitoring system for the state of the transformer bushing, the invention comprises the following steps: the database module is also used for recording alarm records; verifying the validity of the data; data backup and data recovery; user rights and data security management.
As a preferable scheme of the comprehensive monitoring system for the state of the transformer bushing, the invention comprises the following steps: the state evaluation module is also used for sending out alarm information; and analyzing potential faults of the transformer bushing based on the change trend of the state parameter.
In a second aspect, the present invention provides a transformer bushing state online data processing method, including the following steps:
s1: establishing a state parameter set and a state grade set, and collecting measured values of the state parameters;
s2: normalizing the measured value of the state parameter;
s3: the weight of each state parameter is distributed;
s4: calculating a relation fuzzy matrix of each state parameter pair and each state grade;
s5: comprehensively evaluating the state of the transformer bushing according to the weight of each state parameter and the membership degree of each state grade of each state parameter;
s6: predicting the state change trend of the transformer bushing based on the measured values of the state parameters; the method comprises the following steps: firstly, reading the collected measured value of the state parameter, then calculating the predicted value of the state parameter by using a multi-task Gaussian process model, and finally predicting the state change trend of the transformer pipe sleeve based on the predicted value of the state parameter to obtain the predicted value of the state grade of the transformer pipe sleeve.
As a preferable scheme of the transformer bushing state online data processing method, the invention comprises the following steps: the state parameter set is denoted as X, in the form:
；
wherein,for the i-th state parameter, i=1, 2, … …, n, n is the total number of state parameters;
the state level set is denoted as Z, in the form:
；
wherein,indicating a normal state; />Representing an attention state; />Representing an abnormal state; />Indicating a severe condition.
As a preferable scheme of the transformer bushing state online data processing method, the invention comprises the following steps: the method for assigning the weights of the state parameters is as follows:
first, chang Quan weight of each state parameter is calculated based on analytic hierarchy processThen calculate the weight of each state parameter according to the following formula>：
；
Wherein,and->The values are Chang Quan weight and normalized values of the j state parameters, and the value range of i and j is a positive integer of 1-n;
the weight set for each state parameter is constructed and denoted as W, in the form:
W=；
wherein,the value range of i is 1,2, … … and n for the weight of the i-th state parameter;
as a preferable scheme of the transformer bushing state online data processing method, the invention comprises the following steps: the method for constructing the relation fuzzy matrix comprises the following steps:
selecting membership functions, calculating membership degree of each state parameter to each state grade, and recording as，/>，/>，/>Respectively represent state parameters->Status grade->~/>Membership degree of (3);
membership degree is formed into a relation fuzzy matrix R in the following form:
R=。
as a preferable scheme of the transformer bushing state online data processing method, the invention comprises the following steps: the method for comprehensively evaluating the state of the transformer bushing comprises the following steps: constructing a comprehensive state level matrix A, wherein the formula is as follows:
；
a is a 4-dimensional row vector, denoted a=. Wherein (1)>~/>Respectively indicate that the evaluation result is->~/>Probability of (2);
the final state level is determined using the roulette method as follows:
creates an interval (0, 1]The interval (0, 1]Divided into 4 sections, wherein the 1 st section to the 4 th section are respectively%，/>]，(/>，]，(/>，/>+/>]，(/>，1]. Generating a random number b, if b falls in the ith section, the state level of the roulette is +.>The value range of i is [1,2,3,4 ]]；
And repeatedly carrying out roulette, and taking out the state grade with the largest number of occurrences as the final result of the comprehensive evaluation of the state of the transformer sleeve.
As a preferable scheme of the transformer bushing state online data processing method, the invention comprises the following steps: the input of the multi-task Gaussian process model is the actual observed value of each state parameter; the output of the multi-task Gaussian process model is a predicted value of each state parameter at a future time; each state parameter corresponds to a task of the multitasking gaussian process model.
The training set of the multitasking gaussian process model comprises an input set X and an output set Y, wherein X is in the form of:
X=；
wherein,an observation set representing a state parameter is provided in the form:
=/>；
wherein,representing a status parameter +.>The j-th observation value of the state parameter is 1,2, … … n, the j-th observation value of the state parameter is 1,2, … …, m, m represents the number of observation values of each state parameter;
the form of the output set Y is as follows:
Y=；
wherein,representing a status parameter +.>Observation set +.>The observed values of the delta observed points are delayed as shown in the following formula:
；
wherein, the value range of i is 1,2, … …, n and delta represent prediction intervals;
a new set of inputs for prediction is noted asThe form is as follows:
=/>；
=/>；
the output of the model is recorded asThe expression is as follows:
=/>；
wherein,representing the output value of each task of the multitasking Gaussian process model, i.e. each state parameter is +_ relative to the input set>The predicted values of delta observation points are delayed.
As a preferable scheme of the transformer bushing state online data processing method, the invention comprises the following steps: the training and application method of the multi-task Gaussian process model comprises the following steps:
s100: the input training set comprises an input set X and an output set Y;
s200: setting a task relation matrix omega and a kernel function k;
the sizes of the elements in the task relation matrix represent the sizes of the mutual influence among different state parameters, and the form is as follows:
；
wherein n represents the number of tasks, i.e. the total number of state parameters;the value range of the parameter which represents the task relation, i and j is 1,2, … … and n;
selecting a gaussian kernel function as the kernel function of a multi-tasking gaussian process modelThe formula is as follows:
；
wherein,，/>the input sets of the i and j state parameters are respectively represented, and the value ranges of the i and j are 1,2, … … and n; sigma represents a variance parameter, and L represents a feature scale parameter; />Representation matrix->，/>Euclidean distance between them;
s300: the negative log-marginal likelihood function N of the multitasking Gaussian process model is calculated as follows:
；
wherein Y represents an output set in a training set, K is a covariance matrix, and the expression is as follows:
K=；
wherein,=/>the value range of i is 1,2, … …, n and k are kernel functions;
s400: performing parameter optimization to obtain a trained multi-task Gaussian process model;
the parameters to be optimized include task relation parametersVariance parameter sigma, characteristic scale parameter L; the minimization of the negative log-marginal likelihood function N is iterated by a gradient descent method,when the preset maximum iteration times are reached, extracting corresponding parameters in the negative logarithmic marginal likelihood function as task relation parameters +.>The variance parameter sigma and the characteristic scale parameter L are optimized;
s500: predicting the state of the state parameter according to the trained multi-task Gaussian process model;
input set for a new set of inputsCalculating predicted values of each task, i.e. each state parameter, from the trained multitasking gaussian process model>The formula is as follows:
；
wherein Y represents the output set in the training set,an ith column vector representing a task relationship matrix, < >>Is a matrix of n x n representing the new input set +.>The covariance matrix with the input set X in the training set is expressed as follows:
；
wherein,representation matrix->Middle (f)The values of the elements in the i row and the j column are 1,2, … … and n.
Compared with the prior art, the invention has the following beneficial effects: the complex, large-information-quantity and multi-level information sources are used for obtaining reliable conclusion through an intelligent algorithm, and the health state of the transformer bushing is reflected on the whole; the method has the advantages that the state of the transformer bushing is predicted by using historical data and probability inference, the adopted multi-task Gaussian process model can realize simultaneous modeling of a plurality of tasks functionally, besides a data set, the multi-task Gaussian process model can also learn the interrelationship among the tasks autonomously, and can absorb more useful information, so that the prediction accuracy of the model is improved; the prediction result is applied to comprehensive evaluation of the bushing state, so that the reliability and efficiency of the state detection of the transformer bushing are improved, potential problems can be found early, appropriate measures are taken to optimize the maintenance plan of the transformer bushing, including preventive maintenance, repair plan and the like, and the safe operation of the transformer bushing is ensured to improve the service life and reliability of the transformer bushing.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in 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. Wherein:
fig. 1 is a schematic structural diagram of a transformer bushing state integrated monitoring system provided by the present application;
FIG. 2 is a flowchart of a transformer bushing status online data processing method provided by the present application;
FIG. 3 is a functional schematic of a multi-tasking Gaussian process model used in the present application;
FIG. 4 is a flow chart of the training and application of the multi-tasking Gaussian process model provided in the present application.
Detailed Description
The following detailed description of the present invention is made with reference to the accompanying drawings and specific embodiments, and it is to be understood that the specific features of the embodiments and the embodiments of the present invention are detailed description of the technical solutions of the present invention, and not limited to the technical solutions of the present invention, and that the embodiments and the technical features of the embodiments of the present invention may be combined with each other without conflict.
Example 1
The embodiment introduces a comprehensive monitoring system for the state of a transformer bushing, referring to fig. 1, the system includes a data acquisition module, a database module, a state evaluation module, a state parameter prediction module, and a remote management module, wherein:
the data acquisition module is used for acquiring state parameters of the transformer bushing.
The module comprises an electrical test unit and an oil chromatography test unit;
the state parameters collected by the electric test unit comprise capacitance deviation, dielectric loss angle, end screen current and partial discharge capacity;
the state parameters collected by the oil chromatography test unit comprise hydrogen content, acetylene content, methane content, ethylene content and total hydrocarbon content.
The state parameters of the transformer bushing can reflect: insulation degradation conditions, discharge failure, overheat conditions, oil starvation conditions. The system comprehensively considers the state parameters and comprehensively evaluates and predicts the state of the transformer bushing.
The data acquisition mode comprises automatic acquisition and controlled acquisition; wherein:
the controlled acquisition mode is that a data acquisition module receives an instruction, completes the acquisition of the state parameters and transmits the state parameters to a database;
the automatic acquisition mode is that the data acquisition module completes the acquisition of the state parameters in a preset time period according to the configured acquisition mode and frequency.
The database module is used for storing the state parameters, recording alarm records, verifying data validity, backing up and recovering data, and managing user rights and data security.
After the collector finishes the collection and transmission of the data, the system software receives the data transmitted by the collector, and as data errors possibly occur in the data collection and transmission process, the validity verification is to analyze the received data according to the communication protocol of the data and the valid range of the data, and the collector is required to retransmit the data or to acquire the data again if the data errors.
The database periodically backs up the data to ensure the security and integrity of the data. In the event of a system failure or data loss, backup data may be used for recovery and reconstruction.
The database manages user rights and data security. It can define access rights for different users, ensuring that only authorized personnel can view and modify specific data. In addition, the database provides security measures such as data encryption, authentication and the like to protect confidentiality and integrity of sensitive data.
The state evaluation module is used for extracting state parameters from the database, then processing the state parameters to obtain the state grade of the sleeve, and storing the state grade into the database;
when the monitoring data exceeds a set threshold or reaches an early warning condition, corresponding alarm and notification are triggered, and an alarm event is recorded by a database.
The state parameter prediction module is used for extracting continuous state parameters from the database module, predicting the change trend of each state parameter and obtaining a predicted value of each state parameter;
this module may also work in conjunction with a state assessment module to predict potential faults.
The remote management module is used for a manager to inquire and search the state parameters, the state grade and the predicted state grade of the transformer sleeve, and the state of the transformer sleeve is presented to the manager. The manager can obtain the required data according to specific conditions and requirements. The user can inquire according to the time range, the transformer identification, the fault type and other parameters, and obtain relevant monitoring data.
When the sleeve state evaluation result is abnormal, an alarm signal is generated, and the module can timely inform related personnel in a mode of short messages, mails, sound alarms and the like, so that timely measures are ensured to be taken to avoid serious consequences.
Example 2
This embodiment is a second embodiment of the present invention; based on the same inventive concept as embodiment 1, this embodiment describes an online data processing method for transformer bushing status, and referring to fig. 2, the method includes the following steps:
s1: establishing a state parameter set and a state grade set, and collecting measured values of all state parameters;
the state parameter set is denoted as X, in the form:
；
wherein,for the i-th state parameter, i=1, 2, … …, n, n is the total number of state parameters;
the state level set is denoted as Z, in the form:
；
wherein,indicating a normal state; />Representing an attention state; />Representing an abnormal state; />Indicating a severe condition.
S2: and carrying out normalization processing on the measured value of the state parameter, wherein the method comprises the following steps:
setting a first threshold value of each state parameterSecond threshold->；
If it is≤/>Status parameter->The influence on the transformer bushing is negligible;
if it is≥/>Status parameter->Is far beyond the dangerous limit;
normalizing the acquired numerical value of each state parameter according to the following formula:
wherein,representing the initial value of the acquired i-th state parameter,/->Representing the normalized value of the ith state parameter.
S3: the method for distributing the weight of each state parameter comprises the following steps: the method comprises the following steps:
first, chang Quan weight of each state parameter is calculated based on analytic hierarchy processThen calculate the weight of each state parameter according to the following formula>：
；
Wherein,and->The values are Chang Quan weight and normalized values of the j state parameters, and the value range of i and j is a positive integer of 1-n;
the weight set for each state parameter is constructed and denoted as W, in the form:
W=；
s4: the relation fuzzy matrix of each state parameter pair and each state grade is calculated, and the method is as follows:
selecting membership functions, calculating membership degree of each state parameter to each state grade, and recording as，/>，/>，/>Respectively represent state parameters->Status grade->~/>The formula is as follows:
membership degree is formed into a relation fuzzy matrix R in the following form:
R=。
s5: and comprehensively evaluating the state of the transformer bushing according to the weight of each state parameter and the membership degree of each state grade by the state parameters, wherein the method comprises the following steps of: constructing a comprehensive state level matrix A, wherein the formula is as follows:
；
a is a 4-dimensional row vector, denoted a=. Wherein (1)>~/>Respectively represent the evaluation results as/>~/>Probability of (2);
the final state level is determined using the roulette method as follows:
creates an interval (0, 1]The interval (0, 1]Divided into 4 sections, wherein the 1 st section to the 4 th section are respectively%，/>]，(/>，]，(/>，/>+/>]，(/>，1]. Generating a random number b, if b falls in the ith section, the state level of the roulette is +.>The value range of i is [1,2,3,4 ]]；
And repeatedly carrying out roulette, and taking out the state grade with the largest number of occurrences as the final result of the comprehensive evaluation of the state of the transformer sleeve.
If the evaluation result isThe state quantity of the transformer bushing is far smaller than the specified standard limit value, so that the transformer bushing can normally operate for a long time, and maintenance can be temporarily not arranged.
If the evaluation result isThe performance of the individual state quantity of the transformer bushing is reduced, but the continuous operation of the transformer bushing is not influenced, and the maintenance is arranged according to the normal period.
If the evaluation result isIt means that the individual state quantity of the transformer bushing may have exceeded the limit value, which can still continue to operate, but the power outage maintenance should be scheduled as soon as possible.
If the evaluation result isIt means that the performance of the transformer bushing is seriously degraded, and the power outage maintenance is required to be immediately arranged.
S6: predicting the state change trend of the transformer bushing based on the measured values of the state parameters; the method comprises the following steps: firstly, reading the collected measured value of the state parameter, then calculating the predicted value of the state parameter by using a multi-task Gaussian process model, and finally predicting the state change trend of the transformer pipe sleeve based on the predicted value of the state parameter to obtain the predicted value of the state grade of the transformer pipe sleeve.
Example 3
The embodiment provides a method for predicting the state change trend of a transformer bushing. Firstly, reading the collected measured value of the state parameter, then calculating the predicted value of the state parameter by using a multi-task Gaussian process model, and finally predicting the state change trend of the transformer pipe sleeve based on the predicted value of the state parameter to obtain the predicted value of the state grade of the transformer pipe sleeve.
Referring to FIG. 3, the inputs to the multitasking Gaussian process model are actual observations of each state parameter; the output of the multi-task Gaussian process model is a predicted value of each state parameter at a future time; each state parameter corresponds to a task of the multitasking gaussian process model.
The training set of the multitasking gaussian process model comprises an input set X and an output set Y, wherein X is in the form of:
X=；
wherein,an observation set representing a state parameter is provided in the form:
=/>；
wherein,representing a status parameter +.>The j-th observation value of the state parameter is 1,2, … … n, the j-th observation value of the state parameter is 1,2, … …, m, m represents the number of observation values of each state parameter;
the form of the output set Y is as follows:
Y=；
wherein,representing a status parameter +.>Observation set +.>The observed values of the delta observed points are delayed as shown in the following formula:
；
wherein, the value range of i is 1,2, … …, n and delta represent prediction intervals;
a new set of inputs for prediction is noted asThe form is as follows:
=/>；
=/>；
the output of the model is recorded asThe expression is as follows:
=/>；
wherein,representing the output value of each task of the multitasking Gaussian process model, i.e. each state parameter is +_relative to the input set>The predicted values of delta observation points are delayed.
Referring to fig. 4, the training and application method of the multitasking gaussian process model is as follows:
s100: the input training set comprises an input set X and an output set Y;
s200: setting a task relation matrix omega and a kernel function k;
the sizes of the elements in the task relation matrix represent the sizes of the mutual influence among different state parameters, and the form is as follows:
；
；
wherein n represents the number of tasks, i.e. the total number of state parameters;the value range of the parameter which represents the task relation, i and j is 1,2, … … and n;
selecting a gaussian kernel function as the kernel function of a multi-tasking gaussian process modelThe formula is as follows:
；
wherein,，/>the input sets of the i and j state parameters are respectively represented, and the value ranges of the i and j are 1,2, … … and n; sigma represents a variance parameter, and L represents a feature scale parameter; />Representation matrix->，/>Euclidean distance between them;
s300: the negative log-marginal likelihood function N of the multitasking Gaussian process model is calculated as follows:
；
wherein Y represents an output set in a training set, K is a covariance matrix, and the expression is as follows:
K=；
wherein,=/>the value range of i is 1,2, … … and n;
s400: performing parameter optimization to obtain a trained multi-task Gaussian process model;
the parameters to be optimized include task relation parametersVariance parameter sigma, characteristic scale parameter L; iterative minimizing of the negative logarithmic marginal likelihood function N is carried out by a gradient descent method, and when the preset maximum iterative times are reached, the corresponding parameters in the negative logarithmic marginal likelihood function are extracted and used as task relation parameters +.>The variance parameter sigma and the characteristic scale parameter L are optimized;
s500: predicting the state of the state parameter according to the trained multi-task Gaussian process model;
input set for a new set of inputsCalculating predicted values of each task, i.e. each state parameter, from the trained multitasking gaussian process model>The formula is as follows: />
；
Wherein Y represents the output set in the training set,an ith column vector representing a task relationship matrix, < >>Is a matrix of n x n representing the new input set +.>The covariance matrix with the input set X in the training set is expressed as follows:
；
wherein,representation matrix->The value range of the elements i, j in the ith row and the jth column is 1,2, … … and n.
In summary, the multi-task Gaussian process model adopted by the invention can realize simultaneous modeling of a plurality of tasks functionally, and besides the data set, the multi-task Gaussian process can also learn the interrelationship among the tasks autonomously, so that more useful information can be absorbed, and the performance of the model is improved.
The predictions provided by the gaussian process model may be used to optimize the maintenance plan for the transformer bushing. Based on the predicted state change trend and uncertainty information, more effective maintenance strategies including preventive maintenance, repair plans, and the like can be formulated. This will help to improve the life and reliability of the transformer bushing.
It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The embodiments of the present invention have been described above with reference to the accompanying drawings, but the present invention is not limited to the above-described embodiments, which are merely illustrative and not restrictive, and many forms may be made by those having ordinary skill in the art without departing from the spirit of the present invention and the scope of the claims, which are all within the protection of the present invention.

Claims (12)

1. Transformer bushing state integrated monitoring system, its characterized in that: the system comprises a data acquisition module, a database module, a state evaluation module, a state parameter prediction module and a remote management module, wherein:
the data acquisition module is used for acquiring state parameters of the transformer bushing;
the database module is used for storing the state parameters;
the state evaluation module is used for evaluating the state grade of the transformer bushing according to the state parameters;
the state parameter prediction module is used for extracting continuous state parameters from the database module, predicting the change trend of each state parameter and obtaining a predicted value of each state parameter;
the remote management module is used for a manager to inquire the state parameters and the state grade of the transformer bushing.
2. The transformer bushing state integrated monitoring system of claim 1, wherein: the data acquisition module comprises an electrical test unit and an oil chromatography test unit;
the state parameters collected by the electric test unit comprise capacitance deviation, dielectric loss angle, end screen current and partial discharge capacity;
the state parameters collected by the oil chromatography test unit comprise hydrogen content, acetylene content, methane content, ethylene content and total hydrocarbon content.
3. The transformer bushing state integrated monitoring system of claim 2, wherein: the data acquisition mode of the data acquisition module comprises automatic acquisition and controlled acquisition; wherein:
the controlled acquisition mode is that a data acquisition module receives an instruction, completes the acquisition of the state parameters and transmits the state parameters to a database;
the automatic acquisition mode is that the data acquisition module completes the acquisition of the state parameters in a preset time period according to the configured acquisition mode and frequency.
4. A transformer bushing state integrated monitoring system according to claim 3, wherein: the database module is also used for recording alarm records; verifying the validity of the data; data backup and data recovery; user rights and data security management.
5. The transformer bushing state integrated monitoring system of claim 4, wherein: the state evaluation module is also used for sending out alarm information; and analyzing potential faults of the transformer bushing based on the change trend of the state parameter.
6. The transformer bushing state online data processing method is realized based on the transformer bushing state comprehensive monitoring system according to any one of claims 1-5, and is characterized in that: the method comprises the following steps:
s1: establishing a state parameter set and a state grade set, and collecting measured values of the state parameters;
s2: normalizing the measured value of the state parameter;
s3: the weight of each state parameter is distributed;
s4: calculating a relation fuzzy matrix of each state parameter pair and each state grade;
s5: comprehensively evaluating the state of the transformer bushing according to the weight of each state parameter and the membership degree of each state grade of each state parameter;
s6: predicting the state change trend of the transformer bushing based on the measured values of the state parameters; the method comprises the following steps: firstly, reading the collected measured value of the state parameter, then calculating the predicted value of the state parameter by using a multi-task Gaussian process model, and finally predicting the state change trend of the transformer pipe sleeve based on the predicted value of the state parameter to obtain the predicted value of the state grade of the transformer pipe sleeve.
7. The transformer bushing state online data processing method of claim 6, wherein:
the state parameter set is denoted as X, in the form:
；
wherein,for the i-th state parameter, i=1, 2, … …, n, n is the total number of state parameters;
the state level set is denoted as Z, in the form:
；
wherein the method comprises the steps of，Indicating a normal state; />Representing an attention state; />Representing an abnormal state; />Indicating a severe condition.
8. The transformer bushing state online data processing method of claim 7, wherein: the method for assigning the weights of the state parameters is as follows:
first, chang Quan weight of each state parameter is calculated based on analytic hierarchy processThen calculate the weight of each state parameter according to the following formula>：
；
Wherein,and->The values are Chang Quan weight and normalized values of the j state parameters, and the value range of i and j is a positive integer of 1-n;
the weight set for each state parameter is constructed and denoted as W, in the form:
W=；
wherein,the i-th state parameter is weighted, and the value range of i is 1,2, … … and n.
9. The transformer bushing state online data processing method of claim 8, wherein: the method for constructing the relation fuzzy matrix comprises the following steps:
selecting membership functions, calculating membership degree of each state parameter to each state grade, and recording as，/>，/>，/>Respectively represent state parameters->Status grade->~/>Membership degree of (3);
membership degree is formed into a relation fuzzy matrix R in the following form:
R=。
10. the transformer bushing state online data processing method of claim 9, wherein: the method for comprehensively evaluating the state of the transformer bushing comprises the following steps: constructing a comprehensive state level matrix A, wherein the formula is as follows:
；
a is a 4-dimensional row vector, denoted a=The method comprises the steps of carrying out a first treatment on the surface of the Wherein (1)>~/>Respectively represent the evaluation results as~/>Probability of (2);
the final state level is determined using the roulette method as follows:
creates an interval (0, 1]The interval (0, 1]Divided into 4 sections, wherein the 1 st section to the 4 th section are respectively%，/>]，(/>，]，(/>，/>+/>]，(/>，1]The method comprises the steps of carrying out a first treatment on the surface of the Generating a random number b, if b falls in the ith section, the state level of the roulette is +.>The value range of i is [1,2,3,4 ]]；
And repeatedly carrying out roulette, and taking out the state grade with the largest number of occurrences as the final result of the comprehensive evaluation of the state of the transformer sleeve.
11. The transformer bushing state online data processing method of claim 10, wherein: the input of the multi-task Gaussian process model is the actual observed value of each state parameter; the output of the multi-task Gaussian process model is a predicted value of each state parameter at a future time; each state parameter corresponds to a task of the multi-task gaussian process model;
the training set of the multitasking gaussian process model comprises an input set X and an output set Y, wherein X is in the form of:
X=；
wherein,an observation set representing a state parameter is provided in the form:
=/>；
wherein,representing a status parameter +.>The j-th observation value of the state parameter is 1,2, … … n, the j-th observation value of the state parameter is 1,2, … …, m, m represents the number of observation values of each state parameter;
the form of the output set Y is as follows:
Y=；
wherein,representing a status parameter +.>Observation set +.>The observed values of the delta observed points are delayed as shown in the following formula:
；
wherein, the value range of i is 1,2, … …, n and delta represent prediction intervals;
a new set of inputs for prediction is noted asThe form is as follows:
=/>；
=/>；
the output of the model is recorded asThe expression is as follows:
=/>；
wherein,representing the output value of each task of the multitasking Gaussian process model, i.e. each state parameter is +_ relative to the input set>The predicted values of delta observation points are delayed.
12. The transformer bushing state online data processing method of claim 11, wherein: the training and application method of the multi-task Gaussian process model comprises the following steps:
s100: the input training set comprises an input set X and an output set Y;
s200: setting a task relation matrix omega and a kernel function k;
the sizes of the elements in the task relation matrix represent the sizes of the mutual influence among different state parameters, and the form is as follows:
；
wherein n represents the number of tasks, i.e. the total number of state parameters;the value range of the parameter which represents the task relation, i and j is 1,2, … … and n;
selecting a gaussian kernel function as the kernel function of a multi-tasking gaussian process modelThe formula is as follows:
；
wherein,，/>the input sets of the i and j state parameters are respectively represented, and the value ranges of the i and j are 1,2, … … and n; sigma represents a variance parameter, and L represents a feature scale parameter; />Representation matrix->，/>Euclidean distance between them;
s300: the negative log-marginal likelihood function N of the multitasking Gaussian process model is calculated as follows:
；
wherein Y represents an output set in a training set, K is a covariance matrix, and the expression is as follows:
K=；
wherein,=/>the value range of i is 1,2, … …, n and k are kernel functions;
s400: performing parameter optimization to obtain a trained multi-task Gaussian process model;
the parameters to be optimized include task relation parametersVariance parameter sigma, characteristic scale parameter L; iterative minimizing of the negative logarithmic marginal likelihood function N is carried out by a gradient descent method, and when the preset maximum iterative times are reached, the corresponding parameters in the negative logarithmic marginal likelihood function are extracted and used as task relation parameters +.>The variance parameter sigma and the characteristic scale parameter L are optimized;
s500: predicting the state of the state parameter according to the trained multi-task Gaussian process model;
input set for a new set of inputsCalculating predicted values of each task, i.e. each state parameter, from the trained multitasking gaussian process model>The formula is as follows:
；
wherein Y represents the output set in the training set,an ith column vector representing a task relationship matrix, < >>Is a matrix of n x n representing the new input set +.>The covariance matrix with the input set X in the training set is expressed as follows:
；
wherein,representation matrix->The value range of the elements i, j in the ith row and the jth column is 1,2, … … and n.
CN202311417175.8A 2023-10-30 2023-10-30 Comprehensive monitoring system for transformer bushing state and online data processing method thereof Active CN117150934B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202311417175.8A CN117150934B (en) 2023-10-30 2023-10-30 Comprehensive monitoring system for transformer bushing state and online data processing method thereof

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202311417175.8A CN117150934B (en) 2023-10-30 2023-10-30 Comprehensive monitoring system for transformer bushing state and online data processing method thereof

Publications (2)

Publication Number Publication Date
CN117150934A CN117150934A (en) 2023-12-01
CN117150934B true CN117150934B (en) 2024-01-26

Family

ID=88897171

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202311417175.8A Active CN117150934B (en) 2023-10-30 2023-10-30 Comprehensive monitoring system for transformer bushing state and online data processing method thereof

Country Status (1)

Country Link
CN (1) CN117150934B (en)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117409011B (en) * 2023-12-15 2024-03-01 南京中鑫智电科技有限公司 High-voltage sleeve surface pollution monitoring method and system based on target identification
CN117495207B (en) * 2023-12-29 2024-03-22 国网四川省电力公司超高压分公司 Power transformer health state evaluation method
CN118134464B (en) * 2024-05-06 2024-09-10 国网江西省电力有限公司电力科学研究院 Direct current measurement device fault treatment measure generation method and system

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
RU2242830C1 (en) * 2003-05-15 2004-12-20 Александр Николаевич Рассальский Power transformer monitoring device
CN102841245A (en) * 2012-09-24 2012-12-26 广西电网公司电力科学研究院 Device for detecting end screen current parameter of high-voltage transformer bushing in real time, and control method for device
CN103926491A (en) * 2014-04-21 2014-07-16 国家电网公司 Transformer state evaluation method with influence of direct-current magnetic bias taken into account
CN111597757A (en) * 2020-05-13 2020-08-28 山东女子学院 GP model auxiliary SLPSO algorithm based on multi-target point adding criterion
CN112163371A (en) * 2020-09-18 2021-01-01 山东电工电气集团有限公司 Transformer bushing state evaluation method
CN113361751A (en) * 2021-05-17 2021-09-07 同济大学 Firefly algorithm optimization-based product quality prediction method and device for RBF model
WO2021185177A1 (en) * 2020-03-16 2021-09-23 福建省特种设备检验研究院 Method for evaluating health status of petrochemical atmospheric oil storage tank using data from multiple sources
CN113469409A (en) * 2021-05-24 2021-10-01 国网新疆电力有限公司营销服务中心(资金集约中心、计量中心) Gaussian process-based state prediction method and device for electric energy metering device
CN115544793A (en) * 2022-10-26 2022-12-30 国网安徽省电力有限公司经济技术研究院 State evaluation and service life prediction method for power transformer

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
RU2242830C1 (en) * 2003-05-15 2004-12-20 Александр Николаевич Рассальский Power transformer monitoring device
CN102841245A (en) * 2012-09-24 2012-12-26 广西电网公司电力科学研究院 Device for detecting end screen current parameter of high-voltage transformer bushing in real time, and control method for device
CN103926491A (en) * 2014-04-21 2014-07-16 国家电网公司 Transformer state evaluation method with influence of direct-current magnetic bias taken into account
WO2021185177A1 (en) * 2020-03-16 2021-09-23 福建省特种设备检验研究院 Method for evaluating health status of petrochemical atmospheric oil storage tank using data from multiple sources
CN111597757A (en) * 2020-05-13 2020-08-28 山东女子学院 GP model auxiliary SLPSO algorithm based on multi-target point adding criterion
CN112163371A (en) * 2020-09-18 2021-01-01 山东电工电气集团有限公司 Transformer bushing state evaluation method
CN113361751A (en) * 2021-05-17 2021-09-07 同济大学 Firefly algorithm optimization-based product quality prediction method and device for RBF model
CN113469409A (en) * 2021-05-24 2021-10-01 国网新疆电力有限公司营销服务中心(资金集约中心、计量中心) Gaussian process-based state prediction method and device for electric energy metering device
CN115544793A (en) * 2022-10-26 2022-12-30 国网安徽省电力有限公司经济技术研究院 State evaluation and service life prediction method for power transformer

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
EMD法在介损在线测量数据处理中的应用;曹建;杨晓辉;;云南电力技术(第03期);96-100 *
基于局部密度聚类算法的变压器故障状态评估;罗伟明;吴帆;黄业广;吴杰康;覃炜梅;龚杰;金尚婷;;广东电力(第08期);88-97 *
基于模糊综合评判的变压器套管绝缘状态评估;李志超;李卫国;夏喻;张书琦;汤浩;贾鹏飞;陈绍君;;中国电力(第04期);27-31+65 *
电力变压器状态评估及故障诊断方法研究;《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》(第 第02期期);C042-730 *
高斯过程建模方法在工业过程中的应用;肖红军;刘乙奇;黄道平;;华南理工大学学报(自然科学版)(第12期);42-49+58 *

Also Published As

Publication number Publication date
CN117150934A (en) 2023-12-01

Similar Documents

Publication Publication Date Title
CN117150934B (en) Comprehensive monitoring system for transformer bushing state and online data processing method thereof
Hsu et al. Wind turbine fault diagnosis and predictive maintenance through statistical process control and machine learning
Zeng et al. Dynamic risk assessment based on statistical failure data and condition-monitoring degradation data
CN106600115A (en) Intelligent operation and maintenance analysis method for enterprise information system
Song et al. Narrative texts-based anomaly detection using accident report documents: The case of chemical process safety
CN117563184B (en) Energy storage fire control system based on thing networking
CN109492790A (en) Wind turbines health control method based on neural network and data mining
Alvarez Real-time fault detection and diagnosis using intelligent monitoring and supervision systems
CN117689373A (en) Maintenance decision support method for energy router of flexible direct-current traction power supply system
CN118316188A (en) Asset monitoring system and method applied to power grid data class
CN117007977B (en) Energy storage battery health state diagnosis method
Gopakumar et al. Determination of Power Transformer Fault’s Severity Based on Fuzzy Logic Model with GR, Level and DGA Interpretation
Ma Methods and systems for fault diagnosis in nuclear power plants
RU2735296C1 (en) Method for multi-parameter monitoring of complex electrotechnical objects state
D’Atri et al. Improving pipeline risk models by using data mining techniques
CN118411154B (en) Power distribution equipment safety state assessment method and system
CN118153948B (en) Hydrogen energy storage equipment leakage fault early warning method and device based on deep learning
CN118134206B (en) Safety control method and system for hydrogen adding station
CN118312923B (en) Intelligent park-oriented equipment measurement method and computer equipment
CN117994076A (en) Natural gas pipe network scheduling energy-saving assessment method based on big data
Yuan et al. Research on power grid outage risk assessment and early warning model based on intelligent decision algorithm
CN118692219A (en) Environment monitoring and early warning method and system for safety helmet
Чулкін et al. Justification of possibility of ukranian npp’s exploitation time assessment by expert methods
Mandelli et al. Application of Margin-Based Methods to Assess System Health
Basit et al. Predictive Maintenance Using Deep Learning: Enhancing Reliability and Reducing Electrical System Downtime

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