CN111352408B - Multi-working-condition process industrial process fault detection method based on evidence K nearest neighbor - Google Patents

Multi-working-condition process industrial process fault detection method based on evidence K nearest neighbor Download PDF

Info

Publication number
CN111352408B
CN111352408B CN202010164092.2A CN202010164092A CN111352408B CN 111352408 B CN111352408 B CN 111352408B CN 202010164092 A CN202010164092 A CN 202010164092A CN 111352408 B CN111352408 B CN 111352408B
Authority
CN
China
Prior art keywords
sample
evidence
condition
training
samples
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
CN202010164092.2A
Other languages
Chinese (zh)
Other versions
CN111352408A (en
Inventor
贺凯迅
苏照阳
钟麦英
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.)
Shandong University of Science and Technology
Original Assignee
Shandong 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 Shandong University of Science and Technology filed Critical Shandong University of Science and Technology
Priority to CN202010164092.2A priority Critical patent/CN111352408B/en
Publication of CN111352408A publication Critical patent/CN111352408A/en
Application granted granted Critical
Publication of CN111352408B publication Critical patent/CN111352408B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • G05B23/0259—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterized by the response to fault detection
    • G05B23/0262—Confirmation of fault detection, e.g. extra checks to confirm that a failure has indeed occurred
    • 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 a multi-working-condition process industrial process fault detection method based on evidence K nearest neighbor, which belongs to the field of industrial process monitoring. Then calculating the correlation degree of each sample in the training set and each working condition, and establishing an evidence information base; for a sample to be detected, K samples are selected from the training set by the K neighbor model, and evidence information corresponding to the K samples is synthesized by a D-S evidence theory to obtain a final detection result. The fault diagnosis model established based on evidence K nearest neighbor can be well adapted to the multi-working-condition running state of the system; the method can reduce the false alarm rate in the monitoring process and improve the accuracy of fault alarm.

Description

Multi-working-condition process industrial process fault detection method based on evidence K nearest neighbor
Technical Field
The invention belongs to the field of industrial process monitoring, and particularly relates to a multi-working-condition process industrial process fault detection method based on evidence K nearest neighbor.
Background
With the widespread use of DCS, it is becoming easier to collect data from industrial processes. The development of data mining and database technology also provides powerful technical support for the development and application of data-driven modeling methods in industrial processes. Data-driven process monitoring, control and system optimization are therefore widely used.
The boiler is one of the most important parts of a thermal power plant as a complex large energy conversion device, and the stable operation of the boiler is closely related to the quality of electric energy. However, the combustion process of the boiler is complex and changeable, and a precise model of the boiler is difficult to establish by utilizing the traditional mechanism analysis; meanwhile, the load of the boiler changes along with the external requirements, the operation mode of the boiler changes under different loads, and the operation process shows obvious multi-working-condition characteristics.
For conventional fault diagnosis methods, such as Principal Component Analysis (PCA), Partial Least Squares (PLS), etc., have been used for a long time for fault detection and diagnosis of industrial processes. Such methods are also considered to be very effective. However, it should be noted that the methods of PCA, PLS, etc. are formulated under the assumption that the process variables are independent, uniformly distributed and linearly related, and that the system has only a single normal operating condition. In actual industrial production, the assumed conditions are difficult to meet, the monitoring result is inaccurate, and a large number of false alarms occur.
Disclosure of Invention
Aiming at the technical problems, the invention provides a multi-working-condition process industrial process fault detection method based on evidence K nearest neighbor, overcomes the defects of the prior art and has good effect.
In order to achieve the purpose, the invention adopts the following technical scheme:
a multi-working-condition process industrial process fault detection method based on evidence K nearest neighbor comprises the following steps:
step 1: acquiring normal historical working condition data of an industrial process, taking the data as a training set, standardizing the training set, and calculating a standardized value;
step 2: dividing the standardized training set into a plurality of normal working conditions through a clustering algorithm;
and step 3: determining evidence information according to a relevancy criterion, and obtaining the relevancy of each training sample and each normal working condition as each training sample carries the evidence information, thereby establishing an evidence information base; the evidence information base is a set of evidence information carried by all normal samples;
and 4, step 4: calculating discount factor gamma of training samples according to local density of each training sampleiAdapting the fault boundary to the local characteristics of the training sample;
and 5: standardizing the data to be detected, and finding out k samples closest to the data to be detected in a training set;
step 6: and (4) synthesizing the evidence information of the k samples in the step (5) according to a D-S evidence theory Dempster' S integration rule to obtain a fault detection result.
Preferably, in step 1, the normalized value, normal operating condition data set, is calculated according to equation (1)
Figure BDA0002406780310000021
n is the number of normal working condition samples, i represents the ith sample, and p is the dimension of each sample:
Figure BDA0002406780310000022
wherein,
Figure BDA0002406780310000023
is the actual value of the,
Figure BDA0002406780310000024
is a variable xjThe maximum value of (a) is,
Figure BDA0002406780310000025
is a variable xjThe minimum value of (a) is determined,
Figure BDA0002406780310000026
is a normalized value and j represents the sample dimension.
Preferably, the clustering algorithm in step 2 is any one of a K-means clustering method, a gaussian mixture model clustering method, and a density peak value clustering method.
Preferably, the relevancy criterion in step 3 is one or a combination of two methods selected from euclidean distance, mahalanobis distance and local density.
Preferably, in step 3, evidence information is calculated for each training sample according to equation (2):
Figure BDA0002406780310000027
wherein q represents the q-th normal operating condition, diqThe correlation degree of the ith sample and the qth normal working condition is shown, and c represents the total number of the normal working conditions;
Figure BDA0002406780310000028
represents the sum of the correlation of the ith sample with all the working conditions, mi({ωq}) denotes the ithThe probability that the sample belongs to the qth normal condition.
Preferably, in step 4, the discount factor γ of the training samples is calculated according to equation (3)i：
Figure BDA0002406780310000029
Wherein D isiIs a sample xiAnd distance sample xiThe distance between the nearest k training samples is combined into a set, which is a parameter within the quantile function.
Preferably, in step 6, the following steps are included:
step 6.1: calculating an adaptive discount factor gamma according to equation (4)s：
Figure BDA0002406780310000031
Wherein λ is used to adjust γsSize of (a), γiA discount factor for a training sample;
step 6.2: according to equation (5), the weight values of the training samples are calculated:
Figure BDA0002406780310000032
wherein d isiRepresenting Euclidean distance, gamma, of the sample to be measured and the training samplesIs an adaptive discount factor;
step 6.3: and (3) calculating a mass function of the sample to be tested belonging to each working condition according to the formula (6):
Figure BDA0002406780310000033
where α is the confidence of the training set, mi(A) The evidence information of the sample to be detected is obtained;
Figure BDA0002406780310000034
index set representing K training samples closest to the sample to be measured, dsiRepresenting the distance between the sample to be tested and the training sample;
step 6.4: synthesizing a mass function according to the formulas (7), (8) and (9) to obtain a diagnosis result;
Figure BDA0002406780310000035
Figure BDA0002406780310000036
Figure BDA0002406780310000037
wherein, K1As a normalization factor, ms(Ω) represents the failure probability of the test sample, ms({ωq} represents the probability that the test sample is subject to each condition.
The invention has the following beneficial technical effects:
(1) the fault diagnosis model established based on the K neighbor can be well adapted to the multi-working-condition running state of the system; (2) the method can reduce the false alarm rate in the monitoring process and improve the accuracy of fault alarm.
Drawings
FIG. 1 is a flow chart of the method of the present invention.
FIG. 2 is a schematic diagram of a projection of a training set clustering result on a two-dimensional space according to an embodiment of the present invention.
Fig. 3 is a schematic diagram of a fault detection result of the multi-working-condition process industrial process fault detection method based on evidence K nearest neighbor according to the embodiment of the present invention.
Detailed Description
The invention is described in further detail below with reference to the following figures and detailed description:
the example data was generated from one coal mill in a power plant boiler system, for a total of 2880 samples and 28 variables. The fault interval is 2133 and 2500. The first 1000 normal samples are taken as a training set, and the rest samples are taken as a testing set. As shown in fig. 1, the method of the present invention includes the following two major stages: the method comprises the steps of respectively performing off-line modeling and on-line monitoring, wherein the off-line modeling stage specifically comprises the following steps:
step 1: standardizing the training set;
Figure BDA0002406780310000041
in the above formula, the first and second carbon atoms are,
Figure BDA0002406780310000042
is the actual value of the,
Figure BDA0002406780310000043
is a variable xjThe maximum value of (a) is,
Figure BDA0002406780310000044
is a variable xjThe minimum value of (a) is determined,
Figure BDA0002406780310000045
is a normalized value and j represents the sample dimension.
Step 2: projecting a high-dimensional data space composed of process data to a low-dimensional feature subspace by using PCA (principal component analysis), as shown in FIG. 2; observing the optimal clustering number, clustering the training set by a K-means clustering algorithm, wherein a parameter K is 5, and dividing the normal working condition into two working conditions of w1 and w 2;
and step 3: respectively calculating the central points of w1 and w2, calculating the correlation degree of each training set with w1 and w2 to establish an evidence information base, and calculating the discount factor gamma of each training samplei；
Figure BDA0002406780310000046
In the above formula, DiIs a training sample xiModulo K nearest neighborAnd (4) selecting a distance set of the training samples, wherein the parameter is 0.85.
The on-line monitoring stage mainly comprises:
step 1: for a sample x to be measurediFirst, normalization is performed. Finding the 5 nearest neighbor samples in the training set constitutes the set Θ.
Step 2: calculating the sample x to be tested according to the discount factor corresponding to each training set in the thetaiAdaptive discount factor gamma ofs。
Figure BDA0002406780310000051
In the above formula, the parameter λ is 0.2, γiA discount factor for a training sample;
and step 3: calculating a sample x to be measurediAnd the weight between each training set within Θ:
Figure BDA0002406780310000052
in the above formula, diRepresenting Euclidean distance, gamma, of the sample to be measured and the training samplesIs an adaptive discount factor;
and 4, step 4: calculating a mass function of the sample to be detected belonging to each working condition:
Figure BDA0002406780310000053
where the parameter α is 0.95, m is the confidence of the training seti(A) Evidence information of training samples;
and 5: and synthesizing the evidence carried by each training set in the theta by using Dempster's integration rule to obtain a final detection result. When m is specifieds(Ω) ═ 0.2 as the early warning boundary of the fault, msThe fault boundary is set to 0.9 (Ω). The monitoring results are shown in fig. 3;
Figure BDA0002406780310000054
Figure BDA0002406780310000055
Figure BDA0002406780310000056
wherein, K1As a normalization factor, ms(Ω) represents the failure probability of the test sample, ms({ωq} represents the probability that the test sample is subject to each condition.
The above description is not intended to limit the present invention, and the present invention is not limited to the above examples, and variations, modifications, additions and substitutions which may be made by those skilled in the art within the spirit of the present invention are within the scope of the present invention.

Claims (6)

1. A multi-working-condition process industrial process fault detection method based on evidence K nearest neighbor is characterized by comprising the following steps: the method comprises the following steps:
step 1: acquiring normal historical working condition data of an industrial process, taking the normal historical working condition data as a training set, standardizing the training set, and calculating a standardized value;
step 2: dividing the standardized training set into a plurality of normal working conditions through a clustering algorithm;
and step 3: determining evidence information according to a relevancy criterion, and obtaining the relevancy of each training sample and each normal working condition as each training sample carries the evidence information, thereby establishing an evidence information base; the evidence information base is a set of evidence information carried by all normal samples;
and 4, step 4: calculating discount factor gamma of training samples according to local density of each training sampleiAdapting the fault boundary to the local characteristics of the training sample;
and 5: standardizing the data to be detected, and finding out k samples closest to the data to be detected in a training set;
step 6: synthesizing the evidence information of the k samples in the step 5 according to a Dempster' S integration rule of a D-S evidence theory to obtain a fault detection result; the method comprises the following steps:
step 6.1: calculating an adaptive discount factor gamma according to equation (4)s：
Figure FDA0002705629780000011
Wherein λ is used to adjust γsSize of (a), γiA discount factor for a training sample;
step 6.2: according to equation (5), the weight values of the training samples are calculated:
Figure FDA0002705629780000012
wherein d isiRepresenting Euclidean distance, gamma, of the sample to be measured and the training samplesIs an adaptive discount factor;
step 6.3: and (3) calculating a mass function of the sample to be tested belonging to each working condition according to the formula (6):
Figure FDA0002705629780000013
where α is the confidence of the training set, mi(A) The evidence information of the sample to be detected is obtained;
Figure FDA0002705629780000014
index set representing K training samples closest to the sample to be measured, dsiRepresenting the distance between the sample to be tested and the training sample;
step 6.4: synthesizing a mass function according to the formulas (7), (8) and (9) to obtain a diagnosis result;
Figure FDA0002705629780000021
Figure FDA0002705629780000022
Figure FDA0002705629780000023
wherein, K1As a normalization factor, ms(Ω) represents the failure probability of the test sample, ms({ωq} represents the probability that the test sample is subject to each condition.
2. The evidence K nearest neighbor based multi-condition process industrial process fault detection method according to claim 1, wherein: in step 1, the normalized value, normal condition data set, is calculated according to equation (1)
Figure FDA0002705629780000024
n is the number of normal working condition samples, i represents the ith sample, and p is the dimension of each sample:
Figure FDA0002705629780000025
wherein,
Figure FDA0002705629780000026
is the actual value of the,
Figure FDA0002705629780000027
is a variable xjThe maximum value of (a) is,
Figure FDA0002705629780000028
is a variable xjThe minimum value of (a) is determined,
Figure FDA0002705629780000029
is a normalized value and j represents the sample dimension.
3. The evidence K nearest neighbor based multi-condition process industrial process fault detection method according to claim 1, wherein: the clustering algorithm in the step 2 is any one of a K-means clustering method, a Gaussian mixture model clustering method and a density peak value clustering method.
4. The evidence K nearest neighbor based multi-condition process industrial process fault detection method according to claim 1, wherein: and the relevancy criterion in the step 3 is one or a combination of two methods of Euclidean distance, Mahalanobis distance and local density.
5. The evidence K nearest neighbor based multi-condition process industrial process fault detection method according to claim 1, wherein: in step 3, evidence information is calculated for each training sample according to equation (2):
Figure FDA0002705629780000031
wherein q represents the q-th normal operating condition, diqThe correlation degree of the ith sample and the qth normal working condition is shown, and c represents the total number of the normal working conditions;
Figure FDA0002705629780000032
represents the sum of the correlation of the ith sample with all the working conditions, mi({ωqAnd) represents the probability that the ith sample belongs to the qth normal condition.
6. The evidence K nearest neighbor based multi-condition process industrial process fault detection method according to claim 1, wherein: in step 4, the fold of the training sample is calculated according to equation (3)Deduction factor gammai：
Figure FDA0002705629780000033
Wherein D isiIs a sample xiAnd distance sample xiThe distance between the nearest k training samples is combined into a set, which is a parameter within the quantile function.
CN202010164092.2A 2020-03-11 2020-03-11 Multi-working-condition process industrial process fault detection method based on evidence K nearest neighbor Active CN111352408B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010164092.2A CN111352408B (en) 2020-03-11 2020-03-11 Multi-working-condition process industrial process fault detection method based on evidence K nearest neighbor

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010164092.2A CN111352408B (en) 2020-03-11 2020-03-11 Multi-working-condition process industrial process fault detection method based on evidence K nearest neighbor

Publications (2)

Publication Number Publication Date
CN111352408A CN111352408A (en) 2020-06-30
CN111352408B true CN111352408B (en) 2020-12-08

Family

ID=71197282

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010164092.2A Active CN111352408B (en) 2020-03-11 2020-03-11 Multi-working-condition process industrial process fault detection method based on evidence K nearest neighbor

Country Status (1)

Country Link
CN (1) CN111352408B (en)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108171271B (en) * 2018-01-11 2022-04-29 湖南大唐先一科技有限公司 Early warning method and system for equipment degradation
CN112181003A (en) * 2020-08-24 2021-01-05 中国神华能源股份有限公司国华电力分公司 Method for controlling unit pressure and electronic equipment
CN112199829B (en) * 2020-09-27 2022-06-24 浙江大学 Multi-working-condition multi-stage batch process monitoring method based on density peak clustering and instant learning
CN112327189B (en) * 2020-10-14 2023-06-09 北方工业大学 Comprehensive judging method for health state of energy storage battery based on KNN algorithm
CN113449412B (en) * 2021-05-24 2022-07-22 河南大学 Fault diagnosis method based on K-means clustering and comprehensive correlation

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1920511A (en) * 2006-08-01 2007-02-28 东北电力大学 Fusion diagnosing method of centrifugal pump vibration accidents and vibration signals sampling device
CN102750327A (en) * 2012-05-30 2012-10-24 合肥工业大学 K-means cluster diversified searching method on manifold surface and based on geodesic distance
CN104091035A (en) * 2014-07-30 2014-10-08 中国科学院空间应用工程与技术中心 Health monitoring method for effective loads of space station based on data-driven algorithm
CN104600680A (en) * 2015-02-16 2015-05-06 国家电网公司 Intelligent alarming method based on data fusion
CN106022366A (en) * 2016-07-04 2016-10-12 杭州电子科技大学 Rotary mechanical equipment fault diagnosis method based on neighbor evidence fusion
CN107341447A (en) * 2017-06-13 2017-11-10 华南理工大学 A kind of face verification mechanism based on depth convolutional neural networks and evidence k nearest neighbor
CN107436597A (en) * 2017-07-17 2017-12-05 华南理工大学 A kind of chemical process fault detection method based on sparse filtering and logistic regression
CN108388922A (en) * 2018-03-08 2018-08-10 云南电网有限责任公司电力科学研究院 A kind of recognition methods and system for the major-minor equipment state exception of fired power generating unit

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8285438B2 (en) * 2009-11-16 2012-10-09 Honeywell International Inc. Methods systems and apparatus for analyzing complex systems via prognostic reasoning

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1920511A (en) * 2006-08-01 2007-02-28 东北电力大学 Fusion diagnosing method of centrifugal pump vibration accidents and vibration signals sampling device
CN102750327A (en) * 2012-05-30 2012-10-24 合肥工业大学 K-means cluster diversified searching method on manifold surface and based on geodesic distance
CN104091035A (en) * 2014-07-30 2014-10-08 中国科学院空间应用工程与技术中心 Health monitoring method for effective loads of space station based on data-driven algorithm
CN104600680A (en) * 2015-02-16 2015-05-06 国家电网公司 Intelligent alarming method based on data fusion
CN106022366A (en) * 2016-07-04 2016-10-12 杭州电子科技大学 Rotary mechanical equipment fault diagnosis method based on neighbor evidence fusion
CN107341447A (en) * 2017-06-13 2017-11-10 华南理工大学 A kind of face verification mechanism based on depth convolutional neural networks and evidence k nearest neighbor
CN107436597A (en) * 2017-07-17 2017-12-05 华南理工大学 A kind of chemical process fault detection method based on sparse filtering and logistic regression
CN108388922A (en) * 2018-03-08 2018-08-10 云南电网有限责任公司电力科学研究院 A kind of recognition methods and system for the major-minor equipment state exception of fired power generating unit

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
基于K近邻证据融合的故障诊断方法;侯平智等;《控制与决策》(第10期);第1767-1774页 *

Also Published As

Publication number Publication date
CN111352408A (en) 2020-06-30

Similar Documents

Publication Publication Date Title
CN111352408B (en) Multi-working-condition process industrial process fault detection method based on evidence K nearest neighbor
CN112036089A (en) Coal mill fault early warning method based on DPC-MND and multivariate state estimation
CN111159844B (en) Abnormity detection method for exhaust temperature of gas turbine of power station
CN108062565A (en) Double pivots-dynamic kernel principal component analysis method for diagnosing faults based on chemical industry TE processes
CN117008479B (en) Carbon emission optimization control method and system based on biomass gasification furnace
CN109670549B (en) Data screening method and device for thermal power generating unit and computer equipment
CN111752147B (en) Multi-working-condition process monitoring method with continuous learning capability and improved PCA (principal component analysis)
CN107067100A (en) Wind power anomalous data identification method and device for identifying
CN115423158A (en) Predictive analysis method and system for data trend of hydroelectric generating set
JP2019096227A (en) Operation improvement effect calculation apparatus, operation improvement effect calculation method, operation improvement effect calculation program, and recording medium
CN118194026B (en) Gas power generation data analysis system
CN118265275B (en) Air cooling line control cooling optimization control method
CN114255784A (en) Substation equipment fault diagnosis method based on voiceprint recognition and related device
CN114757269A (en) Complex process refined fault detection method based on local subspace-neighborhood preserving embedding
CN112434739A (en) Chemical process fault diagnosis method of support vector machine based on multi-core learning
CN115795389A (en) Method and system for monitoring operation condition of transformer cooling device
CN111859809A (en) Fuzzy theory-based gas turbine system fault mode and influence analysis method
CN112541296A (en) SO2 prediction method based on PSO-LSSVM
You et al. Signal anomaly identification strategy based on Bayesian inference for nuclear power machinery
CN118016202A (en) Chemical equipment operation analysis method and system based on steam-water quality
JP6830414B2 (en) Diagnostic device and diagnostic method
CN113359665A (en) Industrial process fault detection method and system based on weighted key pivot
CN116305733B (en) Quality-related fault detection method based on global and local feature extraction
CN116224887A (en) Multi-station intermittent process fault monitoring method
JP2010237892A (en) Method and apparatus for classifying data

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