CN113240133A - Relay protection equipment familial defect identification method based on artificial intelligence - Google Patents
Relay protection equipment familial defect identification method based on artificial intelligence Download PDFInfo
- Publication number
- CN113240133A CN113240133A CN202110434938.4A CN202110434938A CN113240133A CN 113240133 A CN113240133 A CN 113240133A CN 202110434938 A CN202110434938 A CN 202110434938A CN 113240133 A CN113240133 A CN 113240133A
- Authority
- CN
- China
- Prior art keywords
- defect
- familial
- relay protection
- protection equipment
- equipment
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 230000007547 defect Effects 0.000 title claims abstract description 342
- 238000000034 method Methods 0.000 title claims abstract description 51
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 29
- 238000012216 screening Methods 0.000 claims abstract description 33
- 238000000605 extraction Methods 0.000 claims abstract description 26
- 230000009471 action Effects 0.000 claims abstract description 21
- 238000009826 distribution Methods 0.000 claims abstract description 11
- 238000010606 normalization Methods 0.000 claims abstract description 6
- 238000011426 transformation method Methods 0.000 claims abstract description 3
- 230000002159 abnormal effect Effects 0.000 claims description 35
- 230000006872 improvement Effects 0.000 claims description 28
- 238000011156 evaluation Methods 0.000 claims description 19
- 238000011161 development Methods 0.000 claims description 10
- 230000036541 health Effects 0.000 claims description 10
- 238000004458 analytical method Methods 0.000 claims description 9
- 230000001052 transient effect Effects 0.000 claims description 6
- 238000005065 mining Methods 0.000 claims description 4
- 238000002372 labelling Methods 0.000 claims description 2
- 238000012545 processing Methods 0.000 abstract description 3
- 238000005516 engineering process Methods 0.000 description 7
- 238000012423 maintenance Methods 0.000 description 6
- 238000004519 manufacturing process Methods 0.000 description 4
- 238000011160 research Methods 0.000 description 4
- 230000002950 deficient Effects 0.000 description 3
- 238000003745 diagnosis Methods 0.000 description 3
- 238000007689 inspection Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000007493 shaping process Methods 0.000 description 3
- 230000005856 abnormality Effects 0.000 description 2
- 230000008030 elimination Effects 0.000 description 2
- 238000003379 elimination reaction Methods 0.000 description 2
- 238000007726 management method Methods 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 238000004904 shortening Methods 0.000 description 2
- 238000011425 standardization method Methods 0.000 description 2
- 238000005728 strengthening Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000009412 basement excavation Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000035772 mutation Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
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
- G06Q10/20—Administration of product repair or maintenance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- 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
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Computation (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Primary Health Care (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
A relay protection equipment familial defect identification method based on artificial intelligence belongs to the technical field of relay protection, solves the problem of how to effectively identify the familial defect of running relay protection equipment, obtains defect data by obtaining and arranging the defect information of the relay protection equipment and equipment ledger information, labels the defect data by defect point extraction, information association and data normalization, and analyzes the defect reason and the specific part of the relay protection equipment by combining the defect data of the relay protection equipment; according to the family model of the relay protection equipment, excavating the specific defect distribution condition of the protection equipment, and screening out the family defects of the relay protection equipment; the operation conditions of different familial defect devices in the whole network are combined, the software and hardware risks of incorrect actions are analyzed and protected, a technical transformation method is correspondingly provided, the familial defect identification period of the relay protection device is shortened, the familial defect identification accuracy is improved, the familial defect processing capacity is effectively enhanced, and the operation stability of the power grid is improved.
Description
Technical Field
The invention belongs to the technical field of relay protection, and relates to a relay protection equipment familial defect identification method based on artificial intelligence.
Background
The relay protection equipment has the defect that hidden dangers are buried in the safe operation of a power grid, and particularly, batch equipment defects caused by the same process, the same material, the same design concept and thought and the same software of the same manufacturer can cause great hidden dangers to the safe and stable operation of the power grid. At present, the method mainly aims at identifying the familial defects by means of testing, disintegration analysis and the like of equipment by professionals, and the identification speed is low, the period is long, and the hidden danger equipment runs for a long time.
The rapid development of the artificial intelligence technology provides a good technical basis for solving the problems, the artificial intelligence technology becomes a new research direction for the development of relay protection specialties, and the research of the relay protection state evaluation method and the technology based on the artificial intelligence is imperative by fully utilizing the advantages of incidence relation mining, neural network and the like in data analysis and data mining through the technical means of artificial intelligence deep learning.
The application numbers in the prior art are: the Chinese patent application 'familial defect diagnosis method of a secondary operation and maintenance management system' with 201710402293.X and the publication date of 2017, 9, 19 discloses a familial defect diagnosis method of a secondary equipment intelligent operation and maintenance system, which starts with the frequently-occurring alarms of relay protection secondary equipment, carries out detailed classification on the alarms through a defect management template, and brings the frequently-occurring multiple types of alarms and the frequently-occurring same type of alarms into the familial defect range aiming at the secondary equipment of the same model; although the technical scheme realizes diagnosis of familial defects of the secondary equipment, the health operation state of the secondary equipment can be evaluated in advance, the hidden risk of the secondary equipment of the power system can be found in advance, and a favorable reference value is provided for overhauling of the primary equipment and the secondary equipment of the power system. However, the technical scheme cannot effectively identify the familial defect of the relay protection equipment in operation, so that how to effectively identify the familial defect of the relay protection equipment in operation provides support for key patrol and inspection and key lack elimination of the equipment by an operation and maintenance department, and becomes a problem to be solved urgently.
Disclosure of Invention
The invention aims to effectively identify the familial defects of the relay protection equipment in operation and provide support for key inspection and routing inspection of operation and maintenance departments and key defect elimination of the equipment.
The invention solves the technical problems through the following technical scheme: .
The relay protection equipment familial defect identification method based on artificial intelligence comprises the following steps:
s1, defect digging: acquiring and sorting defect information of the relay protection equipment and equipment standing book information to obtain defect data, labeling the defect data by using defect point extraction, information association and data normalization, associating the type, the running time and the running environment data of the relay protection equipment according to the defect data of the relay protection equipment, and analyzing the defect reasons and specific parts of the relay protection equipment;
s2, familial defect screening: analyzing the number of the same defects according to the characteristics of the family models of the relay protection equipment, excavating the specific defect distribution condition of the protection equipment, and screening out the familial defects of the relay protection equipment;
s3, analyzing the development trend of the equipment health state: the health state and the development trend of the relay protection equipment are analyzed, the running conditions of different familial defect equipment in the whole network are combined, the software and hardware risks of incorrect actions are analyzed and protected, and a technical transformation method is correspondingly provided.
By utilizing the defect data of the relay protection equipment, combining the state evaluation and equipment hidden danger of the relay protection equipment in a normal mode based on artificial intelligence and the research conclusion of the reliability evaluation of the transient state action process of the relay protection in a power grid fault mode, researching the identification of the familial defect of the relay protection and an accurate technological improvement technology, positioning the familial defect from dimensions such as the relay protection equipment of the same model of the same manufacturer, the relay protection equipment of the same type of the same manufacturer and components of the same batch through the description, classification and other information of the equipment defect, shortening the identification period of the familial defect of the relay protection equipment, improving the identification accuracy of the familial defect, further accurately selecting the technological improvement scheme, effectively strengthening the processing capacity of the familial defect and improving the operation stability of the power grid.
As a further improvement of the technical solution of the present invention, the device defect information in step S1 includes: abnormal alarm recording, defect recording, incorrect action recording, early warning abnormal recording and transient analysis abnormal recording.
As a further improvement of the technical solution of the present invention, the device ledger information in step S1 includes: equipment type, equipment model, manufacturer, date of delivery, date of commissioning, voltage class.
As a further improvement of the present invention, the defect data in step S1 includes: defect type, defect point, defect proportion, defect operation environment, defect time type, abnormal time, defect reason, equipment delivery date and equipment commissioning date.
As a further improvement of the technical solution of the present invention, the defect point extraction in step S1 includes: the method for extracting the defect data without the definite defect points comprises the following steps of extracting the definite defect points from defect phenomenon description contents by using a keyword extraction method based on statistical characteristics, wherein the keyword extraction method based on the statistical characteristics specifically comprises the following steps:
s11: defining a professional word bank, a stop word bank, a defect point synonym dictionary and a defect phenomenon synonym dictionary;
s12: inputting defect phenomenon description of a piece of defect data, setting a confidence initial value and a threshold value, removing irrelevant words according to a professional word library and a stop word library, extracting defect points and defect phenomena according to a defect point synonym dictionary and a defect phenomenon synonym dictionary, correspondingly adjusting the confidence value according to an extraction result, finally judging whether the confidence value is greater than the confidence threshold value, judging whether the defect point extraction of the piece of defect data is successful, if the extraction is successful, filling the defect points and defect proportion information, and if not, discarding the piece of defect data.
As a further improvement of the technical solution of the present invention, the step S2 of analyzing the number of the same defects according to the family type characteristics of the relay protection device, mining the distribution of the specific defects of the protection device, and screening the family defects of the relay protection device specifically includes:
s21, finishing the familial defect screening judgment of equipment of the same type and the same manufacturer according to the fact that the equipment of the same type and the same manufacturer have the same logic of software and hardware;
s22, finishing the familial defect screening judgment of the same type of equipment of the same manufacturer according to the logic that the same type of equipment produced by the same manufacturer has the same software and hardware platform;
and S23, statistically analyzing defect data of all relay protection equipment manufacturers according to the fact that the same batch of components supplied by the same relay protection equipment component supplier have the same software and hardware logic, and completing familial defect screening of the same batch of components supplied by the same relay protection equipment component supplier according to the quantity and proportion threshold requirements.
As a further improvement of the technical scheme of the invention, the method for screening and judging the familial defects of the equipment with the same model and the same manufacturer comprises the following steps:
(1) counting and analyzing the defect data of the relay protection equipment of the model of the relay protection manufacturer, and obtaining familial defects according to the quantity requirement;
(2) for unrecoverable abnormal records, forming an abnormal map at a single abnormal moment, searching the number of the records in the whole record according to the map, and obtaining the familial defect corresponding to the unrecoverable abnormal alarm information according to the number requirement;
(3) and integrating the familial defects of the two steps to obtain a familial defect screening result of the relay protection equipment of the type of the manufacturer.
As a further improvement of the technical scheme of the invention, the method for screening and judging the familial defects of the same type of equipment of the same manufacturer comprises the following steps:
(1) counting and analyzing the defect data of the relay protection equipment of the type of the relay protection manufacturer, and obtaining familial defects according to the quantity requirement;
(2) for unrecoverable abnormal records, forming an abnormal map at a single abnormal moment, searching the number of the records in the whole record according to the map, and obtaining the familial defect corresponding to the unrecoverable abnormal alarm information according to a certain number requirement;
(3) and integrating the familial defects of the two steps to obtain a familial defect screening result of the relay protection equipment of the type of the manufacturer.
As a further improvement of the technical solution of the present invention, the method for analyzing the risk of software and hardware protecting incorrect actions described in step S3 specifically includes: judging the improper action risk level of the familial defect by adopting a familial defect improper action risk level evaluation algorithm based on an improved hierarchical clustering method for the familial defect obtained in the familial defect screening step, selecting a technical improvement strategy corresponding to the risk level according to different risk level evaluations, and arranging accurate technical improvement work of the relay protection equipment related to the familial defect.
As a further improvement of the technical scheme of the invention, the familial defect incorrect action risk level assessment algorithm based on the improved hierarchical clustering method specifically comprises the following steps:
s31, standardizing defect sample data; and (3) performing data standardization on each characteristic of the defect sample by adopting a linear standardization method, wherein the formula is as follows:
in the formula: x is the number ofikIs the kth feature of the ith sample, akmin、akminMaximum and minimum values of the kth feature for all samples;
s32, hierarchical clustering based on the improved Euclidean formula; the distance between two objects is calculated using the euclidean formula:
d (i, j) represents an object x composed of m attributesiAnd xjThe distance between them; x is the number ofil，xjlRespectively represent xiAnd xjThe ith attribute value of (1);
euclidean distance has no transitivity, i.e., d (i, k) > t cannot be deduced from d (i, j) > t, d (j, k) > t. Therefore, this method cannot be directly applied; in combination with similarity, an improvement is made to the above method, and a sample consisting of n objects is set as { u }1，u2，…unEach object has m attributes, respectively { a }1，a2，…，anAnd each attribute of the ith object is ui ═ xi1，xi2，…，xin}; then the object uiAnd ujK-th attribute x of (2)ikAnd xjkThe distance between them is:
in the formula: a iskmaxAnd akminRespectively as the kth attribute a of each object of the samplekMaximum and minimum values of;
object uiAnd ujThe comprehensive distance of each attribute is as follows:
object uiAnd ujThe similarity of each attribute is as follows:
s(i,j)＝1-d(i,j)
further obtaining the similarity S between the objects, and intercepting the similarity by using a set threshold value to obtain a corresponding clustering result;
s33, assessing the severity level of the familial defect based on the influence factors; taking the familial defect to be evaluated as a sample, taking various defect characteristics of the familial defect as attributes, and then carrying out familial defect grade evaluation, wherein the evaluation formula is as follows:
in the formula: m is the defect feature number of the family defect; w is aiThe score weight of the ith defect feature is proportional to the importance of the ith defect feature; e.g. of the typeiTaking a characteristic value of the ith defect characteristic;
matching the set threshold value by using the Q value of the influence factor to obtain a preliminary defect grade division R;
s34, hierarchical clustering feedback; and (3) performing secondary judgment on the familial defect equivalent division conclusion by combining the preliminary defect grade division R with the conclusion of the hierarchical clustering algorithm, wherein the judgment formula is as follows:
in the formula, NtrNumber of familial defects belonging to class t and class r, NtNumber of familial defects belonging to the t-th class, k1、k2Is a distribution coefficient;
if the above formula is satisfied, the classification conclusion is considered to be correct; if the defect feature score does not meet the formula, fine-tuning the defect feature score weight within a set range to obtain a weight value meeting the formula; and if all the results cannot be met, the optimal solution of all the conclusion sets is taken.
The invention has the advantages that: by utilizing the defect data of the relay protection equipment, combining the state evaluation and equipment hidden danger of the relay protection equipment in a normal mode based on artificial intelligence and the research conclusion of the reliability evaluation of the transient state action process of the relay protection in a power grid fault mode, researching the identification of the familial defect of the relay protection and an accurate technological improvement technology, positioning the familial defect from dimensions such as the relay protection equipment of the same model of the same manufacturer, the relay protection equipment of the same type of the same manufacturer and components of the same batch through the description, classification and other information of the equipment defect, shortening the identification period of the familial defect of the relay protection equipment, improving the identification accuracy of the familial defect, further accurately selecting the technological improvement scheme, effectively strengthening the processing capacity of the familial defect and improving the operation stability of the power grid.
Drawings
Fig. 1 is a general flowchart of a method for identifying familial defects of an artificial intelligence-based relay protection device according to an embodiment of the present invention;
FIG. 2 is a defect mining flow chart of a relay protection device familial defect identification method based on artificial intelligence according to an embodiment of the present invention;
fig. 3 is a flowchart of a defect point extraction algorithm of the relay protection device familial defect identification method based on artificial intelligence according to the embodiment of the present invention;
fig. 4 is a familial defect screening flowchart of a relay protection device familial defect identification method based on artificial intelligence according to an embodiment of the present invention;
fig. 5 is a flowchart illustrating a precise technique for identifying familial defects of an artificial intelligence-based relay protection device according to an embodiment of the present invention;
fig. 6 is a flowchart of an improved hierarchical clustering method-based familial defect incorrect action risk level assessment algorithm for an algorithm of an artificial intelligence-based relay protection device familial defect identification method according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the embodiments of the present invention, and it is obvious that the described embodiments are some embodiments of the present invention, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The technical scheme of the invention is further described by combining the drawings and the specific embodiments in the specification:
example one
As shown in fig. 1, in the relay protection device familial defect identification method based on artificial intelligence, through information such as description and classification of device defects, suspected familial defects are mined and analyzed, and a corresponding accurate technical improvement strategy is proposed for an analysis and evaluation method of a health state and a development trend of operating devices.
1) And (5) excavating defects. And analyzing the defect reasons and specific parts of the relay protection equipment by combining various types of defect information of the protection equipment and associating data such as equipment types, operation duration, operation environments and the like.
2) And (4) screening familial defects. And analyzing the family models of the relay protection equipment, excavating the specific defect distribution condition of the protection equipment according to the number of the same defects, and screening the family defects of the relay protection equipment.
3) And forming an equipment health state development trend analysis method. And analyzing the health state and the development trend of the equipment, analyzing the principle problem and the software and hardware risks of incorrect actions of relay protection by combining the running condition of the equipment with the defects in the whole network, and correspondingly providing a technical transformation strategy.
The method is characterized in that the familial defect identification and accurate technical improvement technology of the relay protection equipment is researched by utilizing the defect data of the relay protection equipment, and the familial defects are located from dimensions of the relay protection equipment with the same model and the same type of the relay protection equipment and the same batch of components and parts of the same manufacturer through the description, classification and other information of the equipment defects.
1. Defect excavation
As shown in fig. 2, the defect information of the relay protection device is obtained, a plurality of pieces of defect information are collected and sorted, the operation data such as the operation duration, the environment and the like of the corresponding device are associated, and the essential cause and the defect label of the defect are mined.
1) Information acquisition
The initial defect information of the relay protection equipment which needs to be acquired is shown in the following table. Wherein, whether the early warning column is 'NO' indicates that the recording actually occurs, and 'YES' indicates that the recording is obtained through early warning by an artificial intelligence means.
TABLE 1 Equipment Defect information
| Type of information | Whether to give an early warning | Detailed description of the invention |
| Equipment anomaly alarm recording | Whether or not | Software and hardware alarm record of occurrence |
| Monitoring and early warning abnormal record | Is that | Monitoring and early warning moduleEarly warning anomaly record of block prediction |
| Transient analysis anomaly recording | Is that | Action abnormity record judged by transient analysis module |
Secondly, the relevant information of the relay protection equipment ledger needing to be obtained is shown in the following table:
TABLE 2 Equipment ledger and operational information
| Type of information | Detailed description of the invention |
| Equipment standing book | Type of equipment, manufacturer, date of commissioning, etc |
| Operating environment | High temperature, high humidity and other operating environment data |
2) Defect information collation
For the acquired defect information, marking operation is carried out by means of defect point extraction, information association, data normalization and the like, and the defect data after operation at least comprises the following fields:
table 3 details of defect information
As shown in fig. 3, defect point extraction is to extract a definite defect point from the defect phenomenon description content by using a "keyword extraction algorithm based on statistical features" for a defect record without a definite "defect point".
Defining various word libraries
a. Definition professional word library
And defining professional words which are not related to the defect content, such as terms of 'xxx substation', 'xxx interval', 'xxx line protection' and the like.
b. Definition stop word library
Associated words are defined that are independent of the content of the defect, such as words "and", etc.
c. Defining a dictionary of defective point synonyms
For each specific defect point, its synonym dictionary is defined. Synonym dictionaries such as "AD chip" include "AD sampling chip", "sampling chip", and the like.
d. Dictionary for defining synonyms of defect phenomena
For each defect phenomenon of a specific defect point, a synonym dictionary thereof is defined. Such as: synonyms for "incorrect" are "inaccurate", "inconsistent", and the like.
Description of Algorithm
Inputting 'defect phenomenon description' of a defect record, setting an initial confidence coefficient value, and correspondingly adjusting the confidence coefficient according to feature word extraction results of four steps of 'professional word extraction', 'stop word extraction', 'defect point extraction' and 'defect phenomenon extraction'; and finally, judging whether the confidence coefficient is larger than a confidence coefficient threshold value or not to determine whether the defect point extraction of the defect information is successful or not.
And if the extraction is successful, filling the defect points and the defect proportion information. Otherwise, the defect entry is discarded.
3) Standing book information arrangement
For the acquired standing book information, the standing book and the operation data of the equipment to which the defect information belongs are acquired by means of information association and the like:
table 4 standing book information details
| Name of field | Type of field | Detailed description of the invention |
| Type of device | Enumeration | Line protection, main transformer protection and the like |
| Model of the device | Character string | Relay protection model |
| Manufacturer of the product | Enumeration | Relay protection manufacturer |
| Program version | Character string | Relay protection program version |
| Date of manufacture | Long shaping type | Date of manufacture of the apparatus |
| Date of delivery | Long shaping type | Date of commissioning of the device |
| Voltage class | Enumeration | 1000kV, 500kV, 220kV, 110kV, 35kV and the like |
2. Familial defect screening
As shown in fig. 4, threshold judgment analysis is performed on the defect information according to the defect information of the devices of the same manufacturer, the same type, the same model, the same batch, the same defect part, and the like, so as to obtain familial defect information. The method comprises the following three steps:
1) the logic of the same software and hardware is provided according to the same type of equipment produced by the same manufacturer. Finishing the screening and judging of familial defects of equipment with the same model of the same manufacturer by the following two steps:
firstly, counting and analyzing the defect data of the relay protection equipment of the model of the relay protection manufacturer, and obtaining familial defects according to a certain quantity requirement.
And secondly, forming an abnormal map (namely other abnormal alarm information caused by the abnormality) at a single abnormal moment for the unrecoverable abnormal records, searching the number of the records in the whole record according to the map, and obtaining the familial defect corresponding to the unrecoverable abnormal alarm information according to a certain number requirement.
And integrating the familial defects of the two steps to obtain a familial defect screening result of the relay protection equipment of the type of the manufacturer.
2) And the logic of the same software and hardware platform is provided according to the same type of equipment produced by the same manufacturer. Finishing the familial defect screening and judging of the same type of equipment of the same manufacturer by the following two steps:
firstly, counting and analyzing the defect data of the relay protection equipment of the type of the relay protection manufacturer, and obtaining familial defects according to a certain quantity requirement.
And secondly, forming an abnormal map (namely other abnormal alarm information caused by the abnormality) at a single abnormal moment for the unrecoverable abnormal records, searching the number of the records in the whole record according to the map, and obtaining the familial defect corresponding to the unrecoverable abnormal alarm information according to a certain number requirement.
And integrating the familial defects of the two steps to obtain a familial defect screening result of the relay protection equipment of the type of the manufacturer.
If the device is a familial defect of the relay protection equipment, further improving defect information, and establishing detailed information of the relay protection device, wherein the information comprises the following information:
TABLE 5 details of device Defect information
| Name of field | Type of field | Detailed description of the invention |
| Component type | Enumeration | CPU, network port, etc |
| Manufacturer of the product | Character string | Concrete component manufacturer |
| Date of manufacture | Long shaping type | Date of manufacture of components |
3) And the logic of the same software and hardware is provided according to the components of the same batch supplied by the same supplier of the relay protection equipment components. And counting and analyzing defect data of all relay protection equipment manufacturers, and completing the familial defect screening of the components of the same batch supplied by the same relay protection equipment component supplier according to certain quantity and proportion threshold requirements.
3. Equipment health state analysis and accurate technical improvement
As shown in fig. 5, the health state and the development trend of the equipment are analyzed in combination with the familial defect information, the risk of protecting software and hardware of incorrect actions is prompted based on the distribution and the operation condition of the whole network, and an accurate technical improvement strategy is correspondingly proposed.
(1) Familial Defect Attribute
And judging the improper action risk level of the familial defect by adopting an improper action risk level evaluation algorithm of the familial defect based on an improved hierarchical clustering method for the familial defect obtained in the familial defect screening step.
TABLE 6 Defect Attribute for familial Defect
| Defect characteristics | Detailed description of the invention |
| Voltage class | The higher the voltage grade of the defective equipment is, the larger the value is |
| Core level | The core module has a large value. If the CPU fault value is larger than the liquid crystal fault value |
| Extent of influence | The defect value of the cross-interval equipment is larger than that of the single-interval equipment |
| Type of defect | Selecting the component batch faults, the same manufacturer and platform, the same manufacturer and model according to indexes |
| Load rate at defect | When the defect occurs, the higher the equipment load is, the larger the value is |
| Operating environment | The more suitable the operating environment is, the larger the value is |
| Length of normal operation | The shorter the normal operation time before the defect occurs, the larger the value |
| Regional gathering | The wider the distribution of defective equipment, the larger the value |
(2) As shown in FIG. 6, the familial defect incorrect action risk level assessment algorithm based on the improved hierarchical clustering method comprises the following steps:
1) defect sample data normalization
And (3) performing data standardization on each characteristic of the defect sample by adopting a linear standardization method, wherein the formula is as follows:
in the formula: x is the number ofikIs the kth feature of the ith sample, akmin、akminThe maximum and minimum of the kth feature for all samples.
2) Hierarchical clustering algorithm based on improved Euclidean formula
In calculating the distance between two objects, the euclidean formula is typically employed:
d (i, j) represents an object x composed of m attributesiAnd xjThe distance between them; x is the number ofil，xjlRespectively represent xiAnd xjThe ith attribute value of (2).
Euclidean distance has no transitivity, i.e., d (i, k) > t cannot be deduced from d (i, j) > t, d (j, k) > t. Therefore, this method cannot be directly applied.
In combination with similarity, some improvement can be made to the above method, let the sample consisting of n objects be { u }1，u2，…unEach object has m attributes, respectively { a }1，a2，…，anAnd each attribute of the ith object is ui ═ xi1，xi2，…，xin}. Then the object uiAnd ujK-th attribute x of (2)ikAnd xjkThe distance between them is:
in the formula: a iskmaxAnd akminRespectively as the kth attribute a of each object of the samplekMaximum and minimum values of.
Object uiAnd ujThe comprehensive distance of each attribute is as follows:
object uiAnd ujThe similarity of each attribute is as follows:
s(i,j)＝1-d(,j)
and further obtaining the similarity S between the objects, and intercepting the similarity by using a set threshold value to obtain a corresponding clustering result T.
3) Familial defect severity grade evaluation algorithm based on influence factors
Taking the familial defect to be evaluated as a sample, taking various defect characteristics of the familial defect as attributes, and then carrying out familial defect grade evaluation, wherein the evaluation formula is as follows:
in the formula: m is the defect feature number of the family defect; w is aiThe score weight of the ith defect feature is proportional to the importance of the ith defect feature; e.g. of the typeiThe value of the familial defect of the 220kV equipment is larger than that of the familial defect of the 110kV equipment for the characteristic value of the ith defect characteristic, such as the voltage grade defect characteristic.
And matching the set threshold value by using the Q value of the influence factor to obtain a preliminary defect grade division R.
4) And (5) hierarchical clustering feedback. In the preliminary defect classification, due to the influence of factors such as single-feature mutation of the defect and the like, the classification at the critical position is possibly not accurate enough, and the secondary judgment needs to be carried out on the familial defect equivalent classification conclusion by combining the conclusion of the hierarchical clustering algorithm. The judgment formula is as follows:
in the formula, NtrNumber of familial defects belonging to class t and class r, NtNumber of familial defects belonging to the t-th class, k1、k2Is the distribution coefficient.
If the above formula is satisfied, the classification conclusion is considered to be correct; if not, fine tuning the defect feature scoring weight in the set range to obtain the weight value meeting the formula. And if all the results cannot be met, the optimal solution of all the conclusion sets is taken.
The following table is a reference to a precise improvement strategy corresponding to different familial defect risk levels:
TABLE 7 TECHNOLOGY FOR MODIFYING FAMILY DEFECTS AT DIFFERENT RISK LEVELS
| Risk rating | Characteristics of risk | Technical improvement strategy |
| A (high) | Voltage class, core module, failure prone | Priority, active handling of defects |
| B (middle and upper) | Core module, failure prone | Priority, active handling of defects |
| C (middle) | Core module | Active defect handling |
| D (middle and lower) | Broadly distributed, edge assembly | Preferentially cooperate with maintenance work treatment |
| E (Low) | Edge assembly | Is matched with maintenance work treatment |
And further selecting a technical improvement strategy corresponding to the risk grade according to different risk grade evaluations, and arranging accurate technical improvement work of the equipment related to the familial defect.
The above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. The relay protection equipment familial defect identification method based on artificial intelligence is characterized by comprising the following steps:
s1, defect digging: acquiring and sorting defect information of the relay protection equipment and equipment standing book information to obtain defect data, labeling the defect data by using defect point extraction, information association and data normalization, associating the type, the running time and the running environment data of the relay protection equipment according to the defect data of the relay protection equipment, and analyzing the defect reasons and specific parts of the relay protection equipment;
s2, familial defect screening: analyzing the number of the same defects according to the characteristics of the family models of the relay protection equipment, excavating the specific defect distribution condition of the protection equipment, and screening out the familial defects of the relay protection equipment;
s3, analyzing the development trend of the equipment health state: the health state and the development trend of the relay protection equipment are analyzed, the running conditions of different familial defect equipment in the whole network are combined, the software and hardware risks of incorrect actions are analyzed and protected, and a technical transformation method is correspondingly provided.
2. The method for identifying familial defect of artificial intelligence-based relaying protection equipment as claimed in claim 1, wherein said equipment defect information in step S1 comprises: abnormal alarm recording, defect recording, incorrect action recording, early warning abnormal recording and transient analysis abnormal recording.
3. The method for identifying familial defects of artificial intelligence-based relay protection devices of claim 1, wherein the device ledger information in step S1 includes: equipment type, equipment model, manufacturer, date of delivery, date of commissioning, voltage class.
4. The method for identifying familial defect of artificial intelligence-based relay protection equipment according to claim 1, wherein the defect data in step S1 comprises: defect type, defect point, defect proportion, defect operation environment, defect time type, abnormal time, defect reason, equipment delivery date and equipment commissioning date.
5. The method for familial defect identification of artificial intelligence-based relay protection devices according to claim 1, wherein the defect point extraction in step S1 comprises: the method for extracting the defect data without the definite defect points comprises the following steps of extracting the definite defect points from defect phenomenon description contents by using a keyword extraction method based on statistical characteristics, wherein the keyword extraction method based on the statistical characteristics specifically comprises the following steps:
s11: defining a professional word bank, a stop word bank, a defect point synonym dictionary and a defect phenomenon synonym dictionary;
s12: inputting defect phenomenon description of a piece of defect data, setting a confidence initial value and a threshold value, removing irrelevant words according to a professional word library and a stop word library, extracting defect points and defect phenomena according to a defect point synonym dictionary and a defect phenomenon synonym dictionary, correspondingly adjusting the confidence value according to an extraction result, finally judging whether the confidence value is greater than the confidence threshold value, judging whether the defect point extraction of the piece of defect data is successful, if the extraction is successful, filling the defect points and defect proportion information, and if not, discarding the piece of defect data.
6. The method for identifying familial defects of relay protection equipment based on artificial intelligence as claimed in claim 1, wherein the step S2 is performed by analyzing the number of the same defects according to the characteristics of the familial models of relay protection equipment, mining the distribution of the specific defects of the relay protection equipment, and screening the familial defects of the relay protection equipment specifically comprises:
s21, finishing the familial defect screening judgment of equipment of the same type and the same manufacturer according to the fact that the equipment of the same type and the same manufacturer have the same logic of software and hardware;
s22, finishing the familial defect screening judgment of the same type of equipment of the same manufacturer according to the logic that the same type of equipment produced by the same manufacturer has the same software and hardware platform;
and S23, statistically analyzing defect data of all relay protection equipment manufacturers according to the fact that the same batch of components supplied by the same relay protection equipment component supplier have the same software and hardware logic, and completing familial defect screening of the same batch of components supplied by the same relay protection equipment component supplier according to the quantity and proportion threshold requirements.
7. The method for identifying the familial defect of the artificial intelligence-based relay protection equipment according to claim 1, wherein the method for screening and judging the familial defect of the same-model equipment of the same manufacturer comprises the following steps:
(1) counting and analyzing the defect data of the relay protection equipment of the model of the relay protection manufacturer, and obtaining familial defects according to the quantity requirement;
(2) for unrecoverable abnormal records, forming an abnormal map at a single abnormal moment, searching the number of the records in the whole record according to the map, and obtaining the familial defect corresponding to the unrecoverable abnormal alarm information according to the number requirement;
(3) and integrating the familial defects of the two steps to obtain a familial defect screening result of the relay protection equipment of the type of the manufacturer.
8. The method for identifying familial defects of relay protection equipment based on artificial intelligence of claim 1, wherein the method for screening and judging familial defects of equipment of the same manufacturer and type comprises:
(1) counting and analyzing the defect data of the relay protection equipment of the type of the relay protection manufacturer, and obtaining familial defects according to the quantity requirement;
(2) for unrecoverable abnormal records, forming an abnormal map at a single abnormal moment, searching the number of the records in the whole record according to the map, and obtaining the familial defect corresponding to the unrecoverable abnormal alarm information according to a certain number requirement;
(3) and integrating the familial defects of the two steps to obtain a familial defect screening result of the relay protection equipment of the type of the manufacturer.
9. The method for identifying familial defects of artificial intelligence-based relay protection devices according to claim 1, wherein the method for analyzing and protecting software and hardware risks of incorrect actions in step S3 comprises: judging the improper action risk level of the familial defect by adopting a familial defect improper action risk level evaluation algorithm based on an improved hierarchical clustering method for the familial defect obtained in the familial defect screening step, selecting a technical improvement strategy corresponding to the risk level according to different risk level evaluations, and arranging accurate technical improvement work of the relay protection equipment related to the familial defect.
10. The method for identifying familial defects of relay protection equipment based on artificial intelligence according to claim 1, wherein the familial defect incorrect operation risk level assessment algorithm based on the improved hierarchical clustering method specifically comprises:
s31, standardizing defect sample data; the data of each characteristic of the defect sample is normalized by a linear normalization method, and the formula is as follows
In the formula: x is the number ofikIs the kth feature of the ith sample, akmin、akminMaximum and minimum values of the kth feature for all samples;
s32, hierarchical clustering based on the improved Euclidean formula; the distance between two objects is calculated using the euclidean formula:
d (i, j) represents an object x composed of m attributesiAnd xjThe distance between them; x is the number ofil，xjlRespectively represent xiAnd xjThe ith attribute value of (1);
euclidean distance has no transitivity, i.e., d (i, k) > t cannot be deduced from d (i, j) > t, d (j, k) > t. Therefore, this method cannot be directly applied; in combination with similarity, an improvement is made to the above method, and a sample consisting of n objects is set as { u }1，u2，…unEach object has m attributes, respectively { a }1，a2，…，anAnd each attribute of the ith object is ui ═ xi1，xi2，…，xin}; then the object uiAnd ujK-th attribute x of (2)ikAnd xjkThe distance between them is:
in the formula: a iskmaxAnd akminRespectively as the kth attribute a of each object of the samplekMaximum and minimum values of;
object uiAnd ujThe comprehensive distance of each attribute is as follows:
object uiAnd ujThe similarity of each attribute is as follows:
s(i,j)＝1-d(i,j)
further obtaining the similarity S between the objects, and intercepting the similarity by using a set threshold value to obtain a corresponding clustering result;
s33, assessing the severity level of the familial defect based on the influence factors; taking the familial defect to be evaluated as a sample, taking various defect characteristics of the familial defect as attributes, and then carrying out familial defect grade evaluation, wherein the evaluation formula is as follows:
in the formula: m is the defect feature number of the family defect; w is aiThe score weight of the ith defect feature is proportional to the importance of the ith defect feature; e.g. of the typeiTaking a characteristic value of the ith defect characteristic;
matching the set threshold value by using the Q value of the influence factor to obtain a preliminary defect grade division R;
s34, hierarchical clustering feedback; and (3) performing secondary judgment on the familial defect equivalent division conclusion by combining the preliminary defect grade division R with the conclusion of the hierarchical clustering algorithm, wherein the judgment formula is as follows:
in the formula, NtrNumber of familial defects belonging to class t and class r, NtNumber of familial defects belonging to the t-th class, k1、k2Is a distribution coefficient;
if the above formula is satisfied, the classification conclusion is considered to be correct; if the defect feature score does not meet the formula, fine-tuning the defect feature score weight within a set range to obtain a weight value meeting the formula; and if all the results cannot be met, the optimal solution of all the conclusion sets is taken.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110434938.4A CN113240133A (en) | 2021-04-22 | 2021-04-22 | Relay protection equipment familial defect identification method based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110434938.4A CN113240133A (en) | 2021-04-22 | 2021-04-22 | Relay protection equipment familial defect identification method based on artificial intelligence |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113240133A true CN113240133A (en) | 2021-08-10 |
Family
ID=77128843
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110434938.4A Pending CN113240133A (en) | 2021-04-22 | 2021-04-22 | Relay protection equipment familial defect identification method based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113240133A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115033690A (en) * | 2022-05-31 | 2022-09-09 | 国网江苏省电力有限公司信息通信分公司 | Communication defect study and judgment knowledge base construction method, defect identification method and system |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107180267A (en) * | 2017-06-01 | 2017-09-19 | 国家电网公司 | A kind of familial defect diagnostic method of secondary operation management system |
| CN108664538A (en) * | 2017-11-30 | 2018-10-16 | 全球能源互联网研究院有限公司 | A kind of automatic identification method and system of the doubtful familial defect of power transmission and transforming equipment |
| CN112308118A (en) * | 2020-10-15 | 2021-02-02 | 国家电网公司华北分部 | Apriori algorithm-based method and system for analyzing relevance of defect data of relay protection device |
-
2021
- 2021-04-22 CN CN202110434938.4A patent/CN113240133A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107180267A (en) * | 2017-06-01 | 2017-09-19 | 国家电网公司 | A kind of familial defect diagnostic method of secondary operation management system |
| CN108664538A (en) * | 2017-11-30 | 2018-10-16 | 全球能源互联网研究院有限公司 | A kind of automatic identification method and system of the doubtful familial defect of power transmission and transforming equipment |
| CN112308118A (en) * | 2020-10-15 | 2021-02-02 | 国家电网公司华北分部 | Apriori algorithm-based method and system for analyzing relevance of defect data of relay protection device |
Non-Patent Citations (3)
| Title |
|---|
| 徐长宝;应黎明;罗先成;赵立进;高吉普;黄良;谭济;: "基于贝叶斯理论的继电保护设备状态评价权重更新方法", 武汉大学学报(工学版), no. 05 * |
| 王继业;王德林;杨国生;周泽昕;邓春宇;: "大数据技术在继电保护领域的研究与应用", 电力信息与通信技术, no. 12 * |
| 饶威;王凤云;丁坚勇;: "基于改进层次聚类法的电力设备家族缺陷评估", 浙江电力, no. 03, pages 10 - 11 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115033690A (en) * | 2022-05-31 | 2022-09-09 | 国网江苏省电力有限公司信息通信分公司 | Communication defect study and judgment knowledge base construction method, defect identification method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111460167A (en) | Method for positioning pollution discharge object based on knowledge graph and related equipment | |
| CN108304567B (en) | Method and system for identifying working condition mode and classifying data of high-voltage transformer | |
| CN113191509A (en) | Intelligent order dispatching method, equipment, medium and product based on maintenance personnel portrait | |
| CN109472293A (en) | A kind of grid equipment file data error correction method based on machine learning | |
| CN107480731A (en) | A kind of EARLY RECOGNITION method of thermal power plant's automobile assembly welding Iine fault signature | |
| CN113836381A (en) | System scoring coverage degree tuning method | |
| CN111401784B (en) | Fire safety grade assessment method | |
| CN113435759A (en) | Primary equipment risk intelligent evaluation method based on deep learning | |
| CN113342784A (en) | Database design method for risk assessment of main transformer equipment of power grid | |
| Gao et al. | Defect analysis of the same batch of substation equipment based on big data analysis algorithm | |
| Wang et al. | Deep analysis of power equipment defects based on semantic framework text mining technology | |
| CN117150418B (en) | Transformer operation detection period formulation method and system based on state characteristic fault tree | |
| CN111737993A (en) | Method for extracting health state of equipment from fault defect text of power distribution network equipment | |
| CN113240133A (en) | Relay protection equipment familial defect identification method based on artificial intelligence | |
| CN111330871A (en) | Quality classification method and device | |
| CN114662981B (en) | Pollution source enterprise supervision method based on big data application | |
| CN116452054A (en) | Method and device for managing material spot check of electric power system | |
| CN115494431A (en) | Transformer fault warning method, terminal equipment and computer readable storage medium | |
| CN112732773B (en) | Method and system for checking uniqueness of relay protection defect data | |
| CN112258095B (en) | Standard normal distribution based scoring method, device, equipment and storage medium | |
| CN111882289B (en) | Device and method for measuring and calculating project data auditing index interval | |
| CN114386745A (en) | PMS power transformation equipment data checking and identifying method and system | |
| CN113706015A (en) | Data collection system for enterprise asset arrangement based on block chain | |
| CN110703183A (en) | Intelligent electric energy meter fault data analysis method and system | |
| CN111143622A (en) | Fault data set construction method based on big data platform |
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 |