CN111487563B - Transformer state knowledge acquisition method and device based on genetic algorithm and attribute support degree - Google Patents
Transformer state knowledge acquisition method and device based on genetic algorithm and attribute support degree Download PDFInfo
- Publication number
- CN111487563B CN111487563B CN202010411430.8A CN202010411430A CN111487563B CN 111487563 B CN111487563 B CN 111487563B CN 202010411430 A CN202010411430 A CN 202010411430A CN 111487563 B CN111487563 B CN 111487563B
- Authority
- CN
- China
- Prior art keywords
- attribute
- transformer
- decision
- condition
- individual
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 56
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 30
- 230000002068 genetic effect Effects 0.000 title claims abstract description 26
- 230000002159 abnormal effect Effects 0.000 claims abstract description 22
- 238000004364 calculation method Methods 0.000 claims abstract description 12
- 239000004215 Carbon black (E152) Substances 0.000 claims description 24
- 229930195733 hydrocarbon Natural products 0.000 claims description 24
- 150000002430 hydrocarbons Chemical class 0.000 claims description 24
- 230000005856 abnormality Effects 0.000 claims description 16
- 230000009471 action Effects 0.000 claims description 14
- 238000004804 winding Methods 0.000 claims description 13
- BTCSSZJGUNDROE-UHFFFAOYSA-N gamma-aminobutyric acid Chemical compound NCCCC(O)=O BTCSSZJGUNDROE-UHFFFAOYSA-N 0.000 claims description 11
- 238000009413 insulation Methods 0.000 claims description 11
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical group [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 claims description 10
- 230000006870 function Effects 0.000 claims description 10
- 230000008859 change Effects 0.000 claims description 9
- 238000013021 overheating Methods 0.000 claims description 9
- 238000012545 processing Methods 0.000 claims description 8
- 208000024891 symptom Diseases 0.000 claims description 6
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 6
- 239000012774 insulation material Substances 0.000 claims description 3
- 238000004643 material aging Methods 0.000 claims description 3
- 239000002184 metal Substances 0.000 claims description 3
- 229910052751 metal Inorganic materials 0.000 claims description 3
- 230000035772 mutation Effects 0.000 claims description 2
- 238000005457 optimization Methods 0.000 claims description 2
- 238000012423 maintenance Methods 0.000 abstract description 10
- 238000011156 evaluation Methods 0.000 abstract description 8
- 238000011161 development Methods 0.000 abstract description 6
- 230000009286 beneficial effect Effects 0.000 abstract description 4
- 230000007547 defect Effects 0.000 abstract 1
- 238000012544 monitoring process Methods 0.000 abstract 1
- 239000007789 gas Substances 0.000 description 33
- 238000010586 diagram Methods 0.000 description 9
- 238000003745 diagnosis Methods 0.000 description 8
- 238000004590 computer program Methods 0.000 description 7
- 238000001514 detection method Methods 0.000 description 7
- 238000012360 testing method Methods 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 5
- 230000018109 developmental process Effects 0.000 description 5
- 239000000126 substance Substances 0.000 description 5
- 238000007796 conventional method Methods 0.000 description 3
- 230000002349 favourable effect Effects 0.000 description 3
- 230000016507 interphase Effects 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 238000003860 storage Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000010353 genetic engineering Methods 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 239000003990 capacitor Substances 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000007500 overflow downdraw method Methods 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 230000003449 preventive effect Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 150000003839 salts Chemical class 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/50—Testing of electric apparatus, lines, cables or components for short-circuits, continuity, leakage current or incorrect line connections
- G01R31/62—Testing of transformers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- General Engineering & Computer Science (AREA)
- Genetics & Genomics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Physiology (AREA)
- Power Engineering (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a transformer state knowledge acquisition method and equipment based on a genetic algorithm and attribute support. According to the method, the decision table is established through the transformer fault case, the state knowledge under the abnormal information of the specific state of the transformer is solved by combining the genetic algorithm and the attribute support degree, the state evolution rule of the transformer is found, the calculation complexity is greatly reduced compared with that of the traditional method, and the method is beneficial to efficiently extracting the condition attribute which is most relevant to the transformer fault type in the decision table. The method can overcome the defects of the prior art, can be widely used for recognizing the state development rule of the transformer, realizing the state evaluation of the transformer, guiding the state monitoring of the transformer and making a differentiated operation and maintenance strategy, is simple and efficient, and is beneficial to improving the operation safety of the power transformer.
Description
Technical Field
The invention relates to power transformer fault diagnosis and state assessment, in particular to a transformer state knowledge acquisition method and equipment.
Background
In recent years, under the background of energy internet, with the large-scale development of national smart power grids, various state sensing and detection technologies are more and more mature in application, intelligent detection and analysis technologies begin to be completely exposed in various fields of power grids, such as intelligent inspection robots, unmanned inspection helicopters, intelligent analysis and diagnosis systems and the like, and the power grids in China initially have the capability of mastering real-time data such as power equipment operation states, environmental changes, operation and detection information and the like through an informatization means, so that a foundation is provided for carrying out equipment state evaluation, abnormality diagnosis and risk prediction. As one of important devices in an electric power system, a power transformer has been receiving extensive attention for a long time in an efficient and reliable transformer fault diagnosis and state evaluation method. Generally, the state of health of a transformer can be described in terms of electrical, chemical, etc. parameters that can sensitively characterize various aspects of the device. At present, the state evaluation and fault diagnosis of the transformer are roughly carried out from the following three aspects: (1) threshold comparison method: and comparing the measured parameter with a preset standard value, or comparing the parameter change rate with the standard value, and if the measured parameter is higher (or lower) than the preset standard value, judging that the equipment has a fault or reminds the equipment to draw attention. The criteria are largely adopted in corresponding standards and regulations in China. The method has simple form and clear target, has high reasoning efficiency but cannot accurately know the specific fault position and type of the transformer; (2) pattern recognition: for partial detection parameters of the transformer, such as dissolved gas in oil and partial discharge detection, the obtained detection data are not single values, but a group of multiple parameters with complex internal relation, such as the content of multiple gases in oil. For such information characteristics, the fault type can be well identified by an appropriate pattern identification method. However, the dependence of the method on the learning algorithm is large, the algorithm input needs numerous state information, the incidence relation of each state information is unclear, and the influence degree of each state information on the transformer fault cannot be effectively known; (3) and (3) comprehensive evaluation: in order to indicate the position and the severity of the fault as much as possible, the theory of fuzzy mathematics, rough set and the like is introduced, and the information as comprehensive as possible is utilized to obtain a more refined transformer state conclusion. However, the above methods focus only on transformer fault diagnosis and fault cause analysis. In addition, in the existing transformer state evaluation system, most research works only singly score or simply weight results represented by different state information, and extraction and analysis of transformer state knowledge and research of state development rules are less developed.
The following problems exist in the existing transformer fault diagnosis and state evaluation methods: (1) the transformer fault diagnosis model depends on a neural network algorithm, the incidence relation of each state information is unknown, the influence degree of each state information on the transformer fault cannot be effectively known, and the extraction and analysis of the key state information of the transformer are not facilitated; (2) the existing information fusion method only carries out weighting calculation on the multi-dimensional state information of the transformer, cannot find the progressive relation among different abnormal state information and between the abnormal state information and different fault types, and cannot efficiently form transformer state knowledge from fault cases.
Disclosure of Invention
The purpose of the invention is as follows: aiming at the problems in the prior art, the invention provides a transformer state knowledge acquisition method and equipment based on a genetic algorithm and an attribute support degree, which can efficiently acquire transformer state knowledge from a transformer fault case, are simple and efficient, are beneficial to operation and maintenance personnel to know the state evolution rule of the transformer, can provide technical support for differentiated operation and maintenance of the transformer, and improve the operation reliability of the transformer.
The technical scheme is as follows: according to a first aspect of the present invention, there is provided a transformer state knowledge acquisition method based on a genetic algorithm and attribute support, including the following steps:
constructing a decision table S based on a transformer fault case, taking abnormal symptoms of transformer state information as a condition attribute set C, and taking a transformer fault type as a decision attribute set D;
establishing a condition attribute set C 'required by deduction state knowledge according to a decision table S, and calculating an optimal attribute core R comprising the condition attribute set C' according to the condition attribute set C and the decision attribute set D based on a genetic algorithm and an attribute support degree;
and optimizing the decision table S according to the optimal attribute kernel R to form a new decision table S' so as to obtain the transformer state knowledge.
As a preferred embodiment, the transformer state information includes electrical parameters, chemical parameters and relay protection device action conditions for evaluating the transformer state, where the electrical parameters include insulation resistance, direct current resistance, capacitance and dielectric loss, no-load current and no-load loss, short-circuit impedance, partial discharge amount, voltage ratio, iron core grounding current and iron core insulation resistance; the chemical parameters including CO, CO in insulating oil2Rate of change and CO2Ratio of/CO, hydrocarbon gas three ratio (C)2H2/C2H4Content ratio, CH4/H2Content ratio, C2H4/C2H6Content ratio), water content in oil; the action condition of the relay protection device comprises the action condition of a light gas relay and the action condition of a heavy gas relay; a condition attribute set C formed according to the abnormal symptom of the transformer state information is { insulation resistance abnormality, direct current resistance abnormality, capacitance and dielectric loss abnormality, no-load current and no-load loss abnormality, short circuit impedance abnormality, partial discharge quantity abnormality, voltage ratio abnormality, core grounding current and core insulation resistance abnormality, CO in insulating oil2Rate of change and CO2Abnormal ratio of CO, low-temp. overheat (less than or equal to 700 deg.C) of hydrocarbon gas, and high-temp. overheat (>700 ℃), low-energy discharge of hydrocarbon gas with three specific values, high-energy discharge of hydrocarbon gas with three specific values, abnormal water content in oil, action of light gas relay and action of heavy gas relay, { C1,C2,C3,…,C16And obtaining the hydrocarbon gas three-ratio value result according to a gas three-ratio value coding rule table in the national standard GB/T7252. The discretization value of the condition attribute C is {1,2}, and '1' indicates that the corresponding state information in the case is abnormal; "2" indicates that the corresponding state information in the case meets the corresponding condition attribute;
the common fault type of the transformer is taken as a decision attribute set D (normal, winding turn-to-turn short circuit, phase-to-phase short circuit, winding deformation, screen creepage, outgoing line joint fault, coil strand breakage or bare metal overheating fault, iron core fault, insulation material aging and on-load tap-changer fault) (D) }1,D2,D3,D4,D5,D6,D7,D8,D9,D10}；
Establishing a decision table S according to the collected transformer fault cases, wherein the decision table is composed of a plurality of decision rules, each decision rule is formed by processing one transformer fault case, and each decision rule is composed of condition attributes C ═ C1,C2,C3,…,C16Specific values and oneAnd the state information in the transformer fault case forms the specific value of the condition attribute, and the fault type corresponds to the decision attribute.
As a preferred embodiment, let the set of conditional attributes needed to deduce the state knowledge
The calculation of the optimal attribute core R comprising the conditional attribute set C' based on genetic algorithms and attribute support comprises the steps of:
s1, representing each individual code in the genetic algorithm by using a binary string with a length of L ═ card (C) -card (r), where the operator card () represents the number of elements in the computation set, and the code of each position in the string corresponds to each condition attribute in the set L ═ C-C', and when the code is "1", it represents the corresponding position condition attribute in the individual reserved set L, and when the code is "0", it represents that the individual does not reserve the corresponding position condition attribute in the set L; generating initial individuals as initial populations by adopting a borrowing algorithm;
s2, calculating the fitness of each individual according to a fitness function F (r), wherein the fitness function F (r):
wherein lr represents the number of 1 in the individual code, the attribute kernel R 'is composed of a condition attribute set C' and condition attributes of corresponding positions in a reserved set L in the individual code, and γ isR’(D) And expressing the attribute support degree in the following calculation mode:
wherein, PosR’(D) The positive domain of the attribute core R 'and the attribute core R' of the decision attribute set is represented, and card (S) represents the number of decision rules in the decision table;
s3, performing genetic operation on the initial population according to the fitness of the individuals to generate a new generation of individuals, repeatedly calculating the fitness of the new generation of individuals and performing the genetic operation until the fitness is not lower than the iteration number Max, and reserving the optimal attribute kernel R 'in the last generation of individuals as the optimal attribute kernel R, wherein the optimal attribute kernel R' in the last generation of individuals meets the conditions:
as a preferred embodiment, the genetic manipulation comprises:
s3-1, setting a maximum iteration frequency Max, a cross probability pc of a single-point cross method, a variation probability pm of a basic bit variation method, and an iteration frequency f being f + 1;
s3-2, calculating the initial individual fitness F (r), and reserving the optimal individual by a roulette method;
s3-3, calculating all the individual gammaR’(D) Extracting all the individuals satisfying gammaR’(D) 1, to S3-4; if none of the individuals satisfy gammaR’(D) Directly converting to S3-5 when the value is 1;
s3-4, judging whether the iteration frequency f is f +1, and if so, turning to S3-6; if not, go to S3-5;
s3-5, processing the unsatisfied gamma in S3-3 by the single-point crossing method and the basic mutation methodR’(D) The 1-mer forms a new generation of individuals, which also includes all γ -satisfying S3-3R’(D) New generation individuals switched to S3-3 to recalculate all individuals γ for 1 individualR’(D)；
S3-6, extracting the individual with max ((L-lr)/L), wherein the individual codes the condition data of the corresponding position in the reservation set L and the condition attribute set C' form the optimal attribute core R of the algorithm result.
The new decision table S' is formed by only reserving the condition attribute in the optimal attribute core R by each decision rule in the decision table S; the transformer state knowledge consists of two parts, one part is basic state knowledge: the decision rule is composed of a condition attribute set C 'in a new decision table S'; the other part is to expand knowledge: and the decision rule is composed of the condition attributes left after R-C 'in the new decision table S'. The transformer state knowledge logical structure is as follows: if the existing transformer state information can be matched with the condition attribute set C ', the transformer state decision result may be D, and if the matching condition attribute R-C' of the state information is added, the transformer state decision result is determined to be D.
According to a second aspect of the present invention, there is provided a computer apparatus, the apparatus comprising:
one or more processors;
a memory; and
one or more programs, wherein the one or more programs are stored in the memory and configured for execution by the one or more processors, which when executed by the processors implement the steps of the method according to the first aspect of the invention.
Has the advantages that:
the method can be widely used for efficiently and quickly acquiring the transformer state knowledge from the transformer fault case, the calculation complexity is greatly reduced compared with that of the traditional method by adopting the genetic algorithm to solve the optimal attribute kernel, the method is favorable for accurately extracting the key state information of the transformer, the rapid integration and discovery of the transformer state knowledge are realized, the method is simple and efficient, and the method is favorable for operation and maintenance personnel to know the transformer state development process and guide the development of the operation and maintenance work of the transformer.
According to the method, the decision table is established through the transformer fault case, the key information characteristics in the transformer case information are extracted, the decision rule with a unified structure is formed, the transformer case expression can be more efficient, and the support of a data information layer can be provided for transformer state evaluation and knowledge discovery.
According to the method, the transformer state basic knowledge is formed by utilizing a specific condition attribute set, the optimal attribute kernel containing the specific condition attribute is calculated by adopting a genetic algorithm, and the state expansion knowledge is found, so that the state evolution rule of the transformer under a specific abnormal symptom can be found more specifically, the research on the internal evolution relation between the abnormal symptom of the transformer state information and the transformer fault type is facilitated, the theoretical support is provided for realizing the differentiated operation and maintenance of the transformer, the operation and detection efficiency of the transformer is improved, and the operation safety of equipment is improved.
In summary, compared with the traditional method, the method can solve the optimal attribute core of the decision table under specific state information, realizes the hierarchical evolution solution from basic state knowledge to extended knowledge under any state information combination of the transformer, has clear and convenient calculation method, and is beneficial to rapidly summarizing the state knowledge of the transformer and guiding the operation and maintenance work of the transformer.
Drawings
FIG. 1 is a general flow diagram of a method of practicing the present invention.
Detailed Description
The present invention is further illustrated by the following figures and specific examples, which are to be understood as illustrative only and not as limiting the scope of the invention, which is to be given the full breadth of the appended claims and any and all equivalent modifications thereof which may occur to those skilled in the art upon reading the present specification.
Referring to FIG. 1, in one embodiment, the hydrocarbon gas triratio value (C) in the transformer status information is used2H2/C2H4Content ratio, CH4/H2Content ratio, C2H4/C2H6Content ratio) as a condition attribute set required by deducing state knowledge, and realizing the acquisition of the transformer state knowledge according to the following steps.
Step S10, collecting transformer fault cases to construct a decision table S, wherein abnormal symptoms of transformer state information are used as a condition attribute set C, and the transformer fault type is used as a decision attribute set D;
s101, the transformer state information comprises electrical parameters, chemical parameters and action conditions of a relay protection device for evaluating the state of the transformer, wherein the electrical parameters comprise insulation resistance, direct current resistance, capacitance and dielectric loss, no-load current and no-load loss, short-circuit impedance, partial discharge quantity, voltage ratio, iron core grounding current and iron core insulation resistance, the electrical parameters are conventional overhaul test items of the transformer, and the routine overhaul test items can also refer to preventive test regulations DL/T596 of power equipment and handover test standard GB of the power equipmentSelecting the materials at/T50150; the chemical parameters including CO, CO in insulating oil2Rate of change and CO2Ratio of/CO, hydrocarbon gas three ratio (C)2H2/C2H4Content ratio, CH4/H2Content ratio, C2H4/C2H6Content ratio), water content in oil; the action condition of the relay protection device comprises the action condition of a light gas relay and the action condition of a heavy gas relay. The light gas gives an alarm for the gas volume, and the specific alarm value is different according to different equipment and components. The heavy gas is an oil flow speed alarm, and the alarm value is different according to different protection positions.
Conditional attribute set C ═ C formed from abnormal signs of transformer state information1,C2,C3,…,C16The power supply comprises a power supply, a transformer, a capacitor, a medium, a no-load current, a no-load loss, a short-circuit impedance, a local discharge capacity, a voltage ratio, an iron core grounding current, iron core insulation resistance, and CO, CO in insulating oil2Rate of change and CO2Abnormal ratio of/CO (one of the three values meets the requirement), low-temperature overheating of hydrocarbon gas (less than or equal to 700 ℃), and high-temperature overheating of hydrocarbon gas (>700 ℃), hydrocarbon gas three-ratio value is low-energy discharge, hydrocarbon gas three-ratio value is high-energy discharge, water content in oil is abnormal, light gas relay action and heavy gas relay action }, wherein the hydrocarbon gas three-ratio value result is obtained according to a gas three-ratio value coding rule table in the national standard GB/T7252, electrical parameter state abnormity can be obtained through state quantity marked in a fault case table, the low-energy and high-energy discharge standards can refer to the industry standard DL/T722, the water content abnormity values of different devices are different, and formed condition attributes are judged according to actual conditions according to national standards GB/T7600 and GB/T7601 and are shown in table 1. The discretization value of the condition attribute C is {1,2}, and '1' indicates that the corresponding state information in the case is abnormal; "2" indicates that the corresponding status information in the case meets the corresponding condition attribute.
TABLE 1 conditional Attribute set
The common fault type of the transformer is taken as a decision attribute set D (normal, winding turn-to-turn short circuit, phase-to-phase short circuit, winding deformation, screen creepage, outgoing line joint fault, coil strand breakage or bare metal overheating fault, iron core fault, insulation material aging and on-load tap-changer fault) (D) }1,D2,D3,D4,D5,D6,D7,D8,D9,D10And the decision attribute set is shown in table 2.
TABLE 2 decision Attribute set
Establishing a decision table S according to the collected transformer fault cases, wherein the decision table is composed of a plurality of decision rules, each decision rule is formed by processing one transformer fault case, and each decision rule is composed of condition attributes C ═ C1,C2,C3,…,C16The specific value and a corresponding decision attribute, the state information in the transformer fault case forms the specific value of the condition attribute, and the fault type corresponds to the decision attribute. The decision table is shown in table 3.
TABLE 3 decision Table S
S20, according to the decision table S, hydrocarbon gasThe abnormal signs of the three-ratio value are that the three-ratio value of the hydrocarbon gas is low-temperature overheating (less than or equal to 700 ℃), and the three-ratio value of the hydrocarbon gas is high-temperature overheating (>700 deg.c), the hydrocarbon gas three-ratio value is low-energy discharge, the hydrocarbon gas three-ratio value is high-energy discharge, then the condition attribute set C ═ { C ═ required for state knowledge is deduced10,C11,C12,C13And calculating an optimal attribute core R comprising a condition attribute set C 'based on a genetic algorithm and attribute support, wherein corresponding decision attributes can be determined according to the values of the condition attributes, the number of the condition attributes is generally redundant, and the optimal attribute core here represents a condition attribute set which comprises the condition attribute set C' and can correctly distinguish different decision attributes. The specific calculation steps are as follows:
s201, representing each individual code in the genetic algorithm by using a binary string having a length of L ═ card (C) -card (r) -12, and a code correspondence set L ═ C-C ═ C in each position in the string1,C2,C3,C4,C5,C6,C7,C8,C9,C14,C15,C16When the code of each condition attribute in the set L is '1', the condition attribute of the corresponding position in the individual reserved set L is shown, and when the code of each condition attribute in the set L is '0', the condition attribute of the corresponding position in the individual not reserved set L is shown; generating initial individuals (50 individuals) as an initial population by adopting a borrowing algorithm; the individual code range is {000000000000} - {111111111111}, the 1 st individual code {111111111111}, borrow sequentially, the 2 nd individual {111111111110}, the 3 rd individual {111111111101} …, the 14 th individual {111111111100}, the 15 th individual {111111111010} …, the 25 th individual {111111111000}, the 26 th individual {111111110100} …, the 50 th individual {101111110000 }.
S202, calculating the fitness of each individual according to a fitness function F (r), wherein the fitness function F (r):
wherein lr represents the number of 1 in the individual code, and the attribute kernel R 'is composed of a condition attribute set C' and the individual codeReserving conditional attribute composition of corresponding position in set L, attribute support degree gammaR’(D)：
Wherein, PosR’(D) The positive domain of the attribute core R 'and the attribute core R' of the decision attribute set is represented, and card (S) represents the number of decision rules in the decision table; card (Pos)R’(D) For the present example, the number of decision table rows that can be resolved by using different values of the condition attribute in the attribute core R' (which can be directly obtained in the table) is represented, and the calculation method of the positive domain is not described herein again; the attribute support degree is obtained by dividing the number of decision rules completely distinguished by the condition attribute in the attribute core R' by the total number of decision rules, and indicates that the attribute core can distinguish different decision attributes, and when the attribute support degree is equal to 1, indicates that different decision attributes can be completely distinguished by using different condition attribute values in the attribute core.
In this example, l is 12, card(s) is 10;
s203, performing genetic operation on the initial population according to the fitness of the individuals to generate a new generation of individuals, repeatedly calculating the fitness of the new generation of individuals and performing the genetic operation until the fitness is not lower than the iteration number Max, and reserving the optimal attribute kernel R 'in the last generation of individuals as the optimal attribute kernel R, wherein the optimal attribute kernel R' in the last generation of individuals meets the conditions:
the genetic manipulation comprises:
s203-1, setting the maximum iteration frequency Max to be 150, setting the cross probability pc of the single-point cross method to be 0.8, setting the variation probability pm of the basic bit variation method to be 0.02, and setting the iteration frequency f to be f + 1;
s203-2, calculating the initial individual fitness F (r), and reserving the optimal individual by a roulette method;
s203-3, calculating all the individual gammaR’(D) Extracting all the individuals satisfying gammaR’(D) 1, to S203-4; if none of the individuals satisfy gammaR’(D) Directly converting to S203-5 when the value is 1;
s203-4, judging whether the iteration frequency f is f +1, and if so, turning to S203-6, wherein the maximum iteration frequency Max is 150; if not, turning to S203-5;
s203-5, the single-point crossing method and the basic bit variation method process that the gamma is not satisfied in the S203-3R’(D) The 1-mer forms a new generation of individuals, which also includes all γ -satisfying individuals of S203-3R’(D) 1, or a pharmaceutically acceptable salt thereof. New generation individuals are converted into S203-3 to recalculate all individuals gammaR’(D)；
S203-6, extracting an individual with max ((L-lr/L)), wherein the condition data and the condition attribute set C' of the corresponding position in the individual coding seed reservation set L form an optimal attribute kernel R of the algorithm result;
the optimal attribute kernel calculated in this embodiment is R ═ { C ═ C10,C11,C12,C13,C4,C9}。
And S30, optimizing the decision table S according to the optimal attribute kernel R to form a new decision table S', and acquiring the transformer state knowledge. The formed new decision table S' is formed by only reserving the condition attribute in the optimal attribute core R by each decision rule in the decision table S;
TABLE 4 New decision Table S'
The transformer state knowledge consists of two parts, one part is basic state knowledge: the condition attribute set C in the new decision table S' ═ C10,C11,C12,C13The decision rule is composed of the condition attributes of the gray part in the table 4; another part is expansionKnowledge: from the new decision table S' R-C ═ { C4,C9And (4) forming a decision rule formed by the residual condition attributes. The transformer state knowledge logical structure is as follows: if the existing transformer state information can be matched with the condition attribute set C ', the transformer state decision result may be D, and if the matching condition attribute R-C' of the state information is added, the transformer state decision result is determined to be D.
In this embodiment, C is also required, for example, if the hydrocarbon gas is in a high energy discharge (knowledge of the underlying state)4(abnormal no-load current and no-load loss), C9(CO, CO in insulating oil2Rate of change and CO2abnormal/CO ratio) can be distinguished2(winding turn-to-turn short circuit), D3(interphase short circuit), D4(deformation of winding) D6(outlet connection failure).
On the basis of the transformer fault mechanism, when interphase short circuit and winding deformation fault occur, large-area damage can be formed on the winding paper insulation, and CO in the insulating oil are caused2Rate of change and CO2the/CO ratio is abnormal, and in contrast, the winding turn-to-turn short circuit and the joint fault of the outgoing line are not easy to cause large-area damage of the insulating paper. In addition, the transformer no-load test mainly evaluates the turn-to-turn insulation condition of a transformer winding, whether the iron core has deformation or is out of tolerance in size to cause loss increase and the like, so that no-load current and loss abnormity can be caused by turn-to-turn short circuit and interphase short circuit of the winding.
Therefore, when the abnormal high-energy discharge of hydrocarbon gas occurs in the transformer state information, further attention needs to be paid to no-load test data (whether the no-load current and the no-load loss are abnormal), CO in the insulating oil2Rate of change and CO2And whether the/CO ratio is abnormal or not is judged, so that the state of the transformer is accurately determined.
In conclusion, the method can guide the subsequent state parameters of the transformer to be concerned under the condition of paying attention to the specific state basic knowledge, evolve the state development rule of the transformer and guide the differentiated operation and maintenance work of the transformer.
Compared with conventional methods in which traversal is usedCalculating an optimal attribute kernel by using an algorithm, wherein if n condition attributes exist, the number of times of iterative calculation is 2nThe final result can be obtained, and the complexity of calculation increases exponentially with the increase of condition attributes. In addition, in the conventional method for solving the optimal attribute core of the decision table by adopting a genetic algorithm, optimization calculation based on specific condition attributes is not considered. If the operation and maintenance personnel expect to quickly extract the extended knowledge required by distinguishing the fault type of the transformer from the specific abnormal state of the transformer, compared with the conventional method, the method can respond more quickly and efficiently, and is favorable for carrying out differential operation and maintenance work of the transformer.
Based on the same technical concept as the method embodiment, according to another embodiment of the present invention, there is provided a computer apparatus including: one or more processors; a memory; and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, which when executed by the processors implement the steps in the method embodiments.
As will be appreciated by one skilled in the art, 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 present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solutions of the present invention and not for limiting the same, and although the present invention is described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: modifications and equivalents may be made to the embodiments of the invention without departing from the spirit and scope of the invention, which is to be covered by the claims.
Claims (5)
1. A transformer state knowledge acquisition method based on a genetic algorithm and attribute support is characterized by comprising the following steps:
constructing a decision table S according to the transformer fault case, taking abnormal symptoms of the transformer state information as a condition attribute set C, and taking the transformer fault type as a decision attribute set D;
establishing a condition attribute set C 'required by deduction state knowledge according to a decision table S, and calculating an optimal attribute kernel R comprising the condition attribute set C' based on a genetic algorithm and an attribute support degree according to the condition attribute set C and a decision attribute set D, wherein the method comprises the following steps of:
s1, representing each individual code in the genetic algorithm by using a binary string with a length of L ═ card (C) -card (r), where the operator card () represents the number of elements in the computation set, and the code of each position in the string corresponds to each condition attribute in the set L ═ C-C', and when the code is "1", it represents the corresponding position condition attribute in the individual reserved set L, and when the code is "0", it represents that the individual does not reserve the corresponding position condition attribute in the set L; generating initial individuals as initial populations by adopting a borrowing algorithm;
s2, calculating the fitness of each individual according to a fitness function F (r), wherein the fitness function F (r) is as follows:
wherein lr represents the number of 1 in the individual code, the attribute kernel R 'is composed of a condition attribute set C' and condition attributes of corresponding positions in a reserved set L in the individual code, and γ isR’(D) And expressing the attribute support degree in the following calculation mode:
wherein, PosR’(D) The positive domain of the attribute core R 'and the attribute core R' of the decision attribute set is represented, and card (S) represents the number of decision rules in the decision table;
s3, performing genetic operation on the initial population according to the fitness of the individuals to generate a new generation of individuals, repeatedly calculating the fitness of the new generation of individuals and performing the genetic operation until the fitness is not lower than the iteration number Max, and reserving the optimal attribute kernel R 'in the last generation of individuals as the optimal attribute kernel R, wherein the optimal attribute kernel R' in the last generation of individuals meets the conditions:
wherein the step S3 includes:
s3-1, setting a maximum iteration frequency Max, a cross probability pc of a single-point cross method, a variation probability pm of a basic bit variation method, and an iteration frequency f being f + 1;
s3-2, calculating the initial individual fitness F (r), and reserving the optimal individual by a roulette method;
s3-3, calculating all the individual gammaR’(D) Extracting all the individuals satisfying gammaR’(D) 1, to S3-4; if none of the individuals satisfy gammaR’(D) Directly converting to S3-5 when the value is 1;
s3-4, judging whether the iteration frequency f is f +1, and if so, turning to S3-6; if not, go to S3-5;
s3-5, processing the unsatisfied gamma in S3-3 by the single-point crossing method and the basic mutation methodR’(D) The 1-mer forms a new generation of individuals, which also includes all γ -satisfying S3-3R’(D) New generation individuals switched to S3-3 to recalculate all individuals γ for 1 individualR’(D)；
S3-6, extracting an individual with max ((L-lr)/L), wherein the condition data of the corresponding position in the individual encoding seed reservation set L and the condition attribute set C' form an optimal attribute core R of the algorithm result;
forming a new decision table S 'according to the optimal attribute core R optimization decision table S to obtain transformer state knowledge, wherein the new decision table S' is formed by only reserving condition attributes in the optimal attribute core R by each decision rule in the decision table S, and the transformer state knowledge comprises the following steps: basic state knowledge, which is composed of decision rules composed of condition attribute sets C 'in a new decision table S'; and expanding knowledge, which is composed of decision rules composed of condition attributes remaining after R-C 'in the new decision table S'.
2. The method according to claim 1, wherein the conditional attribute set C ═ insulation resistance anomaly, dc resistance anomaly, capacitance and dielectric constant (rc), is selected from the group consisting of a set of parameters consisting of a genetic algorithm, a set of parameters consisting of a set of parameters, and a set of parametersMass loss abnormality, no-load current and no-load loss abnormality, short circuit impedance abnormality, partial discharge amount abnormality, voltage ratio abnormality, core grounding current and core insulation resistance abnormality, and CO, CO in insulating oil2Rate of change and CO2Abnormal ratio of/CO, low-temperature overheating of hydrocarbon gas with three ratios, high-temperature overheating of hydrocarbon gas with three ratios, low-energy discharge of hydrocarbon gas with three ratios, high-energy discharge of hydrocarbon gas with three ratios, abnormal water content in oil, action of light gas relay and action of heavy gas relay, namely { C1,C2,C3,…,C16Wherein the hydrocarbon gas triratio value comprises C2H2/C2H4Content ratio, CH4/H2Content ratio, C2H4/C2H6The content ratio;
the decision attribute set D is { normal, winding turn-to-turn short circuit, phase-to-phase short circuit, winding deformation, screen creepage, outgoing line joint fault, coil strand breakage or bare metal overheating fault, iron core fault, insulation material aging, on-load tap-changer fault } D1,D2,D3,D4,D5,D6,D7,D8,D9,D10}；
The decision table S is composed of a plurality of decision rules, and each decision rule is composed of condition attributes C ═ C1,C2,C3,…,C16The value of which is composed of a corresponding decision attribute.
3. The method for acquiring the state knowledge of the transformer based on the genetic algorithm and the attribute support degree according to claim 2, wherein the discretization value of the condition attribute C is {1,2}, and '1' indicates that no abnormality exists in the corresponding state information in the case; "2" represents the description that the corresponding state information in the case meets the corresponding condition attribute.
4. The method for acquiring transformer state knowledge based on genetic algorithm and attribute support according to claim 1, wherein the logic structure of the transformer state knowledge is as follows: if the existing transformer state information can be matched with the condition attribute set C ', the transformer state decision result may be D, and if the matching condition attribute R-C' of the state information is added, the transformer state decision result is determined to be D.
5. A computer device, the device comprising:
one or more processors;
a memory; and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the programs when executed by the processors implementing the steps of the method of any of claims 1-4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010411430.8A CN111487563B (en) | 2020-05-15 | 2020-05-15 | Transformer state knowledge acquisition method and device based on genetic algorithm and attribute support degree |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010411430.8A CN111487563B (en) | 2020-05-15 | 2020-05-15 | Transformer state knowledge acquisition method and device based on genetic algorithm and attribute support degree |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111487563A CN111487563A (en) | 2020-08-04 |
| CN111487563B true CN111487563B (en) | 2022-02-15 |
Family
ID=71813311
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010411430.8A Active CN111487563B (en) | 2020-05-15 | 2020-05-15 | Transformer state knowledge acquisition method and device based on genetic algorithm and attribute support degree |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111487563B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112557811B (en) * | 2020-11-19 | 2024-01-12 | 安徽理工大学 | Distributed power supply-containing power distribution network fault location based on improved genetic algorithm |
| CN114252807B (en) * | 2021-12-24 | 2023-07-14 | 国网湖北省电力有限公司经济技术研究院 | Transformer life prediction method based on life differentiation phenomenon |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2011017674A (en) * | 2009-07-10 | 2011-01-27 | Tokyo Denki Univ | System and program for estimation of electrical apparatus operation state |
| CN102879677A (en) * | 2012-09-24 | 2013-01-16 | 西北工业大学 | Intelligent fault diagnosis method based on rough Bayesian network classifier |
| CN104298873A (en) * | 2014-10-10 | 2015-01-21 | 浙江大学 | Attribute reduction method and mental state assessment method on the basis of genetic algorithm and rough set |
| CN104360194A (en) * | 2014-11-17 | 2015-02-18 | 国网河南省电力公司 | Fault diagnosis method for smart power grid |
| CN104765810A (en) * | 2015-04-02 | 2015-07-08 | 西安电子科技大学 | Diagnosis and treating rules mining method based on Boolean matrix |
| CN105675802A (en) * | 2014-11-19 | 2016-06-15 | 国网河南省电力公司南阳供电公司 | Transformer fault diagnosis method |
| CN106066432A (en) * | 2016-05-26 | 2016-11-02 | 国网江苏省电力公司电力科学研究院 | A kind of fault detection and fault diagnosis integrated system of power transformer |
| CN106405319A (en) * | 2015-07-30 | 2017-02-15 | 南京理工大学 | Rough set electric power system fault diagnosis method based on heuristic information |
| CN107679368A (en) * | 2017-09-11 | 2018-02-09 | 宁夏医科大学 | PET/CT high dimensional feature level systems of selection based on genetic algorithm and varied precision rough set |
| CN109872249A (en) * | 2019-01-16 | 2019-06-11 | 中国电力科学研究院有限公司 | A kind of method and system based on Bayesian network and genetic algorithm evaluation intelligent electric energy meter operating status |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| FR2908212B1 (en) * | 2006-11-03 | 2008-12-26 | Alcatel Sa | APPLICATIONS FOR THE PROFILING OF TELECOMMUNICATIONS SERVICE USERS |
-
2020
- 2020-05-15 CN CN202010411430.8A patent/CN111487563B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2011017674A (en) * | 2009-07-10 | 2011-01-27 | Tokyo Denki Univ | System and program for estimation of electrical apparatus operation state |
| CN102879677A (en) * | 2012-09-24 | 2013-01-16 | 西北工业大学 | Intelligent fault diagnosis method based on rough Bayesian network classifier |
| CN104298873A (en) * | 2014-10-10 | 2015-01-21 | 浙江大学 | Attribute reduction method and mental state assessment method on the basis of genetic algorithm and rough set |
| CN104360194A (en) * | 2014-11-17 | 2015-02-18 | 国网河南省电力公司 | Fault diagnosis method for smart power grid |
| CN105675802A (en) * | 2014-11-19 | 2016-06-15 | 国网河南省电力公司南阳供电公司 | Transformer fault diagnosis method |
| CN104765810A (en) * | 2015-04-02 | 2015-07-08 | 西安电子科技大学 | Diagnosis and treating rules mining method based on Boolean matrix |
| CN106405319A (en) * | 2015-07-30 | 2017-02-15 | 南京理工大学 | Rough set electric power system fault diagnosis method based on heuristic information |
| CN106066432A (en) * | 2016-05-26 | 2016-11-02 | 国网江苏省电力公司电力科学研究院 | A kind of fault detection and fault diagnosis integrated system of power transformer |
| CN107679368A (en) * | 2017-09-11 | 2018-02-09 | 宁夏医科大学 | PET/CT high dimensional feature level systems of selection based on genetic algorithm and varied precision rough set |
| CN109872249A (en) * | 2019-01-16 | 2019-06-11 | 中国电力科学研究院有限公司 | A kind of method and system based on Bayesian network and genetic algorithm evaluation intelligent electric energy meter operating status |
Non-Patent Citations (3)
| Title |
|---|
| Attribute Reduction Based on Genetic Algorithm for the Coevolution of Meteorological Data in the Industrial Internet of Things;Yong Cheng et.al;《Wireless Communications and Mobile Computing》;20191231;第1-8页 * |
| Knowledge Reduction of Evaluation Dataset Based on Genetic Algorithm and Fuzzy Rough Set;Chengxi Dong et.al;《2008 International Conference on Computer Science and Software Engineering》;20081222;第889-892页 * |
| 基于粒子群和变精度粗糙集的属性约简研究;张霞;《中国优秀博硕士学位论文全文数据库(硕士)信息科技辑》;20190615(第06期);第I140-49页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111487563A (en) | 2020-08-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Huang et al. | Dissolved gas analysis of mineral oil for power transformer fault diagnosis using fuzzy logic | |
| CN109978079A (en) | A kind of data cleaning method of improved storehouse noise reduction self-encoding encoder | |
| Velásquez et al. | Principal components analysis and adaptive decision system based on fuzzy logic for power transformer | |
| CN105894177B (en) | Power equipment operation risk analysis and evaluation method based on decision tree algorithm | |
| CN111487563B (en) | Transformer state knowledge acquisition method and device based on genetic algorithm and attribute support degree | |
| Yu et al. | Oil-immersed power transformer internal fault diagnosis research based on probabilistic neural network | |
| CN105675802A (en) | Transformer fault diagnosis method | |
| Zhao et al. | A novel approach to transformer fault diagnosis using IDM and naive credal classifier | |
| CN110689068A (en) | Transformer fault type diagnosis method based on semi-supervised SVM | |
| CN113987294A (en) | CVT (continuously variable transmission) online fault diagnosis method based on genetic optimization GRU (generalized regression Unit) neural network | |
| CN116205226A (en) | Power equipment state evaluation method, device, equipment and readable storage medium | |
| Islam et al. | Power transformer health condition evaluation: A deep generative model aided intelligent framework | |
| Esmaeili Nezhad et al. | A review of the applications of machine learning in the condition monitoring of transformers | |
| CN110596490A (en) | Intelligent detection method for railway turnout fault | |
| CN116611003A (en) | Transformer fault diagnosis method, device and medium | |
| Patil et al. | DGA Based Ensemble learning and Random Forest Models for Condition Assessment of Transformers | |
| Zakaria et al. | Optimized Artificial Neural Network for the detection of incipient faults in power transformer | |
| CN114626433A (en) | Fault prediction and classification method, device and system for intelligent electric energy meter | |
| Huang | Fault identification of power transformers using genetic-based wavelet networks | |
| Wang et al. | A fault diagnosis method for transformer integrating rough set with fuzzy rules | |
| Alawode et al. | Fuzzy logic and decision tree models for dissolved gas analysis in power transformer fault diagnosis | |
| Wang et al. | Power Transformer Fault Prediction Based on Knowledge Graphs | |
| Abdo et al. | A novel method in classifying and analysing faults in renewable energy power transformers | |
| Strachan et al. | Practical applications of data mining in plant monitoring and diagnostics | |
| Abuanwar et al. | Recurrent Neural Networks RNNs and Decision Tree DT Machine Learning-Based Approaches For Transmission System Faults Diagnosis |
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 |