CN111580506A - Industrial process fault diagnosis method based on information fusion - Google Patents

Industrial process fault diagnosis method based on information fusion Download PDF

Info

Publication number
CN111580506A
CN111580506A CN202010493400.6A CN202010493400A CN111580506A CN 111580506 A CN111580506 A CN 111580506A CN 202010493400 A CN202010493400 A CN 202010493400A CN 111580506 A CN111580506 A CN 111580506A
Authority
CN
China
Prior art keywords
fault diagnosis
data
fault
data set
information fusion
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202010493400.6A
Other languages
Chinese (zh)
Inventor
张胜杰
周凌柯
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.)
Nanjing University of Science and Technology
Original Assignee
Nanjing University of Science and Technology
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 Nanjing University of Science and Technology filed Critical Nanjing University of Science and Technology
Priority to CN202010493400.6A priority Critical patent/CN111580506A/en
Publication of CN111580506A publication Critical patent/CN111580506A/en
Pending legal-status Critical Current

Links

Images

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
    • G05B23/0218—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults
    • G05B23/0243—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults model based detection method, e.g. first-principles knowledge model
    • 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
    • G05B2219/00—Program-control systems
    • G05B2219/20—Pc systems
    • G05B2219/24—Pc safety
    • G05B2219/24065—Real time diagnostics

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Automation & Control Theory (AREA)
  • Testing And Monitoring For Control Systems (AREA)

Abstract

The invention discloses an industrial process fault diagnosis method based on information fusion, which can solve the problem of data nonlinearity by using a data processing method of KPCA; the decision of the fault type is realized by using two machine learning methods, namely SVM based on a statistical learning theory and SOM based on an artificial neural network, and the decision result is subjected to information fusion by using a D-S evidence theory, so that the obtained final diagnosis result has higher fault diagnosis precision. The invention integrates a plurality of methods, replaces the original single method to process the fault diagnosis of the complex industrial process, and improves the multi-fault type diagnosis capability and the fault diagnosis precision.

Description

Industrial process fault diagnosis method based on information fusion
Technical Field
The invention belongs to the field of industrial process control, and particularly relates to an industrial process fault diagnosis method based on information fusion.
Background
With the development of modern production and the progress of scientific technology, modern industrial production equipment is increasingly large-scaled, complicated and automated. Because the actual industrial production process is extremely complex, the operation variables are more, and linear, nonlinear, Gaussian, non-Gaussian and other process data exist, the monitoring of the industrial process is insufficient by using a single diagnosis method, so that the faults possibly generated cannot be diagnosed, and the production safety, efficiency and product quality are possibly influenced badly. Therefore, a better diagnostic method capable of accurately pointing out the process fault problem has become one of the hot problems in the industrial process research.
The actual industrial production process has many uncertain factors and complex process, so that it is difficult to establish an accurate process model, however, a large amount of data generated in the industrial process can be analyzed and judged whether a fault occurs in the process, so a data-driven method is often used for fault diagnosis research. Machine learning, one aspect of the method, is mainly comprised of a neural network and a support vector machine. Self-organizing feature mapping, a method in neural networks, has self-learning and self-organizing capabilities that are unique advantages in the method of fault diagnosis research, but its training requires a large amount of historical sample data, making it constrained in systems where fault samples are difficult to acquire. The support vector machine is based on a statistical learning theory, has advantages in small sample fault decision, but has disadvantages in a fault system with large data scale and multiple classifications.
Due to the limitations of the above methods, a single method has not been able to accommodate all the possible problems of modern industrial practical production systems.
Disclosure of Invention
The invention aims to overcome the limitation of a single method and provide an industrial process fault diagnosis method based on information fusion.
The technical solution for realizing the purpose of the invention is as follows: an industrial process fault diagnosis method based on information fusion comprises the following steps:
(1) the method comprises the following steps that a sensor for monitoring process variables through an industrial system collects sensor data in the normal process and fault process operation of the system to form a training data set and an off-line test set, and the two sets of data sets are stored in a database;
(2) preprocessing and normalizing the training sample data set, namely, obtaining a new data set by the mean value and the variance of each process variable as 0 and 1;
(3) performing dimensionality reduction on the processed data set by using a KPCA dimensionality reduction method to obtain a dimensionality reduced data set and dimensionality reduction model parameters;
(4) training a data set after dimensionality reduction by using two methods, namely an SVM (support vector machine) and an SOM (self-organizing map), establishing different fault diagnosis models, and obtaining corresponding indexes and labels of the models;
(5) normalizing the off-line test data set, and processing by using a trained KPCA model to obtain a new off-line test data set;
(6) respectively using the trained fault diagnosis models to perform fault diagnosis on the new offline test data set to obtain fault diagnosis results and calculate a basic probability assignment function;
(7) collecting new on-line process data, preprocessing and normalizing the new on-line process data, and then processing the new on-line process data by using a dimension reduction model to obtain monitoring data;
(8) respectively diagnosing the monitoring data by using different fault diagnosis models to obtain a label of the monitoring data, and then taking a fault diagnosis result of each method as an evidence body;
(9) and (4) calculating a joint probability assignment function of the fused monitoring data according to the basic probability assignment function of the fault diagnosis result of each method by using a D-S evidence theory, and making a final decision.
Compared with the prior art, the invention has the beneficial effects that: the data processing method of KPCA used in the invention can deal with the problem of data nonlinearity; the decision of the fault type is realized by using two machine learning methods, namely SVM based on a statistical learning theory and SOM based on an artificial neural network, and the decision result is subjected to information fusion by using a D-S evidence theory, so that the obtained final diagnosis result has higher fault diagnosis precision.
Drawings
FIG. 1 is a graph of the hit of the KPCA-SOM method on TE process normal operation test data.
FIG. 2 is a graph of the hit of the KPCA-SOM method on failure 1 test data of a TE process.
FIG. 3 is a graph of the hit results of the KPCA-SOM method on the fault 12 test data of the TE process.
Fig. 4 is a graph of the results of a KPCA-SVM method diagnosing normality, faults 1 and 12 of a TE process.
Detailed Description
The invention relates to an industrial process fault diagnosis method based on information fusion, which comprises the following steps:
step 1, by industryThe system monitors variable sensors, and acquires sensor data in the normal process and fault process operation of the system to form a training data set: x ═ X1；x2；…；xn]Wherein X ∈ Rn×mN is the total number of training samples, m is the number of process variables, R is the set of real numbers, Rn×mThe X satisfies the two-dimensional distribution of n × m, data are collected in the same way to be used as an off-line test set, and the data are stored in a database;
step 2, calling a data set X from the database, preprocessing and normalizing, namely obtaining the mean value of each process variable as 0 and the variance as 1, and obtaining a new data set as
Figure BDA0002521948400000031
Step 3, using a data dimension reduction method to perform dimension reduction operation on the processed data set to obtain the dimension-reduced data set
Figure BDA0002521948400000032
Wherein l is the dimension of the data after dimension reduction;
selecting a Kernel Principal Component Analysis (KPCA) method for data processing, using a radial basis kernel function to map original data to a high-dimensional space to obtain a kernel matrix K, then performing principal component analysis on the kernel matrix K to obtain a characteristic value and a characteristic vector of the kernel matrix K, and using cumulative variance contribution rate (c: (a)>85%) to obtain principal element number l, determining load matrix P, and finally obtaining data matrix set after dimensionality reduction
Figure BDA0002521948400000033
Step 4, calling different fault diagnosis methods, and using the reduced-dimension data set
Figure BDA0002521948400000034
Establishing different fault diagnosis models to obtain corresponding indexes and labels of the models;
2 different fault diagnosis methods were selected, namely self-organizing feature mapping (SOM) neural networks and Support Vector Machines (SVMs):
(1) for the SOM fault diagnosis method, a better-adjusted weight vector W is obtained through iterative learningjLearning rate mu (t), field N (t) and winning neurons, and marking each neuron with a corresponding label after finishing the training of all training data to obtain an SOM fault diagnosis model;
(2) for the SVM fault diagnosis method, a training data set is established, a radial basis kernel function is selected, a Lagrange multiplier and a classification threshold value are calculated, a classification decision function is obtained, an optimal classification hyperplane is established, and an SVM fault diagnosis model is obtained.
Step 5, normalizing the off-line test data set, and processing by using a trained KPCA model to obtain a new off-line test data set;
step 6, respectively using the trained fault diagnosis model to perform fault diagnosis on the new offline test data set to obtain a fault diagnosis result and calculate a basic probability assignment function;
calling different fault diagnosis methods, and calculating corresponding basic probability assignment functions, wherein the formula is as follows:
Figure BDA0002521948400000035
wherein the content of the first and second substances,
Figure BDA0002521948400000036
refers to the element, m, of the ith row and the jth column in the fusion matrix of the a-th fault diagnosis methoda(Ci) Means that the a-th fault diagnosis method classifies the sample as CiThe probability value of the class is also called as a basic probability assignment function value, and G is the number of fault diagnosis methods; according to the method, the basic probability assignment function of each fault class corresponding to different methods can be calculated.
Step 7, collecting new on-line process data, preprocessing and normalizing the new on-line process data, and then processing the new on-line process data by using a dimension reduction model to obtain monitoring data;
step 8, diagnosing the monitoring data by using different fault diagnosis models respectively to obtain labels of the monitoring data, and then taking the fault diagnosis result of each method as an evidence body:
(1) for the self-organizing feature mapping neural network method, calculating the Euclidean distance between a test sample and a weight vector, and determining a label corresponding to the best neuron matched with the closest distance as a final judgment label;
(2) for the support vector machine method, a test sample is substituted into a classification decision function to obtain a sample decision output, and the class of the label is judged according to the output value. Because the method is a binary classification method and the diagnosed fault is unknown, a judgment label is given to each binary classification model, each judgment label is used as a ticket, and voting is carried out to determine to obtain the final judgment label.
Step 9, calculating a combined probability assignment function of the fused monitoring data according to a basic probability assignment function of a fault diagnosis result of each method by using a D-S evidence theory, and making a final decision;
(1) and calling the diagnosis output of the two fault diagnosis methods at the same sampling moment, selecting a corresponding basic probability assignment function of the fault, and solving the final basic probability assignment function of the sample by using a D-S fusion rule:
Figure BDA0002521948400000041
Figure BDA0002521948400000042
wherein the content of the first and second substances,
Figure BDA0002521948400000043
expressing orthogonal sums, as shown in formula (2), sets A, B and C respectively represent different fault class sets, A is the intersection of the sets B and C, k represents the joint probability assignment function value of the intersection as a set, and m is1And m2Respectively representing a first and a second fault diagnosis method, m1,2(A) Is the probability value after the two methods are fused.
(2) And assigning a function value according to the fused basic probability, and selecting the maximum corresponding fault as a final fault diagnosis result by comparing the function values.
The invention fuses a plurality of methods, namely, the information fusion method replaces the original single method to process the fault diagnosis of the complex industrial process, thereby improving the diagnosis capability of multiple fault types and the precision of the fault diagnosis.
The effectiveness of the invention is illustrated in connection with a specific industrial process example.
Examples
The data of the industrial process is from the TE (Tennessee Eastman-Tennessee-Islam) chemical process simulation experiment in the United states, and the prototype is from the actual production process of the Tennessee Eastman chemical company. The TE process is a common standard problem, which can better simulate some typical characteristics of a real complex industrial system, and therefore, is widely used for the research of typical chemical process fault diagnosis. The whole process comprises 5 main units: reactor, condenser, compressor, catch water and stripper. There are 4 reactants, 2 products, an inert component and a by-product in the entire TE process. The whole process contains 53 state variables, which are 12 manipulated variables and 41 measured variables, respectively, but only 52 state variables are actually used. Of these 41 measured variables, 22 consecutive measured variables (which the system will sample every 3 minutes) and 19 component measured variables (artificially adding noise). The whole process presets 21 faults and a normal state, and the faults comprise 16 known faults and 5 unknown faults. Wherein, the faults 1 to 7 are step type faults, the faults 8 to 12 are random variable type faults, the fault 13 is a slow migration type fault, the faults 14 and 15 bit sticky type faults, the faults 16 to 20 are unknown faults, and the fault 21 bit constant position fault. The following detailed description of the implementation steps of the present invention is made in conjunction with specific procedures:
step 1, collecting data of a normal process and a fault process, dividing the data into a training data set and an off-line testing data set, preprocessing and normalizing the data, and storing the preprocessed and normalized data into a database;
step 2, calling a kernel principal component analysis method to perform dimensionality reduction operation on the training data to obtain the number of principal components, a kernel matrix, a dimensionality reduced data set and other parameters;
for training data matrix
Figure BDA0002521948400000051
Establishing a dimension reduction model:
(1) selecting Gaussian radial kernel function, selecting parameters 3000, and calculating
Figure BDA0002521948400000052
Kernel matrix K ∈ Rn×n；
(2) According to
Figure BDA0002521948400000053
Processing the obtained kernel matrix to obtain
Figure BDA0002521948400000054
Figure BDA0002521948400000055
Wherein
Figure BDA0002521948400000056
(3) Decomposing the eigenvalue of the processed kernel matrix and normalizing the eigenvector;
(4) selecting contribution degree of 85%, obtaining principal component number of 23, obtaining new training data matrix
Figure BDA0002521948400000057
Step 3, data set
Figure BDA0002521948400000058
Respectively substituting the fault diagnosis model into an SOM method and an SVM method for training to obtain a fault diagnosis model;
step 4, calling an off-line test data set, substituting the off-line test data set into the trained fault diagnosis model to diagnose the fault, and calculating a basic probability assignment function of each method for fault judgment according to results;
and 5, online process diagnosis.
In order to test the effectiveness of the method, a normal sample and a fault sample are respectively tested. The 3 categories of normal, fault 1 and fault 12 are selected for fault detection, and the diagnosis results of the two fault diagnosis methods of KPCA-SOM and KPCA-SVM are shown in fig. 1, fig. 2, fig. 3 and fig. 4, respectively, wherein the test data hit neurons in fig. 1, fig. 2 and fig. 3 are represented by black box hexagons, and fig. 4 is represented by labels 0, 1 and 2, respectively, as normal, fault 1 and fault 12. It can be seen from the figure that the diagnostic result accuracy varies when different fault diagnosis methods are used for the same test data. The test data were subjected to failure diagnosis using the method of the present invention, and the comparison results are shown in table 1.
Table 1: diagnostic result comparison table of each fault diagnosis method in the invention
KPCA-SOM KPCA-SVM D-S
Is normal 100％ 100％ 100％
Failure
1 88％ 84％ 90％
Fault
12 78％ 68％ 88％
Synthesis of 88.7％ 84％ 92.7％
The result shows that compared with a single fault diagnosis method, the method provided by the invention has the advantage that the fault diagnosis performance is obviously improved.

Claims (8)

1. An industrial process fault diagnosis method based on information fusion is characterized by comprising the following steps:
(1) the method comprises the following steps that a sensor for monitoring process variables through an industrial system collects sensor data in the normal process and fault process operation of the system to form a training data set and an off-line test set, and the two sets of data sets are stored in a database;
(2) preprocessing and normalizing the training sample data set, namely, obtaining a new data set by the mean value and the variance of each process variable as 0 and 1;
(3) performing dimensionality reduction on the processed data set by using a KPCA dimensionality reduction method to obtain a dimensionality reduced data set and dimensionality reduction model parameters;
(4) training a data set after dimensionality reduction by using two methods, namely an SVM (support vector machine) and an SOM (self-organizing map), establishing different fault diagnosis models, and obtaining corresponding indexes and labels of the models;
(5) normalizing the off-line test data set, and processing by using a trained KPCA model to obtain a new off-line test data set;
(6) respectively using the trained fault diagnosis models to perform fault diagnosis on the new offline test data set to obtain fault diagnosis results and calculate a basic probability assignment function;
(7) collecting new on-line process data, preprocessing and normalizing the new on-line process data, and then processing the new on-line process data by using a dimension reduction model to obtain monitoring data;
(8) respectively diagnosing the monitoring data by using different fault diagnosis models to obtain a label of the monitoring data, and then taking a fault diagnosis result of each method as an evidence body;
(9) and (4) calculating a joint probability assignment function of the fused monitoring data according to the basic probability assignment function of the fault diagnosis result of each method by using a D-S evidence theory, and making a final decision.
2. The information fusion-based industrial process fault diagnosis method according to claim 1, wherein the step 1 specifically comprises: monitoring variable sensors through an industrial system, and acquiring sensor data in the normal process and fault process operation of the system to form a training data set: x ═ X1；x2；…；xn]Wherein X ∈ Rn×mN is the total number of training samples, m is the number of process variables, R is the set of real numbers, Rn×mRepresenting that X satisfies a two-dimensional distribution of n × m, and collecting data in the same manner as an off-line test set and storing the data in a database.
3. The information fusion-based industrial process fault diagnosis method according to claim 1, wherein the step 2 specifically comprises: calling a data set X from the database, preprocessing and normalizing, namely obtaining the mean value of each process variable as 0 and the variance as 1, and obtaining a new data set as
Figure FDA0002521948390000011
4. The information fusion-based industrial process fault diagnosis method according to claim 1, wherein the step 3 specifically comprises: selecting a kernel principal component analysis method for data processing, using a radial basis kernel function, mapping original data to a high-dimensional space to obtain a kernel matrix K, then performing principal component analysis on the kernel matrix K to obtain a characteristic value and a characteristic vector of the kernel matrix K, and using cumulative variance contribution rateCalculating to obtain the number l of principal elements, determining a load matrix P, and finally obtaining a data matrix set after dimension reduction
Figure FDA0002521948390000021
5. The information fusion-based industrial process fault diagnosis method according to claim 1, wherein the step 4 specifically comprises: two different fault diagnosis methods are selected, namely a self-organizing feature mapping neural network and a support vector machine:
for the SOM fault diagnosis method, a better-adjusted weight vector W is obtained through iterative learningjLearning rate mu (t), field N (t) and winning neurons, and marking each neuron with a corresponding label after finishing the training of all training data to obtain an SOM fault diagnosis model;
for the SVM fault diagnosis method, a training data set is established, a radial basis kernel function is selected, a Lagrange multiplier and a classification threshold value are calculated, a classification decision function is obtained, an optimal classification hyperplane is established, and an SVM fault diagnosis model is obtained.
6. The information fusion-based industrial process fault diagnosis method according to claim 1, wherein the step 6 specifically comprises:
calling different fault diagnosis methods, and calculating corresponding basic probability assignment functions, wherein the formula is as follows:
Figure FDA0002521948390000022
wherein the content of the first and second substances,
Figure FDA0002521948390000023
refers to the element, m, of the ith row and the jth column in the fusion matrix of the a-th fault diagnosis methoda(Ci) Means that the a-th fault diagnosis method classifies the sample as CiThe probability value of the class is also called as a basic probability assignment function value, and G is the number of fault diagnosis methods; in accordance withAccording to the method, the basic probability assignment function of each fault class corresponding to different methods can be calculated.
7. The information fusion-based industrial process fault diagnosis method according to claim 1, wherein step 7 specifically comprises:
for the self-organizing feature mapping neural network method, calculating the Euclidean distance between a test sample and a weight vector, and determining a label corresponding to the best neuron matched with the closest distance as a final judgment label;
for the support vector machine method, a test sample is substituted into a classification decision function to obtain a sample decision output, and the class of the label is judged according to the output value; because the method is a binary classification method and the diagnosed fault is unknown, a judgment label is given to each binary classification model, each judgment label is used as a ticket, and voting is carried out to determine to obtain the final judgment label.
8. The information fusion-based industrial process fault diagnosis method according to claim 1, wherein step 9 is to calculate a joint probability assignment function of the fused monitoring data according to a basic probability assignment function of the fault diagnosis result of each method by using a D-S evidence theory, and to make a final decision, specifically:
(1) and calling the diagnosis output of the two fault diagnosis methods at the same sampling moment, selecting a corresponding basic probability assignment function of the fault, and solving the final basic probability assignment function of the sample by using a D-S fusion rule:
Figure FDA0002521948390000031
Figure FDA0002521948390000032
wherein the content of the first and second substances,
Figure FDA0002521948390000033
expressing orthogonal sums, as shown in formula (2), sets A, B and C respectively represent different fault class sets, A is the intersection of the sets B and C, k represents the joint probability assignment function value of the intersection as a set, and m is1And m2Respectively representing a first and a second fault diagnosis method, m1,2(A) Is the probability value after the two methods are fused;
(2) and assigning a function value according to the fused basic probability, and selecting the maximum corresponding fault as a final fault diagnosis result by comparing the function values.
CN202010493400.6A 2020-06-03 2020-06-03 Industrial process fault diagnosis method based on information fusion Pending CN111580506A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010493400.6A CN111580506A (en) 2020-06-03 2020-06-03 Industrial process fault diagnosis method based on information fusion

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010493400.6A CN111580506A (en) 2020-06-03 2020-06-03 Industrial process fault diagnosis method based on information fusion

Publications (1)

Publication Number Publication Date
CN111580506A true CN111580506A (en) 2020-08-25

Family

ID=72111345

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010493400.6A Pending CN111580506A (en) 2020-06-03 2020-06-03 Industrial process fault diagnosis method based on information fusion

Country Status (1)

Country Link
CN (1) CN111580506A (en)

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112256686A (en) * 2020-11-12 2021-01-22 北京理工大学 MES-based fault data compression and prediction method and system
CN112613186A (en) * 2020-12-30 2021-04-06 中国航空发动机研究院 Aero-engine gas circuit fault fusion diagnosis method based on statistical distribution characteristics
CN112748717A (en) * 2020-12-08 2021-05-04 华北电力大学（保定） Actuator fault diagnosis method based on evidence fusion
CN112990275A (en) * 2021-02-20 2021-06-18 长春工业大学 High-speed train running gear system fault diagnosis method based on semi-quantitative information fusion
CN113033079A (en) * 2021-03-08 2021-06-25 重庆优易特智能科技有限公司 Chemical fault diagnosis method based on unbalanced correction convolutional neural network
CN113257329A (en) * 2021-04-12 2021-08-13 中国空间技术研究院 Memory fault diagnosis method based on machine learning
EP4148518A1 (en) 2021-09-13 2023-03-15 EVER_EST Srl Kpca-based non-linear chiller fault detection method
CN117348605A (en) * 2023-12-05 2024-01-05 东莞栢能电子科技有限公司 Optimization method and system applied to control system of release film tearing machine
CN117666546A (en) * 2024-01-31 2024-03-08 中核武汉核电运行技术股份有限公司 Distributed control system fault diagnosis method and device

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20140046881A1 (en) * 2012-08-07 2014-02-13 Prüftechnik Ag Method for monitoring of rotating machines
CN103914064A (en) * 2014-04-01 2014-07-09 浙江大学 Industrial process fault diagnosis method based on multiple classifiers and D-S evidence fusion
CN104330255A (en) * 2014-11-04 2015-02-04 徐州隆安光电科技有限公司 Gear fault diagnosing method based on multi-sensor information fusion
CN106371427A (en) * 2016-10-28 2017-02-01 浙江大学 Industrial process fault classification method based on analytic hierarchy process and fuzzy fusion
CN108919059A (en) * 2018-08-23 2018-11-30 广东电网有限责任公司 A kind of electric network failure diagnosis method, apparatus, equipment and readable storage medium storing program for executing
CN109657647A (en) * 2019-01-09 2019-04-19 红河学院 A kind of Fault Diagnosis Method of Hydro-generating Unit based on information fusion technology

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20140046881A1 (en) * 2012-08-07 2014-02-13 Prüftechnik Ag Method for monitoring of rotating machines
CN103914064A (en) * 2014-04-01 2014-07-09 浙江大学 Industrial process fault diagnosis method based on multiple classifiers and D-S evidence fusion
CN104330255A (en) * 2014-11-04 2015-02-04 徐州隆安光电科技有限公司 Gear fault diagnosing method based on multi-sensor information fusion
CN106371427A (en) * 2016-10-28 2017-02-01 浙江大学 Industrial process fault classification method based on analytic hierarchy process and fuzzy fusion
CN108919059A (en) * 2018-08-23 2018-11-30 广东电网有限责任公司 A kind of electric network failure diagnosis method, apparatus, equipment and readable storage medium storing program for executing
CN109657647A (en) * 2019-01-09 2019-04-19 红河学院 A kind of Fault Diagnosis Method of Hydro-generating Unit based on information fusion technology

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
袁海满等: "基于多信息融合的变压器故障诊断", 《高压电器》 *

Cited By (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112256686A (en) * 2020-11-12 2021-01-22 北京理工大学 MES-based fault data compression and prediction method and system
CN112256686B (en) * 2020-11-12 2022-11-11 北京理工大学 MES-based fault data compression and prediction method and system
CN112748717A (en) * 2020-12-08 2021-05-04 华北电力大学（保定） Actuator fault diagnosis method based on evidence fusion
CN112613186B (en) * 2020-12-30 2022-04-12 中国航空发动机研究院 Aero-engine gas circuit fault fusion diagnosis method based on statistical distribution characteristics
CN112613186A (en) * 2020-12-30 2021-04-06 中国航空发动机研究院 Aero-engine gas circuit fault fusion diagnosis method based on statistical distribution characteristics
CN112990275A (en) * 2021-02-20 2021-06-18 长春工业大学 High-speed train running gear system fault diagnosis method based on semi-quantitative information fusion
CN112990275B (en) * 2021-02-20 2022-08-26 长春工业大学 High-speed train running gear system fault diagnosis method based on semi-quantitative information fusion
CN113033079A (en) * 2021-03-08 2021-06-25 重庆优易特智能科技有限公司 Chemical fault diagnosis method based on unbalanced correction convolutional neural network
CN113033079B (en) * 2021-03-08 2023-07-18 重庆优易特智能科技有限公司 Chemical fault diagnosis method based on unbalance correction convolutional neural network
CN113257329A (en) * 2021-04-12 2021-08-13 中国空间技术研究院 Memory fault diagnosis method based on machine learning
EP4148518A1 (en) 2021-09-13 2023-03-15 EVER_EST Srl Kpca-based non-linear chiller fault detection method
CN117348605A (en) * 2023-12-05 2024-01-05 东莞栢能电子科技有限公司 Optimization method and system applied to control system of release film tearing machine
CN117348605B (en) * 2023-12-05 2024-03-12 东莞栢能电子科技有限公司 Optimization method and system applied to control system of release film tearing machine
CN117666546A (en) * 2024-01-31 2024-03-08 中核武汉核电运行技术股份有限公司 Distributed control system fault diagnosis method and device
CN117666546B (en) * 2024-01-31 2024-05-03 中核武汉核电运行技术股份有限公司 Distributed control system fault diagnosis method and device

Similar Documents

Publication Publication Date Title
CN111580506A (en) Industrial process fault diagnosis method based on information fusion
US11586913B2 (en) Power equipment fault detecting and positioning method of artificial intelligence inference fusion
CN103914064B (en) Based on the commercial run method for diagnosing faults that multi-categorizer and D-S evidence merge
CN106355030B (en) A kind of fault detection method based on analytic hierarchy process (AHP) and Nearest Neighbor with Weighted Voting Decision fusion
Ko et al. Fault classification in high-dimensional complex processes using semi-supervised deep convolutional generative models
CN109240274B (en) Industrial process fault diagnosis method based on high-order correlation
CN103901880A (en) Industrial process fault detection method based on multiple classifiers and D-S evidence fusion
CN111860446A (en) Detection system and method for unknown mode of satellite remote measurement time sequence data
CN112904810B (en) Process industry nonlinear process monitoring method based on effective feature selection
Yu et al. Stacked denoising autoencoder‐based feature learning for out‐of‐control source recognition in multivariate manufacturing process
CN111340110A (en) Fault early warning method based on industrial process running state trend analysis
CN111122811A (en) Sewage treatment process fault monitoring method of OICA and RNN fusion model
CN101738998A (en) System and method for monitoring industrial process based on local discriminatory analysis
CN116380445A (en) Equipment state diagnosis method and related device based on vibration waveform
Shajihan et al. CNN based data anomaly detection using multi-channel imagery for structural health monitoring
CN115496108A (en) Fault monitoring method and system based on manifold learning and big data analysis
CN112947649B (en) Multivariate process monitoring method based on mutual information matrix projection
CN117351659B (en) Hydrogeological disaster monitoring device and monitoring method
CN113703422A (en) Gas turbine pneumatic actuator fault diagnosis method based on characteristic analysis processing
CN116108371B (en) Cloud service abnormity diagnosis method and system based on cascade abnormity generation network
CN106250937B (en) A kind of failure modes diagnostic method based on non-index of similarity
CN116720073A (en) Abnormality detection extraction method and system based on classifier
Song et al. Fault diagnosis and process monitoring using a statistical pattern framework based on a self-organizing map
CN114371677B (en) Industrial process state monitoring method based on spectrum radius-interval principal component analysis
CN109389313A (en) A kind of failure modes diagnostic method based on weighting neighbour's decision

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
RJ01 Rejection of invention patent application after publication

Application publication date: 20200825

RJ01 Rejection of invention patent application after publication