CN112966990B - Comprehensive state evaluation method for power transformation equipment - Google Patents

Comprehensive state evaluation method for power transformation equipment Download PDF

Info

Publication number
CN112966990B
CN112966990B CN202110539663.0A CN202110539663A CN112966990B CN 112966990 B CN112966990 B CN 112966990B CN 202110539663 A CN202110539663 A CN 202110539663A CN 112966990 B CN112966990 B CN 112966990B
Authority
CN
China
Prior art keywords
index
state
data
abnormal
evaluation
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
CN202110539663.0A
Other languages
Chinese (zh)
Other versions
CN112966990A (en
Inventor
童超
朱自伟
张益宁
童军心
李唐兵
王华云
张宇
王鹏
万华
刘玉婷
徐碧川
童涛
曾磊磊
周友武
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.)
State Grid Jiangxi Electric Power Co ltd
State Grid Corp of China SGCC
Nanchang University
Electric Power Research Institute of State Grid Jiangxi Electric Power Co Ltd
Original Assignee
State Grid Jiangxi Electric Power Co ltd
State Grid Corp of China SGCC
Nanchang University
Electric Power Research Institute of State Grid Jiangxi Electric Power 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 State Grid Jiangxi Electric Power Co ltd, State Grid Corp of China SGCC, Nanchang University, Electric Power Research Institute of State Grid Jiangxi Electric Power Co Ltd filed Critical State Grid Jiangxi Electric Power Co ltd
Priority to CN202110539663.0A priority Critical patent/CN112966990B/en
Publication of CN112966990A publication Critical patent/CN112966990A/en
Application granted granted Critical
Publication of CN112966990B publication Critical patent/CN112966990B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
    • G06Q10/06311—Scheduling, planning or task assignment for a person or group
    • G06Q10/063114—Status monitoring or status determination for a person or group
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Energy or water supply

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Human Resources & Organizations (AREA)
  • Physics & Mathematics (AREA)
  • Economics (AREA)
  • General Physics & Mathematics (AREA)
  • Strategic Management (AREA)
  • Theoretical Computer Science (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Educational Administration (AREA)
  • Development Economics (AREA)
  • Tourism & Hospitality (AREA)
  • Marketing (AREA)
  • General Business, Economics & Management (AREA)
  • Mathematical Physics (AREA)
  • Pure & Applied Mathematics (AREA)
  • Mathematical Analysis (AREA)
  • Data Mining & Analysis (AREA)
  • Health & Medical Sciences (AREA)
  • Game Theory and Decision Science (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Computational Mathematics (AREA)
  • Mathematical Optimization (AREA)
  • Public Health (AREA)
  • Computing Systems (AREA)
  • Primary Health Care (AREA)
  • Water Supply & Treatment (AREA)
  • Algebra (AREA)
  • General Health & Medical Sciences (AREA)
  • Databases & Information Systems (AREA)
  • Software Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Testing Electric Properties And Detecting Electric Faults (AREA)
  • Supply And Distribution Of Alternating Current (AREA)
  • Housings And Mounting Of Transformers (AREA)

Abstract

The invention discloses a method for evaluating the comprehensive state of power transformation equipment, which imports the processing result of state index data and evaluates the state of a power transformer to represent a data set of the abnormal state of the power transformer
Figure DEST_PATH_IMAGE001
On the basis, classifying the state indexes, processing index data, constructing a layered evaluation system, establishing index weights by using an objective fusion weighting method, and obtaining a state evaluation result of the power transformer according to the index weights; state evaluation of an on-line monitoring device with abnormal data sets caused by abnormalities in the operation of the on-line monitoring device
Figure 524987DEST_PATH_IMAGE002
On the basis, the state evaluation comprises three steps of constructing a characterization relation between data of an abnormal mode and an abnormal state of the online monitoring device, characterizing the probability of the abnormal state, and finishing the state evaluation of the online monitoring device by using a membership function in a fuzzy evaluation theory. The invention respectively establishes the state evaluation models of the power transformer and the online monitoring device, thereby forming a comprehensive state evaluation system of the power transformation equipment based on data driving.

Description

Comprehensive state evaluation method for power transformation equipment
Technical Field
The invention relates to a comprehensive state evaluation method for power transformation equipment, which is used for evaluating a power transformer and a monitoring device respectively according to online monitoring data and belongs to the field of state evaluation of power equipment.
Background
The transformer equipment is used as the hub equipment for energy conversion and transmission of the power system, has a plurality of components, scores all indexes reflecting the operation state of the transformer in the existing national grid power transformer evaluation guide, and mixes the scoring results together to obtain the final state result of the transformer. The evaluation mode is complete, but the relation between the running state and the characteristics of the transformer is ignored, and the evaluation result lacks pertinence. In general, in the operation process of the transformer, the representation degrees of different indicator quantities of state quantity data to the state of the transformer are different, and the accuracy requirements of on-line monitoring and on-line testing on the indicator data are also different, so that in a transformer state evaluation system based on the on-line monitoring data and the on-line testing data, the weights of the indicator quantities in the reflected final evaluation result should be set differently.
The on-line monitoring device is used as real-time measuring equipment for the state index of the transformer and generally comprises a sensor module, a digital-analog signal conversion module and a communication module; the on-line monitoring device is used as a module closest to the actual index state of the transformer, and the running state of the on-line monitoring device is related to the reliability of data received by the system platform. Most of the current researches on the running state evaluation of the on-line monitoring device start from the hardware composition principle of the device, the device needs to be stopped to run an off-line test, the operation is troublesome, and the cost is high; if the operation state of the on-line monitoring device of the regional transformer substation needs to be acquired, the operability is poor. Therefore, the method for evaluating the state of the online monitoring device based on a data-driven mode is provided.
The data source of the current power equipment state evaluation is mainly the operation and maintenance data of the transformer, and part of research relates to the substation equipment state evaluation based on sub-line monitoring data; but the research for evaluating the running state of the on-line monitoring device is rarely developed; therefore, the important research content of the invention is that the operation states of the online monitoring device and the power transformer are respectively evaluated according to the data sets of the different identified abnormal modes, and a comprehensive state evaluation system comprising the online monitoring device and the power transformer is constructed.
Disclosure of Invention
The state evaluation work can rapidly analyze the running state of the equipment according to the current index state of the equipment, and further obtain the state trend of the equipment in a period of time in the future, so that targeted maintenance work is arranged, and the running reliability of the equipment is improved. However, the current state evaluation work on the power transformation equipment is performed based on historical operation and maintenance data, and the evaluation result reliability is low due to the lack of utilization of index online monitoring data. The invention provides a comprehensive state evaluation method of a power transformation device based on-line monitoring data according to the problem that the current state evaluation work of the power transformation device is single.
The invention is realized by the following technical scheme, and the method for evaluating the comprehensive state of the power transformation equipment comprises the following steps:
s1, importing the state index data processing result, dividing the abnormal data into abnormal data set caused by the abnormal work of the on-line monitoring device
Figure 670300DEST_PATH_IMAGE001
Data set for representing abnormal state of power transformer
Figure 221367DEST_PATH_IMAGE002
；
S2, power transformer state evaluation: data set for representing abnormal state of power transformer
Figure 99324DEST_PATH_IMAGE002
On the basis, classifying the state indexes, processing index data, constructing a layered evaluation system, establishing index weights by using an objective fusion weighting method, and obtaining a state evaluation result of the power transformer according to the index weights;
s3, state evaluation of the online monitoring device: dividing the running state of the on-line monitoring device into three grades of normal, abnormal and fault, and using abnormal data set caused by abnormal working of the on-line monitoring device
Figure 153868DEST_PATH_IMAGE001
On the basis, the state evaluation comprises three steps of constructing a characterization relation between data of an abnormal mode and an abnormal state of the online monitoring device, characterizing the probability of the abnormal state, and finishing the state evaluation of the online monitoring device by using a membership function in a fuzzy evaluation theory.
Specifically, the classification of the status index at step S2: and analyzing and inducing the state indexes used in the state evaluation, and dividing the state indexes into quantitative indexes for representing the state by constant value data and sequence indexes for representing the state by continuous change data according to the index sources and the characteristics of the indexes.
Specifically, the processing of the index data at step S2: for quantitative indexes, processing by using a function according to the current threshold value of each grade; the sequence data is processed based on the proximity.
Specifically, the step S2 is to construct a hierarchical evaluation system: considering that the power transformer is complex in structure and various in components, a three-layer power transformer state evaluation system comprising a target layer, an index source layer and a state index layer is established from the perspective of index data sources.
Specifically, in step S2, an index weight is established by using an objective-subjective fusion weighting method, and a state evaluation result of the power transformer is obtained according to the index weight: calculating the basic weight of each index of each state index layer by using an analytic hierarchy process according to a set multilayer state evaluation system, and performing degradation variable weight operation on each index according to the degradation degree of each index on the basis to complete subjective and objective fusion weighting of each state index; and finally, coupling the processed indexes based on the calculated weight to obtain a state evaluation result of the power transformer.
Specifically, in step S3, the characterizing relationship between the data of the abnormal pattern and the abnormal state of the online monitoring device is constructed as follows: analyzing the relationship between the data of each abnormal mode and the abnormal state representation of the online monitoring device, and constructing the representation relationship between the data of the abnormal mode and the abnormal state of the online monitoring device; the abnormal mode comprises five types of interruption data, repeated data, jumping data, undersized data and outlier data.
Specifically, the abnormal state probability characterization in step S3: the occupation ratios of different abnormal mode data in the online data set are used for representing different abnormal occurrence probabilities of the device and representing the degree of membership of the device to a certain abnormal state.
Specifically, in step S3, the membership function in the fuzzy evaluation theory is used to complete the state evaluation of the online monitoring device: and (3) using a membership function quantification device in the fuzzy evaluation theory to quantify the membership degree of different state grades, and completing the state evaluation of the online monitoring device according to the maximum membership principle.
Specifically, the quantitative index processing procedure is as follows: carrying out uniform dimensionalization on the quantitative indexes to adapt to a state evaluation model of the power transformation equipment; respectively constructing quantitative index processing functions suitable for two conditions of negative degradation and positive degradation;
for the negative degradation indicator data, it is processed using the function shown in the formula (1),
Figure 990237DEST_PATH_IMAGE003
（1）
for the positive degradation index data, the index data is processed by using a function shown in formula (2),
Figure 497441DEST_PATH_IMAGE004
（2）
Figure 303723DEST_PATH_IMAGE006
an evaluation score representing the conversion of the index,
Figure 353719DEST_PATH_IMAGE007
the value of the state index is represented,
Figure 220044DEST_PATH_IMAGE008
is an attention value characterizing the upper limit of the fault index characteristic quantity.
Specifically, the sequence index processing procedure is as follows:
the method integrates related index data under the fault, and according to common faults of the power transformer, the method is divided into the following six faults: low-temperature overheating, medium-temperature overheating, high-temperature overheating, partial discharge, low-energy discharge and high-energy discharge; monitoring the index with DGA
Figure 89911DEST_PATH_IMAGE009
、
Figure 699884DEST_PATH_IMAGE010
、
Figure 604386DEST_PATH_IMAGE011
、
Figure 641612DEST_PATH_IMAGE012
、
Figure 123409DEST_PATH_IMAGE013
The five gas concentrations are representative, and the DGA historical data of all faults form a fault data set
Figure 412439DEST_PATH_IMAGE014
；
Calculating data of sampling points and fault data set in DGA index data set
Figure 30502DEST_PATH_IMAGE014
The specific calculation mode of the proximity between the adjacent pixels is shown as formula (3),
Figure 379575DEST_PATH_IMAGE015
（3）
in the formula (I), the compound is shown in the specification,
Figure 348668DEST_PATH_IMAGE016
for DGA index historical data points
Figure 175810DEST_PATH_IMAGE017
And fault data set
Figure 913958DEST_PATH_IMAGE018
The degree of close proximity of (c) to each other,
Figure 292987DEST_PATH_IMAGE019
are data points
Figure 624743DEST_PATH_IMAGE017
And
Figure 380209DEST_PATH_IMAGE018
in
Figure 113810DEST_PATH_IMAGE020
The set of the nearest neighbors of the group,
Figure 929319DEST_PATH_IMAGE021
is the nearest neighbor in the failure data set,
Figure 482791DEST_PATH_IMAGE022
expressed is the euclidean distance in the index data,
Figure 776369DEST_PATH_IMAGE020
refers to a failure data set
Figure 364477DEST_PATH_IMAGE018
Selected subjectively from
Figure 85308DEST_PATH_IMAGE020
A data point;
for fault data set
Figure 516289DEST_PATH_IMAGE018
Computing its internal point relative to the fault data set
Figure 488925DEST_PATH_IMAGE018
The relative neighbor degree of the self-body is calculated, the average value of the neighbor degree is obtained, and a fault data set is obtained
Figure 790593DEST_PATH_IMAGE018
Reference proximity of
Figure 823271DEST_PATH_IMAGE023
；
Failure data set of all data points
Figure 741548DEST_PATH_IMAGE018
The relative neighbor degree is obtained by comparing the neighbor degree with the reference neighbor degree, the expression of the relative neighbor degree is shown as a formula (4),
Figure 252295DEST_PATH_IMAGE024
（4）
finding fault data set existing in DGA monitoring index sequence data
Figure 674049DEST_PATH_IMAGE018
The data points with the maximum relative proximity are used as the basis for intercepting the magnitude of the sequence data; when the relative proximity is higher, the more serious the state of the equipment is, the larger the sequence data quantity required for analyzing the state of the equipment is, and on the contrary, the smaller the intercepted sequence data quantity is; truncated sequence data length
Figure 2262DEST_PATH_IMAGE025
The relation between the relative closeness and the degree of the relative closeness is shown in the formula (5),
Figure 17623DEST_PATH_IMAGE026
（5）
in the formula
Figure 722274DEST_PATH_IMAGE027
Scale length of the analyzed DGA online data;
transformation treatment of sequence data; according to the five gas indexes indicated above, the attention value of the gas is extracted from the current guide rule
Figure 873901DEST_PATH_IMAGE028
Calculating a difference between the index and the attention value for each point in the truncated sequence data, and multiplying a coefficient by a side ratio between the difference and the corresponding attention value
Figure 373015DEST_PATH_IMAGE029
Obtaining scores of various gas indexes
Figure 875672DEST_PATH_IMAGE030
The specific transformation mode is shown as a formula (6),
Figure 587276DEST_PATH_IMAGE031
（6）
in the formula (I), the compound is shown in the specification,iit indicates the kind of the gas or gases,ian integer having a value in the range of 1 to 5,ja sampling point representing an index of the gas,x ji indicates that the gas index isjThe value of the point(s) is,i 1 representing the starting gas sequence points of the analysis,L 1 refers to the length of the gas sequence analyzed.
In step S2, the subjective and objective combination weighting method refers to subjective weighting by the analytic hierarchy process and objective weighting by the index degradation degree, and includes the following steps:
(1) subjective weight setting
When the state of the power transformer is evaluated by using an analytic hierarchy process, according to a layered evaluation system obtained by the previous analysis, combining with expert experience to reasonably give the weight of each scheme in different layers on the basis, judging the advantages and disadvantages among different schemes and sequencing, and taking the basic weight of each index as the evaluation scale of the state of the power transformer;
(2) objective weight setting
In the aspect of objective weight giving, a degradation variable weight theory is introduced, the weight of indexes which are seriously deviated from a normal value is increased, and the index information of partial or local degradation is embodied in the overall evaluation result of the transformer; the degradation is divided into two conditions of forward degradation and reverse degradation, the degradation degree of each index is calculated, the calculation mode of the forward degradation degree is shown as a formula (7), the calculation mode of the reverse degradation degree is shown as a formula (8),
Figure 249201DEST_PATH_IMAGE032
（7）
Figure 794583DEST_PATH_IMAGE033
（8）
in the formula (I), the compound is shown in the specification,z p to indicate the degree of degradation under positive degradation,z n the degradation degree is indicated under negative degradation,
Figure 643591DEST_PATH_IMAGE034
in order to monitor the value of the state quantity,
Figure 299831DEST_PATH_IMAGE035
is a starting index value of the state quantity,
Figure 816263DEST_PATH_IMAGE036
is the value of attention for the indicator,
Figure 391601DEST_PATH_IMAGE037
as the degree of deterioration of the index, ycA negative degradation indicator attention value; coupling with the basic weight according to the deterioration degree of each index to obtain the weight after weight change, wherein the weight change formula is shown as a formula (9),
Figure 880568DEST_PATH_IMAGE038
（9）
in the formula (I), the compound is shown in the specification,
Figure 465133DEST_PATH_IMAGE039
is an index
Figure 711438DEST_PATH_IMAGE041
The weight of the changed weight of (a),
Figure 457677DEST_PATH_IMAGE042
is the basis weight for the index,
Figure 156643DEST_PATH_IMAGE043
is referred to as
Figure 544899DEST_PATH_IMAGE043
The item index is a function of the number of items,
Figure 645710DEST_PATH_IMAGE044
for the number of the index categories, the index number,
Figure 562851DEST_PATH_IMAGE045
for equalizing the coefficients, when the degree of deterioration of the index is not important
Figure 139325DEST_PATH_IMAGE046
On the contrary
Figure 675480DEST_PATH_IMAGE047
。
Specifically, in step S3, the step of completing the state evaluation of the online monitoring device by using the membership function in the fuzzy evaluation theory is as follows:
1) for five data exception modes, respectively
Figure 755432DEST_PATH_IMAGE048
Represents; order on-line monitoring device with monthly as time unit
Figure 249998DEST_PATH_IMAGE049
Index (I)
Figure 48190DEST_PATH_IMAGE050
The amount of data collected is
Figure 247090DEST_PATH_IMAGE051
Respectively statistically satisfy
Figure 322493DEST_PATH_IMAGE048
The data of the features has a ratio of
Figure 112595DEST_PATH_IMAGE052
Indicating the probability of each type of abnormality;
2) definition comment set
Figure 742290DEST_PATH_IMAGE053
Respectively indicating normal, abnormal and fault, setting weight coefficient set of index by using subjective weighting method
Figure 276040DEST_PATH_IMAGE054
;
3) Taking the ratio of different abnormal mode data as independent variable, using triangular membership function to establish the quantization function between the index and the state comment, the expression of the membership function is shown as the following formula,
Figure 205950DEST_PATH_IMAGE055
is an interval threshold of the membership function,
Figure 901373DEST_PATH_IMAGE056
（10）
Figure 408578DEST_PATH_IMAGE057
（11）
Figure 355805DEST_PATH_IMAGE058
（12）
4) the proportion of each type of abnormal data is substituted into the formulas (10) - (12) to obtainEvaluation matrix to fuzzy state
Figure 264856DEST_PATH_IMAGE060
,
Figure 272126DEST_PATH_IMAGE061
（13）
In the formula, the elements in the matrix
Figure 1047DEST_PATH_IMAGE062
Is shown as
Figure 751966DEST_PATH_IMAGE063
Index pair
Figure 781102DEST_PATH_IMAGE064
Membership of status comments;
5) calculating the state evaluation result of the device, multiplying the weight set determined in the foregoing by the fuzzy state evaluation matrix to obtain the model evaluation result,
Figure 552749DEST_PATH_IMAGE065
（14）
according to the maximum membership principle, taking
Figure 175491DEST_PATH_IMAGE066
The state corresponding to the maximum value of the values is used as the evaluation result.
According to the method, based on the result of online data processing, state evaluation models are respectively established for the power transformer and the online monitoring device, so that a data-driven comprehensive state evaluation system of the power transformation equipment is formed. For the power transformer, the state indexes of the transformer are divided into two types of quantitative indexes and sequence indexes, different processing models are respectively established aiming at the quantitative indexes and the sequence indexes, and the problem that online monitoring data are difficult to utilize in a traditional power transformer evaluation model is solved; and establishing a transformer layered state evaluation system based on the processing result of the index data and considering different sources of the index data, and realizing the state evaluation of the transformer by using an objective and subjective combination weighting mode. For the online monitoring device, according to the processing result of the data, the representation between different abnormal states of the device and abnormal data is established, the proportion of the abnormal data is used for representing the abnormal probability of the device, and finally, the state evaluation of the online monitoring device is completed by using a fuzzy state evaluation theory, so that the comprehensive state evaluation of the power transformation equipment is realized.
Drawings
FIG. 1 is a flow chart of the present invention.
Fig. 2 is a diagram of a negative deterioration quantification index processing function.
Fig. 3 is a diagram of a positive deterioration quantitative index processing function.
Fig. 4 is a power transformer layered state evaluation system.
Detailed Description
The present invention will be explained in further detail with reference to examples.
As shown in fig. 1, a method for evaluating the comprehensive state of a power transformation device includes the following steps:
s1, importing the state index data processing result, dividing the abnormal data into abnormal data set caused by the abnormal work of the on-line monitoring device
Figure 589155DEST_PATH_IMAGE067
Data set for representing abnormal state of power transformer
Figure 82584DEST_PATH_IMAGE069
: in the data cleaning stage, abnormal data in the online data are mined, different abnormal mode data existing in the online data of the power transformation equipment are separated, the abnormal mode data are divided into two types according to the types of the abnormal data, and the two types of the abnormal mode data are respectively abnormal data sets caused by the abnormal work of the online monitoring device
Figure 290712DEST_PATH_IMAGE067
Data set for representing abnormal state of power transformer
Figure 135171DEST_PATH_IMAGE069
. The index data is used as the basis of the state evaluation work, and the reliability of the state evaluation work can be increased through the cleaned data. Therefore, the comprehensive state evaluation system of the power transformation equipment is established on the basis that the index online data is subjected to cleaning treatment, and the cleaning result divides the online data into abnormal data sets caused by the abnormal work of the online monitoring device
Figure 86946DEST_PATH_IMAGE067
Data set for representing abnormal state of power transformer
Figure 825095DEST_PATH_IMAGE069
。
S2, power transformer state evaluation: data set for representing abnormal state of power transformer
Figure 345069DEST_PATH_IMAGE069
And on the basis, classifying the state indexes, processing index data, constructing a layered evaluation system, establishing index weights by using an objective fusion weighting method, and obtaining a state evaluation result of the power transformer according to the index weights. The state evaluation work of the invention is considered to be carried out by fusing operation and maintenance test data and online monitoring data, and the online data has strong data flow characteristics. The method is different from the traditional method that only data at a certain moment is used for evaluation, index online data is introduced, the online data characteristics are processed according to the online data characteristics to adapt to state evaluation work, the reliability of the equipment state evaluation work is improved, and the continuous running state of the equipment within a certain period of time is evaluated.
S2-1, classification of state indexes: the invention analyzes the state index used in the inductive state evaluation, and divides the index into a quantitative index representing the state by constant value data and a sequence index representing the state by continuous change data according to the index source and the characteristics.
S2-2, index data processing: different types of indexes have different characteristics, and different processing modes are established for the two types of indexes respectively so as to be suitable for the indexesAnd (5) state evaluation work. For quantitative indexes, processing by using a function according to the current threshold value of each grade; the invention provides a method for processing sequence data based on the proximity. In the state evaluation work, index data processing is generally required to be converted into a scoring form so as to adapt to evaluation; for quantitative indexes from operation and maintenance tests, the method takes the positive degradation and negative degradation characteristics of the indexes into consideration, and respectively uses different functions for processing; for sequence indexes from online monitoring, the invention counts historical fault information and establishes a fault data set of equipment
Figure 801458DEST_PATH_IMAGE070
Extracting a data set
Figure 432291DEST_PATH_IMAGE069
The sequence index is converted into a scoring form by calculating the closeness of the data points in the data base.
S2-3, constructing a layered evaluation system: considering that the power transformer is complex in structure and various in components, the invention establishes a three-layer power transformer state evaluation system comprising a target layer, an index source layer and a state index layer from the viewpoint of index data sources.
S2-4, establishing index weight by using an objective fusion weighting method and obtaining a state evaluation result of the power transformer according to the index weight: calculating the basic weight of each index of each state index layer by using an analytic hierarchy process according to a set multilayer state evaluation system, and performing degradation variable weight operation on each index according to the degradation degree of each index on the basis to complete subjective and objective fusion weighting of each state index; and finally, coupling the processed indexes based on the calculated weight to obtain a state evaluation result of the power transformer. For the constructed multilayer state evaluation system of the power transformer, the basic weight of a state index layer is calculated by using an AHP method; on the basis, considering a plurality of indexes of the transformer, the actual operation state of the equipment is difficult to be reflected only by the basic weight, therefore, the invention carries out degradation weight changing processing on the basic weight based on the degradation degree of the indexes so as to highlight the influence of the degradation indexes on the operation state of the equipment.
S3, state evaluation of the online monitoring device: dividing the running state of the on-line monitoring device into three grades of normal, abnormal and fault, and using abnormal data set caused by abnormal working of the on-line monitoring device
Figure 24946DEST_PATH_IMAGE071
On the basis, the state evaluation comprises three steps of constructing a characterization relation between data of an abnormal mode and an abnormal state of the online monitoring device, characterizing the probability of the abnormal state, and finishing the state evaluation of the online monitoring device by using a membership function in a fuzzy evaluation theory.
S3-1, constructing a characterization relation between the data of the abnormal mode and the abnormal state of the online monitoring device: the analysis result of the online monitoring data shows that the online data set contains abnormal data of various different modes, the invention analyzes the relationship between the data of each abnormal mode and the abnormal state representation of the online monitoring device, and constructs the representation relationship between the data of the abnormal mode and the abnormal state of the online monitoring device. The method respectively counts the characterization relations between the abnormal states of the device and the five abnormalities of the interrupt data, the repeated data, the jump data, the undersized data and the outlier data, and takes the five abnormalities as the indexes for evaluating the state of the device.
S3-2, abnormal state probability characterization: the proportion of various abnormal mode data reflects the probability of abnormal fault of the device from the side surface, and the proportion of different abnormal mode data in an online data set is used for representing the different abnormal probability of the device and simultaneously representing the degree of membership of the device to a certain abnormal state.
S3-3, using membership function in fuzzy evaluation theory to complete the state evaluation of the on-line monitoring device: considering the ambiguity among different state grades of the state, the feasibility of dividing the state of the device by using a threshold value alone is low, so the method uses the membership degree function quantification device in the fuzzy evaluation theory to quantify the membership degree of the device to different state grades, and completes the state evaluation of the online monitoring device according to the maximum membership degree principle.
The specific steps of processing different types of index data in step S2-2 are as follows:
(1) quantitative index
1) The quantitative index is also a numerical index, which means that a specific index numerical value is obtained by measuring actual state quantity, the data of the index has quantitative representation on the quality of the state, but the quantitative index is uniformly scaled by considering the magnitude of the measured data of different indexes and the relation between the measured data of different indexes and the representation of the state degradation direction, so as to adapt to a state evaluation model of the power transformation equipment.
2) In order to adapt to the actual condition of the fault characteristic quantity of the power transformation equipment, the invention constructs quantitative index processing functions which are suitable for the conditions of negative degradation and positive degradation respectively, as shown in figures 2 and 3.
As for the index having a higher measurement value, that is, the negative degradation index data, the function is processed using the function shown in the formula (1), and the function curve is shown in the left side of fig. 2.
Figure 715822DEST_PATH_IMAGE072
（1）
For the index with better measurement value, namely the positive degradation index data, the index data is processed by using the function shown in the formula (2), and the function curve is shown on the right of fig. 3.
Figure 393928DEST_PATH_IMAGE073
（2）
Figure 687506DEST_PATH_IMAGE074
An evaluation score representing the conversion of the index,
Figure 275613DEST_PATH_IMAGE007
the value of the state index is represented,
Figure 262024DEST_PATH_IMAGE075
attention value for characterizing upper limit of fault index characteristic quantityTaking the positive degradation index processing function of the formula (2) as an example, when the closer the index measurement is to the attention value, the more serious the state of the index is, the lower the index score is, and the failure or abnormal state operation is likely to occur; when the index quantity is measured to be close to
Figure 833951DEST_PATH_IMAGE076
And when the index is in a serious deviation state, the index state score is close to or 0, and the transformer equipment needs to be immediately subjected to related test maintenance.
(2) Sequence index
1) The sequence index refers to state data converted from online monitoring data, and the index is usually formed by a series of index values and reflects the state condition of the operation index in a period of time. The type data has strong data flow characteristics, so that the continuity of the data needs to be considered when the type data is processed, and the closer the sampling time is, the more obvious the reflecting effect on the state is; the invention provides a method based on the proximity degree to process online sequence data;
2) the method integrates related index data under the fault, and can be divided into the following six faults according to common faults of the power transformer: low-temperature overheating, medium-temperature overheating, high-temperature overheating, partial discharge, low-energy discharge and high-energy discharge; monitoring the index with DGA
Figure 665640DEST_PATH_IMAGE077
、
Figure 108254DEST_PATH_IMAGE078
、
Figure 265566DEST_PATH_IMAGE079
、
Figure 528051DEST_PATH_IMAGE080
、
Figure 429011DEST_PATH_IMAGE081
The five gas concentrations are representative, and the DGA historical data of all faults constitutes a faultData set
Figure 991711DEST_PATH_IMAGE082
；
3) Calculating data of sampling points and fault data set in DGA index data set
Figure 54345DEST_PATH_IMAGE082
The specific calculation mode of the proximity between the adjacent pixels is shown as formula (3),
Figure 194339DEST_PATH_IMAGE083
（3）
in the formula (I), the compound is shown in the specification,
Figure 508777DEST_PATH_IMAGE016
for DGA index historical data points
Figure 785037DEST_PATH_IMAGE017
And fault data set
Figure 159518DEST_PATH_IMAGE018
The degree of close proximity of (c) to each other,
Figure 786808DEST_PATH_IMAGE084
are data points
Figure 904937DEST_PATH_IMAGE017
And
Figure 301283DEST_PATH_IMAGE018
in
Figure 971299DEST_PATH_IMAGE020
The set of the nearest neighbors of the group,
Figure 695673DEST_PATH_IMAGE021
is the nearest neighbor in the failure data set,
Figure 476547DEST_PATH_IMAGE022
expressed is the European expression in the index dataThe distance between the first and second electrodes,
Figure 868345DEST_PATH_IMAGE020
refers to a failure data set
Figure 709262DEST_PATH_IMAGE018
Selected subjectively from
Figure 779986DEST_PATH_IMAGE020
A data point;
4) for fault data set
Figure 239918DEST_PATH_IMAGE018
Computing its internal point relative to the fault data set
Figure 610856DEST_PATH_IMAGE018
The relative neighbor degree of the self-body is calculated, the average value of the neighbor degree is obtained, and a fault data set is obtained
Figure 232461DEST_PATH_IMAGE018
Reference proximity of
Figure 321640DEST_PATH_IMAGE023
；
5) Failure data set of all data points
Figure 319683DEST_PATH_IMAGE018
The relative neighbor degree is obtained by comparing the neighbor degree with the reference neighbor degree, the expression of the relative neighbor degree is shown as a formula (4),
Figure 545128DEST_PATH_IMAGE085
（4）
6) finding fault data set existing in DGA monitoring index sequence data
Figure 727848DEST_PATH_IMAGE018
The data points with the maximum relative proximity are used as the basis for intercepting the magnitude of the sequence data;when the relative proximity is higher, the more serious the state of the equipment is, the larger the sequence data quantity required for analyzing the state of the equipment is, and on the contrary, the smaller the intercepted sequence data quantity is; truncated sequence data length
Figure 914110DEST_PATH_IMAGE086
The relation between the relative closeness and the degree of the relative closeness is shown in the formula (5),
Figure 574898DEST_PATH_IMAGE087
（5）
in the formula
Figure 784076DEST_PATH_IMAGE089
Scale length of the analyzed DGA online data;
7) transformation treatment of sequence data; according to the five gas indexes indicated above, the attention value of the gas is extracted from the current guide rule
Figure 137697DEST_PATH_IMAGE090
Calculating a difference between the index and the attention value for each point in the truncated sequence data, and multiplying a coefficient by a side ratio between the difference and the corresponding attention value
Figure 811255DEST_PATH_IMAGE091
Obtaining scores of various gas indexes
Figure 275734DEST_PATH_IMAGE092
The specific transformation mode is shown as a formula (6),
Figure 475772DEST_PATH_IMAGE093
（6）
in the formula (I), the compound is shown in the specification,iit indicates the kind of the gas or gases,ian integer having a value in the range of 1 to 5,ja sampling point representing an index of the gas,x ji indicates that the gas index isjThe value of the point(s) is,i 1 representing the starting gas sequence points of the analysis,L 1 refers to the length of the gas sequence analyzed.
The layered state evaluation system provided in step S2-3 of the present invention is specifically a three-layer power transformer comprehensive state evaluation system including a target layer, a data source layer, and a state index layer. The evaluation result of the overall running state of the transformer is a target layer and is also a final target of state evaluation; in the data source layer, setting three categories of an electrical test, an insulating oil test and online monitoring according to different data sources for evaluation; and finally, an index layer reflecting the running state of the transformer, wherein the index layer is composed of state indexes contained in corresponding data sources. Fig. 4 shows a system for evaluating the overall state of the power transformer.
The subjective and objective combination weighting method in step S2-4 of the present invention refers to subjective weighting by the analytic hierarchy process and objective weighting that varies according to the degree of index degradation. The method mainly comprises the following steps:
(1) subjective weight setting
Analytic Hierarchy Process (AHP) is a subjective method of empowerment that organically combines expert experience with mathematics. When the AHP is used for state evaluation of the power transformer, the weight of each scheme in different layers is reasonably given according to the layered evaluation system obtained by the previous analysis and combining with expert experience on the basis, the advantages and disadvantages among different schemes are judged and ranked, and the basic weight of each index is used as the evaluation scale of the state of the power transformer.
According to the operation experience of the existing power transformer, the existing judgment matrix setting criteria are combined, pairwise comparison is carried out on each index in the matrix, the importance degrees in the matrix are respectively recorded as 1-9, and the judgment matrix importance degrees are shown in table 1.
TABLE 1 judgment matrix construction basis
Figure 610081DEST_PATH_IMAGE094
(2) Objective weight setting
Due to the fact that the transformer is complex in structure and large in state quantity, overhaul tests and online monitoring data are independent, when certain index of the transformer is degraded or has local faults, the index is often reflected in the change of a small amount of indexes, if the index is still considered based on a comprehensive integral weight setting method, after basic weight calculation, the final evaluation result of the transformer is still normal, and the real influence of the degradation index on the overall state of equipment cannot be highlighted.
Therefore, in the aspect of objective weight assignment, the invention proposes to introduce a degradation variable weight theory, increase the weight of the index which is seriously deviated from the normal value, and reflect the index information of partial or local degradation in the overall evaluation result of the transformer.
As is known from the foregoing analysis, the degradation is divided into two cases, namely, forward degradation and reverse degradation, and the first step of using the degradation weight varying theory is to calculate the degradation degree of each index. The positive degradation degree is calculated according to equation (7), the negative degradation degree is calculated according to equation (8),
Figure 895569DEST_PATH_IMAGE095
（7）
Figure 304684DEST_PATH_IMAGE096
（8）
in the formula (I), the compound is shown in the specification,z p to indicate the degree of degradation under positive degradation,z n the degradation degree is indicated under negative degradation,
Figure 359228DEST_PATH_IMAGE034
in order to monitor the value of the state quantity,
Figure 789072DEST_PATH_IMAGE035
is a starting index value of the state quantity,
Figure 437223DEST_PATH_IMAGE036
is the value of attention for the indicator,
Figure 509084DEST_PATH_IMAGE037
as the degree of deterioration of the index, ycA negative degradation indicator attention value; coupling with the basic weight according to the deterioration degree of each index to obtain the weight after weight change, wherein the weight change formula is shown as a formula (9),
Figure 27921DEST_PATH_IMAGE097
（9）
in the formula (I), the compound is shown in the specification,
Figure 159825DEST_PATH_IMAGE098
is an index
Figure 29692DEST_PATH_IMAGE099
The weight of the changed weight of (a),
Figure 905244DEST_PATH_IMAGE100
is the basis weight for the index,
Figure 403222DEST_PATH_IMAGE043
is referred to as
Figure 581393DEST_PATH_IMAGE043
The item index is a function of the number of items,
Figure 63190DEST_PATH_IMAGE102
for the number of the index categories, the index number,
Figure 352220DEST_PATH_IMAGE103
for equalizing the coefficients, when the degree of deterioration of the index is not important
Figure 970283DEST_PATH_IMAGE104
On the contrary
Figure 319356DEST_PATH_IMAGE105
。
The step S3-1 of constructing a characterization relationship between the data of the abnormal mode and the abnormal state of the online monitoring device is specifically as follows:
data interruption: the abnormality is usually represented as data loss, the abnormality duration is calculated according to the day, and the abnormality duration is mainly characterized by communication interruption caused by complex operation conditions.
Data repetition: such anomalies are typically negative or extreme values if they are caused by a communication interruption, or a random value within the valid range, typically other than 0, if the sensor fails.
Data jumping: the monitored values of the equipment change in stages, resulting in sudden increases or decreases in the monitored values, usually for a period of time.
The data is too small: generally, due to the fact that the sensitivity of a sensor of the monitoring equipment is reduced, correct feedback on the index of a normal level cannot be achieved.
Outlier data: also called singular points, refer to numerical values with large deviation from the statistical mean, and the change of the operation condition is the main reason for the abnormality.
In the step S3-3, the membership function in the fuzzy evaluation theory is used for completing the state evaluation of the online monitoring device, and the method specifically comprises the following steps:
1) for the five data abnormal patterns analyzed in the preamble, the method comprises
Figure 288449DEST_PATH_IMAGE106
Represents; order on-line monitoring device with monthly as time unit
Figure 240225DEST_PATH_IMAGE107
Index (I)
Figure 853740DEST_PATH_IMAGE108
The amount of data collected is
Figure 498348DEST_PATH_IMAGE109
Respectively statistically satisfy
Figure 564524DEST_PATH_IMAGE110
The data of the features has a ratio of
Figure 319990DEST_PATH_IMAGE111
Is shown byThe probability of each type of anomaly occurring.
2) Definition comment set
Figure 53591DEST_PATH_IMAGE112
Respectively indicating normal, abnormal and fault, setting weight coefficient set of index by using subjective weighting method
Figure 603521DEST_PATH_IMAGE113
。
3) Taking the ratio of different abnormal mode data as independent variable, using triangular membership function to establish the quantization function between the index and the state comment, the expression of the membership function is shown as the following formula,
Figure 281627DEST_PATH_IMAGE114
is an interval threshold of the membership function,
Figure 716151DEST_PATH_IMAGE115
（10）
Figure 428892DEST_PATH_IMAGE116
（11）
Figure 25089DEST_PATH_IMAGE117
（12）
4) substituting the proportion of each type of abnormal data into the formulas (10) - (12) to obtain a fuzzy state evaluation matrix
Figure 190491DEST_PATH_IMAGE118
，
Figure 163126DEST_PATH_IMAGE119
（13）
In the formula, the elements in the matrix
Figure 730374DEST_PATH_IMAGE120
Is shown as
Figure 887686DEST_PATH_IMAGE121
Index pair
Figure 415750DEST_PATH_IMAGE122
Membership of status comments.
5) Calculating the state evaluation result of the device, multiplying the weight set determined in the foregoing by the fuzzy state evaluation matrix to obtain the model evaluation result,
Figure 316710DEST_PATH_IMAGE123
（14）
according to the maximum membership principle, taking
Figure 613831DEST_PATH_IMAGE124
The state corresponding to the maximum value of the values is used as the evaluation result.
Application case
A method for evaluating the comprehensive state of a power transformation device comprises the following steps:
1. the effectiveness of the comprehensive state evaluation method for the 220kV oil-immersed power transformation equipment (SSZ 11-180000/220) is verified, the equipment is put into operation at 2011 for 6 months, and comprehensive evaluation is carried out by combining online monitoring data and historical test data of the equipment.
2. Performing state evaluation on the power transformer, taking 90-quarter group data as a detection object for DGA online indexes, calling the latest offline test data when the data processing result judges that data reflecting the abnormal state of equipment exists, and performing the state evaluation on the transformer by using the method; table 2 shows the data related to the test indexes for a certain time.
TABLE 2 Transformer test data
Figure 942044DEST_PATH_IMAGE125
3. Determining the basic weight of each index by using AHP methodThe random consistent index of each layer is checked to have
Figure 957404DEST_PATH_IMAGE126
And proving that the set judgment matrix meets the requirements, and calculating subjective evaluation weights of all indexes, wherein the subjective evaluation weights are shown in a table 3.
TABLE 3 basic weightings of evaluation indexes of transformers
Figure 130897DEST_PATH_IMAGE127
After data processing, DGA index abnormal values are found in online monitoring data of 7 months and 19 days, and relevant index data are shown in a table 4.
TABLE 4 Online DGA data for transformer 7 months and 19 days
Figure 672736DEST_PATH_IMAGE128
The detection shows that the degree of adjacency of the day-changing DGA data and the fault data set is the maximum and is 0.64%; therefore, according to the formula (5), with the day-changing as a starting point, DGA sequence index data of 14 dates are respectively extracted forwards and backwards and are evaluated and differentiated; meanwhile, scoring treatment is carried out on the corresponding quantitative indexes to obtain index scoring and differentiation results shown in the table 5.
TABLE 5 grading of related indexes for transformer evaluation
Figure 47217DEST_PATH_IMAGE129
The deterioration degrees of all the indexes are calculated according to the expressions (7) to (8), and the deterioration weight change operation is performed on the indexes by using the expression (9) based on the basic weight, and the deterioration degrees and the weight coefficients of the indexes are shown in table 6.
TABLE 6 degradation degree and Final weight coefficient of evaluation index
Figure 408928DEST_PATH_IMAGE130
According to the result of calculation of the evaluation model, the final state of the power transformer is marked as 63.81, the power transformer is in an abnormal state, and the maintenance work should be scheduled as soon as possible. Through operation and maintenance, the actual conditions of the transformer are as follows: the abnormal change of the main-transformer total hydrocarbon is found in the 7-month 17-day oil chromatographic measurement, and the abnormal change and the synchronization are carried out
Figure 527057DEST_PATH_IMAGE132
And
Figure 923403DEST_PATH_IMAGE133
a growing trend also occurs; the inspection shows that the contact surface of the tap switch is locally oxidized and corroded due to poor contact of the tap switch, so that the contact resistance is increased, the transformer is locally overheated, and after the tap switch is replaced in time, all indexes of the transformer return to normal.
4. And (3) evaluating the state of the online monitoring device: taking the annual index monitoring amount of the equipment as an object, and processing the abnormal data set obtained by data processing
Figure 327840DEST_PATH_IMAGE134
The ratios of different abnormal pattern data are respectively counted and used as the representation of the abnormal probability of the device, and the corresponding probability and weight are obtained and shown in table 7.
TABLE 7 device anomaly probability and weight thereof
Figure 317792DEST_PATH_IMAGE135
5. Calculating a masquerade evaluation matrix of each anomaly pair comment set: the matrix information is shown in table 8.
TABLE 8 fuzzy State evaluation matrix
Figure 364246DEST_PATH_IMAGE136
The transformation is finally obtained by the calculation of the formula (14)The membership vector of the online monitor device to the state comment set is
Figure 224886DEST_PATH_IMAGE137
Namely, the on-line monitoring device of the equipment is in an abnormal state, and a manufacturer is required to be contacted to carry out related maintenance work in time.

Claims (5)

1. A method for evaluating the comprehensive state of a power transformation device is characterized by comprising the following steps:
s1, importing the state index data processing result, dividing the abnormal data into abnormal data set caused by the abnormal work of the on-line monitoring device
Figure 886386DEST_PATH_IMAGE001
Data set for representing abnormal state of power transformer
Figure 801121DEST_PATH_IMAGE002
；
S2, power transformer state evaluation: data set for representing abnormal state of power transformer
Figure 289871DEST_PATH_IMAGE002
On the basis, classifying the state indexes, processing index data, constructing a layered evaluation system, establishing index weights by using an objective fusion weighting method, and obtaining a state evaluation result of the power transformer according to the index weights;
classification of the status indicators: analyzing and inducing the state indexes used in the state evaluation, and dividing the state indexes into quantitative indexes representing the state by fixed value data and sequence indexes representing the state by continuous change data according to the index sources and the characteristics of the indexes;
and (3) processing the index data: for quantitative indexes, processing by using a function according to the current threshold value of each grade; processing the sequence index based on a mode of proximity;
constructing a layered evaluation system: establishing a three-layer power transformer state evaluation system comprising a target layer, an index source layer and a state index layer;
the index weight is formulated by using an objective fusion weighting method, and the state evaluation result of the power transformer is obtained according to the index weight: calculating the basic weight of each index of each state index layer by using an analytic hierarchy process according to a set multilayer state evaluation system, and performing degradation variable weight operation on each index according to the degradation degree of each index on the basis to complete subjective and objective fusion weighting of each state index; finally, coupling the processed indexes based on the calculated weight to obtain a state evaluation result of the power transformer;
s3, state evaluation of the online monitoring device: dividing the running state of the on-line monitoring device into three grades of normal, abnormal and fault, and using abnormal data set caused by abnormal working of the on-line monitoring device
Figure 216763DEST_PATH_IMAGE001
On the basis, the state evaluation comprises three steps of constructing a characterization relation between data of an abnormal mode and an abnormal state of the online monitoring device, characterizing the probability of the abnormal state, and finishing the state evaluation of the online monitoring device by using a membership function in a fuzzy evaluation theory;
the characterization relation between the data of the abnormal mode and the abnormal state of the online monitoring device is established as follows: analyzing the relationship between the data of each abnormal mode and the abnormal state representation of the online monitoring device, and constructing the representation relationship between the data of the abnormal mode and the abnormal state of the online monitoring device; the abnormal mode comprises five types of interrupt data, repeated data, jumping data, undersized data and outlier data;
the abnormal state probability characterizes: the occupation ratios of different abnormal mode data in an online data set are used for representing different abnormal occurrence probabilities of the device and representing the degree of membership of the device to a certain abnormal state;
and the state evaluation of the online monitoring device is completed by using a membership function in a fuzzy evaluation theory: and (3) using a membership function quantification device in the fuzzy evaluation theory to quantify the membership degree of different state grades, and completing the state evaluation of the online monitoring device according to the maximum membership principle.
2. The method for evaluating the comprehensive state of the power transformation equipment as claimed in claim 1, wherein the method comprises the following steps: the quantitative index processing process is as follows: carrying out uniform dimensionalization on the quantitative indexes to adapt to a state evaluation model of the power transformation equipment; respectively constructing quantitative index processing functions suitable for two conditions of negative degradation and positive degradation;
for the negative degradation indicator data, it is processed using the function shown in the formula (1),
Figure 474569DEST_PATH_IMAGE003
（1）
for the positive degradation index data, the index data is processed by using a function shown in formula (2),
Figure 310938DEST_PATH_IMAGE004
（2）
Figure 536252DEST_PATH_IMAGE005
an evaluation score representing the conversion of the index,
Figure 749059DEST_PATH_IMAGE006
the value of the state index is represented,
Figure 595792DEST_PATH_IMAGE007
is an attention value characterizing the upper limit of the fault index characteristic quantity.
3. The method for evaluating the comprehensive state of the power transformation equipment as claimed in claim 1, wherein the sequence index processing process is as follows:
the method integrates related index data under the fault, and according to common faults of the power transformer, the method is divided into the following six faults: low temperature overheating, medium temperature overheating, high temperature overheating, localDischarge, low energy discharge, high energy discharge; monitoring the index with DGA
Figure 180226DEST_PATH_IMAGE008
、
Figure 315672DEST_PATH_IMAGE009
、
Figure 394487DEST_PATH_IMAGE010
、
Figure 548256DEST_PATH_IMAGE011
、
Figure 54324DEST_PATH_IMAGE012
The five gas concentrations are representative, and the DGA historical data of all faults form a fault data set
Figure 946842DEST_PATH_IMAGE013
；
Calculating data of sampling points and fault data set in DGA index data set
Figure 563768DEST_PATH_IMAGE013
The specific calculation mode of the proximity between the adjacent pixels is shown as formula (3),
Figure 588356DEST_PATH_IMAGE014
（3）
in the formula (I), the compound is shown in the specification,
Figure 514592DEST_PATH_IMAGE015
for DGA index historical data points
Figure 624631DEST_PATH_IMAGE016
And fault data set
Figure 779669DEST_PATH_IMAGE013
The degree of close proximity of (c) to each other,
Figure 970347DEST_PATH_IMAGE017
are data points
Figure 755901DEST_PATH_IMAGE016
And
Figure 415552DEST_PATH_IMAGE013
in
Figure 561232DEST_PATH_IMAGE018
The set of the nearest neighbors of the group,
Figure 357149DEST_PATH_IMAGE019
is the nearest neighbor in the failure data set,
Figure 110342DEST_PATH_IMAGE020
expressed is the euclidean distance in the index data,
Figure 447170DEST_PATH_IMAGE018
refers to a failure data set
Figure 209589DEST_PATH_IMAGE013
Selected subjectively from
Figure 312543DEST_PATH_IMAGE018
A data point;
for fault data set
Figure 502216DEST_PATH_IMAGE013
Computing its internal point relative to the fault data set
Figure 808564DEST_PATH_IMAGE013
The relative neighbor degree of the self-body is calculated, the average value of the neighbor degree is obtained, and a fault data set is obtained
Figure 358363DEST_PATH_IMAGE013
Reference proximity of
Figure 332135DEST_PATH_IMAGE021
；
Failure data set of all data points
Figure 692709DEST_PATH_IMAGE013
The relative neighbor degree is obtained by comparing the neighbor degree with the reference neighbor degree, the expression of the relative neighbor degree is shown as a formula (4),
Figure 735620DEST_PATH_IMAGE022
（4）
finding fault data set existing in DGA monitoring index sequence data
Figure 574263DEST_PATH_IMAGE013
The data points with the maximum relative proximity are used as the basis for intercepting the magnitude of the sequence data; when the relative proximity is higher, the more serious the state of the equipment is, the larger the sequence data quantity required for analyzing the state of the equipment is, and on the contrary, the smaller the intercepted sequence data quantity is; truncated sequence data length
Figure 464859DEST_PATH_IMAGE023
The relation between the relative closeness and the degree of the relative closeness is shown in the formula (5),
Figure 180355DEST_PATH_IMAGE024
（5）
in the formula
Figure 523612DEST_PATH_IMAGE025
Scale length of the analyzed DGA online data;
transformation treatment of sequence data; according to the aboveThe five gas indexes are extracted from the current guide rules to obtain the attention values
Figure 352897DEST_PATH_IMAGE026
Calculating a difference between the index and the attention value for each point in the truncated sequence data, and multiplying a coefficient by a side ratio between the difference and the corresponding attention value
Figure 97999DEST_PATH_IMAGE027
Obtaining scores of various gas indexes
Figure 738059DEST_PATH_IMAGE028
The specific transformation mode is shown as a formula (6),
Figure 552300DEST_PATH_IMAGE029
（6）
in the formula (I), the compound is shown in the specification,iit indicates the kind of the gas or gases,ian integer having a value in the range of 1 to 5,ja sampling point representing an index of the gas,x ji indicates that the gas index isjThe value of the point(s) is,i 1 representing the starting gas sequence points of the analysis,L 1 refers to the length of the gas sequence analyzed.
4. The method for evaluating the comprehensive state of the power transformation equipment as claimed in claim 1, wherein the method comprises the following steps: in step S2, the subjective and objective combination weighting method refers to subjective weighting by the analytic hierarchy process and objective weighting by the index degradation degree, and includes the following steps:
subjective weight setting
When the state of the power transformer is evaluated by using an analytic hierarchy process, according to a layered evaluation system obtained by the previous analysis, combining with expert experience to reasonably give the weight of each scheme in different layers on the basis, judging the advantages and disadvantages among different schemes and sequencing, and taking the basic weight of each index as the evaluation scale of the state of the power transformer;
objective weight setting
In the aspect of objective weight giving, a degradation variable weight theory is introduced, the weight of indexes which are seriously deviated from a normal value is increased, and the index information of partial or local degradation is embodied in the overall evaluation result of the transformer; the degradation is divided into two conditions of forward degradation and reverse degradation, the degradation degree of each index is calculated, the calculation mode of the forward degradation degree is shown as a formula (7), the calculation mode of the reverse degradation degree is shown as a formula (8),
Figure 998325DEST_PATH_IMAGE030
（7）
Figure 535616DEST_PATH_IMAGE031
（8）
in the formula (I), the compound is shown in the specification,z p to indicate the degree of degradation under positive degradation,z n the degradation degree is indicated under negative degradation,
Figure 658162DEST_PATH_IMAGE032
in order to monitor the value of the state quantity,
Figure 913694DEST_PATH_IMAGE033
is a starting index value of the state quantity,
Figure 163410DEST_PATH_IMAGE034
is the value of attention for the indicator,
Figure 807405DEST_PATH_IMAGE035
as the degree of deterioration of the index, ycA negative degradation indicator attention value; coupling with the basic weight according to the deterioration degree of each index to obtain the weight after weight change, wherein the weight change formula is shown as a formula (9),
Figure 54847DEST_PATH_IMAGE036
（9）
in the formula (I), the compound is shown in the specification,
Figure 109259DEST_PATH_IMAGE037
is an index
Figure 834770DEST_PATH_IMAGE038
The weight of the changed weight of (a),
Figure 408971DEST_PATH_IMAGE039
is the basis weight for the index,
Figure 607740DEST_PATH_IMAGE038
is referred to as
Figure 837864DEST_PATH_IMAGE038
The item index is a function of the number of items,
Figure 163803DEST_PATH_IMAGE040
for the number of the index categories, the index number,
Figure 779461DEST_PATH_IMAGE041
for equalizing the coefficients, when the degree of deterioration of the index is not important
Figure 431022DEST_PATH_IMAGE042
On the contrary
Figure 129201DEST_PATH_IMAGE043
。
5. The method for evaluating the comprehensive state of the power transformation equipment as claimed in claim 1, wherein the method comprises the following steps: in step S3, the step of completing the state evaluation of the online monitoring device using the membership function in the fuzzy evaluation theory is as follows:
1) for five data exception modes, respectively
Figure 258831DEST_PATH_IMAGE044
Represents; order monitoring device with monthly degrees as time units
Figure 542045DEST_PATH_IMAGE045
Index (I)
Figure 285879DEST_PATH_IMAGE046
The amount of data collected is
Figure 287333DEST_PATH_IMAGE047
Respectively statistically satisfy
Figure 158337DEST_PATH_IMAGE044
The data of the features has a ratio of
Figure 810904DEST_PATH_IMAGE048
Indicating the probability of each type of abnormality;
2) definition comment set
Figure 273110DEST_PATH_IMAGE049
Respectively indicating normal, abnormal and fault, setting weight coefficient set of index by using subjective weighting method
Figure 965122DEST_PATH_IMAGE050
;
3) Taking the ratio of different abnormal mode data as independent variable, using triangular membership function to establish the quantization function between the index and the state comment, the expression of the membership function is shown as the following formula,
Figure 420243DEST_PATH_IMAGE051
is an interval threshold of the membership function,
Figure 615732DEST_PATH_IMAGE052
（10）
Figure 779997DEST_PATH_IMAGE053
（11）
Figure 945924DEST_PATH_IMAGE054
（12）
4) substituting the proportion of each type of abnormal data into the formulas (10) - (12) to obtain a fuzzy state evaluation matrix
Figure 955468DEST_PATH_IMAGE055
,
Figure 67781DEST_PATH_IMAGE056
（13）
In the formula, the elements in the matrix
Figure 324319DEST_PATH_IMAGE057
Is shown as
Figure 787661DEST_PATH_IMAGE058
Index pair
Figure 787847DEST_PATH_IMAGE059
Membership of status comments;
5) calculating the evaluation result of the state of the device, and collecting the weight coefficients
Figure 754666DEST_PATH_IMAGE050
Multiplying the fuzzy state evaluation matrix to obtain a model evaluation result,
Figure 260733DEST_PATH_IMAGE060
（14）
according to the maximum membership principle, taking
Figure 132743DEST_PATH_IMAGE061
The state corresponding to the maximum value of the values is used as the evaluation result.
CN202110539663.0A 2021-05-18 2021-05-18 Comprehensive state evaluation method for power transformation equipment Active CN112966990B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110539663.0A CN112966990B (en) 2021-05-18 2021-05-18 Comprehensive state evaluation method for power transformation equipment

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110539663.0A CN112966990B (en) 2021-05-18 2021-05-18 Comprehensive state evaluation method for power transformation equipment

Publications (2)

Publication Number Publication Date
CN112966990A CN112966990A (en) 2021-06-15
CN112966990B true CN112966990B (en) 2021-09-10

Family

ID=76275602

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110539663.0A Active CN112966990B (en) 2021-05-18 2021-05-18 Comprehensive state evaluation method for power transformation equipment

Country Status (1)

Country Link
CN (1) CN112966990B (en)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113486499A (en) * 2021-06-22 2021-10-08 合肥工业大学 Power transmission and transformation equipment fuzzy fault rate calculation method based on state monitoring information
CN113836488B (en) * 2021-09-09 2023-06-20 苏州热工研究院有限公司 Method and device for online data processing and state evaluation of steam turbine
CN114636776A (en) * 2022-03-21 2022-06-17 南京智鹤电子科技有限公司 Transformer fault prediction method based on monitoring of dissolved gas in transformer oil
CN117251804B (en) * 2023-11-17 2024-04-19 天津中电华利电器科技集团有限公司 Substation operation state monitoring data processing system and method

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105868912A (en) * 2016-04-06 2016-08-17 清华大学 Power transformer state evaluate method and apparatus based on data fusion
CN106651169A (en) * 2016-12-19 2017-05-10 国家电网公司 Fuzzy comprehensive evaluation-based distribution automation terminal state evaluation method and system
CN110689234A (en) * 2019-09-05 2020-01-14 国家电网有限公司 Power transformer state evaluation method based on multi-source data fusion

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107016500A (en) * 2017-03-27 2017-08-04 国家电网公司 Transformer fuzzy synthetic appraisement method based on variable weight
CN109299551A (en) * 2018-09-30 2019-02-01 武汉大学 A kind of Condition Assessment for Power Transformer method
CN110175749A (en) * 2019-04-28 2019-08-27 国网辽宁省电力有限公司电力科学研究院 A kind of running state of transformer appraisal procedure based on PMU data
CN110297141A (en) * 2019-07-01 2019-10-01 武汉大学 Fault Locating Method and system based on multilayer assessment models
CN111062500B (en) * 2019-12-05 2022-06-03 国网电力科学研究院武汉南瑞有限责任公司 Power equipment evaluation method based on discrete fuzzy number and analytic hierarchy process
CN112749882A (en) * 2020-12-28 2021-05-04 广东电网有限责任公司佛山供电局 Transformer state evaluation method based on cloud model and fuzzy evidence reasoning

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105868912A (en) * 2016-04-06 2016-08-17 清华大学 Power transformer state evaluate method and apparatus based on data fusion
CN106651169A (en) * 2016-12-19 2017-05-10 国家电网公司 Fuzzy comprehensive evaluation-based distribution automation terminal state evaluation method and system
CN110689234A (en) * 2019-09-05 2020-01-14 国家电网有限公司 Power transformer state evaluation method based on multi-source data fusion

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
The study for electric power equipment supplier evaluation based on rough set and fuzzy grey incidence cluster analysis;Hongsheng Shan，and etc;《 2011 International Conference on Electrical and Control Engineering》;20111024;第459-462页 *
基于状态评价的风力发电设备故障诊断系统的研制;王奇等;《电网与清洁能源》;20141125(第11期);第103-108页 *
基于集对分析和综合赋权的电力变压器套管绝缘状态评估;李志超等;《高压电器》;20180316(第03期);第115-121页 *

Also Published As

Publication number Publication date
CN112966990A (en) 2021-06-15

Similar Documents

Publication Publication Date Title
CN112966990B (en) Comprehensive state evaluation method for power transformation equipment
CN108375476B (en) Hydroelectric generating set health assessment method
CN109766334B (en) Processing method and system for online monitoring abnormal data of power equipment
CN108320043B (en) Power distribution network equipment state diagnosis and prediction method based on electric power big data
CN107358366B (en) Distribution transformer fault risk monitoring method and system
CN110297141A (en) Fault Locating Method and system based on multilayer assessment models
Dukarm Transformer oil diagnosis using fuzzy logic and neural networks
CN103576050A (en) Operating state assessment method of capacitor voltage transformer
CN106447205A (en) Method for evaluating state of distribution automation terminal based on analytic hierarchy process
CN109490726A (en) Electric power transformer insulated state evaluating method based on Clouds theory
CN105242155A (en) Transformer fault diagnosis method based on entropy weight method and grey correlation analysis
CN103810328A (en) Transformer maintenance decision method based on hybrid model
CN109597396B (en) A kind of distribution transforming on-line fault diagnosis method based on high amount of traffic and transfer learning
CN110646685B (en) Comprehensive evaluation method for state of switch cabinet
CN110782157A (en) Maintenance mode making method based on importance of power generation equipment
CN117113166A (en) Industrial boiler fault detection method based on improved integrated learning
CN115372816A (en) Power distribution switchgear operation fault prediction system and method based on data analysis
CN110033181B (en) Power generation equipment state evaluation method based on self-encoder
CN111311135A (en) Transformer substation energy efficiency assessment method
CN105741184B (en) Transformer state evaluation method and device
CN111898794B (en) Abnormal monitoring method for thermal efficiency of large coal-fired boiler
CN110766248B (en) Workshop artificial factor reliability assessment method based on SHEL and interval intuitionistic fuzzy assessment
CN116862478A (en) Transformer health assessment method based on layered architecture and dynamic weighting
CN116628976A (en) Comprehensive evaluation method for state change of hydraulic turbine unit equipment
CN115587331A (en) Power grid equipment operation state diagnosis and prediction method and system

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