CN105843212A - System and method for fault diagnosis of blast furnace - Google Patents
System and method for fault diagnosis of blast furnace Download PDFInfo
- Publication number
- CN105843212A CN105843212A CN201610187770.0A CN201610187770A CN105843212A CN 105843212 A CN105843212 A CN 105843212A CN 201610187770 A CN201610187770 A CN 201610187770A CN 105843212 A CN105843212 A CN 105843212A
- Authority
- CN
- China
- Prior art keywords
- blast furnace
- blast
- fault diagnosis
- attribute data
- hypersphere
- 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
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B23/00—Testing or monitoring of control systems or parts thereof
- G05B23/02—Electric testing or monitoring
- G05B23/0205—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults
- G05B23/0259—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterized by the response to fault detection
Abstract
The present invention provides a system and a method for the fault diagnosis of a blast furnace. The system comprises a historical data acquisition module, an actual data acquisition module, a feature weight matrix construction module, a model building module and a blast furnace fault diagnosis module. The method includes the steps of collecting the actual attribute data, the historical attribute data and the corresponding fault state types of the production condition of a blast furnace; according to the importance degree of each attribute for the fault diagnosis, determining the feature weight of the attribute and constructing a feature weight matrix; establishing a twin hyper-sphere support vector machine model for the feature weighting of the fault diagnosis of the blast furnace; taking the actual attribute data of the production condition of the blast furnace into the above established twin hyper-sphere support vector machine model to obtain the operation fault state type of the blast furnace that corresponds to the actual attribute data of the production condition of the blast furnace; and completing the fault diagnosis of the blast furnace. According to the technical scheme of the invention, the importance degree of each attribute for the fault diagnosis of the blast furnace is quantized. Meanwhile, the importance degree of each attribute is integrated into the constructing process of a learning machine. Therefore, the accuracy of the fault diagnosis is improved.
Description
Technical field
The invention belongs to blast furnace fault diagnosis technology field, be specifically related to a kind of blast furnace fault diagnosis system and method.
Background technology
Blast furnace fault fault diagnosis has great importance for minimizing accident and economic loss.Blast furnace process is pig iron Ore also
Former one-tenth ferrum, is a continuous and production process for complex process；For preventing abnormal conditions from occurring, need in production process is a large amount of
Parameter is monitored, such as, hot blast pathogenic wind-warm, hot blast air quantity, hot blast blast, furnace top pressure, total head are poor, upper pressure, under
Portion's pressure, Rich Oxygen Amount, permeability index, cross temperature, material speed, physical thermal, containing [Si] amount etc.；Therefore the feature of malfunction
It it is the comprehensive embodiment of the signal of high-dimensional multiple features；Therefore, traditional method for diagnosing faults is applied to blast furnace fault diagnosis effect not
It is the most preferable.Method for diagnosing faults based on machine learning and artificial intelligence, owing to not relying on accurate mathematical model, becomes high
The effective ways of stove fault diagnosis.At present, the artificial intelligence approach of blast furnace fault diagnosis be mainly based upon neutral net and support to
Amount machine.Support vector machine, due to its good general magnificent performance, the especially good behaviour of small sample identification problem, becomes blast furnace event
Barrier diagnostic techniques provides new approach to intelligent development.
Twin support vector machine is different from traditional support vector machine, and it, by solving two quadratic programming problems on a small scale, constructs one
To non-parallel planes, rather than as the parallel plane in support vector machine, test result indicate that, on the premise of ensureing precision, learn
Practise efficiency and improve 4 times.Twin Hypersphere Support Vector Machine is the up-to-date improvement of twin support vector machine, and this model is by solving two
Quadratic programming problem on a small scale, constructs two non-concentric hyperspheres in feature space, and each hypersphere comprises a class sample as far as possible,
And away from another kind of sample, similar with twin support vector machine, due to this model solution two quadratic programming problem on a small scale, and
A non-extensive quadratic programming problem, therefore has the higher learning efficiency；Compared to twin support vector machine, twin hypersphere
Support vector machine avoids the computing of two extensive matrix inversions in twin support vector machine, this make twin hypersphere support to
Amount facility have the more efficient learning efficiency；Owing to twin Hypersphere Support Vector Machine uses a pair hypersphere, rather than a pair not parallel super flat
Face, describes two class sample characteristics, and in actual applications, this strategy is more reasonable.Test result indicate that, with support vector machine
Comparing with twin support vector machine, twin Hypersphere Support Vector Machine has the more quickly learning efficiency and nicety of grading, for blast furnace
Fault diagnosis technology provides new thinking.
In blast furnace fault diagnosis, each dimensional feature of fault sample acts on the most different for classification, and difference is the most sometimes
Greatly.Traditional disaggregated model, including support vector machine, twin support vector machine, twin Hypersphere Support Vector Machine, divides building
The difference impact for classification of each feature importance is all have ignored during class model.
Summary of the invention
The problem existed for prior art, the present invention provides a kind of blast furnace fault diagnosis system and method.
The technical scheme is that
A kind of blast furnace fault diagnosis system, including:
Historical data acquisition module: gather blast fumance life history attribute data and the operation of blast furnace malfunction type of correspondence thereof；
Real data acquisition module: gather blast fumance situation actual attribute data；
Feature weight matrix construction module: according to each attribute, the significance level of fault diagnosis is determined to the weight of each attribute, structure
Feature weight matrix；
Model building module: utilize feature weight matrix, raw to any two operation of blast furnace malfunction type and corresponding blast furnace thereof
Occurrence condition historical status data are trained, and set up characteristic weighing twin Hypersphere Support Vector Machine model, and the input of this model is that blast furnace is raw
Occurrence condition historical status data, are output as one of these two operation of blast furnace malfunction types, a pair characteristic weighing in this model
Hypersphere represents this two operation of blast furnace malfunction types respectively, and blast fumance life history attribute data to be sorted is from characteristic weighing
Hypersphere is the nearest, and the probability that corresponding operation of blast furnace malfunction occurs is the highest；
Blast furnace fault diagnosis module: blast fumance situation actual attribute data are brought into each characteristic weighing twin hypersphere support of foundation
In vector machine model, it is thus achieved that operation of blast furnace malfunction type belonging to blast fumance situation actual attribute data, complete blast furnace fault
Diagnosis.
Described blast furnace fault diagnosis system, also includes normalized module, and the blast furnace gathering historical data acquisition module is raw
Occurrence condition historical status data and the operation of blast furnace malfunction type of correspondence, the blast fumance of real data acquisition module collection
Situation actual attribute data are normalized.
Described attribute, including: air quantity, blast, top pressure, pressure reduction, breathability, top temperature, cross temperature, material speed, Si content,
Physical thermal；Described malfunction type, including: Xiang Liang, thermotropism, hanging, collapse material.
The feature weight of described attribute is that comentropy and this attribute of malfunction type breaks down the difference of comentropy of state.
System described in utilization carries out the method for blast furnace fault diagnosis, including:
Gather blast fumance situation actual attribute data, historical status data and the operation of blast furnace malfunction type of correspondence thereof；
According to each attribute, the significance level of fault diagnosis is determined to the feature weight of each attribute, structural feature weight matrix；
Set up the characteristic weighing twin Hypersphere Support Vector Machine model for blast furnace fault diagnosis: utilize feature weight matrix, to appointing
Two operation of blast furnace malfunction types of anticipating and corresponding blast fumance life history attribute data training thereof, set up characteristic weighing twin
Hypersphere Support Vector Machine model, the input of this model is blast fumance life history attribute data, is output as this two operation of blast furnace
One of malfunction type, a pair characteristic weighing hypersphere in this model represents this two operation of blast furnace malfunction types respectively,
Blast fumance life history attribute data to be sorted from characteristic weighing hypersphere more close to, there is the probability of corresponding operation of blast furnace malfunction
The highest；
Blast furnace fault diagnosis: the twin hypersphere of each characteristic weighing that blast fumance situation actual attribute data are brought into foundation supports vector
In machine model, it is thus achieved that operation of blast furnace malfunction type belonging to blast fumance situation actual attribute data, complete blast furnace fault diagnosis.
Operation of blast furnace malfunction type belonging to described acquisition blast fumance situation actual attribute data is: blast fumance situation is real
Border attribute data substitutes in each characteristic weighing twin Hypersphere Support Vector Machine model, adds up the twin hypersphere of each characteristic weighing and supports vector
The operation of blast furnace malfunction type of machine model output, the most operation of blast furnace malfunction type of occurrence number is blast fumance
Operation of blast furnace malfunction type belonging to situation actual attribute data.
Beneficial effect:
Each for blast furnace fault feature importance is quantified by the present invention, and is incorporated by each feature importance in the building process of learning machine, with
Improve the precision of fault diagnosis.The present invention is on the basis of twin Hypersphere Support Vector Machine, by each feature importance i.e. characteristic weighing square
Battle array is incorporated in twin Hypersphere Support Vector Machine, and uses the method for information gain to calculate feature weight, is ensureing fault diagnosis instruction
Practice on the basis of the time, it is to avoid failure diagnostic process is affected by some weak relevant or incoherent features, has more outstanding
Performance of fault diagnosis.
Accompanying drawing explanation
Fig. 1 is specific embodiment of the invention blast furnace fault diagnosis system block diagram；
Fig. 2 is specific embodiment of the invention blast furnace method for diagnosing faults flow chart.
Detailed description of the invention
Below in conjunction with the accompanying drawings the detailed description of the invention of the present invention is elaborated.
In present embodiment, attribute data includes: air quantity (m3/min), blast (Pa), top pressure (MPa), pressure reduction, breathability,
Top temperature (comprising 4 temperature), cross temperature (comprise center and peripheral, unit be DEG C), material speed (unit is batch/hour), Si
Content, physical thermal (unit be DEG C).Malfunction type, including: Xiang Liang, thermotropism, hanging, collapse material.
A kind of blast furnace fault diagnosis system, as it is shown in figure 1, include:
Historical data acquisition module: gather blast fumance life history attribute data and the operation of blast furnace malfunction type of correspondence thereof.
Real data acquisition module: gather blast fumance situation actual attribute data.
Feature weight matrix construction module: according to each attribute, the significance level of fault diagnosis is determined to the feature weight of each attribute,
Structural feature weight matrix.
Model building module: utilize feature weight matrix, raw to any two operation of blast furnace malfunction type and corresponding blast furnace thereof
Occurrence condition historical status data are trained, and set up characteristic weighing twin Hypersphere Support Vector Machine model, and the input of this model is that blast furnace is raw
Occurrence condition historical status data, are output as one of these two operation of blast furnace malfunction types, a pair characteristic weighing in this model
Hypersphere represents that this two operation of blast furnace malfunction types, blast fumance life history attribute data distance feature to be sorted add respectively
Power hypersphere is the nearest, and the probability that corresponding operation of blast furnace malfunction occurs is the highest.
Blast furnace fault diagnosis module: blast fumance situation actual attribute data are brought into each characteristic weighing twin hypersphere support of foundation
In vector machine model, it is thus achieved that operation of blast furnace malfunction type belonging to blast fumance situation actual attribute data, complete blast furnace fault
Diagnosis.
Normalized module: the blast fumance life history attribute data that historical data acquisition module is gathered and the height of correspondence thereof
Stove operation troubles Status Type, the blast fumance situation actual attribute data of real data acquisition module collection are normalized.
Utilize the method that said system carries out blast furnace fault diagnosis, as in figure 2 it is shown, include:
Step 1, collection blast fumance situation actual attribute data, historical status data and the operation of blast furnace malfunction of correspondence thereof
Type；
Present embodiment use Mean-Variance standardized method the data gathered are normalized, and handmarking's sample
This, using the operation of blast furnace malfunction type of the blast fumance life history attribute data gathered and correspondence as training sample set
T={ (x1, y1), (x2, y2) ..., (xn, ym), wherein xi∈ X=RdRepresent i-th historical status data, i.e. i-th training sample
Data vector, in the present embodiment, n value is 14；yi∈ Y={1,2 ..., m} represents the fault shape that i-th training sample is corresponding
State type, m represents failure mode, and value is integer, m=4；RdRepresenting sample space, d represents the intrinsic dimensionality of sample；
Malfunction type totally four class of present embodiment, is to cool, thermotropism, hanging respectively, collapses material；
800 training samples are amounted to, by Gauss distribution N (μ to cool malfunction type1, c1) generate, wherein, it is desirable to
Covariance matrix
Thermotropism malfunction type amounts to 1000 training samples, and wherein 500 by Gauss distribution N (μ21, c21) generate, wherein,
ExpectCovariance matrixRemaining is by N (μ22, c22) generate, wherein, it is desirable toAssociation
Variance matrix
Hanging malfunction type amounts to 500 training samples, by Gauss distribution N (μ3, c3) generate, wherein, it is desirable to
Covariance matrix
Collapse material malfunction type and amount to 400 training samples, by Gauss distribution N (μ4, c4) generate, wherein, it is desirable to
Covariance matrix.
Step 2, according to each attribute, the significance level of fault diagnosis is determined to the feature weight of each attribute, structural feature weight square
Battle array；
Feature weight is the biggest, and attribute is the most important for fault diagnosis, according to the feature weight structural feature weight matrix of each attribute；
The feature weight of attribute is that comentropy and this attribute of malfunction type breaks down the difference of comentropy of state.
Feature weight matrix
Wherein,The comentropy of the malfunction type that attribute is corresponding
For malfunction type, P (Ci) it is malfunction Type CiOccur
Probability；It is characterized ApValue be aiSample set.
Step 3, foundation are used for the characteristic weighing twin Hypersphere Support Vector Machine model of blast furnace fault diagnosis: utilize feature weight square
Battle array, to any two operation of blast furnace malfunction type and corresponding blast fumance life history attribute data training thereof, sets up feature
Weighting twin Hypersphere Support Vector Machine model, the input of this model is blast fumance life history attribute data, is output as these two
One of operation of blast furnace malfunction type, a pair characteristic weighing hypersphere in this model represents this two operation of blast furnace fault shapes respectively
State type, blast fumance life history attribute data to be sorted from characteristic weighing hypersphere more close to, there is corresponding operation of blast furnace fault shape
Probability of state is the highest；
Use between i-th class blast furnace malfunction and the jth class blast furnace malfunction of characteristic weighing twin Hypersphere Support Vector Machine structure
Classifying face, this classifying face is determined by a pair characteristic weighing hypersphere, in order to solve the inseparable problem of hypersphere in former space, will instruction
Practice sample by a certain conversionIt is mapped in high-dimensional feature space Z so that training sample hypersphere in feature space can divide.
In high-dimensional feature space, introduced feature weight matrix, define characteristic weighing hypersphere and the jth class malfunction of the i-th class malfunction
Characteristic weighing hypersphere, respectively as follows:
With
Wherein,It is characterized weight matrix, ci(ij)And Ri(ij)It is respectively the i-th class event in high-dimensional feature space
The centre of sphere of barrier state and radius, cj(ij)And Rj(ij)It is respectively the centre of sphere and the radius of jth class malfunction in high-dimensional feature space.
In high-dimensional feature space, the centre of sphere and the radius of two characteristic weighing hyperspheres determine by solving following quadratic programming problem:
Wherein, ξi(ij)And ξj(ij)For relaxation factor, it is respectively used to retrain positive and negative class singular point, l±For the number of positive and negative class sample, c1、
c2、v1、v2For penalty factor, cross validation method is used to determine.
I(i)=k | yk=i}, I(j)=k | yk=j} represents subscript and the jth of attribute data corresponding to the i-th class malfunction respectively
The subscript of the attribute data that class malfunction is corresponding.
From formula (1) with (2) it can be seen that required characteristic weighing hypersphere surrounds the genus that the i-th class malfunction is corresponding as far as possible
Property data, and away from attribute data corresponding to jth class malfunction, wherein, the biggest attribute of feature weight is for the meter of distance
Calculate the most important, and the contribution that the least attribute of feature weight calculates for distance is the least；Require required characteristic weighing hypersphere simultaneously
Radius is the least so that characteristic weighing hypersphere is the tightest.
For convenience, not direct solution formula (1) and the optimal value of (2), but by solving (1) and the antithesis of (2)
Problem, and then obtain formula (1) and the optimal value of (2).The dual problem of formula (1) and (2) is as follows:
Wherein,WithIt is respectively Lagrange multiplier, Kw(xi, xi) it is referred to as spy
Levy Weighted Kernel function,
Solve dual problem (3) and (4) can obtain
In high-dimensional feature space, the centre of sphere of the characteristic weighing hypersphere of the i-th class malfunction
The radius of the characteristic weighing hypersphere of the i-th class malfunction in high-dimensional feature space
The centre of sphere of the characteristic weighing hypersphere of jth class malfunction in high-dimensional feature space
The radius of the characteristic weighing hypersphere of jth class malfunction in high-dimensional feature space
Wherein,WithIt is respectively dual problem (3) and (4) optimal value,
For any historical status data x, decision function DiiX () i.e. characteristic weighing twin Hypersphere Support Vector Machine model is such as
Under:
This decision function shows: historical status data to be sorted from which centre of sphere more close to, the half of the hypersphere of individual features weighting simultaneously
Footpath is the biggest, then which kind of malfunction type is these historical status data just belong to.
Centre of sphere c for the characteristic weighing ball of actual attribute data x to the i-th malfunctioni(ij)'s
Characteristic weighing distance:
Centre of sphere c for the characteristic weighing ball of actual attribute data x to jth malfunctionj(ij)'s
Characteristic weighing distance:
By making construction feature in aforementioned manners weight twin Hypersphere Support Vector Machine model in any two class malfunctions, permissible
ObtainIndividual characteristic weighing twin Hypersphere Support Vector Machine model Dij(x)。
Step 4, blast furnace fault diagnosis: blast fumance situation actual attribute data are brought into the twin hypersphere of each characteristic weighing of foundation
In supporting vector machine model, it is thus achieved that operation of blast furnace malfunction type belonging to blast fumance situation actual attribute data, complete blast furnace
Fault diagnosis.
Blast fumance situation actual attribute data are substituted in each characteristic weighing twin Hypersphere Support Vector Machine model, adds up each feature
Weight the operation of blast furnace malfunction type of twin Hypersphere Support Vector Machine model output, the operation of blast furnace fault that occurrence number is most
Status Type is operation of blast furnace malfunction type belonging to blast fumance situation actual attribute data.
Step 4-1, initialization ballot array Vote [m]={ 0}；
Step 4-2, by blast fumance situation real data substitute into each characteristic weighing twin Hypersphere Support Vector Machine model DijIn (x).
If step 4-3 DijX ()=i, Vote [i] are from increasing；Otherwise, Vote [j] is from increasing；
Step 4-4, blast fumance situation actual attribute data x finally affiliated malfunction type is the malfunction class that poll is the highest
Malfunction type belonging to type, i.e. actual attribute data x
Claims (6)
1. a blast furnace fault diagnosis system, it is characterised in that including:
Historical data acquisition module: gather blast fumance life history attribute data and the operation of blast furnace malfunction type of correspondence thereof；
Real data acquisition module: gather blast fumance situation actual attribute data；
Feature weight matrix construction module: according to each attribute, the significance level of fault diagnosis is determined to the feature weight of each attribute, structural feature weight matrix；
Model building module: utilize feature weight matrix, to any two operation of blast furnace malfunction type and corresponding blast fumance life history attribute data training thereof, set up characteristic weighing twin Hypersphere Support Vector Machine model, the input of this model is blast fumance life history attribute data, it is output as one of these two operation of blast furnace malfunction types, a pair characteristic weighing hypersphere in this model represents this two operation of blast furnace malfunction types respectively, blast fumance life history attribute data to be sorted from characteristic weighing hypersphere more close to, the probability that corresponding operation of blast furnace malfunction occurs is the highest；
Blast furnace fault diagnosis module: blast fumance situation actual attribute data are brought in each characteristic weighing twin Hypersphere Support Vector Machine model of foundation, it is thus achieved that operation of blast furnace malfunction type belonging to blast fumance situation actual attribute data, complete blast furnace fault diagnosis.
Blast furnace fault diagnosis system the most according to claim 1, it is characterized in that, also include normalized module, the blast fumance life history attribute data of historical data acquisition module collection and the blast fumance situation actual attribute data of the operation of blast furnace malfunction type of correspondence, real data acquisition module collection thereof are normalized.
Blast furnace fault diagnosis system the most according to claim 1, it is characterised in that described attribute, including: air quantity, blast, top pressure, pressure reduction, breathability, top temperature, cross temperature, material speed, Si content, physical thermal；Described malfunction type, including: Xiang Liang, thermotropism, hanging, collapse material.
Blast furnace fault diagnosis system the most according to claim 1, it is characterised in that the feature weight of described attribute is that comentropy and this attribute of malfunction type breaks down the difference of comentropy of state.
5. utilize the method that the system described in claim 1 carries out blast furnace fault diagnosis, it is characterised in that including:
Gather blast fumance situation actual attribute data, historical status data and the operation of blast furnace malfunction type of correspondence thereof；
According to each attribute, the significance level of fault diagnosis is determined to the feature weight of each attribute, structural feature weight matrix；
Set up the characteristic weighing twin Hypersphere Support Vector Machine model for blast furnace fault diagnosis: utilize feature weight matrix, to any two operation of blast furnace malfunction type and corresponding blast fumance life history attribute data training thereof, set up characteristic weighing twin Hypersphere Support Vector Machine model, the input of this model is blast fumance life history attribute data, it is output as one of these two operation of blast furnace malfunction types, a pair characteristic weighing hypersphere in this model represents this two operation of blast furnace malfunction types respectively, blast fumance life history attribute data to be sorted from characteristic weighing hypersphere more close to, the probability that corresponding operation of blast furnace malfunction occurs is the highest；
Blast furnace fault diagnosis: blast fumance situation actual attribute data are brought in each characteristic weighing twin Hypersphere Support Vector Machine model of foundation, it is thus achieved that operation of blast furnace malfunction type belonging to blast fumance situation actual attribute data, complete blast furnace fault diagnosis.
Method the most according to claim 5, it is characterized in that, operation of blast furnace malfunction type belonging to described acquisition blast fumance situation actual attribute data is: blast fumance situation actual attribute data substituted in each characteristic weighing twin Hypersphere Support Vector Machine model, adding up the operation of blast furnace malfunction type of each characteristic weighing twin Hypersphere Support Vector Machine model output, the most operation of blast furnace malfunction type of occurrence number is operation of blast furnace malfunction type belonging to blast fumance situation actual attribute data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610187770.0A CN105843212B (en) | 2016-03-29 | 2016-03-29 | A kind of blast furnace fault diagnosis system and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610187770.0A CN105843212B (en) | 2016-03-29 | 2016-03-29 | A kind of blast furnace fault diagnosis system and method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105843212A true CN105843212A (en) | 2016-08-10 |
| CN105843212B CN105843212B (en) | 2018-09-28 |
Family
ID=56584045
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610187770.0A Expired - Fee Related CN105843212B (en) | 2016-03-29 | 2016-03-29 | A kind of blast furnace fault diagnosis system and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105843212B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107526784A (en) * | 2017-07-27 | 2017-12-29 | 上海电力学院 | A kind of method for diagnosing faults based on matrix fill-in |
| CN109791401A (en) * | 2016-09-16 | 2019-05-21 | 西门子股份公司 | Generate the fault model for being used for embedded analysis and diagnosis/Forecast reasoning |
| CN110059359A (en) * | 2019-03-21 | 2019-07-26 | 江苏东方国信工业互联网有限公司 | A kind of system and method for the control furnace body technique based on big data analysis |
| CN117030264A (en) * | 2023-08-10 | 2023-11-10 | 盐城工学院 | Rolling bearing fault diagnosis method based on feature weighted twin support vector machine |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2009027889A2 (en) * | 2007-08-31 | 2009-03-05 | Koninklijke Philips Electronics N.V. | Uncertainty maps for segmentation in the presence of metal artifacts |
| CN101699432A (en) * | 2009-11-13 | 2010-04-28 | 黑龙江工程学院 | Ordering strategy-based information filtering system |
| CN103745234A (en) * | 2014-01-23 | 2014-04-23 | 东北大学 | Band steel surface defect feature extraction and classification method |
| CN104462846A (en) * | 2014-12-22 | 2015-03-25 | 山东鲁能软件技术有限公司 | Intelligent device failure diagnosis method based on support vector machine |
-
2016
- 2016-03-29 CN CN201610187770.0A patent/CN105843212B/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2009027889A2 (en) * | 2007-08-31 | 2009-03-05 | Koninklijke Philips Electronics N.V. | Uncertainty maps for segmentation in the presence of metal artifacts |
| CN101699432A (en) * | 2009-11-13 | 2010-04-28 | 黑龙江工程学院 | Ordering strategy-based information filtering system |
| CN103745234A (en) * | 2014-01-23 | 2014-04-23 | 东北大学 | Band steel surface defect feature extraction and classification method |
| CN104462846A (en) * | 2014-12-22 | 2015-03-25 | 山东鲁能软件技术有限公司 | Intelligent device failure diagnosis method based on support vector machine |
Non-Patent Citations (2)
| Title |
|---|
| 赵锋云: ""基于改进SVM的流程工业故障诊断方法研究及实现"", 《中国优秀硕士学位论文全文数据库-信息科技辑》 * |
| 邱建坤: ""基于孪生支持向量机的特征选择与多类分类算法研究"", 《中国优秀硕士学位论文全文数据库-信息科技辑》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109791401A (en) * | 2016-09-16 | 2019-05-21 | 西门子股份公司 | Generate the fault model for being used for embedded analysis and diagnosis/Forecast reasoning |
| CN107526784A (en) * | 2017-07-27 | 2017-12-29 | 上海电力学院 | A kind of method for diagnosing faults based on matrix fill-in |
| CN110059359A (en) * | 2019-03-21 | 2019-07-26 | 江苏东方国信工业互联网有限公司 | A kind of system and method for the control furnace body technique based on big data analysis |
| CN117030264A (en) * | 2023-08-10 | 2023-11-10 | 盐城工学院 | Rolling bearing fault diagnosis method based on feature weighted twin support vector machine |
| CN117030264B (en) * | 2023-08-10 | 2024-04-16 | 盐城工学院 | Rolling bearing fault diagnosis method based on feature weighted twin support vector machine |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105843212B (en) | 2018-09-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| AU2020104428A4 (en) | A rolling bearing fault diagnosis method based on GRCMSE and manifold learning | |
| CN105843212A (en) | System and method for fault diagnosis of blast furnace | |
| CN105487526A (en) | FastRVM (fast relevance vector machine) wastewater treatment fault diagnosis method | |
| CN106555788A (en) | Application of the deep learning based on Fuzzy Processing in hydraulic equipment fault diagnosis | |
| CN106971240A (en) | The short-term load forecasting method that a kind of variables choice is returned with Gaussian process | |
| CN105528504A (en) | Rolling bearing health condition evaluation method based on CFOA-MKHSVM | |
| Li et al. | An improved stacking ensemble learning-based sensor fault detection method for building energy systems using fault-discrimination information | |
| CN106371427A (en) | Industrial process fault classification method based on analytic hierarchy process and fuzzy fusion | |
| CN107203785A (en) | Multipath Gaussian kernel Fuzzy c-Means Clustering Algorithm | |
| CN107180155A (en) | A kind of disease forecasting method and system based on Manufacturing resource model | |
| CN110010248A (en) | A kind of readmission's Risk Forecast Method based on cost-sensitive integrated study model | |
| CN101630376A (en) | Soft-sensing modeling method and soft meter of multi-model neural network in biological fermentation process | |
| CN106503731A (en) | A kind of based on conditional mutual information and the unsupervised feature selection approach of K means | |
| CN107133632A (en) | A kind of wind power equipment fault diagnosis method and system | |
| CN109491338A (en) | A kind of relevant method for diagnosing faults of multimode procedure quality based on sparse GMM | |
| CN105243454A (en) | Big data-based electrical load prediction system | |
| CN106548270A (en) | A kind of photovoltaic plant power anomalous data identification method and device | |
| CN109492748A (en) | A kind of Mid-long term load forecasting method for establishing model of the electric system based on convolutional neural networks | |
| CN105425583A (en) | Control method of penicillin production process based on cooperative training local weighted partial least squares (LWPLS) | |
| CN104463251A (en) | Cancer gene expression profile data identification method based on integration of extreme learning machines | |
| CN107194526A (en) | A kind of sales marketization reform progress appraisal procedure based on fuzzy clustering | |
| CN109164794B (en) | Multivariable industrial process Fault Classification based on inclined F value SELM | |
| CN107451684A (en) | Stock market's probability forecasting method based on core stochastic approximation | |
| CN115204041A (en) | Large aircraft approach and approach stage oil consumption prediction method | |
| CN107491656B (en) | Pregnancy outcome influence factor evaluation method based on relative risk decision tree model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180928 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |