CN109033513B - Power transformer fault diagnosis method and power transformer fault diagnosis device - Google Patents
Power transformer fault diagnosis method and power transformer fault diagnosis device Download PDFInfo
- Publication number
- CN109033513B CN109033513B CN201810623024.0A CN201810623024A CN109033513B CN 109033513 B CN109033513 B CN 109033513B CN 201810623024 A CN201810623024 A CN 201810623024A CN 109033513 B CN109033513 B CN 109033513B
- Authority
- CN
- China
- Prior art keywords
- fault diagnosis
- power transformer
- transformer fault
- population
- matrix
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
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
-
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/04—Inference or reasoning models
- G06N5/048—Fuzzy inferencing
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Computing Systems (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Automation & Control Theory (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Fuzzy Systems (AREA)
- Biophysics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Housings And Mounting Of Transformers (AREA)
Abstract
The invention relates to a power transformer fault diagnosis method and a power transformer fault diagnosis device, wherein the power transformer fault diagnosis method comprises the following steps: extracting characteristic vectors according to the original data of the volume fractions of the characteristic gas components obtained by analyzing the dissolved gas in the power transformer oil, inputting the number of data samples to be diagnosed and the number of characteristic variables, and establishing an original data matrix as a fault diagnosis database; carrying out denoising treatment and standardization treatment on sample data in an original data matrix to obtain a standardization matrix; according to the standardized matrix, training and adjusting system parameters of a self-adaptive fuzzy neural inference system by using an improved quantum particle swarm algorithm, and establishing a transformer fault diagnosis model; and diagnosing the test data by adopting the transformer fault diagnosis model. Therefore, the fault type of the power transformer can be rapidly and accurately identified, the intelligent degree and reliability of the fault diagnosis of the transformer are improved, and the diagnosis efficiency is higher and more accurate.
Description
Technical Field
The present invention relates to a power system, and more particularly, to a power transformer fault diagnosis method and a power transformer fault diagnosis apparatus.
Background
Power transformers are one of the vital devices in an electrical power system, the operating conditions of which are directly related to the safe and stable operation of the electrical power system. At present, analysis (dissolved gas analysis, DGA) of dissolved gas in oil is one of important methods for diagnosing faults of a transformer, and can effectively discover latent faults of the transformer in advance and monitor the development of the faults on line at any time. The traditional transformer fault diagnosis method mainly comprises an IEC three-ratio method, a gas graph method, a characteristic gas method and the like based on DGA, and the methods have the advantages of simple principle and simplicity and convenience in operation, but are limited by complex fault characteristics of the transformer, and have great limitation in practical application. With the development of artificial intelligence, various intelligent methods such as artificial neural networks, fuzzy theory, support vector machines and the like are gradually introduced into transformer fault diagnosis research. The intelligent diagnosis method improves the application accuracy and the practicability of fault diagnosis to a certain extent. However, various intelligent methods have the disadvantage of being inefficient in the application of complex transformer fault diagnostics.
Disclosure of Invention
Based on this, it is necessary to provide a power transformer fault diagnosis method and a power transformer fault diagnosis apparatus.
A power transformer fault diagnosis method, comprising the steps of:
extracting characteristic vectors according to the original data of the volume fractions of the characteristic gas components obtained by analyzing the dissolved gas in the power transformer oil, inputting the number of data samples to be diagnosed and the number of characteristic variables, and establishing an original data matrix as a fault diagnosis database;
carrying out denoising treatment and standardization treatment on sample data in an original data matrix to obtain a standardization matrix;
according to the standardized matrix, training and adjusting system parameters of a self-adaptive fuzzy neural inference system by using an improved quantum particle swarm algorithm, and establishing a transformer fault diagnosis model;
and diagnosing the test data by adopting the transformer fault diagnosis model.
According to the power transformer fault diagnosis method, the characteristic quantity of dissolved gas in oil is used as an input variable, the type of transformer fault is used as an output, and the self-adaptive fuzzy neural inference system is optimized by using the improved quantum particle swarm algorithm on the basis of the self-adaptive fuzzy neural inference system, so that the optimizing precision and the calculating efficiency of the algorithm are improved, the type of the power transformer fault can be rapidly and accurately identified, the intelligent degree and the reliability of the transformer fault diagnosis are improved, and the diagnosis efficiency is higher.
In one embodiment, the denoising process includes a gaussian filter denoising process.
In one embodiment, the denoising and normalizing the sample data in the original data matrix includes: sample data in an original data matrix is subjected to denoising treatment, and then the denoised data matrix is subjected to standardization treatment.
In one embodiment, the training and adjusting the system parameters of the adaptive fuzzy neural inference system by using the improved quantum particle swarm algorithm according to the standardized matrix, and establishing a transformer fault diagnosis model, includes the steps of:
s1, each particle of an improved quantum particle swarm is encoded to generate an initial population, and a particle vector consists of system parameters of a self-adaptive fuzzy neural inference system, wherein the system parameters comprise a front part parameter and a back part parameter;
s2, calculating an fitness function of each particle in the initial population according to the standardized matrix, and recording an individual history optimal value and a population global optimal value;
s3, judging whether the current iteration times are larger than the maximum set iteration times, if so, executing the step S4, otherwise, executing the step S5;
s4, taking the global optimal solution of the improved quantum particle swarm algorithm as a system parameter of a self-adaptive fuzzy neural inference system network, training and learning the standardized matrix by using the self-adaptive fuzzy neural inference system optimized by the improved quantum particle swarm algorithm, judging whether accuracy is met, if so, building a transformer fault diagnosis model, otherwise, returning to the step S1;
s5, carrying out position updating on all particles of the improved quantum particle swarm algorithm according to the individual historical optimal position and the population global optimal position, recalculating an fitness function on all particles of the improved quantum particle swarm algorithm after the position updating, counting accumulated algebra of the population global optimal continuous stagnation in each iteration preset time of the improved quantum particle swarm algorithm, judging whether the accumulated algebra is greater than a continuous stagnation threshold, otherwise, returning to the execution step S3, executing mutation operator disturbance population global optimal and average optimal positions according to mutation probability, and returning to the execution step S3.
In one embodiment, the fitness function is a root mean square error between the actual output value and the desired output value of the adaptive fuzzy neural inference system.
In one embodiment, the power transformer fault diagnosis method further includes the steps of: and adjusting the preset times.
In one embodiment, the power transformer fault diagnosis method further includes the steps of: and adjusting the mutation probability.
In one embodiment, the power transformer fault diagnosis method further includes the steps of: and adjusting the continuous stagnation threshold.
In one embodiment, the power transformer fault diagnosis method specifically includes the following steps:
the method comprises the steps of obtaining original data of volume fractions of characteristic gas components obtained by analysis of dissolved gas in transformer oil, extracting characteristic vectors, and establishing an original data matrix as a fault diagnosis database according to the number of data to be diagnosed and the number of characteristic variables;
carrying out Gaussian filtering denoising treatment and standardization treatment on sample data in the original data matrix;
encoding the improved quantum particle swarm algorithm particles to generate an initial population;
calculating fitness functions of all particles;
recording the historical optimal value and the global optimal value of the individual;
judging whether the maximum set iteration number is satisfied,
when the maximum set iteration times are met, the system parameters of the self-adaptive fuzzy neural inference system are optimized by improving the global optimal solution of the quantum particle swarm algorithm;
training and learning a self-adaptive fuzzy neural reasoning system model;
judging whether the error meets the precision requirement;
when the error meets the precision requirement, a transformer fault diagnosis model is established, a test sample is input, and a fault diagnosis result is output;
when the error does not meet the precision requirement, returning to continue to execute the coding of the improved quantum particle swarm algorithm particles to generate an initial population;
when the maximum set iteration times are not met, updating the position of each particle to generate a new population of the improved quantum particle swarm algorithm;
recalculating the historical optimal value and the population global optimal value of the particle fitness function individuals;
counting accumulated algebra of continuous stagnation of global optimal values of the population in iteration of preset times;
judging whether the accumulated algebra is larger than a continuous stagnation threshold;
when the accumulated algebra is larger than the continuous stagnation threshold, calculating mutation probability;
executing mutation operator disturbance population global optimum and average optimum positions, and returning to continuously execute and record individual historical optimum values and global optimum values;
and when the accumulated algebra is not greater than the continuous stagnation threshold, returning to continuously execute recording the historical optimal value and the global optimal value of the individual.
A power transformer fault diagnosis device implemented by the power transformer fault diagnosis method according to any one of the above.
Drawings
FIG. 1 is a flow chart of an embodiment of the present invention.
Fig. 2 is a schematic flow chart of a transformer fault diagnosis method based on an improved quantum particle swarm optimization adaptive fuzzy neural inference system according to another embodiment of the invention.
Detailed Description
In order that the above objects, features and advantages of the invention will be readily understood, a more particular description of the invention will be rendered by reference to the appended drawings. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention. The present invention may be embodied in many other forms than described herein and similarly modified by those skilled in the art without departing from the spirit of the invention, whereby the invention is not limited to the specific embodiments disclosed below.
It will be understood that when an element is referred to as being "fixed" or "disposed" on another element, it can be directly on the other element or intervening elements may also be present. When an element is referred to as being "connected" to another element, it can be directly connected to the other element or intervening elements may also be present. The terms "vertical," "horizontal," "left," "right," and the like are used herein for illustrative purposes only and are not meant to be the only embodiment.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and/or" as used herein includes any and all combinations of one or more of the associated listed items.
Aiming at the problem of low accuracy of the traditional transformer fault diagnosis method, the invention provides a power transformer fault diagnosis method based on an improved Quantum-behaved Particle Swarm Optimization (MQPSO) optimized Adaptive Neuro-Fuzzy Inference System (ANFIS), namely a power transformer fault diagnosis method. According to the power transformer fault diagnosis method, the characteristic quantity of dissolved gas in oil is used as an input variable, the type of transformer fault is used as an output, and an MQPSO algorithm is introduced to optimize the parameters of an ANFIS system on the basis of the ANFIS so as to improve the optimizing precision and speed of the algorithm. The power transformer fault diagnosis method based on the MQPSO-ANFIS can quickly and accurately identify the fault type of the transformer, has higher diagnosis efficiency compared with a single intelligent method, and particularly illustrates the following embodiments.
As shown in fig. 1, a power transformer fault diagnosis method includes the steps of: extracting characteristic vectors according to the original data of the volume fractions of the characteristic gas components obtained by analyzing the dissolved gas in the power transformer oil, inputting the number of data samples to be diagnosed and the number of characteristic variables, and establishing an original data matrix as a fault diagnosis database; carrying out denoising treatment and standardization treatment on sample data in an original data matrix to obtain a standardization matrix; according to the standardized matrix, training and adjusting system parameters of a self-adaptive fuzzy neural inference system by using an improved quantum particle swarm algorithm, and establishing a transformer fault diagnosis model; and diagnosing the test data by adopting the transformer fault diagnosis model. According to the power transformer fault diagnosis method, the characteristic quantity of dissolved gas in oil is used as an input variable, the type of transformer fault is used as an output, and the self-adaptive fuzzy neural inference system is optimized by using the improved quantum particle swarm algorithm on the basis of the self-adaptive fuzzy neural inference system, so that the optimizing precision and the calculating efficiency of the algorithm are improved, the type of the power transformer fault can be rapidly and accurately identified, the intelligent degree and the reliability of the transformer fault diagnosis are improved, and the diagnosis efficiency is higher.
In one embodiment, the denoising processing includes gaussian filtering denoising processing, that is, denoising processing and normalizing sample data in an original data matrix to obtain a normalized matrix, which specifically includes: and carrying out Gaussian filtering denoising treatment and standardization treatment on sample data in the original data matrix to obtain a standardization matrix. In one embodiment, a power transformer fault diagnosis method includes the steps of: extracting characteristic vectors according to the original data of the volume fractions of the characteristic gas components obtained by analyzing the dissolved gas in the power transformer oil, inputting the number of data samples to be diagnosed and the number of characteristic variables, and establishing an original data matrix as a fault diagnosis database; carrying out Gaussian filtering denoising treatment and standardization treatment on sample data in an original data matrix to obtain a standardization matrix; according to the standardized matrix, training and adjusting system parameters of a self-adaptive fuzzy neural inference system by using an improved quantum particle swarm algorithm, and establishing a transformer fault diagnosis model; and diagnosing the test data by adopting the transformer fault diagnosis model. The remaining embodiments and so on. In one embodiment, the denoising and normalizing the sample data in the original data matrix includes: sample data in an original data matrix is subjected to denoising treatment, and then the denoised data matrix is subjected to standardization treatment. In one embodiment, the denoising and normalizing the sample data in the original data matrix includes: sample data in an original data matrix is subjected to Gaussian filter denoising treatment, and then the data matrix subjected to Gaussian filter denoising treatment is subjected to standardization treatment.
In one embodiment, the training and adjusting the system parameters of the adaptive fuzzy neural inference system by using the improved quantum particle swarm algorithm according to the standardized matrix, and establishing a transformer fault diagnosis model, includes the steps of: s1, each particle of an improved quantum particle swarm is encoded to generate an initial population, and a particle vector consists of system parameters of a self-adaptive fuzzy neural inference system, wherein the system parameters comprise a front part parameter and a back part parameter; s2, calculating an fitness function of each particle in the initial population according to the standardized matrix, and recording an individual history optimal value and a population global optimal value; s3, judging whether the current iteration times are larger than the maximum set iteration times, if so, executing the step S4, otherwise, executing the step S5; s4, taking the global optimal solution of the improved quantum particle swarm algorithm as a system parameter of a self-adaptive fuzzy neural inference system network, training and learning the standardized matrix by using the self-adaptive fuzzy neural inference system optimized by the improved quantum particle swarm algorithm, judging whether accuracy is met, if so, building a transformer fault diagnosis model, otherwise, returning to the step S1; s5, carrying out position updating on all particles of the improved quantum particle swarm algorithm according to the individual historical optimal position and the population global optimal position, recalculating an fitness function on all particles of the improved quantum particle swarm algorithm after the position updating, counting accumulated algebra of the population global optimal continuous stagnation in each iteration preset time of the improved quantum particle swarm algorithm, judging whether the accumulated algebra is greater than a continuous stagnation threshold, otherwise, returning to the execution step S3, executing mutation operator disturbance population global optimal and average optimal positions according to mutation probability, and returning to the execution step S3. In one embodiment, the fitness function is a root mean square error between the actual output value and the desired output value of the adaptive fuzzy neural inference system. In one embodiment, the power transformer fault diagnosis method further includes the steps of: and adjusting the preset times. In one embodiment, the power transformer fault diagnosis method further includes the steps of: and adjusting the mutation probability. In one embodiment, the power transformer fault diagnosis method further includes the steps of: and adjusting the continuous stagnation threshold. In this way, the preset times, variation probability and/or continuous stagnation threshold can be flexibly adjusted or set according to specific applications in the implementation process. Further, in one embodiment, the parameters of the fitness function include actual and expected output values of the ANFIS model, and MQPSO population size. Further, in one embodiment, performing a location update on all particles of the modified quantum particle swarm algorithm according to the individual historical optimal location and the population global optimal location, includes: setting weight values for the individual historical optimal positions and the population global optimal positions, and carrying out position update on all particles of the improved quantum particle swarm algorithm according to the individual historical optimal positions and the population global optimal positions by adopting the weight values.
In one embodiment, the power transformer fault diagnosis method specifically includes the following steps: the method comprises the steps of obtaining original data of volume fractions of characteristic gas components obtained by analysis of dissolved gas in transformer oil, extracting characteristic vectors, and establishing an original data matrix as a fault diagnosis database according to the number of data to be diagnosed and the number of characteristic variables; carrying out Gaussian filtering denoising treatment and standardization treatment on sample data in the original data matrix; encoding the improved quantum particle swarm algorithm particles to generate an initial population; calculating fitness functions of all particles; recording the historical optimal value and the global optimal value of the individual; judging whether the maximum set iteration times are met, and optimizing the system parameters of the self-adaptive fuzzy neural inference system by using the global optimal solution of the improved quantum particle swarm algorithm when the maximum set iteration times are met; training and learning a self-adaptive fuzzy neural reasoning system model; judging whether the error meets the precision requirement; when the error meets the precision requirement, a transformer fault diagnosis model is established, a test sample is input, and a fault diagnosis result is output; when the error does not meet the precision requirement, returning to continue to execute the coding of the improved quantum particle swarm algorithm particles to generate an initial population; when the maximum set iteration times are not met, updating the position of each particle to generate a new population of the improved quantum particle swarm algorithm; recalculating the historical optimal value and the population global optimal value of the particle fitness function individuals; counting accumulated algebra of continuous stagnation of global optimal values of the population in iteration of preset times; judging whether the accumulated algebra is larger than a continuous stagnation threshold; when the accumulated algebra is larger than the continuous stagnation threshold, calculating mutation probability; executing mutation operator disturbance population global optimum and average optimum positions, and returning to continuously execute and record individual historical optimum values and global optimum values; and when the accumulated algebra is not greater than the continuous stagnation threshold, returning to continuously execute recording the historical optimal value and the global optimal value of the individual.
The embodiments provide a power transformer fault diagnosis method based on an improved quantum particle swarm optimization adaptive fuzzy neural inference system, which is a novel power transformer fault diagnosis method based on a traditional method for analyzing DGA (differential gas analysis) of dissolved gas in oil, and can analyze and diagnose according to the DGA data of the transformer, rapidly and accurately judge the fault type of the transformer, and further improve the accuracy of power transformer fault diagnosis in continuous operation.
The power transformer fault diagnosis method based on the improved quantum particle swarm optimization adaptive fuzzy neural inference system is further described below with reference to specific application. The power transformer fault diagnosis method can also be called as a power transformer fault diagnosis method based on an improved quantum particle swarm optimization adaptive fuzzy neural inference system, and is specifically realized according to part or all of the following steps.
Step one, extracting characteristic vectors according to data of volume fractions of characteristic gas components obtained by analysis of dissolved gas in power transformer oil, inputting the number m of data samples to be diagnosed and the number n of characteristic variables, and establishing an original data matrix X m×n As a fault diagnosis database;
step two, carrying out Gaussian filtering denoising treatment and standardization treatment on sample data in an original data matrix to obtain a standardized matrix, wherein the specific process comprises the following steps of:
(a) Denoising the original data matrix obtained in the step one by using a Gaussian filtering method, wherein the denoising is realized by adopting the following formula:
where x is the data sample and σ is the sample variance.
(b) The data matrix after Gaussian filter denoising is subjected to standardization processing, and the data matrix is realized by adopting the following formula:
wherein x is a data sample; x is x max 、x min Respectively the maximum and minimum values in the data samples; x' is the normalized data sample.
Training and adjusting system parameters of an adaptive fuzzy neural inference system ANFIS by using an improved MQPSO algorithm, and establishing a fault diagnosis model of the MQPSO-ANFIS transformer, wherein the specific process comprises the following steps of:
(1) The MQPSO particles are encoded to generate an initial population, the particle vector is composed of system parameters of the ANFIS, the system parameters comprise a front piece parameter (namely a, b, c and d in membership functions) and a back piece parameter (p, q and r), and the MQPSO particle structure is obtained as follows:
x＝[a,b,c,d,p,q,r]
(2) Calculating the fitness function f of each particle in the MQPSO population according to the standardized matrix in the second step, and recording the individual historical optimal value p id And population global optimum p gd . The fitness function f is expressed as the root mean square error between the actual and expected output values of the ANFIS, and the formula is as follows:
wherein Y, T is the actual output value and the expected output value of the ANFIS model respectively; m is the MQPSO population size.
(3) Judging whether the current iteration number is greater than the maximum set iteration number N max . If yes, executing the step (7); if not, executing the steps (4) to (6).
(4) And carrying out position update on all the particles of the MQPSO according to the historical optimal position of the individual and the global optimal position of the population, wherein the update formula is as follows:
p id (t+1)＝a·p id (t)+(1-a)·p gd (t)
wherein p is id A historic optimal position for particle i; p is p gd The global optimal position of the population; m is M best The average optimal position of all particles in the population; x is x id Is the position of the ith particle;m is population scale; t is the current iteration algebra; alpha is the shrinkage expansion coefficient of QPSO; a and b are [0,1 ]]Random numbers uniformly distributed over the interval.
(5) Recalculating the fitness function f by the MQPSO particles after updating the positions;
(6) Statistics of MQPSO per iteration N c (0＜N c ≤N max ) In the second time, the population global optimum p gd Cumulative algebra N of continuous arrest pla (0＜N pla ≤N c ) And judge N pla Whether or not it is greater than the continuous stagnation threshold N pla0 If yes, according to variation probability P m Executing mutation operator disturbance population global optimum p gd And average optimum position M best The method comprises the steps of carrying out a first treatment on the surface of the If not, not executing the mutation operation; returning to the step (3).
Wherein, according to the mutation probability P m Executing mutation operator disturbance population global optimum p gd And average optimum position M best The method is realized by adopting the following calculation formula.
The variation probability calculation formula is as follows:
wherein P is m The current variation probability of the population; p (P) 0 Is [0,1]Constants within the range.
The formula of the mutation operator is as follows:
p gd (t)＝p gd (t)+μD 1 (·)
M best (t)＝M best (t)+γD 2 (·)
wherein mu and gamma are [0,1 ]]Random numbers uniformly distributed in interval D 1 (·)、D 2 (. Cndot.) is a random variable that obeys the Cauchy distribution.
(7) Taking the global optimal solution of the MQPSO algorithm as a system parameter of the ANFIS network;
(8) Training and learning the standardized data matrix in the second step by utilizing the MQPSO optimized ANFIS, judging whether accuracy is met, and if so, obtaining a fault diagnosis model of the MQPSO-ANFIS transformer; otherwise, returning to the step (1).
And step four, diagnosing the test data by using the power transformer fault diagnosis method.
Further, after the power transformer fault diagnosis method is applied to diagnose the test data or the transformer fault diagnosis model is adopted to diagnose the test data, the power transformer fault diagnosis method further comprises the following steps: and outputting a diagnosis result. In one embodiment, the diagnostic results are output on a display, a workspace, and/or a mobile terminal of the registered user. Further, after the diagnosis result is output, the power transformer fault diagnosis method further includes the steps of: and processing the power transformer fault according to the diagnosis result. Further, after processing the power transformer fault according to the diagnosis result, the power transformer fault diagnosis method further includes the steps of: and acquiring feedback information for processing the power transformer faults according to the diagnosis results. Further, after obtaining feedback information of the power transformer fault processed according to the diagnosis result, the power transformer fault diagnosis method further includes the steps of: and adjusting the transformer fault diagnosis model according to the feedback information. Therefore, the power transformer fault diagnosis method can be continuously learned and updated, so that the transformer fault diagnosis model and the power transformer fault diagnosis method can be continuously optimized, and more accurate and more effective diagnosis results can be obtained.
In one embodiment, as shown in fig. 2, the power transformer fault diagnosis method specifically includes the following steps:
obtaining characteristic gas component volume fraction data obtained by analyzing dissolved gas in transformer oil, and establishing an original data matrix according to the data number to be diagnosed and the characteristic variable number;
carrying out Gaussian filtering denoising and standardization on the data of the original data matrix;
encoding the MQPSO particles to generate an initial population;
calculating fitness functions f of all particles;
recording the most historic of individualsFigure of merit p id And global optimum p gd ；
Determine if the maximum set number of iterations is met?
When the maximum set iteration times are met, optimizing the system parameters of the ANFIS by the MQPSO global optimal solution;
training and learning an ANFIS model;
is the determination error meets the accuracy requirement?
When the error meets the precision requirement, inputting a test sample and outputting a fault diagnosis result;
when the error does not meet the precision requirement, returning to continue to execute the encoding of the MQPSO particles to generate an initial population;
when the maximum set iteration times are not met, updating the position of each particle to generate an MQPSO new population;
recalculating particle fitness functions f and p id 、p gd ；
Statistics of N c Population global optimum p in multiple iterations gd Cumulative algebra N of continuous arrest pla ；
Judging N pla Whether or not it is greater than the continuous stagnation threshold N pla0 ？
N pla Greater than the continuous stagnation threshold N pla0 When calculating the mutation probability p m ；
Executing mutation operator to disturb global optimal and average optimal positions of population, and returning to continuously executing and recording historical optimal values p of individuals id And global optimum p gd ；
N pla Not greater than continuous stagnation threshold N pla0 When the record of the individual history optimal value p is returned to be continuously executed id And global optimum p gd 。
The embodiment of the invention also comprises a power transformer fault diagnosis device which is realized by adopting the power transformer fault diagnosis method according to any embodiment, namely, the power transformer fault diagnosis device which adopts the power transformer fault diagnosis method according to any embodiment has better optimizing precision and speed, can quickly and accurately identify the fault type of the transformer, improves the intelligent degree and reliability of the fault diagnosis of the transformer, and has higher efficiency and more accurate diagnosis efficiency.
It should be noted that, other embodiments of the present invention further include a power transformer fault diagnosis method and a power transformer fault diagnosis device that are formed by combining the technical features of the foregoing embodiments, and aim at the deficiency of the traditional adaptive fuzzy neural inference system ANFIS, and optimize the system parameters of the ANFIS by using an improved quantum particle swarm algorithm, so as to effectively improve the calculation speed and avoid falling into a locally optimal solution. The power transformer fault diagnosis method based on the improved quantum particle swarm optimization adaptive fuzzy neural inference system can further improve the accuracy of power transformer fault diagnosis. Meanwhile, according to the characteristics of the faults of the transformer, the fault diagnosis method fused by various intelligent algorithms is researched, so that the intelligent degree and reliability of the fault diagnosis of the transformer are improved, and the method has important theoretical research significance and practical value.
The technical features of the above-described embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above-described embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.
The above examples illustrate only a few embodiments of the invention, which are described in detail and are not to be construed as limiting the scope of the invention. It should be noted that it will be apparent to those skilled in the art that several variations and modifications can be made without departing from the spirit of the invention, which are all within the scope of the invention. Accordingly, the scope of protection of the present invention is to be determined by the appended claims.
Claims (6)
1. A power transformer fault diagnosis method, comprising the steps of:
raw gas component volume fraction obtained by analyzing dissolved gas in power transformer oilInitial data, extracting characteristic vectors, inputting the number m of data samples to be diagnosed and the number n of characteristic variables, and establishing an original data matrix X m×n As a fault diagnosis database;
carrying out denoising treatment and standardization treatment on sample data in an original data matrix to obtain a standardization matrix;
according to the standardized matrix, training and adjusting system parameters of a self-adaptive fuzzy neural inference system by using an improved quantum particle swarm algorithm, and establishing a transformer fault diagnosis model;
diagnosing the test data by adopting the transformer fault diagnosis model;
according to the standardized matrix, training and adjusting system parameters of a self-adaptive fuzzy neural reasoning system by using an improved quantum particle swarm algorithm, and establishing a transformer fault diagnosis model, wherein the method comprises the following steps of:
s1, each particle of an improved quantum particle swarm is encoded to generate an initial population, a particle vector is composed of system parameters of a self-adaptive fuzzy neural inference system, and the system parameters comprise front piece parameters a, b, c, d and back piece parameters p, q and r; the resulting MQPSO particle structure was as follows:
x＝[a,b,c,d,p,q,r]
s2, calculating the fitness function f of each particle in the initial population according to the standardized matrix, and recording the historical optimal value p of the individual id And population global optimum p gd The method comprises the steps of carrying out a first treatment on the surface of the The fitness function f is expressed as the root mean square error between the actual and expected output values of the ANFIS, and the formula is as follows:
wherein Y, T is the actual output value and the expected output value of the ANFIS model respectively; m is the MQPSO population scale;
s3, judging whether the current iteration number is greater than the maximum set iteration number N max If yes, executing step S4, otherwise executing step S5;
s4, taking the global optimal solution of the improved quantum particle swarm algorithm as a system parameter of a self-adaptive fuzzy neural inference system network, training and learning the standardized matrix by using the self-adaptive fuzzy neural inference system optimized by the improved quantum particle swarm algorithm, judging whether accuracy is met, if so, building a transformer fault diagnosis model, otherwise, returning to the step S1;
s5, carrying out position update on all particles of the improved quantum particle swarm algorithm according to the individual historical optimal position and the population global optimal position, wherein the update formula is as follows:
p id (t+1)＝a·p id (t)+(1-a)·p gd (t)
wherein p is id A historic optimal position for particle i; p is p gd The global optimal position of the population; m is M best The average optimal position of all particles in the population; x is x id Is the position of the ith particle; m is population scale; t is the current iteration algebra; alpha is the shrinkage expansion coefficient of QPSO; a and b are [0,1 ]]Random numbers uniformly distributed on the interval;
recalculating the fitness function f of all particles of the improved quantum particle swarm algorithm after the position updating, and counting the preset times N of each iteration of the improved quantum particle swarm algorithm c Cumulative algebra N of global optimum continuous arrest of population in (a) pla ，0＜N c ≤N max And 0 < N pla ≤N c Judging cumulative algebra N pla Whether or not it is greater than the continuous stagnation threshold N pla0 Otherwise, returning to the step S3, if yes, according to the variation probability P m Executing mutation operator disturbance population global optimum p gd And average optimum position M best Returning to the execution step S3; wherein, according to the mutation probability P m Executing mutation operator disturbance population global mostOptimal p gd And average optimum position M best The method is realized by adopting the following calculation formula:
p gd (t)＝p gd (t)+μD 1 (g)
M best (t)＝M best (t)+γD 2 (g)
wherein P is m The current variation probability of the population; p (P) 0 Is [0,1]Constants within the range; mu, gamma are all 0,1]Random numbers uniformly distributed in interval D 1 (g)、D 2 (g) Is a random variable subject to the Cauchy distribution.
2. The power transformer fault diagnosis method according to claim 1, wherein the denoising process includes a gaussian filter denoising process.
3. The power transformer fault diagnosis method according to claim 1, wherein the denoising and normalizing the sample data in the raw data matrix comprises: sample data in an original data matrix is subjected to denoising treatment, and then the denoised data matrix is subjected to standardization treatment.
4. The power transformer fault diagnosis method according to claim 1, further comprising the step of: and adjusting the preset times.
5. The power transformer fault diagnosis method according to claim 1, further comprising the step of: and adjusting the mutation probability.
6. The power transformer fault diagnosis method according to claim 1, further comprising the step of: and adjusting the continuous stagnation threshold.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810623024.0A CN109033513B (en) | 2018-06-15 | 2018-06-15 | Power transformer fault diagnosis method and power transformer fault diagnosis device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810623024.0A CN109033513B (en) | 2018-06-15 | 2018-06-15 | Power transformer fault diagnosis method and power transformer fault diagnosis device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109033513A CN109033513A (en) | 2018-12-18 |
| CN109033513B true CN109033513B (en) | 2023-05-19 |
Family
ID=64609923
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810623024.0A Active CN109033513B (en) | 2018-06-15 | 2018-06-15 | Power transformer fault diagnosis method and power transformer fault diagnosis device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109033513B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110108806B (en) * | 2019-04-04 | 2022-03-22 | 广东电网有限责任公司广州供电局 | Transformer oil chromatographic data representation method based on probability information compression |
| CN113051700B (en) * | 2019-12-27 | 2023-12-29 | 北京国双科技有限公司 | Equipment reliability monitoring method and device |
| CN111426816A (en) * | 2020-04-10 | 2020-07-17 | 昆明理工大学 | Method for predicting concentration of dissolved gas in transformer oil based on PSO-L STM |
| CN111682570B (en) * | 2020-06-30 | 2022-02-15 | 广东电网有限责任公司电力科学研究院 | Method, system and equipment for optimizing operation efficiency of power system transformer |
| CN111931423A (en) * | 2020-08-12 | 2020-11-13 | 国网吉林省电力有限公司吉林供电公司 | Power transformer fault diagnosis method based on fuzzy proximity of state data |
| CN112085084B (en) * | 2020-08-24 | 2023-12-15 | 宁波大学 | Transformer fault diagnosis method based on multi-feature fusion common vector |
| CN112347317B (en) * | 2020-10-23 | 2022-07-12 | 四川长虹电器股份有限公司 | Equipment fault diagnosis method based on particle swarm optimization improved negative selection algorithm |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104730378A (en) * | 2015-02-13 | 2015-06-24 | 国家电网公司 | Internal transformer composite-defect fuzzy diagnostic method based on gas dissolved in oil |
| CN106548230A (en) * | 2016-10-14 | 2017-03-29 | 云南电网有限责任公司昆明供电局 | Diagnosis Method of Transformer Faults based on Modified particle swarm optimization neutral net |
-
2018
- 2018-06-15 CN CN201810623024.0A patent/CN109033513B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104730378A (en) * | 2015-02-13 | 2015-06-24 | 国家电网公司 | Internal transformer composite-defect fuzzy diagnostic method based on gas dissolved in oil |
| WO2016127598A1 (en) * | 2015-02-13 | 2016-08-18 | 国家电网公司 | Transformer internal composite defect fuzzy diagnosis method based on gas dissolved in oil |
| CN106548230A (en) * | 2016-10-14 | 2017-03-29 | 云南电网有限责任公司昆明供电局 | Diagnosis Method of Transformer Faults based on Modified particle swarm optimization neutral net |
Non-Patent Citations (3)
| Title |
|---|
| 基于DGA的QPSO-BP模型变压器故障诊断方法研究;程加堂等;《高压电器》;20160216(第02期);全文 * |
| 基于改进粒子群优化神经网络的变压器故障有效诊断;乔维德;《职教与经济研究》;20160630(第02期);全文 * |
| 基于组合核相关向量机和量子粒子群优化算法的变压器故障诊断方法;付华等;《高压电器》;20171016(第10期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109033513A (en) | 2018-12-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109033513B (en) | Power transformer fault diagnosis method and power transformer fault diagnosis device | |
| CN111222290B (en) | Multi-parameter feature fusion-based method for predicting residual service life of large-scale equipment | |
| Yassin et al. | Binary particle swarm optimization structure selection of nonlinear autoregressive moving average with exogenous inputs (NARMAX) model of a flexible robot arm | |
| CN109145516B (en) | Analog circuit fault identification method based on improved extreme learning machine | |
| CN112508243B (en) | Training method and device for multi-fault prediction network model of power information system | |
| CN111428201B (en) | Prediction method for time series data based on empirical mode decomposition and feedforward neural network | |
| CN113962314A (en) | Non-invasive enterprise load decomposition method based on federal learning | |
| CN110442911B (en) | High-dimensional complex system uncertainty analysis method based on statistical machine learning | |
| CN111414943B (en) | Anomaly detection method based on mixed hidden naive Bayes model | |
| Kong et al. | Research on real time feature extraction method for complex manufacturing big data | |
| CN117272102A (en) | Transformer fault diagnosis method based on double-attention mechanism | |
| CN113295399B (en) | Converter valve element state evaluation method, device, electronic equipment and storage medium | |
| CN114648060A (en) | Fault signal standardization processing and classification method based on machine learning | |
| CN110688484A (en) | Microblog sensitive event speech detection method based on unbalanced Bayesian classification | |
| CN117170980A (en) | Early warning method, device, equipment and storage medium for server hardware abnormality | |
| WO2014173271A1 (en) | Optimization method and system for the number of monitoring units of digital man-machine interface | |
| CN116303786B (en) | Block chain financial big data management system based on multidimensional data fusion algorithm | |
| CN110244690B (en) | Multivariable industrial process fault identification method and system | |
| CN110969238A (en) | Method and device for calibrating electricity consumption data | |
| CN116542701A (en) | Carbon price prediction method and system based on CNN-LSTM combination model | |
| JP2016520220A (en) | Hidden attribute model estimation device, method and program | |
| CN115130546A (en) | Abnormal parameter detection method and device, computer readable medium and electronic equipment | |
| CN114707414A (en) | Robot health state assessment method, device, equipment and storage medium | |
| CN113052255A (en) | Intelligent detection and positioning method for reactor | |
| CN112423031A (en) | KPI monitoring method, device and system based on IPTV |
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 | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20200918 Address after: 510620 Tianhe District, Guangzhou, Tianhe South Road, No. two, No. 2, No. Applicant after: Guangzhou Power Supply Bureau of Guangdong Power Grid Co.,Ltd. Address before: 510620 Tianhe District, Guangzhou, Tianhe South Road, No. two, No. 2, No. Applicant before: GUANGZHOU POWER SUPPLY Co.,Ltd. |
|
| TA01 | Transfer of patent application right | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |