CN110110784B - Transformer fault identification method based on transformer related operation data - Google Patents

Transformer fault identification method based on transformer related operation data Download PDF

Info

Publication number
CN110110784B
CN110110784B CN201910363981.9A CN201910363981A CN110110784B CN 110110784 B CN110110784 B CN 110110784B CN 201910363981 A CN201910363981 A CN 201910363981A CN 110110784 B CN110110784 B CN 110110784B
Authority
CN
China
Prior art keywords
data
transformer
weighting
feature
classifier
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201910363981.9A
Other languages
Chinese (zh)
Other versions
CN110110784A (en
Inventor
李诗勇
姜龙
薛静
张霖
丁健
谢荣斌
张丽娟
杨超
施艳
汪德军
靳斌
申峻
杨俊秋
吴冕之
李俊文
何愈杰
余鹏程
张英
赵世钦
潘云
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Guizhou Power Grid Co Ltd
Original Assignee
Guizhou Power Grid Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Guizhou Power Grid Co Ltd filed Critical Guizhou Power Grid Co Ltd
Priority to CN201910363981.9A priority Critical patent/CN110110784B/en
Publication of CN110110784A publication Critical patent/CN110110784A/en
Application granted granted Critical
Publication of CN110110784B publication Critical patent/CN110110784B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/10—Pre-processing; Data cleansing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
    • G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
    • G06V30/10—Character recognition
    • G06V30/19—Recognition using electronic means
    • G06V30/192—Recognition using electronic means using simultaneous comparisons or correlations of the image signals with a plurality of references
    • G06V30/194—References adjustable by an adaptive method, e.g. learning

Abstract

The invention discloses a transformer fault identification method based on transformer related operation data, which comprises the following steps of: 1) reading transformer fault information; 2) numbering the data of the transformer fault information according to the data name; 3) preprocessing the data; 4) obtaining a weight coefficient by using a principal component analysis method; 5) training data by using a weighting support vector machine method to obtain a classifier model; 6) weighting the support vector machine; 7) inputting relevant operation data and fault types of the training classifier into the classifier obtained in the step 6), continuously inputting test data for parameter adjustment and improvement, and finally inputting data required to be predicted for fault prediction; according to the transformer fault identification method, more related operation data are added, reasonable training is conducted on the classifier according to the influence weight of the data on fault identification, the parameters and the accuracy of the classifier are adjusted by inputting related training data according to requirements, and the transformer fault identification is more accurate.

Description

Transformer fault identification method based on transformer related operation data
Technical Field
The invention belongs to the technical field of power transformer state evaluation, and particularly relates to a transformer fault identification method based on transformer related operation data.
Background
The reliability of the electrical equipment directly affects the safe operation of the power system. However, because the transformer is a closed whole integrating multiple disciplinary technologies such as mechanical, electrical, chemical and thermodynamic technologies, the reasons for influencing the fault are complicated, the fault diagnosis needs multiple data and knowledge, and the subjective underground conclusion on one or more aspects inevitably leads to misjudgment or missed judgment. The current common fault diagnosis methods are a three-ratio judgment method, an over-temperature discharge diagram judgment method, an HAE triangular diagram judgment method, a characteristic gas method and the like. Each single diagnostic method, although highly specific for certain faults, has some disadvantages. The defects are mainly reflected in that the consideration factor is single, the accuracy rate is reduced, and in addition, the set threshold value is not accurate enough due to different operation factors, and the defects are not favorable for identifying the fault of the transformer. Therefore, the transformer fault identification method which can consider various influence factors and adjust parameters according to the operation conditions is required by the operation and maintenance and fault diagnosis of the transformer at the present stage.
Disclosure of Invention
The technical problem to be solved by the invention is as follows: the transformer fault identification method based on the related operation data of the transformer is provided to solve the problems in the prior art.
The technical scheme adopted by the invention is as follows: a transformer fault identification method based on transformer related operation data comprises the following steps:
reading transformer fault diagnosis information including transformer oil test data, insulation oil data and fault data from a database server; the specific data content is as shown in table 1:
numbering the data types of the transformer oil test data, the insulation test data and the insulation oil data according to data names, and respectively setting the transformer oil test data, the insulation test data and the insulation oil data to A, B, C, wherein the data types are respectively according to A1～An、B1～Bm、C1～CkNumbering; setting the fault type of the transformer to be F, and pressing the fault type of the transformer to be F1～FtNumbering;
step three, data preprocessing is carried out on the data with the serial number A, B, C, namely missing value filling and denoising processing are carried out on the data, and then normalization processing is carried out on the data according to the type;
fourthly, carrying out weight analysis on different data types of data by using a principal component analysis method to obtain weight coefficients, and solving a characteristic root Evalaue and a characteristic vector Eectror according to the covariance matrix, wherein the value after the characteristic root normalization is the weight omega of the relevant operating data of each transformerjiI.e. the contribution rate of each feature root, is similarly applicable to a variety of data (a1, a2 … An) of different types (A, B, C);
fifthly, training data by using a weighting support vector machine method to obtain a classifier model;
step six, weighting the support vector machine in the step five;
and step seven, inputting the relevant operation data (A, B, C) of the transformer used by the training classifier and the corresponding fault type (F) into the weighting support vector machine obtained in the step six to obtain the classifier, continuously inputting test data to adjust and improve the parameters of the classifier, and finally inputting the data required to be predicted to predict the fault.
The invention has the beneficial effects that: compared with the prior art, the transformer fault identification method provided by the invention has the advantages that more related operation data are added, the classifier is reasonably trained according to the influence weight of the data on fault identification, and the related training data can be input according to requirements to adjust the parameters and accuracy of the classifier, so that the transformer fault is reasonably identified finally.
Drawings
FIG. 1 is a flow chart of a transformer fault identification method based on transformer related operational data;
FIG. 2 is a flow chart of a principal component analysis acquisition weight method;
FIG. 3 is a schematic diagram of a support vector machine;
fig. 4 is a flow chart of a classifier based on a weighted support vector machine.
Detailed Description
The invention is further described with reference to the accompanying drawings and specific embodiments.
Example (b): as shown in fig. 1 to 4, a transformer fault identification method based on transformer-related operation data includes the following steps:
reading transformer fault diagnosis information including transformer oil test data, insulation oil data and fault data from a database server; the specific data content is as shown in table 1:
TABLE 1 Transformer related data
Figure GDA0002335718160000031
Numbering the data types of the transformer oil test data, the insulation test data and the insulation oil data according to data names, and respectively setting the transformer oil test data, the insulation test data and the insulation oil data to A, B, C, wherein the data types are respectively according to A1～An、B1～Bm、C1～CkNumbering; setting the fault type of the transformer to be F, and pressing the fault type of the transformer to be F1～FtNumbering as shown in table 2;
TABLE 2 Transformer data numbering case
Figure GDA0002335718160000032
Figure GDA0002335718160000041
Step three, data preprocessing is carried out on the data with the serial number A, B, C, namely missing value filling and denoising processing are carried out on the data, and then normalization processing is carried out on the data according to the type;
the normalization processing process is explained by A1 data, and n pieces of data are set;
Figure GDA0002335718160000042
Figure GDA0002335718160000043
i is the number of data collected for each type of data,
Figure GDA0002335718160000044
is A1Average value of Medium data, A'1iIs A1Carrying out normalization processing on the data;
step four, carrying out weight analysis on the data of different data types by using a principal component analysis method to obtain a weight coefficient, wherein a flow chart of a method for obtaining the weight by the principal component analysis is shown in fig. 2, and the principal component analysis method is defined as follows:
(1) arranging original data according to rows to form a matrix M;
(2) carrying out data standardization on M to enable the mean value of M to become zero;
(3) solving a covariance matrix Cov of M;
(4) obtaining a feature vector T and a feature root M of the feature vector by using Cov;
the contribution rate V of each characteristic root is calculated by the following formulai；Vi＝Mi/(M1+M2+.....) the contribution rate is a weighting factor for each data type.
Where M is the original data matrix for all A, B, C sample data combinations:
M＝{A1，A2，…，A5，B1，B2，…，B8，C1，C2，…，C6}
as can be seen from the above, the first two steps in the step (4) are completed in the step (3), and then the covariance matrix Cov of the M matrix formed by the data is obtained as follows (assuming a is provided)1、A2、A3Three sets of data):
Figure GDA0002335718160000051
the diagonal of the matrix is A1、A2、A3Rather than the diagonal being covariance. Covariance is a measure of the degree of change in which two variables change simultaneously. A covariance greater than 0 means that if one of the two quantities increases, the other increases; less than 0 indicates one increase and one decrease. If the two quantities are statistically independent, then the covariance between the two is 0; but the covariance is 0 and does not indicate that the two quantities are independent. The larger the absolute value of the covariance is, the larger the influence of the two on each other is, and the smaller the influence is otherwise;
solving a characteristic root M and a characteristic vector T according to the covariance matrix, wherein the value after the characteristic root is normalized is the weight omega of the relevant operation data of each transformerjiI.e. contribution rate of each feature root, the same appliesA plurality of data (A1, A2 … An) at different types (A, B, C);
step five, training data by using a weighted support vector machine method to obtain a classifier model, wherein the support vector machine is defined as follows:
a schematic diagram of a support vector machine is shown in fig. 3, the support vector machine (support vector machines) is a binary model, and aims to find a hyperplane to segment a sample, and the segmentation principle is interval maximization, and finally is converted into a convex quadratic programming problem to solve. When the training samples are linearly separable, learning a linearly separable support vector machine through hard interval maximization; when the training samples are linearly irretrievable, a nonlinear support vector machine is learned through the maximization of kernel skills and soft intervals, and a classifier model is as follows:
Figure GDA0002335718160000052
Figure GDA0002335718160000061
in the formula: k is a radical ofiIs hyperplane normal vector, C is penalty factor, n is sample number, ξiIs a relaxation factor and represents the allowable error rate under the linear irreducible condition; y isiIs the sample output, and yi∈{-1,1}；xiThe sample input quantity is; b is a threshold value;
and (3) introducing a Lagrange multiplier algorithm to solve the problem to obtain an optimized objective function:
Figure GDA0002335718160000062
α in the above formulai、αjIs Lagrange multiplier, xi、xjAs sample input amount, yi、yjFor sample output, the following equation is a constraint.
Figure GDA0002335718160000063
SVM by introducing non-linear mapping
Figure GDA0002335718160000064
Rn→ H, map the samples to a new data set
Figure GDA0002335718160000065
Can transform the optimized objective function into
Figure GDA0002335718160000066
In the formula
Figure GDA0002335718160000067
The kernel function is selected as follows through comparison tests:
Figure GDA0002335718160000068
in the formula xi、xjFor the sample input, γ is the radial basis kernel function vector whose value determines the classification accuracy of the support vector machine.
Step six, weighting the support vector machine in the step five, wherein the definition of the weighted support vector machine and the k weighting steps are as follows:
the characteristic weighting is given to each characteristic in the data set according to the criterion, the characteristic weighting is called as characteristic weighting, the performance of the algorithm can be improved by applying the characteristic weighting, and a formula for expanding the standard Euclidean distance by utilizing a characteristic weight vector omega is as follows:
Figure GDA0002335718160000071
wherein d isω(xi,xj) Representing two samples xiAnd xjWeighted euclidean distance of, xikRepresents a sample xiOf the kth feature, ω ═ ω (ω ═ ω)1，ω2…ωn) Is a weight vector, ωk0(k ═ 1, … n) is the importance weight corresponding to each feature;
the support vector machine constructed based on the feature weighting kernel function is called a feature weighting support vector machine, and the feature weighting kernel function is defined as follows:
let K be the kernel function defined at X,
Figure GDA0002335718160000072
p is an n-th order linear transformation matrix for a given input space, where n is the dimension of the input space, a feature weighting kernel function KpIs defined as
Figure GDA0002335718160000073
The linear transformation matrix P, also called the feature weighting matrix, is of the form:
Figure GDA0002335718160000074
omega in the above formulajiThe weight of each type of data obtained in the step (3) and the step (4) is obtained;
the weighting form of the kernel function, i.e. the radial basis function weighting form, is selected as follows:
Figure GDA0002335718160000075
namely, it is
Kp(xi,xj)＝exp(-γ((xi-xj)TPPT(xi-xj))2)；
And step seven, inputting the relevant operation data (A, B, C) of the transformer used by the training classifier and the corresponding fault type (F) into the weighting support vector machine obtained in the step six to obtain the classifier, continuously inputting test data to adjust and improve the parameters of the classifier, finally inputting the data required to be predicted to predict the fault of the classifier, and obtaining a flow chart of the classifier based on the weighting support vector machine as shown in fig. 4.
The transformer fault identification method can add more related operation data, reasonably train the classifier according to the influence weight of the data on fault identification, and input related training data according to requirements to adjust the parameters and the accuracy of the classifier, thereby finally playing a role in reasonably identifying the transformer fault.
The invention can also be combined with other functions required by users, such as the acquisition and storage functions of related data, the fault early warning function and the like.
The conditions that influence factors related to transformer faults are too many, the types of the transformer faults are not distinguished obviously and the like are considered, the weight analysis can be carried out on the related operation data of the transformer, and a support vector machine in a weighting form is adopted for classification. The method has the advantages that the weight adjustment is carried out according to actual data, the fault types are intelligently classified by a support vector machine, and the like, namely, the classification factors are not only single threshold judgment, but also related data are mapped to a higher-dimensional space for reasonable classification. The method can provide guidance for early warning and maintenance of the transformer, can reasonably allocate the operation of the transformer, moderately reduces the load rate of the early warning transformer under possible conditions, arranges the maintenance of the early warning transformer as soon as possible, and has good application prospect.
The above description is only an embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can easily conceive of changes or substitutions within the technical scope of the present invention, and therefore, the scope of the present invention should be determined by the scope of the claims.

Claims (1)

1. A transformer fault identification method based on transformer related operation data is characterized in that: the method comprises the following steps:
reading transformer fault diagnosis information including transformer oil test data, insulation oil data and fault data from a database server;
numbering the data types of the transformer oil test data, the insulation test data and the insulation oil data according to data names, and respectively setting the transformer oil test data, the insulation test data and the insulation oil data to A, B, C, wherein the data types are respectively according to A1～An、B1～Bm、C1～CkNumbering; setting the fault type of the transformer to be F, and pressing the fault type of the transformer to be F1～FtNumbering;
step three, missing value filling and denoising are carried out on the data with the serial number A, B, C, and then normalization processing is carried out on the data according to the types;
fourthly, carrying out weight analysis on different data types of data by using a principal component analysis method to obtain weight coefficients, and solving a characteristic root Evalaue and a characteristic vector Eectror according to the covariance matrix, wherein the value after the characteristic root normalization is the weight omega of the relevant operation data of each transformerjiI.e. the contribution rate of each feature root;
fifthly, training data by using a weighting support vector machine method to obtain a classifier model;
the classifier model is as follows:
Figure FDA0002335718150000011
Figure FDA0002335718150000012
in the formula: k is a radical ofiIs hyperplane normal vector, C is penalty factor, n is sample number, ξiIs a relaxation factor and represents the allowable error rate under the linear irreducible condition; y isiIs the sample output, and yi∈{-1,1}；xiThe sample input quantity is; b is a threshold value;
introducing a Lagrange multiplier algorithm for solving to obtain an optimized objective function:
Figure FDA0002335718150000021
α in the above formulai、αjIs Lagrange multiplier, xi、xjAs sample input amount, yi、yjTaking the output quantity of the sample, and taking the following formula as a constraint condition;
Figure FDA0002335718150000022
SVM by introducing non-linear mapping
Figure FDA0002335718150000023
Rn→ H, map the samples to a new data set
Figure FDA0002335718150000024
Transforming an optimization objective function into
Figure FDA0002335718150000025
In the formula
Figure FDA0002335718150000026
The kernel function is selected as follows through comparison tests:
K(xi,xj)＝exp(-γ||xi-xj||2)
in the formula xi、xjTaking the sample input quantity as gamma, and taking the gamma as a radial basis kernel function vector;
step six, weighting the support vector machine in the step five;
wherein the steps of defining and k-weighting the weighted support vector machine are as follows:
giving a certain weight to each feature in the data set according to the criterion is called feature weighting, and the formula for expanding the standard Euclidean distance by using the feature weight vector omega is as follows:
Figure FDA0002335718150000027
wherein d isω(xi,xj) Representing two samples xiAnd xjWeighted euclidean distance of, xikRepresents a sample xiOf the kth feature, ω ═ ω (ω ═ ω)1，ω2…ωn) Is a weight vector, ωk0(k ═ 1, … n) is the importance weight corresponding to each feature;
the support vector machine constructed based on the feature weighting kernel function is called a feature weighting support vector machine, and the feature weighting kernel function is defined as follows:
let K be the kernel function defined at X,
Figure FDA0002335718150000031
p is an n-th order linear transformation matrix for a given input space, where n is the dimension of the input space, a feature weighting kernel function KpIs defined as
Figure FDA0002335718150000032
The linear transformation matrix P, also called the feature weighting matrix, is of the form:
Figure FDA0002335718150000033
omega in the above formulajiThe weight of each type of data obtained in the step (3) and the step (4) is obtained;
the weighting form of the kernel function, i.e. the radial basis function weighting form, is selected as follows:
Figure FDA0002335718150000034
namely, it is
Kp(xi,xj)＝exp(-γ((xi-xj)TPPT(xi-xj))2)；
And step seven, inputting the relevant operation data of the transformer used by the training classifier and the fault type corresponding to the relevant operation data into the weighting support vector machine obtained in the step six to obtain the classifier, continuously inputting test data to adjust and improve the parameters of the classifier, and finally inputting the data required to be predicted to predict the fault.
CN201910363981.9A 2019-04-30 2019-04-30 Transformer fault identification method based on transformer related operation data Active CN110110784B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910363981.9A CN110110784B (en) 2019-04-30 2019-04-30 Transformer fault identification method based on transformer related operation data

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910363981.9A CN110110784B (en) 2019-04-30 2019-04-30 Transformer fault identification method based on transformer related operation data

Publications (2)

Publication Number Publication Date
CN110110784A CN110110784A (en) 2019-08-09
CN110110784B true CN110110784B (en) 2020-03-24

Family

ID=67487930

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910363981.9A Active CN110110784B (en) 2019-04-30 2019-04-30 Transformer fault identification method based on transformer related operation data

Country Status (1)

Country Link
CN (1) CN110110784B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111638427B (en) * 2020-06-03 2021-05-28 西南交通大学 Transformer fault detection method based on nuclear capsule neuron coverage
CN115293383A (en) * 2022-10-09 2022-11-04 西南石油大学 Game theory fused transformer risk cause analysis method

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105629109A (en) * 2015-12-30 2016-06-01 西安工程大学 ARTI-neural network-based oil-immersed transformer fault diagnosis method
CN107831300A (en) * 2017-10-20 2018-03-23 广东电网有限责任公司河源供电局 A kind of transformer insulation oil based on three-dimensional trapezoidal Probabilistic Fuzzy collection deteriorates appraisal procedure
CN108680811A (en) * 2018-06-29 2018-10-19 广东工业大学 A kind of transformer fault state evaluating method
CN109239516A (en) * 2018-09-05 2019-01-18 国网山西省电力公司检修分公司 Merge the transformer fault layering diagnostic method of a variety of intelligent diagnostics models
CN109669087A (en) * 2019-01-31 2019-04-23 国网河南省电力公司 A kind of method for diagnosing fault of power transformer based on Multi-source Information Fusion

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105335759B (en) * 2015-11-12 2019-07-16 南方电网科学研究院有限责任公司 A kind of transformer fault detection method based on generating probability model
CN106934421B (en) * 2017-03-16 2020-11-06 山东大学 Power transformer fault detection method based on 2DPCA and SVM
CN109030790A (en) * 2018-08-21 2018-12-18 华北电力大学（保定） A kind of method for diagnosing fault of power transformer and device
CN109164343B (en) * 2018-08-30 2020-11-06 西华大学 Transformer fault diagnosis method based on characteristic information quantization and weighted KNN
CN109670676A (en) * 2018-11-26 2019-04-23 安徽继远软件有限公司 Distributing net platform region method for prewarning risk and system based on Support Vector data description

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105629109A (en) * 2015-12-30 2016-06-01 西安工程大学 ARTI-neural network-based oil-immersed transformer fault diagnosis method
CN107831300A (en) * 2017-10-20 2018-03-23 广东电网有限责任公司河源供电局 A kind of transformer insulation oil based on three-dimensional trapezoidal Probabilistic Fuzzy collection deteriorates appraisal procedure
CN108680811A (en) * 2018-06-29 2018-10-19 广东工业大学 A kind of transformer fault state evaluating method
CN109239516A (en) * 2018-09-05 2019-01-18 国网山西省电力公司检修分公司 Merge the transformer fault layering diagnostic method of a variety of intelligent diagnostics models
CN109669087A (en) * 2019-01-31 2019-04-23 国网河南省电力公司 A kind of method for diagnosing fault of power transformer based on Multi-source Information Fusion

Also Published As

Publication number Publication date
CN110110784A (en) 2019-08-09

Similar Documents

Publication Publication Date Title
CN110132598B (en) Fault noise diagnosis algorithm for rolling bearing of rotating equipment
Fan et al. Deep learning-based feature engineering methods for improved building energy prediction
Hassani et al. Forecasting European industrial production with singular spectrum analysis
CN108898251B (en) Offshore wind farm power prediction method considering meteorological similarity and power fluctuation
CN108051660A (en) A kind of transformer fault combined diagnosis method for establishing model and diagnostic method
CN111860658A (en) Transformer fault diagnosis method based on cost sensitivity and integrated learning
Li et al. A new flood forecasting model based on SVM and boosting learning algorithms
CN109669009A (en) The adaptive extreme learning machine electronic nose drift rejection method in domain based on domain correction
CN112507479B (en) Oil drilling machine health state assessment method based on manifold learning and softmax
CN110110610B (en) Event detection method for short video
CN110110784B (en) Transformer fault identification method based on transformer related operation data
CN108830417B (en) ARMA (autoregressive moving average) and regression analysis based life energy consumption prediction method and system
CN111178585A (en) Fault reporting amount prediction method based on multi-algorithm model fusion
CN114297918A (en) Aero-engine residual life prediction method based on full-attention depth network and dynamic ensemble learning
Wang et al. A novel sliding window PCA-IPF based steady-state detection framework and its industrial application
CN111079861A (en) Power distribution network voltage abnormity diagnosis method based on image rapid processing technology
Bo et al. An integrated method of independent component analysis and support vector machines for industry distillation process monitoring
CN112633556A (en) Short-term power load prediction method based on hybrid model
CN115438726A (en) Device life and fault type prediction method and system based on digital twin technology
CN112330153A (en) Non-linear orthogonal regression-based industry scale prediction model modeling method and device
CN115017970A (en) Migration learning-based gas consumption behavior anomaly detection method and system
CN114266297A (en) Semantic knowledge base of thermal power equipment, construction method and zero sample fault diagnosis method
Li et al. A CM&CP framework with a GIACC method and an ensemble model for remaining useful life prediction
Wang et al. Combining autoencoder with similarity measurement for aircraft engine remaining useful life estimation
Ghassemi et al. Optimal surrogate and neural network modeling for day-ahead forecasting of the hourly energy consumption of university buildings

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