CN113449456A - Health state assessment method for power transformer under incomplete multi-mode information - Google Patents
Health state assessment method for power transformer under incomplete multi-mode information Download PDFInfo
- Publication number
- CN113449456A CN113449456A CN202110766718.1A CN202110766718A CN113449456A CN 113449456 A CN113449456 A CN 113449456A CN 202110766718 A CN202110766718 A CN 202110766718A CN 113449456 A CN113449456 A CN 113449456A
- Authority
- CN
- China
- Prior art keywords
- power transformer
- attribute
- weight
- data
- information
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 37
- 230000036541 health Effects 0.000 title claims abstract description 26
- 230000009467 reduction Effects 0.000 claims abstract description 41
- 238000011156 evaluation Methods 0.000 claims abstract description 21
- 230000010354 integration Effects 0.000 claims abstract description 16
- 238000012545 processing Methods 0.000 claims abstract description 15
- 230000007246 mechanism Effects 0.000 claims abstract description 6
- 239000002245 particle Substances 0.000 claims description 60
- 238000012544 monitoring process Methods 0.000 claims description 14
- 230000003044 adaptive effect Effects 0.000 claims description 10
- 230000006870 function Effects 0.000 claims description 10
- 230000003862 health status Effects 0.000 claims description 8
- 238000009413 insulation Methods 0.000 claims description 6
- 238000013506 data mapping Methods 0.000 claims description 5
- 230000001133 acceleration Effects 0.000 claims description 4
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 3
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 claims description 2
- 230000009286 beneficial effect Effects 0.000 claims description 2
- 238000010606 normalization Methods 0.000 claims description 2
- 230000009466 transformation Effects 0.000 abstract description 6
- 238000012423 maintenance Methods 0.000 abstract description 4
- 238000001514 detection method Methods 0.000 abstract description 3
- 238000004804 winding Methods 0.000 description 9
- 238000003745 diagnosis Methods 0.000 description 7
- 238000012360 testing method Methods 0.000 description 6
- 230000008569 process Effects 0.000 description 5
- 238000011160 research Methods 0.000 description 5
- 238000007726 management method Methods 0.000 description 4
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical group [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 3
- 238000013480 data collection Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000002474 experimental method Methods 0.000 description 3
- 238000007689 inspection Methods 0.000 description 3
- 230000008439 repair process Effects 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 238000004587 chromatography analysis Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000004927 fusion Effects 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013075 data extraction Methods 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 230000006866 deterioration Effects 0.000 description 1
- 238000002405 diagnostic procedure Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000012854 evaluation process Methods 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012806 monitoring device Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000009666 routine test Methods 0.000 description 1
- 238000010187 selection method Methods 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/25—Design optimisation, verification or simulation using particle-based methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/08—Probabilistic or stochastic CAD
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)
- Health & Medical Sciences (AREA)
- Strategic Management (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Development Economics (AREA)
- Educational Administration (AREA)
- General Health & Medical Sciences (AREA)
- Entrepreneurship & Innovation (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Artificial Intelligence (AREA)
- Game Theory and Decision Science (AREA)
- Geometry (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- Computer Hardware Design (AREA)
- Primary Health Care (AREA)
- Life Sciences & Earth Sciences (AREA)
- Operations Research (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Monitoring And Testing Of Nuclear Reactors (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention discloses a method for evaluating the health state of a power transformer under the condition of incomplete multi-mode information, which comprises the following steps: acquiring parameter information of a power transformer, wherein the parameter information contains a plurality of characteristic indexes; collecting fault cases of the power transformer, discretizing each modal characteristic according to the difference that the power transformer with different voltage grades, rated capacity and operation life has the requirement on whether multiple characteristic indexes of the power transformer exceed the threshold value, and establishing an information decision table of each modal characteristic by using a rough set; solving the NP-Hard problem of attribute reduction under the multi-modal background by adopting a PSO algorithm, and acquiring the nuclear characteristics of the health state of the power transformer; completing core feature integration by using a TFG-T task flow mechanism; and evaluating the state of the power transformer through an integrated nuclear characteristic and information decision table. The method is used for processing the state evaluation problem of the equipment under the incomplete information, and fusing the multimode of the detection data, thereby providing technical support for intelligent operation and maintenance and digital transformation of the power transformer.
Description
Technical Field
The invention relates to the technical field of transformer state evaluation, in particular to a method for evaluating the health state of a power transformer under the condition of incomplete multi-mode information.
Background
With the continuous deepening of the digital transformation of the power system, the sensing network and the intelligent measurement technology are more and more applied, and powerful data support is provided for the accurate evaluation of the equipment state under the panoramic monitoring data. The power transformer is a core device in the power system, and the quality of the operation state of the power transformer directly affects whether the power system can safely and stably operate.
"should repair and" intelligent repair "become the inevitable trend of equipment repair. When the state of the power transformer is evaluated, the index system is complicated, the data types are various, and multi-mode characteristics are presented to a certain extent, namely, a certain specific object is described through multiple channel information such as images, numerical values, texts, videos and the like. However, due to the problems of communication equipment failure, data specification and the like, a large number of vacancy values, abnormal values and the like exist in related monitoring indexes, and as seen from the research and analysis of the large data quality of a certain transformer intelligent operation and inspection control platform, indexes related to a power transformer from 1986 to 2016 are as follows: the total recorded number of the oil chromatography experiment, the insulating oil experiment, the winding insulation resistance experiment, the top layer oil temperature and the like is 1384565, the non-empty recorded number is 565401, and the average non-empty proportion of the data is only 40.80%.
At present, the state evaluation of the power transformer has certain limitations:
1) the existing state evaluation guide is still dependent on the traditional artificial knowledge experience, most of the existing state evaluation guide adopts a threshold value discrimination mode, the evaluation process is greatly influenced by subjectivity, and the accurate rule between the fault and the characteristic is difficult to objectively reflect;
2) the relevant research focuses on paying attention to the complementarity of each characteristic index, and the subjective and objective weights of the indexes are obtained through methods such as an incidence relation method and an analytic hierarchy process, so that the comprehensive health state of the power transformer is obtained through weighting analysis;
3) the industry judges the state evaluation method of the power transformer in a classification mode by initial single parameters, gradually turns to a multi-parameter combined action, and considers the multi-factor methods of the running environment, the bad working condition information and the like of the transformer, and the research on the construction method of the key evaluation indexes of the power equipment is still less at present;
4) in a traditional power industry mode, most of characteristic data reflecting the operation health state of a power transformer are dispersedly stored in each subsystem database by different logic entities or data characteristics, and the problem of information island is prominent.
Disclosure of Invention
The invention aims to provide a health state assessment method of a power transformer under incomplete multi-mode information, which is used for solving the problem of state assessment of equipment under the incomplete information, fusing the multi-mode state of detection data and providing technical support for intelligent operation and maintenance and digital transformation of the power transformer.
In order to achieve the purpose, the invention provides the following scheme:
a health state assessment method of a power transformer under incomplete multi-mode information comprises the following steps:
s1) parameter information of the power transformer is obtained, wherein the parameter information contains a plurality of characteristic indexes and allows missing data to exist;
s2) collecting fault cases of the power transformer, discretizing each modal characteristic according to the difference that whether a plurality of characteristic indexes of the power transformer with different voltage grades, rated capacity and operation years meet the standard threshold requirement, and establishing an information decision table of each modal characteristic by using a rough set;
s3) solving the NP-Hard problem of attribute reduction under the multi-modal background by adopting a PSO algorithm, and acquiring the kernel characteristic of the health state of the power transformer;
s4) completing the core feature integration of the health state of the power transformer by using a TFG-T task flow mechanism;
s5) evaluating the state of the power transformer through the integrated core features and the information decision table in the step S2).
Optionally, the parameter information of the power transformer in step S1) at least includes one or more of parameter information of the power transformer about mechanical properties, insulation properties, thermal properties, and operating environment.
Optionally, in step S3), the NP-Hard problem of attribute reduction in the multimodal background is solved by using a PSO algorithm, and the kernel feature of the health state of the power transformer is obtained, which specifically includes:
s301), inputting an information decision table, and obtaining the kernel attribute Kernel (tx), the reduction set Z (tx) and the importance of each feature of the power transformer fault by using a relative attribute reduction algorithm; if there is kernel (tx) ═ z (tx), the result is minimal reduction, otherwise step S302 is executed;
s302) initializing the particle group for the obtained reduction set z (tx): m particles (20 is more than or equal to m and less than or equal to 40) are initialized randomly according to the weight of each attribute, and the greater the importance of the feature is, the greater the corresponding weight is, and the greater the possibility that the corresponding position feature takes 1 is; since the core attribute belongs to the minimum reduction, the core attribute weight value takes its importance, as shown in formula (1):
weightk=σcd (1)
in the formula: weightkRepresenting the weight, σ, corresponding to the kernel attributecdIndicating the importance of the attribute;
and (3) carrying out normalization processing on the rest characteristic weights, wherein the formula is as follows (2):
in the formula: weightotherWeight, normalized to represent the uncore attributekWeight, representing the current attribute in the uncore attributeminRepresents the minimum weight value, weight, in the uncore attributemaxRepresents the largest weight value in the uncore attributes;
the corresponding position of the initialized particle is denoted as pkAs in formula (3):
in the formula: rand () is a random number of (0,1), weightkRepresenting the weight value corresponding to the current attribute;
s303) in order to make the number of the features in the particle equal to the contribution of the dependency function of the feature set to the adaptive value, and simultaneously, the reduction of the number of the features or the increase of the dependency are both beneficial to the increase of the total adaptive value, the adaptive value function of the particle is as follows:
in the formula, gammaP/γcFor determining whether the same reduction, k1Indicating that a certain particle is more important in the reduction result, k2Representing that the number of attributes in the reduction result is more important, the card () function represents the number of elements contained in a certain set, C is all attribute sets, and p is a specific attribute;
s304) updating each particle swarm, updating the positions and the speeds of the particles through the formulas (5) and (6), calculating the adaptive values of the updated particles, and updating the global optimal particle gbest and each local optimal particle pbest in time:
pbesti=max(pbesti,ffitness(i)) (5)
gbest=max(pbesti,ffitness(i)) (6)
in the formula: gbest is the globally optimal particle, pbest is the locally optimal particle, ffitness(i) Is the adaptive value calculated by step S303);
s305) in order to ensure that the particles find the optimal solution through iteration, the current speed and the current position of the particles need to be continuously updated according to pbest and gbest, and the updating methods of the speed and the position are as shown in the formulas (7) and (8):
xi=xi+vi (8)
in the formula: i is 1,2,3 … N-1, N, N is the total number of particles in the population, viThe velocity of the particle is represented by range () between (0,1)Random number of (2), xiIs the current particle position, c1Acceleration item weight, c, representing movement of a particle to a historical location pbest, as a local learning factor2The global learning factor represents the weight of the acceleration direction when the particles move to the gbest, and w is the inertia of the previous movement direction of the particles in the current direction and takes a non-negative value;
s306) the PSO algorithm processing result is a continuous value, the processing result is converted into a particle form, such as a formula (9) and a formula (10), and the nuclear characteristics of the health state of the power transformer are obtained:
sig(x)=(1+exp(-x))-1 (10)
where rand () is a random number between (0,1), and sig (x) is distributed between (0, 1).
Optionally, in step S4), completing core feature integration of the health state of the power transformer by using a TFG-T task flow mechanism, specifically including:
s401) analyzing the XML format request list to obtain a database access path, a table field and data related to secondary model processing of an evaluation task, requesting each module in a URL form, processing each operator according to task distribution conditions, and requesting original data from a CIM mode center by the operator;
s402) each CIM mode center completes the storage iteration of the latest monitoring data of the equipment according to different data updating frequencies, when the CIM mode center receives a related request list, the CIM mode center performs data resource conversion URI on the type and the object, completes data mapping by using the mapping rule of the object and the data, packages the data into the request list, and returns the data to the request task body, thus completing the nuclear characteristic integration of the health state of the power transformer.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects: the method for evaluating the health state of the power transformer under the condition of incomplete multi-mode information can effectively process the characteristic selection of the power transformer under the condition of modal data loss, and further carry out equipment evaluation, fault diagnosis and other problems; the invention completes the fault diagnosis of the equipment by using the matching method of the information decision table, and simultaneously constructs the core characteristics of the equipment by using the multi-mode attribute reduction of the PSO, the reduction algorithm can enable the PSO algorithm to find the global optimal solution instead of converging to the local optimal solution prematurely, has the characteristics of high convergence speed, good stability and strong optimizing capability, and can comprehensively utilize the multi-mode information of the equipment to diagnose the fault; the TFG-T task flow integration method well solves the problem of cross-system integration of index data during equipment evaluation, can effectively decouple customer requirements, and can perform data mapping access on a model of a CIM center by modifying access nodes related to corresponding XML (extensive makeup language), so that the system is more flexible; the research result of the invention has good reference value for technical requirements under the background of digital transformation of the power grid, comprehensive and considerable equipment state under multi-mode data fusion and fault diagnosis.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without inventive exercise.
Fig. 1 is a flowchart of a method for evaluating a health status of a power transformer without completing multi-modal information according to an embodiment of the present invention;
FIG. 2 is a diagram of data collection for a power transformer with multi-modal characteristics according to an embodiment of the present invention;
FIG. 3 is a diagram of data collection for a power transformer with multi-modal characteristics according to an embodiment of the present invention;
FIG. 4 is a flow chart of a power transformer information decision table under a multi-mode established based on a fault case library according to an embodiment of the present invention;
FIG. 5 is an XML representation of a single TFG according to an embodiment of the present invention;
fig. 6 is a schematic diagram of a TFG workflow under multiple tasks according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the 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 invention aims to provide a health state assessment method of a power transformer under incomplete multi-mode information, which is used for solving the problem of state assessment of equipment under the incomplete information, fusing the multi-mode state of detection data and providing technical support for intelligent operation and maintenance and digital transformation of the power transformer.
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in further detail below.
As shown in fig. 1, the method for evaluating the health status of a power transformer without multi-modal information according to the embodiment of the present invention includes the following steps:
s1) arranging and installing each sensor on a corresponding monitoring position of equipment, such as gas monitoring, vibration monitoring, sound monitoring and the like, and acquiring parameter information of the power transformer, wherein the parameter information contains a plurality of characteristic indexes and allows missing data to exist; as shown in fig. 2, the data collection distribution is that a plurality of data such as electric signals, vibration signals, operating environments, load information and the like of the power transformer are completed by using an internet of things sensing, edge processing and cloud decision mode;
s2) as shown in fig. 3, collecting fault cases of the power transformer, discretizing each modal characteristic according to the difference of whether the power transformer with different voltage classes, rated capacities and operation years has the requirement of exceeding the threshold of multiple characteristic indexes, and establishing an information decision table of each modal characteristic by using a rough centralized tolerance relationship, an equivalence relationship, etc.; compared with the traditional feature selection method, the method has the advantages that incomplete and inconsistent incomplete information in the information system can be better processed from a knowledge level by utilizing the rough set for attribute reduction, implicit knowledge is obtained from the incomplete information, and potential rules are revealed; in the information decision table, "+" indicates that information is missing, "1" indicates that the numerical value is normal, "2" indicates that the numerical value exceeds the standard, and examples are shown in table 1, table 2 and table 3:
TABLE 1 Transformer body Fault information decision List Condition attributes
TABLE 2 Transformer body Fault information decision List Condition attributes
TABLE 3 Transformer core and clip information decision-making Table
Taking the multipoint grounding fault type of the iron core and the clamping piece as an example, the information decision table STX (U) under the multipoint grounding of the iron core and the clamping piece is inputTX,QTX,VTX,FTX) (ii) a Obtaining the kernel attribute kernel (tx) and the reduction set z (tx) of the fault and the importance of each feature by using a relative attribute reduction algorithm (the algorithm flow is shown in fig. 4); if kernel (tx) ═ z (tx), the result is the minimum reduction, the algorithm is ended, otherwise, particle swarm initialization is performed on the reduction set;
s3), solving the NP-Hard problem of attribute reduction under the multi-modal background by adopting a PSO algorithm, and acquiring the kernel characteristics of the health state of the power transformer, wherein the specific process is as follows:
performing particle swarm initialization on the obtained reduction set: m particles (m is more than or equal to 20 and less than or equal to 40) are initialized randomly according to the weight of each attribute, and the greater the feature importance is, the greater the corresponding weight is, and the greater the possibility that the corresponding position feature takes 1 is; since the core attribute must belong to the minimum reduction, the core attribute weight value takes its importance as shown in formula (1):
weightk=σcd (1)
in the formula: weightkRepresenting the weight, σ, corresponding to the kernel attributecdIndicating the importance of the attribute;
the other feature weights need to be normalized, as shown in formula (2):
in the formula: weightotherWeight, normalized to represent the uncore attributekWeight, representing the current attribute in the uncore attributeminRepresents the minimum weight value, weight, in the uncore attributemaxRepresents the largest weight value in the uncore attributes;
the corresponding position of the initialized particle is denoted as pkAs in formula (3):
in the formula: rand () is a random number of (0,1), weightkRepresenting the weight value corresponding to the current attribute;
in order to make the number of features in the particle equal to the contribution of the dependency function of the feature set to the fitness value, and simultaneously, the decrease of the number of features or the increase of the dependency should be favorable for the increase of the total fitness value, the fitness value function of the particle is as follows:
in the formula, gammaP/γcFor determining whether the same reduction, k1Indicating that a certain particle is more important in the reduction result, k2Representing that the number of attributes in the reduction result is more important, the card () function represents the number of elements contained in a set, C is all attribute sets, and p is a certain itemAttributes of the body;
updating each particle group, updating the position and speed of the particle by equations (5) and (6), and calculating the adaptive value f of the updated particlefitness(i) Updating the global optimal particle gbest and each local optimal particle pbest in time according to the flow chart;
pbesti=max(pbesti,ffitness(i)) (5)
gbest=max(pbesti,ffitness(i)) (6)
in order to ensure that the particles find the optimal solution through iteration, the current speed and position of the particles need to be continuously updated according to pbest and gbest, and the updating method of the speed and the position is shown as the following formulas (7) and (8):
xi=xi+vi (8)
wherein i is 1,2,3 … N-1, N, N represents the total number of particles in the population, viRepresenting the velocity of the particle, and the rand () function represents a random number, x, between (0,1)iIndicating the current particle position. c. C1An acceleration item weight representing a local learning factor, representing the movement of the particle to the historical location pbest, c2The global learning factor is represented, the weight of the acceleration direction of the particles moving to the gbest is represented, w represents the inertia of the movement direction of the particles in the current direction, the value is a non-negative value, and the larger the value is, the stronger the global optimization capability is indicated; in general, c1=c2∈[0,4],w∈[0.4,0.9]In the present invention, c1=c2=2,w=0.435;
Since the PSO processing results are also continuous values, the results obtained from the PSO algorithm still need to be converted into particle form, which is expressed by equations (9), (10):
sig(x)=(1+exp(-x))-1 (10)
where rand () represents a random number between (0,1), and sig (x) represents a distribution between (0, 1);
attribute reduction (Attribute reduction), also known as feature selection, has been difficult to implement by many studies, and based on the above power transformer information decision table, a related knowledge reduction definition needs to be further provided.
DL/T1685-
The related equipment state evaluation and overhaul file standards stipulate the state information classification, the state evaluation classification, the basic state evaluation requirements, the quantitative standards of state quantity, components and the overall evaluation method of the AC oil immersed transformer (reactor) in operation, and have an auxiliary comprehensiveness effect on the establishment of an information decision table of the power transformer and the like;
power transformer data distribution is mainly divided into two types: a. online data: the on-line monitoring data mainly comprises oil chromatography, iron core grounding current, top layer oil temperature and the like, wherein the related monitoring data of the transformer is obtained in real time through an on-line monitoring device (sensor) associated with the transformer; b. offline data: the method mainly comprises the steps of obtaining test results of transformer standing book-related delivery test reports, handover acceptance reports, routine test reports and diagnostic test reports, routing inspection data and the like; the test mainly comprises an oil chromatogram, a winding insulation resistance, a winding direct-current resistance, a winding voltage ratio, a winding dielectric loss and capacitance, a winding direct-current leakage current, a no-load test and the like; in addition, other related data comprise unstructured data such as winding frequency response test, sleeve partial discharge, infrared test, routing inspection data and the like; in an actual application scenario, online monitoring data is mainly in an HBASE library of a management and control system, offline data is mainly in an ORACLE library of the management and control system, and data of the management and control system is extracted from a source data system to the ORACLE library and the HBASE library of the management and control system in a data extraction mode, wherein the number of related database tables is about 14;
s4) completing the core feature integration of the health state of the power transformer by using a TFG-T task flow mechanism; when equipment evaluation is proposed, carrying out task decomposition on the whole evaluation; integrating the core features related to each subtask, designing a plurality of single task flow unit bodies in an XML form, as shown in FIG. 5, recording information such as relevant databases, data tables, data fields and the like in a form of task bodies for the distribution of core index data in a system, and using a core feature integration method under a multitask body as shown in FIG. 6;
s5) evaluating the state of the power transformer through the integrated core characteristics and the information decision table in the step S2); when a user proposes to evaluate equipment, decomposing the indexes related to the step S4), and performing comparison analysis according to the equipment data collected in the step S1) and the related power transformer information decision table obtained in the step S2) so as to judge the health state and the fault type of the equipment, wherein the specific steps comprise:
a) after the evaluation task is sent out, uniformly setting an accessed task queue, carrying out queue time marking on the task, and continuously polling and inquiring related queue tasks from the message queue and identifying by the server side; at the moment, an asynchronous mode is adopted, and RESTFulWebService is used for completing the separation of an access interface and a database;
b) after receiving the list, the RESTFul WebService background firstly analyzes the list of the XML format request to obtain a database access path, a table field and possibly related secondary model processing, if some indexes possibly relate to threshold judgment, image identification and the like, all modules need to be requested in a URL form, and all operators process according to task distribution conditions; in addition, each operator needs to request original data from a CIM mode center;
c) each CIM center needs to complete the storage iteration of the latest monitoring data of the equipment according to different data updating frequencies, and for certain gas content, different time scales can be appointed to obtain the mean value of the gas; after the CIM global mode center receives the relevant request list, the data resource conversion URI is carried out on the type and the object, the data mapping is completed by using the mapping rule of the object and the data, the data is packaged into the request list, and the data is returned to the request task body; and completing data integration of each system.
The parameter information of the power transformer in the step S1) at least includes one or more of parameter information of the power transformer about mechanical property, insulation property, thermal property and operation environment.
For example, a certain type of transformer with SFPSZ 4-150000-.
a) From step 2 to step 9, the core characteristic relating to the ontology is C1、C2、C3、C5、C6、C11、C12、C13、C14、C15、C16、C17When 12 index parameters are met, the multi-mode monitoring data integration result is shown in the table 4:
TABLE 4 multimodal monitoring data integration results table
b) The transformer C can be obtained by utilizing the index threshold information of the operation nodes O3 and O4 in the TFG-T2H4Excessive content of CO2The content exceeds the standard, the insulation resistance value exceeds the standard, and the winding capacitance exceeds the standard;
c) the decision list is available based on the information of step b, i.e. C3、C8、C12、C14Out of limits, C1、C2、C4When the index is normal, corresponding to the fault type D5;
d) The fault type [ normal, attention, abnormity and severity ] membership degree vector is [0.363307,0.05586,0.116846 and 0.463987] can be obtained by processing the index through a deterioration degree and Gaussian cloud model, namely the equipment generates body winding deformation, the deformation is compared with a field maintenance result, the diagnosis result is consistent with the actually generated fault type, and the effectiveness and the feasibility of the method are verified.
The method for evaluating the health state of the power transformer under the condition of incomplete multi-mode information can effectively process the characteristic selection of the power transformer under the condition of modal data loss, and further carry out equipment evaluation, fault diagnosis and other problems; the invention completes the fault diagnosis of the equipment by using the matching method of the information decision table, and simultaneously constructs the core characteristics of the equipment by using the multi-mode attribute reduction of the PSO, the reduction algorithm can enable the PSO algorithm to find the global optimal solution instead of converging to the local optimal solution prematurely, has the characteristics of high convergence speed, good stability and strong optimizing capability, and can comprehensively utilize the multi-mode information of the equipment to diagnose the fault; the TFG-T task flow integration method well solves the problem of cross-system integration of index data during equipment evaluation, can effectively decouple customer requirements, and can perform data mapping access on a model of a CIM center by modifying access nodes related to corresponding XML (extensive makeup language), so that the system is more flexible; the research result of the invention has good reference value for technical requirements under the background of digital transformation of the power grid, comprehensive and considerable equipment state under multi-mode data fusion and fault diagnosis.
The principles and embodiments of the present invention have been described herein using specific examples, which are provided only to help understand the method and the core concept of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, the specific embodiments and the application range may be changed. In view of the above, the present disclosure should not be construed as limiting the invention.
Claims (4)
1. A method for evaluating the health state of a power transformer under the condition of incomplete multi-mode information is characterized by comprising the following steps:
s1) parameter information of the power transformer is obtained, wherein the parameter information contains a plurality of characteristic indexes and allows missing data to exist;
s2) collecting fault cases of the power transformer, discretizing each modal characteristic according to the difference that whether a plurality of characteristic indexes of the power transformer with different voltage grades, rated capacity and operation years meet the standard threshold requirement, and establishing an information decision table of each modal characteristic by using a rough set;
s3) solving the NP-Hard problem of attribute reduction under the multi-modal background by adopting a PSO algorithm, and acquiring the kernel characteristic of the health state of the power transformer;
s4) completing the core feature integration of the health state of the power transformer by using a TFG-T task flow mechanism;
s5) evaluating the state of the power transformer through the integrated core features and the information decision table in the step S2).
2. The method for assessing the health status of a power transformer without multi-modal information as recited in claim 1, wherein the parameter information of the power transformer in the step S1) at least comprises one or more of the parameter information of the power transformer regarding mechanical properties, insulation properties, thermal properties and operating environment.
3. The method for assessing the health status of a power transformer without multi-modal information as claimed in claim 1, wherein in step S3), the method for assessing the health status of a power transformer using a PSO algorithm to solve the NP-Hard problem of attribute reduction in a multi-modal context and obtaining the kernel signature of the health status of the power transformer specifically comprises:
s301), inputting an information decision table, and obtaining the kernel attribute Kernel (tx), the reduction set Z (tx) and the importance of each feature of the power transformer fault by using a relative attribute reduction algorithm; if there is kernel (tx) ═ z (tx), the result is minimal reduction, otherwise step S302 is executed;
s302) initializing the particle group for the obtained reduction set z (tx): m particles (20 is more than or equal to m and less than or equal to 40) are initialized randomly according to the weight of each attribute, and the greater the importance of the feature is, the greater the corresponding weight is, and the greater the possibility that the corresponding position feature takes 1 is; since the core attribute belongs to the minimum reduction, the core attribute weight value takes its importance, as shown in formula (1):
weightk=σcd (1)
in the formula: weightkRepresenting the weight, σ, corresponding to the kernel attributecdIndicating the importance of the attribute;
and (3) carrying out normalization processing on the rest characteristic weights, wherein the formula is as follows (2):
in the formula: weightotherWeight, normalized to represent the uncore attributekWeight, representing the current attribute in the uncore attributeminRepresents the minimum weight value, weight, in the uncore attributemaxRepresents the largest weight value in the uncore attributes;
the corresponding position of the initialized particle is denoted as pkAs in formula (3):
in the formula: rand () is a random number of (0,1), weightkRepresenting the weight value corresponding to the current attribute;
s303) in order to make the number of the features in the particle equal to the contribution of the dependency function of the feature set to the adaptive value, and simultaneously, the reduction of the number of the features or the increase of the dependency are both beneficial to the increase of the total adaptive value, the adaptive value function of the particle is as follows:
in the formula, gammaP/γcFor determining whether the same reduction, k1Indicating that a certain particle is more important in the reduction result, k2Representing that the number of attributes in the reduction result is more important, the card () function represents the number of elements contained in a certain set, C is all attribute sets, and p is a specific attribute;
pbesti=max(pbesti,ffitness(i)) (5)
gbest=max(pbesti,ffitness(i)) (6)
in the formula: gbest is the globally optimal particle, pbest is the locally optimal particle, ffitness(i) Is the adaptive value calculated by step S303);
s305) in order to ensure that the particles find the optimal solution through iteration, the current speed and the current position of the particles need to be continuously updated according to pbest and gbest, and the updating methods of the speed and the position are as shown in the formulas (7) and (8):
vi=w×vi+c1×rand()×(pbesti-xi)+c2×rand()×(gbesti-xi) (7)
xi=xi+vi (8)
in the formula: i is 1,2,3 … N-1, N, N is the total number of particles in the population, viFor the velocity of the particle, rand () is a random number between (0,1), xiIs the current particle position, c1Acceleration item weight, c, representing movement of a particle to a historical location pbest, as a local learning factor2The global learning factor represents the weight of the acceleration direction when the particles move to the gbest, and w is the inertia of the previous movement direction of the particles in the current direction and takes a non-negative value;
s306) the PSO algorithm processing result is a continuous value, the processing result is converted into a particle form, such as a formula (9) and a formula (10), and the nuclear characteristics of the health state of the power transformer are obtained:
sig(x)=(1+exp(-x))-1 (10)
where rand () is a random number between (0,1), and sig (x) is distributed between (0, 1).
4. The method for assessing the health status of a power transformer based on incomplete multi-modal information as claimed in claim 1, wherein in step S4), the core feature integration of the health status of the power transformer is completed by using a TFG-T task flow mechanism, which specifically comprises:
s401) analyzing the XML format request list to obtain a database access path, a table field and data related to secondary model processing of an evaluation task, requesting each module in a URL form, processing each operator according to task distribution conditions, and requesting original data from a CIM mode center by the operator;
s402) each CIM mode center completes the storage iteration of the latest monitoring data of the equipment according to different data updating frequencies, when the CIM mode center receives a related request list, the CIM mode center performs data resource conversion URI on the type and the object, completes data mapping by using the mapping rule of the object and the data, packages the data into the request list, and returns the data to the request task body, thus completing the nuclear characteristic integration of the health state of the power transformer.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110766718.1A CN113449456B (en) | 2021-07-07 | 2021-07-07 | Health state assessment method for power transformer under incomplete multi-mode information |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110766718.1A CN113449456B (en) | 2021-07-07 | 2021-07-07 | Health state assessment method for power transformer under incomplete multi-mode information |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113449456A true CN113449456A (en) | 2021-09-28 |
CN113449456B CN113449456B (en) | 2024-03-26 |
Family
ID=77815198
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110766718.1A Active CN113449456B (en) | 2021-07-07 | 2021-07-07 | Health state assessment method for power transformer under incomplete multi-mode information |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113449456B (en) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2015149593A1 (en) * | 2014-04-04 | 2015-10-08 | 袁志贤 | Iec61850-based communication simulation method for transformer grounding on-line monitoring device |
CN106529124A (en) * | 2016-10-14 | 2017-03-22 | 云南电网有限责任公司昆明供电局 | A transformer insulation state assessment method based on principal component analysis and support vector machines |
CN110333414A (en) * | 2019-08-02 | 2019-10-15 | 华北电力大学(保定) | The multi-level state evaluating method of power transformer |
CN110866331A (en) * | 2019-10-28 | 2020-03-06 | 国网河北省电力有限公司电力科学研究院 | Evaluation method for quality defects of power transformer family |
CN112596496A (en) * | 2020-12-08 | 2021-04-02 | 中国船舶重工集团公司第七0四研究所 | Health management platform and management method for ship electric propulsion system |
-
2021
- 2021-07-07 CN CN202110766718.1A patent/CN113449456B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2015149593A1 (en) * | 2014-04-04 | 2015-10-08 | 袁志贤 | Iec61850-based communication simulation method for transformer grounding on-line monitoring device |
CN106529124A (en) * | 2016-10-14 | 2017-03-22 | 云南电网有限责任公司昆明供电局 | A transformer insulation state assessment method based on principal component analysis and support vector machines |
CN110333414A (en) * | 2019-08-02 | 2019-10-15 | 华北电力大学(保定) | The multi-level state evaluating method of power transformer |
CN110866331A (en) * | 2019-10-28 | 2020-03-06 | 国网河北省电力有限公司电力科学研究院 | Evaluation method for quality defects of power transformer family |
CN112596496A (en) * | 2020-12-08 | 2021-04-02 | 中国船舶重工集团公司第七0四研究所 | Health management platform and management method for ship electric propulsion system |
Non-Patent Citations (1)
Title |
---|
陈发广;周步祥;曾澜钰;: "基于多信息融合的变压器运行状态评估模型", 电力系统及其自动化学报, no. 04 * |
Also Published As
Publication number | Publication date |
---|---|
CN113449456B (en) | 2024-03-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110826228B (en) | Regional power grid operation quality limit evaluation method | |
CN112906764B (en) | Communication safety equipment intelligent diagnosis method and system based on improved BP neural network | |
CN111563682A (en) | Test evaluation method for distribution automation equipment | |
CN110361609B (en) | Extra-high voltage equipment monitoring system and method | |
CN114444910A (en) | Electric power Internet of things-oriented edge network system health degree evaluation method | |
CN113627735A (en) | Early warning method and system for safety risk of engineering construction project | |
CN116933010A (en) | Load rate analysis and evaluation method and system based on multi-source data fusion and deep learning | |
CN114757392A (en) | Power grid system comprehensive evaluation index system optimization method based on game weight method | |
CN111724071A (en) | TOPSIS method-based intelligent box-type substation operation state evaluation method | |
CN116579640A (en) | Power marketing service channel user experience assessment method and system | |
Zhang et al. | Quality attributes assessment for feature-based product configuration in software product line | |
Wang et al. | Component-oriented reassembly in remanufacturing systems: Managing uncertainty and satisfying customer needs | |
MansourLakouraj et al. | Event classification in active distribution grids using physics-informed graph neural network and PMU measurements | |
Wang et al. | Power equipment fault information acquisition system based on Internet of things | |
Wang et al. | Human Operating Risk Assessment for Outdoor Terminal Box of Electric Power | |
Wenjie et al. | A multi-index evaluation method of voltage sag based on the comprehensive weight | |
CN113449456A (en) | Health state assessment method for power transformer under incomplete multi-mode information | |
Rocha et al. | Fast and flexible design of optimal metering systems for power systems monitoring | |
CN116842936A (en) | Keyword recognition method, keyword recognition device, electronic equipment and computer readable storage medium | |
CN110766248A (en) | Workshop human factor reliability evaluation method based on SHEL and interval intuition fuzzy evaluation | |
Yu et al. | Knowledge reasoning of transmission line component detection using CRITIC and TOPSIS approaches | |
Zheng et al. | [Retracted] Application Based on Artificial Intelligence in Substation Operation and Maintenance Management | |
Chen et al. | A new approach to the application of condition-based maintenance technology of power equipment in smart grid | |
Chen et al. | BP neural network-based model for evaluating user interfaces of human-computer interaction system | |
Huang | Static security assessment of a power system using query-based learning approaches with genetic enhancement |
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 |